Foundations Of Pragmatics.pdf

Foundations of Pragmatics HoPs 1

Handbooks of Pragmatics

Editors

Wolfram Bublitz Andreas H. Jucker Klaus P. Schneider Volume 1

De Gruyter Mouton

Foundations of Pragmatics

Edited by

Wolfram Bublitz Neal R. Norrick

De Gruyter Mouton

ISBN 978-3-11-021425-3 e-ISBN 978-3-11-021426-0 Library of Congress Cataloging-in-Publication Data Foundations of pragmatics / edited by Wolfram Bublitz, Neal R. Norrick. p. cm. ⫺ (Handbook of pragmatics; 1) Includes bibliographical references and index. ISBN 978-3-11-021425-3 (alk. paper) 1. Pragmatics. I. Bublitz, Wolfram. II. Norrick, Neal R. P99.4.P72F68 2011 4011.45⫺dc22 2011013980

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de. ” 2011 Walter de Gruyter GmbH & Co. KG, Berlin/Boston Cover image: Vetta Collection/iStockphoto Typesetting: Dörlemann Satz GmbH & Co. KG, Lemförde Printing: Hubert & Co. GmbH & Co. KG, Göttingen ⬁ Printed on acid-free paper Printed in Germany www.degruyter.com

Preface to the handbook series Wolfram Bublitz, Andreas H. Jucker and Klaus P. Schneider The series Handbooks of Pragmatics, which comprises nine self-contained volumes, provides a comprehensive overview of the entire field of pragmatics. It is meant to reflect the substantial and wide-ranging significance of pragmatics as a genuinely multi- and transdisciplinary field for nearly all areas of language description, and also to account for its remarkable and continuously rising popularity in linguistics and adjoining disciplines. All nine handbooks share the same wide understanding of pragmatics as the scientific study of all aspects of linguistic behaviour. Its purview includes patterns of linguistic actions, language functions, types of inferences, principles of communication, frames of knowledge, attitude and belief, as well as organisational principles of text and discourse. Pragmatics deals with meaning-in-context, which for analytical purposes can be viewed from different perspectives (that of the speaker, the recipient, the analyst, etc.). It bridges the gap between the system side of language and the use side, and relates both of them at the same time. Unlike syntax, semantics, sociolinguistics and other linguistic disciplines, pragmatics is defined by its point of view more than by its objects of investigation. The former precedes (actually creates) the latter. Researchers in pragmatics work in all areas of linguistics (and beyond), but from a distinctive perspective that makes their work pragmatic and leads to new findings and to reinterpretations of old findings. The focal point of pragmatics (from the Greek prãgma ,act’) is linguistic action (and inter-action): it is the hub around which all accounts in these handbooks revolve. Despite its roots in philosophy, classical rhetorical tradition and stylistics, pragmatics is a relatively recent discipline within linguistics. C.S. Peirce and C. Morris introduced pragmatics into semiotics early in the twentieth century. But it was not until the late 1960s and early 1970s that linguists took note of the term and began referring to performance phenomena and, subsequently, to ideas developed and advanced by Wittgenstein, Ryle, Austin and other ordinary language philosophers. Since the ensuing pragmatic turn, pragmatics has developed more rapidly and diversely than any other linguistic discipline. The series is characertised by two general objectives. Firstly, it sets out to reflect the field by presenting in-depth articles covering the central and multifarious theories and methodological approaches as well as core concepts and topics characteristic of pragmatics as the analysis of language use in social contexts. All articles are both state of the art reviews and critical evaluations of their topic in the light of recent developments. Secondly, while we accept its extraordinary complexity and diversity (which we consider a decided asset), we suggest a definite structure, which gives coherence to the entire field of pragmatics and provides

vi

Wolfram Bublitz, Andreas H. Jucker and Klaus P. Schneider

orientation to the user of these handbooks. The series specifically pursues the following aims: – it operates with a wide conception of pragmatics, dealing with approaches that are traditional and contemporary, linguistic and philosophical, social and cultural, text- and context-based, as well as diachronic and synchronic; – it views pragmatics from both theoretical and applied perspectives; – it reflects the state of the art in a comprehensive and coherent way, providing a systematic overview of past, present and possible future developments; – it describes theoretical paradigms, methodological accounts and a large number and variety of topical areas comprehensively yet concisely; – it is organised in a principled fashion reflecting our understanding of the structure of the field, with entries appearing in conceptually related groups; – it serves as a comprehensive, reliable, authoritative guide to the central issues in pragmatics; – it is internationally oriented, meeting the needs of the international pragmatic community; – it is interdisciplinary, including pragmatically relevant entries from adjacent fields such as philosophy, anthropology and sociology, neuroscience and psychology, semantics, grammar and discourse analysis; – it provides reliable orientational overviews useful both to students and more advanced scholars and teachers. The nine volumes are arranged according to the following principles. The first three volumes are dedicated to the foundations of pragmatics with a focus on micro and macro units: Foundations must be at the beginning (volume 1), followed by the core concepts in pragmatics, speech actions (micro level in volume 2) and discourse (macro level in volume 3). The following three volumes provide cognitive (volume 4), societal (volume 5) and interactional (volume 6) perspectives. The remaining three volumes discuss variability from a cultural and contrastive (volume 7), a diachronic (volume 8) and a medial perspective (volume 9): 1. Foundations of pragmatics Wolfram Bublitz and Neal R. Norrick 2. Pragmatics of speech actions Marina Sbisà and Ken Turner 3. Pragmatics of discourse Klaus P. Schneider and Anne Barron 4. Cognitive pragmatics Hans-Jörg Schmid 5. Pragmatics of society Gisle Andersen and Karin Aijmer

Preface to the handbook series

6. Interpersonal pragmatics Miriam A. Locher and Sage L. Graham 7. Pragmatics across languages and cultures Anna Trosborg 8. Historical pragmatics Andreas H. Jucker and Irma Taavitsainen 9. Pragmatics of computer-mediated communication Susan Herring, Dieter Stein and Tuija Virtanen

vii

Acknowledgements

This initial, foundational volume in this series of handbooks represents the work of a large number of individuals over a fairly long period of time. We must, of course, thank our contributors for producing high quality articles reflecting scholarly rigor and a strong sense of what our readers seek in a handbook of this scope. Without conscientious, willing reviewers, projects like this are simply impossible. Our sincere thanks go out to all our reviewers, though we cannot identify them by name. We editors are surrounded by competent, trustworthy staff in Augsburg and Saarbrücken, Germany. We own a debt of gratitude to Claudia Enzweiler for proofreading and Sylvia Monzon for her clerical services, to Katharina Rüters, Katrin Stuis and Claudia Rieger for their help with the desk editing and compilation of the indexes. Finally, we are grateful to the De Gruyter team: Barbara Karlson, Wolfgang Konwitschny and Anke Beck.

Table of contents

Preface to the handbook series . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v ix

Introduction: the burgeoning field of pragmatics Wolfram Bublitz and Neal R. Norrick . . . . . . . . . . . . . . . . . . . . .

1

Part I – Conceptual foundations 1.

2.

3.

4.

Pragmatics as a linguistic concept Anita Fetzer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

23

Micropragmatics and macropragmatics Piotr Cap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

51

Pragmalinguistics and sociopragmatics Sophia Marmaridou . . . . . . . . . . . . . . . . . . . . . . . . . .

77

Metapragmatics Axel Hübler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

Part II – Theoretical foundations 5.

The rise of pragmatics: a historiographic overview Wataru Koyama . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

6.

Semiotic foundations of pragmatics Winfried Nöth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

7.

Pragmatics in modern philosophy of language Nikola Kompa and Georg Meggle . . . . . . . . . . . . . . . . . . . 203

8.

Foundations of pragmatics in functional linguistics Saskia Daalder and Andreas Musolff . . . . . . . . . . . . . . . . . . 229

9.

Foundations: ethnomethodology and Erving Goffman Christine Domke and Werner Holly . . . . . . . . . . . . . . . . . . 261

xii

Table of contents

10.

Pragmatics in Habermas’ Critical Social Theory Maeve Cooke . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

289

Part III – Key topics in pragmatic description 11.

12.

13.

14.

Deixis and indexicality William F. Hanks . . . . . . . . . . . . . . . . . . . . . . . . . . . .

315

Reference and anaphora Monika Schwarz-Friesel and Manfred Consten . . . . . . . . . . . .

347

Speech acts Elena Collavin . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

373

Types of inference: entailment, presupposition, and implicature Yan Huang . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

397

Part IV – The place of pragmatics in the description of discourse 15.

16.

17.

18.

Pragmatics and grammar Arnulf Deppermann . . . . . . . . . . . . . . . . . . . . . . . . . .

425

Pragmatics and semantics John Saeed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

461

Pragmatics and prosody: prosody as social action Elizabeth Couper-Kuhlen . . . . . . . . . . . . . . . . . . . . . . .

491

Pragmatics and literature Jacob L. Mey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

511

Part V – Methods and tools 19.

20.

Approaching the data of pragmatics Monika Bednarek . . . . . . . . . . . . . . . . . . . . . . . . . . .

537

Experimental pragmatics Richard Breheny . . . . . . . . . . . . . . . . . . . . . . . . . . . .

561

Table of contents

xiii

21.

Corpus-based pragmatics I: qualitative studies Gisle Andersen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 587

22.

Corpus-based pragmatics II: quantitative studies Christoph Rühlemann . . . . . . . . . . . . . . . . . . . . . . . . . . 629

23.

The transcription of face-to-face interaction Roger J. Kreuz and Monica A. Riordan . . . . . . . . . . . . . . . . 657

About the authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 681 Name index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 691 Subject index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 706

xiv

Table of contents

Introduction: the burgeoning field of pragmatics Wolfram Bublitz and Neal R. Norrick 1.

The aim of the handbook

The present volume, as the initial handbook in the nine volume series Handbooks of Pragmatics, provides a comprehensive overview of the foundations of pragmatics. Its in-depth articles cover the roots and evolution of those central theories and approaches as well as key concepts and topics that are characteristic of twenty-first century pragmatics as an approach to the means and ways of using language in authentic social contexts. The articles provide reliable orientational overviews useful to researchers, students, and teachers. They offer both state-of-the-art reviews of their topics and critical evaluations in the light of on-going developments. Thus, topics are considered not only within their contemporary scholarly context but are also critically evaluated from one or more current perspectives. As the opening volume in the series, Foundations of Pragmatics provides historical, conceptual, theoretical and methodological vantage points from which the articles in the following eight volumes can be related to each other and to the development of the entire field. The series as a whole seeks to be reasonably comprehensive to account for the exceptionally vast, unusually heterogeneous and still rapidly expanding field of pragmatics. At the same time, it sets out to give structure and coherence to the field through its organization and choice of topics as well as by revealing traits and contours that have evolved during the past half-century. The nine volumes are meant to reflect the substantial and wide-ranging significance of this transdisciplinary subject for nearly all areas of language description, to trace its origins and influences, as well as to account for its remarkable and continuously rising popularity in linguistics and adjoining disciplines. 2.

The history of pragmatics

Even though its roots can be traced back to early classical traditions of rhetoric and stylistics, to Immanuel Kant’s conception of pragmatics as empirical and purposive and to William James, who pointed out its practical nature, modern pragmatics is a fairly recent discipline. Its inauguration as an independent field of study within semiotics took place early in the 20th century by C. Morris, R. Carnap and ultimately C.S. Peirce. The classic division between syntax, semantics, and pragmatics goes back to Morris, who distinguished three separate “dimensions of semiosis” within his science of signs:

2

Wolfram Bublitz and Neal R. Norrick One may study the relations of signs to the objects to which the signs are applicable. This relation will be called the semantical dimension of semiosis, symbolized by the sign ‘DSEM’; the study of this dimension will be called semantics. Or the subject of study may be the relation of signs to interpreters. This relation will be called the pragmatical dimension of semiosis, symbolized by the sign ‘DP’; the study of this dimension will be named pragmatics. One important relation of signs has not yet been introduced: the formal relations of signs to one another. […] This third dimension will be called the syntactical dimension of semiosis, symbolized by the sign ‘DSYN’, and the study of this dimension will be named syntactics. (Morris 1938: 21–22)

Syntax studies the relations signs bear to other signs, semantics the relation between signs and objects, and pragmatics the relation between signs and their interpreters. Of course, there were and are differences of opinion on where exactly to draw the line between semantics and pragmatics, as many of the chapters in this volume will show. Some thirty years elapsed before pragmatics finally made its way into modern linguistics in the late 1960s, when linguists began to explore so-called performance phenomena. To this end, they adopted ideas developed and advanced by L. Wittgenstein, G. Ryle, P. Strawson, J.L. Austin and other eminent (ordinary or natural) language philosophers. It seems safe to claim that the ensuing pragmatic turn was most notably induced by J.L. Austin, J.R. Searle and H.P. Grice, who were interested in utterance meaning rather than sentence or word meaning, i.e. in studying unique historical events created by actual speakers to perform linguistic acts in actual situational contexts in order to accomplish specific goals. Other scientific movements that nourished pragmatics include anthropology (B. Malinowski, P. Wegener, A. Gardiner), contextualism (J.R. Firth) and functionalism (K. Bühler, R. Jakobson, D. Hymes), ethnomethodology (H. Garfinkel, E. Goffman, H. Sacks) and European sociology (J. Habermas). This volume addresses all of these influences. Since the pragmatic turn, pragmatics has arguably developed more rapidly and diversely than any other linguistic discipline. Since the 1970s, the early AngloAmerican framework of pragmatic linguistic study has been immensely expanded and enhanced by research in Continental Europe and elsewhere. With historiographic hindsight, it can be seen that the broadening, i.e. the interdisciplinary expansion, of the field of pragmatics has been a cumulative process; the broader conception of pragmatics chronologically (and causally) followed the narrower one. We can easily detect a first shift from the binarity of early structuralist concepts (such as lexical/word meaning versus causal/sentence meaning) to the multiplicity of speech act related concepts. A further step towards a (conceptual as well as methodological) widening of the field took place in the 1970s and 1980s, when structure and action oriented pragmatics linked up with the emerging interactional paradigms in sociology (in general) and ethnomethodology (in particular). The ensuing developments have seen inter alia a realignment of context (from a static and au-

Introduction: the burgeoning field of pragmatics

3

tonomous to a dynamic and collaborative concept, which is cognitively, situationally and socio-culturally much more refined), and a shift from the concept of the unilateral, i.e. the individual speaker’s (speech) act (as advocated by pioneers like Austin) to the interactionally expanded concept of the bi- or (in some types of interactive computer-mediated forms of communication) multi-lateral ‘inter-act’. As such, contemporary ‘broad’ pragmatics takes account of the ‘interactional turn’ that can be observed in the most recent development of interactive (Web 2.0-based) media formats. Taken all in all, the establishment and development of linguistic pragmatics has been an authentic success story.

3.

The definition of pragmatics

Despite its scientific acclaim, the notion of pragmatics remains somewhat enigmatic and is still difficult to define. This holds for its readings in everyday discourse as well as in scholarly contexts. Nonetheless, when we refer to attitudes and modes of behavior as pragmatic, we mean that they have a factual kind of orientation in common. People who act pragmatically or take a pragmatic perspective generally have a preference for a practical, matter of fact and realistic rather than a theoretical, speculative and idealistic way of approaching imminent problems and handling everyday affairs. To put it differently, they share a concrete, situation-dependent approach geared to action and usage rather than an abstract, situation-independent and system-related point of view. To assume a pragmatic stance in everyday social encounters as well as in political, historical and related kinds of discourse, means to handle the related affairs in a goal-directed and object-directed, common-sense and down to earth kind of way. Such an understanding of pragmatics as an attitude in non-scientific discourse has obviously left its traces in scientific definitions of the term. By and large we can say that in semiotics and philosophy, pragmatic characterizes those theoretical and methodological approaches that are oriented toward use and context rather than toward some system, and that they regard use and context as creating a high degree of analytical surplus. While essentially the same is true for linguistics in general, there is no commonly accepted definition of pragmatics in linguistics which would refer to a single, unified and homogeneous field of study. In contemporary linguistics, we can identify a narrow and a broad way of delineating pragmatics (of which the former is sometimes allocated to an “Anglo-American” and the latter to a “Continental [European]” tradition of pragmatics, cf. Huang 2007: xi). According to the narrow view, pragmatics is understood as the systematic investigation of what and how people mean when they use language as a vehicle of action in a particular context and with a particular goal in mind. Thus, the context-dependency of utterance meaning is the central component of more narrowly defined accounts of prag-

4

Wolfram Bublitz and Neal R. Norrick

matics, which focus on a few key issues that can be juxtaposed with related issues in other modules of language theory such as grammar and semantics. Those issues include indexicality/deixis (versus anaphora), presuppositions, implicatures (versus entailments) and speech acts (versus types of sentences), to name only the most conspicuous topics. In this volume, as indeed in the whole Handbooks of Pragmatics series, we adopt a much broader point of view and understand pragmatics as the scientific study of all aspects of linguistic behavior. In particular, pragmatics includes patterns of linguistic actions, language functions, types of inferences, principles of communication, frames of knowledge, attitude and belief, as well as organisational principles of text and discourse. Pragmatics deals with meaning-in-context, which for analytical purposes can be viewed from different perspectives (that of the speaker, the recipient, the analyst, etc.). It bridges the gap between the system side of language and the use side, and relates both of them at the same time. Unlike syntax, semantics, sociolinguistics and other linguistic disciplines, pragmatics is defined by its point of view more than by its objects of investigation. The former precedes (actually creates) the latter. Researchers in pragmatics work in all areas of linguistics (and beyond), but from a distinctive perspective that makes their work pragmatic and leads to new findings and to reinterpretations of old findings. The focal point of pragmatics (from the Greek prãgma ‘act’) is linguistic action (and inter-action); it is the hub around which all accounts in these handbooks revolve. (Preface to the handbook series)

Pragmatics is fundamentally concerned with communicative action in any kind of context. The multifaceted research paradigm of pragmatics has provided new directions and perspectives in the arts and humanities, philosophy, cognitive science, computer science and the social sciences. Pragmatic perspectives have been employed in information technology and in the social sciences, particularly in economics, politics and education. In the pragmatic perspective, language use and language users in interaction are primary, as opposed to language as a system of signs or a set of rules. The pragmatic perspective scrutinizes neither just individual words nor sentences nor even isolated texts, but rather whole speech events or language games in real social contexts, considering both the present state of affairs and its connectedness with prior and succeeding actions. It rejects a localization of language in a limited segment of the acts of speaking, understanding and responding or within the conscious of the individual. It supplants a view of language as an abstraction without variation by speaker, region or time, of language as a non-cultural, non-social, static, depersonalized fact independent of context and discourse. Pragmatics goes beyond the perspective of written texts with their carefully marshalled grammatical sentences to embrace everyday talk and the “messiness” of language in real embodied human contexts, where participants with personalities, feelings and goals interact in complex ways with physical objects and other participants within institutions and communities.

Introduction: the burgeoning field of pragmatics

5

That pragmatics overcomes any narrow focus on language as a system of signs or a set of rules does not mean, of course, that pragmatics (in contradistinction to other scientific disciplines) attempts to describe language without recourse to the systematic level of theoretical abstraction. Ultimately and in order to create and secure a sound scientific foundation, pragmatics, like any other scientific theory, has to accomplish the transfer from the level of token to the level of type. For the pragmatic scholar, specific concrete linguistic events (and their contexts) are only relevant in as far as their properties and constraints can be integrated under analogous conditions into a general concept of language and communication. While pragmatics is on a par with other scientific theories in this respect, it transcends them in various ways, in particular by adopting a constructivist point of view which allows for the inclusion of new contextual, situational and cognitive variables. Describing language pragmatically thus means going beyond the description of language as an autonomous, type-restricted principle and taking into account extra-linguistic phenomena and conditions emanating from the context and concrete situation of language use. Turning contextual and related conditions into prototypical conditions (and thus theorizing them) constitutes the pragmatic surplus.

4.

The purview of pragmatics

If pragmatics is more a perspective or orientation toward language use than a particular theory or specific research object, then the data practitioners choose to analyze assume great importance. In linguistic research, as Saussure realized, it is the viewpoint that determines the object rather than the other way around. In the tradition growing out of Natural Language Philosophy the intuitions and casual observations of the individual writer provide the stuff for analysis. Introspection counts as data. All the early work on speech acts, presupposition and inference was based on introspection and invented examples in imagined contexts (Austin, Searle, Strawson and Grice). Other traditional areas of pragmatic research were based on authentic data from the start. Indexicality/deixis, discourse markers, particles and the like are ubiquitous in texts, and examples for study are thus easy to collect. As corpora of spoken language became increasingly available and reliable, the problem of evidence versus introspection began to take care of itself. Scholars no longer needed to guess about distributions and frequencies of linguistic items. One might begin with a small pre-selected corpus representing a specific set of items or types of interaction to identify a range of functions for a particular linguistic feature before embarking on a general search in a large corpus to determine distributions and frequencies in the corpus at large only to return to a small corpus representing specific contexts for careful qualitative analysis. By contrast, investigations of large corpora might provide the impetus for research, illustrating a range of items or contexts not found

6

Wolfram Bublitz and Neal R. Norrick

in a small corpus and suggesting hypotheses to be tested. For statistical purposes, larger was presumably always better, because scholars tended to correlate frequency with typicality (the more frequent the more typical). On the other hand, increasing interest in multi-modal data and description in research on spoken interaction brought researchers back to small, carefully annotated corpora. By its very nature, research on prosody, gaze, gesture and the details of interaction with objects in a specific physical setting must proceed from a narrow recording and transcript of a particular event: Large corpora are extremely difficult to manage in these circumstances. Many pragmaticians maintain that language data must be as authentic as possible, recorded in real life contexts where something is at stake for the participants. But the whole matter of recording, often with camcorders as well as microphones, raises the spectre of the Observer’s Paradox. How natural is the language interaction, when we require technical apparatus to record it? Some pragmatic scholars ascribe to the notion that language use must be understood in terms of embedded practices, as part of the social world in which listening and speaking are modalities of action, in line with the treatment of speech production and reception as embodied social facts (Hanks 1996), and practices as shared ways of doing things, ways of talking, beliefs and values as a function of engagement in an activity (Wenger 1998; Eckert 2000). Conceptualizing speaking and listening in terms of practices affords a natural purchase on nonverbal features of participant behavior, including gaze, posture and gesture in face-to-face interaction (Goodwin 2003). This has consequences for what counts as data, how much context must be considered, what must be transcribed and in how much detail. At the same time, certain questions will always go unanswered on the basis of corpus data, no matter how careful the transcription and no matter how large the corpus, since every individual exchange is participant designed and context dependent. With some kinds of structures like prosody, discourse markers and interjections, there is the problem of how to get enough examples to make sensible generalizations. Experimental procedures could complement corpus-based research in order to generate parallel structures and contexts to test for recurrent speaker and listener practices. Indeed, Experimental Pragmatics represents a new area of research at the intersection of traditional linguistic pragmatics and psycholinguistics. It seeks to apply experimental techniques from psycholinguistics to solve problems defined in linguistic pragmatics (see Breheny this volume). So far such experiments have focused on issues surrounding the conversational maxims and implicatures of Grice (1975) and his followers (both Neo-Griceans like Horn 1984 and Levinson 2000, and Post-Griceans like Sperber and Wilson 1986 and Carston 2002), but creative experiment design could extend this new research paradigm to all sorts of questions about spoken language. All the same, investigations of large corpora provide a body of natural data against which experimental results can be evaluated.

Introduction: the burgeoning field of pragmatics

7

The discussions between scholars of different persuasions about what constitutes appropriate data are ongoing. The historical development of different data sources and areas of inquiry as well as the continuing debates about their usability and validity resonate in this handbook. Indeed, the authors contributing to this handbook were chosen to represent various positions in these debates.

5.

Organization of this Handbook

To distinguish pragmatics as a linguistic discipline from pragmatics as an attitude or a mode of behavior (a practical rather than an idealistic approach to problems and dealings) as used in everyday, in political, in economic and in other kinds of discourse, the present handbook Foundations of Pragmatics opens with a section on Conceptual Foundations consisting of the description of pragmatics as a linguistic concept. Fundamental concepts covered are: micropragmatics, i.e. the pragmatics of utterance based concepts such as speech acts etc. (which are covered in more detail in vol. 2 of the series) as opposed to macropragmatics, i.e. the pragmatics of discourse or text based concepts such as topics, discourse markers etc. (cf. a more detailed overview in vol. 3 of the series), pragmalinguistics (dealing with forms) as opposed to sociopragmatics (dealing with pragmatic strategies), and metapragmatics. The second major section is dedicated to the Theoretical Foundations of pragmatics. It begins with a historiographic overview of how pragmatics was established as a linguistic field of study, how it developed and spread, and how it eventually became an international success story. After all, since the so-called pragmatic turn, linguistic pragmatics has, arguably, developed more rapidly and diversely than any other linguistic field. The scientific disciplines that nourished pragmatics in the first place range from semiotics (Peirce, Morris, Carnap) and (Natural Language) philosophy (Wittgenstein, Strawson, Austin, Searle, Grice) to anthropology (Malinowski, Wegener, Gardiner), contextualism (Firth) and functionalism (Bühler, Jakobson, Hymes), from ethnomethodology (Garfinkel, Goffman, Sacks) to European sociology (Habermas). The third major section, Key Topics in Pragmatic Description, investigates those central concepts of pragmatic description which were the first to make their way from grammar into pragmatics (indexicality and deixis, followed by reference and anaphora), on the one hand, and those notions that rapidly became the focal descriptive construct in mainstream pragmatics (speech acts and types of inference such as entailment, presupposition, implicature), on the other. In the fourth major section, the Place of Pragmatics in the Description of Discourse, the scope of pragmatics within linguistic description is delimited vis-à-vis grammar, semantics, prosody. A final chapter on pragmatics and literature underscores the contention that the social aspect of language use is as prominent in literary contexts as in any non-literary context.

8

Wolfram Bublitz and Neal R. Norrick

Methods and Tools constitutes the fifth and final major section. It contains an initial chapter comparing various methodological approaches to pragmatics such as intuition, introspection, ethnographic field work, elicitation and corpus analysis, before it then proceeds with articles on experimental pragmatics, corpus-based pragmatics and the transcription of face-to-face interaction. No handbook can include every individual perspective and specific topic even regarding the foundations of an area of study. Our decisions regarding which fields and issues to include and how they should be addressed reflect both our own views of pragmatics and the expertise and interests of our authors. As authors were recruited for the handbook project, joined us, and began to think and write about their areas, topics were redefined, expanded and narrowed, titles were adjusted, some boundaries were sharpened, others redrawn. Recent developments in pragmatics, the general availability of new kinds of multi-media data and analytic procedures along with dynamically evolving attitudes toward topics and research methods helped shape the contents of this volume. Inevitably, the exigencies of scheduling and the special interests of the contributing authors conspire to yield a final set of articles representing individual views at a specific point in time. We embrace this particular character of the snapshot offered by the articles in this handbook.

6.

Part I – Conceptual foundations

In the lead article, “Pragmatics as a linguistic concept”, Anita Fetzer seeks to limit the concept of linguistic pragmatics. The chapter compares pragmatics as a linguistic concept with pragmatics as used in everyday, in political, in economic and other kinds of discourse. It discusses what is special about linguistic pragmatics, covering General Pragmatics as well. ‘Pragmatics’ as a common-sense notion is functionally synonymous with ‘practical’ or ‘just right at that stage’, describing a language-bound and/or action-based phenomenon within a particular situation, as is reflected in its usage in everyday, in political, in economic and other kinds of discourse. In linguistics, semiotics or philosophy, pragmatics refers to a theoretical concept, comprising contextual, generally usage-related constraints and requirements, for instance presuppositions, indexical expressions, felicity conditions, common ground and background assumptions. These are necessary conditions for communication to be felicitous in context. Thus, the perspective is shifted from the analysis of the language system and its constitutive parts to that of its rule-governed instantiation and embeddedness in context. In General Pragmatics the focus of investigation is on universals of communicative action, such as directness versus indirectness. These are refined in Linguistic Pragmatics, where they undergo language- and culture-specific modification. That change in perspective has important consequences for methodology. To account for the parts-whole connectedness, a relational frame of reference is required which accommodates the relation

Introduction: the burgeoning field of pragmatics

9

between the language system and language use, and interlocutors and what they do with language in context, comprising linguistic context (or co-text), social context and cognitive context. In the second article, “Micropragmatics and macropragmatics”, Piotr Cap addresses the central distinction between micropragmatics and macropragmatics. This chapter distinguishes micropragmatics as referring to speech acts/actions etc. from macropragmatics as directed at the discourse/text level. Defining pragmatics by its general perspective, rather than by its specific object of investigation, entails looking at the analytic scope of pragmatics in terms of micro- and macro-level concepts. These have been traditionally (cf. Mey 1993) associated with the opposition between the analysis of speech act force at an utterance level and the analysis of global intentionality at the level of a discourse/text. Following this division, as well as its later refinements, the article, first, reviews the locutionary, illocutionary and perlocutionary constituents of force of an utterance, relating them to pragmatic concepts which characterize the process of encoding and decoding its message (deixis, presupposition, implicature etc.). Second, it demonstrates how individual utterances comprising different sets of constituents and markers of force, including syntactic markers of explicitness and inferable carriers of implicitness, can form sequences to shape the global intentionality of a discourse/text. Invoking such notions as speech event and macro speech act, the article shows which pragmatic concepts utilized in (“micro-”) analysis of individual utterances are essentially complementary, i.e. which markers of force denoted by these concepts can collectively generate complex – “macro” – illocutions responsible for accomplishing global discourse/text goals. The important distinction between “Pragmalinguistics and sociopragmatics” forms the topic of chapter 3 by Sophia Marmaridou. In Principles of Pragmatics (1983), Leech introduced a distinction between pragmalinguistics (roughly, what form to use to achieve an intended pragmatic effect) and sociopragmatics (roughly, when to use a particular pragmatic strategy). This chapter describes past and current research on this distinction in pragmatics. Within a broad definition of pragmatics as the scientific study of all aspects of linguistic behavior, the distinction between pragmalinguistics and sociopragmatics is intended to focus on two methodological approaches. Pragmalinguistics concerns the study of the particular resources that a given language provides for conveying pragmatic meaning (illocutionary and interpersonal), whereas sociopragmatics, as the “sociological interface of pragmatics” (Leech 1983: 10), relates pragmatic meaning to an assessment of participants’ social distance, the language community’s social rules and appropriateness norms, discourse practices, and accepted behaviors (Thomas 1995). The above methodological distinction to some extent relates to earlier work in Marxist pragmatics (Mey 1979; Prucha 1983), foreshadows work in Critical Discourse Analysis (Fairclough 1995), and intersects with studies in sociolinguistics. A critical assessment of the two concepts and the respective methodologies emerges from

10

Wolfram Bublitz and Neal R. Norrick

current research practices. Even though the borderline between the two concepts often seems fuzzy, their relevance to language learning, teaching and testing has been enhanced (McNamara and Roever 2006) and so has its significance in historical corpus linguistics (Archer and Culpeper 2003). In the fourth chapter, Axel Hübler addresses the subject of “Metapragmatics”. He focuses on metapragmatics concerning the actual management of discourse. Key tasks in metapragmatics are the investigation of that area of speakers’ competence which reflects the judgments of appropriateness of their own and other people’s communicative behavior and the exploration of the ‘know how’ regarding the control and planning of, as well as feedback on, the ongoing interaction. The human capacity for reflecting about language and language use/communication is a feature of language-related competence that no animal is endowed with. Part of this capacity extends to areas which linguists have become accustomed to referring to as metapragmatics. The term is fairly new, and its usage not yet fully regulated; it accommodates various systematic perspectives and pertains to a wide range of language-reflexive phenomena. The article provides a comprehensive survey along two lines: (1) what linguists (and experts from other disciplines such as philosophy or social-psychology) have observed about how people act and interact and what they express about their practice; and (2) what these experts have observed about their own descriptive, analytical and theorizing practice. Line (1) includes theories of pragmatics as prefiguring observation and resulting from it; and it takes account of what could be categorized as (implicit or explicit) meta-practice by members of a speech community (e.g., monitoring, remedying, evaluating by, e.g., quoting, topicalizing, metaphorizing). Line (2) comprises meta-theoretical considerations, which (critically) reflect upon and compare varying theoretical stances or elaborate on the difference to be drawn between a scientific view on pragmatics and folk theories.

7.

Part II – Theoretical foundations

The opening chapter 5 in this section, “The rise of pragmatics: a historiographic overview”, by Wataru Koyama traces the history of the still young field of linguistic pragmatics. This chapter stresses the success story as well as the newness of the field of pragmatics. Especially after the artificial impoverishment of linguistics due to structuralism and generative grammar, there was clearly a need for pragmatics as a field to amalgamate insights from rhetoric, anthropology, functionalism etc. When and by whom was linguistic pragmatics ‘invented’? Who were its (extra-linguistic) predecessors? What was the nature of the scientific and, particularly, the linguistic environment, in which pragmatics could arise (the 1950s; logical positivism etc.). Then the author considers further developments: how and why pragmatics developed, slowly at first, abruptly later in the 1980s (Levinson 1983,

Introduction: the burgeoning field of pragmatics

11

International Pragmatics Association, Journal of Pragmatics etc.). Finally, he looks at how today’s pragmatics differs from earlier stages of pragmatics. In the sixth article, “Semiotic foundations of pragmatics”, Winfried Nöth relates pragmatics to its foundations in semiotics. The chapter shows how semiotics influenced the development of the theoretical foundations of pragmatics, focusing on the contributions of Peirce, Morris and Carnap. Semiotics, the general study of signs and sign processes, is fundamental to pragmatics, the study of language use, speech acts, and communicative processes, and it extends the framework of language use to include nonverbal, visual, and other signs in the context of verbal behavior. According to the general framework outlined by Carnap and Morris, pragmatics is one of the three branches of semiotics besides syntax and semantics, but twentieth century structuralist semiotics tended to neglect the pragmatic dimension in its study of language and discourse. Central issues of pragmatics, such as the questions concerning intention and purpose, meaning and reference, word and object, interpretation and communication, or the relationship between language and action, find fundamental answers in the writings of the founder of general semiotics and philosophical pragmatism (or pragmaticism), Charles S. Peirce. Chapter 7, “Pragmatics in modern philosophy of language” by Nikola Kompa and Georg Meggle, traces the origins of linguistic pragmatics in the Philosophy of Language, particularly in Natural Language Philosophy in the work of Wittgenstein, Strawson, Austin, Searle, Grice. There are two paradigms of Pragmatics in modern studies in the Philosophy of Language. In the first, semantics is taken as given and the aim is to incorporate as many context-sensitive aspects of language as possible into that semantics; in the second, research is trying to spell out the semantics on a pragmatical (action-theoretical) basis itself. Accordingly, the difference between the two paradigms could be considered that between semantic pragmatics on the one side and pragmatic semantics on the other side. Characteristic contributions to the first approach are those by Frege, Carnap, Kaplan and Stalnaker; characteristic contributions to the second are by Wittgenstein, Austin, Grice, Lewis and Bennett. This article considers the basic problems as well as the present state of the art in both paradigms of pragmatics – and attempts to achieve some clarity about their deeper (dis-) connection in particular In chapter 8, “Foundations of pragmatics in functional linguistics”, Saskia Daalder and Andreas Musolff trace the roots of pragmatics in functional approaches to linguistics. Pragmatics addresses the functions of language as found in the work of Bühler, Jakobson, Hymes. This chapter traces the historical connections from these sources to contemporary thinking in pragmatics. The study of linguistic pragmatics received a major impetus from the systematic analysis of “language functions” as started by Karl Bühler, who distinguished three fundamental dimensions of the linguistic sign: it functions at the same time as symbol, symptom and signal. The model has proved its enduring value in the development of functionalist language theories, via its modification and augmentation into a six-func-

12

Wolfram Bublitz and Neal R. Norrick

tions-model by Roman Jakobson. His model in turn informed subsequent functionally oriented theories of language and communication, e.g. Dell Hymes’ Ethnography of Speaking (1962) and Michael A.K. Halliday’s concept of Language as Social Semiotic (1978). In retrospect, Bühler’s model (and, to a lesser extent, Jakobson’s modified model) can be judged to suffer from its forced integration with the structuralist concept of the linguistic sign, which initially hampered a fully fledged pragmatic interpretation. In their chapter 9, “Foundations: Ethnomethodology and Erving Goffman”, Christine Domke and Werner Holly discuss two related questions: What is Ethnomethodology – or rather what are the crucial questions and assumptions of this qualitative research approach? And how did Ethnomethodology impact linguistic pragmatics? At first they overview the studies of Harold Garfinkel which contain the groundbreaking thoughts about the most significant question: How is social reality constructed and organized by activities of the society’s members? Accordingly, it is shown how conversation analysts (e.g. Sacks, Schegloff) are applying Garfinkel’s fundamental assumption of the social reality accomplished through people’s everyday life activities in analyzing talk-in-interaction. Subsequently, the main focus lies on the interface between social order, which Ethnomethodology considers as produced by people interacting in their daily life, and the fundamental pragmatic question what people are doing using speech and talking. How the idea of “(re-)produced social order” is in line with and can be linked to the pragmatic focus on the role of language/verbal interaction in different forms of daily interaction. The work of Habermas has been central in the development of pragmatics in Europe. In her contribution in chapter 10, “Pragmatics in Habermas’ Critical Social Theory”, Maeve Cooke considers the influence of Habermas on linguistic pragmatics. She deals with the pragmatic turn that Jürgen Habermas introduced to critical social theory with his program of universal pragmatics in the 1970s. With this program, subsequently referred to as “formal pragmatics”, Habermas investigated everyday linguistic interaction from the point of view of its orientation towards validity. His investigations were guided by the critical intentions of his social theory. His principal intentions were to provide a normative basis for a conception of practical rationality without reference to “otherworldly” authorities or entities and to develop a corresponding normative account of social action. However, his formal pragmatics has profound implications for social theory in general and, indeed, for anyone concerned with questions of meaning and validity. This article sets out the key elements of Habermas’ formal pragmatics, shows its implications for social theory and beyond, and discusses some important objections that can be raised against him.

Introduction: the burgeoning field of pragmatics

8.

13

Part III – Key topics in pragmatic description

Pragmatics first made inroads into linguistics through the pioneering work of Yehoshua Bar-Hillel. In this chapter 11, William Hanks describes past and current research on “Deixis and indexicality” from a pragmatic perspective. He starts from a historical sketch of indexicality in the semiotics of Peirce and Morris, and traces the development of the concept by philosophers, linguists, linguistic anthropologists, sociolinguists (via style, register, “contextualization”) and conversation analysts. Called “shifters” by Jespersen and Jakobson, deictics encode a relational structure projecting from an indexical ground (a context of utterance) to a referential figure (the object denoted) via a relation (proximity, perceptibility, anaphora, etc.). Deictic systems differ cross-linguistically in ways just beginning to emerge. Pronouns are a related type of shifter and are closely tied into “participation frameworks” as treated by Goffman, conversation analysts and linguistic anthropologists. Recent debate is summarized regarding the role of spatial proximity vs. cognitive-social accessibility as a basis for deixis. The final section outlines empirical approaches to pragmatic research on deixis. Reference and anaphora are inherently pragmatic in nature and their treatment in linguistic description helped bring pragmatic notions into prominence. Chapter 12 on “Reference and anaphora” by Monika Schwarz-Friesel and Manfred Consten concerns reference, the area in scientific research which deals with the ancient problem of how language relates to the world. In the tradition of philosophical logics and formal semantics, reference has been defined as a static relationship between expressions and the things they denote. In the course of pragmatic change in linguistics, models of reference have arisen that (to different extents) take cognitive properties of language users into account. In the framework of the latter notion, anaphora is a specific kind of reference that has to be resolved mainly through the activation of discourse based knowledge established by the preceding text. Thus, anaphora resolution is analysed as a process of reactivating pre-mentioned referents. The traditional view of anaphora as a means of referential continuity and the availability of given information has to be revised. From a procedural perspective, anaphora may add new information to the mental files of the referents in the text-world model. Elena Collavin addresses the key topic of “Speech acts” in chapter 13. The chapter provides an overview of the notions of act/action (versus behavior) and speech act: how they developed, how they made their way into linguistics, how they became the core notions of various (speech act) theories, how the latter developed etc. Collavin presents an account of the notion of speech act, which includes an illustration of the main formulations of the theory, its limitations, and its current relevance and application in the broad field of pragmatic research. She traces the origins of this notion and describes the main formulations of Speech Act Theory, focusing on conventionality, intentionality, and action as the key aspects of speech

14

Wolfram Bublitz and Neal R. Norrick

acts around which similarities and incompatibilities in different versions of the theory are established. The discussion then turns to the criticisms moved to various notions of speech act. The relevance of speech acts for the empirical analysis of language use in interaction has been strongly debated. Collavin discusses how some notion of speech act can relate to conceptualizations nowadays fundamental in the study of language usage such as frame, voicing, indexicality, multimodality, shared cognition and embodiment. Chapter 14 provides an overview of the varieties of inference recognized in pragmatics, how they developed, how they made their way into linguistics, how they became core notions of various theories, how the latter developed etc. In this chapter, “Types of inference: entailment, presupposition, and implicature”, Yan Huang provides a critical survey of three fundamental types of inference recognized in semantics and pragmatics, namely entailment, presupposition and implicature, covering their origin, development, and the roles they play in various semantic/pragmatic theories in linguistics and the philosophy of language. Of the three categories of inference, entailment is essentially semantic in nature, though the notion of pragmatic entailment is discussed. Presupposition straddles the semanticspragmatics boundary, resulting in semantic and pragmatic presupposition. Finally, with regard to the two subtypes of implicature, while conversational implicature is pragmatic in nature, conventional implicature can be considered either as semantic or pragmatic, depending on how the distinction between semantics and pragmatics is drawn. Notions parallel to conversational implicature (such as explicature) and the argument that conversational implicature should not be treated as inference are also considered. Finally, comparisons between entailment and presupposition on the one hand, and between presupposition and implicature on the other, are presented.

9.

Part IV – The place of pragmatics in the description of discourse

Part IV opens with chapter 15 “Pragmatics and grammar” by Arnulf Deppermann. This chapter details the delineation of pragmatics and grammar and the integration of pragmatics and grammar into linguistic theory in relation to each other, focusing on issues, methods and areas covered. Deppermann discusses theoretical approaches, linguistic phenomena and empirical methods concerning the relationship between grammar and pragmatics. He sketches functional approaches to grammar, which claim that grammar is shaped and constrained by cognition, text and/or social interaction (functional grammar, systemic-functional linguistics, construction grammar, cognitive grammar, interactional linguistics). Historical studies on the paths of grammaticalization, psycholinguistic insights from usage-based approaches, and conversation analysis are introduced as methods for the study of the relationship between grammar and pragmatics. The relevance of grammar for

Introduction: the burgeoning field of pragmatics

15

pragmatics is discussed concerning a broad range of issues, such as cognitive processes (implicature, inference, information structure), the expression of cognitive states (epistemic stance), conversational organization (repair, turn-construction and turn-taking, collaborative action, intersubjective grounding), action structure (speech acts, genres), bodily action (gesture) and socio-structural concerns (politeness, power, social styles). In chapter 16 on “Pragmatics and semantics”, John Saeed discusses some contrasting views of the relationship between semantics and pragmatics. His chapter details the delineation of pragmatics and semantics in linguistic theory, focusing on issues, methods, areas covered. Within the philosophical and formal traditions, he considers the growth of contextualism, that is the increasing emphasis on the underspecification of meaning in linguistic form, illustrated by phenomena beyond traditional deixis (indexicals) such as gradable and comparative adjectives, quantifiers, possessives etc. The article also looks at the defense against this from semantic minimalism (e.g. Cappelen and Lepore 2005). As an example of contextualism, Saeed discusses aspects of Relevance Theory (Sperber and Wilson 1986, Carston 2002), with its identification of pragmatic principles reaching deep into what earlier accounts viewed as semantic content. In particular, he considers the theory’s notions of explicatures and higher level explicatures, and looks at its view of lexical pragmatics. Finally, he brings into the discussion the quite different view from cognitive linguistics, where the very distinction between semantics and pragmatics is questioned. So in theories like Fauconnier’s, mechanisms of mental spaces and conceptual blending seek to model phenomena from across the traditional divide, including e.g. tense, modality, presupposition and metaphor. An interesting question is whether a pragmatic theory like Relevance Theory and the cognitive semantics of writers like Lakoff, Fauconnier and Langacker are in fact compatible. Elisabeth Couper-Kuhlen describes the role of prosody in pragmatics, and the integration of prosody into linguistic theory. In chapter 17 “Pragmatics and prosody”, she contemplates what has been achieved so far in the field of prosody and interaction and what still remains to be done. Instead of terms such as suprasegmentals, sentence types and illocutions, she speaks of prosodic resources for single- and multi-turn construction, action formation and floor management. Comparing the use of prosodic resources across turns-at-talk, she acknowledges the contribution of prosody to marking the continuation of same-speaker turns, to organizing sequences of turns, to cueing problems, negotiating agreement and signaling affiliation in talk. But recent developments also present challenges to researchers in the field of prosody and interaction: (i) What counts as prosodic? and (ii) What is prosody’s place in the analysis of multimodal interaction? The chapter discusses these issues and ways of dealing with them in the future. In chapter 18, Jacob Mey discusses “Pragmatics and literature”. According to Mey (2000), pragmatics is concerned with the way humans use language in a social

16

Wolfram Bublitz and Neal R. Norrick

context: When people use language in a literary context, the social aspect is no less prominent. Literary pragmatics, therefore, concerns itself with the relation between the language users in the context of literature, that is, the ‘producers’ (the authors) and the ‘consumers’ (the readers). In the context of literary production, the processes of producing and consuming are paramount. The product (the written text) can only be evaluated properly in terms of its social conditions and the concomitant roles of (authorial) production and (readerly) consumption. Here, the key concept of ‘voice’ (originally due to Bakhtin) plays a major role. The author assumes a ‘narrative’ voice, while retaining a distinct personal voice. Conversely, the readers have a voice, inasmuch as they identify with, and co-construct, the narrative personae; in this way, they take part in the building of the literary universe and making it come alive. Some implications of this pragmatic perspective are followed up through analysis of snippets of literary texts.

10.

Part V – Methods and tools

The initial chapter 19 in this final part by Monika Bednarek, “Approaching the data of pragmatics”, provides an overview of how various aspects of language use are studied using different methodological approaches: intuition, introspection, ethnographic field work, elicitation (and experimentation), and corpus analysis. It compares critically the various methodologies that are employed in areas such as conversation analysis, politeness theory, and speech act theory as well as in related functional approaches. The chapter covers both quantitative and qualitative methodologies ranging from intuition and introspection to interviews, discourse completion tasks, ethnographic field work, elicited data and corpus analysis. It takes a broad approach to ‘the data of pragmatics’ in that it not only includes research in key pragmatic areas but also disciplines related to pragmatics such as ethnography and sociolinguistics. It describes different methodologies and compares the insights gained from them as well as the different perspectives they provide on linguistic data. In chapter 20, “Experimental pragmatics”, Richard Breheny surveys recent research in this emerging field. Experimental pragmatics aims to use the methods of experimental psychology to test claims found in the more theoretical linguistics literature. The strategy of testing theoretical ideas experimentally has been around for a while and is notably exemplified in the work of Clark, Sanford, Gibbs and Glucksberg. More recently, there has been an upturn in the interest in experimental methods, coinciding with renewed interest in pragmatic questions such as the nature of generalized implicatures, presupposition, metaphor and the like. Breheny reviews studies involving children, adolescents and adults from clinical populations as well as typically developing groups using the full range of methods to be found in psycholinguistics and developmental psychology (including both beha-

Introduction: the burgeoning field of pragmatics

17

vioral and neuro-scientific methodologies). The contribution also contains a discussion of the difficulty of drawing empirically testable predictions from claims made in the theoretical literature. The final three chapters in the volume all address the role and forms of corpus data in linguistic pragmatics. The first two consider corpus-based research in pragmatics, first qualitative then quantitative studies, and the third discusses methods of transcription. In chapter 21, “Corpus-based pragmatics I: qualitative studies”, Gisle Andersen describes the practice and potentials of corpus-based research in pragmatics, focusing on qualitative approaches to language data. He considers the development of spoken and written corpora like the Brown Corpus, the London-Lund Corpus, the Helsinki Corpus, the British National Corpus, the Santa Barbara Corpus of Spoken American English etc., and their growing influence on research in pragmatics over the last thirty years, paying attention to special corpora, large general corpora, continually growing monitor corpora and web-based corpora. Corpora allow for in-depth studies of patterns of frequency, distribution and collocation, thus facilitating statistically based observations of innovative language use and variation. Dynamic and ever-growing corpora allow for the observation of emerging repetitive patterns such as the routinization of forms with discourse marking functions, ongoing grammaticalization and the development of new abstract and interactional meanings from old forms. In contrast to quantitative, statistical approaches to corpus data, qualitative studies see corpora primarily as a source of natural data and a way of overcoming the fallible introspection of ‘armchair linguistics’. In the area of qualitative pragmatic studies, corpora have provided a constantly growing pool of data, allowing researchers to test their hypotheses about individual items and constructions, as well as practices and usage within particular groups, but also suggesting new avenues of research. Diachronically oriented corpora are responsible for the rapid progress made in historical pragmatics. At the same time, corpora representing specific groups, genres, registers or stages of a language have led to research of various kinds, including the language of adolescents, foreign language learners and other user groups. In the companion chapter 22, “Corpus-based pragmatics II: quantitative studies”, Christoph Rühlemann takes up quantitative corpus-based research. Due to the massive dependence of pragmatic phenomena on context, corpora, as a relatively decontextualized medium, have long been seen by some researchers as unfit for use in pragmatic research. Nonetheless, corpus linguistic analyses, both qualitative and quantitative in orientation, have produced a wealth of new insights into key pragmatic phenomena. The aim of this article is to illustrate key quantitative corpus studies into phenomena of pragmatic interest. The article is divided into four sections. Following a brief introduction that addresses general issues in corpus linguistic research on pragmatic phenomena, such as the question of how much or little context-sensitive corpora are, the second section introduces quanti-

18

Wolfram Bublitz and Neal R. Norrick

tative studies carried out on the small number of corpora that are pragmatically annotated. The third, and largest, section discusses key corpus linguistic research into pragmatic phenomena such as speech acts, deixis, evaluation, discourse marking, backchannelling and vagueness. Further, as a borderline case between pragmatics and semantics, research into semantic prosody will be reviewed. The concluding section looks to the future, outlining inter alia recent attempts at building multimodal corpora that will potentially shed light on the interplay between utterance and kinesics. The final contribution 23 deals with various methods of transcribing face-toface interaction and their significance for pragmatic theorizing. In this chapter, “The transcription of face-to-face interaction”, Roger J. Kreuz and Monica A. Riordan consider issues in the representation of linguistic data. Researchers in pragmatics frequently transcribe audio- or videotaped recordings of conversational participants. A comprehensive transcription coding system would help to unify diverse lines of research and make collaboration and interpretation of results easier. However, a comprehensive system currently does not exist. Kreuz and Riordan review coding schemes that have been previously proposed and assess their strengths and weaknesses. Most systems have focused solely on the acoustic (i.e., lexical and intonational) aspects of discourse between adult speakers, but they also review systems that have been developed to transcribe the speech of other populations (e.g., children, aphasics and cognitively impaired individuals). In addition, they make recommendations about how non-acoustic information can be coded. Specifically, the challenges and utility of coding for facial and gestural information are addressed. It is hoped that a more comprehensive coding system will make it easier for researchers to describe and study the multimodal aspects of conversational interaction. Taken together, these chapters constitute an outline of the fundamental issues and trends in the study of pragmatics at the beginning of the twenty-first century. They can be taken as a kind of extended definition of the term “pragmatics” as practiced today, along with critical attempts to set down the history of the discipline and identify the sorts of questions taken to be important and the sorts of answers being proposed. We hope in this way to contribute to the definition and coherence of the field, and thereby to foster research in the generation to come.

References Archer, Dawn and Jonathan Culpeper 2003 Sociopragmatic annotation: New directions and possibilities in historical corpus linguistics. In: Andrew Wilson, Paul Rayson, and Tony McEnery (eds.), Corpus Linguistics by the Lune: A Festschrift for Geoffrey Leech, 37–58. Frankfurt am Main: Peter Lang.

Introduction: the burgeoning field of pragmatics

19

Cappelen, Herman and Ernie Lepore 2005 Insensitive Semantics. Oxford: Blackwell. Carston, Robyn 2002 Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell. Eckert, Penelope 2000 Linguistic Variation as Social Practice. Malden, MA: Blackwell. Fairclough, Norman 1995 Critical Discourse Analysis. London: Longman. Goodwin, Charles 2003 Pointing as situated practice. In: Sotaro Kita (ed.), Pointing: Where Language, Culture, and Cognition Meet, 217–242. Mahwah, NJ: Erlbaum. Grice, Herbert Paul 1975 Logic and Conversation. In: Peter Cole and Jerry L. Morgan (eds.). Syntax and Semantics. Vol. 3: Speech Acts, 41–58. New York: Academic Press. Halliday, Michael A.K. 1978 Language as Social Semiotic. London: Arnold. Hanks, William 1996 Language and Communicative Practices. Boulder, CO: Westview Press. Horn, Lawrence R. 1984 Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature. In: Deborah Schiffrin (ed.), Meaning, Form, and Use in Context: Linguistic Applications, 11–42. Washington DC.: Georgetown University Press. Huang, Yan 2007 Pragmatics. Oxford: University Press. Hymes, Dell 1962 The ethnography of speaking. In: Thomas Gladwin and William C. Sturtevant (eds.), Anthropology and Human Behavior, 13–53. Washington: Anthropological Society. Leech, Geoffrey 1983 Principles of Pragmatics. London: Longman. Levinson, Stephen 1983 Pragmatics. Cambridge: University Press. Levinson, Stephen 2000 Presumptive Meanings. The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. McNamara, Tim and Carsten Roever 2006 The social dimension of proficiency: How testable is it? Language Learning 56/2: 43–79. Mey, Jacob L. 1979 Introduction. In: Jacob L. Mey (ed.), Pragmalinguistics. Theory and Practice, 10–17. The Hague: Mouton. Mey, Jacob L. 1993 Pragmatics. An Introduction. Oxford: Blackwell. Mey, Jacob L. 2000 When Voices Clash: A Study in Literary Pragmatics. Berlin: Mouton de Gruyter.

20

Wolfram Bublitz and Neal R. Norrick

Morris, Charles 1938 Foundations of the Theory of Signs. In: Otto Neurath, Rudolf Carnap and Charles Morris (eds.). International Encyclopaedia of Unified Science, 77–138. Chicago: University of Chicago Press. Prucha, Jan 1983 Pragmalinguistics: East European Approaches. Amsterdam: Benjamins. Sperber, Dan and Deirdre Wilson 1986. Relevance: Communication and Cognition. Oxford: Blackwell. Thomas, Jenny 1995 Meaning in Interaction: An Introduction to Pragmatics. London: Longman. Wenger, Etienne 1998 Communities of Practice: Learning, Meaning, and Identity. Cambridge: Cambridge University Press.

Part I Conceptual foundations

1.

Pragmatics as a linguistic concept1 Anita Fetzer

1.

Introduction

Pragmatics is fundamentally concerned with communicative action and its felicity in context, investigating action with respect to the questions of what action is, what may count as action, what action is composed of, what conditions need to be satisfied for action to be felicitous, and how action is related to context. These research questions and the object of research require action in general and communicative action in particular to be conceived of as relational concepts, relating action and context, relating action and communicative action, relating communicative action and interlocutors, and relating interlocutors with the things they do with words in context. The inherent parts-whole perspective requires the explicit accommodation of a theory of context, capturing the dynamic processes from the connectedness between parts and wholes thus transcending clearly delimited frames of investigation. The multifaceted research paradigm of pragmatics has provided new directions and perspectives in arts and humanities, philosophy, cognitive science, computer science and social science. Pragmatic perspectives have been employed in information technology and in the social sciences, particularly in economics, politics and education. In line with Wittgenstein’s conception of language games and Levinson’s conception of activity types (Levinson 1979; Wittgenstein 1958), game theory (Lewis 1979) has been adapted to their particular contextual constraints and requirements. The change of perspective from objects to interlocutors (or users), and the accommodation of their wants and needs is reflected in the paradigm-anchored collocations of pragmatic thinking and learning, pragmatic software, pragmatic design, pragmatic modelling and pragmatic technology on the one hand, and pragmatic foreign policy, pragmatic politicians, pragmatic sanction and pragmatic nationalism on the other. Hence, it is no longer solely individual actions and their perlocutionary effects, which are at the heart of analysis, but rather whole games, considering both the present state of affairs and its connectedness with prior and succeeding actions. In philosophical contexts, pragmatics is related closely to William James, who pointed out its practical nature, and to Charles Peirce, who referred back to Immanuel Kant’s conception of pragmatics as experimental, empirical and purposive thought (e.g., The New Encyclopaedia Britannica 1976; cf. also Koyama this volume). Pragmatics is also frequently connected with Greek ‘pra˘gma’, meaning ‘action’ and ‘instructive’ or ‘useful’ (e.g., Bublitz 2009; Cummings 2005).

24

Anita Fetzer

The connectedness between ‘action’ and ‘purpose’ is also reflected in its ordinary-language meaning ‘learning by doing’ (cf. Longman DCE 2003; Collins Cobuild 1990). As a non-scholarly, ordinary-language notion, pragmatics is functionally synonymous with ‘practical’ or ‘just right at that stage’, describing a surface-bound and action-based phenomenon within a particular situation. It occurs primarily in its adjectival form, foregrounding practical and unexpected, rather than theoretical and dogmatic considerations. For instance, politicians and economists may argue for a case on wholly pragmatic grounds, brushing aside traditionally established borders, which tend to result from presupposed and taken-forgranted ideology. Whenever an action, decision or solution is assigned the status of a pragmatic action, a pragmatic decision or a pragmatic solution, it is contrasted implicitly with its theory-based and abstract-principles-driven counterparts. Against this background, the adjective ‘pragmatic’ in its function as a modifier anchors the head of a noun phrase to the here and now of discourse. To employ Gricean terminology from his cooperative principle, it signals that the ongoing discourse is “such as is required, at the current stage at which it occurs, by the accepted purpose or direction of the talk exchange” (Grice 1975: 45). Pragmatics in its non-scholarly ordinary-language meaning thus meets with one of the most fundamental premises of scholarly pragmatics, the Gricean cooperative principle. There is a further bridging point between non-scholarly pragmatics and scholarly pragmatics: the differentiation between brute facts and institutional facts. According to Searle (1995), facts, that is brute facts, mental facts, social facts and institutional facts, constitute and regulate the world. To have a good knowledge of the present situation, interlocutors need to have a good knowledge of what regulates the world and what constitutes the world. Only then may they construct alternative visions of a present situation – or alternative ‘social realities’.

2.

Pragmatics in context

Pragmatics is frequently conceptualized as the science of language use, the study of context-dependent meaning and the study of speaker-intended meaning, presupposing the existence of language, language user and context on the one hand, and context-independent meaning on the other. To capture its multifaceted nature, definitions tend not to concentrate on the questions of what pragmatics is and what it does, but rather on what it is not and what it does not do. The majority of conceptualizations pay tribute to Charles Morris’s definition as “the study of the relation of signs to interpreters” (Morris 1971: 6) and to Austin’s differentiation between constative and performative (Austin 1971), foregrounding the performance of communicative action and the effects the uttering of words may have. Against this background, pragmatics is considered to be the study of invisible meaning (Yule 1996: 127), the science of the unsaid (Mey 2001: 194), the study of meaning as it

Pragmatics as a linguistic concept

25

“emerges in language use” (Marmaridou 2000: 1), and the study of linguistic acts and communicative action and their appropriateness (Bublitz 2009; van Dijk 2008; Fetzer 2004, 2007). There have been various attempts to systematize the multifaceted and heterogeneous field of pragmatics, or to bring some order into the pragmatic wastebasket (Bar-Hillel 1971), as is reflected in the differentiation between the broad fields of scholarly pragmatics: general (or non-linguistic) pragmatics and linguistic pragmatics. Implicit in the philosophy-anchored general-action and linguistic-actionbased paradigms is the premise that pragmatics is conceptualized best as a perspective, which comprises a general pragmatic perspective, a social perspective, a compositional perspective and a relational perspective. These are introduced below. 2.1.

The pragmatic perspective

The pragmatic-perspective paradigm does not represent a clearly delimited field of research but rather offers a general perspective towards the object to be examined. It provides a point of view for analysis, in which not the object in isolation is examined but rather the nature of its connectedness with users and other objects, or with the theoretical framework to which it is anchored. The perspective is thus shifted from an analysis of the language system and its constitutive parts to that of its rulegoverned instantiation and embeddedness in context, considering its generalized and particularized conditions of use. That change in perspective has important consequences for methodology. To account for the parts-whole connectedness, a relational frame of reference is adopted, which accommodates the relation between the language system, language use and language users, and what they do with words in context, comprising linguistic context (or co-text), social context and cognitive context. The pragmatic-perspective paradigm provides a general cognitive, social and cultural perspective on linguistic phenomena in relation to their usage in forms of behaviour, accounting for the dynamics of language and language use, as is reflected in the premises that meaning is not a product and given but rather dynamic, multifaceted and negotiated in context. Language use is dependent on variability, and language users adapt to contextual constraints accommodating them not only in the formulation of utterances but also in their interpretation (e.g. Mey 2001; Verschueren 1999). As in functional grammar and systemic functional grammar (Givón 1993; Halliday 1994), the dynamic construct of meaning is a necessary condition for making linguistic choices. Following Verschueren (1999: 55–58), the choices are made at every possible level of structure as regards form and strategy. In line with the sociolinguistic premise of linguistic variation, which claims that speaker-intended meaning can be formulated with different linguistic surfaces (Brown 1995), these choices are not equivalent in status. They carry along their alternatives and are always seen and interpreted against the background of other,

26

Anita Fetzer

more or less marked possibilities. While unmarked variants are seen as more frequent in distribution, more conventional in formulation and easier to process, marked variants are less frequent in distribution, less conventional in formulation and harder to process (Fetzer 2002, 2007). The pragmatic perspective is not only reflected in the more general approach examined above, considering linguistic communication as well as linguistic behaviour. It is also found in the particularized, ‘hyphenated’ fields of pragmatics, as in cognitive pragmatics (cf. vol. 4 Cognitive pragmatics of this series Handbooks of Pragmatics), empirical pragmatics (cf. Bednarek this volume, Andersen this volume, Rühlemann this volume), experimental pragmatics (cf. Breheny this volume), formal pragmatics, historical pragmatics (cf. vol. 8 Historical pragmatics of this series Handbooks of Pragmatics), intercultural pragmatics (cf. vol. 7 Pragmatics across languages and cultures of this series Handbooks of Pragmatics), interlanguage pragmatics, sociopragmatics (cf. vol. 5 Pragmatics of society of this series Handbooks of Pragmatics) and theoretical pragmatics. 2.2.

The social perspective

An analysis of the rule-governed and strategic use of language in context anchored in a social-perspective goes beyond traditional linguistic analyses. It also considers social and sociocultural aspects of communication in an explicit manner, assigning them the status of a constitutive part of communicative action, thus overlapping with the research paradigms of anthropological linguistics (Duranti 1997; Hanks 1996), interactional sociolinguistics (Auer and Di Luzio 1992; Gumperz 1992, 1996, 2003), (critical) discourse analysis (Fairclough 1992; van Dijk 2008) and ethnomethodological conversation analysis (Garfinkel 1994; Heritage 1984). Societal pragmatics examines social parameters which affect the production and interpretation of utterances, placing language use in an external relation to language users. It is based on the premise that language use and social structure are connected dialectically, (re)constructing social and sociocultural context by confirming (or disconfirming) social values in interaction, for instance gender, power and social status. Consequently, the language user and her/his social roles and identities are at the heart of societal pragmatics, analysing social and sociocultural context on the micro level of face-to-face interaction as well as on the macro level of institution. In interaction, language users interactionally organize their social roles and identities while at the same time constituting social context. To use Heritage’s (1984: 242) terminology, the production of talk is doubly contextual. An utterance relies upon the existing context for its production and interpretation, and it is, in its own right, an event that shapes a new context for the action that will follow. The social perspective to pragmatics has shifted the focus of investigating context-dependent meaning from semantics-based methodologies and their premise of

Pragmatics as a linguistic concept

27

semantic well-formedness and propositional truth to social context-based approaches and their premise of negotiated communicative meaning and sociocultural appropriateness (Fetzer 2007), thus adding a further layer of meaning: social and sociocultural meaning. In this perspective, meaning is not given per se and carried as such into the utterance, metaphorically speaking. Rather, meaning, viz. lexical meaning and utterance meaning, is brought into the utterance and its diverse layers of meaning are brought out in the utterance, foregrounding salient meanings while backgrounding non-salient ones (to employ interactional-sociolinguistic terminology, Gumperz 1992, 1996). Language is assigned the status of a sociocultural construct which is used strategically by rational language users in context, considering possible perlocutionary effects their utterances may trigger as regards negative and positive politeness (Brown and Levinson 1987), for instance. The social-meaning perspective also prevails in ethnographic studies, sociolinguistics and interactional sociolinguistics, as well as in ordinary language philosophy (Brandom 1994; Recanati 2004; Searle 1969, 1975, 1995). It is in disaccordance with frameworks conceptualizing language and language use as mutually exclusive concepts, as is the case with the dichotomies of linguistic competence and communicative performance, I(nternalized)-language and E(xternalized)-language, or langue and parole. In the Anglo-American research tradition, the social perspective and the pragmatic perspective are referred to as European-Continental pragmatics. This field defines pragmatics “in a far broader way, encompassing much that goes under the rubric of sociolinguistics, psycholinguistics, and discourse analysis” (Huang 2007: 4), thus “not falling strictly within formal linguistic theory” (Horn and Ward 2004: xi). The Anglo-American tradition tends to favour the compositional perspective. 2.3.

The compositional perspective

In the pragmatics-as-component perspective pragmatics is assigned the status of a necessary component (or module) of grammar, thus anchoring pragmatics firmly to linguistics (Birner and Ward 2006; Carston 2002; Huang 2007; Horn 1984; Jaszczolt 2005; Levinson 2000; Sperber and Wilson 1995). Pragmatics is conceived of as an individual component of grammar, interacting with the components of semantics, morphology, syntax and phonology, as is the case in its most prominent representatives: theoretical pragmatics and formal pragmatics. Radical pragmatics departs from the premise of strict modularity or discrete compositionality by merging semantics and pragmatics. It is based on the premise of sense generality and contextualism, presuming that the output of semantic meaning is underdetermined and assigned determinate meaning in context through inference and pragmatic enrichment. The compositional perspective is based on the premise that language is a theoretical construct focussing on language-internal rules and regularities, and on in-

28

Anita Fetzer

ferencing processes, thus considering language use and context in an explicit manner. In contrast to societal pragmatics, however, its conception of context tends to be restricted to linguistic context. In the case of deixis and its realization as deictic expressions, e.g., temporal and local deixis, participant deixis and discourse deixis (Hanks 1996 and this volume; Levinson 1983), the compositional perspective needs to accommodate physical components anchored in the social context, e.g., speaker and hearer, and time and place. Participants tend to be conceptualized as the prototypical constructs of Model Speaker, Model Addressee or Model Person, viz. ideally competent interlocutors in the domains of linguistic competence, communicative competence and rationality (Brown and Levinson 1987; Jaszczolt 2005), abstracting from their social, discursive and interactional roles. The same holds for the conceptualization of time and place. At the heart of compositional-perspective investigations are pragmatic universals, viz. deixis, reference, non-literal meaning, indefiniteness, presupposition, information structure, discourse marker, speech act and implicature. To account for their function, pragmatics differentiates between an abstract construct and its linguistic realization (or representation). For instance, the pragmatic universal of deixis is examined with respect to deictic expression, and the universal of speech act is analysed with respect to its constitutive acts and their direct, indirect, and conventionalized indirect realizations (Sbisà 2002; Searle 1995; Brown and Levinson 1987). Closely connected with indirect speech acts are speaker- and heareroriented cognitive operations: the former are conventional implicature, generalized and particularized conversational implicature (Levinson 2000; Horn 1989; Huang this volume), and impliciture (Bach 2006), and the latter are inference and abduction (Cohen, Morgan and Pollack 1992; Givón 1989, 2005). Inferencing is also indispensable for the definition of intentionality of communicative action and non-literal meaning, for indefiniteness and discourse markers, and for presupposition and information structure. Another pragmatic universal is face (Brown and Levinson 1987; Goffman 1981, 1986), which is also of key importance to the social perspective. More recently, the compositional perspective has undergone a shift in focus and methodology, calling for the accommodation of methodological compositionality to bridge the gap between two competing paradigms (Jaszczolt 2005). Only then is it possible to account for speech act and uptake, speech act type and cognitive reality, and speech act type and situation (mis)match. The change in direction from monological, speaker-centred methodologies to dialogical, co-participant-centred interaction requires the explicit consideration of the connectedness amongst world knowledge, linguistic knowledge, discourse knowledge, the properties of the human inferential system and stereotypes and presumptions about society and culture, as is accommodated in the relational perspective.

Pragmatics as a linguistic concept

2.4

29

The relational perspective

The three perspectives adopted towards the heterogeneous field of pragmatics all share a particular perspective towards their object of investigation, accounting for its connectedness with context as regards utterance meaning and conditions of use. For this reason, they adopt a more or less explicit relational frame of reference. They go separate ways, however, in their accommodation of social context. While the social perspective considers social context to be a key concept, the compositional perspective tends to disregard it, concentrating on linguistic context and cognitive context. All of them are based on the premise that language is as a rulegoverned system, and all of them intend to account for form, function and meaning. The types of meaning under examination and their conceptualizations are different, though. In a relational frame of reference, language may be examined from a parts perspective, analysing the form, function and meaning of individual parts, viz. sounds, morphemes and constituents. It may be investigated from a holistic viewpoint, viz. phonological words and lexical words, and propositions and sentences, and it may be analysed from a parts-whole perspective considering the nature of their connectedness. The latter perspective is intrinsically dynamic, focussing not on the object as such, but rather on its embeddedness in context, its conditions of use and its appropriateness. The examination of linguistic form does not generally cause delimitation problems, while the analyses of linguistic meaning and function have become increasingly messy, necessitating the explicit accommodation of a pragmatic wastebasket for non-fits and other-fits. Pragmatics is generally delimited from semantics by shifting the focus of investigation from what is said, viz. semantic, literal or truth-conditional meaning, to what is meant, viz. what is communicated implicitly, what is implicated, utterance meaning or meaningnn (e.g., Grice 1975; Jaszczolt 1999, 2005; Levinson 1983, 2000; Recanati 2004; Saeed this volume). While the object of research is meaning in semantics as well as in pragmatics, accounting for world-to-language, languageto-world and language-to-world-to-mind relations, the delimitation of the object, the focus of investigation and the methodologies employed differ. In semantics the meaning of a lexeme, proposition or discourse unit is defined from a unit-of-investigation-internal perspective, and in pragmatics, the meaning of an utterance, speech act or conversational contribution is defined from unit-of-investigation-external and unit-of-investigation-internal perspectives. Metaphorically speaking, semantics examines what is in the container, and pragmatics examines what is brought out of the container (Reddy 1979). The following figure attempts to bring some order in the pragmatic wastebasket by systematising the two closely related but yet distinct fields of research, which intersect and overlap in their analysis of meaning in context.2 While meaning is conceptualised as conventional, truth-conditional meaning with discrete bound-

30

Anita Fetzer

Figure 1. Meaning in context

Pragmatics as a linguistic concept

31

aries in semantics, pragmatic meaning is seen as speaker-intended, non-truthconditional meaning with fuzzy boundaries, as is indicated by the dotted lines. Furthermore, meaning is autonomous and represented by well-formed context-independent sentences in semantics, and it is inferred and represented by appropriate context-dependent utterances in pragmatics. Semantics and pragmatics meet at the interpretation interface, where the calculation of utterance meaning is informed by the linguistic code and at the same time by pragmatic inferencing. While the former is a unidirectional operation as is signified by single arrows, the latter is dialectical and interactive in nature as is indicated by double arrows, capturing the dynamics of pragmatic meaning: The performative and its constitutive parts, viz. meaningnn, non-truth-conditional meaning and presupposition, are among the most important pillars of pragmatics. They are also relevant to the definition of speech acts and their realization in context as utterance meaning, which is a linguistic-pragmatics phenomenon par excellence. It is at the interpretation interface where context-dependent implicatures are calculated and where pragmatics interacts with semantics, syntax, morphology and phonology. In the following, the research domain of general pragmatics is introduced, concentrating on its bridging points with linguistic pragmatics.

3.

General pragmatics

General pragmatics is set firmly in the research paradigm of philosophy, considering cooperation, action theory, intentionality, rationality and context, thus providing relevant bridging points for language-anchored linguistic pragmatics. Both research paradigms analyse pragmatic universals and pragmatic mechanisms, such as speech act and implicature (e.g. Austin 1971; Bach 1992, 2006; Grice 1975; Sbisà 2002; Searle 1975; Vanderveken and Kubo 2002), indexical and deixis (Recanati 2004; Levinson 1983), common ground and presupposition (Akman et al. 2001; Bouquet et al. 1999), focussing on the connectedness between parts and wholes, and between content and context (Stalnaker 1999). While general pragmatics adopts an action-based frame of reference, linguistic pragmatics examines the form and function of speech acts in various languages. In general pragmatics, action and cooperation are analysed in the framework of an ideal pragmatic situation in which communication is neither distorted by social factors, e.g. institutional constraints, unequal social hierarchies and other forms of oppression (e.g., Habermas 1998, cf. Cooke this volume), nor by interpersonal considerations, such as face-wants and face needs (Brown and Levinson 1987), politeness or etiquette (Watts, Ide and Ehlich 1992). The high degree of idealization inherent in this research paradigm is also reflected in its conception of language and language use, which does not refer to natural language as spoken by or-

32

Anita Fetzer

dinary language users in context but rather an ideal language built in accordance with the language-system specific constraints and requirements. In an ideal pragmatic situation, all necessary felicity conditions obtain, such as normal input and output conditions, propositional content and preparatory conditions, sincerity and essential conditions (Searle 1969). Consequently, rational agents have ideal linguistic, communicative and cognitive competence as well as an ideal world knowledge (cf. Cohen, Morgan and Pollack 1992). Pragmatic universals can be assigned the status of basic building blocks of communication. Among the most prominent ones are the speech act and its constitutive acts, viz. propositional act composed of further constitutive acts and illocutionary act (Searle 1969; Vanderveken and Kubo 2002), indexicals, for instance pure and impure indexicals, intentionals and perspectivals (Recanati 2004), and presuppositions and presupposition triggers (Huang 2007; Levinson 1983; Stalnaker 1999). Another basic ‘building block’ is the rational agent or model user, who employs building blocks in a strategic manner to speaker-intend communicative meaning, thus achieving particular goals in context. The model user can be a natural user, that is speaker and hearer of a natural language, or an artificial user, such as a robot or a dialogue system. In accordance with the fundamental pragmatic premises of rationality and intentionality, the model user is equipped with a reasoning device, accommodating pragmatic inferencing and default inferencing, top-down reasoning and bottom-up reasoning, and abduction. This is a necessary condition for a model user to calculate and formulate speaker-intended implicatures and to perform the hearer-intended logical operations of inference and abduction (Cohen, Morgan and Pollack 1992; Givón 1989, 2005; Huang this volume; Sperber and Wilson 1995). 3.1.

Presupposition and common ground

Presupposition refers to a proposition or an inference whose validity is taken for granted for a sentence to be true or for a speech act to be felicitous. Philosophy and linguistics distinguish between pragmatic presupposition and semantic presupposition. The satisfaction of semantic presupposition is a necessary condition for the truth value of a sentence, and the satisfaction of pragmatic presupposition is necessary for a speech act to be appropriate in context. Pragmatic presuppositions are accommodated in speech act theory’s felicity conditions, which are considered as linguistic and social context categories and their satisfaction is assigned the status of a default configuration (Sbisà 2002). Presuppositions are generated by presupposition triggers, for instance definite descriptions, factive verbs, aspectual verbs, change of state verbs, iteratives and clefts (cf. Huang 2007, this volume; Levinson 1983). Some contexts allow presupposition neutralisation, and more recently presupposition triggers have been classified as hard presupposition triggers and soft pre-

Pragmatics as a linguistic concept

33

supposition triggers. Presuppositions are connected intrinsically with propositions and assumptions which are taken for granted and therefore do not need to be made explicit. They are organized and administered in the framework of context sets (Stalnaker 1999), which serve as common ground in communication. Consequently, pragmatic presupposition may be seen as a restriction on common ground. Common ground is indispensable for philosophical and cognitive conceptions of knowledge where it serves as background for reasoning and for retrieving speaker-intended meaning and other types of implicit meaning, such as indexical expression or implicature. In the field of computer science, common ground and world knowledge are frequently conceptualized as a database. It is seen as comprising a set of propositions, which serve as a resource for the understanding of utterances. Common ground is implicit but can be made explicit via propositions, and model users presuppose its validity and fall back on it when they retrieve implicatures (cf. Bublitz 2006). Common ground in the sense of background also plays a fundamental role in dialogue system modelling. According to Vanderveken and Kubo (2002), model users negotiate the compatibility of background with utterances and their felicity and satisfaction conditions in and through the process of communication. However, background and context are not identical because possible contexts of utterance can have different backgrounds. As a consequence of this, background contains not only mutual knowledge of facts about the conversational background but also knowledge about the world and of the world, such as ethical norms and sociocultural values, transcending the common sense notion of context. Searle (1995) considers background to be a necessary condition both for literal and non-literal meaning, thus assigning it the status of a basic premise for felicitous communication. He defines background as an open-ended set of skills, pre-intentional assumptions and practices, which are not representational but rather enable intentional acts and states to be made manifest. In Relevance Theory (Sperber and Wilson 1995), common ground is conceived of as a common set of premises for inference rules. The relevance-theoretic conception differs from the traditional notion of mutual knowledge by its attempt to avoid the logical consequence of infinite regress, which follows from the code model of communication. To avoid infinite regress, Sperber and Wilson base their theoretical framework on an approximation of mutual knowledge, namely cognitive environments and mutual manifestness. Common ground is a presupposed common knowledge base, which is a necessary condition for felicitous communication. To that base, model users anchor their communicative action and they fall back on it, should they require further information, which may not be encoded explicitly. The overall encompassing concept of common ground has been given a context-dependent interpretation as conversational record (Thomason 1992) based on Lewis’s concept of accommodation (Lewis 1979). Conversational record contains public information, presumptions,

34

Anita Fetzer

and an update operator. Other context-dependent subsets of common ground are Clark’s notions of personal common ground, which contains a model user’s subjective experience, and communal common ground, which stores social experience (Clark 1996), and Fetzer’s dialogue common ground, which is differentiated in individual dialogue common ground, containing a model user’s representation of her/his dialogue common ground, and collective dialogue common ground, viz. the model user’s representation of the set of model users’ representation of dialogue common ground (Fetzer 2004). All of the subsets need to be interconnected, constantly updated and, if necessary, revised. 3.2.

Context

The analysis of context-dependent meaning is at the heart of pragmatics, and for this reason context is one of its key objects of investigation. The theory, practice and implementation of context are also of relevance to diverse fields of investigation, ranging from philosophy and computer-mediated communication to cognitive science, in particular dialogue management, artificial speech production, artificial intelligence, distributed knowledge representation, robotics and information technology. The heterogeneous nature of context and the context-dependence of the concept itself have made it almost impossible for the scientific community to agree upon one commonly shared definition or theoretical perspective, and frequently, only a minute aspect of context is described, modelled or formalized (cf. Akman et al. 2001; Blackburn et al. 2003; Bouquet et al. 1999). Because of its multifaceted nature and inherent complexity, context is no longer considered an analytic prime but rather seen from a parts-whole perspective as an entity containing sub-entities (or sub-contexts). The multilayered outlook on context contains a number of different perspectives. First, context is conceived as a frame of reference whose job it is to frame content by delimiting the content while at the same time being framed and delimited by less immediate adjacent frames. The nature of the connectedness between the different frames is a structured whole composed of interconnected frames (Goffman 1986). The gestalt-psychological figure-ground scenario prevails in psychological and psycholinguistic perspectives on context. It is also adopted in cognitive pragmatics as is reflected in the relevance-theoretic conception of context as an onion, metaphorically speaking. The individual layers are interconnected and their order of inclusion corresponds to their order of accessibility (Sperber and Wilson 1995), as is reflected in inferencing and other kinds of reasoning. Second, context is seen as a dynamic construct, which is interactionally organized in and through the process of communication. This view prevails in ethnomethodology (Garfinkel 1994; Goodwin and Duranti 1992; Heritage 1984), interactional sociolinguistics (Gumperz 1996, 2003) and sociopragmatics (Bublitz 2009;

Pragmatics as a linguistic concept

35

Fetzer 2007, Schmid 2003), where context is assigned the dual status of process and product. The dynamic outlook is based on the premise of indexicality of social action, and the joint construction of context. In the primarily qualitatively oriented paradigms, context is connected intrinsically with adjacency pair, conditional relevance and the turn-taking system on the micro level, and with institution on the macro level, whose order is captured through context-independent and contextsensitive constraints and requirements. Closely related to the conception of context as a dynamic construct is its relational conception, which conceives it as relating communicative actions and their surroundings, relating communicative actions, relating individual participants and their individual surroundings, and relating the set of individual participants and their communicative actions to their surroundings (Fetzer and Akman 2002). Third, context is seen as given as is reflected in the presuppositional approach to context, which is also referred to as common ground or background information (Stalnaker 1999). Here, context is seen as a set of propositions, which participants take for granted in interaction. This allows for two different conceptions of context: a static conception in which context is external to the utterance, and an interactive one, in which context is imported into the utterance while at the same time invoking and reconstructing context. The context-dependence of context is reflected in its status as given and external to the utterance, reconstructed and negotiated in communication, indexical, and never saturated. A further classification of context is anchored to a holistic conception of context embedding its constitutive parts of model user, conversational contribution, surroundings and their presuppositions, viz. cognitive context, linguistic context and social context (Fetzer 2002, 2004). Linguistic context comprises language use and is delimited by the constraints and requirements of genre. Language is composed of linguistic constructions (or parts) embedded in adjacent linguistic constructions (further parts), composing a whole clause, sentence, utterance, turn or text. Linguistic context is functionally synonymous to co-text (de Beaugrande and Dressler 1981; Widdowson 2004), denoting a relational construct composed of local and global adjacency relations. Analogously to speech-act-theoretic constitutive rules and regulative rules (Searle 1969), linguistic constructions constitute text. The production and interpretation of linguistic constructions is constrained by the rules of grammar, and the production and interpretation of speech acts are constrained by felicity conditions. Cognitive context is the foundation on which inference and other forms of reasoning are based. Constitutive elements of cognitive context are mental representations, propositions, contextual assumptions and factual assumptions. Since cognitive contexts are anchored to an individual but are also required for a cognitively based outlook on communication, they need to contain assumptions about mutual cognitive environments. Thus, cognitive context is not only defined by representations but also by meta-representations. In the social-psychological paradigm,

36

Anita Fetzer

cognitive context is conceptualized along the lines of the gestalt-psychological distinction between figure and ground and the related metacommunicative concepts of frame and framing (Bateson 1972; Goffman 1986). Frame is seen as a delimiting device, which “is (or delimits) a class or set of messages (or meaningful actions)” (Bateson 1972: 187). Because of its delimiting function, “psychological frames are exclusive, i.e. by including certain messages (or meaningful actions) within a frame, certain other messages are excluded” and they are “inclusive, i.e. by excluding certain messages certain others are included” (ibid.). This also holds for context which, analogously to frame is also structured and metacommunicative, or to use Bateson’s words: “the hypothesis depends upon the idea that this structured context also occurs within a wider context – a metacontext if you will – and that this sequence of contexts is an open, and conceivably infinite, series” (Bateson 1972: 245). Social context comprises the context of a communicative exchange and is defined by deducting linguistic context and cognitive context from a holistic conception of context. Constituents of social context are, for instance, model users, the immediate concrete, physical surroundings including time and location, and the macro contextual institutional and non-institutional domains. The connectedness between language and language use on the one hand, and linguistic context, social context and cognitive context on the other is reflected in deixis, viz. temporal deixis, local deixis, participant deixis, discourse deixis and social deixis (Hanks 1996). Furthermore, the category of model user (as speaker and hearer) can no longer be conceived of as an analytic prime but needs to be refined by the accommodation of footing (Goffman 1981; Levinson 1988). The importance of social context to communication is spelled out by Hanks as follows: “Hence it is not that people must share a grammar, but that they must share, to a degree, ways of orienting themselves in social context. This kind of sharing – partial, orientational and socially distributed – may be attributed to the habitus, or relatively stable schemes of perception to which actors are inculcated” (Hanks 1996: 235). 3.3.

Cooperation

The pragmatic principle of cooperation is connected intrinsically with coordination and collaboration on the one hand (Clark 1996; Cummings 2005; Grosz and Sidner 1992), and with the pragmatic principles of rationality and intentionality (Cohen, Morgan and Pollack 1992; Searle 1983) on the other. The principles are examined in philosophy as regards their contents, function and status (Brandom 1994; Penco 1999), and they are analysed in psychology as mental representations, mental states and beliefs (Bouquet et al. 1999; Sperber and Wilson 1995). Against this background, language and language use are seen as fundamentally cooperative, and subsequently as intrinsically discursive and social (e.g., Grice 1975; Lewis 1979; Clark 1996).

Pragmatics as a linguistic concept

37

The connectedness between cooperation and intentionality, and the differentiation between I-intentions and we-intentions are at the heart of Searle’s conception of speech act theory. To use his own words: “Collective intentionality presupposes a Background sense of the other as a candidate for cooperative agency; that is, it presupposes a sense of the others as more than just conscious agents, indeed as actual or potential members of a cooperative activity” (Searle 1991: 414). In his work, Searle is very explicit about the nature of the parts-whole connectedness of individual speech acts and I-intentions on the one hand, and of a conversation or discourse and we-intentions on the other: “The reason that we-intentions cannot be reduced to I-intentions, even I-intentions supplemented with beliefs and beliefs about mutual beliefs, can be stated quite generally. The notion of a we-intention of collective intentionality implies the notion of cooperation” (Searle 1991: 406). In the Gricean approach to natural-language communication, in which language is seen as both purposive and social, and rational and cognitive, cooperation counts as the fundamental premise of communication. As a philosophical concept, cooperation is anchored to the premise of rationality and intentionality. Consequently, conversations are rational endeavours and cooperative efforts, in which model users realize and recognize a common purpose or set of purposes, or, at least, some mutually accepted direction. This is reflected in the Gricean cooperative principle which reads as follows: “Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged” (Grice 1975: 45). The cooperative principle is a necessary requirement for the calculation of pragmatic meaning or meaningnn, and is for this reason not open for negotiation. Model users may, however, employ the pragmatic strategy of opting out, which is generally accompanied by explicit references to some higher-order moral principle, such as confidentiality or trust. When they opt out, they generally spell out the reasons for their non-compliance, thus acting in accordance with the cooperative principle on a deeper level. The cooperative principle subsumes four conversational maxims, the maxim of quality, the maxim of quantity, the maxim of relation, and the maxim of manner (Grice 1975: 45–46). Further constitutive parts of the cooperative principle are the conventional and conversational implicatures, which are dealt with below in the section 4.2 on linguistic pragmatics. A cooperation-anchored outlook on language and language use not only is manifest in a conversation as a whole, but also in its building blocks of speech act and in the act of referring and predicating. Reference operates on the phrasal level and involves a speaker’s use of a linguistic expression, which tends to be a noun phrase, to induce a hearer to access or create some entity in her/his mental model of the discourse. Reference may be performed directly with a noun phrase and indirectly as deferred reference in the contexts of pars-pro-toto configurations, meaning transfer and figurative meaning. Referring seen from a cooperation viewpoint is not restricted to the semantic domain of a proposition but also applies to the in-

38

Anita Fetzer

terpersonal and interactional domains, viz. negotiation of meaning and common ground. The status of cooperation as a universal pragmatic principle has been examined extensively in general pragmatics, and there seems to be a general agreement on its validity. However, the claim that the Gricean cooperative principle is universal has been under severe attack. There have been modifications as regards the adoption of a further constitutive pragmatic principle on the same level, the politeness principle (Leech 1983), and there have been various modifications of the number of conversational maxims required for the calculation of pragmatic meaning as well as their speaker- and hearer-oriented conceptualizations and interpretations (Horn 1984; Lakoff 1973; Levinson 2000). Pragmatics is a field of research which intersects with various paradigms. As regards meaning, it cannot but consider semantics. As regards rationality, it cannot but consider cognitive science, and as regards cooperation, it cannot but consider philosophy. Thus, pragmatics needs linguistics – and linguistics needs pragmatics. For a thorough investigation of meaning, we need to examine how things are done with words, and how conversational contributions are structured to achieve particularized goals in context. This is done in linguistic pragmatics, where generalpragmatic principles and pragmatic universals undergo language- and culture-specific modification.

4.

Linguistic pragmatics

Linguistic pragmatics is defined as the science of language use, and “[i]n the same way as human actions change existing reality, linguistic actions also change the world” (Marmaridou 2000: 22). Linguistic pragmatics and general pragmatics share almost identical goals: general pragmatics examines pragmatic principles, mechanisms and universals in the context of action theory, rationality and intentionality, while linguistic pragmatics focuses on their instantiation in language and language use. Hence linguistic pragmatics overlaps with general pragmatics, sharing its generalized principles, mechanisms and universals, and departs from the generalized framework by concentrating on language as a general construct and on languages as particularized instantiations. In linguistics, language tends to be examined in its own right, accounting for language use, but not for the model user. The language system comprises the constitutive subsystems of phonology, morphology, syntax and semantics. These may be assigned the status of autonomous modules, as is the case in the formal paradigms, they may be conceived of as connected subsystems, as is the case in systemic functional grammar (Halliday 1994), or as dialectically connected subsystems with fuzzy boundaries, as for instance the cognitive paradigm (Givón 1993, 2005), where language use is connected dialectically with the language system and

Pragmatics as a linguistic concept

39

the model user. Language is seen as a dynamic construct changing through patterned use in context, in which grammatical and pragmatic aspects of interpretation interface (Ariel 2008: 257–259). The dialectical connectedness between language’s constitutive subsystems requires the explicit accommodation of pragmatic interfaces. The pragmatic perspective adopted towards the object to be examined does not consider the object in isolation but rather its connectedness with its conditions of use and with the user. This is reflected in grammaticalization or grammaticization (Traugott 1988) and in pragmaticalization (Aijmer 1997). Pragmatic principles shape the lexicon, as is reflected in Horn’s division of pragmatic labour, viz. his hearer-oriented Q-principle and speaker-oriented R-principle (Horn 1984), and their impact on the structure and use of lexical items, e.g. pragmatic strengthening, scalar implicature or polarity items. It is also manifest in deictic expressions and tense and aspect markers, whose meanings are constant but whose referents vary with the speaker and the hearer, and with time and place of the utterance, and on the style, register or purpose of a speech act (Ariel 2008; Huang 2007). Furthermore, pragmatic principles have a strong influence on the information status of constituents within and across sentences regarding focus, topic (or theme) and comment (or rheme). Deviations from the default generate implicatures, such as emphasis implicating a set of alternatives (Birner and Ward 1998; Lambrecht 1994). 4.1.

Speech act

Speech act theory is connected intrinsically with J.L. Austin’s groundbreaking lectures on ordinary language philosophy, edited posthumously (Austin 1971). Austin scrutinized the then prevailing view that language describes the world in terms of true and false, pointing out its shortcomings if applied to natural (or ordinary) language. He compared and contrasted the constative view, according to which sentences are true or false, with the performative approach, according to which sentences are used to perform speech acts, and came to the conclusion that model users do not just describe the world, but rather do things with words, even when describing the world. This is further refined by J.R. Searle’s modifications and refinement in his research on speech acts (Searle 1969, 1991), systematizing the cognitive, linguistic and social contexts, in which speech acts are performed by differentiating between constitutive rules, which are a necessary part of the language game and make speech acts count as particularized speech acts, e.g. promise, threat or rejection, and regulative rules, which are normative in nature. Considering language use from an action-theoretic frame of reference has paved the way for investigating linguistic meaning as well as speaker-intended meaning. It has shifted the focus of investigation from the rigid framework of formal semantics to the action-theoretic premises of rationality, intentionality and communication.

40

Anita Fetzer

Speech act theory’s basic unit of investigation is the speech act, and depending on the frameworks employed, speech acts are divided into locutionary, illocutionary and perlocutionary acts (Austin 1971), or propositional and illocutionary acts (Searle 1969), and their respective sub-acts, viz. reference and predication in the Searlean paradigm, and the phonetic act, phatic act and rhetic act in Austin’s frame of reference. Speech acts and their constitutive acts are performed simultaneously, and a speech act may be realized as a direct speech act, as an indirect speech act and as a conventionally indirect speech act. To account for a speech act’s perlocutionary effects as regards the social and interpersonal planes of discourse, speech acts have been further refined and redefined as face-threatening acts, in which the impact of a speech act’s perlocutionary effects on the interpersonal plane is calculated with respect to the model person’s, to employ Brown and Levinson’s terminology, negative and positive face wants (Brown and Levinson 1987). Speech act theory is anchored in action theory on the one hand, and in language and language use on the other. Speech acts are described in various ways: they are classified as explicit performative utterances, viz. utterances in which the speech act contains a performative formula realized by a first-person pronoun and a speech-act verb, e.g. I/we hereby close the meeting or we hereby request you to pay your dues; they are classified as first-order speech act and second-order speech act, accounting for the differentiation between direct and indirect speech acts; they are classified as micro speech act and macro speech act, accounting for the larger unit of genre (van Dijk 1981; Fetzer 2002), and as nth order speech acts, capturing the multi-layered nature of performing speech acts in context. Furthermore, speech acts are categorized with respect to possible directions of fit between language and the world, that is the words-to-world direction of fit for assertion, the world-towords direction of fit for directive and commissive, the double direction of fit for declaration and the empty direction of fit for expressive (Searle 1979). In addition to the illocutionary-point anchored taxonomies, speech act verbs are classified and systematized according to the way they are used in a particular language (cf. Collavin this volume). Speech act theory combines language-internal and language-external factors by systematizing them in the framework of felicity conditions, which are classified with regard to propositional content conditions, preparatory conditions, essential conditions and sincerity conditions. While propositional content conditions generally specify restrictions on the content of the proposition, preparatory, essential and sincerity conditions explicate language-external conditions concerning the language game as well as speaker- and hearer-specific requirements. Speech acts are neither true nor false but rather felicitous or infelicitous. To be felicitous, they need to be produced (and interpreted) in accordance with generalized felicity conditions, viz. normal input and output conditions, the essential and sincerity conditions, and with particularized felicity conditions, viz. proposition content condition.

Pragmatics as a linguistic concept

41

“Speaking a language” is seen as “engaging in a (highly complex) rule-governed form of behavior” (Searle 1969: 12). The rule-governed behaviour is specified in speech act theory’s felicity conditions and in the requirement that the model user in her/his role as speaker intends that her/his utterance will produce in a model user in her/his role as hearer a belief that the sincerity and essential conditions obtain by means of the recognition of the intention to produce this belief, and that the speaker intends that recognition to be achieved by means of the recognition of the utterance as one conventionally used to produce such beliefs. Secondly, that the semantic rules of the language spoken by the model users are such that the utterance is uttered sincerely and such as is required just in case all of the previous conditions obtain. 4.2.

Implicature

The cooperative principle counts as a universal principle in pragmatics, where it represents the solid base to which communication in general and the formulation and interpretation of communicative action in particular are anchored. This holds especially for the calculation of context-dependent communicative meaning and the necessary processes of inferencing required for the contextualization and enrichment of underspecified conversional contributions. To use Ariel’s words, “[u]nderdeterminacy is an inherent characteristic of human language, since no natural language sentence can encode interlocutors’ intended statements fully” (Ariel 2008: 265). In a similar vein, but more explicitly, Levinson (1995) argues that intentionality is a fundamental premise of natural-language communication. In his words, “human interaction, and thus communication depends on intentionascription. Achieving this is a computational miracle: inferences must be made way beyond the available data. It is an abductive process of hypothesis formation, yet it appears subjectively as fast and certain – the inferences seem determinate, though we are happy to revise them when forced to do so” (Levinson 1995: 241). Underdeterminacy and the necessary processes of inferencing are connected with the Gricean paradigm and its differentiation between what is said and what is meant, which both refer to utterances produced in context (Grice 1975); and they are also connected with Relevance Theory (Sperber and Wilson 1995). In both frames of reference, communication is seen as a context-dependent endeavour, in which communicative meaning may go beyond the level of what has been said. Hence, what is said cannot be equated with pure linguistic meaning but rather is “closely related to the conventional meaning of the words (the sentence) […] uttered” (Grice 1975: 44). Unlike the rather controversial status of ‘what is said’ in semantics and pragmatics, ‘what is meant’ has always been equated with ‘what is implicated’. Relevance Theory differentiates between a code model of communication and an inference model. While the former is conventional, the latter is inferential and

42

Anita Fetzer

hence pragmatic by definition. Sperber and Wilson define inference as follows: “Inference is the process by which an assumption is accepted as true or probably true on the strength of the truth or probable truth or other assumptions. It is thus a form of fixation of belief” (Sperber and Wilson 1995: 68). Their theory uses demonstrative inference as is employed in various types of logic, and non-demonstrative inference, which is employed in the process of inferential comprehension. Unlike local deductive reasoning, inferential comprehension is global and based on inductive rules. Pragmatic inferencing occurs at various levels of comprehension. It is connected with the differentiation between explicature, which is “a combination of linguistically encoded and contextually inferred conceptual features” (Sperber and Wilson 1995: 182), and different kinds of implication, viz. trivial implication, analytic implication, synthetic implication and contextual implication. Against this background, implicature “is a contextual assumption or implication which a speaker, intending her utterance to be manifestly relevant, manifestly intended to make manifest to the hearer. We will distinguish two kinds of implicatures: implicated premises and implicated conclusions” (Sperber and Wilson 1995: 194–195). Implicated premises “must be supplied by the hearer, who must either retrieve them from memory or construct them by developing assumption schemas retrieved from memory”, and implicated conclusions “are deduced from the explicatures of the utterance and the context. What makes it possible to identify such conclusions as implicatures is that the speaker must have expected the hearer to derive them, or some of them, given that she intended her utterance to be manifestly relevant to the hearer” (1995: 195). With these tools, Relevance Theory explicates the relevant inferencing processes involved in communication. They delimit themselves from the Gricean paradigm where, from their perspective, “the successes of human non-demonstrative inference must be explained by appealing not to logical processes of assumption confirmation, but to constraints on the formation and exploitation of assumptions” (1995: 81), viz. the cooperative principle and the maxims. Grice (1975: 43–44) differentiates between implicate and the related nouns implicature (cf. implying) and implicatum (cf. what is implied). He distinguishes between two basic types of implicature: conventional implicature and conversational implicature. The latter is subdivided into generalized conversational implicature and particularized conversational implicature. Generalized conversational implicature is also referred to as default implicature or pragmatic regularities (Bach 2006). This is in line with Levinson’s claim that “utterance-types carry generalized implicatures […]: rational speakers meannn both what they say (except in non-literal uses of language) and what that saying implicates; different layers of meaning all come under the umbrella of meaningnn” (Levinson 2000: 373). While conventional implicature is connected closely with linguistic form, for instance with connectives (e.g., but), implicative verbs (e.g., manage, forget to), honorifics or nonrestrictive relative clauses, conversational implicatures are essentially connected

Pragmatics as a linguistic concept

43

with certain general features of discourse, viz. the Gricean conversational maxims. Both conversational and conventional implicatures can be suspended: conventional implicatures are detachable, but not cancellable and conversational implicatures are cancellable but not detachable (for a thorough analysis of implicature, cf. Huang this volume). Conversational implicatures are generated or triggered by model users exploiting a maxim. That is to say, a model user gets in a conversational implicature if s/he flouts a maxim, blatantly failing to fulfil it. Grice (1975: 49–50) characterizes the notion of conversational implicature as follows: A man [or person, A.F.] who, by (in, when) saying (or making as if to say) that p has implicated that q, may be said to have conversationally implicated that q, PROVIDED THAT (1) he is to be presumed to be observing the conversational maxims, or at least the cooperative principle; (2) the supposition that he is aware that, or thinks that, q is required in order to make his saying or making as if to say p (or doing so in THOSE terms) consistent with this presumption; and (3) the speaker thinks (and would expect the hearer to think that the speaker thinks) that it is within the competence of the hearer to work out, or grasp intuitively, that the supposition in (2) IS required.

The conversational implicature is characterized by some degree of indeterminateness because it is, by definition, defeasible, non-detachable, calculable and nonconventional (cf. Bach 2006; e.g., Levinson 1979, 1983). Moreover, “conversational implicatures are universal and reinforceable” (Ariel 2008: 13). That is to say, they can be asserted explicitly without bringing about a redundancy effect. It needs to be pointed out, however, that the universality claim only holds for the inferential mechanism and not for the relevant background assumptions. Implicature is a cognitive mechanism anchored to model users in their roles as speakers or producers of conversational contributions, and inference is a cognitive mechanism anchored to model users in their roles as hearers. Cognitive processes can be local, they can be global or they can be both. Not only natural model users, but also non-natural intelligent agents may perform the cognitive operations of implicating and inferring, thus drawing invited inferences. For the operations to be felicitous, compatible contexts for reasoning and drawing analogies are required (cf. e.g., Bouquet et al. 1999; Cohen, Morgan and Pollack 1992). In Gricean pragmatics, the conversational maxims are seen as specifications of some unmarked communicative context representing the ground on which the conversational implicature is calculated. For this reason, deviations, however common, are seen as special or marked and signify that the speaker intends to communicate conversationally implicated meaning. The signalling is only possible against the background of the Gricean premise of a reflexive intention, which is succinctly explicated by Levinson: “[w]hat distinguishes a Gricean reflexive intention from other kinds of complex reflexive intention is that the communicators’ goal or intention is achieved simply by being perceived: recognition exhausts or realizes the intention” (Levinson 1995: 228).

44

Anita Fetzer

5.

Outlook

Neither general pragmatics nor linguistic pragmatics examine its objects of investigation in isolation but rather focus on their conditions of use, the connectedness with their surroundings, and the necessary and sufficient conditions which assign the object, e.g. intentionality, rationality, model user or action, the status of a particular object and make it count as that object. While general pragmatics concentrates on the analysis of these fundamental premises of practical action, identifying their necessary and sufficient conditions, linguistic pragmatics establishes the explicit connection between those foundations and their language-specific and language-use specific constraints and requirements. Pragmatics is more of a perspective towards an object under investigation than the examination of the object as such. For this reason, it needs to touch on and interface with neighbouring disciplines, in particular philosophy, cognitive science, neuroscience, linguistics and the social sciences. In spite of the diversity of the field, the key research question of pragmatics is concerned with communicative action, particularly with the expression and interpretation of meaning in context. To tackle that question felicitously, pragmatics needs to accommodate extra-linguistic world knowledge, cultural and social stereotypes, and situation of discourse on the one hand, and word meaning, sentence structure and the cognitive system, especially inference and abduction on the other. The generalized knowledge is stored as typed world knowledge and typed discursive knowledge. These defaults represent pillars, against the background of which the argument from ignorance holds: The argument from ignorance, it thus emerges, is an integral part of the inferential mechanism by means of which implicatures are recovered in conversation. This argument permits us to conclude that conversational principles and maxims are being observed on the grounds that there is no reason to believe that they are not being observed. Such a conclusion is at all possible because these principles and maxims have the status of presumptions in communication – principles that stand in the absence of counterindications. (Cummings 2005:109)

Presumptions or defaults concern both linguistic and extra-linguistic knowledge, which need to be stored as social and cultural world-knowledge defaults on the one hand, and as linguistic and discursive defaults on the other. Only then is it possible to account for a pragmatic perspective which accommodates the inherent dialogic and pragmatic principles of speech act and uptake, speech act type and cognitive reality, and speech act type and situation (mis)match, which is calculated against the background of the argument of ignorance. If a conversational contribution is in accordance with the presumptions, no counterindications are found. Consequently, the contribution is assigned the status of a match and interpreted accordingly. If a conversational contribution is in disaccordance, there are counterindications and a particularized process of inferencing is generated to calculate nonce meaning.

Pragmatics as a linguistic concept

45

The dialogue-oriented pragmatic mechanisms of inferencing and abduction require not only a dynamic framework accounting for the parts-whole connectedness, but also a conscious model user who is accountable for her/his actions (Fetzer 2004). At the same time, s/he makes use of automatic processing and default interpretations which figure as salient and strong interpretative probabilities, unless counterindications in the context signify a non-default match. Only an integrated framework, making allowances for methodological compositionality and synthesizing relevant findings (Fetzer 2004; Jaszcolt 2005; Marmaridou 2000) may capture the multifaceted nature of pragmatics.

Notes 1. I am grateful to the anonymous reviewer and the editors of this volume for helpful comments on the first version of this article, and I would like to thank Varol Akman and Teun van Dijk for discussing previous versions of this chapter. 2. The seemingly clear-cut distinction between semantics and pragmatics, though useful for analytic reasons, is highly idealized and has been strongly contested (cf. Bublitz 2009).

References Aijmer, Karin 1997 I think – an English modal particle. In: Toril Swan and Olaf Jansen (eds.), Modality in Germanic Languages. Historical and Comparative Perspectives, 1–47. Berlin: de Gruyter. Akman, Varol, Paolo Bouquet, Richmond Thomason and Roger A. Young (eds.) 2001 Modeling and Using Context. Third International and Interdisciplinary Conference, CONTEXT 2001, Dundee, UK, July 2001, Proceedings. Springer: Heidelberg. Ariel, Mira 2008 Pragmatics and Grammar. Cambridge: Cambridge University Press. Auer, Peter and Aldo di Luzio (eds.) 1992 The Contextualization of Language. Amsterdam: Benjamins. Austin, John L. 1971 How to Do Things with Words. Cambridge: Cambridge University Press. Bach, Kent 1992 Communicative intentions, plan recognition, and pragmatics: comments on Thomason and Littman and Allen. In: Cohen, Philip R., Jerry Morgan and Martha E. Pollack (eds.), Intentions in Communication, 389–400. Cambridge, MA: MIT Press. Bach, Kent 2006 The top 10 misconceptions about implicature. In: Betty Birner and Gregory Ward (eds.), Drawing the Boundaries of Meaning. Neo-Gricean Studies in Pragmatics and Semantics in Honour of Laurence R. Horn, 21–30. Amsterdam: Benjamins.

46

Anita Fetzer

Bar-Hillel, Yehoshua 1971 Out of the pragmatic wastebasket. Linguistic Inquiry 2: 401–407. Bateson, Gregory 1972 Steps to an Ecology of Mind. Toronto: Chandler Publishing Company. Beaugrande, Robert de and Wolfgang Dressler 1981 Einführung in die Textlinguistik. Tübingen: Niemeyer. Birner, Betty and Gregory Ward 1998 Information Status and Noncanonical Word Order in English. Amsterdam: Benjamins. Birner, Betty and Gregory Ward (eds.) 2006 Drawing the Boundaries of Meaning. Neo-Gricean Studies in Pragmatics and Semantics in Honour of Laurence R. Horn. Amsterdam: Benjamins. Blackburn, Patrick, Chiara Ghidini, Roy M. Turner, and Fausto Giunchiglia (eds.) 2003 Modeling and Using Context. 4th International and Interdisciplinary Conference, Context 2003, Proceedings. Heidelberg: Springer. Bouquet, Paolo, Patrick Brezillon, Massimo Benerecetti, Francesca Castellani and Luciano Serafini (eds.) 1999 Modeling and Using Context. Second International and Interdisciplinary Conference on Modeling and Using Context, Context’99, Trento, Italy, September 1999, Proceedings. Heidelberg: Springer. Brandom, Robert B. 1994 Making it Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, MA: Harvard University Press. Brown, Penelope 1995 Politeness strategies and the attribution of intentions: the case of Tzeltal irony. In: Esther Goody (ed.), Social Intelligence and Interaction, 153–174. Cambridge: Cambridge University Press. Brown, Penelope and Stephen C. Levinson 1987 Politeness. Some Universals in Language Usage. Cambridge: Cambridge University Press. Bublitz, Wolfram 2006 It utterly boggles the mind: knowledge, common ground and coherence. In: Hanna Pishwa (ed.), Language and Memory: Aspects of Knowledge Representations, 359–386. Berlin: de Gruyter. Bublitz, Wolfram 2009 Englische Pragmatik. (2nd rev. ed.). Berlin: Erich Schmidt. Carston, Robyn 2002 Thoughts and Utterances. The Pragmatics of Explicit Communication. Oxford: Blackwell. Clark, Herbert H. 1996 Using Language. Cambridge: Cambridge University Press. Cohen, Philip R., Jerry Morgan and Martha E. Pollack (eds.) 1992 Intentions in Communication. Cambridge, MA: MIT Press Collins Cobuild Essential English Dictionary 1990 Glasgow and London: Collins. Cummings, Louise 2005 Pragmatics: A Multidisciplinary Perspective. Edinburgh: Edinburgh University Press.

Pragmatics as a linguistic concept

47

Duranti, Alessandro 1997 Linguistic Anthropology. Cambridge: Cambridge University Press. Fairclough, Norman 1992 Discourse and Social Change. Cambridge: Polity Press. Fetzer, Anita and Varol Akman 2002 Contexts of social action: guest editors’ introduction. Language and Communication 22(4): 391–402. Fetzer, Anita 2002 Communicative intentions in context. In: Anita Fetzer and Christiane Meierkord (eds.), Rethinking Sequentiality: Linguistics Meets Conversational Interaction, 37–69. Amsterdam: Benjamins. Fetzer, Anita 2004 Recontextualizing Context: Grammaticality Meets Appropriateness. Amsterdam: Benjamins Fetzer, Anita 2007 Non-acceptances in context. Intercultural Pragmatics 4(4): 493–520. Garfinkel, Harold 1994 Studies in Ethnomethodology. Cambridge: Polity Press. Givón, Talmy 1989 Mind, Code and Context. Hillsdale: Laurence Erlbaum. Givón, Talmy 1993 English Grammar: A Function-Based Introduction. Amsterdam: Benjamins. Givón, Talmy 2005 Context as Other Minds. Amsterdam: Benjamins. Goffman, Erving 1981 Forms of Talk. Oxford: Blackwell. Goffman, Erving 1986 Frame Analysis. Boston: Northeastern University Press. Goodwin, Charles and Alessandro Duranti 1992 Rethinking context: an introduction. In: Alessandro Duranti and Charles Goodwin (eds.), Rethinking Context. Language as an Interactive Phenomenon, 1–42. Cambridge: Cambridge University Press. Grice, Herbert Paul 1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and Semantics, 41–58. New York: Academic Press. Grosz, Barbara J. and Candace L. Sidner 1992 Plans for discourse. In: Philip R. Cohen, Jerry Morgan and Martha E. Pollack (eds.), Intentions in Communication, 417–444. Cambridge, MA: MIT Press. Gumperz, John J. 1992 Contextualization and understanding. In: Alessandro Duranti and Charles Goodwin (eds.), Rethinking Context: Language as an Interactive Phenomenon, 229–252. Cambridge: Cambridge University Press. Gumperz, John J. 1996 The linguistic and cultural relativity of inference. In: John J. Gumperz and Stephen C. Levinson (eds.), Rethinking Linguistic Relativity, 374–406. Cambridge: Cambridge University Press.

48

Anita Fetzer

Gumperz, John J. 2003 Response essay. In: Susan L. Eerdmans, Carlo L. Prevignano and Paul J. Thibault (eds.), Language and Interaction. Discussions with John J. Gumperz, 105–126. Amsterdam: Benjamins. Habermas, Jürgen 1998 On the Pragmatics of Communication. Cambridge, MA: MIT Press. Halliday, Michael A.K. 1994 Introduction to Functional Grammar. London: Arnold. Hanks, William F. 1996 Language form and communicative practices. In: John J. Gumperz and Stephen C. Levinson (eds.), Rethinking Linguistic Relativity, 232–270. Cambridge: Cambridge University Press. Heritage, John 1984 Garfinkel and Ethnomethodology. Cambridge: Polity Press. Horn, Laurence R. 1984 Toward a new taxonomy for pragmatic inference: Q-based and R-based Implicature. In: Deborah Schiffrin (ed.), Meaning, Form and Use in Context: Linguistic Applications, 11–42. Washington: Washington University Press. Horn, Laurence R. 1989 A Natural History of Negation. Chicago: Chicago University Press. Horn, Laurence R. and Gregory Ward (eds.) 2004 The Handbook of Pragmatics. Oxford: Blackwell. Huang, Yan 2007 Pragmatics. Oxford: Oxford University Press. Jaszczolt, Kasia 1999 Discourse, Beliefs, and Intentions: Semantic Defaults and Propositional Attitude Ascription. Oxford: Elsevier Science. Jaszczolt, Kasia 2005 Default Semantics: Foundations of a Compositional Theory of Acts of Communication. Oxford: Oxford University Press. Lakoff, Robin 1973 The logic of politeness: or minding you p’s and q’s. Papers from the 9th Regional Meeting of the Linguistic Society of Chicago: 292–305. Lambrecht, Knud 1994 Information Structure and Sentence Form. Topic, Focus and the Mental Representation of Discourse Referents. Cambridge: Cambridge University Press. Leech, Geoffrey 1983 Principles of Pragmatics. London: Longman. Levinson, Stephen C. 1979 Activity types and language. Linguistics 17: 365–399. Levinson, Stephen C. 1983 Pragmatics. Cambridge: Cambridge University Press. Levinson, Stephen C. 1988 Putting linguistics on a proper footing: explorations in Goffman’s concepts of participation. In: Paul Drew and Anthony Wootton (eds.), Erving Goffman. Exploring the Interaction Order, 161–227. Cambridge: Cambridge University Press.

Pragmatics as a linguistic concept

49

Levinson, Stephen C. 1995 Interactional bias in human thinking. In: Esther Goody (ed.), Social Intelligence and Interaction, 221–260. Cambridge: Cambridge University Press. Levinson, Stephen C. 2000 Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Lewis, David 1979 Scorekeeping in a language game. Journal of Philosophical Logic 8: 339–359. Longman Dictionary of Contemporary English 2003 Harlow: Pearson Education Limited. Marmaridou, Sophia 2000 Pragmatic Meaning and Cognition. Amsterdam: Benjamins. Mey, Jacob 2001 Pragmatics. An Introduction. Oxford: Blackwell. Morris, Charles 1971 Writings on the General Theory of Signs. The Hague: Mouton. Penco, Carlo 1999 Objective and cognitive context. In: Paolo Bouquet et al. (eds.), 2nd International and Interdisciplinary Conference on Modeling and Using Context (Context’99), 270–283. Heidelberg: Springer. Recanati, François 2004 Literal Meaning. Cambridge: Cambridge University Press. Reddy, Michael 1979 The conduit metaphor. In: Andrew Ortony (ed.), Metaphor and Thought, 284–324. Cambridge: Cambridge University Press. Sbisá, Marina 2002 Speech acts in context. Language and Communication 22(4): 421–436. Schmid, Hans-Jörg 2003 Introduction: context and cognition – a concern shared by cognitive sciences. In: Ewald Mengel, Hans-Jörg Schmid and Michael Steppat (eds.), Anglistentag 2002 Bayreuth. Proceedings, 363–367. Trier: Wissenschaftlicher Verlag Trier. Searle, John R. 1969 Speech Acts. Cambridge: Cambridge University Press. Searle, John R. 1975 Indirect speech acts. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and Semantics, 59–82. New York: Academic Press. Searle, John R. 1979 Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press. Searle, John R. 1983 Intentionality. Cambridge: Cambridge University Press. Searle, John R. 1991 Collective Intentions and actions. In: John R. Searle, Herman Parret and Jef Verschueren (eds.), (On) Searle on Conversation, 401–415. Amsterdam: Benjamins. Searle, John R. 1995 The Construction of Social Reality. New York: The Free Press.

50

Anita Fetzer

Sperber, Dan and Deidre Wilson 1995 Relevance: Communication and Cognition. Oxford: Blackwell. Stalnaker, Robert 1999 Context and Content. Oxford: Oxford University Press. The New Encyclopaedia Britannica. 1976 The University of Chicago, Vol 14. Chicago: Helen Hemmingway Benton. Thomason, Richmond H. 1992 Accommodation, meaning, and implicature: interdisciplinary foundations for pragmatics. In: Philip R. Cohen, Jerry Morgan and Martha E. Pollack (eds.), Intentions in Communication, 325–363. Cambridge: MIT Press. Traugott, Elizabeth Closs 1988 Approaches to Grammaticalization. Amsterdam: Benjamins. Vanderveken, Daniel and Susumo Kubo (eds.) 2002 Essays in Speech Act Theory. Amsterdam: Benjamins. Vanderveken, Daniel and Susumo Kubo 2002 Introduction. In: Daniel Vanderveken and Susumo Kubo (eds.), Essays in Speech Act Theory, 1–21. Amsterdam: Benjamins. Van Dijk, Teun 1981 Studies in the Pragmatics of Discourse. The Hague: Mouton. Van Dijk, Teun 2008 Discourse and Context. A Sociocognitive Approach. Cambridge: Cambridge University Press. Verschueren, Jef 1999 Understanding Pragmatics. London: Arnold. Watts, Richard, Sachiko Ide and Konrad Ehlich (eds.) 1992 Politeness in Language. Berlin: de Gruyter. Widdowson, Henry 2004 Text, Context, and Pretext. Critical Issues in Discourse Analysis. Oxford: Blackwell. Wittgenstein, Ludwig 1958 Philosophical Investigations. Oxford: Blackwell. Yule, George 1996 The Study of Language. Cambridge: Cambridge University Press.

2.

Micropragmatics and macropragmatics Piotr Cap

1.

Introduction

The present chapter discusses micropragmatics (the pragmatics of utterance based concepts such as deixis, anaphora, presupposition, etc.) and macropragmatics (the pragmatics of discourse based concepts such as speech events, global intentionality or macro speech acts). Its aim is to show how micropragmatics and macropragmatics (have come to) complement each other, both conceptually and in terms of the division of labor across the field of contemporary linguistic pragmatics (henceforth: pragmatics).1 After a brief overview of the perspectivist approach to pragmatics in section 2., I turn to the discussion of ‘context’ (section 3.) and discuss how its potential and actual manifestations affect pragmatic inferences at different levels of the utterance and discourse structures. In 4., the perspectivist approach and the considerations of context are bridged to constitute two crucial factors defining micro- and macropragmatics, in addition to several other factors and developments that have led to such a distinction. In section 5., I deal with the four concepts central to the methodology of pragmatics in general: deixis, presupposition, implicature and speech acts. I discuss them from a micropragmatic perspective, showing their contribution to the proposition of an utterance, its illocutionary force, and the range and kinds of effects the utterance is able to bring about. At the same time, however, I introduce the macropragmatic approach, by raising three observations. The first is the different potential each of the four concepts possesses to contribute to the description of function and effects of an utterance. The second is the fact that, in the process of describing an utterance, some of the concepts ‘go together’ and form more sophisticated methodological tools, better than others. The third is the different range of utterance contexts each of the concepts is able to account for, which opens up prospects for their hierarchical application in analysis. In 6., I take these observations as prompts for an integrated study of intentionality at the macro level of discourse. Addressing such notions as speech event and macro speech act, I revisit the potential of individual speech acts for combining interacting discourse topics into larger functional units. I discuss the conceptual infiniteness pertaining to speech events in terms of their composition and hierarchy, and show the resulting relativity of the concept of the macro speech act.

52 2.

Piotr Cap

The pragmatic perspective

Pragmatics offers a unique, function-based perspective on all aspects of human (linguistic) communication (cf. Verschueren 1999; Fetzer 2002). It is concerned with all facets of a communicative act or a series of acts, such as the speaker, his/ her background knowledge and contextual assumptions, the lexical and grammatical constituents of an utterance, the hearer’s interpretations and patterns of inferencing, etc. All these are studied against a network of social factors, preconditions, norms and expectations that govern communication, both within a culture and across cultures. Since communicative acts involve linguistic units, whose choice is dictated by language-internal rules, as well as their interpersonal, social and cultural embedding, pragmatic studies bridge the system and the use side of language. They examine what is lexically and grammatically available for a speaker to accomplish a communicative goal, and at the same time explore the ways in which the linguistic potential is realized in a specific social context. Since the arsenal of language forms and the catalogue of social embeddings in which these forms are used in communication are virtually infinite and heterogeneous, ‘pragmatic’ scholarship naturally draws upon a number of diverse disciplines in linguistics (and beyond). It is generated collectively by phonologists, morphologists, syntacticians, discourse analysts – but also psychologists, sociologists and other scientists in the Humanities – as long as their work has a functional-communicative bent. Pragmatics is thus defined by its perspective more than by a set number of its objects of investigation.2 It offers explanations which apply to different lexical and structural levels of language – word, phrase, sentence, discourse.

3.

Context

The conception of pragmatics as a (functional) perspective foregrounds and endorses its preoccupation with context. Different types of context are naturally salient at different levels of communication (utterance, series of utterances, discourse, genre) but there are also manifestations of different aspects of context at any specific level. Hence, context is a fundamental concept to account for in describing the micro-macro dialogue in pragmatic studies. In a nutshell (cf. e.g. Bazzanella 2002), one can speak of a ‘static’ and a ‘dynamic’ conception of context. The static view (cf. e.g. Halliday and Hasan 2000; Cummings 2005) sees context as a set of variables (linguistic, cognitive, social) that ‘surround’ strips of text. They constitute an a priori, static infrastructure of resources which are there for interlocutors to determine the meaning of utterances at hand. On the dynamic view (cf. e.g. Duranti and Goodwin 1992; Bublitz 2003, 2006), context is never ‘given’ and interlocutors keep creating current contexts for current utterances – the moment a sentence is uttered “it becomes part of the envi-

Micropragmatics and macropragmatics

53

ronmental resources on which the contextual interpretation of the following as well as the preceding utterances has to draw” (Bublitz 2003: 383). As the next section will show in detail, the dynamic view of context is especially important in elucidating the relationship between micropragmatics and macropragmatics – the prerequisite being, of course, the move from a discourse-participant to a discourse-analytic perspective. The complementarity of micropragmatics (‘pragmatics of the utterance’) and macropragmatics (‘pragmatics of the discourse’) consists in the micro-level analysis getting verified from the macro standpoint, a procedure often leading to the redefinition of the original analytic track. Text and context are, not only in interlocution but also in analysis, constantly matched with and against each other, both prospectively and retrospectively. Any micropragmatic analysis of an utterance, reshaped as a result of macropragmatic considerations of the discourse to which the utterance contributes, is thus an example of ‘reinterpretation from hindsight’ – endorsed by the dynamic view of context. Such a reinterpretation is often indicative of the change in the number and kind of textual cues which are used before and after the reshaping takes place. It is likely that in the ‘original’ micro-pragmatic analysis the reliance on the most prominent textual cues will be significant, since the macro factors and constraints are, as yet, disregarded. However, the ‘reshaped’ micro-analysis may well see some of the initial cues brushed aside, often for the benefit of the cues originally neglected (see section 4.). All in all, the dynamic notion of context is a vital concept for micro- and macropragmatics since it allows speculations over, a) the extent to which the interpretation of meaning is guided by text; b) what factors may cause such an interpretation to be insufficient; c) at which level of analysis such a deficit becomes visible. The dynamic conception of context recognizes one more important premise for the existence of the micro-macro dialogue, i.e. the different degrees of accessibility of contextual cues, at different stages of interpretation.3 The linguistic cues (often referred to as ‘co-text’) are immediately accessible, but some of the cognitive and social cues might not be. This, altogether, makes discourse participants abstract from the ‘holistic’ context (the context which is often defined at the macro level, according to the characteristics of an attitudinal frame, genre, speech event, discourse type) the cues that seem actually relevant to a given communicative situation (micro-level). The above process has been extensively described in terms of a number of methodologies, which account for the relation between context and subcontexts (cf. Goffman 1986), figure and ground (cf. Langacker 1987, 2001), frame and framing (cf. Bateson 1972; Fetzer 2007), generalized (default, unmarked) context and particularized (marked) context, global context and local context, etc. Their common assumption is that the concept of context is best investigated from a parts-whole perspective, which elucidates its function of delimiting content. Such a perspective is naturally in line with the perspectivist view of pragmatics in general and, at the same time, congruent with the scope of micropragmatics and macropragmatics.

54 4.

Piotr Cap

Micropragmatics and macropragmatics: conceptual origins and scope

As has been indicated before, micropragmatics can be defined as the study of illocutionary force at the utterance level. In contrast, the focus of macropragmatics is not on the utterance, but on a series or sequence of utterances which form discourses/texts, seen as bearers of global intentionality of the speaker (i.e. the intentionality resulting from different speech act configurations, often referred to as speech events) and as initiators of complex effects. The distinction between micro- and macropragmatics follows from several factors as well as research developments. First, it has to do with flexible definitional boundaries of pragmatics. If pragmatics involves the study of meaning in context, then context is not limited in formal ways, such as the length of sentence/utterance or the amount of its discourse history. Contextual considerations which are used for determining the function and effect of an utterance may or may not stop within the boundaries of the language form used. Thus, we need a conceptual handle on the interface between smaller and larger functional units of discourse (e.g. speech act versus speech event), coded in smaller or larger forms (e.g. utterance versus text). Second, since context is complex, dynamic and multilayered (cf. section 3.), the accessibility of contextual cues is a matter of degree (cf. Marmaridou 2000). The more access to contextual (including co-textual) embedding of an utterance or a series of utterances, the better profiled the meaning. Hence a tendency on the part of the analyst to pile up contextual cues and to add to the picture as much of the ‘prehistory’ of the discourse situation (cf. Mey 1993) as possible (and feasible). But since access to contextual cues is each time different and invariably limited by a dynamic network of constraints, the analyst can hardly ever claim to have exhausted ‘all context’ to determine the meaning. In this sense, the micro-macro dichotomy involves a relation between the minimal and the maximal contextual input into analysis, irrespective of its scope, i.e. the length of the language form under scrutiny. The domains of micro- and macropragmatics are complementary in terms of their contribution to analytic labor. There is no micropragmatic analysis that would not provoke a macropragmatic extension of scope; similarly, there is no macropragmatic study that would not question, retrospectively, its micropragmatic components, thus prompting revision or modification of the original analytic track. Altogether, we arrive at a bottom-top-bottom cycle of upgrades on the explanatory power of both micropragmatic and macropragmatic concepts. This cycle involves looking at different aspects of context (linguistic, cognitive, social) from two different perspectives, a more linguistic bottom-up perspective and a more social topdown perspective. By way of illustration, consider a sample analytic procedure whose objective is to account for the pragmatics of speeches of the American cold war presidents (say,

Micropragmatics and macropragmatics

55

Harry S. Truman or Dwight D. Eisenhower). Imagine that a micro-level analysis of utterances making up these speeches points to a large number of deictic, referential and anaphoric markers, embedded in a multitude of direct speech acts. In contrast, some other units or phenomena explored in such an analysis, for instance implicatures and presuppositions, turn out to be underrepresented. The next step is to classify these findings under controlling categories, such as speech events or macro speech acts, to establish the performative denominators and thus postulate global function(s) of the discourse genre (i.e. the presidential cold war speeches) as a whole. This macro-analytic task requires inspecting the extralinguistic context, which informs and complements the linguistic observations. The likely outcome of the extralinguistic considerations is that the cold war presidents address a world that is ideologically divided and respond to a need to clearly define to their audience the essence of this divide, which eventually leads them to construct the wellknown ‘us and them’ opposition. Hence the rhetorical urge for directness and clarity, rationalizing the opulent use of indexicals or anaphoric expressions and banning most forms of implicit communication such as implicatures or presuppositions. Reaching this conclusion means that the analyst was able to use micropragmatic considerations as building blocks of a functional hypothesis, which he or she then tested against contextual factors (including his or her expert knowledge), in order to generate a macro-function of the genre. This is, roughly, how the analytic curiosity about regularities governing the salience of some phenomena observed at an utterance level can provoke the ‘macropragmatic extension of scope’. However, the upgrade cycle does not stop here, there is still room for a reverse procedure, which has the macropragmatic findings relate to the original component premises. Accordingly, the analyst may want to return to the bottom level of an utterance to look for more data, to further strengthen the macro conclusion(s). This may mean leaving out some of the now ‘useless’ forms (e.g. markers of implicitness) and delving more deeply into those which have passed the extralinguistic context verification. As a result, further micropragmatic activity can follow, with a view to, either narrowing down the study of the original parameters, e.g. deixis or direct speech acts, to only those aspects which are in line with the macropragmatic conclusion, or, possibly, identifying more bottom-level forms which are relevant, for instance more cohesive devices (cf. Halliday and Hasan 1976), relational propositions (cf. Mann and Thompson 1983), etc. Which is where the first round of the bottom-top-bottom cycle ends, but, of course, more rounds can still take place, with the micro-macro conclusions refining each other virtually ad infinitum. The existence of such a dialogue is probably among the most significant methodological arguments for having the two concepts, micropragmatics and macropragmatics, in the theoretical framework of the field. The distinction between micro- and macropragmatics has been strongly influenced by work in disciplines which pragmatics addressed while defining its scope, and upon which it imposed its functional perspective. The ethnography of

56

Piotr Cap

communication (cf. Hymes 1974; Saville-Troike 1989; Gumperz 1992) has provided insights into communicative acts seen as socially situated cultural forms, and micro-specifications of the more general social and linguistic contexts. Interactional socioliguistics (cf. Gumperz 2003) distinguishes between two levels of inference in analyses of interpretation processes. The higher level accommodates global inferences of what the exchange is about and what mutual rights and obligations apply, what topics can be brought up and what reply is desired by the cultural macro-convention, and the lower level contains local inferences concerning what is intended with the speaker’s own move and what is required by way of a response. Research in corpus linguistics (cf. Biber 1988) has adapted social and institutional factors to the examination of variation in language, both spoken and written. It is argued (cf. Biber 1988) that conversational and discourse goals comprise both conventional social and institutional goals, as well as personal microgoals. Furthermore, the micro-macro interface has been at the heart of studies in activity types (cf. Levinson 1979) and communicative projects (cf. Linell 1998). These studies have paved the way for bridging the micro with the macro on the plane of the explicit or implicit accommodation of context in interpretative processes. As it seems, the only way in which a distinction between the macro and the micro domains can be attempted is indeed by looking into the different degrees of lexical and grammatical coding of context in an utterance or a discourse vis a vis its extralinguistic coding, i.e. the import of mostly non-indexical cues. Such analytic endeavors are central to the most recent approaches to the analysis of communicative genres (cf. Bauman 1992; Luckmann 1995; Thibault 2003; Fetzer 2007). Genre (recall our sample analysis above) is a classificatory concept; it specifies typical ways in which lexicogrammatical resources are deployed to enact a communicative goal, by matching them against its own delimiting frame. Thus, micro analysis is an analysis of individualistic phenomena (tokens), which macro analysis verifies against the generic requirements (type). The token contributions are not merely cumulative, they can get encapsulated in one another or form a hierarchy – as will be shown in the next section.

5.

A micropragmatic perspective on deixis, presupposition, implicature and speech acts – with implications for macropragmatics

While the basic scope of micropragmatics is an utterance, any ‘micropragmatic’ analysis is, ultimately, a journey into the macro domain. This follows from the conceptual and methodological characteristics of the most common parameters utilized in description: deixis, presupposition, implicature, speech act. They are never static ‘components’ of the analytic enterprise; instead, they prompt an analysis which is essentially cumulative, interactional and hierarchical. At the bottom of the hierarchy is the input obtained from the interpretation of the lexical and grammati-

Micropragmatics and macropragmatics

57

cal coding of the utterance (immediate) context. This input (from deixis, presupposition lexical triggers, etc.) is usually insufficient for the utterance interpretation since it does not include non-textual cues. Accordingly, the analysis proceeds to the study of inferencing (pragmatic presuppositions, implicatures), which bridges the linguistic cues with the extralinguistic ones. Finally, the speech act level sees the function of the utterance emerge from the premises and interpretations (speaker-, hearer-, content-oriented) that have accumulated. Such a function remains, however, tentative, until a wider discourse context has been addressed for verification (viz. the sample analysis in 4.). Thus, the category of the speech act sits at the methodological borderline between the micro and the macro domain, which the latter involves generic, institutional, etc., characteristics and constraints. The speech act is at the same time an umbrella category for deixis, presupposition, implicature, and, potentially, a component category in the macro considerations. Altogether, micropragmatics feeds sequentially into macropragmatics because the consecutive-collective use of ‘micropragmatic’ tools in analysis entails a continual extension of scope, which eventually straddles the original utterance boundaries. The four subsections below unfold this sequence, showing, ultimately, the inherent relativity of the micro-macro distinction. 5.1.

Deixis

The outset of micropragmatic analysis is marked by establishing a ‘tangible’, lexicogrammatical architecture of the utterance, and especially the explicit manifestations of the relation between the words used and the context. The analyst’s task is to identify the lexicalized ‘pointers’ which indicate who utters the words to whom to accomplish what, when and where. This task endorses the status of deixis as the initiator of the (micro)pragmatic analysis since deixis is the concept that captures the relationship between the language form and the context in the most evident and direct manner. It can be viewed as the main phenomenon whereby features of context are encoded in utterances by primarily lexical (e.g. demonstratives) but also grammatical (e.g. tense) means.4 While the analysis of deixis ‘sets the stage’ for a more complex scrutiny of the utterance, its potential to interact with the other parameters of description (presupposition, implicature) is limited. As a concept involving lexicalized or grammaticalized forms of expression, it makes a relatively small contribution to the illocutionary force of the utterance, especially in implicit communication. Naturally, there are exceptions. For instance, deictic ‘projections’ (the cues are processed not from the speaker’s own perspective, but from the perspective projected on the hearer, cf. Levinson 1983; Fillmore 1997; Marmaridou 2000; Manning 2001) enable deictic expressions to generate implicatures. If, at 4:01 p.m. on Friday, I send an e-mail to my publisher which says:

58

Piotr Cap

(1) I will submit the manuscript on Saturday. I have good reason to expect the publisher (who works until 4 p.m. and only on weekdays) to read the e-mail no earlier than Monday and, seeing no attachment containing the ms. in the mailbox, to be puzzled – did I mean the Saturday that has just passed or the one to come? Of course, I retain the rhetorical comfort of denying whichever interpretation. In fact the ‘comfort’ starts the moment the Friday e-mail is sent out – I can feel free to submit the manuscript on whichever of the two Saturdays I prefer. There is no way I can be blamed for not submitting the ms. on the first of the two Saturdays, as I can always respond that I was taking my addressee’s perspective (i.e. receiving time), not mine. Most types of deixis (person, time, place, for an overview cf. Hanks this volume) are analyzed at the utterance level. However, a notable exception is discourse deixis, which crosses the utterance-discourse boundary, thus becoming a macropragmatic worktool. Discourse deixis involves the use of a lexical item within an utterance to point to the preceding or following utterances in the same discourse situation (speech event). The backward and forward reference can be illustrated by (2) and (3), respectively, where already refers to the earlier stretch of discourse and here anticipates the upcoming stretch: (2) As already indicated, all languages possess deictics. (3) Here goes my argument. There is no rule to how much textual distance should hold between the deictic expression and its referent. While it is reasonable to expect that the referent of here will be the immediately following utterance(s), the referent of already is surely not the closest, preceding utterance, but rather an utterance made much earlier in the unfolding discourse. This brings us to an important conclusion regarding the pragmatics of discourse deixis. By using a specific number of deictic expressions, the speaker is able to control the overt connectedness of discourse and, in consequence, its comprehension by the hearer. The presence of deictic markers in utterances – which make up a discourse situation where a specific topic is pursued – usually contributes to explicitness and clarity. On the other hand, the speaker may choose to withhold the use of deixis to purposefully obscure the message. These exceptions (deictic projections contributing to implicatures; discourse deixis as controllers of discourse comprehension), however important, do not detract from the core function of deixis in pragmatic studies. Deixis is an essentially micropragmatic phenomenon which sets analysis of the utterance ‘in motion’, but, by itself, adds little to the characteristics of the speech act(s) residing in it. The deictic framework of the utterance is sketched to establish the basic, lexically encoded relations between the referent(s) of the utterance and the ‘common ground’ of knowledge the utterance assumes to exist between the speaker and the hearer. Yet, a full account of these relations (and of the utterance function(s) the relations

Micropragmatics and macropragmatics

59

contribute to) needs complementation from concepts that go beyond lexically encoded relations (e.g. presupposition, implicature). 5.2.

Presupposition

Presupposition can be defined as a mechanism whereby the speaker addresses a body of knowledge and experience, involving both linguistic and non-linguistic contexts, which he or she assumes to be common to him-/herself and the hearer. The assumption of the existence of the shared knowledge may cause the speaker not to grammaticalize (or lexicalize) it in the utterance. This characterization takes presupposition to be a phenomenon lying at several intersections: the encoded and the assumed, the semantic and the pragmatic, (or even) the linguistic and the nonlinguistic.5 Presupposition comes in contact with deixis on the plane of its partial anchoring in lexical and structural forms. However, since many instances of presupposition can only be approached with reference to (non-linguistic) context, presupposition also reaches out in the direction of the implicit, constituting, in a sense, a shared knowledge prerequisite for communicating messages whose final destination is their inference by the hearer. Hence its feasible combination with the apparatus of implicature and, altogether, its relevance to the hierarchy of micropragmatic analysis, which derives its output from both accumulation and interaction of descriptions offered by the individual conceptual tools. As one of the latter, presupposition targets the communicative act at the stage where it develops ‘upwards’ from the lexicogrammatical coding of context to its further abstraction and elaboration by the speaker, with a view to producing a speech act. Throughout this stage, the speaker ‘decorates’ the deictic framework of the utterance with instantiations of knowledge shared by the speaker and the hearer with regard to all entities indexed, referred to, or implied in the utterance. Traditionally, the more a presupposition was linked with a lexical item or a linguistic construction generating it, the more it was treated as a semantic phenomenon, the other cases deemed ‘pragmatic’ and worth less attention precisely because of the absence of fixed language forms responsible for enacting particular presuppositions.6 This view has produced multiple typologies of presupposition, based on its embedding in lexicogrammatical forms called presupposition triggers. Furthermore, a number of properties have been assigned to presuppositions, including cancellability (a possibility of denying a presupposition, usually by adding more content to the utterance where it occurs) and constancy under negation (negating the predication carrying a presupposition does not detach or change the presupposition) (cf. e.g. Beaver 2001). From the perspective of micropragmatic analysis oriented toward the speech act characterization of the function of the utterance, as well as the macropragmatic perspective of the discourse, a rigid distinction between semantic and pragmatic

60

Piotr Cap

presupposition seems far from necessary. It offers little explanatory power compared to an integrated, global view of presupposition as a concept which should be studied against the utterance and discourse goals it serves. Such a view is quite naturally pragmatic, because even though there are specific lexical items (in fact, many) associated with specific assumptions (for instance, the assumptions of existence), their descriptive capacity does not expire within the structural boundaries of the linguistic expression. On the contrary, their significance goes much beyond as they are able to combine with primarily experiential premises and thus successfully contribute to a network of contextual, often non-linguistic beliefs making up the entire load of knowledge shared by the speaker and the hearer (see fn. 5). Most communicative goals served by presupposition (relevant to both the micro- and macropragmatic analysis) have to do, in one way or another, with economy of expression, though the latter is rarely the only goal sought. If a speaker could not rely on shared assumptions, the lexical and grammatical load of his/her utterance would grow in size, potentially obstructing its comprehension. Although economizing on the linguistic form for the benefit of unobstructed communication could be an utterance goal in itself, it is often a contribution to a larger utterance or discourse function. This becomes clear when analyzing, at the macro level, a series of utterances containing consecutive acts of adjustment to shifting discourse expectations. If I am overweight and say “I started jogging after visiting my doctor” and my confession meets with a blatant attack like “So it took you going to the doctor’s to work out like you should”, I can always follow a defense line in continuing “Well, to be honest, I tried to do some jogging a few times before, but now I do it regularly”. Since the short form of the initial utterance makes it underdetermined in terms of meaning, the denial of the ‘only then and never before’ presupposition comes rather easy, contributing to the overall explanation and justification (see Huang 2007 on cancellability of presuppositions). 5.3.

Implicature

The relevance of the analysis of implicature7 to micropragmatics (and, potentially, macropragmatics) is twofold. First, implicature studies account for the further processing of information that has been encoded by the speaker based on his/her presuppositions. Second, in doing the latter, they eventually recognize the contribution implicature makes to the update of the utterance-discourse context. From the analytic standpoint, the explanatory powers of presupposition and implicature are inherently complementary, shedding light on both the speaker and the hearer side of the speech act formation. They are suited to cover, in combination, the whole process of encoding messages by speakers and decoding them by their hearers. This process is essentially a continuum, where making a presupposition paves the way for the utterance before it takes on a linguistic form, in which the presupposition is lexically or non-lexically salient. From that point on, i.e. the

Micropragmatics and macropragmatics

61

point of making the utterance by the speaker, the recovery of the implicature(s) by the hearer may begin – of course, if the hearer senses a prompt to search for it/ them. The inference of the implicature(s), whether in accordance with the speaker’s expectations or not, finalizes the entire process, thus updating the status of the interaction and creating a new contextual basis on which to build presuppositions for further utterances in the exchange. The cycle in question corroborates the dynamic view of context (see section 3.) and endorses the intrinsic relativity of the micro-macro dichotomy. While the update takes place, technically speaking, ‘within the utterance’, its effect is on the prospective discourse. Delving deeper, implicatures created within the boundaries of the utterance (the micro level) are often ‘returned to’ or ‘readdressed’ purposefully later on in the unfolding discourse (the macro level). Since implicature is rarely encoded in language form,8 it involves a virtually indeterminable number of more or less complex contextual inferences. As such, it constitutes a valuable rhetorical tool whereby the speaker can control the flow of discourse, adopting his or her consecutive utterances to the current goals. This is due to a central property of implicature, cancellability (cf. Sadock 1978; Levinson 2000, etc.), which makes it possible for the speaker to deny, at any moment of speech situation, any implicature he or she apparently created. We have seen this property as partly relevant to presuppositions, but it is implicature that permits its broadest manifestation. Indeed, a great many implicatures are cancelled to re-establish adherence to the conversational norms (see the Cooperative Principle and the Maxims of Conversation by Grice (1975, 1989) as well as their numerous reformulations and supplements9) but also to ‘play’ with the addressee, pull a trick on him/her, or simply annoy him/her. Following on this note, many implicatures are cancelled for ironic or sarcastic effects, which is well documented in humor studies (cf. Attardo 1990, 1993). Another area where implicatures and their cancellations surface, is public (especially political) discourse. Implicatures contribute to the rhetorical safety of public speakers, who, on the one hand, wish to make statements that are universally acceptable, but on the other, want to retain the possibility of ‘refining’ or ‘fine-tuning’ these statements for hearers who hold different expectations of the meanings conveyed in them (cf. Cap 2008, 2010). Altogether, the phenomenon of the cancellability of implicature belongs to macropragmatics, since, first, the context that determines the cancellations is made up of a heterogeneous number of social and institutional factors, second, the ‘distance’ between implicature and its cancellation is a matter of discourse, rather than utterance. 5.4.

Speech acts – towards macropragmatics

The three brief subsections above have shown that deixis, presupposition and implicature make their distinctive micropragmatic contributions to understanding how an utterance is built, what its referents are and how they are encoded, what as-

62

Piotr Cap

sumptions are made before the utterance is produced, what effects can be expected after it has been produced and what inferential processes determine these effects. They partake in the process of enacting goal(s) of the utterance, from the speaker’s intention to realize its envisaged function via application of specific indicators of force, to the hearer’s successful (or not) recognition of this function and its results. A procedure this complex needs a controlling, ‘umbrella’ parameter of description. It needs a conceptual tool that is able to cover both speaker and hearer related aspects of the utterance function, and, while doing so, draw upon and thus systematize the particular contributions from deixis, presupposition, and implicature in order to make them fit for macropragmatic work at the discourse level. The concept of the speech act seems an excellent theoretical candidate to take up this task.10 The orientation of speech acts to both parties of a verbal exchange, as well as to its linguistic matter, is visible at a glance from the traditional distinction between the locutionary, the illocutionary, and the perlocutionary aspects of a speech act.11 While the locutionary aspect is the most ‘objective’ since it concerns the stable language form of the utterance, the illocutionary and perlocutionary aspects involve a dynamic negotiation of meaning between the speaker and the hearer. In saying “It’s hot in here” a speaker may be producing an (implicit) illocutionary act requesting the hearer to open the window, and the perlocutionary act (effect) might be that the hearer indeed opens it, but it might also be that he or she turns on the air-conditioning instead. Thus, the illocutionary-perlocutionary relation not only mirrors the complex process of meaning evolution as sketched at the beginning of this subsection; it also inscribes in the distinction between explicit (direct) and implicit (indirect) ways of communicating a speech act. Consequently, it invokes the notions of deixis, presupposition, and implicature, since they all situate themselves at some specific yet different points of the conceptual axis which links ‘what is said’ with ‘what is effected’. The classificatory, controlling power of the speech act is further reflected in its network of felicity conditions, i.e. the conditions that underlie a successful, logical, ‘felicitous’ production of different acts.12 For example, a speaker cannot make a successful order if he or she does not sincerely want the order to be followed, or if he or she deems the hearer incapable of following it. These two felicity conditions are excellent illustrations of the connection that holds between the concept of the speech act and the other ‘micropragmatic’ concepts – a relation we have postulated at the beginning of this subsection. The speaker’s awareness of cognitive and social context obtaining at the moment of producing a speech act gives rise to pragmatic presuppositions underlying the utterance that contains this act. Then, once the act is accomplished, the speaker’s presuppositions can be assessed against the effectiveness of implicatures they helped to create. While this part is naturally the domain of implicature studies, we do get an extra insight from the ‘cover’ framework of the speech act theory and, specifically, from its contribution to the research in direct versus indirect acts.

Micropragmatics and macropragmatics

63

Ironically, the greatest advantage of this research is that it almost removes one of the two groups of acts from our scope of interest, at both the micro and the macro level of description. As pointed out in several studies, most speech act usages are indirect (cf. Bertolet 1994; Holdcroft 1994, etc.) – and indeed, even the apparently straightforward act of asking for salt in “Can you pass the salt?” is, formally speaking, indirect. This, however, should not be discouraging. The orientation of the speech act theory to indirect usages is only reflective of the visible emphasis on indirectness that is salient in the other (micro-)pragmatic concepts; after all, our discussion of deixis, presupposition, and implicature has been mostly preoccupied with ‘hidden’ meanings. Thus, speech act theory turns out to be nothing but congruent with the other apparatuses. Moreover, it extends over all of their fields of application, from the stage where knowledge is assumed to get encoded in the utterance, to the stage where the utterance is interpreted. Conceivably, a single speech act can be ‘realized’ through recourse to an x number of presuppositions, an x number of deictic markers in the utterance, which the latter could produce an x number of implicatures. We have thus arrived at a micropragmatic hierarchy of analysis, with the speech act constituting a category superordinate over the other micropragmatic categories. At the same time, however, the speech act should not be considered the top-most variable of description; in order to account for discourses, rather than individual utterances, we are in need of yet higher-rank concepts.

6.

Macropragmatics

Most speech acts residing in individual utterances tend to combine into larger functional units, thus paralleling the combination of utterances into texts and discourses. This process can be seen from the analysis of (4). Assume that instead of reading out students’ names from the attendance list at the beginning of a seminar, I simply ask: (4) Hello, are we all here? In uttering (4), I perform two direct speech acts (greeting and asking), as well as an indirect act of requesting my students to reveal the names of the absentees. The speech acts involve a deictic anchoring (e.g., “here” recognized as classroom), presuppositions (e.g., of some students still missing), and the indirect act involves a process of inference. This is, with some simplification, where a micropragmatic analysis of (4) can get us. However, being myself the producer of (4), I can enjoy the privilege of stating with absolute certainty that, most of the time, I do not open the class with this utterance for the mere sake of greeting, asking, and requesting, as described above. That would be unlikely considering that (4) happens regularly and has thus become

64

Piotr Cap

(a part of) expectable routine, a pattern that arises, usually, from some kind of a larger intention.13 Mark an analogy: If I follow the routine of buying each morning a newspaper (an action which involves micro-actions analogical to the acts defined in (4)), the ‘larger intention’ can be described as a general aspiration to stay updated with the current news. The larger intention in (4) is to begin the seminar, making sure all things are in place for a productive meeting. This finding does not sound terribly original in itself, but we will take it as a starting point to show, eventually, that the nature of intentionality is such that it cannot be a matter of the individual utterance and thus the scope of inquiry must be extended accordingly. More often than not, I may need to say more than just (4), to successfully begin the seminar. I may choose from an apparently infinite catalogue of other utterances which are able to serve the same intention. Some typical cases may be: “It’s getting late”, “Would you close the doors please?”, “Now, Adam, listen up, will you?”, “Right, shall we begin?”, “And the chalk is missing as usual!”, “I can’t remember when we last started on time”, etc. All these seemingly diverse utterances contribute, in one way or the other, to realization of the principal intention. Interestingly, they do so while performing different speech acts (assertives, directives, expressives) whose force may also be different, direct or indirect. Consequently, each of these acts may accommodate a different input from deixis, presupposition, and implicature. The routine of performing individual speech acts (which may in turn involve the other micropragmatic categories) to realize a more general intention, can be defined as a speech event.14 Speech events may contain, as we have seen, a potentially infinite number of utterances, depending on how many are actually necessary to carry out the intention. If, for instance, my seminar group enters the classroom discussing a just-finished exam in which they participated, I may be in for a longer stretch of talk to ‘set up the stage’ for my own class. Otherwise, it may take a few short remarks to cover the technicalities and initiate the topic proper. Whatever happens in actuality is thus dictated, as anything in pragmatics, by context. A speech event can comprise just as many utterances (and speech acts) as needed to match the contextual preconditions. These utterances do not have to come in a monologic pattern. I may keep performing the principal speech event by producing a few utterances in a row, then pausing, then, possibly, reacting to a question that comes in the meantime, then resuming the monologue, etc. Thus, my performance is a genuine ‘discourse’ performance, which exists in and responds to a social setting, though, of course, some parts of it can mirror ‘textual’ units and their connectedness patterns. The concept of infiniteness resurfaces once we go on to discuss speech events as ‘building blocks’ for realization of a still ‘larger’ intention. To readdress the case initiated by (4), the speech event which realizes the intention to successfully begin the seminar is not only an umbrella category for a series of individual acts, but may itself be subordinate to a larger discourse goal – be it, for instance, con-

Micropragmatics and macropragmatics

65

ducting a productive seminar meeting as a whole. This goal entails that an apparently infinite number of speech events are carried out ‘on the way’, from a speech event of providing a complex, multi-act explanation to a problem that has arisen during the session, to a speech event of assigning homework, which may again involve a number of component speech acts. The latter (event) may include: an expressive (e.g., reprimanding students for not completing the previous assignment), a commissive (e.g., threatening to fail students at the end of the course, if they keep neglecting their assignments), a directive (e.g., telling students to do a particular task for the next meeting), an assertive (e.g., describing a rationale for the task), etc. The variety and diversity of the acts involved is, here, no smaller than in the case of the speech event performed to begin the seminar. In this clearly bottom-up fashion, we have approached the problem of the ‘uppermost’ or ‘global’ category of intentionality enactment, one whose promise would be to encompass all the subordinate intentions realized in speech events and their attendant acts. Such a promise is partly fulfilled by the classical concept of the macro speech act, which is, in van Dijk’s words, a global speech act performed by the utterance of a whole discourse, and executed by a sequence of possibly different speech acts.15 Still, while doing some useful job in the way of systematizing speech events, the macro speech act suffers from a problem of an inherent relativity as regards the range of its own operation. The definition of the macro act only corroborates the problem; we do not get to know how much is ‘a whole discourse’. Is it, to return to our example above, the whole body of discourse produced within the duration of the seminar, controlled by the global intention to make it a productive meeting? Then the relationship between the macro speech act and the component speech events (and their individual acts) seems analytically elegant. But, does this account exhaust the potential of the macro speech act to combine with further macro speech acts, to serve a yet-more-global intention? Apparently not. The intention to carry out an academically rewarding seminar can be considered subordinate to the intention to conduct the entire course as planned, which in turn partakes in the intention to perform my (academic) duties properly as a whole, etc., etc., – which, naturally, makes the consecutive macro acts accumulate accordingly. To envisage the highest-rank intention, pursued in the highest-rank speech act, turns, then, a philosophical undertaking. One of the undisputed methodological values of the search for the clear-cut categories signposting the particular levels of intentionality and its enactment lies, paradoxically, in the recognition of limits to which intentionality could be accounted for in larger stretches of discourse. Thus, endeavors like the macro speech act theory should not be carelessly brushed off, as they eventually motivate research, however minimalist, in better demarcated and better empirically equipped areas. Apart from the aforementioned research in genres, a prominent example of such an area is Conversation Analysis.16 Conversation analysts have elaborated an impressive arsenal of techniques for the description of speech act deployment,

66

Piotr Cap

though in necessarily limited contexts, often just co-texts. As a result, the apparatus of macropragmatics has been endowed with a number of relevant concepts, such as ‘floor’, ‘topic’, ‘turn’, ‘turn-taking’, ‘transition relevance place’, etc., each of which contributes its share of insight in how people manage their intentions and goals in particularized (conversational) settings. A related and very important significance is that a ‘macro’ perspective on the pragmatics of discourse makes the analyst approach intentionality as continually re-shaped and updated by extralinguistic context and thus work out methodological tools to handle the variability of enactment of the speaker’s intention(s) over an extensive period of time. Let us illustrate this phenomenon with an example from political discourse, a domain rich in speakers’ attempts to continue with an enactment of a global intention, though often in consecutively modified patterns, responding to the changing extralinguistic reality. Consider an excerpt from one of G.W. Bush’s speeches legitimizing US involvement in the recent Iraq war: (5) By advancing freedom in the greater Middle East, we help end a cycle of dictatorship and radicalism that brings millions of people to misery and brings danger to our own people. By struggling for justice in Iraq, Burma, in Sudan, and in Zimbabwe, we give hope to suffering people and improve the chances for stability and progress. […] Had we failed to act, the dictator’s programs for weapons of mass destruction would continue to this day. For all who love freedom and peace, the world without Saddam Hussein’s regime is a better and safer place.17 This speech comes eight months into the war, which started on the assumption that the Iraqi regime (and its alleged terrorist allies) had developed access to weapons of mass destruction (WMD), thus becoming a world threat and unwilling to disarm unless forced to. In his address, Bush attempts to maintain the aura of legitimization of the US intervention, against the increasingly evident collapse of the original premise – no weapons of mass destruction were found in Iraq ever since the US and the coalition troops entered it in March 2003. Since he can no longer employ the single WMD argument to execute his global intention (to keep the legitimization process intact), he deftly switches to a more universal, ideological rationale. Instead of invoking a direct threat from a destructive impact of WMD, he concentrates on building up an ideological framework for a potential growth of such a threat in the future and presents the US strike as part of the necessary means to ensure that the antagonistic ideologies (“dictatorship”, “radicalism”, “regime”) do not materialize in the form of physical impact. The localization of these ideologies in more countries than Iraq itself (viz. “Burma”, “Sudan”, “Zimbabwe”) is in fact a clever rhetorical ploy. By extending the spectrum of the ideological conflict and, thus, the spectrum of US foreign involvement, he encourages the construal of the WMD intelligence failure in terms of an isolated incident, fully justifiable given the range of the American mission as a whole. Altogether, Bush

Micropragmatics and macropragmatics

67

keeps up with the global intention to continue with legitimization of the Iraq war, now on ideological grounds, but the pragmatic patterns (and, consequently, lexical choices) responsible for realizing this intention within a macro-temporal reality need to be updated to meet the extralinguistic developments. In section 4., I argued that macropragmatic analyses often lead to redefinitions of the micropragmatic input that has been utilized to build up their tentative versions. They also tend to review the micropragmatic concepts and point to those whose descriptive potential has not yet been fully exhausted. The analysis of Bush’s speech is a good case in point. At its macro-level, we consider the general patterns of adaptation of the political speaker to contextual conditions. This leads us to the more focused question which of the micro-level concepts are able to account for such processes in the possibly richest lexical way, i.e. which of them are most frequently reflected in lexical items and structures that eventually make up the matter of the macroanalysis. The analysis of (5) reveals, for instance, that much explanatory power is yet to be drawn from a careful scrutiny of implicature forms in the text. If we gather from macro-contextual considerations that Bush is forced to switch to a new argumentative strategy, yet, as seems logical, without discrediting the previous argumentation, what better way to accomplish the goal than through implied meanings, which are always subject to cancellation as the speaker sees fit? Recall the phrase “programs for weapons of mass destruction”. It is flexible enough to concede that Saddam did not indeed possess WMD understood as a ‘product ready for use’, at the outset of the war, but it does not completely detract from the original assumption that he did. It is quite likely that a microanalysis of this implicature (as well as of many other bottom-level forms in (5)) could be overlooked but for the prompt from the macropragmatic approach, involving a vast range of contextual factors affecting the entire discourse of the Iraq war. In this way we have returned to the question of the ‘micro-macro dialogue’, which I defined in the fourth section of this chapter as one of ‘the most significant methodological arguments for having the [distinction between the] two concepts, micropragmatics and macropragmatics, in the theoretical framework of the field’. It should be said in closing that prompts for such a dialogue need not come from ‘pragmatic’ analyses alone; in fact, the dialogic relationship between micropragmatics and macropragmatics is kept alive by insights from approaches which, at least definitionally, go much beyond linguistics-based disciplines or perspectives. An example of such an approach is Critical Discourse Analysis (CDA; cf. e.g. Fairclough 1995; Wodak and Chilton 2005), which extends onto domains traditionally associated with various branches of sociology and social psychology. Rooted in the conception that discourse is just as much reflective of the existing reality as it is potentially constitutive of a new reality, CDA offers a number of useful ways for the analysis of ‘the reality’ (which pragmatics would rather term extralinguistic aspects of context). In so doing, it provides (macro-)pragmatic considerations with socio-cognitive grounding of discourse (cf. van Dijk 1995, 2002), thus signposting

68

Piotr Cap

the ‘downward’, top-down analytic procedure. Two examples of how such a procedure could operate have been, in this article, the suggested refinements and extensions of the analysis of deixis and direct speech acts in the genre of the cold war presidents’ speeches, as well as the just discussed hint for a fully fledged study of implicature in (5). Of course, CDA has much more to offer, as does any approach or discipline that recognizes the paramount importance of (holistic, dynamic) context in human (linguistic) behavior. In whatever kind of acting, including pragmatic acting, we adapt ourselves to a context ‘as is’, but at the same time we change it with a view to future effects. As Mey (1993: 263) writes, all this is done through the intermediate use of language as the tool helping us to select the relevant features of any situation in the total context. Micropragmatics and macropragmatics are two perspectives which, on the one hand, ‘statically’ account for the different size manifestations of this ‘total’, or ‘holistic’, context. On the other hand, they ‘dynamically’ inform each other (and thus the pragmatic analysis as a whole) of, first, which instances of the utterance-level use of language are worth a broader contextual verification to determine their function, second, which parts of the broader contextual check are applicable as determiners of further analytic activity at the utterance level.

Notes 1. See Fetzer (this volume) for the distinction between ‘general (non-linguistic) pragmatics’ and ‘linguistic pragmatics’. 2. See Fetzer (this volume) or Huang (2007) for a discussion of perspectivist versus componential views of pragmatics. 3. Cognitive pragmatics (e.g. Sperber and Wilson 1986) illustrates this by the ‘onion metaphor’. The pattern of inclusion of the individual layers, which are interconnected, reflects their order of accessibility in the process of inferencing. 4. Foundations of research in deixis are presented in Bühler (1934/1982). Among the most useful modern overviews are Nunberg (1993), Fillmore (1997), and Levinson (2004); cf. also Hanks (this volume). 5. The standard view of presupposition (see Huang this volume) is that (i) it involves a proposition whose truth is taken for granted in the utterance of a sentence and remains in force even if the sentence is negated, and that (ii) it is engendered by certain lexical and structural triggers. This definition goes back to philosophical and semantic views of Frege (1892) and Strawson (1952), later enriched with semantic-pragmatic analyses (cf. Karttunen 1973; Stalnaker 1973; Gazdar 1979). Though these characterizations of presupposition are still generally accepted (cf. Saeed 2009), many linguists have pointed to the existence of counter-examples deeming the mainstream approach (necessarily) “imperfect” (cf. Huang this volume). In the present chapter, presupposition is treated as a phenomenon which does not rely for its existence on linguistic form alone, but on a complex relation between the linguistic form, the speaker’s and the hearer’s (common) “understanding of the world” (cf. Grundy 1995: 78), as well as the speaker’s fulfillment of

Micropragmatics and macropragmatics

6.

7.

8. 9.

10.

11. 12. 13.

14.

15.

69

felicity conditions for performance of an act involving presupposition(s). This stance integrates the standard view with several other research perspectives: the speech act approach to presupposition (cf. Grice 1981; Green 1989), the cognitive approach (cf. Carston 1996), the experiential approach (cf. Marmaridou 2000), or the common ground approach (cf. Grundy 1995). Examples of ‘battles’ over the status of presupposition are Stalnaker (1973), Soames (1989); while Horn (1996), Atlas (2004) and Huang (2007) offer well-balanced overviews. In the present chapter the use of the term ‘implicature’ is synonymous with ‘conversational implicature’. See Huang (this volume) for a comprehensive account of various types of implicature. We do not deal with ‘conventional implicatures’ here (cf. e.g. Davis 1998; Bach 1999). See reductionist attempts such as Horn’s (1984) Q-R model or Levinson’s (2000) Q-I-M principles; for supplements and general overviews see Levinson (2000), Horn (2007, 2009), Huang (2009a, b and this volume). See Austin (1962/1975) and Searle (1975, 1979) for the foundations of the Speech Act Theory. Tsohatzidis (1994), Thomas (1995), Marmaridou (2000), and Sadock (2004) are excellent, objective overviews of both the classical ideas and later developments. See also Collavin (this volume). See Austin (1962/1975) and a review of Austin’s ideas in e.g. Alston (1994). Cf. Austin’s (1962/1975) account of felicity conditions; see also Searle (1979). Despite a temptation to the contrary, I save the term global for the discussion of macro acts. I refer to this intention as ‘larger’ not just to say that it is ‘more indirect’ than the intentions to greet, grasp students’ attention, or work out the number of absentees. My point is that all of the latter can accumulate into a bigger intentional schema and that there is no limit to the number of micro-intentional components making up such a schema. Hence the use of the comparative ‘larger’, which allows for a ‘downward’ specification (i.e. the specification of what the ‘larger intention’ is supposed to comprise), as well as an ‘upward’ specification (i.e. the specification of what higher or highest level of intentionality it could contribute to). Interestingly, van Dijk’s (1980) speech event approach to intentionality levels, as well as the activity type and genre oriented approaches to intentions and context (cf. Levinson 1979; Bauman 1992; Luckmann 1995; Thibault 2003; discussed in section 4), are far from elucidating this bi-directional possibility, which is an important element of the micro-macro interface. See van Dijk (1980) and Mey (1993). Levinson (1979) uses the term activity type. The most frequent, however, is the term genre (cf. fn. 13). Recall (section 4) that “[genre] specifies [multiple] typical ways in which lexicogrammatical resources are deployed to enact a communicative goal”, so even such short one-utterance forms as (4) apply, since the ways in which they could be supported by other forms (recall from the above the potentially infinite catalogue of expressions such as “it’s getting late”!) are controlled by common, conventional, and thus generic constraints. Unlike speech events, which combine into macro speech acts (cf. van Dijk 1980), genres do not presuppose a methodologically superordinate concept. Thus, they constitute a more elegant framework in terms of their ‘downward’ specification, but fall short of describing the infiniteness of consecutive layers in the whole intentionality structure. See mainly van Dijk’s work, from classical contributions to text grammar (cf. van Dijk 1977, 1980), to recent research in critical discourse analysis (cf. e.g. van Dijk 2008).

70

Piotr Cap

See also his joint work with Walter Kintsch (e.g. van Dijk and Kintsch 1983), which in my view represents the best step ever taken to integrate the concepts of ‘discourse’ and ‘text’ in a ‘pragmatic’ analysis. 16. Its protagonists include Emanuel Schegloff, Harvey Sacks, Gail Jefferson, Jonathan Potter; cf. Ten Have (2007) for an overview. 17. The Whitehall Palace address, November 19, 2003. The analysis of (5) shows that macropragmatic studies often proceed in a top-down manner (cf. section 4); they draw on social and cognitive context accessible to the analyst and employ the analyst’s expert knowledge to an a priori formulation of the functional thesis. The source of the analyst’s expert knowledge is not only his/her cognitive experience, but also the data provided by disciplines outside the principal domain of linguistics. Thus, macropragmatics fosters interdisciplinarity, a notable example of which is Critical Discourse Analysis.

References Alston, William 1994 Illocutionary acts and linguistic meaning. In: Savas Tsohatzidis (ed.), Foundations of Speech Act Theory: Philosophical and Linguistic Perspectives, 29–49. London: Routledge. Atlas, Jay 2004 Presupposition. In: Laurence Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 29–52. Oxford: Blackwell. Attardo, Salvatore 1990 The violation of Grice’s maxims in jokes. In: Kira Hall, Michael Meacham and Richard Shapiro (eds.), Proceedings of the 16th Annual Meeting of the Berkeley Linguistics Society, 355–362. Berkeley, CA: University of California Press. Attardo, Salvatore 1993 Violation in conversational maxims and cooperation: The case of jokes. Journal of Pragmatics 19: 532–558. Austin, John 1962/1975 How to Do Things with Words. Oxford: Oxford University Press. Bach, Kent 1999 The myth of conventional implicature. Linguistics and Philosophy 22: 327–366. Bateson, Gregory 1972 Steps of an Ecology of Mind. Toronto: Chandler Publishing Company. Bauman, Richard 1992 Contextualization, tradition, and the dialogue of genres: Icelandic legends of the kraftskald. In: Alessandro Duranti and Charles Goodwin (eds.), Rethinking Context: Language as an Interactive Phenomenon, 125–146. Cambridge: Cambridge University Press. Bazzanella, Carla 2002 The significance of context in interpretation: The we case. In: Bruce Edmonds and Varol Akman (eds.), Context in Context. Special issue of Foundations of Science 7, 239–254.

Micropragmatics and macropragmatics

71

Beaver, David 2001 Presupposition and Assertion in Dynamic Sentences. Stanford: CSLI. Bertolet, Rod 1994 Are there indirect speech acts? In: Savas Tsohatzidis (ed.), Foundations of Speech Act Theory: Philosophical and Linguistic Perspectives, 335–349. London: Routledge. Bezuidenhout, Anne 1997 Pragmatics, semantic underdetermination and the referential/attributive distinction. Mind 106: 375–409. Bezuidenhout, Anne 2002 Truth-conditional pragmatics. Philosophical Perspectives 16: 105–134. Biber, Douglas 1988 Variation across Speech and Writing. Cambridge: Cambridge University Press. Blakemore, Diane 1992 Understanding Utterances: An Introduction to Pragmatics. Oxford: Blackwell. Bublitz, Wolfram 2003 Emotive prosody: how attitudinal frames help construct context. In: Ewald Mengel, Hans-Jörg Schmid and Michael Steppat (eds.), Anglistentag 2002 Bayreuth, Proceedings, 381–391. Trier: Wissenschaftlicher Verlag. Bublitz, Wolfram 2006 It utterly boggles the mind: knowledge, common ground and coherence. In: Hanna Pishwa (ed.), Language and Memory: Aspects of Knowledge Representations, 359–386. Berlin: de Gruyter. Bühler, Karl 1934/1982 Sprachtheorie. Die Darstellungsfunktion der Sprache. Stuttgart/New York: Fischer. Cap, Piotr 2008 Towards the proximization model of the analysis of legitimization in political discourse. Journal of Pragmatics 40: 17–41. Cap, Piotr 2010 Axiological aspects of proximization. Journal of Pragmatics 42: 392–407. Carston, Robyn 1996 Metalinguistic negation and echoic use. Journal of Pragmatics 25: 309–330. Carston, Robyn 2002 Thoughts and Utterances. The Pragmatics of Explicit Communication. Oxford: Blackwell. Cole, Peter (ed.) 1978 Syntax and Semantics 9: Pragmatics. London: Academic Press. Cummings, Louise 2005 Pragmatics: A Multidisciplinary Perspective. Edinburgh: Edinburgh University Press. Davis, Wayne 1998 Implicature: Intention, Convention, and Principle in the Failure of Gricean Theory. Cambridge: Cambridge University Press.

72

Piotr Cap

Duranti, Alessandro and Charles Goodwin (eds.) 1992 Rethinking Context. Language as an Interactive Phenomenon. Cambridge: Cambridge University Press. Elugardo, Reinaldo and Robert Stainton 2004 Shorthand, syntactic ellipsis, and the pragmatic determinants of what is said. Mind and Language 19: 442–71. Fairclough, Norman 1995 Critical Discourse Analysis: The Critical Study of Language. London: Longman. Fetzer, Anita 2002 Communicative intentions in context. In: Anita Fetzer and Christiane Meierkord (eds.), Rethinking Sequentiality: Linguistics Meets Conversational Interaction, 37–69. Amsterdam: Benjamins. Fetzer, Anita (ed.) 2007 Context and Appropriateness: Micro Meets Macro. Amsterdam: Benjamins. Fillmore, Charles 1997 Lectures on Deixis. Stanford: CSLI. Frege, Gottlob 1892 Über Sinn und Bedeutung. Zeitschrift für Philosophie und Philosophische Kritik 100: 25–50. Trans. as ‘On sense and reference’. In: Peter Geach and Max Black (eds.), 1952 Translations from the Philosophical Writings of Gottlob Frege. Oxford: Blackwell. Gazdar, Gerald 1979 Pragmatics: Implicature, Presupposition and Logical Form. London: Academic Press. Goffman, Erving 1986 Frame Analysis. Boston: Northeastern University Press. Green, Georgia 1989 Pragmatics and Natural Language Understanding. Hillsdale, NJ: Lawrence Erlbaum. Grice, Paul 1975 Logic and conversation. In: Peter Cole and Jerry Morgan (eds.), Syntax and Semantics 3: Speech Acts, 41–58. New York: Academic Press. Grice, Paul 1981 Presupposition and conversational implicature. In: Peter Cole (ed.), Radical Pragmatics, 183–198. New York: Academic Press. Grice, Paul 1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press. Grundy, Peter 1995 Doing Pragmatics. London: Edward Arnold. Gumperz, John 1992 Contextualization and understanding. In: Alessandro Duranti and Charles Goodwin (eds.), Rethinking Context: Language as an Interactive Phenomenon, 229–252. Cambridge: Cambridge University Press. Gumperz, John 2003 Response essay. In: Susan Eerdmans, Carlo Prevignano and Paul Thibault (eds.), Language and Interaction. Discussions with John J. Gumperz, 105– 126. Amsterdam: Benjamins.

Micropragmatics and macropragmatics

73

Halliday, Michael and Ruqaiya Hasan 1976 Cohesion in English. London: Longman. Halliday, Michael and Ruqaiya Hasan 2000 System and text: making links. Text 20: 201–210. Holdcroft, David 1994 Indirect speech acts and propositional content. In: Savas Tsohatzidis (ed.), Foundations of Speech Act Theory: Philosophical and Linguistic Perspectives, 350–364. London: Routledge. Horn, Laurence 1984 Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature. In: Deborah Schiffrin (ed.), Meaning, Form, and Use in Context: Linguistic Applications, 11–42. Washington, DC: Georgetown University Press. Horn, Laurence 1996 Presupposition and implicature. In: Shalom Lappin (ed.), The Handbook of Contemporary Semantic Theory, 299–320. Oxford: Blackwell. Horn, Laurence and Gregory Ward (eds.) 2004 The Handbook of Pragmatics. Oxford: Blackwell. Horn, Laurence 2007 Toward a Fregean pragmatics: Voraussetzung, Nebengedanke, Andeutung. In: Istvan Kecskes and Laurence Horn (eds.), Explorations in Pragmatics: Linguistic, Cognitive and Intercultural Aspects, 39–69. Berlin: Mouton de Gruyter. Horn, Laurence 2009 WJ-40: Implicature, truth, and meaning. International Review of Pragmatics 1: 3–34. Huang, Yan 2007 Pragmatics. Oxford: Oxford University Press. Huang, Yan 2009a Neo-Gricean pragmatics and the lexicon. International Review of Pragmatics 1: 118–153. Huang, Yan 2009b Neo-Gricean pragmatic theory of conversational implicature. In: Bernd Heine and Heiko Narrog (eds.), The Oxford Handbook of Linguistic Analysis, 669–696. Oxford: Oxford University Press. Hymes, Dell 1974 Foundations in Sociolinguistics: An Ethnographic Approach. Philadelphia: University of Philadelphia Press. Karttunen, Lauri 1973 Presuppositions of compound sentences. Linguistic Inquiry 4: 169–193. Lakoff, George and Mark Johnson 1999 Philosophy in the Flesh: The Embodied Mind and its Challenge to Western Thought. New York: Basic Books. Langacker, Ronald 1987 Foundations of Cognitive Grammar. Vol. I: Theoretical Prerequisites. Stanford, CA: Stanford University Press. Langacker, Ronald 2001 Discourse in Cognitive Grammar. Cognitive Linguistics 12: 143–188.

74

Piotr Cap

Levinson, Stephen 1979 Activity types and language. Linguistics 17: 365–399. Levinson, Stephen 1983 Pragmatics. Cambridge: Cambridge University Press. Levinson, Stephen 2000 Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Linell, Per 1998 Approaching Dialogue. Amsterdam: Benjamins. Luckmann, Thomas 1995 Interaction planning and intersubjective adjustment of perspectives by communicative genres. In: Esther Goody (ed.), Social Intelligence and Interaction: Expressions and Implications of the Social Bias in Human Intelligence, 175– 188. Cambridge: Cambridge University Press. Lyons, John 1995 Linguistic Semantics. Cambridge: Cambridge University Press. Mann, William and Sandra Thompson 1983 Relational propositions in discourse. Technical Report ISI/RR-83–115. University of Southern California: Information Sciences Institute. Manning, Paul 2001 On social deixis. Anthropological Linguistics 43: 54–100. Marmaridou, Sophia 2000 Pragmatic Meaning and Cognition. Amsterdam: Benjamins. Mey, Jacob L. 1993 Pragmatics. Oxford: Blackwell. Nunberg, Geoffrey 1993 Indexicality and deixis. Linguistics and Philosophy 16: 1–44. Recanati, François 1993 Direct Reference: From Language to Thought. Oxford: Blackwell. Recanati, François 2001 What is said. Synthese 128: 75–91. Recanati, François 2004 Literal Meaning. Cambridge: Cambridge University Press. Saeed, John 2009 Semantics. 3rd edition. Oxford: Blackwell. Sadock, Jerrold 1978 On testing for conversational implicature. In: Peter Cole (ed.), Syntax and Semantics 9: Pragmatics, 281–298. London: Academic Press. Sadock, Jerrold 2004 Speech acts. In: Laurence Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 53–73. Oxford: Blackwell. Saville-Troike, Muriel 1989 The Ethnography of Communication. Oxford: Blackwell. Searle, John 1975 A taxonomy of speech acts. In: Keith Gunderson (ed.), Minnesota Studies in the Philosophy of Science 9: 344–369.

Micropragmatics and macropragmatics

75

Searle, John 1979 Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press. Soames, Scott 1989 Presupposition. In: Dov Gabbay and Franz Guenthner (eds.), Handbook of Philosophical Logic, 553–616. Dordrecht: Reidel. Sperber, Dan and Deirdre Wilson 1986 Relevance: Communication and Cognition. Oxford: Blackwell. Stalnaker, Robert 1973 Presuppositions. Journal of Philosophical Logic 2: 447–457. Ten Have, Paul 2007 Doing Conversation Analysis. A Practical Guide. Thousand Oaks: Sage. Strawson, Peter 1952 Introduction to Logical Theory. London: Methuen. Thibault, Paul 2003 Contextualization and social meaning-making practices. In: Susan Eerdmans, Carlo Prevignano and Paul Thibault (eds.), Language and Interaction. Discussions with John J. Gumperz, 41–62. Amsterdam: Benjamins. Thomas, Jenny 1995 Meaning in Interaction: An Introduction to Pragmatics. London: Longman. Tsohatzidis, Savas (ed.) 1994 Foundations of Speech Act Theory: Philosophical and Linguistic Perspectives. London: Routledge. Van Dijk, Teun 1977 Text and Context. London: Longman. Van Dijk, Teun 1980 Macrostructures. An Interdisciplinary Study of Global Structures in Discourse, Interaction, and Cognition. Hillsdale, NJ: Erlbaum. Van Dijk, Teun 1995 Discourse, semantics and ideology. Discourse and Society 6: 243–259. Van Dijk, Teun 2002 Ideology: Political discourse and cognition. In: Paul Chilton and Christina Schaeffner (eds.), Politics as Text and Talk, 33–57. Amsterdam: Benjamins. Van Dijk, Teun 2008 Discourse and Context. A Sociocognitive Approach. Cambridge: Cambridge University Press. Van Dijk, Teun and Walter Kintsch 1983 Strategies of Discourse Comprehension. New York: Academic Press, Verschueren, Jef 1999 Understanding Pragmatics. London: Edward Arnold. Wodak, Ruth and Paul Chilton (eds.) 2005 A New Agenda in (Critical) Discourse Analysis: Theory, Methodology and Interdisciplinarity. Amsterdam: Benjamins.

3.

Pragmalinguistics and sociopragmatics Sophia Marmaridou

1.

Introduction

Within a broad definition of pragmatics as the scientific study of all aspects of linguistic behaviour, the distinction between pragmalinguistics and sociopragmatics is intended to focus on two methodological approaches to pragmatic analysis. Pragmalinguistics typically concerns the study of the particular resources that a given language provides for conveying pragmatic meaning (illocutionary and interpersonal), whereas sociopragmatics relates pragmatic meaning to an assessment of participants’ social distance, the language community’s social rules and appropriateness norms, discourse practices, and accepted behaviours. This methodological distinction, initially launched by Leech (1983) and Thomas (1983) and widely adopted in subsequent work in pragmatics, is distinct from, though not totally unrelated to, earlier work in Marxist pragmatics, in which the term pragmalinguistics essentially marks the pragmatic turn in linguistic analysis and is broadly associated, or even identified, with sociolinguistics, stylistics, or text linguistics (see Mey 1979; Prucha 1983). Moreover, the term sociopragmatics has been used in contrast to psychopragmatics (Dascal and Françozo 1989), also independently from the previous distinction just mentioned. In this paper I shall argue that, as defined above, the distinction between pragmalinguistics and sociopragmatics suggests a conceptual dichotomy in our understanding of language either as a system used for communication purposes, or as an action domain reflecting, or co-constructing, social order in culturally sanctioned ways. Regardless of whether this conceptual dichotomy is systematically reflected in the methodological distinction it motivates, the latter has been particularly productive. In fact, it has both affected the internal organisation and theoretical refinement of other concepts, such as pragmatic failure and pragmatic transfer, and also informed particular methodological approaches to second language teaching and testing. Moreover, it has been of particular relevance to historical pragmatics and historical corpus linguistics. The apparent usefulness of the distinction between pragmalinguistics and sociopragmatics seems to be at odds with the fact that the borderline between the two is often fuzzy, to the extent that they are often thought of as the two end points of a continuum. Given the variety of studies straddling different points along this continuum, I propose to use the figure/ground schema afforded us by psychology to further argue that the two concepts under investigation essentially reflect a methodological choice in current practices to focus on one area of pragmatic research by conferring background status to another. The use of the

78

Sophia Marmaridou

figure/ground relationship in this context seeks to explain the bibliographically attested fluidity of the content of these concepts and to present them as distinct, but interacting, instances of perspectivisation. It has become obvious from the above that the discussion of the concepts underlying the terms under investigation will adopt a “historical” approach to their analysis. This is expected on the one hand to place pragmalinguistics and sociopragmatics in the wider framework of pragmatic study that was flourishing in the two crucial decades of its theoretical formation, and, on the other, to show how these concepts have gradually permeated more recent trends in pragmatic study. In the next section reference will be made to the ways in which the terms pragmalinguistics and sociopragmatics have been used independently of each other and prior to the methodological distinction they have been mainly associated with as introduced by Leech (1983) and Thomas (1983). In particular, the term pragmalinguistics as used in Marxist linguistics in the late seventies and early eighties will be shown to be broadly equivalent to pragmatic linguistics or societal pragmatics (Mey 1993). In this framework, the term covers research which emphasises functional aspects of language use within various paradigms (Marxist, or otherwise). Moreover, sociopragmatics, as a theory of the ways in which the non-linguistic environment affects language use, will be contrasted with psychopragmatics as a theory of how the linguistic environment affects thought (Dascal and Françozo 1989). In the third section the distinction between pragmalinguistics and sociopragmatics will be discussed from the perspective of Leech (1983) and Thomas (1983). More specifically, the conceptual underpinnings of this distinction will be set off against established research practices concerning issues of contrastive linguistics, pragmatic failure, pragmatic transfer, pragmatic development, and the teaching and testing of pragmatics in second language learning settings. In the fourth section the contribution of the pragmalinguistic and sociopragmatic perspective to historical pragmatics and historical corpus linguistics will be examined with a view to highlighting more recent trends in pragmatic research. In the last section of this paper an overall assessment of the concepts of pragmalinguistics and sociopragmatics will be attempted and their distinction will be evaluated with respect to the fields of study in which it has been adopted.

2.

The borders between pragmalinguistics and sociopragmatics

In order to appreciate the conceptual dichotomy encoded in the terms pragmalinguistics and sociopragmatics, one should be reminded of the two main trends that have dominated the field of pragmatics ever since its inception as a distinct, established field of linguistic enquiry. In the Anglo-American tradition, pragmatics is mainly concerned with speaker’s meaning as intentionally communicated to an addressee. As Jucker (2008: 894) observes, pragmatics in this tradition is “a rescue

Pragmalinguistics and sociopragmatics

79

plan for syntax and semantics that deals with a systematic account of implicit meanings and deals with intentional human action”. Levinson (1983: 7–32), while espousing this tradition, has identified, discussed, and amply demonstrated problems of defining pragmatics in this narrow sense, to conclude that the best way to delimit this field of study is to actually observe what practitioners do, or provide a list of the phenomena which a pragmatic theory must account for, namely deixis, implicature, presupposition, speech acts, and aspects of discourse structure (Levinson 1983: 27). He also points out that “the upper bound of pragmatics is provided by the borders of semantics and the lower bound by sociolinguistics (and perhaps psycholinguistics too)”, and admits that “drawing a boundary between sociolinguistic and pragmatic phenomena is likely to be an exceedingly difficult enterprise” (Levinson 1983: 29). Within the tradition of Levinson’s “topical” definition of pragmatics, the term pragmalinguistics has been used in the area of literary linguistics, or stylistics, to refer to the study of Gricean maxims, speech acts and politeness phenomena in a literary work. An example is Haverkate’s (1994) study of the dialogues of Don Quixote de la Macha in Cervantes’s well known novel. Apparently, this pragmalinguistic frame of reference provides new insights into the interactional roles and personality traits of Don Quixote and Sancho Panza, which traditional literary theory does not typically address. The Anglo-American perspective on pragmatics has been criticised as leaving out, or backgrounding, “the larger social context of communication, which is of central importance in the European tradition” (Jucker 2008: 895). The latter derives more directly from Morris’s (1938) definition of semiotics, distinguishing between syntactics as the formal relation of signs to one another, semantics as the relation of signs to the objects to which the signs are applicable, and pragmatics as the relation of signs to interpreters. In Morris’s account, the user/speaker is seen in a wider perspective, since the human being, as an interpreter of signs, does not only have a mental identity, but also a biological and social identity, which affects her interpretation of signs (Marmaridou 2000: 18). In fact, Morris acknowledges the dependence of semiosis on the human being as an entity that is simultaneously the source and the subject of all psychological, biological and sociological phenomena that have to do with the functioning of signs comprising language use: It is a sufficiently accurate characterization of pragmatics to say that it deals with the biotic aspects of semiosis, that is, with all the psychological, biological, and sociological phenomena which occur in the functioning of signs. (Morris 1938: 108)

Within the context of these two traditions in pragmatics, the terms pragmalinguistics and sociopragmatics have both been used to refer broadly to pragmatic phenomena that characterise the use of language for communication purposes. In what follows, it will be shown that, while pragmalinguistics has been associated with a societal and Marxist perspective on language use, sociopragmatics generally refers

80

Sophia Marmaridou

to the study of the external pragmatic factors affecting language use. In this sense, sociopragmatics has also been juxtaposed to psychopragmatics, the latter focusing on internal, cognitive operations such as reasoning and problem solving. It should be pointed out that pragmalinguistics and sociopragmatics, as defined above, constitute two different and unrelated projects. Therefore they do not reflect the distinction and complementariness they have been associated with in the work of other scholars, such as Leech (1983) and Thomas (1983) (see section 3). Rather, the use of these two terms in the frameworks discussed below is indicative of the various developments of pragmatics within the two decades of its initial expansion. Focusing on sociological phenomena, Mey (1979) uses the term pragmalinguistics to redefine the object and goal of linguistic enquiry in a Marxist framework as the study of language (the linguistic object) in a variety of pragmatic settings. In his analysis he identifies pragmatics with the description of action in general and pragmatic linguistics (or pragmalinguistics) with the study of linguistic action as determined by the social conditions for speaking and understanding. As he claims, “from a pragmatic viewpoint, the conditions for speaking and understanding, for ‘production’ and ‘consumption’ of language, cannot be divorced from the conditions of production and consumption in the society in general” (Mey 1979: 11). The totality of conditions that are active in the production and consumption of texts, including co-text, constitute the context, which is dynamic, rather than static, and creative, rather than passive. Following Mey’s understanding of pragmalinguistics, Olsen (1979) raises the issue of linguistic meaning as located in, and defined by, language use in its social context. As he observes, the meanings of words and utterances were not universals which merely were differently labelled in different languages. The meanings were dependent on, and part of, the different social systems […]. Utterances are produced and understood within the definition of the situation and it is the rules by which we define the situation that are shared by the members of a social system (Olsen 1979: 247).1

That linguistic meaning is a function of language use and that it is being constructed all the time by the process of interaction are facts that, according to Olsen, must be accounted for by pragmatic linguistics. It transpires that, in his analysis, pragmalinguistics is to be broadly understood as the pragmatic turn, or a pragmatic perspective on linguistics as informed by social theory. As already mentioned, pragmalinguistics is also understood to generally cover the study of functional aspects of language use as practised in Eastern Europe in the late seventies. The relevant studies spread over the areas of functional stylistics, text linguistics, rhetorics, psycholinguistics, sociolinguistics, social communication theory, and semiotics (see Prucha 1983). Within such a broad framework, pragmalinguistics emerges as a theory of the pragmatic component of a grammar and as a theory of verbal activity in social intercourse. These two per-

Pragmalinguistics and sociopragmatics

81

spectives are reflected in the analysis of the meaning of utterances as comprising contextual and situational aspects respectively, as in Bang and Døør’s (1979) work. Contextual meaning is determined by the lexicon and syntax, style and genre, and textual implications, and it is to a high degree reproducible and general. Situational meaning expresses an interpretation of the world, society and the human being, and carries an evaluative orientation. To accept a situational meaning is to subscribe to a certain set of values. In this understanding, contextual meaning belongs to the pragmatic component of grammar, whereas situational meaning develops from verbal activity in social intercourse. These two kinds of utterance meaning are said to have a dialectal relation to one another and to define a continuum ranging from the unique and unrepeatable (situational) to the general and reproducible (contextual) (Bang and Døør 1979: 46). It transpires from the above that pragmalinguistics in the Eastern European and Marxist traditions encompasses aspects of language use which range from grammatically and textually encoded elements to socially constructed discursive practices. The emphasis is typically placed on the latter, which figures prominently in this tradition against the background of the grammatical and textual aspects of language. This theoretical turn is to be interpreted both as a reaction to explicitly formal accounts of language in the various structural approaches of the time (see, for example, the work of Chomsky), and as an attempt to articulate an ideological stance towards the epistemological foundations of linguistics. The realisation that utterance meaning, as the domain proper of pragmalinguistic investigation, is a function of linguistic means in interaction with, or in a dialectal relation to, the social conditions of language use creates a conceptual frame for what is later to be established as a distinction between pragmalinguistics and sociopragmatics, albeit outside explicitly stated ideological agendas. Apart from the societal perspective on language and the association of pragmatics with the social conditions of language use, the cognitive perspective on language contributes to its construal as an instrument of thought. The latter line of investigation has led to the use of the terms sociopragmatics and psychopragmatics to carry the distinction between the use of language for communication purposes and the use of language as an instrument of thought, respectively. More specifically, in their attempt to shed light on the pragmatic turn in psycholinguistics, Dascal and Françozo (1989) refer to pragmatics as the field of study concerned with “conversational maxims, politeness requirements, contextual circumstances of utterances, and other factors that impose constraints on the selection and understanding, in a given context, of a particular linguistic expression for a certain cognitive content” (Dascal and Françozo 1989: 5). In many cases, the pragmatic turn in psycholinguistics amounts to the study of the acquisition of pragmatics by young children and, in particular, the study of how children acquire and perform speech acts and how they establish interpersonal communicative relations (Bates 1976). However, Dascal and Françozo (1989) observe that language plays an important

82

Sophia Marmaridou

role in our mental processes and higher level cognitive processes or states, such as reasoning, problem solving, storing and retrieving information, believing, etc. They further claim that the use of language for such purposes also belongs to pragmatics: we use language not only to communicate, but also to do our reasoning, to solve problems, etc. They therefore use the term psychopragmatics to refer to the study of such mental use of language and distinguish it from the study of its use for communicative purposes, i.e. from sociopragmatics. By the former term they refer to internal pragmatic factors that affect the performance of the cognitive operations themselves which in some way involve language. By sociopragmatics they refer to external pragmatic factors that concern the perception and the production of linguistic signs in a particular situation, such as indirectness in the performance of speech acts. It follows that psychopragmatics is a theory of the way in which “the linguistic environment of thought influences thought, whereas sociopragmatics is a theory of the ways in which the non-linguistic environment (which includes both thought and situation) affects language interpretation and production” (Dascal and Françozo 1989: 13). The proposed distinction between sociopragmatics and psychopragmatics serves to compare and contrast language use in jokes and dreams, topics mainly addressed in the framework of Freud’s writings (see Freud 1900; 1905). More specifically, Dascal (1985: 98) argues that jokes depend for their effectiveness on the existence of the sociopragmatic phenomenon of indirectness; they are intended to be understood as such by an audience and therefore perform a social function. As opposed to jokes, dreams are mental objects, not meant to be communicated to an audience, i.e. they have an asocial character. Therefore, they are not socially constrained, their only constraint being the existence of some kind of associative link, perhaps arbitrary and fortuitous, that provides some path for the expression of a problematic content. In this respect “the psychopragmatic role of language in dreams is […] similar to its use for another cognitive purpose namely, as a mnemonic device” (Dascal 1985: 104). In this analysis, Dascal’s sociopragmatics covers the broader area of general pragmatics in both the Anglo-American and the European traditions with the aim of distinguishing it from psychopragmatics, the latter illustrating a pragmatic turn in psycholinguistics.

3.

The pragmalinguistics/sociopragmatics distinction in a cross-linguistic and cross-cultural perspective

The distinction between pragmalinguistics and sociopragmatics was essentially established within the framework of applied linguistics studies on which it was to exert a strong influence. It was initially introduced in the work of Thomas (1981), further explored in Thomas (1983), and explicitly spelt out as a methodological approach in pragmatics in Leech (1983). In the latter’s programme, General Prag-

Pragmalinguistics and sociopragmatics

83

matics is that aspect of human language which concerns the use of language in communication. More specifically, the understanding of the nature of language depends both on the study of grammar, as the abstract, formal system of language (comprising phonology, syntax and semantics), and on the study of pragmatics, as the interaction of the Cooperative and Politeness Principles partly characterizing language use. Moreover, Leech (1983: 7) views semantics as a level of contact between grammar and pragmatics, so that linguistic meaning can be fruitfully studied by a combination of semantic and pragmatic approaches, the former targeting the sense of an utterance and the latter its force. Within Leech’s project, General Pragmatics is a fairly abstract study of the general conditions of the communicative use of language, excluding more specific ‘local’ conditions on language use: “The latter may be said to belong to a less abstract field of SOCIO-PRAGMATICS, for it is clear that the Cooperative Principle and the Politeness Principle operate variably in different cultures or language communities, in different social situations, among different social classes, etc.” (Leech 1983: 10). Alongside sociopragmatics, Leech introduces PRAGMALINGUISTICS as “the study of the more linguistic end of pragmatics – where we consider the particular resources which a given language provides for conveying particular illocutions” (Leech 1983: 11). It follows that sociopragmatic studies are culture-specific, while pragmalinguistic studies are language-specific. Apparently, Leech’s observation that sociopragmatics is the sociological interface of pragmatics is implicitly an attempt to address an issue that Levinson (1983: 29) also raises when he observes that the boundary between sociolinguistic and pragmatic phenomena is hard to define. Since Leech (1983) is concerned with establishing the Cooperative and Politeness Principles as the two cornerstones of General Pragmatics and with arguing their universality, he does not pursue the pragmalinguistics/sociopragmatics distinction any further. Nevertheless, he views both principles as the natural constraints of universality. For example, while claiming the universality of the principles, he concedes that they do not apply in an identical manner in all societies: Indeed, one of the main purposes of sociopragmatics […] is to find out how different societies operate maxims in different ways, for example by giving politeness a higher rating than cooperation in certain situations, or by giving precedence to one maxim of the Politeness Principle rather than another (Leech 1983: 80).

Similarly, he observes that cultural stereotyping (e.g. the British are more tactful that the Americans) is related to “pragmalinguistic strategies such as strategies of indirectness, and the norms observed in the performance of these strategies in different speech communities” (Leech 1983: 231). Leech’s (1983) distinction between pragmalinguistics and sociopragmatics is the natural outcome of his claim about the universality of the principles of pragmatics he proposes. By prioritizing conversational principles, he proposes a rhe-

84

Sophia Marmaridou

torical model of pragmatics referring to the study of the effective use of language in communication and focusing on the goal-oriented speech situation, whereby the speaker uses language in order to produce a particular effect in the mind of the hearer. Leech’s principles of general pragmatics are to be held constant and form the basis, or the ground, against which all other language- and culture-specific pragmatic variables can be studied as distinct figures. This position lends itself to cross-linguistic and cross-cultural comparisons of communicative behaviour with obvious applications for language teaching/learning and testing. Moreover, the distinction proposed within this framework seems compatible with the aims and methods of the study of discourse and historical pragmatics in particular. In what follows, these areas of investigation will be related to the conceptual and methodological distinction between pragmalinguistics and sociopragmatics as introduced in this section.

4.

Applications of the pragmalinguistics/sociopragmatics distinction

4.1.

Preliminary investigations in contrastive linguistics

The impact of the pragmalinguistics/sociopragmatics distinction has been particularly felt in the conceptual organisation of some theoretical constructs it has motivated, such as pragmatic failure, pragmatic transfer, and pragmatic development. This development in the area of applied linguistics is to some extent foreshadowed by Riley (1979) advocating a contrastive pragmalinguistics. Taking the broad view of pragmalinguistics as pragmatic linguistics, Riley (1979) defines its aim as the study of communicative acts. These may be realised verbally, paralinguistically, or non-verbally, and rest on a theory of illocution; illocutionary force is extended in this system to all kinds of communicative acts. Focusing on linguistic acts, he notes that what is central to contrastive pragmalinguistics is “language functions rather than linguistic structures, discourse, not grammar, the communicative act in context, not the sentence in isolation” (Riley 1979: 57). Drawing in this way a methodological line between form and function, he highlights instances of same form and different function (e.g. the possibility of using “You’re not going out” to express the illocutions of prohibiting, confirming, threatening, expressing surprise, or stating), and instances whereby the same function is performed by different forms (e.g. the function of agreeing as expressed by the forms “Yes, of course”, “I agree”, nodding, or repetition of the previous speaker’s utterance). In his model, form-function pairings constitute the ground against which cross-linguistic contrasts are to be drawn. In order to place this model in a wider framework encompassing all aspects of pragmatic study, Riley (1979) further proposes cross-linguistic contrasts across what he calls discourse structure, consisting of three sub-levels of structure:

Pragmalinguistics and sociopragmatics

85

Firstly, formal structure includes all aspects of message-bearing elements, e.g. verbal, paralinguistic, or non-verbal, in a particular situation. Secondly, illocutionary structure consists of sequences of illocutionary acts, e.g. inviting – accepting, confirming – thanking, etc. Thirdly, interactive structure refers to linguistic organization in terms of interactional tactics such as turns (e.g. opening, reply, closing, address, exchange, transaction, etc.). Evidently, Riley’s work is an attempt to include in his model all topics of pragmatic interest and put them to the forefront of contrastive linguistics. However, his efforts to systematise the proposed model lead to a schematic and idealised understanding of communicative acts. For example, the sequences of illocutionary acts seem to refer to idealised prototypes. In reality, invitations are typically, but not always, followed by acceptances, or refusals, confirmations by thanks, etc. This apparent theoretical weakness of Riley’s model, though, is counterbalanced by its applicability in contrastive analysis and language teaching in particular. More specifically, the model allows for the identification and the teaching of interactional styles and illocutionary point in relation to sociolinguistic variables such as roles, status, formality, and discursive control, as for example, in a typical teacher – student illocutionary sequence, or in a typical asking-for-street-instructions exchange between strangers. In this way, Riley’s proposal addresses newly created needs in the methodology of communicative language teaching research that was being carried out at the time (see, e.g., Candlin, 1976; Brumfit and Johnson, 1979; Widdowson, 1979). Perhaps the most important contribution of Riley’s work in introducing contrastive pragmalinguistics is that he creates an awareness of the significance of pragmatic knowledge to language teaching and emphasises the systematicity of such knowledge so that it can be actually taught. His suggestions clearly encourage a methodological turn, to be further explored in the context of the pragmalinguistics/sociopragmatics distinction introduced by Thomas (1981; 1983) and Leech (1983). 4.2.

Pragmatic failure

Following Leech’s (1983) division of linguistics into grammar and pragmatics and adopting his goal-oriented understanding of language use (see above), Thomas (1983) first construes linguistic competence as consisting of grammatical competence (abstract knowledge of phonology, syntax and semantics) and pragmatic competence as “the ability to use language effectively in order to achieve a specific purpose and to understand language in context” (Thomas 1983: 92). Within this framework, pragmatic failure is defined as “an inability to recognise the force of the speaker’s utterance when the speaker intended that this particular hearer should recognise it” (Thomas 1983: 94). Notably, whereas grammatical error indicates the speaker’s restricted linguistic proficiency, pragmatic failure reflects badly on her as a person. It follows that, in the context of cross-cultural communication, pragmatic

86

Sophia Marmaridou

failure may lead to a misunderstanding of the foreign speaker’s intentions. Awareness-raising tasks aiming to develop foreign students’ metapragmatic ability to consciously analyse language use are commendable in this case, but inadequate in Thomas’s (1983: 99) view, unless informed by the distinction between pragmalinguistic and sociopragmatic failure. The proposed categorisation of types of pragmatic failure, consistent with Leech’s (1983) dyad of language vs. culture specificity, is not intended to impose rigid boundaries. Thomas (1983: 101) explicitly states that the two concepts form a continuum from what is language-specific to what is culture-specific. Pragmalinguistic failure results from the speaker’s mapping of pragmatic force to an utterance in a way that is systematically different from a native speaker’s. Sociopragmatic failure results from the speaker’s miscalculation of the social conditions placed on language in use, such as the size of the speaker’s imposition on the addressee in performing a particular speech act, a cost/benefit scale of acts in the particular culture, the social distance between interlocutors, and relative, culture-specific rights and obligations of interlocutors to each other in specific situations. As one moves from the communicative intent of a speaker’s utterance to the interpersonal dynamics between speaker and addressee, one is also moving from the pragmalinguistic to the sociopragmatic end of the continuum. In providing examples of pragmalinguistic failure, Thomas (1983: 101) concedes to the systematicity of form–function pairings in a particular language and culture in the sense that native speakers assign pragmatic force to certain utterances fairly predictably. Unlike Riley (1979), she seems to restrict the scope of this systematicity to highly conventionalised utterances, such as the often-quoted “can you do X” English utterance used to issue a request rather than ask for one’s ability. The fact that in other languages, such as French or Russian, the opposite is true may lead the foreign language learner to an inappropriate transfer of a form–function pair, and hence to pragmalinguistic failure. A further cause of pragmalinguistic failure may be the inappropriate transfer of speech act strategies from L1 to L2, such as the use of direct vs. indirect forms of issuing requests politely. For example, it is reported that polite, indirect requests for street instructions (e.g. “excuse me, could you please tell me how to …”) would be counterproductive in Russian, while they are pretty standard in English. Hence, the Russian learner of English using direct means to perform a request might be misunderstood as being impolite (Thomas 1983: 102). Politeness is indeed an area in which pragmalinguistic failure merges with sociopragmatic failure. Whereas the choice of polite forms and strategies seems to belong to the pragmalinguistic end of the continuum, when and to whom to be polite concerns the sociopragmatics end. In Thomas’s view, the size of imposition of a speech act, the awareness of taboos in a particular cultural context, and differing assessments of relative power or social distance between interlocutors may lead to sociopragmatic failure. One may add to these sources of sociopragmatic failure the

Pragmalinguistics and sociopragmatics

87

orientation of a culture to positive or negative politeness (see Sifianou 1992). However, it is not always clear whether an instance of pragmatic failure can be assigned to the pragmalinguistic or to the sociopragmatic level. An example that Thomas provides has to do with the conception of what constitutes ‘free goods’ in the former Soviet Union as opposed to Britain. Asking for a cigarette in the Soviet Union requires a minimal degree of politeness, as these items are ‘free goods’ in this cultural setting. A Russian asking for a cigarette in Britain and using a similar strategy may have either encoded wrongly the ‘correct’ amount of politeness she intended, or misjudged the amount of imposition of her request. The former case would represent an instance of pragmalinguistic, while the latter of sociopragmatic failure. The effect of either of these two types of pragmatic failure may be the same for the addressee and their causes difficult for the analyst to distinguish. However, the distinction is essential in the context of language teaching, as it reflects two different types of decision-making. As Thomas observes, pragmalinguistic decisions are language-specific and the teacher can correct them in a fairly straightforward manner. Sociopragmatic ones are culture-specific and sensitive to the learners’ own system of values; hence “they should not be ‘corrected’, but only pointed out and discussed” (Thomas 1983: 109). It is therefore clear that the discussion of pragmatic failure in the context of language teaching relies heavily on the distinction between pragmalinguistics and sociopragmatics. However, given that sociopragmatic decisions are social before they are linguistic (Thomas 1983: 104), the importance of the distinction between sociopragmatics and sociolinguistics becomes apparent. The identification of social roles and stereotypes, power relations between interlocutors relating to age, social status, or dialect, socioculturally-based evaluations of situations, and linguistic value judgements have been typically considered social variables and have traditionally concerned sociolinguistic research (see, e.g. Holmes 2008). How are all these to be differently researched in sociopragmatics? Indeed, this question is often raised, at least implicitly, in the relevant literature and is differently approached (see the discussion in Bou-Franch and Garcés-Conejos 2003; Roever 2006; Archer and Culpeper 2003, below). It transpires that the application of the pragmalinguistics/sociopragmatics distinction to pragmatic failure on the one hand enriches language teaching methodology, but on the other, raises the issue of the theoretical distinction between sociopragmatics and sociolinguistics, to be addressed in the next section. 4.3.

Pragmatic transfer

The pragmalinguistic/sociopragmatics distinction is also reflected in the study of pragmatic transfer within the wider context of second language acquisition research. The relevant literature strongly supports the position that non-native speakers’ understanding and production of linguistic action is influenced by their L1

88

Sophia Marmaridou

pragmatic knowledge (Kasper 1992; Kasper and Rose 2002). Prominent in this kind of research is the concept of interlanguage as a stage on a continuum representing the second language learner’s development towards the L2 system. In Selinker’s (1972) definition, interlanguage is a psycholinguistic phenomenon which mainly refers to the learning process, rather than the outcome of this process. However, the dominant practice in interlanguage pragmatics concerns the outcome, i.e. the collection and comparison of data from native speakers’ linguistic performance in the target language, the learner’s use of her native language, and the learner’s interlanguage (i.e. her use of the target language at a particular stage of development).2 In this context, pragmatic transfer concerns the possible influences from the learner’s L1 on her interlanguage performance. Even though in practice pragmatic transfer has been typically associated with transfer of speech act knowledge, indirectness, and politeness, it may also make reference to sociolinguistic, or discourse parameters in the same instance of language use. For example, transfer of address terms and register concern social variation in language use and are therefore to be considered in the domain of sociolinguistics. Yet, their strategic exploitation in pursuing some illocutionary goal places their investigation in the area of pragmatics proper. Moreover, some speech acts are characterized by their position in discourse: greetings occur discourse-initially and refusals occur as ‘seconds’ in a conversational exchange. Furthermore, the speech act force of an utterance may become evident only after the utterance is responded to by the interlocutor, which points to its negotiable character and conversational grounding. As in the case of pragmatic failure, the pragmalinguistics/sociopragmatics distinction has enabled a more detailed discussion of pragmatic transfer and has added clarity to this concept by maintaining the corresponding form/function dyad. In Kasper’s (1992) definition, “pragmalinguistic transfer shall designate the process whereby the illocutionary force or politeness value assigned to particular linguistic material in L1 influences learners’ perception and production of form-function mappings in L2” (Kasper 1992: 209). However, as shown in the discussion of pragmatic failure, politeness value does not only depend upon linguistic material, but also upon social parameters of the speech event, such as the relative social distance of interlocutors and the power relations between them. For example, L2 learners may be familiar with address terms of varying degrees of deference or intimacy in L2, but their use depends upon the learners’ assessment of social roles, settings, etc., which may be based on the their own culture. Therefore, politeness is not only a candidate for pragmalinguistic, but also for sociolinguistic transfer. As Kasper (1992: 209) observes, the latter is “operative when the social perceptions underlying language users’ interpretation and performance of linguistic action in L2 are influenced by their assessment of subjectively equivalent L1 contexts”. It follows that the form-function pair does not guarantee appropriate constraints on the discussion of pragmalinguistic and sociopragmatic transfer. Whether politeness

Pragmalinguistics and sociopragmatics

89

transfer occurs at the pragmalinguistic or the sociopragmatic level is largely a matter of perspectivisation. If construed as a transfer of (in)appropriate forms, the latter are foregrounded against the background of the social conditions of politeness. If the social conditions of politeness are focused upon, the formal aspects constitute the background to the analysis. The issue of pragmalinguistic and sociopragmatic transfer is also associated with the distinction between positive and negative transfer. Positive transfer of elements from L1 in L2 performance is methodologically difficult to identify, given that acceptable L2 performance may have sources other than positive transfer of L1 elements, for example the application of general pragmatic principles, generalisation of L2 specific knowledge, and teacher input, among others. As Kasper (1992: 213) correctly points out “performance data alone cannot tell us whether […] learners did indeed consult their L1 knowledge or operated strictly on their interlanguage pragmatic competence”. More specifically, positive transfer has been associated with the possibility of pragmatic universals (as in Grice 1975; Leech 1983; Brown and Levinson 1987) that are assumed to function cross-culturally. Interlanguage research has shown that L2 learners have access to the same range of strategies as native speakers in implementing linguistic actions, such as requests, suggestions, invitations, refusals, complaints, compliments, etc. (Wolfson 1981; Trosborg 1987; Beebe and Takahashi 1989; Blum-Kulka 1991; Olshtain and Weinback 1992). This has been taken as evidence of the existence of pragmalinguistic universals. Similarly, L2 learners’ sensitivity to social factors such as interlocutors’ relative status, or degree of imposition, etc., possibly indicates the availability of sociopragmatic universals. However, in spite of the fact that the evidence is based on a variety of languages and cultures, the existence of pragmatic universals at the pragmalinguistic or sociopragmatic level is by necessity an empirical matter, especially since the universality of theoretical concepts in general pragmatics, such as politeness, or the co-operative principle, have been seriously questioned (see, e.g. Keenan 1976; Rosaldo 1982; Wierzbicka 1987; Eelen 2001). Negative pragmalinguistic transfer occurs when inappropriate forms or strategies are used in performing particular speech acts affecting the politeness value of the utterance, or even its illocutionary force. For example, it has been reported that, in expressing refusals, Japanese learners of English use strategies such as generalisations (‘I never yield to temptation’), statements of philosophy (‘to err is human’), or suggestions for alternative action (‘why don’t you ask someone else?’) that are atypical of native English speakers (Beebe et al. 1990). At the sociopragmatic level, L2 learners’ perception of contextual factors may originate in their L1 and affect politeness style. It is reported that Japanese learners of English selected their refusal strategies on the basis of whether the refuser’s status was higher or lower than the interlocutor’s, whereas American native speakers’ choice was made on the basis of status-equal or status-unequal speaker – addressee relationships. In spite of the great number of studies that address the issue of negative pragmatic transfer, the

90

Sophia Marmaridou

systematicity of the relevant experiments, and the careful assessment of their results, L2 learners’ inappropriate performance may also be a function of other factors, such as learning context, shifts in the available target variety of the L2, little exposure to the target language and culture, varying L2 development stages, and lack of contextualisation cues, among others. To the extent that the pragmalinguistics/sociopragmatics distinction shows the need for more precise experimental techniques in the study of pragmatic transfer, it can contribute to the validation of hypotheses and is therefore to be understood as a useful methodological tool in interlanguage pragmatics. 4.4.

Pragmatic development

A further contribution of the pragmalinguistics/sociopragmatics distinction to pragmatic research concerns the issue of the development of grammar and pragmatics in interlanguage research. An important argument in interlanguage pragmatics evolves round the primacy of grammatical or pragmatic knowledge in interlanguage development. While focus on the development of L2 learners’ pragmatics as an autonomous component of communicative competence has had a serious impact on second language acquisition research, its relation to other aspects of communicative ability has also attracted attention (see Bardovi-Harlig 1999 and references therein). In this context, the pragmalinguistics/sociopragmatics distinction serves to investigate what happens when existing grammatical knowledge (such as the semantics of the progressive as a grammaticalised aspectual category) is not put to target-like pragmalinguistic use (e.g. to mitigate imposition), as in “I was wondering if I could have a word” (Kasper and Rose 2002: 180); moreover, this distinction serves to investigate what happens when grammatical knowledge leads to target-like pragmalinguistic use, but non-target-like sociopragmatic use. Robinson (1992) reports that, although a Japanese learner of English had the appropriate knowledge to form and use the “I would like to” form to express preference, she nevertheless used “I want to” on the sociopragmatic assumption that the former form would be “too polite” when addressing her American friend. Several studies show that L2 learners may have knowledge of a particular grammatical structure, which, however, they do not put to native-like pragmalinguistic use. For example, Bodman and Eisenstein (1988) report on L2 learners’ use of sophisticated grammatical knowledge which they put to unconventional L2 pragmalinguistic use, as in expressions of thanking like “May God grant you a long life”. It is hypothesised that such instances also testify to L1 pragmatic transfer phenomena. Similarly, Takahashi (1996; 2001) has found that advanced Japanese EFL students avoided bi-clausal “I was wondering whether you could VP” and favoured the mono-clausal “Could you (please) VP?” even though they knew the semantics of the bi-clausal structure. On this evidence Takahashi (2001: 173) con-

Pragmalinguistics and sociopragmatics

91

cludes that “the Japanese EFL learners lack the L2 pragmalinguistic knowledge that an English request can be mitigated to a greater extent by making it syntactically more complex”. Sophisticated as the above studies may be, they cannot exclude the possibility that the particular adult L2 learners’ production is constrained by sociopragmatic, rather than pragmalinguistic, considerations. For example, it could be argued that Bodman and Eisenstein’s subjects may have chosen the particular non-native-like forms of thanking as a way of maintaining their cultural identity. Besides, it is possible that the Japanese EFL students may not have assessed the particular request(s) as requiring a comparatively higher level of mitigation than others. It is therefore obvious that the lack of sharp boundaries between pragmalinguistics and sociopragmatics is also testified in pragmatic development studies. Knowledge of a grammatical structure and its pragmalinguistic functions, accompanied by lack of familiarity with the sociopragmatic conditions of target-like use, is also reported in the relevant literature. Bardovi-Harlig and Hartford (1991) show that, in a student-advising session at an American University, a Japanese student wishing to reject the faculty advisor’s suggestion uses a series of questions as a strategy of indirectly conveying negative intent. In the researchers’ analysis, the student clearly had the grammatical knowledge to form questions and the pragmalinguistic knowledge to use questions in order to indirectly convey intent, but did not have the sociopragmatic knowledge to assess when the particular strategy is appropriate and effective. It is also reported that this question strategy was not observed in the rejections of American-English speaking students. The particular data fall short of providing evidence for the appropriate use and effectiveness of the question strategy in conveying intent by L2 learners or by native speakers of the L2. For example, they do not show whether American-English speaking students employ the question strategy to indicate negative intent in other social settings, or to what extent this can happen. If it turned out that the particular strategy to convey negative intent is not favoured by native speakers of the L2, then it is possible that the Japanese student has erroneously associated a grammatical form with a pragmatic strategy, which is evidence for insufficient development of her pragmalinguistic rather than her sociopragmatic knowledge. Moreover, even if the question strategy was associated with the same function by American speakers, it is not clear whether the understanding of negative intent would result from the operation of sociopragmatic factors or from on-line inferencing heuristics, i.e. pragmalinguistic factors. As previously noted with respect to pragmatic failure and pragmatic transfer, it is sometimes difficult to diagnose the source of a learner’s inappropriate use of a speech act strategy at any particular stage of L2 development.

92 4.5.

Sophia Marmaridou

Second language teaching and testing

The most widely stated aim of L2 teaching is to enable successful communication in the foreign language. This involves not only the mastery of the phonology, syntax and semantics of the L2, but also its pragmatics. The question that is typically addressed in this context is whether pragmatic knowledge can be taught in the classroom, i.e. not necessarily by cultural immersion in the native environment, and how it can be taught to learners of different ages, educational backgrounds, etc. The applications of the pragmalinguistics/sociopragmatics distinction in second language acquisition research and the consequent refinement of the concepts of pragmatic failure and pragmatic transfer relate in an obvious way to second language teaching and testing. If this distinction is useful in, perhaps selectively, identifying causes of pragmatic failure and areas of pragmatic transfer, then L2 teaching and testing should be oriented towards the development of pragmalinguistic and sociopragmatic appropriateness (Harlow 1990: 328). Following Leech (1983), Kasper and Rose (2001) claim that communicative action includes “not only using speech acts (such as apologising, complaining, complimenting, and requesting), but also engaging in different types of discourse and participating in speech events of varying length and complexity”. More importantly, speakers do not only want to perform speech acts in order to achieve their goals, but also wish to attend to interpersonal relationships with the other speech participants in the communicative event. Within this context, pragmalinguistics refers to the resources for conveying communicative acts and encoding interpersonal meanings. According to Kasper and Rose (2001: 2), these resources include pragmatic strategies such as directness and indirectness, speech act routines, and the intensification or softening of communicative acts. The social perceptions underlying participants’ interpretation and performance of communicative action fall within the area of sociopragmatics and refer to speakers’ and hearers’ social distance and social power, their respective rights and obligations, and the degree of imposition involved in particular communicative acts. It transpires from the above that in language teaching research, as in second language acquisition research, pragmalinguistics and sociopragmatics focus on the area of speech acts and, more specifically, on their linguistic realisation and their strategic use in particular social circumstances. In this view, teaching L2 pragmatics primarily involves teaching the means and strategies of performing particular speech acts, i.e. the pragmalinguistics of L2 (see McNamara and Roever 2006), and teaching politeness, i.e. teaching how to assess social distance between interlocutors, social power and degrees of imposition in the L2 sociocultural setting (see Bou-Franch and Garcés-Conejos 2003). Clearly, to be pragmatically competent in L2, learners need knowledge both at the pragmalinguistic and the sociopragmatic levels. For example, learners need to know not only how to express a particular speech act indirectly, but also the relation between indirectness and po-

Pragmalinguistics and sociopragmatics

93

liteness, as well as when and to whom to be polite, the degree of politeness that is required in a particular context of interaction, etc. Apart from the linguistic expressions themselves, strategies of communicative actions may vary not only according to context within a particular language, but also across languages and cultures. However, as Kasper and Rose (2001) observe, adult learners get a considerable amount of L2 pragmatic knowledge for free, since some pragmatic knowledge is universal, while other aspects of such knowledge may be successfully transferred from the learners’ L1. At the pragmalinguistic level, L2 teaching is facilitated by the cross-cultural availability of the major categories of speech acts (e.g. Searle’s (1979) directives, commissives, expressives, etc.), and specific acts such as greetings, leave-takings, etc. Moreover, it has been found that the major realisation strategies of some speech acts occur across several ethnolinguistically distant speech communities (see e.g. House and Kasper 1987; Faerch and Kasper 1989; Rose 2000, among others). It is shown that in early learning states, learners may not have acquired the linguistic means necessary to implement such strategies, but once linguistic obstacles are removed by teaching, they will use the appropriate strategies without instruction. Similarly, corresponding form-function pairings between L1 and L2 used in the same contexts facilitate pragmalinguistic performance. Positive transfer and awareness of universally existing social variables, such as social distance, social power and degree of imposition, often affected by other social parameters such as sex, age, and familiarity between interlocutors (Harlow 1990), can also facilitate the learner’s task in acquiring sociopragmatic knowledge in L2. However, pragmalinguistic and sociopragmatic conventions make reference to the lexical and grammatical constructions of particular languages, while successfully performing communicative acts often depends on the degree to which a strategy is conventionalised in a particular language and culture. These parameters, then, point to the necessity of classroom instruction in L2 pragmalinguistics and sociopragmatics. Especially concerning the latter, it has been argued that the promotion of sociopragmatic awareness is also likely to improve production and comprehension of pragmalinguistic resources (Bou-Franch and GarcesConejos 2003). The apparent usefulness of the pragmalinguistics/sociopragmatics distinction in the study of L2 teaching is not proportionately reflected in L2 testing research. As in the discussion of pragmatic failure above, it is often difficult in practice to determine whether a given error is due to pragmalinguistic or sociopragmatic deficits (McNamara and Roever 2006: 55). For example, if politeness markers are missing from an utterance, this could be either because the learner does not know these markers, or because she is not aware that these markers should be used in the particular situation. The former indicates pragmalinguistic, whereas the latter sociopragmatic, inadequacy. Therefore, it is difficult to devise a test that would assess pragmalinguistic knowledge to the exclusion of sociopragmatic knowledge, or the

94

Sophia Marmaridou

reverse. The problem seems more acute when testing pragmatic strategies, such as indirectness. While the linguistic expressions used to encode indirectness belong to the pragmalinguistic level, the social target to which they are addressed is a sociopragmatic matter. Moreover, sociopragmatic appropriateness seems to be less dependent on linguistic proficiency than pragmalinguistic knowledge.3 L2 learners with a high general L2 competence are likely to build their pragmalinguistic competence more easily, but high general L2 competence does not guarantee correct judgements about sociopragmatic appropriateness. The problem of testing pragmalinguistics and sociopragmatics independently from each other also relates to the types of tests devised to test pragmatic knowledge in general, as there is “a significant tension between the construction of authentic assessment tasks and practicality” (McNamara and Roever 2006: 54). The two most frequent types are discourse completion tests (DCT’s) and role plays. In attempting to test sociopragmatic knowledge, multiple choice DCT’s were shown not to be reliable, as it is extremely difficult to devise distractors (i.e. ‘wrong’ items) that are totally unacceptable by all members of the target language community without being wrong in a very obvious way (Brown 2001). Open DCT’s can only provide evidence of the L2 learner’s sociopragmatic ability concerning the particular items in the test (e.g. apologies, requests, formulaic implicatures, etc.) rather than her general sociopragmatic ability. Additionally, they do not provide evidence of the learner’s ability to participate in conversations in real time, or to take turns, especially given that speakers normally distribute strategies over various turns, monitoring their interlocutor’s reactions and adjusting their talk accordingly. Although role plays are more reliable in this respect, they are not preferred as testing devices because they take substantial time to conduct, they are difficult to monitor, and they require multiple rating. Besides, role play cannot establish context in the way that authentic communication does in the real world, nor does it represent the speaker’s or the addressee’s actual face needs in real terms. However, it is reported that role plays can assess learners’ ability to make judgments about sociopragmatic appropriateness at discourse level (Cook 2001). By contrast, it is possible to test L2 pragmalinguistic knowledge practically and reliably, even though creating items for pragmalinguistic instruments is not easy. Roever (2006), for example, constructed a test to assess interpretation of implicature, recognition of routine formulae, and knowledge of speech act strategies. While these three types of knowledge differ, they are based on general pragmatic and linguistic input. Interpretation of implicature was expectedly found to rely on the learners’ proficiency and deduction abilities. The effect of L2 learners’ proficiency in their knowledge of routines was almost negligible, given that routines are not creatively constructed, but rather learned as holistic items to be used in specific situations. As opposed to routines, the production of speech acts and the pragmalinguistic strategies employed therein correlated with level of L2 proficiency, as expected. Some of the limitations of assessing L2 learners’ pragmalinguistic

Pragmalinguistics and sociopragmatics

95

knowledge are similar to the ones of testing sociopragmatics. The written format of the tests restricts the number of contextual cues that would be available in a real encounter. The lack of extended negotiation of speech acts in the context of a DCT is another limiting factor. However, studies of pragmalinguistic assessment in L2 can benefit both from the development of testing techniques in general and from further cross-cultural study of pragmalinguistic parameters of communicative acts in particular. It becomes obvious from all the above that the pragmalinguistics/sociopragmatics distinction is better exploited in the area of L2 teaching methodology than in studies of pragmatic failure, pragmatic transfer, pragmatic development and L2 testing. What the latter have in common is their diagnostic methodology in assessing L2 learners’ approximations to the target language. The sources of particular approximations are not easy to detect. By contrast, a theoretical categorisation into pragmalinguistic and sociopragmatic parameters of language use can fairly easily be mapped onto teaching methods, materials and awareness raising tasks in which L2 learners may be involved.

5.

Historical sociopragmatics4

The concepts of pragmalinguistics and sociopragmatics have also informed the field of study generally known as Historical Pragmatics. Still a relatively young field of linguistics, historical pragmatics takes a broad, sociologically based view of patterns of intentional human interaction, as these are determined by the conditions of societies of earlier periods (see Jucker 2006 for a brief overview).5 One approach to historical pragmatics, pragmaphilology, describes the contextual aspects of historical texts and is essentially synchronic. It refers to the identities of speakers and addressees, their social and personal relationship, the physical and social setting of text production and reception and the goals of the text (Jacobs and Jucker 1995). Another approach, diachronic pragmatics, concerns form-to-function and function-to-form mappings across different historical stages of the same language. Apparently, form-to-function mappings correspond to a pragmalinguistic perspective on the analysis of phenomena such as deixis (Fries 1993), discourse markers (Brinton 1990), or interjections (Taavitsainen 1995). Diachronic function-to-form studies, adopting a sociopragmatic perspective, compare the realisation of speech acts, politeness formulae, text types, or forms of dialogue, at different points in the development of a language (Arnovick 1999; Jucker and Taavitsainen 2008; Fritz 1995; Biber and Finegan 1992). Since the mid-1990’s the investigation of the relevant topics is based on available historical corpora (e.g. the Helsinki Corpus of English Texts), which are either read and annotated in an old-fashioned and time consuming way, or analysed by corpus-linguistic methods of data retrieval. Apart from written genres, such as essays, fiction, personal letters, and manuals for good

96

Sophia Marmaridou

behaviour, speech-based genres have also been analysed, e.g. dialogues from plays and fiction, court proceedings, political debates, public speeches, and sermons (Biber and Finegan 1992; Jacobs and Jucker 1995). Within this framework, historical sociopragmatics focuses on the interaction between specific aspects of social context and particular historical language use that leads to pragmatic meanings. More specifically, it concerns language use in its situational, local context and the ways in which situational contexts generate norms which interlocutors employ or exploit for pragmatic purposes. A synchronic approach to historical sociopragmatics consists in showing how language use shapes and is shaped by context at a particular historical period, while a diachronic perspective involves the investigation of how shifts in language use affect shifts in contexts, or how shifts in contexts shape language use over time. An important line of investigation in historical sociopragmatics concerns the reconstruction of contexts on the basis of historical texts; the latter are viewed as carrying evidence of, or projecting, their own contexts (Culpeper 2009: 182–183). As already mentioned, the local context, or the sociological context, of language use is of primary concern to sociopragmatics. For example, Nevalainen and Raumolin-Brunberg (1995) have studied the sociopragmatics of terms of address in Early English correspondence. Given that the way people address each other in interaction depends on social variables such as age, status, dialect, etc., the question arises as to the relation between sociopragmatics and sociolinguistics. According to Nevalainen and Raumolin-Brunberg (2003), sociolinguistics comprises three areas of research, namely, social dialectology, interactional sociolinguistics and the sociology of language. Parallel to this distinction, Culpeper (2009: 180) identifies three levels of what may constitute the context against which texts may be understood: the most local, immediate text and co-text of interlocutors, the medial level of social situation (including speech events, activity types, frames, etc.), and the most general level, making reference to national and regional cultures, institutional cultures, etc. As the author notes (Culpeper 2009: 181) sociopragmatics should primarily, though not exclusively, concern itself with the medial context and the phenomena that constitute it. Social situations can provide a link between micro, more linguistically-oriented considerations (the typical focus of pragmalinguistics), and macro, more sociologically-oriented considerations (the typical focus of a field such as Critical Discourse Analysis).6

According to this definition, sociolinguistics is concerned with mapping regular patterns of usage in interaction, while sociopragmatics is concerned with the ways in which these regular patterns are used and exploited in particular interactions. Given the concern of historical sociopragmatics for interactional data and the obvious non-availability of recorded forms of interaction of earlier periods, the question arises as to the availability of historical corpora that lend themselves to sociopragmatic research and the search tools required for this type of investigation.

Pragmalinguistics and sociopragmatics

97

Recent developments in historical corpus linguistics have allowed more pragmatic research questions in a way that is amenable to corpus-linguistic methodologies (see, e.g., Kohnen 2008; Jucker et al. 2008); but the need is still felt for the appropriate pragmatic tagging of relevant pragmatic units, such as those of interest to sociopragmatic research. Within this framework, Culpeper and Kytö’s compilation of the Sociopragmatic Corpus, a subsection of the Corpus of English Dialogues: 1560–1760, has made sociopragmatic annotation possible. Facing the issue of the sociopragmatic variability of context, Archer and Culpeper (2003) devise and implement a tagging system that allows the annotation of speech changes that potentially affect the social meaning of interaction in drama and trial proceedings.7 The variables include relatively static, sociolinguistic values, such as status (nobility, gentry, professional, ordinary commoners, lowest groups) and age (young, adult, old), as well as dynamic values, such as activity role (witness, defendant, customer), kinship role (father, daughter, mother, son, father-in-law), social role (surgeon, baker, friend), and dramatic role (fool, villain, seducer, etc.). The significance of this type of work lies in capturing the utterance-by-utterance interaction between speakers and their addressees in terms of sociopragmatic variables, which enhance the dynamic aspects of the analysed texts. Working with the same corpus, Archer and Culpeper (2009) develop the notion of keyness as the identification of keywords, key parts-of-speech and key semantic fields that are statistically characteristic of the speech of dyads in interaction, e.g. master/mistress with servant and examiner with examinee in trials. Examples include the use of imperative verbs directed to servants by their masters in the partsof-speech category and the use of the semantic domain of documents and writing in the speech of the same dyad. These social role dyads are a specific part of particular social situations, the latter constituting the local context relative to all the social roles and situations that make up the two genres under examination, i.e. drama and trial proceedings. Therefore, keyness analysis is a tool for the identification of the characteristic discourse norms of particular local contexts. While the authors use the term sociophilology to refer to their approach, their concern for the particular type of annotation and the analysis of the local conditions of language use indicates the deployment of a sociopragmatic orientation to their work. The concept of sociopragmatics has also informed further studies focusing on the social conditions that affect the use of texts. Adopting the model of critical discourse analysis and frame analysis, Wood (2009) examines personal letters from late 15th century English taken from the two volumes of the Paston Collection, and Margaret Paston’s letters in particular. Wood (2009: 188) addresses the issue of authorship, which is important in historical sociopragmatics: “since some of Margaret’s letters were penned by her own sons, might we not be justified in ascribing the language of those letters to them, rather than representative of Margaret Paston herself?” Supporting the view that local contexts are culture specific, she shows that local context does not only make reference to a particular interactional dyad,

98

Sophia Marmaridou

but also to a more extended speech community, to the discursive practices of that community, the text’s distribution and its consumption. Within this framework, she shows that on the one hand these letters evidence the manipulation of power structures within a family and, on the other, they indicate Margaret Paston’s control over the form and content of the letters, thereby confirming authorship. Viewed broadly, historical sociopragmatics includes the macro-level of social, socio-cultural and sociological factors as well as the micro-level of personal, situational and stylistic factors. Nevala (2009) discusses the concept of person reference and social deixis in Late Modern English letters and journals. Her data consist of the correspondence of some members of the Lunar Society of Birmingham and the letters and journals of Agnes Porter, an eighteenth century governess. In particular, she concentrates on the interactional aspects of the referential term friend as used by ‘public’ figures and ordinary people, as well as on the way in which selfand addressee-oriented third person reference is used to convey contextual shifts in interpersonal distance and authority. She demonstrates that the writers in the data strategically use different terms in order to alternate between specific social positionings, but they also take into consideration the prevalent social and societal constraints. Examining the letters subcorpus of the Network of Eighteenth-century English Texts within a relevance theoretic perspective, Fitzmaurice (2009) focuses on the sociopragmatic construction of implicature and inference in the illicit courtship correspondence between Edward Wortley and Mary Pierrepont. She argues that key historical and cultural reference points are necessary in understanding how communicative practices are embedded in the local context. Such practices are typically associated with certain activity types and social rituals including courting. Therefore, the understanding of implicated meanings in these letters makes reference to the cotext created by the discourse of the letter, the situational context constructed by the exchange of letters, and the broader historical context and the social constraints in which the correspondence is embedded. Fitzmaurice (2009) shows that it is at the discourse cotext that the participants in the interaction enact their sociopragmatic roles and fight for dominance of their relationship. It transpires from all the above that the concepts of pragmalinguistics and sociopragmatics have been particularly useful in historical pragmatics and enhance the possibilities of historical corpus linguistics. Also notice that, within the framework of historical sociopragmatics, pragmalinguistic concerns are also accommodated, as, for example, the study of the forms used to perform a particular function in Nevala’s (2009) work. On this understanding, sociopragmatics is broadly construed as referring to both the micro and macro levels of pragmatic analysis. It is possible that the success in the application of the two concepts under investigation to historical pragmatics depends partly on the availability of appropriate corpora and partly on appropriate retrieval methods that facilitate the analysis.

Pragmalinguistics and sociopragmatics

6.

99

Concluding remarks

As already shown, apart from Leech’s (1983) and Thomas’s (1983) contrast and distinction between pragmalinguistics and sociopragmatics, these terms have been variously construed in the relevant literature and, used independently of each other, they have emphasised the social conditions in which utterances are produced and understood as verbal contributions in acts of communication. Regardless of whether the pragmalinguistics/sociopragmatics distinction is theoretically justified, analytically possible, or ideologically desirable, it has motivated the corresponding methodological distinction in L2 acquisition research and L2 teaching and testing in particular. The concepts of pragmatic failure, pragmatic transfer and pragmatic development, relevant in these fields of research, explicitly make reference to this distinction. The research questions in the pragmatics of L2 teaching and testing have to do with defining L2 learners’ verbal approximations to L2 as detected in their performance and then attributing them to pragmalinguistic or sociopragmatic sources. The focus is on whether L2 learners’ utterances are appropriate communicative contributions in the social context in which they occur and whether their inferencing is based on an understanding of social context. Appropriateness is an essentially evaluative term (see Fetzer 2004; 2007). In the framework of L2 acquisition research it refers to the learners’ speech actions, which are constantly assessed as being in accordance with the linguistic norms and contextual constraints and requirements of particular communicative acts in the target language and culture. The appropriateness of L2 learners’ performance is not always easy to evaluate on the basis of the pragmalinguistics/sociopragmatics distinction. More specifically, the absence of sharp boundaries between these two concepts weakens their methodological usefulness in diagnosing the source of pragmatic failure, pragmatic transfer, or the stage of L2 learners’ pragmatic development, especially in the context of L2 testing, as already mentioned. It seems easier to maintain the pragmalinguistics/sociopragmatics distinction in the context of L2 teaching, in the sense that the methodology adopted, the materials used, and the tasks in which learners are engaged can be designed so as to focus on this distinction and raise the L2 learners’ awareness of the linguistic and social issues involved in their use of L2. In the area of historical pragmatics and historical corpus linguistics the situation is different. The texts examined belong either to the same period (as in synchronic studies) or to different periods (as in diachronic studies).8 Pragmalinguistic diachronic studies proceed from form to function, whereas sociopragmatic diachronic studies follow the function-to-form process. Historical sociopragmatics focuses on the interaction of context and language use either synchronically or diachronically. In the context of historical pragmatics and historical corpus linguistics the distinction between pragmalinguistics and sociopragmatics can be fruitfully maintained in spite of the lack of sharp boundaries between the two. His-

100

Sophia Marmaridou

torical texts are viewed as definitive and complete instantiations of a complex configuration of parameters, such as coparticipants and their social, interactional and discursive roles, communicative action, genre, and cultural norms and strategies of a speech community. Unlike historical pragmatics research, L2 acquisition research is based on texts as approximations of the target language, which makes it extremely difficult to diagnose just which specific parameter is at issue in evaluating appropriateness. On the basis of the above, it is fair to say that, even though the pragmalinguistics/sociopragmatics distinction was initially launched as a level of cross-linguistic application of the general pragmatic principles of cooperation and politeness (as in Leech 1983), and even though it was systematically taken up in the context of L2 acquisition research, it seems to have been more relevant in the area of historical pragmatics and historical corpus linguistics as a reliable methodological tool in the investigation of the pragmatics of historical texts. Needless to say, further theoretical justification of the concepts of pragmalinguistics and sociopragmatics, and of the distinction between them, ultimately depends on future research in pragmatics and related areas of linguistic enquiry.

Notes 1. It is interesting to note here that Olsen’s views remind one of Wittgenstein’s (1958) position that linguistic meaning resides in language use and Wierzbicka’s (1992) introduction of a Natural Semantic Metalanguage to deal with the cultural relativity of word-meaning pairings across languages. Evidently, the dependence of linguistic meaning on use in particular socio-cultural settings is a recurring theme in linguistic theory. 2. In fact, interlanguage, as an interim system of L2 learners, has some features of the L1 and some of the L2, but also features that are independent of their L1 and L2 (see Yule 2006: 167). 3. This is compatible with the view that sociopragmatic awareness is likely to improve pragmalinguistic performance. 4. See also Jucker A. H. and I. Taavitsainen (eds.) 2010. Historical Pragmatics, Vol. 8 of the series Handbooks of Pragmatics (Berlin: Mouton de Gruyter). 5. Within the narrow, Anglo-American tradition to pragmatics, Traugott (2004: 538) defines historical pragmatics as a usage-based, pragmatically motivated, approach to language change. 6. Critical discourse analysis is “fundamentally concerned with analysing opaque as well as transparent structural relationships of dominance, discrimination, power and control as manifested in language” (Wodak 2001: 2). 7. These two text types offer interactive, face-to-face, speech-related data that only approximate authentic discourse. Drama consists of imaginary constructed dialogue and trial proceedings constitute the record of a prior speech event. 8. Relevant to this point is the distinction between diachronic pragmatics and pragmaphilology (Jacobs and Jucker 1995) and diachronic sociopragmatics and sociophilology (Archer and Culpeper 2009).

Pragmalinguistics and sociopragmatics

101

Acknowledgements I would like to thank the three reviewers for their helpful suggestions and comments. I am also grateful to Jonathan Culpeper, Marcelo Dascal, Andreas Jucker, and Carsten Roever for making their work available to me. This paper has benefited from very useful stylistic suggestions made by Eleni Antonopoulou, to whom I am greatly indebted.

References Archer, Dawn and Jonathan Culpeper 2003 Sociopragmatic annotation: New directions and possibilities in historical corpus linguistics. In: Andrew Wilson, Paul Rayson, and Tony McEnery (eds.), Corpus Linguistics by the Lune: A Festschrift for Geoffrey Leech, 37–58. Frankfurt am Main: Lang. Archer, Dawn and Jonathan Culpeper 2009 Identifying key sociophilological usage in plays and trial proceedings (1640– 1760). An empirical approach via corpus annotation. Journal of Historical Pragmatics 10/2: 286–309. Arnovick, Leslie K. 1999 Diachronic Pragmatics. Seven Case Studies in English Illocutionary Development. Amsterdam: Benjamins. Bang, Jørgen Chr. and Jørgen Døør 1979 Language, theory, and conditions for production. In: Jacob L. Mey (ed.), Pragmalinguistics. Theory and Practice, 21–50. The Hague: Mouton. Bardovi-Harlig, Kathleen 1999 Exploring the interlanguage of interlanguage pragmatics: A research agenda for acquisitional pragmatics. Language Learning 49: 677–713. Bardovi-Harlig, Kathleen and Beverly Hartford 1991 Saying “no” in English: Native and nonnative rejections. In: Lawrence F. Bouton and Yamura Kachru (eds.), Pragmatics and Language Learning, 41–57. Urbana-Champaign, IL: Division of English as an International Language, University of Illinois. Bates, Elizabeth 1976 Language and Context. The Acquisition of Pragmatics. New York: Academic Press. Beebe, Leslie M. and Tomoko Takahashi 1989 Sociolinguistic variation in face-threatening speech acts. In: Miriam Eisenstein (ed.), The Dynamic Interlanguage, 199–218. New York: Plenum. Bebee, Leslie M., Tomoko Takahashi, and Robin Uliss-Weltz 1990 Pragmatic transfer in ESL refusals. In: Robin C. Scarcella, Elaine Andersen, and Stephen C. Krashen (eds.), Developing Communicative Competence in a Second Language, 55–73. New York: Newbury House. Biber, Douglas and Edward Finegan 1992 The linguistic evolution of five written and speech-based English genres from the 17th to the 20th centuries. In: Matti Rissanen, Ossi Ihalainen, Terttu Neval-

102

Sophia Marmaridou

ainen, and Irma Taavitsainen (eds.), History of Englishes. New Methods and Interpretations in Historical Linguistics, 55–73. Berlin: Mouton de Gruyter. Blum-Kulka, Shoshana 1991 Interlanguage pragmatics: The case of requests. In: Robert Phillipson, Eric Kellerman, Larry Selinker, Michael Sharwood-Smith, and Merrill Swain (eds.), Foreign/Second Language Pedagogy Research, 255–272. Clevedon: Multilingual Matters. Bodman, Jean and Miriam Eisenstein 1988 May God increase your bounty: The expression of gratitude in English by native and non-native speakers. Cross Currents 15: 1–21. Bou-Franch, Patricia and Pilar Garcés-Conejos 2003 Teaching linguistic politeness: A methodological proposal. International Review of Applied Linguistics 41/1: 1–22. Brinton, Laurel 1990 The development of discourse markers in English. In: Jacek Fisiak (ed.), Historical Linguistics and Philology, 45–71. Berlin: Mouton de Gruyter. Brown, James Dean 2001 Six types of pragmatics tests in two different contexts. In: Kenneth Rose and Gabriele Kasper (eds.), Pragmatics in Language Teaching, 301–325. New York: Cambridge University Press. Brown, Penelope and Stephen C. Levinson 1987 Politeness: Some Universals in Language Usage. New York: Cambridge University Press. Brumfit, Chris and Keith Johnson 1979 The Communicative Approach to Language Teaching. Oxford: Oxford University Press. Candlin, Chris N. 1976 Communicative language teaching and the debt to pragmatics. In: Clea Rameh (ed.), 27th Round Table Meeting, 237–256. Washington, DC: Georgetown University Press. Cook, Haruko Minegishi 2001 Why can’t learners of Japanese as a foreign language distinguish polite from impolite speech styles? In: Kenneth R. Rose and Gabriele Kasper (eds.), Pragmatics in Language Teaching, 80–102. New York: Cambridge University Press. Culpeper, Jonathan 2009 Historical sociopragmatics: An introduction. Journal of Historical Pragmatics 10/2: 179–186. Dascal, Marcelo 1985 Language use in jokes and dreams: sociopragmatics vs. psychopragmatics. Language and Communication 5/2: 95–106. Dascal Marcelo and Edson Françozo 1989 The pragmatic turn in psycholinguistics: Problems and perspectives. Theoretical Linguistics 15/1,2: 1–23. Eelen, Gino 2001 A Critique of Politeness Theories. Manchester: St Jerome.

Pragmalinguistics and sociopragmatics

103

Faerch, Claus and Gabriele Kasper 1989 Internal and external modification in interlanguage request realization. In: Shoshana Blum-Kulka, Juliane House, and Gabriele Kasper (eds.), Cross-cultural Pragmatics, 221–247. Norwood, NJ: Ablex. Fetzer, Anita 2004 Recontextualizing Context: Grammaticality Meets Appropriateness. Amsterdam: Benjamins. Fetzer, Anita 2007 If I may say so: Indexing appropriateness in dialogue. In: Anita Fetzer (ed.), Context and Appropriateness, 115–145. Amsterdam: Benjamins. Fitzmaurice, Susan M. 2009 The sociopragmatics of a lovers’ spat. The case of the eighteenth-century courtship letters of Mary Pierrepont and Edward Wortley. Journal of Historical Pragmatics 10/2: 215–237. Freud, Sigmund 1900 The Interpretation of Dreams. London: Harvester Press, Hassocks. Freud, Sigmund 1905 Jokes and their Relation to the Unconscious. London: The Hogarth Press. Fries, Udo 1993 Towards a description of text deixis in Old English. In: Klaus R. Grinda and Klaus-Dieter Wetzel (eds.), Anglo-Saxonica. Festschrift für Hans Schabram zum 65. Geburtstag, 527–540. München: Fink. Fritz, Gerd 1995 Topics in the history of dialogue forms. In: Andreas H. Jucker (ed.), Historical Pragmatics. Pragmatic Developments in the History of English, 469–498. (Pragmatics and Beyond New Series 35). Amsterdam: Benjamins. Grice, Paul 1975 Logic and conversation. In: Peter Cole and Jerry Morgan (eds.), Syntax and Semantics, Volume 3: Speech Acts, 41–58. New York: Academic Press. Harlow, Linda L. 1990 Do they mean what they say? Sociopragmatic competence and second language learners. The Modern Language Journal 74: 328–351. Haverkate, Henk 1994 The dialogues of Don Quixote de la Mancha: A pragmalinguistic analysis within the framework of Gricean maxims, speech act theory, and politeness theory. Poetics 22: 219–241. Holmes, Janet 2008 An Introduction to Sociolinguistics. 3rd ed. London/New York: Longman. House, Juliane and Gabriele Kasper 1987 Interlanguage pragmatics: Requesting in a foreign language. In: Wolfgang Lörscher and Rainer Schulze (eds.), Perspectives on Language in Performance, 1250–1288. Tübingen: Narr. Jacobs, Andreas and Andreas H. Jucker 1995 The historical perspective in pragmatics. In: Andreas H. Jucker (ed.), Historical Pragmatics. Pragmatic Developments in the History of English. (Pragmatics and Beyond New Series 35), 3–33. Amsterdam: Benjamins.

104

Sophia Marmaridou

Jucker, Andreas H. 2006 Historical pragmatics. In: Jan-Ola Östman and Jef Verschueren (in collaboration with Eline Versluys) (eds.), Handbook of Pragmatics 2006, 1–14. Amsterdam: Benjamins. Jucker, Andreas H. 2008 Historical pragmatics. Language and Linguistics Compass 2/5: 894–906. Jucker, Andreas H. and Irma Taavitsainen (eds.) 2008 Speech Acts in the History of English. Amsterdam: Benjamins. Jucker, Andreas H. and Irma Taavitsainen (eds.) 2010 Historical Pragmatics. Handbooks of Pragmatics, Vol. 8. Berlin: De Gruyter. Jucker, Andreas H., Gerold Schneider, Irma Taavitsainen, and Barb Breustedt 2008 Fishing for compliments: Precision and recall in corpus-linguistic compliment research. In: Andreas H. Jucker and Irma Taavitsainen (eds.), 2008 Speech Acts in the History of English, 273–274. Amsterdam: Benjamins. Kasper, Gabriele 1992 Pragmatic transfer. Second Language Research 8/3: 203–231. Kasper, Gabriele and Kenneth R. Rose 2001 Pragmatics in Language Teaching. In: Kenneth R. Rose and Gabriele Kasper (eds.), Pragmatics in Language Teaching, 1–10. New York: Cambridge University Press. Kasper, Gabriele and Kenneth R. Rose 2002 Pragmatic Development in a Second Language. Oxford: Blackwell. Keenan, Edward L. 1976 The universality of conversational postulates. Language in Society 5: 67–80. Kohnen, Thomas 2008 Tracing directives through text and time: Towards a methodology of a corpusbased diachronic speech-act analysis. In: Andreas H. Jucker and Irma Taavitsainen (eds.), 2008 Speech Acts in the History of English, 295–310. Amsterdam: Benjamins. Leech, Geoffrey N. 1983 Principles of Pragmatics. London/New York: Longman. Levinson, Stephen C. 1983 Pragmatics. Cambridge: Cambridge University Press. Marmaridou, Sophia S. A. 2000 Pragmatic Meaning and Cognition. Amsterdam: Benjamins. McNamara, Tim and Carsten Roever 2006 The social dimension of proficiency: How testable is it? Language Learning 56/2: 43–79. Mey, Jacob L. 1979 Introduction. In: Jacob L. Mey (ed.), Pragmalinguistics. Theory and Practice, 10–17. The Hague: Mouton. Mey, Jacob L. 1993 Pragmatics. An Introduction. Oxford: Blackwell. Morris, Charles W. 1938 Foundations of the theory of signs. In: Otto Neurath, Rudolf Carnap and Charles W. Morris (eds.), International Encyclopaedia of Unified Science. Volume 1, 77–138.

Pragmalinguistics and sociopragmatics

105

Nevala, Minna 2009 Altering distance and defining authority. Person reference in Late Modern English. Journal of Historical Pragmatics 10/2: 238–259. Nevalainen, Terttu and Helena Raumolin-Brunberg 1995 Constraints on politeness: The pragmatics of address formulae in Early English correspondence. In: Andreas H. Jucker (ed.), Historical Pragmatics. Pragmatic Developments in the History of English, 57–76. Amsterdam: Benjamins. Nevalainen, Terttu and Helena Raumolin-Brunberg 2003 Historical Sociolinguistics. Harlow: Pearson. Olsen, Svend Erik 1979 Psychopathology, interaction, and pragmatic linguistics. In: Jacob L. Mey (ed.), Pragmalinguistics. Theory and Practice, 233–253. The Hague: Mouton. Olshtain, Elite and Liora Weinback 1992 Interlanguage features of the speech act of complaining. In: Gabriele Kasper and Shoshana Blum-Kulka (eds.), Interlanguage Pragmatics, 195–208. New York: Oxford University Press. Prucha, Jan 1983 Pragmalinguistics: East European Approaches. Amsterdam: Benjamins. Riley, Philip 1979 Towards a contrastive pragmalinguistics. Papers and Studies in Contrastive Linguistics 10: 57–78. Robinson, Mary Ann 1992 Introspective methodology in interlanguage pragmatics research. In: Gabriele Kasper (ed.), Pragmatics of Japanese as native and target language Technical report no. 3: 27–82. Honolulu: University of Hawai’I at Manoa, Second Language Teaching and Curriculum Centre. Roever, Carsten 2006 Validation of a web-based test of ESL pragmalinguistics. Language Testing 23/2: 229–256. Rosaldo, Michelle Z. 1982 The things we do with words: Ilongot speech acts and speech act theory in philosophy. Language in Society 11: 203–237. Rose, Kenneth R. 2000 An exploratory cross-sectional study of interlanguage pragmatic development. Studies in Second Language Acquisition 22: 27–67. Searle, John R. 1979 Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press. Selinker, Larry 1972 Interlanguage. International Review of Applied Linguistics 10: 209–230. Sifianou Maria 1992 Politeness Phenomena in England and Greece. A Cross-Cultural Perspective. Oxford: Clarendon Press. Taavitsainen, Irma 1995 Interjections in Early Modern English: From imitation of spoken to conventions of written language. In: Andreas H. Jucker (ed.), Historical Pragmatics. Pragmatic Developments in the History of English, 439–465. Amsterdam: Benjamins.

106

Sophia Marmaridou

Takahashi, Satomi 1996 Pragmatic transferability. Studies in Second Language Acquisition 18: 189– 223. Tahahashi, Satomi 2001 The role of input enhancement in developing pragmatic competence. In: Kenneth R. Rose and Gabriele Kasper (eds.), Pragmatics in Language Teaching, 171–199. New York: Cambridge University Press. Thomas, Jenny 1981 Pragmatic failure. M.A. dissertation, Department of Language and Linguistics, University of Lancaster. Thomas, Jenny 1983 Cross-cultural pragmatic failure. Applied Linguistics 4/2: 91–112. Traugott, Elizabeth Closs 2004 Historical Pragmentics. In: Lawrence R. Horn and Gregory Ward (eds.), Handbook of Pragmatics, 530–561. Oxford: Blackwell. Trosborg, Anita 1987 Apology strategies in natives/non-natives. Journal of Pragmatics 11: 147–167. Widdowson, Henry G. 1979 Explorations in Applied Linguistics. Oxford: Oxford University Press. Wierzbicka, Anna 1987 English Speech Act Verbs: A Semantic Dictionary. Sydney/New York: Academic Press. Wierzbicka, Anna 1992 Semantics, Culture, and Cognition. Universal Human Concepts in CultureSpecific Configurations. New York/Oxford: Oxford University Press. Wittgenstein, Ludwig 1958 Philosophical Investigations. Oxford: Blackwell. Wodak, Ruth 2001 What CDA is about. A summary of its history, important concepts and its developments. In: Ruth Wodak and Michael Meyer (eds.), Methods of Critical Discourse Analysis, 1–13. London: Sage Publications. Wolfson, Nessa 1981 Compliments in cross-cultural perspective. TESOL Quarterly 15: 117–124. Wood, Johanna L. 2009 Structures and expectations. A systematic analysis of Margaret Paston’s formulaic and expressive language. Journal of Historical Pragmatics 10/2: 187–214. Yule, George 2006 The Study of Language. 3rd ed. Cambridge: Cambridge University Press.

4.

Metapragmatics Axel Hübler

Introduction The term ‘metapragmatics’ is slippery, many meanings have been ascribed to it, too many perhaps. A restrictive or selective attitude, however, does not seem appropriate for a handbook article. On the contrary, the survey should accommodate as many perspectives as are possible and justifiable. The problems mainly arise from two interrelated facts. (1) The Greek loan prefix meta- has different meanings (in between, after, later or beyond, among others) and thus lends itself to forming all sorts of terms. (2) The formative meta- is popular. Wordspy, the internet window administered by Random, lists 75 new ‘meta’-formations.1 Its attraction presumably derives from its particular air of something intellectual, detached and with an impressively broad view. To reduce the risk of a sell-out of this formative device, it may be prudent to control its deployment and avoid it where it is not necessary. A case in point is the use of the term metapragmatics, where the term pragmatics actually suffices. If we define ‘pragmatics’ as the discipline which has interpersonal communication as its object, then it goes without saying that all abstractions and theoretical conclusions which help explain communicative behavior form part of this discipline, in fact they are defining features of the discipline. This holds true for all the answers given to questions such as how to cooperate, how to be kind, polite, etc. and all the conditions found to explain aspects of interpersonal communication, i.e., the conditions of ‘thinkability’ (e.g., the constitutive rules of Searle’s speech acts), the conditions of feasibility (e.g., the Gricean maxims) and the conditions of recognizability (e.g., the linguistic and paralinguistic expressive means) (cf. Caffi 2006: 84). To separate such theoretical stances or constructs from the discipline that developed them seems to make little sense. ‘Communication’ can be act and object, as Hagemann (1997: 32) subtly observes, ‘pragmatics’ cannot. It is certainly true that communication shows properties that can be labeled ‘pragmatic’, but such labeling originates in an external view. This then is a reading of metapragmatics that deserves mentioning,2 but will not be part of the focus of the article; the theoretical views and notions as such, however, will, of course, be used where necessary. What remains to be unpacked are the following four readings of the term and the four corresponding forms of linguistic practice: metapragmatics as the study of explicit metacommunication (section 1); metapragmatics as the study of implicit metacommunication (section 2); metapragmatics as the study of people’s abstract-

108

Axel Hübler

ing from interacting (section 3.1); and metapragmatics as metatheory of pragmatics (section 3.2).

1.

Metapragmatics as the (pragmatic) study of explicit metacommunication

The object of a (meta)pragmatic analysis is metacommunication. In a first go, it can be defined as communication about (selected aspects of) communication. Usually, the communication talked about is neither communication ‘as such’ (the abstract and general possibilities and impossibilities of communication) nor some definite previous or future communication, but the ongoing communication. Paul (1999: 98) characterizes it immanently as ‘practical speech reflection’. Quite original is Anton’s view contrasting it with communication. [W]e routinely and commonly attend not to speech but simply from it. That is, when we listen with [sic!] one another in our everyday encounters, we routinely listen from our speech to the thought so intended. (Anton 1998: 198–199)

This behavior contrasts with what happens in particular circumstances, where we explicitly attend to the speech itself. In Leder’s [1990] terminology, moments of breakdown manifest speech ‘dys-appearance.’ Reconsider the case of reading a novel: I may be engrossed in the story, simply attending from the sentences to the story, but then, a misspelled word, a foreign word, or perhaps a too-recondite one appears. Now, I consciously and explicitly reflect to the word […]. Speech, then, is routinely an absent body, an intentional arc which disappears for the sake of the meaning so intended. (Anton 1998: 199)

What Anton’s circumscription brings out nicely is a kind of break separating communication from metacommunication, a sort of strained dialogical structure, which is not attributable to a change in speakers (or voices), but to a change in the level of communication. This is mainly brought about by a functional change from practical acting to reflecting.3 The change in level is often notable in that it involves a change from ordinary language to metalanguage. Metalanguage is commonly understood as the language referring to language; it represents what Verschueren (1998: 55) calls the object notion of metalanguage.4 It is, however, not an autonomous language (not even where linguists use a metalanguage). When members of a speech community engage in metacommunication, the metalanguage used does not consist of complicated terminology but of ordinary words, mainly nouns (such as chat or denial) and verbs (such as promise or argue), whose feature that distinguishes them from other words is their function to refer not to something in the world out there but to some aspect of speech.5

Metapragmatics

1.1.

109

On full forms of metacommunication

Though metacommunication is always embedded in primary communication, its extension can vary considerably. For the purpose of this article, it will be sufficient to simply differentiate between full and reduced (or abbreviated) metacommunicative forms. A full form consists of at least one metacommunicative utterance; anything below utterance level can no longer be considered metacommunication proper, but is, of course, still metacommunicative.6 If primary communication is the on-line event, then metacommunicative utterances (and more extended forms) cause such an articulate deviation from it that they appear as ‘off-line’, as it were (cf. Hübler and Bublitz 2007: 12). The boundary between primary communication and metacommunication is usually not marked in strictly formal terms (except for those cases in which a speaker refers to some previous utterance within the current communicative situation by using some speech reporting frame (I said …, you said …); for references to utterances outside the current situation, cf. below). The shift in level is identified, rather, in pragmatic terms. The following questions are meant to delineate the main pragmatic parameters that apply for analyzing metacommunication. They pertain to the issue/topic chosen, to the intention pursued, and to some frame conditions of metacommunication (for a more detailed treatment and examples of full-length analyses, cf. Bublitz and Hübler (eds.) 2007).7 (1) What are the topics of metacommunication? Though always referring to particular speech events, either single or serial, the topical scope can be generic as well as specific. Metacommunicative topics of a more generic kind may, for instance, concern the problem of effability (I don’t have words which could adequately express how I feel or ‘Love’ is just a gross label), the mode chosen (You don’t have to touch me when you want something from me or Could you speak up a bit?), or questions of participating in a conversation (Why do you always want to dominate conversation? or You never take part in this kind of talk). They may be comments that operate with general principles, norms and maxims; thus Grice’s maxims find rich application (e.g., the maxim of quality in Are you telling the truth?, the maxim of quantity in Just give us the gist of it, the maxim of relevance in Come to the point, please! and the maxim of manner in Let me just finish this story before we leave) as well as Leech’s politeness principle (Stop being so patronizing!). More specific metacommunicative topics may concern illocutionary functions (Is that a complaint?), acts of reference (Who are you talking about?) or predication (That’s a very euphemistic formulation or What do you mean by ‘insinuation’?). Or they may concern matters of discourse organization. These include performing or holding back certain speech acts (Give me your permission to do so, please), announcing a speech act (And now I will disclose my next plans), or steering an ongoing discussion (We’ve got three requests to speak; first Lila, then John,

110

Axel Hübler

then Karen). Comments relating to the text structure (such as But the real sensation is coming now) may occur within longer turns. (2) What for or why? These are the central questions for any pragmatic approach to communication. The purposes that metacommunicative utterances in particular pursue are in part prefigured by their topic.8 Thus they concern the organization of discourse, including the textual structure of longer stretches of utterance (So far we have been describing a state of affairs; now we may turn to …); they help secure an adequate understanding, including attempts at rendering a speech act function more precise (Is this a complaint or an accusation?) and other kinds of monitoring. Their sense is to prevent or repair misunderstandings. Another common purpose of metacommunicative comments is evaluative, mainly critically, in nature, in as far as what is topicalized marks or implies a negative deviance from what is to be expected (But I am repeating myself). Where such comments are self-critical, they often represent a preventive measure anticipating criticism from others. Besides general communicative norms and principles, which provide the standard for (metacommunicatively) evaluating a given utterance, other standards seem to have been availed of recently as well. They are based on what has become known under the label of critical language awareness (cf. Fairclough ed. 1992), which Coupland and Jaworski (2004) characterize (rather critically in tone) thus: The growing obsession with “good” communication (Cameron 2000) is filled by a communication industry happily producing prescriptive orders of metadiscourse which are constitutive of new social orders. These meta-discourses are based on various regimes of verbal hygiene (Cameron 1995) and codifications of communication skills. (Coupland and Jaworski 2004: 39)

The underlying intentions are certainly honorable: to contribute – on the mental level – to the social goal of emancipation. Farther reaching objectives, i.e., objectives that go beyond the associative link with the communicative aspect topicalized, may consist in securing the participants’ receptivity or even in raising conflict, in a constructive or destructive key. The last example (But I am repeating myself) and its variant (You are repeating yourself) can well illustrate these purposes. And in a still wider perspective, such metacommunicative comments may be used for defending or questioning communicative norms (the speaker subscribing to a culture-specific norm of brevity and conciseness) or even creating/modifying personal identities as communicators (profiling him- or herself as a critical, assertive person with a social status that entitles him or her to utter things of that sort). (3) Who is the target? Though comments on utterances may appear completely depersonalized, the speaker of the utterance commented on is nonetheless the ultimate target of such comments. In principle, any participant in a communication can be the target; we thus can roughly differentiate between comments directed at

Metapragmatics

111

oneself and comments directed at some other speaker (cf. Hübler and Bublitz 2007: 15–16). In principle, metacommunicative actions are reciprocal and egalitarian, wherever symmetrical relations between participants in an interaction are given – or should be given (cf. Habermas 1971). There are interactive frames, however, which do not allow for symmetry; in the classroom or the courtroom, for example, the rights for metacommunicative utterances are asymmetrically distributed. This pertains, of course, primarily to the right of commenting on the contribution of the other party. But even self-commenting may show restrictions, especially where the possibility for a certain communicative move (such as changing the topic, or launching into a longer discourse) is not readily available; one may have to choose the form of asking for permission to add something not really to the point instead of simply announcing that one is going to do that. (4) What are the forms of realization? The examples illustrating the various points have already provided a variety of examples of how a speaker’s metacommunicative intentions can be realized. Since any metacommunicative activity is based on a reflexive and diagnostic analysis, a commenting utterance seems to be the most direct equivalent, either a categorical (I am talking too much or You are talking too much respectively) or a modalized one (Do I talk too much? Maybe, I am talking too much or I think you are talking too much respectively). Another option is to operationalize the analysis in terms of a corresponding action to be taken (I should not talk that much or Stop talking that much, Could you stop talking too much? respectively). Where the target is not the speaker him/herself, such metacommunicative utterances, especially if they are critical, are likely to provoke some reaction (That’s ok or I’m always supposed to keep my mouth shut!), and this could even lead to extended sequences. As to the position of metacommunicative clauses, it is most common that they follow the utterance that they refer to. Only where the speaker takes him/herself as target may we find a reversed order, in which the utterance referred follows. This restriction, of course, is not surprising; it is, after all, only in this circumstance that the speaker of a metacommunicative utterance knows what it will be that s/he (cataphorically) refers to. A systematic treatment of how the various aspects interact would certainly be desirable, but is currently not feasible. Certain meta-phenomena have so far not been taken into account – for good reasons: Though they consist of reflexive language (in the sense of Lucy 1993) and though they contribute to the ongoing communication, they do not result in metacommunication. Their common denominator consists in their being reportive in some way or another, i.e., they refer to instances of direct speech outside the current communicative situation and either reproduce them or sum them up.9 The most common forms are direct and reported speech, which are usually framed (introduced or rounded off) by a phrase containing a verb of saying (He

112

Axel Hübler

said, “ …” or She told me that …), a performative verb (He promised to stop smoking), or one of a more interpretative character (He insisted that …). They lack, to quote Meyer-Hermann (1978: 128), the defining feature that “the object referred to and predicated about is a communicative interaction (sequence) or part belonging to the same interactional unit as the metacommunicative speech act”. These ‘metarepresentations’, as Wilson (2000) calls them, refer to speech events outside a given communicative situation and can therefore be considered ‘extracommunicative’ (cf. Ungeheuer 1970). Due to their extracommunicative character, these speech references get absorbed in the primary discourse either as instrument or object. Newspaper articles are a common source for illustration; examples of another kind are literary analyses or philosophical language essays. A phenomenon that is not extracommunicative and still represents hardly an instance of metacommunication is the so-called echo-utterance, where the speaker echoes the preceding speaker by repeating some linguistic material yet giving a specific turn to it; cf. Sperber and Wilson (1986), Wilson (2000); Graf (2007). Echo-questions (You’ve got frequent-flyer status with which company?) try to reelicit a certain part of the preceding utterance that the addressee either did not understand or feels provoked by. Echo-statements such as in the following example He: It’s a lovely day for a picnic [They go for a picnic and it rains] She: (sarcastically): It’s a lovely day for a picnic, indeed. (Sperber and Wilson 1986: 239)

usually just convey attitudes10 towards the propositional state of affairs quoted/ echoed. In neither case are we confronted with instances of metacommunication; the echo-utterances are information-oriented (in that they elicit information or evaluate information) and do not concern some formal aspect of communication. Where they cause a break, as in the information-seeking reading of the echo-question above, they do not transpose the ongoing communication onto the meta-level. Another quotative variant, which is not extracommunicative either but functions within an ongoing communicative situation, is at least a strong candidate for metapragmatics. The reference is to parody, which is defined as imitating the characteristic style of a speaker (or an author or a work of his/hers) with the intention to ridicule or achieve a comic effect. We usually associate a literary genre with it, but parody also occurs in ordinary life and everyday conversation. The illustration that follows is taken from a piece of literature, but the parodist here is not the author himself but one personage in his (realistic) drama. Nick: George: Nick: George: Nick: George

I try not to … Get involved. Um? Isn’t that right? Yes … that’s right. I’d imagine not. I find it embarrassing. (sarcastic): Oh, you do, hunh?

Metapragmatics Nick: George

113

Yes. Really. Quite. (mimicking him): Yes. Really. Quite. (Then aloud, but to himself:) It’s disgusting! (Edward Albee, Who’s Afraid of Virginia Woolf? New York: Pocket Books, 1966: 90)

The parodistic mimicking performs the metacommunicative act inexplicitly, but could be made explicit – as is illustrated in the given example by the aside with which George’s turn ends; both modes convey his disgust at the way in which Nick expresses himself. As is also shown by the example, parody – at least in everyday conversation where it is used only occasionally – shows something similar to, but weaker than, the break which marks the boundary between communication and metacommunication: the ongoing dialogic flow comes to a halt. With Bally (1965), we could characterize this inexplicit type of metacommunication as mode vécu, contrasting with the mode pur of explicit metacommunication (cf. below). Stylization, as treated by Coupland (2004), and some forms of irony provide other instances of such inexplicit metacommunication. They represent doublevoiced utterances in the sense of Bakhtin (e.g., 1981) in that they are moving on two different levels (the level of primary communication and the level of metacommunication, which consists in exaggerated imitation) and yet are united in one utterance. 1.2.

On abbreviated forms of metacommunication

The subtitle already links this part directly to section 1.1 where full forms of metacommunication were tackled. The same (meta-)pragmatic aspects are applicable here as well. They will, consequently, be expounded by the same guiding questions; merely the order will be slightly altered, starting this time with the most distinctive feature, i.e., the formal make-up. (1) What are the forms of realization? The means allowing for abbreviated metacommunication are quite easy to delineate. They mainly consist in adverbials that typically operate on the whole proposition of the given utterance. The most concise forms are disjuncts (such as frankly); slightly more ‘voluminous’ are prepositional phrases/clauses (such as in short), infinitive clauses (such as to tell you in a few words) and participial clauses (such as bluntly speaking). They usually can be expanded to a fully fledged metacommunicative utterance (I’ll be very frank with you; I will sum it up in a few words; I don’t want to fuss around but say bluntly what I have in mind). Being briefer than their elaborated counterparts, they are better integrated into the dialogical flow; the communicative break they cause is more subtle, and at best syntactically marked off from the main clause providing the propositional content of what is primarily communicated. While the full forms of metacommunication were characterized as off-line phenomena (seen from the perspective of ongoing primary communication), the abbreviated variants keep the

114

Axel Hübler

communication on-line, in a manner of speaking. It is therefore no wonder that they are unlikely to open up metacommunicative (side-)sequences. (2) What are the topics that the adverbials cover? Quirk et al.’s (1985) classifications of adverbs/adverbial provide a comprehensive overview. Of prime relevance are style disjuncts and – with some restrictions – conjuncts. According to Quirk et al., style disjuncts convey “the speaker’s comment on the style and form of what he is saying, defining in some way under what conditions he is speaking as the ‘authority’ for the utterance” (Quirk et al. 1985: 615). The metalinguistic character can (often) be brought forth explicitly by transforming the disjunct into a clause in which the adverbial (turning into a process adjunct) is linked to a verb of speaking with ‘I’ as subject. (e.g., Frankly, I am tired – I tell you frankly, I am tired).11 In line with Quirk et al. (1985: 615 ff.), we can subclassify style disjuncts. One group of adverbs that is of interest focuses on the manner in which an utterance is made; with their help, a speaker can either refer to the style of his/her utterance (e.g., bluntly, briefly, simply) or include some modal overtones (e.g., truthfully). Recall that longer variants are also available, such as prepositional phrases, or (in-)finitive and participial clauses (cf. above). The other group of interest consists of adverbs (and adverbial expressions) which characterize the make-up of a formulation, whether it has to be taken, say, figuratively or literally. Again, some expressions may be tinged with modal overtones. Thus, an utterance like Hawkins was not, strictly speaking, a traitor may address not only the issue of whether the word traitor is a suitable term for Hawkins’s behavior but also the issue of whether or not Hawkins is a traitor. In addition, there are some adverbs of degree (belonging to the category of subjuncts) that can serve the same metalinguistic purpose: compromisers (such as kind of, sort of) and approximators (such as almost, virtually) provide indications as to the reliability and adequacy of the lexical expression chosen. In uttering, for example, He is kind of a traitor or He virtually stole the money the speaker may want to let the listener know that the verdictive terms ‘traitor’ and ‘steal’ respectively are not to be taken literally, but come close. These adverbials can, in general, be linked up again with some conversational principle, norm or maxim. A speaker using truthfully in his/her utterance echoes Grice’s maxim of quality, using briefly his maxim of quantity, while with figuratively he alludes to his maxim of relevance (by flouting it), and with frankly to his maxim of manner.12 Adverbs like sort of pay tribute to Leech’s politeness principle in as far as they may render the utterance more palpable. The second set of adverbials that – in part – can be interpreted in a metacommunicative key are conjuncts/conjunctions. They are text-structural markers that specify how an utterance fits into the surrounding discourse “in which what is to follow is systematically connected to what has gone before” (Halliday and Hasan 1976: 227). Not all of the 7 conjunctive roles (and sub-roles) between which Quirk et al. (1985: 634 ff.) distinguish (i.e., listing, summative, appositional, resultive, in-

Metapragmatics

115

ferential, contrastive, transitional) are metacommunicative. The ones that do not qualify for inclusion are those that establish the co-textual link purely in terms of content, not of form. The ones whose use allow for a pragmatic interpretation as being metacommunicative are adverbs of ‘enumerative listing’ (such as first, second, third; to begin with, next, to conclude), ‘summative’ expressions (such as in conclusion, to summarize), contrastive adverbs with a ‘reformulatory’ (such as more accurately, alias, in other words) or ‘replacive’ meaning (as in alternatively, rather; better, worse) and finally, adverbials, such as incidentally or by the way, which mark ‘discoursal transition’.13 The preceding classification of metacommunicative adverbials shows – more openly perhaps than the treatment of full metacommunicative utterances does – a clear-cut topical dividing line between personal and textual thematic foci. With respect to narratives, McNeill (1992: 185 ff.) differentiates – in my opinion quite felicitously – between two levels outside the narrative proper: the meta-level, which includes the language clues characterizing and structuring the narrative as text, and the para-level which relates to the narrator – listener axis, comprising elements of an interpersonal nature (e.g., providing orientation to the listener, comments and attitudes of the narrator etc.) Applied to communication in general, we could distinguish between a meta-communicative level that would pertain to the structural aspects of dialogues and turns and the para-communicative level that would account for those non-primary forms of communication which are person-centered and focus on the interactional aspects involving the speaker and/or hearer.14 More recent treatments draw similar distinctions, though less categorically. Ädel (2006) starts out from the notion of ‘metadiscourse’ embracing various configurations that enfocus either the text, the writer (or speaker), and the reader (or hearer) and distinguishes in a second step between ‘metatext’, which relates to the text or code proper, and ‘writer-reader interaction’. Also Hyland (2005) uses ‘metadiscourse’ as cover-term; he defines it as basically interpersonal and then draws a functional distinction between means that help to guide the reader (or listener) through the text (interactive function) and those that involve the reader (or listener) in the text (interactional category). Some such terminological differentiation could help to disentangle the complexity of the concept at issue. (3) The question as to who is the target (and who has got the right to freely select the target) is easy to answer for the adverbials with a metacommunicative function. Where a speaker uses style disjuncts, s/he has necessarily him/herself as target in that his/her metacommunicative comment refers pro- or retrospectively to his or her own primary conversational contribution. Whereas the speaker as metacommunicative target is oftentimes quite transparent with such style disjuncts (cf. frankly speaking), s/he seems to withdraw from being target where conjuncts that are reflexive of discourse structure come into play. The sociological question concerning the right to select the target, in this case, boils down to the question of whether any speakers in any communicative situation may take the opportunity to

116

Axel Hübler

comment on him- or herself as communicator in such brief (adverbial) terms. This is such an elementary (and minimalistic) question that it can either be answered in strictly categorical moral terms along the line of Habermas or in empirical terms on the basis of rich data sets. (4) The final question to face concerns, again, the raison d’etre, the why or what for? The purposes that a pragmatic view can ascribe to these reduced forms of metacommunication cannot be expected to be essentially different from those identified for the extended forms with the speaker as target. But they can be expected to be somewhat reduced, according to the iconic principle of a proportionality between quantity of form and quantity of meaning. This seems, indeed, to be the case. In one way or another, they help the speaker optimize the ongoing primarylevel conversation. The style disjuncts serve him/her as (prophylactic) lubricant to avoid clashes with interactants (and subsequent repairs). The textual conjuncts facilitate the reception of the primary message emitted.

2.

Metapragmatics as the study of implicit metacommunication

Auer (1986: 22) has hailed it as a new ‘research paradigm’ and Verschueren (1998: 60) holds that it is the central area, the ‘proper domain’ of metapragmatic studies: people’s metapragmatic communicative behavior based on their metapragmatic awareness of situation and context as co-determinants of communication. No specific explicit repertoire is available to meet this meta-function. The approach conceives metalanguage as a dimension of language use and does not relate it to the object notion of metalanguage (Verschueren 1998: 55); cf. section 1. What we look for and find in utterances, instead, are indicators that testify to a metapragmatic awareness of speakers, a kind of self-monitoring activity which ‘constantly calibrates’ (to use a Silversteinian term) the communicative intention with situational and contextual conditions. The significance is highlighted by Silverstein, to whom Verschueren repeatedly refers in his argumentation. Without a metapragmatic function simultaneously in play with whatever pragmatic function(s) there may be in a discursive interaction, there is no possibility of interactional coherence, since there is no framework of structure – here, interactional text structure – in which indexical origins or centerings are relatable one to another as aggregated contributions to some segmentable, accomplishable event(s). (Silverstein 1993: 36–37)

The corresponding type of reflexive language are indexical forms, which “change their value depending on the actual event of speaking” (Lucy 1993: 10). The specific segment of reflexive language consists of shifters, contextualization cues (prosodic patterns, code switching, etc.) or implicit ‘voices’ (cf. Verschueren 1998: 61).15

Metapragmatics

117

Communicatively speaking, there is, consequently, nothing that could be paraphrased in explicit metacommunicative terms (cf. 1.2). And likewise, no neat break or boundary between a primary and a secondary level of communication can be found, and no turn to off-line communication. But at least one metacommunicative feature is preserved: a reader/listener will be lead to attend to the words instead of attending from them, to recall Anton’s (1998) subtle play with prepositions (quoted in section 1). The attending to words, more specifically to indexical words and their metapragmatic dimension, shall be further elaborated and illustrated in two points. The first point concerns the so-called shifters themselves, e.g., personal pronouns, deictic expressions, or tenses whose meanings shift due to their dependency on the individual single text and situation.16 The pronoun you, for example, acquires its specific referential meaning according to the communicative situation. This situation-dependent meaning assignment then is the result of the metapragmatic reasoning, on the speaker’s as well as the receiver’s side. More interesting and complex are such metapragmatical assignments in the case of some text-cohesive devices, i.e., reference, substitution, or ellipsis. Halliday and Hasan (1976) provide characterizations in which the metapragmatics behind their use and understanding is (unvoluntarily) well highlighted. “All three are forms of presupposition, devices for identifying something by referring it to something that is already there – known to, or at least recoverable by, the hearer” (Halliday and Hasan 1976: 144).17 With respect to reference, we read that “[b]oth exophoric and endophoric reference embody an instruction to retrieve from elsewhere the information necessary for interpreting the passage in question” (Halliday and Hasan 1976: 33). More concretely, a reference item signals ‘supply the appropriate instantial meaning, the referent in this instance, which is already available (or shortly to become available)’; and one source of its availability is the preceding (or following) text. (Halliday and Hasan 1976: 227)

In substitution and ellipsis (as a special case of substitution), the presupposition is at the level of words and structures. “When a substitute is used, it signals that the actual item required, the particular word or group or clause, is recoverable from the environment” (Halliday and Hasan 1976: 145). The second point pertains to contextualization cues, whose metapragmatic dimension is – as Gumperz (1982: 131) states – “rarely consciously noted.” He gives the following definition: [A] contextualization cue is any feature of linguistic form that contributes to the signaling of contextual presuppositions. Such cues may have a number of such linguistic realizations depending on the historically given linguistic repertoire of the participants. The code, dialect and style switching processes, some […] prosodic phenomena […] as well as choice among lexical and syntactic options, formulaic expressions, conversational […] strategies can all have similar contextualizing functions (Gumperz 1982: 131).

118

Axel Hübler

Among his examples, we find one where a graduate student is supposed to interview a black housewife in a low income, inner city neighborhood. After an appointment has been arranged over the phone by someone from the university, the student goes to the black home and is met by the husband, who opens the door, smiles and says to him So y’re gonna check out ma ol lady, hah? What he utters is a formulaic phrase “identifiable through co-occurrent selections of phonological, prosodic, morphological and lexical features” (Gumperz 1982: 133); it is marked for ethnicity and signals “both expectations about what is to be accomplished and about the form that replies must take” (Gumperz 1982: 134). That such expectations exist comes best to the fore where they are not met by the interactant and subsequently cause irritation, break-down or the like. In the example, the student does not meet the husband’s expectations by (inappropriately) replying Ah, no, I only came to get some information. They called from the office. Thereupon the husband dropped his smile and disappeared without a word while calling his wife. Sometimes, the metapragmatic reasoning behind an utterance may not be interactionally oriented, but may rather serve the maintenance and construction of identity (cf. Tajfel 1982). Identity is not only to be understood as a fixed set of properties or operations residing in the individual’s cognitive make-up, where language is regarded “as a largely docile medium through which dimensions of a person’s identity […] may receive occasional expression” (Wooffitt and Clark 1998: 107); it is also (and above all) an accomplishment “with respect to interactional and inferential concerns generated by the trajectory of verbal exchanges […], [whose] ascription is inextricably tied to the details of talk-in-interaction” (Wooffitt and Clark 1998: 107). Take the utterance Hang on, love. The term of address used could be indicative of an informal (cross-gender) relationship between speaker and addressee. In the actual instance, however, the addressee is not the speaker’s wife, but a (female) mediator the couple has turned to in their divorce case. The speaker here “constitutes his relationship with [the addressee] in a way which is at odds with the mediator-disputant relationship they have thus far collaboratively sustained” (Greatbatch and Dingwall 1998: 128). With this utterance, the speaker is questioning the addressee’s professional identity. What the metapragmatic devices under consideration all have in common – and whereby they differ from the reflexive types taken into consideration in the preceding two sections – is the mode of reflection. Bally (1965) distinguishes between the mode pur and the mode vécu; it is the latter one which we have been confronted with in the present section, whereas we were concerned with the former one previously. Bally’s distinction is worth quoting in greater detail. Si le langage n’est pas une création logique, c’est que la vie dont il est l’espression n’a que faire des idées pures. Si l’on me dit que la vie est courte, cet axiome ne m’intéresse pas en lui-meme, tant que je ne le sens pas, tant qu’il n’est pas vécu […] (Bally 1965: 15)

Metapragmatics

119

‘If language is not a logical creation, it is due to the fact that life, whose expression it represents, does not simply produce pure thoughts: If I am told that life is short, I am not interested in this idea unless I feel it, unless I experience it.’

The following illustration fits well in our context, since the Italian pur(e) is a discourse marker. Le jugement intellectuel La terre tourne se change en jugement de valeur dans la bouche de Galilée s’eciant devant ses juges: E pur si muove! Ce n’est plus une vérité scientifique, c’est l’affirmation d’une valeur attachée à cette vérité […] (Bally 1965: 15) ‘The logical judgement The globe revolves becomes an evaluative judgement in the mouth of Galileo screaming into the face of his judges E pur si muove! Then it is no longer a simple scientific truth, rather it is the affirmation of a value attached to this truth.’

Uncovering the metapragmatic dimension of language and language use in indexical forms is the objective of the corresponding field of scientific activity, which, in turn, derives its name from the very objective. Strictly speaking, metapragmatics is in this sense then the study of the metapragmatic dimension of language and speech rather than the study of metacommunication. If we were to continue in the attempt at terminologically distinguishing the various conceptualizations of ‘metapragmatics’, the term syncommunicative could well serve this purpose; the prefix syn- (of Greek origin like the prefixes meta- and para- in the complementary expressions) conveys quite adequately the fact that this reflexive level is not isolatable but inextricably linked to the primary level of communication.18

3.

Beyond interacting

Leaving ongoing conversations and taking a more generic view instead opens up perspectives to be dealt with now. Observing communication in its general design and different patterns and describing and systematizing the observations is the perspective that constitutes a first-level step away from conversation; theorizing about the conditions, presuppositions and implications of such endeavors constitutes a second-level step. While the second level is reserved for the professional linguist (cf. 3.2), the first level is open to both the professional and the ordinary conversationalists (cf. 3.1). 3.1.

Metapragmatics as the study of people’s abstracting from interacting

Through their participation in all sorts of conversations, members of a speech community gather in the course of time knowledge about conversation in general, their genres, patterns, styles, norms, etc. Since this knowledge relates to the pragmatic dimension of communication, it may be specified as metapragmatic knowledge. Some similar metapragmatic knowledge may be obtained by the scientist who

120

Axel Hübler

studies the conversations of this speech community as an outsider yet in the role of a participating observer. In other words, it is hardly possible to clearly distinguish between the scientific metapragmatic knowledge and the metapragmatic knowledge of the users of a language. The assumption of a continuum between both stances, as made by Caffi (2006: 87), might therefore seem quite appropriate;19 the difference consists merely in a smaller or greater distance from the object, viz.: the use of the language and a lower or higher degree of explicitness. Folk theories, however, call for the mediating linguist, the linguist, that is, who elicits folk knowledge or elaborates on their shorthand versions. Gumperz (1982) identifies as mediating linguists in particular those working in the fields of discourse analysis or ethnography of communication. Folk-linguistics is a separate discipline (cf. Niedzielski and Preston 2000), but has roots in both. In any case the task is “to come to grips with the varying ways in which linguistic behavior is conceptualized by those engaged in it” (Verschueren 1998: 60–61). Of the folk-linguistic topics, speakers’ awareness of and knowledge about pragmatic issues is, of course, particularly relevant for the current context. Two areas and modes (cf. Preston 1996) shall be considered in greater detail. (1) A major source of information about people’s metapragmatic knowledge could be the lexical repertoire they have at their disposal for referring to aspects of speech; it is not an autonomous language, but a subset of the ordinary language. Speech act verbs are a favorite research area (where folk taxonomies merge with the empirical-conceptual approach of speech act theories and the ethnography of speaking, cf. Caffi 2006: 85). But other lexical expressions relating to other selected aspects of communication could also be common ground where folk- and empirical-conceptual linguistics meet. Lexical expressions, in general, can be interpreted as a means that a speech community uses for coding cognitive and cultural models, which consist in more or less coherent sets of concepts for structuring experience (cf. also Gee 1999: 40 ff.). Experience also includes acting by communicating, and this, in turn, calls for adequate conceptualizations of a metapragmatic kind and their storage in words. One of the most comprehensive collections of English speech act verbs is the dictionary compiled by Ballmer and Brennenstuhl (1981).20 Its comprehensiveness derives from two decisions. (a) They apply a much wider definition of ‘speech act verb’ by including “all those verbs which designate (aspects of) speech activities” (Ballmer and Brennenstuhl 1981: 3), no matter whether they can be used performatively or not; thus lie and persuade, for example, are speech act verbs just as well as admit or promise. (b) They include literal as well as metaphorical expressions, as, for example, contradict and reject. Of particular interest is the authors’ ordering of the gathered material into a huge word field and its many subfields, because these can be understood as identifying the metapragmatic aspects of communication that the English speaking community takes recourse to. The word fields in their entirety unfold a global metapragmatic conceptualization whose

Metapragmatics

121

main constituents delineate models entering into an integral system of speech activities. Each model has as its linguistic correlate a set of verbs, which again have substructures. The ‘enaction model’, for example, consists of verbs, of which wish, aim at, or anticipate express VOLITION , alert s. o., indicate, or refer to PUTTING INTO FOCUS , while expressions of offering, asking, commissioning, ordering, warning etc. are instances of ENACTION , and verbs like accept, refuse, or obey instances of REACTIONS ON ENACTIONS . At the most refined level, we find sets of verbs that show a high degree of synonymy; for the enactment category SURPRISING , for instance, the following verbs are listed: amaze, astonish, astound, bewilder, disconcert, flabbergast, shock, stagger, startle, strike, surprise. Taken together they delineate, if we will, how the English language community conceptualizes acts of surprising somebody by words; seen in a differentiating way, each single item contributes a special nuance to the overall-concept, in comparison to the others. Along these lines, Goossens (1987) has made an analysis of some such nearsynonymous verbs. His treatment of the speech act verbs say, tell, talk and speak can well illustrate how much an in-depth analysis is able to reveal about people’s metapragmatic models, though it explicitly tries just “to answer a typically Fillmorean question, namely, how do linguistic action verbs frame the scene of linguistic (inter)action?” (Goossens 1987: 95). Thus, native English speakers assign different roles to receivers, encoding in tell and say the receptor who plays a more passive part, and in talk the interactor who plays a more active, participative part; in the case of speak, the English speaker accounts for the two roles by using distinct prepositions, with where the receptor acts as interactor, to where as receptor (cf. Goossens 1987: 103). The four verbs reveal, furthermore, that native English speakers differentiate messages according to the degree to which they have been condensed. This is reflected in the facts that, for example, say “freely combines with direct enunciations,” while tell “is considerably restricted in taking them” (Goossens 1987: 103); the differentiation is in principle even reflected where the distributional pattern is reversed, in other words where more extreme condensations go with tell, but not with say (e.g., tell the truth vs. *say the truth). Lexical expressions that belong to certain metaphorical concepts offer particularly rich information about people’s cognitive metapragmatic models.21 Metaphors have in general attracted much attention by linguists because they seem to offer insights into how the human mind operates. As Lakoff and Johnson state in their pivotal study of 1980: “[M]etaphor is pervasive in everyday life, not just in language but in thought and action. Our ordinary conceptual system, in terms of which we both think and act, is fundamentally metaphorical in nature” (Lakoff and Johnson 1980: 3); and metaphors that conceptualize communication in one respect or another have been paid attention to right from the very beginning. For an illustration, we may turn to two of the most widely known examples, i.e., the so-called conduit metaphor, identified by Reddy (1979) and differentiated later by Vanparys (1995) and Semino (2006), and the ‘argument is war’ concept

122

Axel Hübler

proposed by Lakoff and Johnson (1980) and modified by Grady (1998) and Semino (2006). The results of these investigations, derived from a close analysis of sets of metaphorical expressions in general use and subsequently supplemented by corpus data, can be understood as reconstructions of native English speakers’ conceptualizations of certain aspects of how they communicate and what their form of communication is like. These conceptualizations are metapragmatic in nature. As far as communication between communicators in general is concerned, the dominant meta-pragmatic model people have developed shows three main constituents: the speaker/sender puts an idea-object into a word-container and sends it (through a conduit) to the receiver, who extracts from the word-container the ideaobject. Discussion/argumentation represents a specific form of communication/discourse. The corresponding metaphors widely in circulation reveal that English speakers share a (metapragmatic) model of argumentative discourse that centers around physical conflict; debatants have two positions, attacking the other’s opinion, while defending one’s own; we have phases of maneuvering and retreat; the outcome amounts to truce or victory/surrender. (2) Another important source for gaining insights into people’s understanding of communication would consist in information that researchers are able to elicit directly from members of the speech community; this is a method which is often applied by discourse analysts, where “the aim is to produce ideotypical descriptions [of concrete discourses] that can be dissected into significant components” (Gumperz 1982: 157). But the results are on the whole quite unsatisfying; ordinary conversationalists are obviously “poor descriptors” (Gumperz 1982: 157),22 and this should hold true even more, were they confronted with a more abstract and complicated task. Alternatively, an experimental way of eliciting metapragmatic knowledge may be feasible (similar to the method that sociologists, e.g., Goffman, occasionally deploy, for discovering certain norms which people subscribe to in ‘relations in public,’ i.e., by confronting them with transgressions), but has not, to my knowledge, been pursued. The use of questionnaires seems to be the relatively more reliable method. The study by Simon-Vandenbergen (1995) shall be used to further illustrate the point at stake, even if it is not exactly made for such a purpose. Its aim was to obtain assessments of linguistic behavior from English native speakers. The method consisted in having natives evaluate selected aspects and forms of communicative behavior through evaluating corresponding (metaphorical) expressions, which, as we saw, embody parts and aspects of folk models of behavior. The questionnaire applied included expressions covering social functions of talk (e.g., conversation, chat, or prattle), turn-taking (e.g., run on, cut someone short, step in), topic management (e.g., bring up, move unto, meander), and manner of speaking (e.g., rattle, babble, or air, thunder, or tell someone flat out, wrap up one’s meaning). The questions/tasks that were to stimulate statements about people’s metapragmatic views

Metapragmatics

123

had three formats: 1. (social) assessments in terms of positive – negative – neutral, 2. (aesthetic) assessments in terms of light – heavy – neither, and 3. the same as format 1 but supplemented by (a) a question asking why the subject rated the expression the way s/he did and (b) the request to give an ‘example sentence’. While certainly all tasks can be interpreted as throwing light on people’s understanding of communication, it is the supplementary task (a) which should elicit the most; unfortunately, Simon-Vandenbergen does not provide any cue as to what and how subjects met this task. Nevertheless, some of the results she obtained allow conclusions. Speech tempo is a good case in point. This is the picture emerging for native English conversationalists (cf. Simon-Vandenbergen 1995: 108): Speed is an important criterion; high and low speeds are often negatively evaluated as ‘too fast’ and ‘too slowly’ respectively, in which case they tend to be talked about in metaphorical terms. One could, of course, object and claim that the negative evaluations elicited by the questionnaire were just triggered by the metaphoricity of the expressions themselves. Any such objection, however, would not be justified. As Simon-Vandenbergen already seems to imply, it is the perceived marked value of the communicative phenomenon in relation to norms and standards in a given community that triggers metaphorical expressions and assessments, and not the metaphorical expression that triggers a corresponding evaluation. Besides, in the case of speed, there is even some independent support from social-psychological research for the claim that the negative evaluations relate, indeed, to speed and are not simply triggered by the metaphoricity of the lexical expression. Smith et al. (1975), for example, had subjects listen to recordings of speakers performing one and the same text with different speeds; in reality, it was just one speaker whose natural speech tempo had been technically manipulated. The subjects had to judge the ‘speakers’ as to ‘benevolence’, a social criterion which aligns well with Simon-Vandenbergen’s positive-negative scale. The authors found that “[t]he benevolence/rate plot reveals an inverted U-relationship, with the mean for normal voices […] being rated more benevolent” (Smith et al. 1975: 150), while high- and low-speed ‘speakers’ scored significantly low in benevolence. The insights obtained by metapragmatic research outlined so far can be enriched (and maybe even relativized) by widening the scope so as to also include possibilities of variation, always applying the same analytical parameters.23 Of the three main types of variation, i.e., social, developmental, and historical, the social variant has not yet received due attention. Illustrations will therefore be restricted to the other two, turning to the lexicological source first. As Geeraerts and Grondelaers (1995) point out, if cognitive models are also cultural models, they are also cultural institutions, and as such, they carry their history along with them: their institutional nature implies their historical continuity. It is only by investigating their historical origins and their gradual transformation that their contemporary form can be properly understood. (1995: 177)

124

Axel Hübler

In line with this stance, Arnovick (1999) has examined changes in the conceptualization of promising: The frequency of promises rendered defective at inception or follow-through has created a practical (if not moral) dilemma, a problem with promises. People do not always trust the promises they hear. As a result, the need to convince the hearer is often assumed by the speaker. To perform a promise that satisfies the hearer, the speaker often has to emphasize his or her commitment. (Arnovick 1999: 60)

She notes that even a performative use of promise, which is already a very strong form of promising (cf. Traugott 1997), is nowadays often felt to be insufficient and to require some further declarations, such as those observed with George Bush: Read my lips, I guarantee it, Believe you me. “A simple ‘I will’ nowadays works only in a promissory situation and an extra-linguistic institution for its force” (Arnovick 1999: 58), in contrast to OE times when both the use of sculan (expressing a subject’s obligation) and wilan (expressing a subject’s intention) were binding on the speaker. Nowadays, promises usually undergo relativizations and rationalizations: promises merely express honorable intentions; changing circumstances can render vows unrealistic; insincerity is politically motivated, deception is committed ‘for the sake of the task’ pursued. Such views are further warranted by references to independent proposals (by Muck 1989 and Rawls 1955) for constructing “theories of promising that incorporate a limit on speaker responsibility” (Arnovick 1999: 60). The historical and culture-bound relativity of metaphorical concepts on the pragmatics of language can well be illustrated by two investigations of mine. In Hübler (1998), I showed that the predominant spatial metaphorization relating to the notion of ‘express,’ which represents the first constituent within the conduit paradigm resulted from a change that had taken place around the turn from the 16th to the 17th centuries. The argument centered around the change in the use of prepositions going with the notion of ‘express;’ it was a change from the instrumental ‘with’ to the spatial ‘in’ (words). The former conceptualization aligns well with and foreshadows the tools-paradigm proposed by Reddy (1979) as a possible alternative to the predominance of the conduit metaphor. And in Hübler (ms), I examined data from the 16th and 17th centuries which show that argument was conceived during that period in less fierce and more playful terms, more specifically in terms of a tennis match; Grady’s (1998) findings can be taken as a diluted late version of this early concept. A discoursal source is tapped on by Gotti (2006). He shows for the 16th/17th century that neologizing (in contrast to the practice nowadays) was still a matter to talk and write about; it was the time when a scientific language developed in England, by taking over elements from the common code. The following passage can be read as an instance of the public discourse about it, particularly about using new words of foreign origin in English.

Metapragmatics

125

But now as touching myne entent in writing this treatise in the English. Though this cause might seme sufficient to satisfy many men that I am an englysh man, and therefore may more easely and plainly write in my natyue tonge, rather then in any other: yet vnto them that know the hardness of the mater, this answer shuld seme vnlykely: considering that it is more harder to translate into such a tonge, wherein the arte hath not ben written before, then to write in those tongues that are accustomed, and (as I might say) acquainted with the termes of the science (Recorde 1547, quoted in Gotti 2006: 218)

Recorde was a physician and mathematician. His topic in this passage (probably from the preface) is metalinguistic and concerns the (un-)availability of English words for certain concepts. But the discourse itself could be a case for metapragmatics (in the sense discussed in section 1); the passage is metacommunicative in that the writer raises the problem of effability in English and thus justifies his use of un-English words in the treatise to follow. In the current context, however, the quotation can be seen as documenting metapragmatic knowledge about the limits and chances of informing an uninformed readership about specific and new states of affairs. Representing the metapragmatic knowledge of a member of an educated 16th-century English speech community, the document could also give grounds for comparing it with (in-)effability treatments by speakers from different social-cultural backgrounds; but I do not know of any such investigation. The developmental type of variation in metapragmatic knowledge is, in contrast, well investigated. For the purpose of illustration, the article by Bernicot and Laval (1996) has been chosen, because it focuses, again, on promises. Their subjects are children of three age groups, around 3, around 6 and around 10 years of age. The children’s metapragmatic knowledge is defined as their capacity to talk and think about acts of keeping and breaking promises and tested by examining their verbal comments stimulated by a series of cartoon stories; each story consists of a promise (by a child or a parent) and the subsequent fulfilment or non-fulfilment; the effect of sadness (as sign of dissatisfaction) and happiness (as sign of satisfaction) respectively, which the (non-)fulfilment has on the person to whom the promise was made, has then to be determined by the subjects themselves and subsequently justified. The authors summarize their results thus: Metapragmatic knowledge was found to evolve with age. At the ages of 3 and 6, children’s metapragmatic knowledge mainly concerns the execution of the action. At age 6, the listener’s desires start being added in cases where [this] preparatory condition is not satisfied. At age 10, explanations pertaining to action accomplishment completely disappear, and explanations about the speaker’s intentions alone or about both the speaker’s intentions and the listener’s desires appear. (Bernicot and Laval 1996: 120)

The metapragmatic knowledge of adults would at least also comprise the sincerity condition of promises, i.e., that the speaker intends to accomplish the future action (cf. Gibbs and Delaney 1987).

126 3.2.

Axel Hübler

Metapragmatics as metatheory of pragmatics

So far, we have dealt with cases where speakers provide the data/material, which linguists then process, refine (as was the case with metacommunication) or elaborate on (as in the case of speakers’ theoremes). We now leave the sphere where linguists occupy themselves with other people’s products and turn to their own activity, subjecting it to reflection.24 That portion of such reflections which becomes manifest constitutes a professional theoretical dialogue (mainly in the written medium) about pragmatics, with the aim of defining the field, supporting, supplementing, or justifying certain theoretical stances or challenging them. Such theoretical discussions show, in principle, two orientations. They either operate introspectively or involve external considerations, and sometimes these perspectives merge. Since pragmatics as such is already the most complex language-related discipline, it is easy to understand that the meta-discourse about it is immense and open-ended; problems abound. The few topics that will subsequently be addressed can, therefore, do nothing else but give some indications. The problems already start on the most elementary of levels, including even the question what pragmatics is all about. What, in other words, is the object of pragmatics? Is it human communication or just the verbal part of it? The answer may depend on the answer given to an even more elementary question: How to define the scope of activity for a linguist working in the field of pragmatics? Is it the discipline that defines the scope or is it the object as such? In that particular case, pragmatics originated in philosophy but was subsequently fully absorbed by linguistics; should it therefore remain a linguistic discipline and restrict its activities accordingly? According to what? Was it not one of the major figures of linguistics, de Saussure, who embedded linguistics in semiotics, which consequently would license any occupation with the non-verbal as well? If, on the other hand, it is the object which defines for scholars of pragmatics the scope of their activities, then the non-verbal with its phonetic and kinesic domains would be included anyway because natural communication, even any single utterance, can only be fully understood as a trimodal phenomenon. It is not even justifiable to claim that the verbal part plays the main role because intonation and/or gestures oftentimes modify in an important way what is verbally said; they sometimes even override the verbally mediated meaning. Even if one were to subscribe to the view that, ideally speaking, the nonverbal modes should be attended to as well, it could still be argued that such a holistic approach is simply not practicable; several disciplines should therefore participate and join in a common endeavor. But how much compartmentalizing does the subject matter tolerate? Interdisciplinary research may be a good answer, but transdisciplinarity may still be unavoidable up to a certain degree. Linguistics and communication research did not fall out of the sky; its predecessor was rhetoric, the art of persuasion. Traces of this tradition of thinking can

Metapragmatics

127

still be found in Searle (1969) in that he keeps the link alive that exists between a speech act and its effect, the perlocution. Though the ensuing research paradigm widened the scope so as to treat the speech act as part of speech act sequences, it may cause wonder why perlocution never played a role. Why has research in indirect communication strategies been restricted to what the speaker ‘plots’ – with the addressee underdetermined; why has research not systematically investigated the (perlocutionary) effects that certain strategies are likely to cause in receivers? Because social psychology is doing this job? At some US universities, departments of ‘Communication’ (not tantamount to mass communication) have been established whose activities are at the cross-roads of various lines of research, including linguistics, rhetoric and persuasion, and social psychology. Is that a more adequate discipline for coming to terms with interpersonal communication? Or what do they lack that the (European) language-biased approach has to offer? Questions of involving other disciplines do not only represent “boundary issues” (Caffi); they may also be raised and pursued with pleasure when they help ‘dignify’ central views or important findings of one’s own discipline. The dignifying effect results from their proving that the views or results are not isolated, but align well with what is considered to be more fundamental in the sense that it provides the foundations of one’s own discipline, its overarching rationale. A good case in point is the ever growing theoretical horizon associated with pragmatics. The cognitive turn in the humanities necessitated a cognitive foundation also for pragmatics. Since the cognitive sciences, however, have (neuro-)biological foundations, to what degree are they to be paid tribute? Is it enough to allude every now and then to such deep connections? Is it acceptable to do so only when it seems opportune, i.e., when it serves the argument at issue, and neglect them otherwise? McNeill’s treatment of (everyday conversational) narratives provides an interesting example. A central part of his production theory, which (to varying degrees) takes all three modes of realization into consideration (i.e., language, intonation, and gestures), operates with a model of self-organization that obviously has its roots in the theory of ‘autopoietic’ (or self-referential, self-organizing) systems, which goes back to Maturana (1970a, 1970b). McNeill, however, does not discuss (metapragmatically!) these roots. Was it because the theory is difficult for students of interpersonal communication to ‘digest’?25 Where are the limits, and how far do we have to go, in how much detail? Many of the questions raised (as well as others that could be raised) still await answers. Whatever the answers are or will be, the fundamental nature of these issues (and even the questions themselves) are rarely free of any ideological bias, whether one wants to admit it or not. A sensibility for the ideological dimension of research in pragmatics cannot be taken for granted; it seems, in fact, that it still needs to be called in. “[A] constant monitoring of linguistic rhetoric in view of the ideological underpinnings of theories and analyses is not a luxury but a prerequisite for the advancement of linguistic pragmatics” (Verschueren 1998: 67). How

128

Axel Hübler

subtle the dangers of ideological pitfalls can sometimes be is well highlighted by Caffi. [O]ne has to avoid the risk of new forms of hypostasis […], brought about by certain, apparently harmless, metalinguistic uses. For instance, the use of a definite noun phrase can be idealizing: the risk of an idealization exists whenever one refers to ‘the speaker,’ who could end up by being an hyperuranian subject, or a lunar effet-sujet. (Caffi 2006: 83)

But more tangible issues have, of course, to be subjected to such self-examination as well. What is the researcher’s stance to the speech community that s/he wants to investigate about, even towards the individual informant? Does the stance have any political implications? And on a methodological level, are the questions asked (orally or by means of a questionnaire) politically correct? A subspecies of this kind of metatheoretical problem consists in the question of whether the methods of tackling pragmatically relevant data is neutral enough to cope with different yet comparable sets of data. The problem of culturally biased approaches is addressed here as well. Among the English speech act verbs, for example, we find none that would have lexicalized any specification as to time and location (as in Tenejapa Tzeltal, where concepts like ‘talk that has taken place just a short time ago’ or ‘talk occurring in a grassy area’ are lexicalized), but we do find English speech act verbs of directing, such as appoint, convene or ordain, which include specifications as to social setting and institution (cf. Verschueren 1987: 131–133). A simple speech act classification would not do because it would neglect the specificities. In the face of such shortcomings, Verschueren proposes to “avoid attempts at classifying linguistic action verbs (and the corresponding acts) and […] concentrate instead on the careful contrastive analysis of the semantic dimensions involved in the lexicalization of linguistic action” (Verschueren 1987: 126). The descriptive framework he offers is supposed to allow for accommodating (by way of comparing) intra-linguistic, cross-linguistic and cross-cultural differences. A (partial) translation of Verschueren’s program into action is Mey’s analysis of a (by now familiar) speech act, ‘promise’. [T]he context in which a promise is made is of the utmost importance for its status as a promise and for its binding effects. Take the case of a young person promising his or her parents not to smoke before the age of eighteen. In this case, the societal conditions surrounding the execution of such a promise can be exceedingly difficult (peer-group pressure, work conditions, etc.). (Mey 1993: 170)

But it serves the author to underline a metapragmatic argument: the view that a philosophical speech act theory à la Austin (21990) and Searle (1969) may not suffice for linguistic purposes. Both Austin and Searle operate on the ‘one sentence – one case principle’. That is to say, in order to illustrate their theory, they use examples that are characteristic of what they

Metapragmatics

129

see as the ‘case’ under discussion, such as, e.g., a particular kind of speech act. I will, in the following, show the shortcomings of the ‘case approach’ (Mey 1993: 170).

The example also demonstrates that metapragmatic discourse does not always have to be the raison d’etre for an entire (theoretical) text. It often has, rather, a more limited purpose, i.e., to profile one’s own theoretical assumptions against the backdrop of what is widely held to be the case. This pattern is pervasive and can be found in almost all scholarly work in whatever discipline; but it is of course only in pragmatics that such metatheoretical discourse is metapragmatic. The illustration to follow is taken from a prominent source. A more radical difference between Grice’s approach and relevance theory [as proposed by Sperber and Wilson, A.H.] is this. Grice’s principle and maxims are norms which communicators and audience must know in order to communicate adequately. Communicators generally keep to the norms, but may also violate them to achieve particular effects; and the audience uses its knowledge of the norms in interpreting communicative behaviour. The principle of relevance, by contrast, is a generalisation about ostensiveinferential communication. Communicators need no more know the principle of relevance to communicate than they need to know the principles of genetics to reproduce […] (Sperber and Wilson 1986: 162).

The quotation is an excerpt from a short metapragmatic passage of one and a half pages, where Sperber and Wilson confront their concept of relevance with that of Grice (e.g. 1975). The metapragmatic discourse in this case is even sparked off by a metalinguistic problem, i.e., the fact that both parties use one and the same term, but associate it with different concepts. The passage, however, is not metacommunicative because concepts, methods and the like are the issue, not the communicative form or aspects of it. This is a feature that generally applies to any form in which pragmatics is theoretically reflected. It is a type of reflection that is at best accidentally language-reflexive. This sets it apart from all the other phenomena that have been taken account of in this article and aligns it with the genre metatheory. ‘Metapragmatics’ may be, in the end, an all too generous term.

Notes 1. On July 31, 2009. 2. This is basically what Caffi (2006) takes account of as ‘metapragmatics 2’. 3. Applying the typology of metamessages proposed by Jakobson (1971 [1957] 130–133), we are mainly concerned with type (M/M), i.e., messages about messages. Certain attention will also be given to type (C/M), that is to those metamessages in which so-called shifters (as the code involved) incorporate a message about a message. Messages about code (M/C) as well as the meta-message type (C/C), where proper names are the protagonists, fall outside the scope of metacommunication proper. Cf. also Lucy (1993: 16). 4. In section 2, a different concept will be introduced, which conceptualizes metalanguage as a dimension.

130

Axel Hübler

5. Metalanguage in the sense outlined, however, is part of but not equivalent to reflexive language as described by Lucy (1993) – just as speech reflexive activity/ reflexive language use is not tantamount to metacommunication. 6. Such reduced forms will be tackled separately below. 7. Quite extensive studies along these lines have been conducted earlier by Meyer-Hermann (1980 [unpublished]). Lumping together topic and function, he identified 72 metacommunicative speech act types (on the basis of French and Portuguese material). The 11 most frequently used types are: requesting the performance of speech acts, acknowledging the performance of speech acts, announcing the performance of speech acts, describing communicative interaction, guiding discussions, criticizing communicative interaction, restating a communicative function more precisely, eliciting receptive readiness, securing comprehension, evaluating communicative interaction, and anticipating sanctions prophylactically (cf. Welte 1990: 178–180). They all have been included in one way or another in the present outline. 8. Meyer-Hermann (1980), in fact, tackles functions only, the topics being implied. 9. Quoting, in fact, is the phenomenon that Jakobson’s meta-message type (M/M) was obviously meant to mainly accommodate. 10. As Wilson (2000: 148) points out, “[the] attitudes conveyed by echoic utterances can be very rich and varied: the speaker may indicate that she agrees or disagrees with the original, is puzzled, angry, amused, intrigued, skeptical etc., or any combination of these.” 11. Often the disjunct may have a paraphrase with a verb of saying taking the form of a finite clause (If I may say so without offence, your writing is immature). 12. Hagemann (1997) conducted a detailed investigation on German ‘diction-characterizing’ adverbs in Gricean terms. 13. I am not sure about the status of adverbs such as well which can either function as discourse initiators (cf. Quirk et al. 1995) or continuation markers (cf. Halliday and Hasan 1976). 14. The topical distinction as such has been highlighted, under different labels, by various authors; cf. Bamford and Biondi (2005: XVII), Hagemann (1997: 27–28), Geissner (1981: 206). Watzlawick et al. (1968) and Bateson (1972) focus exclusively on the interpersonal dimension and reserve for this aspect the term ‘metacommunication’. 15. In Jakobson’s typology (cf. footnote 3), they would represent metamessages of the type (C/M). 16. Jakobson (1971: 131), who adopted the concept from Jespersen (1922: 123), discusses “shifters” at some length. 17. Halliday and Hasan’s use of the term ‘presupposition’ differs, of course, from its usual semantic-pragmatic reading; they use it in the wider sense that “something is to be supplied, or understood” (Halliday and Hasan 1976: 144). 18. Cf. also analogous term formations such as ‘synsemantic.’ 19. But caution is advised. “Speakers’ awareness […] does not have to match the linguist’s metapragmatic descriptions. Naively confusing these two may contribute to the furthering of a specific kind of folk metapragmatics inherent in western linguistics itself. Assuming that such naïve confusion can be avoided, at least some of my further comments will be based on results obtained in this line of research” (Verschueren (1998: 60–61). 20. In certain respects complementary, but of a narrower scope, is Schmid (2000) with his investigation of ‘shell’ nouns, a subset of which overlaps with notions taken account of by Ballmer and Brennenstuhl (1981).

Metapragmatics

131

21. For obvious reasons, I simply take both kinds for granted, though it is certainly a matter worth discussing; cf. Kövecses (2005: 193 ff.) and his discussion of the question whether cultural models can exist without prior metaphorical understanding, as claimed by Quinn (1991). 22. “When participants report on actual verbal encounters, they tend to do so by mentioning some item of content, or by referring to what people were getting at …” (Gumperz 1982: 157). 23. For questioning the parameters, cf. section 3. 24. This overlaps with what Caffi characterizes (under ‘metapragmatics 1’) as “concern for the epistemological implications of what researchers assume pragmatics is about” (Caffi 2006: 83). 25. The reader may test it, in the original or in a version such as Benseler et al. (eds.) 1980, which has communication scientists as its target audience.

References Ädel, Annelie 2006 Metadiscourse in L1 and L2 English. Amsterdam: Benjamins. Anton, Corey 1998 ‘About talk’: The category of talk-reflexive words. Semiotica 121/3,4: 193–212. Arnovick, Leslie K. 1999 Diachronic pragmatics. Seven case studies in English illocutionary development. Amsterdam: Benjamins. Auer, Peter 1986 Kontextualisierung. Studium Linguistik 19: 22–47. Austin, John L. 1990 How to do things with words. Edited by James O. Urmson and Marina Sbisà. 2nd. ed. Oxford: Oxford University Press. Bakhtin, Michail M. 1981 The dialogic imagination. Edited by Michael Holquist translated by Emerson, Caryl and Michael Holquist. Austin, TX: University of Texas Press. Ballmer, Thomas and Waltraud Brennenstuhl 1981 Speech act classification. A study in the lexical analysis of English speech activity verbs. Berlin: Springer. Bally, Charles 1913 [1965] Le langage et la vie. Genève: Ed. Atar, [reprinted Genève: Librairie Droz]. Bamford, Julia and Marina Biondi 2005 Introduction. In: Julia Bamford and Marina Biondi, Dialogue within discourse communities. Metadiscursive perspectives on academic genres. Tübingen: Niemeyer. I–XXIV. Bateson, Gregory 1972 [1955] Steps to an ecology of mind. New York: Balantine. Benseler, Frank, Peter M. Hejl and Wolfram K. Köck (eds.) 1980 Autopoiesis, communication and society. The theory of autopoietic system in the social sciences. Frankfurt, New York: Campus.

132

Axel Hübler

Bernicot, Josie and Virginie Laval 1996 Promises in French children: Comprehension and metapragmatic knowledge. Journal of Pragmatics 24: 101–122. Bublitz, Wolfram and Axel Hübler (eds.) 2007 Metapragmatics in use. Amsterdam: Benjamins. Caffi, Claudia 2006 Metapragmatics. Encyclopedia of Language and Linguistics. Amsterdam: Elsevier. 82–88. Cameron, Deborah 1995 Verbal hygiene. London: Routledge. Cameron, Deborah 2000 Good to talk? Living and working in a communication culture. London: Sage. Coupland, Nikolas 2004 Stylised deception. In: Adam Jaworski, Nikolas Coupland and Dariusz Galasinski (eds.), Metalanguage. Social and ideological perspectives, 249–274. Berlin: Mouton de Gruyter. Coupland, Nikolas and Adam Jaworski 2004 Sociolinguistic perspectives on metalanguage: Reflexivity, evaluation and ideology. In: Adam Jaworski, Nikolas Coupland and Dariusz Galasinski (eds.), Metalanguage. Social and ideological perspectives, 15–51. Berlin: Mouton de Gruyter. Fairclough, Norman (ed.) 1992 Critical language awareness. London: Longman. Gee, James P. 1999 Discourse analysis. London, New York: Routledge. Geeraerts, Dirk and Stefan Grondelaers 1995 Looking back at anger: Cultural traditions and metaphorical patterns. In: John R. Taylor and Robert E. MacLaury (eds.), Language and the cognitive construal of the world, 153–179. Berlin: Mouton de Gruyter. Geissner, Hellmut 1981 Sprechwissenschaft. Theorie der mündlichen Kommunikation. Königstein im Taunus: Scriptor. Gibbs, Raymond W. and Suzanne M. Delaney 1987 Pragmatic factors in making and understanding promises. Discourse Processes 10: 107–126. Goffman, Erving 1972 Relations in public: Microstudies of the public order. New York: Harper and Row. Goossens, L. 1987 Dealing with linguistic action verbs in depth. In: Jef Verschueren (ed.), Linguistic action: Some empirical-conceptual studies. Norwood, NJ: Ablex. 95–106. Gotti, Maurizio 2006 Metalinguistic considerations in 17th-century specialized discourse. In: Anna Duszak and Urszula Okulska (eds.), Bridges and barriers in metalinguistic discourse, 217–231. Frankfurt: Peter Lang.

Metapragmatics

133

Grady, Joseph 1998 The ‘Conduit’ metaphor revisited: A reassessment of metaphors for communication. In: Jean-Pierre Koenig (ed.), Discourse and cognition: Bridging the gap. Stanford, CA: CSLI Publications. 205–218. Greatbatch, David and Robert Dingwall 1998 Talk and identity in divorce mediation. In: Charles Antaki and Sue Widdicombe (eds.), Identities in talk, 121–132. London: Sage. Graf, Claudia 2007 Echo und Zitat. In: Elke Brendel, Jörg Maibauer and Markus Steinbach (eds.), Zitat und Bedeutung. Linguistische Berichte, Sonderheft 15: 89–109. Grice, H. Paul 1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and semantics. Vol. 3: Speech acts. New York: Academic Press. Gumperz, John J. 1982 Discourse strategies. Cambridge: Cambridge University Press. Habermas, Jürgen 1971 Vorbereitende Bemerkungen zu einer Theorie der kommunikativen Kompetenz. In: Jürgen Habermas and Niklas Luhmann, Theorie der Gesellschaft oder Sozialtechnologie, 101–141. Frankfurt: Suhrkamp. Hagemann, Jörg 1997 Reflexiver Sprachgebrauch: Diktumscharakterisierung aus Gricescher Sicht. Opladen: Westdeutscher Verlag. Halliday, Michael A.K. and Ruquaia Hasan 1976 Cohesion in English. London: Longman. Hübler, Axel 1998 The expressivity of grammar. Grammatical devices expressing emotion across time. Berlin: Mouton de Gruyter. Hübler, Axel o.J. The conceptualization of argument in terms other than war. A socio-cultural case study. (Ms) Hübler, Axel and Wolfram Bublitz 2007 Introducing metapragmatics in use. In: Wolfram Bublitz and Axel Hübler (eds.), Metapragmatics in use, 1–26. Amsterdam: Benjamins. Hyland, Ken 2005 Metadiscourse. Exploring interaction in writing. London, New York: Continuum. Jakobson, Roman 1971 [1957] Shifters, verbal categories, and the Russian verb. In: Selected writings II: Word and language, 130–147. The Hague: Mouton. Jespersen, Otto 1922 Language. London: Allen and Unwin. Kövesces, Zoltan 2005 Metaphor in culture. Universality and variation. Cambridge: Cambridge University Press. Lakoff, George and Mark Johnson 1980 Metaphors we live by. Chicago: The University of Chicago Press.

134

Axel Hübler

Leder, Drew 1990 The absent body. Chicago: The University of Chicago Press. Lucy, John A. 1993 Reflexive language and the human disciplines. In: John A. Lucy (ed.), Reflexive language. Reported speech and metapragmatics. Cambridge: Cambridge University Press. 9–32. Maturana, Humberto R. 1970a Biology of cognition. Report 9. Biological Computer Laboratory, Department of Electrical Engineering. University of Illinois, Ill. Maturana, Humberto R. 1970b Neurophysiology of cognition. In: Paul L. Garvin (ed.), Cognition: A multiple view, 3–23. New York: Spartan. McNeill, David 1992 Hand and mind. What gestures reveal about thought. Chicago, London: The University of Chicago Press. Mey, Jacob L. 1993 Pragmatics. An introduction. Oxford, Cambridge: Blackwell. Meyer-Hermann, Reinhard 1978 Aspekte der Analyse metakommunikativer Interaktionen. In: Reinhard MeyerHermann (ed.), Sprechen – Handeln – Interaktion, 103–142. Tübingen: Niemeyer. Meyer-Hermann, Reinhard 1980 Studien zur Funktion von Metakommunikation am Beispiel portugiesischer und französischer Alltagssprache. Habilitationsschrift Universität Bielefeld. (Ms.) Muck, Bruce 1989 Promising rules and intentions. PhD Diss. University of Wisconsin. Niedzielski, Nancy A. and Dennis R. Preston 2000 Folk linguistics. Berlin: Mouton de Gruyter. Paul, Ingwer 1999 Praktische Sprachreflexion. Tübingen: Niemeyer. Preston, Dennis R. 1996 Whaddayaknow? The modes of folk linguistic awareness. Language Awareness 5: 40–77. Quinn, Naomi 1991 The cultural basis of metaphor. In: James Fernandez (ed.), Beyond metaphor: The theory of tropes in anthropology. Stanford, CA: Stanford University Press. 56–93. Quirk, Randoph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik 1985 A comprehensive grammar of the English language. London, New York: Longman. Rawls, John 1955 Two concepts of rules. Philosophical Review 64/1: 3–32. Reddy, Michael 1979 The conduit metaphor: a case of frame conflict in our language about language. In: Andrew Ortony (ed.), Metaphor and thought, 164–201. Cambridge: Cambridge University Press.

Metapragmatics

135

Schmid, Hans-Jörg 2000 English abstract nouns as conceptual shells. Berlin: de Gruyter. Searle, John R. 1969 Speech acts. Cambridge: Cambridge University Press. Semino, Elena 2006 A corpus-based study of metaphors for speech activity in British English. In: Anatol Stefanowitsch and Stefan T. Gries (eds.), Corpus-based approaches to metaphor and metonymy, 36–62. Berlin: Mouton de Gruyter. Silverstein, Michael 1993 Metapragmatic discourse and metapragmatic function. In: John A. Lucy (ed.), Reflexive language, 33–58. Cambridge: Cambridge University Press. Simon-Vandenbergen, Anne-Marie 1995 Assessing linguistic behaviour. A study of value judgements. In: Louis Goossens, Paul Pauwels, Brygida Rudzka-Ostyn, Anne-Marie Simon-Vandenbergen and Johan Vanparys (eds.), By word of mouth, 71–124. Amsterdam: Benjamins. Smith, Bruce L., Bruce L. Brown, William J. Strong and Alvin C. Rencher 1975 Effects of speech rate on personality perception. Language and Speech 18: 145–152. Sperber, Dan and Deirdre Wilson 1986 Relevance. Communication and cognition. Oxford: Blackwell. Tajfel, Henri 1982 Social identity and intergroup relations. Cambridge: Cambridge University Press. Traugott, Elizabeth 1997 Subjectivication and the development of epistemic meaning: The case of promise and threaten. In: Toril Swan and Olaf J. Westvik (eds.), Modality in Germanic languages: Historical and comparative perspectives, 185–210. Berlin: de Gruyter. Ungeheuer, Gerold 1970 Kommunikative und extrakommunikative Betrachtungsweisen in der Phonetik. In: Proceedings of the Sixth International Congress of Phonetic Sciences, 73–86. (Prag 1967). Prague. Vanparys, Johan 1995 A survey of metalinguistic metaphors. In: Louis Goossens, Paul Pauwels, Brygida Rudzka-Ostyn, Anne-Marie Simon-Vandenbergen and Johan Vanparys (eds.), By word of mouth, 1–34. Amsterdam: Benjamins. Verschueren, Jef 1987 Metapragmatics and universals of linguistic action. In: Jef Verschueren (ed.), Linguistic action. Some empirical-conceptual studie,125–147. Norwood, NJ: Ablex. Verschueren, Jef 1998 Notes on the role of metapragmatic awareness in language use. In: Adam Jaworski, Nikolas Coupland and Dariusz Galasinski (eds.), Metalanguage. Social and ideological perspectives, 53–73. Berlin: Mouton de Gruyter. Watzlawick, Paul, Janet H. Beavin and Don D. Jackson 1968 Pragmatics of human communication: A study of interactional patterns, pathologies, and paradoxes. London: Faber and Faber.

136

Axel Hübler

Welte, Werner 1990 Alltagssprachliche Metakommunikation im Englischen und Deutschen. Frankfurt: Lang. Wilson, Deirdre 2000 Metarepresentation in linguistic communication. In: UCLWPL 11: 127–161. Wooffitt, Robin and Clark Colin 1998 Mobilizing discourse and social identities in knowledge talk. In: Charles Antaki and Sue Widdicombe (eds.), Identities in talk, 107–120. London: Sage.

Part II Theoretical foundations

5.

The rise of pragmatics: a historiographic overview Wataru Koyama

1.

Introduction

This article investigates the theoretical, philosophical, social institutional, and interpersonal genealogies of several key concepts of contemporary linguistic pragmatics, such as presupposition, implicature, speech act, and performativity, as well as perhaps more fundamental ones, viz., context, indexicality, function, referential and non-referential practice, metapragmatics, and reflexivity. More specifically, after briefly dealing with the genealogy of the concepts “context” and “indexicality” in Section 2, the article discusses the genealogies of “presupposition” and “effect” (presupposing and creative functions) in Section 3, “function” (multifunctionality) in Sections 4, 5, and 6, and “performativity”, “metapragmatics”, and “reflexivity” in Section 7. In so doing, this article tries to articulate a historiographic overview of how linguistic pragmatics was established and developed before it became an institutionalized international discipline. It specifically focuses on two critical developments since the late nineteenth century that led to the rise of pragmatics in the 1970s: i.e., the “linguistic turn”, to be followed by the “pragmatic turn”. The former, which took place from the 1880s to the 1930s, gave rise to structural-functionalist linguistics. This historical transformation was closely linked to the neoKantian movement, marking the empirical turn of Kantian philosophy, and the phenomenological movement, focusing on the synchronic “here and now” of communicative events (see Sections 5 and 6). On the other hand, the linguistic turn also led to the emergence of analytic philosophy, with its focus on the propositional, logical-semantic aspects of language. Structural linguistics, interacting with analytic philosophy, took the “formalist turn” in the 1940s, turning away from context and (communicative) function and moving to concentrate on the (morpho)phonological, (morpho)syntactic, and semantic aspects of language. Gradually, however, especially in the 1950s and ’60s, there was increasing awareness of the importance of context, often discussed under the rubrics of indexicality and pragmatics, even in logico-analytic philosophy (see Section 2). This led to the pragmatic turn and the institutional establishment of linguistic pragmatics in the ’70s, linguistics and philosophy moving from the analytically-oriented study of de-contextualizable regularities of language toward the empirical study of the contextualized use of language. Thus, since the 1970s, the theoretical notions articulated or suggested by the older traditions of structural-functional linguistics, such as context, indexical-

140

Wataru Koyama

ity, function, and metalanguage, have been re-discovered and further explored. In addition, the pragmatic turn has allowed the convergence of pragmatics with the empirical social sciences, which are traditionally concerned with actions and interpretations in context. As is described in Section 3, this convergence characterizes the current state of linguistic pragmatics, where the earlier emphasis on regularities in language use is increasingly supplemented with the focus on the pragmatically achieved (trans)formation of group identities and power-relations of language users in specific sociocultural contexts. In such a way, this article tries to present a historiographic account showing that the true home of linguistic pragmatics lies in “the Continental tradition” (cf. Huang 2007: 4–5), which tries to provide a contextual, pragmatic view of language and communication in our life-worlds, not in the “Anglo-American tradition”, which sees pragmatics as a narrowly delimited component supplementary to the ahistorical, universalizing studies of syntax and semantics, abstractable from referential regularities. As this article will demonstrate, pragmatics came out of the Continental tradition in the late nineteenth century and, after the formalist lapse at the mid-twentieth century, the discipline has been moving back from the AngloAmerican to the Continental tradition, further exploring referential and non-referential practices, i.e., what we do, with or without words, in our life-worlds (see Section 5 for details).

2.

Genealogy of the concepts “context” and “indexicality”

Let us begin our historiographic reconstruction with one of the most basic notions of pragmatics, namely, context, which is closely related to the phenomenological notion of “life-world” and which includes not only the tokens of linguistic types but also communicative-event participants and their pragmatic acts (cf. Mey [1993] 2001: 206–235). It would not be too far-fetched to say that “pragmatic” means “contextual(ized)”, as the discipline itself is differentiated by its focus on context, from structural linguistics and semantic logic, which are mainly concerned with de-contextualized regularities abstractable from referential practices. Clearly, the historiography of pragmatics must be written around this theme: how context has been discovered as a theoretical topic, and theoretically dealt with (cf. Nerlich and Clarke 1996). Of course, all empirical studies of language or communication have been at least implicitly concerned with context, so that we might argue that modern theorization on context(uality) started with the emergence of modern empiricism, as suggested by John Locke’s (1632–1704) treatise on “human understanding” and communication (cf. Locke [1689] 1975), which anticipated Thomas Reid’s (1710–96) pragmatic “commonsense” philosophy and thus paved the way for the development of modern semiotics, pragmati(ci)sm, and pragmatics.1

The rise of pragmatics: a historiographic overview

141

Yet, from the linguistic (structural) point of view, which focuses on de-contextualizable linguistic categories, it is, of course, personal and demonstrative pronouns, moods, modalities, tenses, and other deictic categories (deixis) or indexical types that most clearly show contextuality, often called “context-dependency” (cf. Fillmore 1975; Levinson 1983: 54–96); hence, the specifically linguistic “origins” of today’s linguistic pragmatics might be found in empirical investigations and theorizations of the deictic categories. Thus, the “roots” of linguistic pragmatics may be sought, as Nerlich and Clarke (1996) have done, in the structural-functionalist traditions: i.e., the (Austro-)German tradition, which led to Karl Bühler’s (1879–1963) theory of indexical expressions (Zeigwörter), the deictic field (Zeigfeld; vs. Symbolfeld), and Organon model, developed in 1933–34; the French one, culminating in Benveniste’s (1902–76) theory of pronouns, subjectivity, and sémantique, or what would become known as énunciation (Benveniste, 1966); and the British tradition, which eventually led to Lyons’s (1977) “semantics” (in the broad sense, including linguistic pragmatics) and Halliday’s text linguistics, the latter synthesizing the Firthian theory (“London School”) and the Praguean theory of referential (informational) communication and multi-functionalism (see Section 6).2 We shall follow these functionalist movements in Europe, anticipating the worldwide rise of pragmatics in the latter half of the twentieth century, later in this article (see Sections 4–6). Here, we should take a brief look at the development on the other side of the Atlantic. That is, in the United States, the Praguean functionalist theory became combined with the Peircean theory of semiotic pragmati(ci)sm by Jakobson (1896–1982), who articulated a contextually-anchored, that is, speech event-centered account of grammatical categories, including what Bally called “actualisateurs” or Jespersen (1860–1943) named “shifters” (Jespersen 1922), which Jakobson (1957) characterized as “denotational-indexicals”. As is well known, the notion of indexicality came from Peirce (see note 1), who defined it as the kind of relationship between signs (e.g., tokens of linguistic types) and their objects (e.g., contextual referents) that is pragmatically motivated by the empirical principle of contiguity in the context of signification (communication). Note that indexicality is a general “semiotic” notion, characterizing not only deixis, but also pointing gestures, gazes, spatial arrangements (e.g., proximity), and any other modes of communicative practices based on the principle of contiguity in the contexts of communicative events (cf. Peirce [1893–1910] 1955). Clearly, indexicality is a key to the proper theorization of context and contextualized communication, as duly recognized by the micro-sociologist (ethnomethodologist) Garfinkel (see Section 3) and, later, the social pragmatician Mey ([1993] 2001: 52–60). The significance of indexicality for pragmatics had been also acknowledged, in a different disciplinary context, in Bar-Hillel (1954), which would lead to Richard Montague’s (1930–71) attempt at a formal pragmatics. As this indicates, Bar-Hillel was an analytic philosopher generally working in the “logical syntax” framework

142

Wataru Koyama

of Rudolf Carnap (1891–1970), closely associated with the behavioristic pragmatist Charles Morris (1901–79), i.e., two figures involved in the logical-empiricist project of Unified Science (Einheitswissenschaft), or the Unity of Science Movement, led by Otto Neurath (1882–1945; a member of the Vienna Circle), in the early-to-mid-twentieth century (cf. Carnap [1934] 1937; Morris 1938). In particular, Morris, partially following Peirce’s semiotic pragmati(ci)sm (see immediately above), had articulated, in 1937, the famous semiotic triad of syntax, semantics, and pragmatics,3 to be adopted by Carnap. Coming out of such a tradition, Bar-Hillel’s work on indexicality, pragmatics (as “wastebasket” of syntax and semantics), and “implying” (what Grice would later call “implicature”) in the 1940s and 1950s (cf. Bar-Hillel 1946; Chapman 2005: 92) anticipated, in retrospect, not only Montague and Grice, but also the pragmatic turn in analytic philosophy in the 1960s and ’70s, which saw Kripke’s (1972) theory of “rigid designation”, “causal theory of reference” on proper names (cf. Putnam 1975), “speaker’s reference” (“true reference”; Donnellan 1966, 1978; Kripke 1979), and Kaplan’s (1978) theory of “Dthat” on demonstratives. With these developments, by which analytic philosophy became increasingly focused on indexicality and pragmatics, the discipline turned “post-analytic” and came to merge with neo-pragmatism in the 1970s. Also, whereas the Carnapian “logical syntax” tradition was associated with neo-Bloomfieldian structural linguistics and its derivatives like Chomskyan generativism, the newer, post-analytic developments became associated with Generative Semantics and the emerging discipline of linguistic pragmatics in the 1970s (cf. Donnellan 1978; Davis 1991). Thus, indexicality, along with its associated concepts, has been a key element for the constitution of pragmatics and the proper theorization of “context”. Further, the Peircean notion of “indexicality” and Bühler’s ideas of “deictic field” (Zeigfeld) and “deictic center” (origo) became theoretically combined in the linguisticanthropological tradition, derived from not only Boas (1858–1942) and Sapir (1884–1939), but also Peirce, Jakobson, Benveniste, and Kuryłowicz (1895– 1978), and led by a disciple of Jakobson’s: namely, Silverstein (b. 1945). As documented by Pressman’s (1994) contribution to the historiography of pragmatics (also, Caton 1987), Silverstein (1976a) demonstrated, in his famous Jakobsonian work on case-marking and the Noun Phrase (NP) hierarchy, that grammatical categories are universally, systematically anchored onto the deictic center, located at the center of the context of communicative interaction (i.e., Jakobson’s (1957) “speech event”). Further, Silverstein (1976b) made clear that this pragmatic, indexicality-based theory of linguistic structure was part of the general linguistic-anthropological theory of sociocultural communication, covering not only linguistic structure, but also referential and non-referential communicative practices, the latter pertaining to “social indexicality”, or the indexing of the group identities and power-relations of communicative-event participants. Thus, indexicality has been shown to be a key concept in the pragmatic study of contextuality, across the struc-

The rise of pragmatics: a historiographic overview

143

tural, referential, and non-referential domains of language and communication (see Section 6, below, for more on indexicality; cf. Hanks 1990; Duranti and Goodwin 1992; Agha 2007).

3.

Genealogy of the concepts “(contextual) presupposition” and “effect”

Having briefly seen how the notions of “indexicality” and “deictic center” have been made use of to account for the communicative event-centered constitution of context and linguistic structure, we now move to explore historical developments in the theoretical understanding about pragmatic processes that take place in, and (re-)create, context. Here, it appears that the intuitive notions of situational (contextual) “appropriateness” and “effectiveness” (or efficacy) of actions have led to, and still lie behind, more elaborate theorizations of this theme, in terms of, for example, “presupposition” and “perlocution(ary effect)”, respectively. Let us first deal with theoretical notions related to contextual appropriateness. As noted immediately above, the concept of “presupposition”, be it semantic or pragmatic (cf. Karttunen 1974), is a theoretical elaboration of the intuitive notion “appropriateness”. The concept was developed by Frege (1848–1925; see Section 5) and Russell (1872–1970; a leading figure in the logicist tradition of British analytic philosophy), and was later picked up by Peter Strawson (1919–2006; a major figure in the Oxford “Ordinary Language School”). Subsequently, the school’s theoretical framework and concepts were discovered and integrated into linguistics in the late 1960s and ’70s by (post-Chomskyan) Generative Semanticists: viz., George Lakoff, McCawley, Robin Lakoff, Ross, Sadock, et al. (cf. Lakoff 1989; Harris 1993; Koyama 2000). As these were immediate predecessors of linguistic pragmatics (which became established in the 1970s; cf. Mey [1993] 2001: 19–27),4 “presupposition” became a main theme in the new discipline, especially in the 1970s and early 1980s (cf. Gazdar 1979: 89–128; Levinson 1983: 167–225). Generally speaking, since the 1970s, the studies of presupposition have been characterized by the increasing focus on pragmatic, rather than semantic, presupposition and the increasing awareness of the centrality of pragmatic context in matters of presupposition in general. Similarly, the notion of “felicity condition”, apparently concerning contextual appropriateness and derived from the works of Austin (1911–60) and Searle (b. 1932), became a prominent theme in the new discipline. (Austin was the leading figure of the Oxford philosophy of language, and Searle was an American-born student of Austin and Strawson at Oxford in the 1950s). As Austin must have been at least dimly aware, the theme of “felicity condition”, which concerns the contextual conditions that must obtain in order for a (speech) act to occur validly, was actually an old theme in law (see note 7), and, we should also observe, in other disciplines concerned with (judicial or other) rituals, such as sociology and anthropol-

144

Wataru Koyama

ogy, which had the long tradition of investigating the social or cultural (contextual) appropriateness of rituals and other actions (see Section 7). Thus, in the 1970s and early ’80s, “felicity condition”, an old theme recently re-thematized by Austin’s disciple Searle, became a topic not only in the new discipline of pragmatics, but also in the social sciences (see immediately below), thereby facilitating the interaction and partial merger between pragmatics, micro-sociology, and linguistic anthropology. In other words, it appears that this theme may have helped the new discipline to move in the direction of “sociocultural”: that is, toward the microsociology of “conversation analysis” (cf. Levinson 1983: 284–370) and, beyond that, linguistic anthropology, social semiotics, social pragmatics, critical pragmatics, and critical discourse analysis, which have been gaining prominence in pragmatics since the late 1970s and especially ’80s and ’90s (cf. Mey [1993] 2001). Note that, in the 1970s and early ’80s, when “felicity condition” was hotly discussed, the micro-sociologist Goffman (1922–82) wrote an important article on it (Goffman 1983), and a variety of concepts coming out of his œuvre, such as “face” (cf. Brown and Levinson [1978] 1987), “frame” (cf. Tannen 1986), and “footing” (cf. Irvine 1996), began to exercise great influence among pragmaticians, discourse analysts, and linguistic anthropologists (cf. Goffman 1967, 1974, 1981, 1983). On the other hand, in the same period, the more individual-psychological aspects of felicity condition, such as the speaker’s intention and sincerity, articulated by Searle, became critically examined from cross-cultural perspectives by linguistic anthropologists (cf. Rosaldo 1982; Du Bois 1987; Duranti 1993), working in the tradition of “ethnography of speaking/communication” (see Section 6, below, for details; cf. Gumperz and Hymes [1972] 1986; Bauman and Sherzer [1974] 1989). This move anticipated the increasing appreciation of crosscultural differences in pragmatics especially since the 1980s (cf. Mey [1993] 2001: 262–288). Indeed, the social anthropological concern with situational (contextual) appropriateness of actions goes back to a founder of British social anthropology, Malinowski (1884–1942), who advanced the notions of “context of situation” and “context of culture”,5 adopted in the 1930s by his close associate, J. R. Firth (1890–1960), i.e., the leading figure of the London school of functional linguistics. The modern anthropological idea, advanced by Malinowski, Radcliff-Brown (1881–1955), and Boas (1858–1942) in the early twentieth century, that any action, be it linguistic or non-linguistic, should be understood in relation to the situational or cultural context in which the action takes place, was intertwined with the anthropological doctrine of cultural relativism, which urged that actions be seen and understood “from natives’ point of view”, “from within”, or “emically” in the terminology of Pike (1967: 37–72). Beyond the disciplinary matrix of twentieth-century social/cultural anthropology, the idea of the cultural relativity of actions, especially linguistic actions, was also shared by Wittgenstein (1889–1951) and Peter Winch (1926–97), the former sometimes considered a precursor of the Ordinary

The rise of pragmatics: a historiographic overview

145

Language School, including Austin, Grice, and Searle, and subsequently, linguistic pragmatics (cf. Levinson 1983: 227). Importantly, the (situational or cultural) contextualist doctrine, advanced by modern anthropology and later adopted by much of linguistic pragmatics, not only emphasized the role of context in determining the “meanings” of actions, but also tended to take context to be an independent macro-social variable determining the meanings of micro-social actions, the latter considered as dependent variables. In other words, the doctrine failed to pay sufficient attention to how actions create and change the contexts (situations and cultures) in which they take place. This “static” view of culture was perhaps most pronounced in Radcliff-Brown’s Durkheimian functionalism, where the “meanings” of actions were explained by the “(macro-)social functions” they served. In this view, which came close to macrosocial determinism, the “meanings” of micro-social actions (“what we do”, with or without words) were equated with the “functions” they served in the macro-social order (contextual “reality”) presupposed to exist independently of actions. To be sure, in Malinowski’s functionalism, which was less Durkheimian and more Freudian than Radliff-Brown’s (cf. Kuklick 1991), it was not the macro-social order, but the psychological needs of socially contextualized individuals that were taken as presupposed variables, teleologically explaining the “meanings” (functions) of their actions; all the same, the “meanings” of actions were functionally explained by the conformity (“fit”, or appropriate correspondence) of the actions with their presupposed ends, be they psychological or (macro)sociological. The functionalist doctrine was espoused by much of linguistic functionalism and the nascent discipline of linguistic pragmatics in the 1970s. For instance, in these disciplines, actions (utterances, or speech acts) were often explained or theorized by appealing to the notions such as (social or psychosocial) “maxims” and “conditions”, which were, as in Malinowskian functionalism, taken to be independent variables explaining the meanings (implicatures) or efficacy of utterances and other practices. This functionalist way of seeing speech and other social actions had been dominant in the post-World War Two behavioral sciences, which became challenged in the social sciences in the 1960s and ’70s, when the idea that society and culture are contingently (re-)created in and by daily actions and, thus, can be changed by actions, gained momentum, especially in the anti-establishment, youth, peace, and (second-wave) feminist movements. In the social sciences, even before the 1960s, the ethnomethodologist Harold Garfinkel (b. 1917) had begun frontally to attack the functionalist doctrine, attributed to Durkheim (1858–1917), and explore how micro-social, contextualized actions create, and indeed lie at the foundation of, social organization (cf. Heritage 1984). Interacting with the microsociological studies carried out by Goffman (mentioned above, in this Section), ethnomethodology subsequently saw a new development called “conversation analysis”, explored by Sacks, Jefferson, Schegloff, and their followers, and influencing pragmatics (cf. Levinson 1983: 284–370). Meanwhile, since the 1970s,

146

Wataru Koyama

second-wave feminism, critical of the functionalist view of society, has given rise to women’s studies, gender studies, and later, queer studies, a series of developments in which feminist linguistics/pragmatics have been involved (cf. Lakoff 1973; Cameron and Kulick 2003), and in which the term “performativity”, taken from Austin, has played a crucial role (cf. Butler 1990). As this indicates, the speech act theories of Oxonian philosophy and early pragmatics were not exclusively concerned with the question of contextual appropriateness. Rather, they investigated “conditions”, “maxims”, and other presupposed contextual variables, at least partially so as to explain the efficacy, or the creative (performative) function, of actions, the intuitive appreciation of which was, indeed, the point of departure for them. And, it was this insight of “performativity” that attracted theoretical interests of many literary critics, historians, social scientists, and feminists, to pragmatics (see Section 7 for more on performativity). Thus, interacting with them, pragmaticians moved to construct pragmatic theories focusing on the creative, rather than presupposing, function of language use, especially as related to powerrelations and group identities of language users. As a result, not only feminist pragmatics, but also critical linguistics/pragmatics, associated with the Frankfurt School, the New Left, and neo-Marxism, came into being in the late 1970s (cf. Mey 1979, 1985, [1993] 2001), in critical opposition to the functionalist view of society. Eventually, the 1980s and ’90s saw the rise of critical discourse analysis in Western and Central Europe. Even sociolinguistics, in the wake of criticisms against the functionalist doctrine (cf. Williams 1992), has begun to move in the direction of social constructivism, investigating how social realities and systems are dynamically created by discursive interactions (cf. Coupland, Sarangi, and Candlin 2001).

4.

Genealogy of the idea “(communicative) functions” (multi-functionality)

Thus, appropriateness and effectiveness, that is, the presupposing and creative functions of interactional practices, constitute one dimension of pragmatic processes in context, around which the history of pragmatics can be described. Crosscutting this is another dimension of pragmatic processes in context: viz., the dimension of what one might call “the functional (indexical) orientations of communication”, such as the referential (descriptive), expressive (emotional), and “appeal” functions (cf. Bühler [1934] 1990). Jakobson’s (1960) “six-functions” model, one of the most solid models of the multi-functionality of communication, identified the emotive, conative, referential, metalingual (metalinguistic), poetic, and phatic functions, respectively characterized by the six “factors” constituting the event of communicative practice, i.e., addresser, addressee, referent, code, message, and channel (contact; see Section 6 for more details). Indeed, the term “function” has been usually used in this sense in linguistic functionalism and prag-

The rise of pragmatics: a historiographic overview

147

matics, and the history of pragmatics has been often written by tracing the history of studies dealing with these functions in linguistics, philosophy, psychology, law, and the social sciences (cf. Nerlich and Clarke 1996). Although Austin’s theory of speech act seems to have arisen more or less independently of these functionalist studies, which had blossomed from the 1880s to the 1930s (see Section 5), it is clear that they anticipated the basic insight of Wittgenstein and the Ordinary Language School (including Austin, Warnock, Urmson, Grice, Strawson), to be absorbed by Generative Semantics and linguistic pragmatics from the late 1960s onward, that language includes many functions other than the referential function or propositional meaning, which appears to have been the exclusive concern of logic, structural linguistics, and generative syntax. That is, as Nerlich and Clarke (1996) have shown, the social, psychological, linguistic, and philosophical studies of the non-referential functions (multi-functionality) of language and communication, were vigorously pursued especially from the 1880s to the 1930s. This was the period of the “linguistic turn” (cf. Rorty 1967; Apel 1994), before the rise of formalist structural linguistics, which concentrated on the formal and semantic, but not pragmatic or indexical aspects of language, in the 1940s. To be sure, the functionalist traditions were kept alive in the works of Jakobson, Benveniste, Kuryłowicz, Pragueans, and (neo-)Firthians, to name but a few important figures and “schools”, but the dominance of formalists (such as neo-Bloomfieldians and Chomskyans) in structural and generative linguistics, associated with the philosophical project of logical syntax, in the United States and beyond, continued until the 1970s (see Section 2). As a result, the earlier functionalist theories and their achievements became largely forgotten, so that, when Generative Semanticists re-discovered the importance of context (pragmatics) and started to rebel against the orthodox Chomskyan doctrine of autonomous syntax in the late 1960s, they made appeal to the philosophical theories of Austin, Grice, Strawson, and Searle, but not to the older functionalist traditions (cf. Harris 1993). Linguistic pragmatics, which arose in the 1970s in the wake of Generative Semantics, was no less strongly influenced by the philosophical theories, but it also actively interacted with the offspring of the older functionalist traditions, i.e., a variety of functionalist approaches to grammar, text, and discourse, which saw a renaissance in the 1970s (cf. Bolinger 1968; Comrie [1981] 1989; Halliday and Hasan 1976; Leech 1983: 46–78). Thus, the importance of the notions advanced by the older traditions, especially “deictic field” (Zeigfeld), “indexicality”, and “metapragmatics” (the latter two derived in part from Jakobson’s (1957, 1960) “code-and-message” model and “metalingual function”, respectively; cf. Silverstein 1976b), has become more fully appreciated, and a number of attempts have been made to explore and elaborate on the implications of these important notions, especially in the recent literature of pragmatics (see Sections 2 and 7; cf. Mey [1993] 2001; Verschueren 1999; Bublitz and Hübler 2007).

148 5.

Wataru Koyama

The linguistic turn and the philosophy of language: functionalism, phenomenology, and neo-Kantianism

The older traditions of functionalism, as noted above, blossomed from the 1880s to the 1930s, the period when philosophy took the linguistic turn. Here, focusing on the disciplinary development of philosophy, which was closely linked to the rise of semantics and pragmatics, we note that the late nineteenth century saw the “returnto-Kant” movement, as can be seen in the Kant-inspired pragmatic philosophy of Peirce in the United States and, especially, the epistemology of neo-Kantians in Germany, whose opponents included the logico-mathematician Frege (cf. Coffa 1991). This suggests that logico-analytic philosophy, other kinds of the philosophy of language, neo-Kantianism, and phenomenology were, directly or indirectly, all involved in the linguistic turn, and interacted with each other, as well as with the older traditions of functionalism (cf. Richardson 1998; Friedman 2000). Let us briefly explore historical connections among them, taking Frege as our starting point. Needless to say, Frege (1848–1925) was one of the founders of modern predicate logic and semantics, and his distinction between sense (Sinn) and reference (Bedeutung) anticipated the later division of semantics and pragmatics, articulated by Morris and Bar-Hillel (see Section 2, above). No less important, Frege’s differentiation between proposition and force (Kraft), anticipated by Condillac among others,6 was a precursor of the distinctions between constatives and performatives, and between locutionary, illocutionary, and perlocutionary forces, advanced by Austin, who was a translator of Frege’s Die Grundlagen der Arithmetik (‘The Foundations of Arithmetic’). As is well known, these distinctions underlay the formula “F(p)” (where “p” and “F” stand for proposition and illocutionary force, respectively), articulated by Austin’s student, Searle (cf. Nerlich and Clarke 1996: 22, 76, 202–204). Interestingly, the later Searle (1983) advanced a phenomenologically-oriented theory of speech, as suggested by the title “Intentionality” (a notion derived from Husserl (1859–1938), Brentano (1838–1917), and, ultimately, the Scholastic concept of intentio); and Searle’s master, Austin, intended his philosophy of language to be a kind of “linguistic phenomenology”.7 Thus, there appears a link between Frege’s logic(ism) and Husserlian phenomenology, although Husserl, who started his career as a philosopher of mathematics and logic, moved away from logic towards transcendental phenomenology and, eventually, the phenomenology of the life-world (Lebenswelt). Subsequently, phenomenology inspired various kinds of the philosophy of language: e.g., not only Austin’s and Searle’s, but also the French dialectician Merleau-Ponty’s (1960), who exchanged his views with Oxonians, viz., Ryle, Austin, and Urmson, at a colloquium in 1958. Further, phenomenology also influenced the critical social theory of Habermas (b. 1929; cf. Mey [1993] 2001: 315), who has adopted the notion of Lebenswelt, and, earlier, the sociology of Alfred Schutz (1899–1959), who carried

The rise of pragmatics: a historiographic overview

149

it from Vienna to New York. In North America, phenomenological sociology was then carried out by Berger and Luckmann (cf. Mey [1993] 2001: 115) and by ethnomethodologists, including Garfinkel (cf. Giddens [1976] 1993; Bourdieu 1980). As linguistic pragmatics has been variously influenced by ethnomethodology, conversation analysis, and Critical Theory, it can be said that the phenomenological orientation has entered linguistic pragmatics not only from the Ordinary Language School (see immediately above), but also from phenomenological sociology. Hence, the phenomenological strand is fairly strong in linguistic pragmatics. After all, the discipline was established in reaction against formalist structuralism in linguistics, which was, unlike the phenomenologically-oriented functional theories of Bühler, Jakobson, and the Prague School (cf. Holenstein 1977), non- or even antiphenomenological. Thus, although analytic logic and phenomenology may appear polar opposites, there is a linkage between them, as indicated by the careers of Husserl and Austin, as well as by the history of linguistic pragmatics, which has been, since the late 1970s, moving from de-contextualized logic and maxims towards contextualized events, i.e., social interactions in historical and cultural life-worlds. Further, this strongly suggests a certain affinity between linguistic pragmatics and the philosophical movement of neo-Kantianism. Here, observe that, just like pragmatics, neo-Kantianism seems located between logic and phenomenology, as witnessed by the fact that Peirce, who belonged to the Return-to-Kant Movement, invented a semiotically-based predicate logic, distinct from Frege’s arithmetic one, and called his pragmati(ci)sm “phenomenology” and “phaneroscopy”.8 Also, one may note that a neo-Kantian philosopher par excellence, namely Cassirer (1874–1945), who belonged to the Marburg School of neo-Kantianism and emigrated to New York, contributed an important paper (Cassirer 1945) on the status of structural linguistics to the New York-based journal Word, run by Martinet (1908–99), Jakobson, and other structural-functionalists, who were opposed to native-born neo-Bloomfieldian formalists in the United States. This points to the linkage between neo-Kantianism and functionalist linguistics, and the latter was, as we have seen, intertwined with phenomenology, especially in the works of Jakobson and Bühler.9 As a matter of fact, Cassirer’s career, characterized by the application of (neo)Kantian philosophy, in particular epistemology, onto empirical investigations into myths, languages, and the history of ideas, was part of the historical transformation, or empirical turn, of Kantian philosophy, which moved from “pure” and “practical” (praktisch in Kant’s terminology) reason, i.e., critical-philosophical, transcendental principles of epistemology and praxiology, towards more empirical, historical, anthropological, pragmatic (pragmatisch) studies of knowledge and actions, at the turn of the twentieth century. Crucially, we should see that, in general, when knowledge and actions are studied empirically, it becomes clear that they are mediated by language and communication. Here, recall from Section 2, that Locke

150

Wataru Koyama

(1632–1704), one of the founders of modern empiricism, felt compelled to write a treatise on human understanding, in particular, language, communication, and semiotics; and Thomas Reid (1710–96), deeply influenced by Locke (as well as by Hume, a Scotsman like Reid), was a precursor of modern pragmatics and, perhaps, phenomenology (cf. Mulligan 1987: 33–34). Subsequently, Peirce (1839–1914), influenced by Locke, Reid, and Kant (1724–1804), developed a semiotic philosophy of communication, which he called “pragmati(ci)sm”, based on Kant’s term pragmatisch (see immediately above). The term was adopted by the behavioristic pragmatist Morris, who coined the term “pragmatics” in his famous trichotomy of communication (see Section 2). We still operate with Morris’s terminology, i.e., syntax, semantics, and pragmatics; and this fact suggests the intermingling of not only the terminological but also the conceptual genealogies of today’s linguistic pragmatics and the neo-Kantian movement in philosophy. Thus, the empirical turn of (neo)Kantianism seems to lie at the roots of linguistic pragmatics. Here, notice that the heyday of neo-Kantianism in philosophy coincided with the linguistic turn, i.e., the period in which the older traditions of functionalism blossomed, from the 1880s to the 1930s (cf. Nerlich and Clarke 1996). This cannot be by chance. By the linguistic turn, “judgment”, a concept integral to Kantian philosophy, became replaced with “proposition” in post-Fregean logic, and, more generally, with “language” in twentieth-century philosophy, including logico-analytic philosophy and the Ordinary Language School;10 at the same time, (neo)Kantianism turned empirical, giving rise to pragmati(ci)sm and the empirical science of language, i.e., linguistic functionalism, studying the synchronic, living state of language in action, context, and communication. This movement, replacing the nineteenth-century historical “comparative linguistics” (over)systematized by Neo-Grammarians, was what was called (functionalist) structuralism (cf. Jakobson 1973), which would become overshadowed by formalist structuralism, the latter turning away from context and pragmatic functions to formal structure and semantics in the 1940s and ’50s. We shall soon come back to “the synchronic turn” in the studies of language, corresponding to the linguistic turn in philosophy: i.e., the rise of structuralism as part of linguistic functionalism, between the 1880s and the 1930s. But, first, we must observe that the empirical turn of (neo)Kantian philosophy meant the emergence of the empirical sciences of not just language, but also the mind, society, and culture: i.e., modern twentieth-century psychology, sociology, and anthropology. Note that William James (1842–1910), Wundt (1832–1920), Max Weber (1864– 1920), Simmel (1858–1918), Durkheim, Mauss, Boas, Malinowski, and RadcliffBrown, who are usually counted as founders of these disciplines, worked in this period, from the 1880s to the 1930s, and some of them were explicitly associated with pragmatism and philosophical neo-Kantianism, while others directly or indirectly communicated with them. Further, whether or not they were “pragmatists” or “neo-Kantians” in the narrow sense, their empirical sciences and theories gen-

The rise of pragmatics: a historiographic overview

151

erally shared the epistemological framework of neo-Kantianism, such as the methodologically-based distinction between nomothetic sciences, investigating laws and other regularities, and idiographic (or historical) sciences, exploring unique and singular, sociohistorically contextualized phenomena (cf. Stocking 1996). Together with the related opposition between collectives and individuals, this distinction underlay, among others, the Saussurean notions of langue and parole, the proto-Weberian Methodenstreit concerning the status of economics as a socio-historical (idiographic) or ahistorical (nomothetic, classical Newtonian-Laplacean) science, and the Durkheimian understanding of “sociology” in opposition to individualist psychology. Thus, (neo-)Kantianism was the philosophical “foundation” for the rise of the human sciences (cf. Foucault 1966), including linguistics and, later, pragmatics. The linguistic turn in philosophy and the (functionalist) structuralist turn in linguistics were part of the (broadly conceived) neo-Kantian movement, which also included the rise of phenomenology. That is, in the studies of language, the linguistic turn at the beginning of the twentieth century consisted of the move towards the “synchronic”, living state of language in action (i.e., pragmatics); and the rise of structuralism should be understood as part of the functionalist, phenomenological movement, oriented towards actions and other happenings taking place hic et nunc, or the dectic center (origo) of such pragmatic occurrences (see Section 2). As Jakobson was never tired of pointing out (cf. Jakobson 1973; Holenstein 1977), structuralism, i.e., functionalist (vs. formalist) structuralism, was not opposed to phenomenology; they constituted a unified whole (holistic Gestalt), as it were. The philosophy underlying structuralism was the neo-Kantian one, represented by the idea of “the primacy of (human) (ap)perception”, advanced against nineteenthcentury positivism and mechanical empiricism, which had culminated in the NeoGrammarians’ comparative (historical) linguistics in the late nineteenth century, and which had earlier served to suppress the “proto-linguistic turn”, achieved by Bernhardi and Humboldt in Germany, Condillac and the Idéologues in France, and Reid and the followers of his commonsense philosophy such as Dugald Stewart (1753–1828) at the turn of the eighteenth to the nineteenth century (see notes 1 and 2, above; cf. Nerlich and Clarke 1996). The primacy of human perception, or subjective apperception, clearly underlies the structuralist notions of “phoneme” and “sound pattern”, advanced in contradistinction to “mechanical” physical sounds, studied by the empiricist, positivist science of nineteenth-century phonetics. Further, phonemes were considered part of not just subjective (individual), but also intersubjective, sociocultural “reality” (see Section 6). The insiders of a community sharing (by degrees) a linguistic structure, i.e., members of a linguistic community, (ap)perceive the (more or less) “same” phonemes, but not phones; they have phonemes as part of their “psychological reality” (cf. Sapir 1933), that is, intersubjective phenomenological reality, a pattern of (ap)perception not shared by the outsiders of the (Durkheimian) socio-cultural, linguistic-structural commu-

152

Wataru Koyama

nity. Thus understood, it becomes less mysterious why the phenomenologist Merleau-Ponty (1945, 1960) appreciated both Gestalt psychology and Saussure’s linguistics, celebrating the latter’s “discovery” of synchrony and parole, and why Saussure was closely associated with Bréal (1832–1915), a founder (precursor) of modern twentieth-century semantics and pragmatics, in Paris (cf. Aarsleff 1982; Nerlich 1992). For all of them, including even Saussure, language was, au fond, an intersubjective, phenomenological reality, pragmatically unfolding in front of our eyes, that is, emerging at the synchronic hic et nunc of communication. Thus, the historical and theoretical roots of modern structural linguistics and linguistic pragmatics converge in the rise of structuralism, i.e., functional structuralism, in the period from the 1880s to the 1930s, that is, the days of the older traditions of linguistic functionalism, to be eclipsed when structural linguistics became dominated by neo-Bloomfieldian, Chomskyan, and other kinds of dogmatic formalism after the achronic (vs. synchronic) turn in the 1940s.

6.

Intersubjectivity and interaction: functionalism, Ordinary Language philosophy, and “the social turn” in pragmatics

“Intersubjectivity” involves not only (1) the speaker (addresser) and (2) the hearer (addressee), but also (3) the channel (contact) between them, (4) the sign (message) that is sent from the former to the latter, (5) the (intersubjectively) shared code(s) for interpretations of the sign, and (6) the (intersubjectively) understood referent(s) of the sign, located in the micro- or macro-context of the interaction between the speech- or communicative-event participants, such as the speaker and addressee. This is, of course, the contextual, pragmatic matrix of communication, including language, which was articulated in the Praguean functionalist theory of Jakobson (1960), mentioned in Section 4. As part of the code, linguistic structure is embedded in this functional matrix of communication. In other words, structure emerges through communicative practices; there is no structure (langue) without pragmatics (parole). According to this Jakobsonian theory, as noted earlier, there are six basic functions of communication, corresponding to the six aspects of the contextual matrix: viz., (1) emotive, (2) conative, (3) phatic, (4) poetic, (5) metalingual (metalinguistic), and (6) referential functions.11 Although linguistic structure, as part of the code, is primarily used to interpret the sign (token) to identify (i.e., intersubjectively construct) its referent(s) and thus primarily involved in the metalinguistic, poetic, and referential functions (respectively corresponding to the code, sign, and referents), it inevitably interacts with the other functions as well, because it exists only as part of the code pragmatically used for intersubjective interpretations of the sign indexically exchanged in the context of communication. Thus, language, including linguistic structure, i.e., (morpho)phonology, (morpho)syntax, and semantics, should be understood, theorized, and analyzed by tak-

The rise of pragmatics: a historiographic overview

153

ing into proper consideration the six functions, as well as the six aspects, of communicative practices (cf. Silverstein 1976a, 1976b, 1993). But, of course, the contextual, functional matrix of communication, articulated in the six-functions model, is not limited to, or centered around, linguistic structure. For example, the code includes not just linguistic structure, but also “maxims”, “interpretive rules of thumb”, “passing theories”, metapragmatic frames, and models of interpretation vis-à-vis (not necessarily linguistic) signs, including gestures, facial expressions, and other “paralinguistic” phenomena (cf. Bolinger 1968), all of which occur in specific, sociohistorical contexts of communicative practices (see Section 7; cf. Mey [1993] 2001: 206–235). And such contextualized practices index12 not just referents (in the referential function) and the semantic and pragmatic “codes” being used (in the metalinguistic function). They also index the social identities of, and power-relations among, communicative-event participants (e.g., in the emotive and conative functions). That is, contextualized practices index not just referents, but also social identities and power-relations, which can be “read” in the emergent “texts” of sociocultural practices of communication. Such texts are not to be understood as static “objects”, but as the dynamically emergent, socioculturally interpretable organizations of referential and non-referential (interactional) practices (mainly corresponding to the poetic, phatic, and “metapragmatic” functions), that is, organizations studied in conversation analysis (cf. adjacency pairs; Levinson 1983: 284–370), discourse analysis (cf. “frames” of interaction; Goffman 1974), linguistic anthropology (cf. Silverstein and Urban 1996; Blommaert 2005), and, of course, pragmatics (cf. Mey [1993] 2001). Thus, in a sense, the Praguean/Jakobsonian functional theory provided a general framework in which today’s empirical approaches to language, listed immediately above, operate (also see Section 2). As this indicates, the rise of structuralism in the late nineteenth and early twentieth centuries served to prepare the proper theorization of the functional, contextual matrix of language, which would emerge in the 1970s after the formalist lapse from the 1940s to ’60s. This lapse, however, made obscure the pragmatic, functionalist foundation of structural linguistics. Indeed, by the late 1960s, when Generative Semanticists, who were immediate predecessors of linguistic pragmaticians, started to rebel against formalist structuralism (Chomskyan generativism) and delve into the realm of pragmatics, North American linguistics had become more closely linked to the philosophy of language, especially logico-analytic philosophy, than to the early twentieth-century functionalist traditions of structuralism, so that, in trying to find theoretical tools to deal with pragmatic phenomena, Generative Semanticists discovered and adopted the theories of speech act, implicature, and presupposition, articulated by Ordinary Language philosophers like Austin, Grice, and Strawson (see Section 3, above; cf. Levinson 1983). The publication of Austin’s (1962; second edition, 1975) William James Lectures, originally delivered in 1955 at Harvard, which is located in the same city as MIT (Cambridge, Mass.), i.e., the home of generativism, made the

154

Wataru Koyama

theories of the Ordinary Language philosophers more visible to generativists interested in pragmatic matters (also cf. Grice 1967; Chapman 2005: 100). And Searle (1969) provided a desideratum, i.e., the desired systematization of speech act theory, making it amenable to the linguists’ formal, syntactic treatment of pragmatic phenomena.13 As Generative Semantics yielded to linguistic pragmatics in the 1970s, the early years of the new discipline were marked by the preoccupation with the issues articulated by the Ordinary Language philosophers, and many of the culture-specific assumptions of their philosophy, i.e., post-Lockean modern British philosophy, concerning human communication, seem to have been directly borrowed into linguistic pragmatics. Here, recall that Locke, a founder of modern British empiricism, discussed, like Saussure, the “arbitrary” nature of human language (communication), but, unlike the latter, Locke meant by the term the volitional, intentional, “willful” character of the individual’s use of language, whereas Saussure meant the conventional (non-volitional), extensionally unmotivated (intensional) yet socially existing (i.e., Durkheimian) association of the signifiant and the signifié. As this indicates, British and Anglo-American philosophy, especially the empiricist tradition thereof (see Section 2), has tended to theorize “communication” (more generally, human practice) in terms of volitional, intentional actions or, in sociological terms of Max Weber, the “purposive” function (cf. Silverstein 1976b; Apel 1994: 63). Ordinary Language philosophy, an empiricist philosophy of language and communication concerned with “what we do with words” and “what is meant”, is no exception. It sees language and communication as purposive activities, as explicitly articulated by G. E. M. Anscombe, G. H. von Wright, and Grice, who tried to explain the social, interactional fact of communication in terms of what we intend to do or what we intend to mean, that is, individual’s conscious intention (cf. Chapman 2005). Notwithstanding, as Weber and other social scientists have shown, actions have consequences (perlocutionary effects) which are unintended by agents (subjects), and which may be more important, in terms of historical consequences (i.e., pragmatic effects), than actors’ mere intentions (cf. Silverstein 1976b). Further, any practice at least partially consists of unintentional, non-conscious elements; obviously, we are not conscious of everything we do. For example, we, as ordinary language users, are usually more aware of surface “segments” (continuous, segmentable units), such as phonemes, words, and expressions, than of allophones and discontinuous syntactic units (cf. Lucy 1992); yet, the latter two are often part of dialectal/sociolectal varieties and thus index their users’ social identities, pertaining to class, gender, status, etc., and power-relations among them, as can be readily seen in the sociolinguistic literature, Labovian or otherwise. Thus, there are nonpurposive pragmatic functions in discourse, which are too important to be left out in any theory of pragmatics. In addition, cross-cultural comparisons suggest that many societies place much less value on speakers’ intentions in the evaluation of (sociocultural) “meanings”

The rise of pragmatics: a historiographic overview

155

of their linguistic practices than the culture of Ordinary Language philosophers does, a point poignantly raised and substantiated by linguistic anthropologists, who have been critical of the intentionalist and individualist “bias” of the modern “Standard Average European” speech act theories ever since the 1970s and early ’80s.14 Subsequently, partly as a result of this cross-cultural (anthropological) critique, linguistic pragmatics has come to appreciate and focus on cross-cultural differences in “speech acts” and, more generally, linguistic practices and their metapragmatic understandings, the latter now discussed under the rubric of “language ideology”.15 No less important, the individualist assumption of the speech act theories was undermined by empirical studies of discourse, which showed that the meaning(s) of an action (speech act) cannot be determined out of context, i.e., without considering a series of interactions which contains the action (cf. Hancher 1979; Levinson 1983). This means that, to the extent that the “meaning(s)” of an action is determinable, it is determined by the contextualized interaction of speech-event participants, not solely by the individual speaker’s intention. Contributions of conversation analysis (cf. Levinson 1983: 284–370) and discourse analysis (cf. Mey [1993] 2001: 191), focusing their attention on interaction rather than on solitary individual actors’ minds, have been important in this “interactional turn” (social turn) of pragmatics, which has characterized the discipline since the 1980s or so (cf. Thomas 1996). Pragmatics today is much more interaction- or discourse-centric than in the 1970s, when the new discipline was more intention- and individualcentric under the influence of Ordinary Language philosophy and modern British empiricism.

7.

Genealogy of “performative utterances” and other multi-functional co-incident happenings in the speech event

Let us finally move to the question of performativity, another central theme in linguistic pragmatics. Clearly, this is a complex notion, only a part of which is accounted for by the creative function (“efficacy”), discussed in Section 3. That is, in addition to the creative function, performativity involves “self-referentiality” (cf. Recanati 1979),16 which is itself a composite notion, made up of “reflexitivity” and “(transparent) referentiality”. The latter means the (relatively) transparent relationship between the referential function (what is said) and the non-referential, interactional function (what is done), as observed when we use a referentially explicit metapragmatic utterance, such as “When I said that, I meant it to be a compliment” (cf. Caffi 1984; Mey [1993] 2001: 173–205). On the other hand, “reflexivity” as such is just the mutual indexing of some semiotic elements, linguistic or non-linguistic, co-occurring in the event of communication. For example, a pointing gesture and a token of the deictic expression “this book”, occurring together, mutually

156

Wataru Koyama

index each other and function reflexively, as the gesture and the linguistic utterance coincide in the same context (deictic, indexical field; see Section 2). Mutual, reflexive indexing also obtains when a defiant smile, or a serious demeanor, occurs with an utterance “I’m sorry”, or when a judge, properly dressed or completely undressed, utters “I declare you guilty”. Clearly, these are examples of what Gregory Bateson ([1955] 1972) called “metacommunication” (cf. Ruesch and Bateson 1951; Caffi 1984) and what Gumperz (1982) called “contextualization cue” (cf. Auer and Luzio 1992; Duranti and Goodwin 1992), notions coming out of the anthropological, psychiatric, and ethological traditions, independently of Austin’s discussion of performatives. Austin’s (1962) notion of performativity, following the discussions by the phenomenologist Adolf Reinach (see note 7) and others in the study of law, did not clearly separate these two dimensions of “performativity” or explore each dimension in depth. Originally, linguistic pragmatics was under the influence of Austin’s way of formulating the problem, but it has gradually moved to see that the problem of performativity should be dealt with by separating out different principles underlying the problem, such as the principle of metapragmatics, which is becoming a major topic in our discipline (cf. Ducrot 1984; Verschueren 1999; Mey [1993] 2001). Like the cognate notion “metacommunication”, the concept of “metapragmatics” came out of the anthropological literature, in particular Silverstein (1976b), who coined the term to designate the pragmatic (vs. semantic) aspect of the “metalingual” function,17 articulated by his teacher, Jakobson (1957, 1960), who, in turn, derived it from the discussion of “metalanguage” (vs. “object language”) by the mathematician David Hilbert (1862–1943), Kurt Gödel (1906–78), Tarski and other logicians, e.g., Russell, the early Wittgenstein, and Carnap (cf. Apel 1994). The metapragmatic function obtains when a sign, be it an utterance, a gesture, or whatever, indexes another sign (speech, non-linguistic action, etc.) in context. When the two signs co-occur (coincide) in the same context, they are said to show “reflexively calibrated” metapragmatics (cf. Lucy 1993; Mey [1993] 2001). According to Silverstein (1993), the reflexive metapragmatic function serves to create texts, be they referential or non-referential (see Section 6 on “text”). For instance, the reflexive co-indexing of referential signs, such as (thematic) anaphor and its (rhematic) antecedent, in the same unit of discourse creates referential textuality, as studied in Praguean functional structuralism, text linguistics, and discourse analysis (cf. Halliday and Hasan 1976). Similarly, the reflexive co-indexing of non-referential signs, such as the two units of an adjacency pair (e.g., request-compliance or -rejection; cf. Levinson 1983: 284–370), in the same phase of discourse creates interactional, non-referential textuality. Clearly, the idea of “textualization” came from, among other places, Bühler’s notion of Sprachwerk ‘language work’ and, in particular, Jakobson’s “poetic function” (see Section 6), i.e., the communicative function that is focused on messages (i.e., signs) occurring in the context of interaction and thus gives textual organiz-

The rise of pragmatics: a historiographic overview

157

ation to them, as can be most clearly seen in parallelisms and repetitions abundantly observed in poetry and other rhetorically elaborate verbal arts (e.g., folklore and myths; cf. Bauman 2004). Importantly, basic units of poetic text are often nonreferential segments, such as meters and rhymes, in contradistinction to referential units such as rhemes, themes, and syntactic segments like phrases, clauses, sentences. Thus, poetic texts are, at least partially, non-referential texts, i.e., texts of non-referential interaction. And hence, poetic texts tend to become endowed with non-referential “meanings”, such as emotional (emotive) and persuasive (conative) “forces” (functions). The same characteristics are found in rituals, which also typically show the abundance of parallelisms and repetitions in the linguistic and non-linguistic, interactional modes (cf. Caton 1990). To be sure, in “ordinary conversations” too, we also find such interactional poetic parallelisms and repetitions especially when they contain what Goffman (1967) called “interaction rituals” (cf. greetings and other adjacency pairs, mentioned above). Yet, the characteristic is most salient in rituals, in which, we should recall, Austin (1962) discovered (explicit) performatives: e.g., christening rituals, legal rituals, etc. Performative utterances are the part of rituals characterized by the use of little ritual one-liners, i.e., performative sentences, which are, thus, socially recognized formulaic types, recursively co-occurring with the non-linguistic elements of ritual settings (cf. Volosˇinov [1929] 1973; Silverstein 1993). Thus, performatives show “reflexivity”, i.e., the reflexive metapragmatic function, which is observed more broadly in rituals, poetry, metacommunication, contextualization cues, etc. On the other hand, performatives involve another principle, that is, “referentiality”, or relatively pronounced degrees of transparency between the referential function (what is said) and the non-referential, interactional function (what is done), as noted above. Clearly, it is this principle that has been the principal focus of pragmatic discussions on performativity and speech act, most clearly observable in Searle’s theory of “indirect speech act” and Grice’s theory of “implicature”. These two notions were formulated to account for non-explicit or implicit relationships between the referential and non-referential functions of utterances, and they have been main themes of linguistic pragmatics since the 1970s. Much of linguistic pragmatics still operates with these concepts. For example, Brown and Levinson’s ([1978] 1987) politeness theory and, inter alia, Sperber and Wilson’s Relevance Theory seem squarely located in this tradition. Clearly, the theme of referential explicitness vs. implicitness goes back to the old tradition of rhetoric, where “literal” expressions and tropes were understood in such terms (cf. Ricœur [1975] 1977). When Austin “discovered” non-referential functions in explicit performatives, he did not depart from this age-old tradition, as he focused on non-referential functions only insofar as they appear together with the referential function, as in performatives, where the two kinds of function happen to coincide. Just as in ancient rhetoric, Austin and his followers moved

158

Wataru Koyama

from the referential function, as studied in logic and formalist structuralism, to non-referential functions. That is, Austin’s was a reference-centric approach to pragmatic functions, the severe limit of which is indicated by the comparison with multi-functionalist approaches to pragmatics found in, e.g., Jakobson’s theory (see Section 6), its developments in linguistic anthropology (see Section 2 and immediately above), and the theory of “pragmatic act”, advanced in social pragmatics (cf. Mey [1993] 2001: 206–235; also see Section 3, above). The rise of social pragmatics and linguistic anthropology in linguistic pragmatics since the 1990s seems to suggest that the discipline is not only recovering but also further elaborating the multi-functionalist, social interactionist insights of the older traditions of linguistic functionalism, which blossomed from the 1880s to the 1930s (see Section 2).

Notes 1. More specifically, Reid’s philosophy influenced Benjamin Smart (1786–1872), Adolphe Garnier (1801–64), and Henry Sidgwick (1838–1900), and developed into Charles Peirce’s (1839–1914) semiotic pragmati(ci)sm, Charles Morris’s (1901–79) behaviorist pragmatism, Sidgwick’s student George Stout’s (1860–1944) empiricist and behavioristic philosophy, the latter’s student G. E. Moore’s (1873–1958) commonsense empiricism, and Grice’s (1913–88) theory of “non-natural meaning” and “implicature” in the nineteenth and twentieth centuries (cf. Nerlich and Clarke 1996: 20–21, 105–106; Chapman 2005: 22, 25, 71–72). 2. The German tradition ran from Bernhardi (1769–1820) and Wilhelm von Humboldt (1767–1835) to Philipp Wegener (1848–1916), the later Karl Brugmann (1849–1919), Anton Marty (1847–1914), and Bühler, whereas the French one ran from the Idéologues to Bréal, (1832–1915), Meillet (1866–1936), Mauss (1872–1950), Bally (1865–1947), Guillaume (1883–1960), and Benveniste, to mention only a few of the major figures. 3. This was preceded by the medieval trivium (artes sermocinales), i.e., grammar, logic, and rhetoric. Note that Morris’s predecessor Peirce was thoroughly familiar with Scholastic philosophy. 4. The Journal of Pragmatics was launched in 1977 (cf. Haberland and Mey 1977). Subsequently, major textbooks of pragmatics became published in the early 1980s (cf. Leech 1983; Levinson 1983), the International Pragmatics Association (IPrA) was founded in 1986, and Pragmatics was started in 1991, while the “second wave” of major, and more Continental-European (vs. Anglo-American), textbooks appeared in the 1990s (cf. Mey [1993] 2001; Verschueren 1999). 5. These ideas had been suggested and, in part, advanced by Philipp Wegener (1848–1916). 6. For instance, before Condillac (1714–80), Grammaire générale et raisonnée had made the distinction between dictum and modus, roughly corresponding to proposition and force, respectively. 7. Also, note that Adolf Reinach (1883–1917), whose theory of “social acts” was one of the precursors of Austin’s theory of “speech acts” and “performativity”, was a phenomenologist. Reinach was a student of Husserl’s and, with Johannes Daubert (1877–1947), belonged to the “Munich School” of phenomenology. – The Scottish commonsense philos-

The rise of pragmatics: a historiographic overview

8.

9.

10.

11. 12. 13. 14. 15. 16.

17.

159

ophy of Thomas Reid (see note 1), which underlay Ordinary Language philosophy and much of twentieth-century British empiricism, also exercised influence over the early phenomenologists, such as Brentano (1838–1917), one of the predecessors of Reinach. (Reinach was under the influence of the (individualistic) functionalist Anton Marty (1847–1914), a student of Brentano’s.) Further, it should be underlined that Reid, Reinach, and Austin all used the philosophy of law as a common source of reference for their philosophies of language (see Section 7, below; cf. Mulligan 1987: 33–34; Nerlich and Clarke 1996: 12, 104, 211–215). Peirce and Husserl were both influenced by Bolzano (1781–1848), who had been, in turn, influenced by Johann Lambert (1728–77) and Locke; the latter two also directly influenced Peirce (cf. Nerlich and Clarke 1996: 21). Bühler was also closely associated with the Vienna Circle, including Carnap. As this suggests, and as Friedman (2000) has shown, logical positivism (of the Vienna Circle) and phenomenology shared common themes (e.g., intersubjectivity, construction of the world, etc.) and common roots (e.g., neo-Kantianism). For instance, the early Carnap made appeal to Gestalt psychology and epistemological holism, which were intimately associated with structuralism, phenomenology, and especially, the neo-Kantian notion of “constitution”. Generally speaking, the “linguistic turn” in philosophy was made up of (1) the Russellian critique of ordinary language by means of logico-mathematical analysis, (2) the Carnapian critique of metaphysical dogmas (later, epistemology), such as realism and phenomenalist positivism, by means of logic, and (3) the Austinian critique of philosophy, including logic, by means of the “commonsense” analysis of ordinary language. Austin’s philosophy of language was anticipated by Scottish commonsense philosophy, especially Reid, who influenced Peirce, Moore, and Grice, among others (see note 1, above). See Section 7, below, for the metalinguistic (more specifically, metapragmatic) and poetic functions. See Section 2 for indexicality. Recall Ross’s and other Generative Semanticists’ “performative hypothesis” (cf. Sadock 1974; Levinson 1983: 247–263, 271–273). See Section 3, above; cf. Silverstein 1976b; Rosaldo 1982; Du Bois 1987; Duranti 1993. Cf. Schieffelin, Woolard, and Kroskrity [1992] 1998; Verschueren 1999; Mey [1993] 2001: 173–205, 262–288; Jaworski, Coupland, and Dariusz 2004. As Nerlich and Clarke (1996) have pointed out, the property of self-reference (soiréférence) had been discovered, in 1929, by Erwin Koschmieder (1895–1977), who started to discuss Koinzidenzfall ‘the case of coincidence’ in the 1940s. See Section 6, above.

References Aarsleff, Hans 1982 From Locke to Saussure: Essays on the Study of Language and Intellectual History. Minneapolis, MN: University of Minnesota Press. Agha, Asif 2007 Language and Social Relations. Cambridge: Cambridge University Press.

160

Wataru Koyama

Apel, Karl-Otto 1994 Karl-Otto Apel: Selected Essays, Volume 1: Towards a Transcendental Semiotics. Edited by Eduardo Mendieta. Atlantic Heights, NJ: Humanities Press. Auer, Peter and Aldo di Luzio (eds.) 1992 The Contextualization of Language. Amsterdam: Benjamins. Austin, John L. 1962 How to Do Things with Words: The William James Lectures Delivered at Harvard University in 1955. Edited by James O. Urmson. Oxford: Clarendon Press. Bar-Hillel, Yeoshua 1946 Analysis of “correct” language. Mind 55: 328–340. Bar-Hillel, Yeoshua 1954 Indexical expressions. Mind 63: 359–376. Bateson, Gregory [1955] 1972 A theory of play and fantasy. In: Gregory Bateson, Steps to an Ecology of Mind, 177–193. New York: Ballantine Books. Bauman, Richard 2004 A World of Others’ Words: Cross-Cultural Perspectives on Intertextuality. Oxford/Malden, MA: Blackwell. Bauman, Richard and Joel Sherzer (eds.) [1974] 1989 Explorations in the Ethnography of Speaking, 2nd edition. Cambridge: Cambridge University Press. Benveniste, Emile 1966 Problèmes de linguistique générale, I. Paris: Gallimard. Blommaert, Jan 2005 Discourse: A Critical Introduction. Cambridge: Cambridge University Press. Bolinger, Dwight 1968 Aspects of Language. New York: Harcourt Brace Jovanovich. Bourdieu, Pierre 1980 Le sens pratique. Paris: Minuit. Brown, Penelope and Stephen C. Levinson [1978] 1987 Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press. Bublitz, Wolfram and Axel Hübler (eds.) 2007 Metapragmatics in Use. (Pragmatics and Beyond, n. s. 165.) Amsterdam: Benjamins. Bühler, Karl [1934] 1990 Theory of Language: The Representational Function of Language. Amsterdam: Benjamins. [Sprachtheorie: Die Darstellungsfunktion der Sprache. Jena: Fischer.] Butler, Judith 1990 Gender Trouble: Feminism and the Subversion of Identity. London: Routledge. Caffi, Claudia (ed.) 1984 Special Issue on “Metapragmatics”. Journal of Pragmatics 8(4). Cameron, Deborah and Don Kulick 2003 Language and Sexuality. Cambridge: Cambridge University Press.

The rise of pragmatics: a historiographic overview

161

Carnap, Rudolf [1934] 1937 The Logical Syntax of Language. London: Kegan Paul. [Logische Syntax der Sprache. Vienna: Springer.] Cassirer, Ernst A. 1945 Structuralism in modern linguistics. Word 1: 99–120. Caton, Steven C. 1987 Contributions of Roman Jakobson. Annual Review of Anthropology 16: 223– 260. Caton, Steven C. 1990 “Peaks of Yemen I Summon”: Poetry as Cultural Practice in a North Yemeni Tribe. Berkeley/Los Angeles/London: University of California Press. Chapman, Siobhan 2005 Paul Grice, Philosopher and Linguist. Basingstoke/New York: Palgrave Macmillan. Coffa, J. Alberto 1991 The Semantic Tradition from Kant to Carnap: To the Vienna Station. Cambridge: Cambridge University Press. Comrie, Bernard [1981] 1989 Language Universals and Linguistic Typology: Syntax and Morphology, 2nd edition. Chicago: University of Chicago Press. Coupland, Nikolas, Srikant Sarangi and Christopher N. Candlin (eds.) 2001 Sociolinguistics and Social Theory. Harlow: Pearson Education. Davis, Steven (ed.) 1991 Pragmatics: A Reader. Oxford/New York: Oxford University Press. Donnellan, Keith S. 1966 Reference and definite descriptions. Philosophical Review 75: 281–304. Donnellan, Keith S. 1978 Speaker reference, descriptions and anaphora. In: Peter Cole (ed.), Syntax and Semantics, Volume 9: Pragmatics, 47–68. New York: Academic Press. Du Bois, John W. 1987 Meaning without intention: Lessons from divination. IPrA Papers in Pragmatics 1: 80–122. Ducrot, Oswald 1984 Langage, métalangage et performatifs. In: Le dire et le dit, 117–148. Paris: Minuit. Duranti, Alessandro 1993 Intentions, self, and responsibility. In: Jane H. Hill and Judith T. Irvine (eds.), Responsibility and Evidence in Oral Discourse, 24–47. Cambridge: Cambridge University Press. Duranti, Alessandro and Charles Goodwin (eds.) 1992 Rethinking Context: Language as an Interactive Phenomenon. Cambridge: Cambridge University Press. Fillmore, Charles J. 1975 Santa Cruz Lectures on Deixis, 1971. Bloomington, IN: Indiana University Linguistics Club. Foucault, Michel 1966 Les mots et les choses: une archéologie des sciences humaines. Paris: Gaillmard.

162

Wataru Koyama

Friedman, Michael 2000 A Parting of the Ways: Carnap, Cassirer, and Heidegger. Chicago/La Salle, IL: Open Court. Gazdar, Gerald 1979 Pragmatics: Implicature, Presupposition, and Logical Form. New York: Academic Press. Giddens, Anthony [1976] 1993 New Rules of Sociological Method: A Positive Critique of Interpretative Sociologies, 2nd edition. Stanford: Stanford University Press. Goffman, Erving 1967 Interaction Ritual: Essays on Face-to-Face Behavior. New York: Anchor Books. Goffman, Erving 1974 Frame Analysis: An Essay on the Organization of Experience. New York: Harper and Row. Goffman, Erving 1981 Forms of Talk. Philadelphia: University of Pennsylvania Press. Goffman, Erving 1983 Felicity’s condition. American Journal of Sociology 1: 1–53. Grice, H. Paul 1967 Logic and conversation. Unpublished manuscript of the William James Lectures, Harvard University. Published in Studied in the Way of Words. Cambridge, MA: Harvard University Press [1989]. Gumperz, John J. 1982 Discourse Strategies. Cambridge: Cambridge University Press. Gumperz, John J. and Dell Hymes (eds.) [1972] 1986 Directions in Sociolinguistics: The Ethnography of Communication. Oxford: Basil Blackwell. Haberland, Hartmut and Jacob L. Mey 1977 Editorial: Pragmatics and linguistics. Journal of Pragmatics 1(1): 1–16. Halliday, Michael A. K. and Ruqaiya Hasan 1976 Cohesion in English. London: Longman. Hancher, Michael 1979 The classification of cooperative illocutionary acts. Language in Society 8: 1–14. Hanks, William F. 1990 Referential Practice: Language and Lived Space among the Maya. Chicago: University of Chicago Press. Harris, Randy Allen 1993 The Linguistic Wars. Oxford: Oxford University Press. Heritage, John 1984 Garfinkel and Ethnomethodology. Cambridge: Polity Press. Holenstein, Elmar 1977 Jakobson’s contribution to phenomenology. In: Daniel Armstrong and Cornelis H. van Schooneveld (eds.), Roman Jakobson: Echoes of his Scholarship, 145–162. Lisse: The Peter de Ridder Press.

The rise of pragmatics: a historiographic overview

163

Huang, Yan 2007 Pragmatics. Oxford/New York: Oxford University Press. Irvine, Judith T. 1996 Shadow conversations: The indeterminacy of participant roles. In: Michael Silverstein and Greg Urban (eds.), Natural Histories of Discourse, 131–159. Chicago: University of Chicago Press. Jakobson, Roman 1957 Shifters, verbal categories, and the Russian verb. Cambridge, MA: Harvard University Russian Language Project. Jakobson, Roman 1960 Closing statement: Linguistics and poetics. In: Thomas A. Sebeok (ed.), Style in Language, 350–377. Cambridge, MA: MIT Press. Jakobson, Roman 1973 Main Trends in the Science of Language. London: George Allen and Unwin. Jaworski, Adam, Nikolas Coupland and Galasin´ski Dariusz (eds.) 2004 Metalanguage: Social and Ideological Perspectives. Berlin: Mouton de Gruyter. Jespersen, Otto 1922 Language. London: Allen and Unwin. Kaplan, David 1978 Dthat. In: Peter Cole (ed.), Syntax and Semantics, Volume 9: Pragmatics, 221–243. New York: Academic Press. Karttunen, Lauri 1974 Presupposition and linguistic context. Theoretical Linguistics 1: 181–194. Koyama, Wataru 2000 How to be a singular scientist of words, and worlds, and other (possibly) wonderful things. Journal of Pragmatics 32: 651–686. Kripke, Saul 1972 Naming and Necessity. Cambridge, MA: Harvard University Press. Kripke, Saul 1979 Speaker’s reference and semantic reference. In: Peter A. French, Theodore E. Uehling Jr. and Howard K. Wettstein (eds.), Contemporary Perspectives in the Philosophy of Language, 6–27. Minneapolis, MN: University of Minnesota Press. Kuklick, Henrika 1991 The Savage Within: The Social History of British Anthropology, 1885–1945. Cambridge: Cambridge University Press. Lakoff, Robin T. 1973 Language and woman’s place. Language in Society 2(1): 45–80. Lakoff, Robin T. 1989 The way we were. Journal of Pragmatics 13: 939–988. Leech, Geoffrey N. 1983 Principles of Pragmatics. London: Longman. Levinson, Stephen C. 1983 Pragmatics. Cambridge: Cambridge University Press. Locke, John [1689] 1975 Essay on Human Understanding. Edited by Peter H. Nidditch. Oxford: Oxford University Press.

164

Wataru Koyama

Lucy, John A. 1992 Language Diversity and Thought: A Reformulation of the Linguistic Relativity Hypothesis. Cambridge: Cambridge University Press. Lucy, John A. (ed.) 1993 Reflexive Language: Reported Speech and Metapragmatics. Cambridge University Press: Cambridge. Lyons, John 1977 Semantics (2 volumes). Cambridge: Cambridge University Press. Merleau-Ponty, Maurice 1945 Phénoménologie de la perception. Paris: Gallimard. Merleau-Ponty, Maurice 1960 Signes. Paris: Gallimard. Mey, Jacob L. 1979 Zur kritischen Sprachtheorie. In: Jacob L. Mey (ed.), Pragmalinguistics: Theory and Practice, 411–434. The Hague: Mouton. Mey, Jacob L. 1985 Whose Language?: A Study in Linguistic Pragmatics. Amsterdam: Benjamins. Mey, Jacob L. [1993] 2001 Pragmatics: An Introduction, 2nd edition. Oxford/Malden, MA: Blackwell. Morris, Charles W. 1938 Foundations of the theory of signs. In: Otto Neurath, Rudolf Carnap, and Charles W. Morris (eds.), International Encyclopedia of Unified Science, Volume 1, No. 2, 77–137. Chicago: University of Chicago Press. Mulligan, Kevin (ed.) 1987 Speech Act and Sachverhalt: Reinach and the Foundations of Realist Phenomenology. Dordrecht: Martinus Nijhoff. Nerlich, Brigitte 1992 Semantic Theories in Europe: From Etymology to Contextuality. Amsterdam: Benjamins. Nerlich, Brigitte and David D. Clarke 1996 Language, Action, and Context: The Early History of Pragmatics in Europe and America, 1780–1930. Amsterdam: Benjamins. Peirce, Charles S. [1893–1910] 1955 Logic as semiotic: The theory of signs. In: Justus Buchler (ed.), Philosophical Writings of Peirce, 98–119. New York: Dover. Pike, Kenneth Lee 1967 Language in Relation to a Unified Theory of the Structure of Human Behavior, 2nd, rev. edition. The Hague: Mouton. Pressman, Jon F. 1994 Pragmatics in the late twentieth century: Countering recent historiographic neglect. Pragmatics 4(4): 461–489. Putnam, Hilary 1975 Philosophical Papers, Volume 2: Mind, Language, and Reality. Cambridge: Cambridge University Press. Recanati, François 1979 La transparence et l’énonciation: pour introduire à la pragmatique. Paris: Seuil.

The rise of pragmatics: a historiographic overview

165

Richardson, Alan 1998 Carnap’s Construction of the World: The Aufbau and the Emergence of Logical Empiricism. Cambridge: Cambridge University Press. Ricœur, Paul [1975] 1977 The Rule of Metaphor: Multi-Disciplinary Studies of the Creation of Meaning in Language. Toronto: University of Toronto Press. Rorty, Richard M. (ed.) 1967 The Linguistic Turn: Recent Essays in Philosophical Method. Chicago: University of Chicago Press. Rosaldo, Michelle Z. 1982 The things we do with words: Ilongot speech acts and speech act theory in philosophy. Language in Society 11: 203–237. Ruesch, Jürgen and Gregory Bateson 1951 Communication: The Social Matrix of Psychiatry. New York: Norton. Sadock, Jerrold M. 1974 Toward a Linguistic Theory of Speech Acts. New York: Academic Press. Sapir, Edward 1933 La réalité psychologique des phonèmes. Journal de Psychologie Normale et Pathologique 30: 247–265. Schieffelin, Bambi B., Kathryn Woolard and Paul V. Kroskrity (eds.) 1998 Language Ideologies: Practice and Theory. New York/Oxford: Oxford University Press. [Originally, 1992, Special Issue on “Language Ideologies”, Pragmatics 2(3).] Searle, John R. 1969 Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Searle, John R. 1983 Intentionality: An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press. Silverstein, Michael 1976a Hierarchy of features and ergativity. In: Robert M. W. Dixon (ed.), Grammatical Categories in Australian Languages, 112–171. Canberra: Australian Institute of Aboriginal Studies. Silverstein, Michael 1976b Shifters, linguistic categories, and cultural description. In: Keith H. Basso and Henry A. Selby (eds.), Meaning in Anthropology, 11–55. Albuquerque, NM: University of New Mexico Press. Silverstein, Michael 1993 Metapragmatic discourse and metapragmatic function. In: John A. Lucy (ed.), Reflexive Language: Reported Speech and Metapragmatics, 33–58. Cambridge: Cambridge University Press. Silverstein, Michael and Greg Urban (eds.) 1996 Natural Histories of Discourse. Chicago: University of Chicago Press. Stocking, Jr., George W. (ed.) 1996 Volksgeist as Method and Ethic: Essays on Boasian Ethnography and the German Anthropological Tradition. (History of Anthropology, Volume 8.) Madison, WI: University of Wisconsin Press.

166

Wataru Koyama

Tannen, Deborah (ed.) 1986 Framing in Discourse. Oxford/New York: Oxford University Press. Thomas, Jenny 1996 Meaning in Interaction: An Introduction to Pragmatics. London: Longman. Verschueren, Jef 1999 Understanding Pragmatics. London: Arnold. Volosˇinov, Valentin N. [1929] 1973 Marxism and the Philosophy of Language. Cambridge, MA: Harvard University Press. Williams, Glyn 1992 Sociolinguistics: A Sociological Critique. London: Routledge.

6.

Semiotic foundations of pragmatics Winfried Nöth

1.

Semiotics as a general framework of pragmatics

Levinson (1983: 1–5) gives the standard account of how semiotics, the general study of signs, is considered to be fundamental to linguistic pragmatics. The author of the Cambridge Textbook on Pragmatics begins with Charles W. Morris (1901–1979), the founder of a general semiotics first outlined in 1938. Morris’s “science of signs” (1938: 1–2) is both broader and narrower than the scope of modern linguistic pragmatics.1 It is broader insofar as it is not restricted to the study of verbal communication but deals with “signs in all their forms and manifestations, whether in animals or men, whether normal or pathological, whether linguistic or nonlinguistic, whether personal or social” (1964: 1), acoustic, visual, olfactory, gustatory, or tactile. It is also broader since pragmatics, as conceived by Morris, comprises branches of language studies which today fall into the domain of other disciplines of linguistics, such as “psycholinguistics, sociolinguistics, neurolinguistics, and much besides” (Levinson 1983: 2). It is narrower insofar as Morris aims at a science of signs “on a biological basis and specifically within the framework of the science of behavior” (1946: 80) indebted to the positivist paradigm of psychological and social behaviorism. As Levinson (1983: 2–4) points out, the scope of pragmatics became still narrower in the frameworks of logical positivism and analytical philosophy of ordinary language: Carnap (1942) restricts pragmatics to the logical study of language use, for Bar-Hillel (1954) pragmatics is the study of language use involving indexical words, and Montague (1970: 68) conceives pragmatics as a contextual theory of truth conditions founded on intensional logic. Sign behavior, according to Morris, involves three main factors: “that which acts as a sign [the sign vehicle], that which the sign refers to [the designatum], and that effect on some interpreter in virtue of which the thing in question is a sign to that interpreter” [the interpretant] (1938: 3). Based on this triad, Morris (1938: 6–7) defines semiotics as a field of study of the following three domains corresponding to three well-known branches of modern linguistics: syntax (or syntactics), the study of the relation between sign vehicles, semantics, the study of the relations between sign vehicles and their designata, and pragmatics, the study of the relation between sign vehicles and their interpreters (cf. Posner 1985). 20th century semiotics has developed in many directions, and some of the schools of semiotic research have largely ignored the pragmatic dimension of sign processes. This is especially true of structuralist semiotics in the tradition of Fer-

168

Winfried Nöth

dinand de Saussure (1857–1913) and Louis Hjelmslev (1899–1965), whose focus is entirely on language as a system and not on language use. It is no wonder, therefore, that Jacob L. Mey, once Hjelmslev’s PhD student in Copenhagen, in his Introduction to Pragmatics, disregards semiotics entirely as a framework of linguistic pragmatics and that Dan Sperber and Deirdre Wilson, while reducing semiotics to “a theory of codes”, dismiss its relevance to pragmatics altogether (Sperber and Wilson 1986: 7, 2). The structuralist and behaviorist heritage of semiotics has been an important obstacle to a more fruitful interdisciplinary dialogue between general semiotics and linguistic pragmatics. For a long time, it has impeded the spread of the seminal ideas which the semiotic writings of Charles S. Peirce (1839–1914) offer to linguistics in general (Nöth 2002a) and pragmatics in particular. Only relatively recently has the relevance of these ideas to pragmatics become acknowledged,2 but so far mostly only by philosophers of language and hardly by linguists (e. g., Levinson 1983: 1, fn. 1). When Morris conceived the triadic subdivision of semiotics into syntactics, semantics, and pragmatics, he was clearly influenced by Peirce. Peirce had also proposed a triadic framework for the study of signs, but the three branches of semiotics which Peirce set up describe a much wider field of study. Peirce was inspired by the medieval canon of the three liberal arts, the trivium of grammar, logic, and rhetoric.3 In contrast to the Saussurean tradition, in which linguistics is a branch of semiology (cf. Nöth 2000: 72–73), Peirce considers linguistics and semiotics as two different branches of the system of sciences. Whereas linguistics, in Peirce’s “detailed classification of the sciences” of 1902 (CP4 1.203–283), is a branch of the empirical or special sciences, semiotic(s), “the quasi-necessary, or formal, doctrine of signs”, belongs to the normative and formal sciences (CP 2.227, c. 1897). In contrast to the empirical sciences, which study “what is in the actual world”, semiotics, as a normative science, aims at inquiring into “what must be the character of all signs” and “what would be true of signs in all cases” (CP 2.227). The normative approach is particularly evident in logic since to decide whether a statement or conclusion is true or false is a normative decision. This is why, to Peirce, semiotics was “only another name for logic” (CP 2.227, c. 1897). The three branches of semiotics, in Peirce’s redefinition of the trivium of the liberal arts, are the following: (1.) speculative (or formal) grammar, (2.) logic (proper), and (3.) pure, formal or speculative rhetoric.5 Speculative grammar, the precursor of Morris’s syntactics, is the study of signs as such; it inquires into “the general conditions of signs being signs” (CP 1.444, c. 1896), and it is a “general theory of the nature and meanings of signs, whether they be icons, indices, or symbols” (CP 1.191, 1903). Of the branches of modern linguistics, not only syntax, but also phonology and morphology belong to this branch of semiotics. Logic proper, the second branch of semiotics, which Peirce also calls “critic”, is the precursor of Morris’s semantics and certainly also comprises linguistic semantics. According to

Semiotic foundations of pragmatics

169

Peirce, it deals with the relations of signs to the objects that they represent, being “the science of what is quasi-necessarily true of signs […] in order that they […] may be true; [… it] is the formal science of the conditions of the truth of representations” (CP 2.229, c. 1897). Speculative rhetoric, the precursor of Morris’s pragmatics and with it of linguistic pragmatics, deals with the relations of signs to their producers and interpreters; it studies the effects created by signs in the participants of the process of semiosis. Speculative rhetoric is “the study of the necessary conditions of the transmission of meaning by signs from mind to mind, and from one state of mind to another” (CP 1.444, c. 1896); it deals with “the formal conditions of the force of symbols, or their power of appealing to a mind” (CP 8.342, 1908). Speculative rhetoric is broader than Morris’s pragmatics in two respects. First, to study how signs are transmitted “from one state of mind to another” means that this branch of semiotics does not only study signs in communication but also in thought not communicated to any other interpreter. Thinking has pragmatic dimensions since it “always proceeds in the form of a dialogue – a dialogue between different phases of the ego” (CP 4.6, 1906). Second, speculative rhetoric is also conceived as a rhetoric of fine art, a rhetoric of “practical persuasion”, and a rhetoric of science (EP 26: 329, 1904). Peirce had the vision that the rhetoric of science, which he also called methodeutics (Santaella 1999), was “destined to grow into a colossal doctrine which may be expected to lead to most important philosophical conclusions” (CP 3.454, 1896).

2.

Semiotic foundations of communication

Linguistic pragmatics studies how language is used in communication, especially, how verbal messages are exchanged between speakers and hearers in dialogue. Merely incidentally transmitted utterances, monologues, and other forms of language use without an addressee are of no concern to the field of study (cf. Levinson 1983: 16). Communication is a central issue of semiotics; its study is also called semiotics of communication (Santaella and Nöth 2004). The general study of signs and sign processes may thus provide a theoretical framework for linguistic pragmatics. Among its topics of relevance to linguistic pragmatics are: (1) models and functions of communication, (2) the instrumental view of language in communication, (3) the problem of intentionality, (4) the nature of cooperation in conversation, and (5) the dialogical nature of the sign and its relevance to the study of communication.

170 2.1.

Winfried Nöth

Models and functions of communication

Most theories of pragmatics take a very rudimentary model of communication as the point of departure of their analyses of speech acts. With reference to Grice (1957), Levinson considers the triad S-U-R, sender (speaker, writer), utterance, and receiver (hearer, reader), to be fundamental: “Communication consists of the ‘sender’ intending to cause the ‘receiver’ to think or do something, just by getting the ‘receiver’ to recognize that the ‘sender’ is trying to cause that thought or action” (Levinson 1983: 16). The utterance, in this triad, is a “linguistic token”; U has meaning, which becomes “mutual knowledge” to S and R in the process of communication, but initially, U may have two kinds of meaning, possibly in conflict, a “speaker meaning” and a “sentence meaning” (Levinson 1983: 16). Leech, in his Principles of Pragmatics, proposes a flow diagram of communication according to which an S uttering U to H is engaged in three kinds of semiotic transaction: (1) an interpersonal transaction, or discourse, (2) an ideational transaction or message transmission, and (3) a textual transaction or text (Leech 1983: 59). As to its three constituents, S-U-H, and its notion of message transmission, this model is certainly inspired by the technical model of the “flow of information” from S to H by Shannon and Weaver (1949). This model is inadequate for the study of human communication, since it is linear, unidirectional, and assigns a merely passive role to the “recipient” of the message, making the sender the only active agent in communication. It disregards H’s autonomy in the process of interpreting S’s message as well as the important elements of circularity by which S and H are connected in communication through feedback loops by which H contributes to the production of S’s message from the moment of its articulation on (cf. Nöth 2000: 244–46). A much cited communication model in linguistic pragmatics is the semiotic model of the six factors and functions of communication which Roman Jakobson proposed in extension of Bühler’s triadic organon model of language (Bühler 1933). Jakobson defines the factors involved in a process of communication as follows: An addresser sends a message to an addressee. This message refers to a context (the “referent” of the verbal message) “seizable by the addressee”. Addresser and addressee have a common code, such as the vocabulary and grammar of the language in which they communicate, and finally, there is a contact which serves as “a physical channel and psychological connection between the addresser and the addressee, enabling both of them to enter and stay in communication” (Jakobson 1960: 353). Each of these six factors is the determinant of a specific communicative function; one and the same message may have several functions, but one is usually predominant. Messages which focus on the referent (context), such as informative, descriptive, or narrative texts, have a predominantly referential function. The emotive or expressive function is predominant when the focus is on the speaker’s attitude in relation to the meaning of the message. The use of interjections, a tremb-

Semiotic foundations of pragmatics

171

ling voice, or the frequent use of the pronouns I or me are examples of signs of an utterance in which the emotive function of language is predominant. The conative function is oriented toward the addressee; its most typical grammatical expressions are the vocative and the imperative. The phatic function predominates in messages which, without necessarily conveying any particular meaning at all, serve “to establish, to prolong, or to discontinue communication, to check whether the channel works (‘Hello, do you hear me?’), to attract the attention of the interlocutor or to confirm his continued attention (‘Are you listening?’)” (1960: 355). The metalinguistic function predominates when language is used to refer to language and verbal communication; metalanguage is language about language, e.g., as in grammar books or in the discussion of questions of terminology. Finally, the poetic function, according to Jakobson, predominates when the focus of verbal communication is on the qualities of language irrespective of its message, e.g., on the sound pattern, rhythm, or meter of a text. Although Peirce considers the model of the sign to be more fundamental than the model of communication, it is not true that the founder of modern semiotics had little to say about the topic of communication, as Habermas (1995) believes. Elements of a Peircean theory of communication have meanwhile been reconstructed.7 In his manuscript on Pragmatism of 1907, Peirce defines communication of a speaker with a hearer as follows: “Signs mostly function between two minds, or theatres of consciousness, of which one is the agent that utters the sign (whether acoustically, optically, or otherwise) while the other is the patient mind that interprets the sign” (EP 2: 403). In this description, the characterization of communication as a theater, a staged scenario of acting agents, is more than a mere metaphor. Actors on stage do not act autonomously, but they represent actions and act on behalf of other agents. But who should be the agents in communication if not S and H? This question is the topic of sections 2.2 and 2.3. 2.2.

Are verbal signs instruments of those who communicate?

Most approaches to linguistic pragmatics are based on the assumption that verbal messages are instruments by which senders seek to influence receivers. Language use is regarded as a means to some end. Searle (1969: 48), e.g., writes: “The sentence” uttered by a speaker who knows its meaning “then provides a conventional means of achieving the intention to produce certain illocutionary effects in the hearer.” Leech (1983: 36), too, postulates that speaking is a means to an end and hence an instrument for a purpose: spoken or written messages are “transferred” or “transmitted” from a speaker to a hearer, and the speaker’s purpose in this “transaction” is to “convey a particular illocutionary force to the hearer” (Leech 1983: 58–60). The view that verbal signs are used as instruments has been defended since antiquity. It is central to one of the earliest treatises on the philosophy of language,

172

Winfried Nöth

Plato’s Cratylus. In this dialogue on the relation between words and the objects for which they stand, Socrates compares the practical utility of the craftsman’s tools with the utility which words have for a speaker. A tool, says Socrates, is naturally adequate and appropriate to the task for which it serves. To the weaver, the “right and natural” (391E) instrument is the shuttle, to the blacksmith, the anvil and the hammer serve best to shape the iron, and to the shoemaker, the awl is the natural tool for piercing leather. As any practical action requires its proper tool, so does the act of speaking. The speakers’ appropriate instruments for communicating their ideas are their words (388a-389c). Practical instruments have a practical utility, whereas words serve a semiotic utility. As the weaver’s shuttle is useful for separating the warp from the woof, so is the word useful to the speaker who wants to communicate or to organize his or her ideas. In the history of semiotics, the instrumental theory of the sign has lived on in many variants (cf. Nöth 2009). In the context of pragmatics, Wittgenstein’s (1953) revival of the topic is of special interest. “Instrument”, or “tool”, are appropriate characterizations of verbal signs in Wittgenstein’s philosophy of language since these terms illustrate that the proper study of verbal signs is the study of their use. What language and instruments have in common is that they are both used and have functions: “Think of the tools in a box: there is the hammer, pliers, a saw, screwdriver, a glue-pot, nails and screws. – The functions of words are as diverse as the functions of these objects” (§ 11). Whereas Socrates focuses on the utility of verbal signs in communication, Wittgenstein’s focus is on their meaning and the way meaning is revealed in the use of signs. Not only is language in general an instrument, but also “its concepts are instruments” (§ 569). Instruments as well as language have their sense in their employment: “Look at a sentence as an instrument and its sense as its employment” (§ 421). Although the view that language serves as an instrument in communication is a proverbial commonplace, Peirce contests it. A tool is only a means, i.e., an efficient cause, but not an end, i.e., not a final cause, in the craftsman’s process of producing an effect. Words used in communication, by contrast, are more than mere tools since they have purposes of their own, which is not their speakers’ purpose. What a word means is not determined by its user. The truth conditions of a proposition are not determined by the utterer of the sentence which it contains either; they are determined by the reality to which it refers and by the consequences the utterance has. Whether an argument is valid or not is determined by the logic of language and not by the person who argues, and whether a speech act is felicitous or not depends on whether it fulfills certain felicity conditions that cannot be ignored by the utterer. Hence, to use words means to be constrained by restrictions determined by language and the consequences of its use. Instead of saying that it is the speaker who produces the utterance in accordance with the rules of syntax, semantics, and pragmatics, we can also say that these rules determine the speaker to speak the way he or she speaks, but the pragmatist

Semiotic foundations of pragmatics

173

account of language and meaning goes a step further. According to Peirce’s pragmatic maxim (see 3.1), it is not the rules of language and logic that determine our utterances, but the anticipated feedback which the success or failure of our utterances gives us whenever we speak. The anticipation of the effects of our words makes us utter them in this or that way. Nor can it be argued that the utterer’s thoughts in which his or her intention to speak is conceived use the ensuing utterance as instruments to put the speaker’s intentions into effect. Utterances which express thoughts are interpretants of these thoughts, which make the interpretants the final causes of the thoughts. Since the utterance is itself a sign with the purpose of creating further interpretants in the hearer’s mind it cannot be merely instrumentally (or efficiently) caused by the thoughts in which they were conceived. Nothing could be less in accordance with Peirce’s semiotic premises than a dualist assumption of internal thoughts becoming merely externalized as spoken or written words in communication. Sperber and Wilson (1986: 1), for example, make this assumption when they reflect on the relationship between their own thoughts and their expression in writing as follows: what we “put down on paper are little dark marks […]. As for our thoughts, they remain where they always were, inside our brains”. Peirce, by contrast, argues that signs exert their semiotic agency inside as well as outside of human brains, and perhaps even more so outside than inside. He underlines the latter view by saying that it is even “much more true that the thoughts of a living writer are in any printed copy of his book than that they are in his brain” (CP 7.364, 1902). Hence, thought-signs do not remain contained in the container of a human brain just as written words do not remain fixed in their meaning once they are written down. Words are symbols which live and grow outside of human brains in the semiotic effects which they have, and this growth takes place both in thought and in communication. What they have in common is their dialogical nature: Before the sign was uttered, it already was virtually present to the consciousness of the utterer, in the form of thought. But […] a thought is itself a sign, and should itself have an utterer (namely, the ego of a previous moment) to whose consciousness it must have been already virtually present, and so back. Likewise, after a sign has been interpreted, it will virtually remain in the consciousness of its interpreter, where it will be a sign […] and, as a sign should, in its turn, have an interpreter, and so on forward. (EP 2: 403, 1907; see below, 2.4 for the argument of the infinite regress.)

In sum, if signs have purposes, or final causes, and exert a semiotic agency of their own, they cannot be the mere instruments, that is, instrumental (or efficient) causes, of other signs; and since both thoughts and utterances are signs, utterances cannot be the instruments of the thoughts in which they were first conceived. Although the sign is not an instrument, instruments are necessary for the dissemination of signs. We need a pen and paper to write down words, and our voices are necessary instruments by which our words reach their hearers, but these words themselves are not our instruments. Not words are our semiotic instruments, but

174

Winfried Nöth

our voices, tongues and lips, pens, paper, typewriters, telephones, or computers are. The instruments by which thought-signs exert their semiotic agency are the neurons, the brain lobes, and the perceptual organs, eyes and ears, by means of which thoughts are processed. This is probably Peirce’s most radical thesis: our brain is not the only cause of our utterances; it is merely the efficient cause of the thoughts and of their expression. The ideas on which this approach to semiosis is based have parallels in current quasi-biological positions of the philosophy of mind known as memetics and teleosemantics; the biolinguistic assumption is here that words in culture spread like genes spread in biological species (cf. Nöth 2009). In order to deconstruct the autonomy of the speaking subject, Peirce, for the sake of the argument, may exaggerate here and there when he speaks of the purpose of the sign which is not the purpose of the user. Elsewhere, he acknowledges two purposes, the purpose of the sign and the purpose of the sign user, in a mutual feedback control circuit. In an evolutionary perspective, he describes the loop in which sign users using signs and signs using sign users interact in terms of a process of co-evolution: Man makes the word, and the word means nothing which the man has not made it mean, and that only to some man. But since man can think only by means of words or other external symbols, these might turn round and say: “You mean nothing which we have not taught you, and then only so far as you address some word as the interpretant of your thought.” In fact, therefore, men and words reciprocally educate each other. (CP 5.313, 1868)

There is a certain affinity between this Peircean account of the relationship between the signs and their users on the one hand and Saussure’s theories of the immutability of the verbal sign and the determination of each signifier by the system of which it is part on the other. As Stetter (1989: 159) has pointed out in his Saussurean interpretation of the above quote: we are not the masters of our signs “who determine the interpretation of terms but it is rather determined by presuppositions inherent in the system of language”, but Peirce’s object of the sign is of course more than the system of which the sign is a part. The co-evolution of signs and their users can only take place if purposes are ascribed to both. But this does not mean that the sign user is the autonomous agent external to the signs as which the intentionalists describe him (see 2.3). Peirce solves the apparent paradox of the conflict between the purposes of the sign user and the one of the sign with the argument that sign users are themselves objects of signs (see 2.5) so that their intentions are, in turn, purposes of signs. By no means do sign users pursue nonsemiotic purposes, for they themselves are and act as signs in the flow of thought-signs and public signs which constitute them as signs. This is the line of argument of the famous passage in which Peirce defines the human being as a sign:

Semiotic foundations of pragmatics

175

There is no element whatever of man’s consciousness which has not something corresponding to it in the word; and the reason is obvious. It is that the word or sign which man uses is the man himself. For, as the fact that every thought is a sign, taken in conjunction with the fact that life is a train of thought, proves that man is a sign; so, that every thought is an external sign, proves that man is an external sign. That is to say, the man and the external sign are identical, in the same sense in which the words homo and man are identical. Thus my language is the sum total of myself; for the man is the thought. (CP 2.314, 1868)

If the sign user is a sign, his or her intentions are signs, too, and in the speaker’s utterances the underlying intentions have to be interpreted as objects of signs. As the object determines the sign so does the sign determine its interpretants. Words “only stand for the objects they do, and signify the qualities they do, because they will determine, in the mind of the auditor, corresponding signs” (CP 2.92, 1902). As any object of a sign determines the sign which it represents so do the speaker’s intentions, since they are objects of signs. This is how the speaker’s intention participates in the creation of an interpretant. 2.3.

Intentionality of communication and the purpose of verbal signs

The classics of linguistic pragmatics agree in considering intentionality as a prerequisite of verbal communication. Levinson (1983: 16) summarizes as follows: “Communication consists of the ‘sender’ intending to cause the ‘receiver’ to think or do something, just by getting the ‘receiver’ to recognize that the ‘sender’ is trying to cause that thought or action. So communication is a complex kind of intention that is achieved or satisfied just by being recognized.” Searle (1983: 164–66) postulates two intentions. First, there must be an intention to represent on the speaker’s side, but since it is possible to represent without the purpose to address or to influence anybody, a second intention is required: the intention to communicate. Searle (1983: 166) explains that it is impossible to inform a hearer that it is raining without intending that this message represents truly or falsely the weather condition expressed. Not only the speaker’s intention but also the hearer’s awareness of this intention is a prerequisite of communication. Furthermore, the belief produced in the hearer’s mind as the result of the message must have its cause in the speaker’s intention (cf. Sperber and Wilson 1986: 23). In structural semiotics, intentionality is the criterion by which communication is distinguished from signification. The distinction has its roots in the writings of Buyssens (1943).8 All signs which humans can interpret, including the signs of physical nature, have signification, but not all of them are intended; only those signs constitute acts of communication which have an addresser whose intention it is to convey a message to an addressee. In the tradition of analytical language philosophy, Grice (1957) interprets the difference in question in terms of the dichot-

176

Winfried Nöth

omy of “natural” (i.e., unintentional) and “nonnatural” (i.e., intended) meanings. The latter term is unfortunate, though. To call the signs of natural language “nonnatural”, as Grice does, fails to do justice to the continuity of semiosis in the evolution of life from nature to culture (cf. Nöth 2008). Prieto (1966: 33–38) anticipates Searle’s theory of the two communicative intentions when he distinguishes between two indications which signs have in communication, a notificative and a significative indication. Significative indication is the one by which a sign indicates whatever it means (i.e., its signification). Notificative indication, by contrast, indicates that the sign is being communicated with the intention of a sender. The content of a notificative indication is so to speak: “Attention! This sign is intended to convey a message to you!” (cf. Hervey 1982: 71). If signs can only be intended or not, intentionality cannot be a matter of degree. However, intentions are volitions, and wishes and desires may vary in strength. Addressers may be divided as to their intentions, they may have doubts and hesitate whether they should communicate or not, or they may merely pretend to communicate and act as if they communicated (cf. Parret 1994). Such factors of intentionality are rarely taken into consideration in linguistic pragmatics. Psychoanalysis, by contrast, has shown how verbal and nonverbal messages may also communicate hidden intentions which the utterer may want to conceal but fails to do, e.g., when he or she, impatient with the hearer, loses control over his or her voice, sounding angry despite the intention to appear calm. In such a scenario, the speaker’s loss of control over his or her voice is the sign of a conflict between two intentions, one determined by the unconscious mind, the other by the conscious self. The conflict of intentions is thus the symptom of a divided self in which a conscious mind aims at concealing what the unconscious mind intends to express. Linguistic pragmatics also disregards what the communication theorists Watzlawick, Beavin, and Jackson (1967) say about the ubiquity of communication in everyday life. In their study Pragmatics of Human Communication, the authors postulate the metacommunicative axiom that it is impossible not to communicate, be it verbally or nonverbally: “Behavior has no opposite. There is no such thing as nonbehavior […]. One cannot not behave”, and any form of behavior in social interaction is of semiotic relevance since “one cannot not communicate” (Watzlawick et al. 1967: 48–49). Even silence and doing nothing are messages. Peirce’s semiotic theory of intentionality (cf. Kappner 2004) and his theory of the purpose of signs in sign processes are rather incompatible with the views of intentionality held in mainstream linguistic pragmatics.9 Peirce defines the sign as a medium (MS10 339: 283r, 1906), but not as one which mediates between an addresser and an addressee; instead, the sign logically mediates between its object, which it represents, and its interpretant, which is the interpretative effect it creates. The object of the sign is the verbal or nonverbal knowledge, feeling, behavior, or habit which the sign user associates with the sign. The knowledge of its object is a

Semiotic foundations of pragmatics

177

prerequisite for the interpretation of the sign. While the object of the sign includes aspects of what linguists define as meaning (see 3.3), the interpretant includes the illocutionary force and the perlocutionary effect of a speech act, as defined by Austin (1962). In Peirce’s definition, the interpretant is the cognitive, emotional, or behavioral effect which results from the interpretation of the sign. According to Peirce, the sign is a semiotic agent with its own intention (cf. 2.2). Signs continue to exert their agency in the absence of those who first produced them. This topic is also discussed by Derrida (1988: 8), who uses the metaphor of the machine for the agency which the written sign exerts irrespective of its producer: “To write is to produce a mark that will constitute a sort of machine which is productive in turn, and which my future disappearance will not, in principle, hinder in its functioning, offering things and itself to be read and to be rewritten.” Peirce’s view of the agency of the sign goes beyond the mere assumption of its interpretability in the absence of its writer. Signs have a life of their own, which can be studied in processes of language change and in the way texts are interpreted in new ways as time passes and signs change. In 1904, Peirce writes about the purpose of the symbol: “The symbol, by the very definition of it, has an interpretant in view. Its very meaning is intended. Indeed, a purpose is precisely the interpretant of a symbol” (EP 2: 308). With purpose, Peirce thus does not mean the sign user’s intention, but the sign’s intention to represent its object and to create and interpretant, i.e., to “be interpreted in another sign” (MS 1476, 1904). Purpose is thus a semiotic teleology inherent in the sign. Not only uttered or written signs have purposes, but also thought-signs. Their purpose is to act in a mental dialogue in which one thoughtsign is “translated or interpreted in a subsequent one” (CP 5.284, 1868). On the one hand, the sign represents its object to the mind of an interpreter, on the other, the object represented by the sign is “in a sense the cause, or determinant, of the sign even if the sign represents its object falsely” (CP 6.347, 1909; cf. Parmentier 1985). To justify the autonomy of the sign by its being determined sounds contradictory, but it describes the process of an autopoietic self-determination of the sign. Since the object of the sign is not external, but internal to the sign, to say that the sign is determined by its object means that it is determined by a dynamics inherent in itself and not by an agency outside of the sign (cf. Colapietro 1995). The way in which a sign is determined by its object depends on its modes of relation to its object and its interpretants. Rhematic signs, such as words in isolation, are determined to represent their objects as such, without affirming, questioning, or negating their existence. Dicentic signs, such as affirmations, are determined by the truth of what they represent, and arguments are determined by the laws of logic and reasoning in general. Symbols are determined by the habits of their users, by which they have become symbolic signs. Indices are determined by their objects because they are related to them by causal, temporal, or spatial relations, whereas icons represent their objects by means of qualities of their own. All these factors

178

Winfried Nöth

determine the sign independently of their use. For example, a rhematic symbolic legisign such as the word apple, uttered by a speaker, is determined by the habits of the speakers of English to represent the fruit of the apple tree by means of this word; the word represents its object as a general kind without affirming the existence of any particular fruit. The sign user may try to escape from his or her determination by the object of this sign and use this word to refer to a pear, but in the long run, words cannot be used in this way because the speech acts will fail. A sign user may also lie or argue against the laws of logic, but in the long run, to ignore the truth and the laws of reasoning will not pay. Furthermore, verbal signs are determined by biolinguistic and evolutionary constraints, which have restricted the speakers’ possibilities of articulation and perception (cf. Deacon, 1997: 116). 2.4.

Conversational cooperation, commonage, and the common ground

The model of communication underlying the classical theories of linguistic pragmatics is one of rational discourse.11 The aim of dialogic communication is guided by the ideal of mutual consensus; the speaker’s purpose must be achieved. Communication is modeled as being cooperative; its goal is not only comprehension but also the full acceptance of the intended message on the part of the hearer; speech acts can only succeed or fail, and successful speech acts fulfill felicity conditions; the speakers’ intentions are described as being sincere, and the messages are always relevant. Such accounts of verbal exchange between speakers and hearers model communication according to the ideals of a desirable symmetry between the agents interacting in dialogue. The underlying assumption is that a full congruence between the speakers’ intended meaning and the hearers’ understanding of it is not only possible but also desirable. To reach consensus is indeed an ideal of dialogic interaction in rational discourse. The very concept of communication implies, in its root, the notion of making something common to those who communicate. Morris (1946: 195) expresses this normative goal of communicative interaction in a definition according to which communication is “the use of signs to establish a commonage of signification”. The assumption that commonage is an essential characteristic of communication is not undisputed. In everyday language, e.g., it is contradicted by the military metaphors by which speakers represent their arguments as “positions” which need be “defended” (Lakoff and Johnson 1980: 7). In the sociological theory of communication, Luhmann (1984) starts from the basic assumption that the differences between the utterer’s and the interpreter’s frames of mind instead of their common ground, is the source of all dialogic interaction, whereas Habermas (1981) founds his Theory of Communicative Action on the assumption that consensus is constitutive of rational communicative interaction. As a prerequisite for, and means of, reaching the normative ideal of dialogic commonage, linguistic pragmatics recognizes such principles as Grice’s maxims

Semiotic foundations of pragmatics

179

of conversational cooperation and its derivative, the principle of relevance advocated by Sperber and Wilson. A precursor in the study of these normative principles is Peirce (Pietarinen 2004, Short 2007: 213–14). Although by no means restricted to the study of rational discourse (Oehler 1995: 267–68), Peirce’s speculative rhetoric describes conversational cooperation as follows: Honest people, when not joking, intend to make the meaning of their words determinate, so that there shall be no latitude of interpretation at all. That is to say, the character of their meaning consists in the implications and non-implications of their words; and they intend to fix what is implied and what is not implied. They believe that they succeed in doing so, and if their chat is about the theory of numbers, perhaps they may. (CP 5.447, 1905)

However, immediately after this description of rational conversational cooperation, Peirce also points out that commonage can only serve as a regulative idea, an ideal which can rarely be reached in ordinary conversation. The sentence after the above passage goes on to say: “But the further their topics are from such presciss12, or ‘abstract’, subjects, the less possibility is there of such precision of speech. Insofar as the implication is not determinate, it is usually left vague” (CP 5.447, 1905). Furthermore, Peirce also argues that, despite all efforts of cooperation, dialogues also evince a fundamental divergence of interest between the participants. From this perspective, Peirce calls the participants of dialogue “opponents” (MS 515: 25, s. d.; cf. Hilpinen 1995: 293) and describes dialogic interaction in terms of a metaphorical scenario of war: “The utterer is essentially a defender of his own proposition and wishes to interpret it so that it will be defensible. The interpreter, not being so interested, and being unable to interpret it fully without considering to what extreme it may reach, is relatively in a hostile attitude, and looks for the interpretation least defensible” (MS 9: 3–4, c.1903). Relevance, so much discussed in linguistic pragmatics since Grice, is another principle about which Peirce has written (cf. Pietarinen 2004). Peirce discusses it as follows: “If the utterer says ‘Fine day!’ he does not dream of any possibility of the interpreter’s thinking of any mere desire for a fine day that a Finn at the North Cape might have entertained on April 19, 1776. He means, of course, to refer to the actual weather, then and there, where he and the interpreter have it near the surface of their common consciousness” (MS 318: 32–33, 1907). Whereas Grice would describe the elliptic utterance “Fine day” as one which requires filling the gap of what remains unsaid by means of nonconventional conversational implicatures, Peirce would consider most conversational implicatures unnecessary since communication, studied from the semiotic perspective, is not only verbal communication. What linguistic pragmatics describes as gaps between what is said and what is left out may be a gap in the verbal message, but interpreters of a dialogue do not only interpret words; they also interpret nonverbal signs and the situational circumstances of the utterance. What appears a gap in the verbal message may not be a gap in the multi-channel message of the dialogic exchange. Furthermore, not

180

Winfried Nöth

everything that is left open and undetermined in conversation needs to be filled by implicatures either. It is wrong to assume that the speaker’s message is always clearly determined and that the interpreter’s only task is to fill the gaps in order to restore the intentions which seem unclear. Vagueness and the impossibility of resolving it may be the speaker’s purpose and the interpreter’s source of creative thinking (see 3.3). In Peirce’s speculative rhetoric, commonage is not only a normative goal; it is also its prerequisite of the dialogue. As the prerequisite of communication, it belongs to the domain of the object of the sign; as its goal, it pertains to the interpretant created by the sign. The object of the sign can only be known if both the utterer and the interpreter have collateral experience of the Real or Dynamical Object of a sign. In order to know what a sign represents, we need to have “previous acquaintance with what the sign denotes” (CP 8.179, s. d.). The sign itself cannot express its dynamical object, “it can only indicate and leave the interpreter to find out by collateral experience” what it represents (EP 2: 198, 1909). There is nothing “nonconventional” in this process of interpretation, as Grice assumes with his implicatures. Interpretation, according to Peirce is always a natural process of abductive, inductive, and deductive inferencing. An aspect of the object of the sign, i.e., that which must be presupposed for communication to be successful, is the common ground of experience and knowledge which the utterer and the interpreter share. Both Grice and Searle have described it as a prerequisite of conversation. Pietarinen (2004: 302–305) shows in which respect Grice’s and Searle’s theories of the preconditions of successful communication were anticipated by Peirce. In 1908 Peirce formulates it as follows: “No man can communicate the smallest item of information to his brother-man unless they have […] common familiar knowledge […] such that each knower knows that every other familiarly knows it, and familiarly knows that every other one of the knowers has a familiar knowledge of all this” (MS 614: 1–2, in Pietarinen 2006: 438). To make clear that no circulus vitiosus is involved in this endless reciprocity, Peirce underlined that this infinite regress is a logical and not a psychological one: “Of course, two endless series of knowings are involved; but knowing is not an action but a habit, which may remain passive for an indefinite time” (MS 614: 1–2, in Pietarinen 2006: 438). Insofar as commonage is the goal of communication, it pertains to the interpretant of the sign. Peirce calls the meaning effects common to the speaker and the hearer the cominterpretant of the sign. At the same time, he recognizes that there may be effects of meaning which are specific to the utterer and the interpreter, the former being the intentional, the latter the effectual interpretant:

Semiotic foundations of pragmatics

181

There is the Intentional Interpretant, which is a determination of the mind of the utterer; the Effectual Interpretant, which is a determination of the mind of the interpreter; and the Communicational Interpretant, or say the Cominterpretant, which is a determination of that mind into which the minds of utterer and interpreter have to be fused in order that any communication should take place. This mind may be called commens. It consists of all that is, and must be, well understood between utterer and interpreter at the outset, in order that the sign in question should fulfil its function. (EP 2: 478, 1906)

The commind, as Peirce calls the fusion of the two minds in one and the same commens elsewhere, is a normative ideal. In practice, the ideal of a commind can never be fully reached since mutual understanding is only possible in a fragmentary way. In 1907 (MS 318), Peirce points out that no interpreter can be said to have access to the utterer’s mind. Knowledge about what a speaker means can only be obtained in fragments, and such fragments are mere “copies of a scrap torn out of another’s life”. By such scraps we can only “supplement the ideas of [our own] life” (MS 318: 194, 1907; cf. Johansen 1993: 198–202). As interpreters we must match those fragments found in the signs of the utterance with our own discourse universe and find out where they can be “inserted or recopied” in our own “panorama of universal life” (MS 318: 194, 1907). Interpretation is thus a semiotic patchwork put together by abductive, inductive and deductive reasoning. 2.5.

Dialogue as the prototype of verbal communication, and the dialogical nature of the sign

Peirce not only considered the dialogue as “a wonderfully perfect kind of signfunctioning” (EP 2: 391, c. 1906), he also had derived his first model of the sign together with the first version of his three categories, the I, the IT, and the THOU from the three elementary constituents of a dialogue (W 113: xxvii, 174, 1865).14 Fisch explains how the model of the sign is derived from the model of the dialogue: “Peirce began where most of us begin, with a model, which, taken by itself, would suggest too narrow a definition; namely the model of conversation between two competent speakers of the same natural language – say, English. […] It goes without saying that words are signs; and it goes almost without saying that phrases, clauses, sentences, speeches, and extended conversations are signs” (Fisch 1986: 357). In Peirce’s fully developed semiotics, “a sign, or representamen, is something which stands to somebody for something in some respect or capacity. It addresses somebody, that is, creates in the mind of that person an equivalent sign, or perhaps a more developed sign. That sign which it creates I call the interpretant of the first sign. The sign stands for something, its object” (CP 2.228, c.1897). Sign, object, and interpretant must be understood as following each other in time. The object comes first insofar as it precedes the sign; it is that which must be known before the sign can be understood. The sign comes next since it represents the object to an in-

182

Winfried Nöth

terpretant, and the interpretant follows last since it is created by the sign. This temporal sequence makes clear how the three correlates of the sign correspond to the three correlates of a dialogue, addresser (or sender), message, addressee (or receiver): the sign (or Representamen) is the message; it addresses the interpretant as its receiver, and the sender of this message is its object, which determines the sign (cf. Bergman 2003: 11, Pietarinen 2006: 426). Despite the close association of the interpreter and the interpretant in the process of interpretation, the interpretant must not be confused with the mind of the addressee of a message. In its full definition, the interpretant is any “significant effect of the sign” (CP 2.303, 1902), be it a feeling, action, or thought, and since significant effects of signs are not restricted to individuals and to the present moment of sign interpretation, the interpretant can also be a memory, a habit, a fashion, an innovation, a revolution, a war, or a peace treaty. In all cases, the interpretant is itself a sign. The object of the sign corresponds to its sender in an equally impersonal understanding (cf. 2.2 on the utterer’s intention as an object of the sign). Ransdell (1977) explains why Peirce equates the two instances of conversation which linguists are used to distinguishing as the referent of the sign on the one hand and the sender of a message on the other: to the degree that the reality, the phenomena which we perceive, experience, or think about, are interpretable phenomena, they are objects of signs. Reality is a sign producer since it is the cause of the signs which we have of it; after all, we interpret reality by means of its signs, which in turn create interpretants. Both the sender of a message who, although not a creation of this message, is nevertheless the source of its distribution, and the sign’s object precede the moment and are the causes of the production of the sign. Not only the sender but also the sign itself has an agency in the process of semiosis, as shown above (2.2–3). Peirce calls it the “action of the sign” (e.g., CP 5.472–73, 1905), but again, this agency, which is due to the purpose of the sign, is neither restricted to the purpose of a human being nor must it be the effect of a verbal sign. There are even signs which have no sender at all. These are evidently the natural signs (cf. Nöth 2008). Since “there can be no isolated sign”, Peirce attributes a Quasi-utterer to (natural) signs without a real utterer (CP 4.551, 1906). There are also signs without an interpreter but a Quasi-interpreter (CP 4.551, 1906), for example, “if nobody takes the trouble to study the record” (EP 2: 404, 1907). According to Peirce’s dialogic semiotics, it is essential that the utterer or quasi-utterer as well as the interpreter or quasi-interpreter are not instances exterior to, but can be found within the sign, in which “they must nevertheless be distinct” although “they are, so to say, welded” in the sign (CP 4.551, 1906). A summary of the resulting simultaneity of the presence and absence of the utterer in the dialogical sign is the following:

Semiotic foundations of pragmatics

183

The action of a sign generally takes place between two parties, the utterer and the interpreter. They need not be persons; for a chameleon and many kinds of insects and even plants make their livings by uttering signs, and lying signs at that. Who is the utterer of signs of weather, which are not remarkably veracious, always? However, every sign certainly conveys something of the general nature of thought, if not from a mind, yet from some repository of ideas, or significant forms, and if not to a person, yet to something capable of some how “catching on”. (MS 318, S. 17, 1907)

In the dialogic process of conversation, the continuous alternation between the role of the utterer and the one of the hearer involves a permanent transformation of signs into interpretants and interpretants into the objects of new signs creating new interpretants. This continuous process of alternations and transformations is a source of creativity in semiosis. Since not only the utterers but also the objects of the signs participate in the dialogue in the sense that they determine its interpretants, not only the voices of two participants but a plurality of voices is audible. Dialogue is a Bakhtinian polyphonic chorus with a plurality of participants speaking the voices of many discourses (cf. Nöth 2009).

3.

Pragmatism and Peirce’s semiotics of meaning and reference

Peirce’s semiotics is also relevant to issues of linguistic pragmatics such as the relationship between and the respective scopes of semantics and pragmatics, the difference between speaker, hearer, and utterance meaning, the nature of reference and indexicality in language, and the theory of speech acts. Even the concept of pragmatics has its origin in Peirce’s writings. 3.1.

Pragmatics, pragmatism, and pragmaticism

The term pragmatics, as introduced by Morris (see 1.), was clearly inspired by the philosophy of pragmatism, whose main representatives in the first decades of the 20th century were William James and John Dewey, and whose founder was Peirce (cf. Murphy 1990). In a letter of 1912, Peirce explained why and how he coined the term in the 1870s (Houser 1998: xxi-xxii): “When I gave the doctrine of pragmatism the name it bears […] I derived [… it] from , – ‘behavior’, – in order that it should be understood that the doctrine is that the only real significance of a general term lies in the general behavior which it implies” (in: Eisele 1987: 95). Morris’s change of the concept of pragmatism to pragmatics made the term more compatible with the names of the neighboring fields of pragmatics in linguistics syntactics and semantics as well as with the term linguistics itself. One of the main aims of pragmatism is to overcome the dualism between thought and action (Colapietro 1992: 430–31). Peirce formulated the essence of pragmatism, as he conceived it, in several versions of his “pragmatic maxim” for-

184

Winfried Nöth

mulated in several versions since 1878. In 1902, the maxim stated the following: “Consider what effects, that might conceivably have practical bearings, we conceive the object of our conception to have. Then, our conception of these effects is the whole of our conception of the object” (CP 5.9, c. 1905). Peirce gives the following account of how this maxim should be applied: “Pragmatism is the principle that every theoretical judgment expressible in a sentence in the indicative mood is a confused form of thought whose only meaning, if it has any, lies in its tendency to enforce a corresponding practical maxim expressible as a conditional sentence having its apodosis in the imperative mood” (CP 5.18, 1903). Colapietro (1992: 433) gives an example of how this principle can be followed in practice: “To learn most clearly the meaning of an assertion – e.g., of ‘Sugar is soluble’ – requires us to translate the assertion into a maxim such as ‘If you place this substance into water or similar liquids it will dissolve.’” Disappointed with how his word coinage became “abused in the merciless way that words have to expect when they fall into literary clutches” (CP 5.414, 1905) in a trend that culminated in the “deterioration of pragmatism into behaviorism” (Nadin 1993: 223), Peirce, in 1905, began to replace the term pragmatism with the neologism pragmaticism, a word “ugly enough to be safe from kidnappers”, as its author remarked (CP 5.414, 1905). Parret (1983: 3) believes that there are three main lessons which the philosophy of pragmatism has taught to linguist pragmatics: “Meaning is context bound, rationality is discourse bound, and signifying should be seen as understanding”. What linguistic pragmatics has learned from the tradition of pragmatism is certainly the importance of the study of language with respect to its consequences for human action and behavior. A common ground of Peirce’s semiotics and the classics of linguistic pragmatics is the normative approach to communication. Among the other insights which Peirce has contributed to linguistic pragmatics is the pragmatic “principle of expressibility” often attributed to Searle. Searle’s (1969: 220) definition is: “For any meaning X and any speaker S whenever S means (intends to convey, wishes to communicate in an utterance, etc.) X then it is possible that there is some expression E such that E is an exact expression of or formulation of X.” Peirce reduces it to the form: “thought and expression are really one” (CP 1.349, 1903). 3.2.

Reference and indexicality

Reference and deixis are key issues of linguistic pragmatics; in Peirce’s semiotics, in which the term reference is hardly used (cf. Nöth 2006), these topics are studied as indexical signs. Searle (1969: 26–29, 72–96) studies reference as a speech act. Expressions may be used to refer or they may not be used to refer. A man came is used to refer to ‘a man who came’; John is a man is not used referentially, but predicatively (Searle 1969: 27). Necessary conditions of a referring speech act are:

Semiotic foundations of pragmatics

185

“whatever is referred to must exist” (1969: 77); “There must exist one and only one object to which the speaker’s utterance of the expression applies”, and: “The hearer must be given sufficient means to identify the object from the speaker’s utterance of the expression” (Searle 1969: 82). Existence, singularity, and referential selectivity are criteria of the genuine index in Peirce’s semiotics. In fact, all of Searle’s conditions of reference are conditions of the genuinely indexical sign: “A genuine Index and its Object must be existent individuals (whether things or facts)” (CP 2.283, c. 1902), and indexical signs are “directions of what to do to find the object” (CP 2.289, c. 1893). In contrast to Searle, Peirce also studies to what degree nonverbal signs can be genuine indices and to what degree language may be insufficient for the purpose of referring to an individual object. Furthermore, he also studies another kind of index, which he calls degenerate index. Such indices do not refer directly to individual existing object; instead, they refer to symbols or mental objects (cf. Nöth 2000: 187). Their prototype is the relative pronoun, which is a sign referring to another sign. Since not all expressions are used to refer, according to Searle, not all of them have a referent. In Peirce’s semiotics, too, only genuine indices have a referent in the sense of referring indexically to an individual existent object, but although not all signs have a referent, all signs have an object.15 Peirce’s semiotics is utterly incompatible with the dualistic theories of referential semantics which split verbal reference into two mutually independent domains, one being external to the human mind, the other being mental and hence internal to it. According to Peirce, not reference and sense are the correlates of the sign, but the object and the interpretant, and all signs have an object as well as an interpretant. All words are signs which represent an object and determine an interpretant, even the conjunction and. Savan (1994: 189) describes its object as our experience of how of two or more entities are combined to form a group, a set, or otherwise a whole, but since the object of the sign need not be a material thing, we should add that the idea of a symmetrical conjunction is among the objects of this sign. A sign without an object would be a contradiction in terms in Peirce’s semiotics. In contrast to the referent of logical semantics, which is defined as an individual or class of existents, Peirce’s object of the sign may also be a feeling, an experience, a cognition, a thought, an imagination, or even a fictional event. Unlike the sense or meaning of a word, the interpretant is not necessarily a mental concept, a thought, or an idea corresponding to the sign; it can again be a feeling, an action, or even be a material thing, provided it is a result of a sign. For example, the cake baked according to a cooking recipe is among the interpretants of the recipe practically interpreted by the cook. The object of the sign may also be something “believed formerly to have existed or expected to exist” or “something of a general nature” (CP 2.232, 1910). Only rarely is it a “thing”. It is rather that information, knowledge, or experience which an observer of the sign must have in

186

Winfried Nöth

order to be able to interpret the sign. Peirce had adopted the term object from 13th century scholastic terminology where objectum meant “a creation of the mind in its reaction with a more or less real something […] upon which cognition is directed” (MS 693A: 33; Pape 1996b: 115). The index is defined in opposition to the symbol and to the icon as a class of signs comprising natural but also conventional signs. An index is “really affected” (CP 2.248, c. 1903) by, or “physically connected with, its object; the sign and its object make an organic pair, but the interpreting mind has nothing to do with this connection except remarking it, after it is established” (CP 2.299, c. 1895). Among the features of Peirce’s index are that it directs the interpreter’s attention towards the object and that it asserts nothing, but shows the object (cf. Goudge 1965: 53–54). Examples of verbal and nonverbal indexical signs are: a weathercock, a yardstick, a photograph, a rap on the door, a pointing finger, any other gesture that indicates a present emotion, a cry for help, deictic words, proper names, possessive, relative, personal, and selective pronouns. All verbal utterances contain explicit or implicit indices referring to, or identifying, the situational circumstances, the time, place, or persons in whose context they were uttered. The adequate interpretation of these indices is necessary to distinguish successfully between the relevant and irrelevant interpretation of a sign (cf. 2.4): If, for example, a man remarks, ‘Why, it is raining!’ it is only by some such circumstances as that he is now standing here looking out at a window as he speaks, which would serve as an Index (not, however, as a Symbol) that he is speaking of this place at this time, whereby we can be assured that he cannot be speaking of the weather on the satellite of Procyon, fifty centuries ago. (CP 4.544, 1906)

Morris did not adopt the term index, but his category of identifiors corresponds to Peirce’s index (1946: 154, 362). But in contrast to Peirce, Morris restricted the class of identifiors to spatio-temporal deixis. Identifiors signify “locations in space and time (locata) and direct behavior toward a certain region of the environment”. The identifior “has a genuine, though minimal, sign status; it is a preparatorystimulus influencing the orientation of behavior with respect to the location of something other than itself”. Morris distinguished three kinds of identifiors: indicators, “which are non-language signals”, descriptors, which “describe a spatial or temporal location”, and namors, which are “language symbols, and hence, substitute signs synonymous with other identifiers”. 3.3.

Meaning in dialogue and its vagueness

Several meanings are involved in the utterance of a speaker addressed to a hearer. Searle (1969: 45–48) distinguishes between intended and conventional meanings, Leech (1983: 6) between pragmatic and semantic meanings, and Levinson (1983:

Semiotic foundations of pragmatics

187

16–17) between speaker meanings and utterance or sentence meanings. The various kinds of meaning distinguished by these authors partly overlap, but some of the criteria according to which these distinctions are made differ significantly. Leech’s “semantic” meaning roughly corresponds to Levinson’s sentence meaning and Searle’s conventional meaning. All three kinds of meaning are opposed to what is studied in linguistic pragmatics, i.e., how meaning is intended by a speaker and understood by a hearer. Leech defines “semantic meaning” as a meaning “defined purely as a property of expressions in a given language, in abstraction from particular situations, speakers, or hearers” (1983: 6). Can such a meaning abstracted from the circumstances of its use be called “conventional”, as Searle suggests when he opposes “conventional” to “intended meaning”? The idea that intended meaning is opposed to conventional meaning is rejected by Strawson (1971), who points out that conventional meaning is not only semantic meaning but pertains to pragmatic meaning, too, since speech acts, like words and sentences, are also determined by conventions. The study of pragmatic meaning, according to Leech (1983: 14), begins with the study of utterances: “In fact we can correctly describe pragmatics as dealing with utterance meaning and semantics as dealing with sentence meaning.” With this definition, Leech’s dichotomy of semantic vs. pragmatic meaning is no longer compatible with Levinson’s dichotomy of utterance meaning and speaker meaning, which suggests that the study of speaker meaning is the domain of pragmatics and not the study of utterance meaning. Furthermore, Levinson’s distinction between speaker meaning and utterance meaning raises the question of how speaker meaning can differ from utterance meaning at all. Is not the utterance meaning the meaning conveyed by the utterer, i.e., speaker? Can an utterer, by means of an utterance, convey any other meaning than his own or her own meaning? The distinctions drawn by the three authors thus do not only overlap, they also contradict each other. If speaker meaning is the utterer’s and thus also the intended meaning, one can no longer oppose it to utterance meaning (as Levinson does). Speaker meaning can only be intended meaning since whatever a speaker utters must be intended, but if utterance meaning is speaker meaning and speaker meaning is intended meaning, the congruence of theses meanings makes Leech’s distinction between semantic and pragmatic meanings invalid. On the other hand, if speaker and utterance meanings are intended meanings, and if utterances are determined by conventions, too, Searle’s distinction between intended and conventional meanings is invalid. If there is speaker meaning there should also be hearer meaning. Grice (1989) takes it into consideration but the dichotomies set up by Leech, Searle and Levinson neglect it. Sbisà and Fabbri (1980: 314) criticize this neglect for reducing “the interpretive work of the addressee […] to the mere recognition of a set of coherent intentions on the part of the agent/speaker”. The neglect of the category of a hearer

188

Winfried Nöth

meaning in linguistic pragmatics is due to its normative approach to the study of conversation. Like Chomsky, who pursues a normative approach to syntax when he postulates the idealized competent speaker as one who commits no mistakes in uttering sentences, Searle (1969: 20) explicitly adopts a normative approach to the description of the speaker’s meanings in conversation when he declares that “cases where the speaker does not say exactly what he means – the principal cases of which are nonliteralness, vagueness, ambiguity, and incompleteness – are not theoretically essential to linguistic communication”. Speakers can “mean more than they say” (Leech 1983: 9). According to Searle (1969: 45) the difference between what a speaker means and what he or she says is a difference between the meanings intended by the speaker and those conventionally inherent in his or her utterances. “Meaning is more than a matter of intention; it is also at least sometimes a matter of convention,” says Searle (1969). The theory of implicatures is based on the assumption of such differences between meaning and saying. Peirce’s theory of meaning is pragmatic throughout (cf. Hilpinen 1995: 297). Meaning is the idea which the sign “attaches to its object” (CP 5.6, 1905), but meanings are not a priori inherent in signs; they can only reveal themselves in processes of semiosis (Pape 1996a: 308). Words only have meaning “in so far as we are able to make use of it in communicating our knowledge to others and in getting at the knowledge that these others seek to communicate to us” (CP 8.176, 1903). Thus, no “semantic” or “conventional” meaning can be found in Peirce’s assumptions. In accordance with the pragmatic maxim, the meaning of an utterance reveals itself in its interpretations as well as in the actions and habits which are its consequences. In one of his later definitions, Peirce concludes that “the meaning of any [assertion] is the meaning of the composite of all the propositions which that [assertion] would under all circumstances empower the interpreter to scribe” (MS 280, ca. 1905; cf. Pietarinen 2005: 1769). Since utterances have no inherent meaning independent of their use in speech acts, Peirce does not distinguish between utterance and speaker meaning. Furthermore, he distinguishes between many more kinds of meaning. Some aspects of meaning are associated with the object of the sign insofar as the sign presupposes collateral knowledge of its object and a common ground of knowledge shared by the speaker and the hearer. Other meanings pertain to the interpretant, the effect of meaning created by the sign (cf. Bergman 2000: 234 for both). In the theory of the interpretant discussed above, we find the intentional interpretant as a counterpart to speaker meaning and the effectual interpretant as the counterpart to Grice’s hearer meaning, but Peirce distinguishes other kinds of interpretants (cf. Johansen 2002: 48). His immediate interpretant, the meaning which the sign indicates before it is interpreted, comes close to Searle’s conventional meaning, but the interpretant is not always the meaning of a sign in the linguistic sense. For example Peirce distinguishes between an emotional, and energetic, and a logical interpretant. The first is an emotional, the sec-

Semiotic foundations of pragmatics

189

ond a bodily reaction to a sign. Only the third comes close to the kind of meaning which linguists usually describe since logical meaning is the meaning of a concept but since not all words are concepts, words also have meanings which are not logical interpretants. Peirce also acknowledges the possibility of differences between saying and meaning so much discussed in linguistic pragmatics, but his account of how meaning is conveyed in the dialogue of an utterer and an interpreter is based on other premises. Meaning is not determined by rules, conventions, and speakers’ intentions, as Searle (1969: 43–45) sees it, but by habits, common grounds, and the collateral experience of the object of the sign. Meaning is not only vague by exception, but it is essentially vague (Nöth and Santaella 2009). If meaning is fuzzy in language because of the generality and vagueness of words, it is even vaguer if the feelings, desires, and inner conflicts are considered which speakers and hearers have when they communicate. Emotions, e.g., are never certain and clear; they are necessarily vague, merely expressible and interpretable in the form of allusions, conjectures, surmises, guesses, intuitions, and communicative negotiations. Therefore, “no communication of one person to another can be entirely definite, i.e., nonvague. […] Wherever degree or any other possibility of continuous variation subsists, absolute precision is impossible. Much else must be vague, because no man’s interpretation of words is based on exactly the same experience as any other man’s. Even in our most intellectual conceptions, the more we strive to be precise, the more unattainable precision seems” (CP 5.506, c. 1905). Vagueness and generality are the two major causes of indeterminacy in conversation, but this indeterminacy leads not only to misunderstandings or even failures of communication. Instead, indeterminate signs are also a source of creativity to both utterers and interpreters. The interpreter can and must exert his or her own imagination in determining what the vague or general sign means. He or she can specify what is left undetermined by associating it with meanings from his or her own textual universe. The utterer of a vague sign has the advantage that he or she can suggest meanings without being held responsible for them. Hence vagueness is not only a risk that endangers communication; it also offers a useful potential of creativity to those who communicate. Peirce develops these ideas of the advantages and risks of taking and giving freedom of semantic choice with an example of a generality and one of referential vagueness: A sign […] is objectively general in so far as it extends to the interpreter the privilege of carrying its determination further. Example: “Man is mortal.” To the question, What man? the reply is that the proposition explicitly leaves it to you to apply its assertion to what man or men you will. A sign that is objectively indeterminate in any respect is objectively vague in so far as it reserves further determination to be made in some other conceivable sign, or at least does not appoint the interpreter as its deputy in this office. Example: “A man whom I could mention seems to be a little conceited.” The suggestion here is that the man in view is the person addressed; but the utterer does not authorize

190

Winfried Nöth

such an interpretation or any other application of what she says. She can still say, if she likes, that she does not mean the person addressed. Every utterance naturally leaves the right of further exposition in the utterer; and therefore, in so far as a sign is indeterminate, it is vague, unless it is expressly or by a well-understood convention rendered general. (CP 5.447, 1905)

Elsewhere, Peirce gives indefinite pronouns as the example of the risks and chances which generality and vagueness bring to the semantic choices made by the utterer and the interpreter: “‘Some’ means that the speaker is to select an instance, while ‘Every’ or ‘Any’ means that a second person is to perform the selection. Of course, it is easier to satisfy the conditions of a statement if one can select one’s own examples, except for this, that he who undertakes to find an example guarantees that there is one, while if he leaves the selection to another, and there is none, his statement is not broken down” (CP 2.523, fn 1, 1893).

4.

Semiotics of speech acts: Peircean and Greimasian contributions

The theory of speech acts is the branch of linguistic pragmatics to which semiotics has both contributed as a precursor and to which it has offered alternative approaches. The semiotic precursor of speech act theory is again Peirce; the alternative approach to be presented in the following is an approach based on the semiotics of A. J. Greimas. Only passing reference can be made to the semiotic elements of a pragmatic theory of speech acts which may be found in the typology of discourse of Charles Morris (1946: 203–205; cf. Nöth 2000: 94) and to the rather different elements of a semiotics of some speech acts developed in the framework of M. A. K. Halliday’s systemic semiotics by Ventola (1987). 4.1.

Peirce’s contributions to speech act theory

Peirce’s contributions to speech act theory have been the focus of interest of several studies.16 In wider contexts they have also been discussed by Hilpinen (1995, 1998) and Pietarinen (2004, 2006). Actually, Peirce has even anticipated key terms and issues of speech act theory, among them Austin’s thesis that by speaking we “do” things and that this doing is an “act”. In a fragment of 1908, Peirce describes the speech act of taking an oath in these terms: “Taking an oath is not mainly an event of the nature of a […] Vorstellung, or representing. It is not mere saying, but is doing. The law, I believe, calls it an ‘act’” (CP 5.546). In anticipation of the distinction between an illocutionary act and its propositional content of an utterance later drawn by Searle (1969: 23–29), Peirce defines a proposition as the logical content of a sentence, irrespective of the particular “mood” in which it is uttered (cf. Brock 1981a: 281). Sentences differ from utterances as types differ from tokens:

Semiotic foundations of pragmatics

191

A sentence, in the sense here used, is a single object. Every time it is copied or pronounced, a new sentence is made. But a proposition is not a single thing and cannot properly be said to have any existence. Its mode of being consists in its possibility. A proposition which might be expressed has all the being that belongs to propositions although nobody ever expresses it or thinks it. It is the same proposition every time it is thought, spoken, or written, whether in English, German, Spanish, Tagálog, or how. A proposition consists in a meaning, whether adopted or not, and however expressed. (MS 599: 5–6, ca. 1902)

Among the speech acts which Peirce describes as moods of propositions are assertion, judgment, question, denial, doubt, command, teaching, taking an oath, laying a wager, lying, poetic language, and forms of politeness.17 Expressing a proposition in one of these moods “involves an act, an exertion of energy, and is liable to real consequences, or effects” (CP 5.547, ca. 1908). The “exertion of energy” associated with these moods is evidently equivalent to Searle’s illocutionary force of speech acts. Examples and illustrations of the consequences which such speech acts have for their utterers and interpreter are: “An assertion is an act by which a person makes himself responsible for the truth of a proposition. Nobody ever asserted that the moon is made of green cheese; yet this is a familiar proposition” (MS 599: 5, ca. 1902). The person who took an oath has only made an assertion if the denotation of the proposition is a fact (CP 4.500, ca. 1903): [An oath] would be followed by very real effects, in case the substance of what is asserted should be proved untrue. This ingredient, the assuming of responsibility, which is so prominent in solemn assertion, must be present in every genuine assertion. For clearly, every assertion involves an effort to make the intended interpreter believe what is asserted, to which end a reason for believing it must be furnished. But if a lie would not endanger the esteem in which the utterer was held, nor otherwise be apt to entail such real effects as he would avoid, the interpreter would have no reason to believe the assertion. (CP 5.546, 1908)

By contrast, a conversation solely determined by the conventions of politeness has no consequences as to its truth: “Nobody takes any positive stock in those conventional utterances, such as ‘I am perfectly delighted to see you,’ upon whose falsehood no punishment at all is visited” (CP 5.546, 1908). Among the speech acts whose consequences depend on the truth of the proposition expressed are not only assertions and oaths, but also laying a wager. Peirce describes the consequences of these speech acts as follows: What is the difference between making an assertion and laying a wager? Both are acts whereby the agent deliberately subjects himself to evil consequences if a certain proposition is not true. Only when he offers to bet he hopes the other man will make himself responsible in the same way for the truth of the contrary proposition; while when he makes an assertion he always (or almost always) wishes the man to whom he makes it to be led to do what he does. Accordingly in our vernacular ‘I will bet’ so and so, is the phrase expressive of a private opinion which one does not expect others to share, while ‘You bet’ is a form of assertion intended to cause another to follow suit. (CP 5.31, 1902)

192

Winfried Nöth

Peirce considers not only the conditions of publicly expressed speech acts but also such mental acts, as a judgment or a doubt. “A judgment is a mental act by which one makes a resolution to adhere to a proposition as true, with all its logical consequences” (CP 5.31, 1902). Among the mental acts expressing propositions which are not asserted is the doubt: “I may state it to myself and worry as to whether I shall embrace it or reject it, being dissatisfied with the idea of doing either” (CP 5.31, 1902). Peirce also specifies several of those conditions of speech acts which Searle later called “preparatory conditions” (cf. Brock 1981a: 282). For example, Peirce’s formulation of what Searle calls the “preparatory condition” of a speech act is: “The only way to get a sufficient understanding of a proposition is to imagine one person communicates in some language to a second. The utterer does not, properly speaking, communicate it to the interpreter unless the latter is at the time ignorant (or oblivious) of its truth” (MS 284: 41, 1905; Brock 1981a). The distinction between the illocutionary and the perlocutionary effects of a speech act has almost no counterpart in Peirce’s theory of speech acts. A consequence of the pragmatic maxim is that speech acts do not only sometimes have perlocutionary effects; they have always such effects. The study of these effects is the study of the interpretant, which is a necessary constituent of all signs (see 2.4). 4.2.

Speech acts as narrative scenarios: The Greimasian approach

Elements of the structural semiotics of A. J. Greimas18 are the foundation of a theory of speech acts developed by Sbisà and Cooren.19 Sbisà (1983: 100–101) criticizes Searle’s speech act theory for being based on a causal theory of action (cf. Davidson 1980) which results in a too strictly speaker centered approach to the study of dialogues. According to Sbisà, the causal approach to speech acts means a regrettable “return of the subject, the triumph of the empirical speaker” and the “resuscitation of the use of introspection as a criterion of sense” (1983: 100–101). To liberate pragmatics from its mentalist bias, she proposes a semiotic conception of acts and actions. In accordance with the definition proposed by Greimas and Courtés (1979: 5), Sbisà defines an act as a “fair-être”, a ‘make-be’, i.e., “the producing of a state (or of a change of state)”. She argues that the agent of an act thus defined is not the cause of the speech act but simply its presupposition. Instead of the study of speech acts, Sbisà proposes the study of interactions between speakers and hearers. The focus should be on the hearer as an active participant of the dialogue; it should be taken into consideration that the hearer has the choice “(i) to select an acceptable interpretation of the speech act, and (ii) to either accept the speech act, under such an interpretation, as a successful act, or to completely or partly reject it as more or less inappropriate and ‘unhappy’” (Sbisà and Fabbri 1980: 305). Instead of investigating the utterer’s agency in dialogic interaction, Sbisà (1983: 102, 1987: 259) deals with action as an “ascribed concept”: the interpreter

Semiotic foundations of pragmatics

193

“ascribes a certain change-in-the-world to the responsibility of the [utterer as the] agent”. In this perspective, agency shifts from the utterer to the interpreter, whose decision it is to ascribe intentions and meanings to an utterer or not. With this shift of perspective, the analysis can take into consideration ascriptions of agency to quasi-utterers in Peirce’s sense (see 2.5). According to Sbisà (1987: 259), the utterer of a sign, whose agency is now merely an agency ascribed by the hearer, can thus “be a human individual, an infra-individual instance (e.g., the Unconscious), a super-individual construction (class, party, society, church, and the like), and even a personified natural agent.” Hence, in a semiotic speech act theory adopting these premises, not only speakers are agents, but even “beings other than humans can be said to do things with words” (Cooren 2006: 8). Two common examples of assertive and commissive speech acts in which agency is ascribed to nonhuman agents are: “This decision contests his result” and “This signature commits you to payment” (Cooren 2006: 17). Based on her assumption that acts result in transformations of states, Sbisà (1985: 529) arrives at a redefinition of basic concepts of speech act theory: (1) “A speech act is the production of an utterance insofar as it brings about a change for which the utterer is taken to be responsible […].” (2) “An illocutionary act is the production of a change of context at the conventional level of the addresser’s and the addressee’s modal competences.” (3) “A perlocutionary act is the production of a change of context at the material (psychological or praxeological) level of the addressee’s reactions.” An example of her Greimasian approach to the illocutionary force of speech acts is Sbisà’s (1985: 530) study of the discursive figure of manipulation, which she defines as the agency of a manipulating addresser who puts the manipulated addressee into a position of “not being able not to do”. The application of Greimas’s theory of modalities leads Sbisà (1984: 100) to a typology of modal changes of the speakers’ and hearers’ states in speech acts, e.g.: in exercitives (such as granting or permitting), the speaker addresses the addressee in the deontic mood of ‘being allowed’ (can), and the effect in the addressee is one of the modality of obligation (ought); in commissives, these two modal ascriptions are inverse. Taking up Sbisà’s and Greimas’s elements of a semiotic pragmatics, Cooren (2000: 81–169) proposes a model based on the premise that speech acts involve a narrative scenario typically involving three actants: the speaker as the agent, the hearer as the recipient, and the utterance as the object of transfer (Cooren 2000: 86). Uttering is seen as an act of production in which “the object produced by the agent”, i.e., the utterance, “is irremediably severed or dissociated from its origin” so that the producer “no longer has control over the object produced” and acquires an agency of its own (Cooren 2000: 82–83). A speech act is always embedded in a plurality of verbal and nonverbal texts; it does not only refer to, it also produces a situation. Continuing to act in absence of its utterer, the speech act becomes “the product of multiple texts that ‘speak on behalf’ of their producer, creating one or several different situations” (Cooren 2000: 87).

194

Winfried Nöth

Following Greimas’s discourse grammar, which describes narrative roles by means of the actantial model of verb valency of Tesnière (1959), Cooren develops a typology of illocutionary acts based on verb valencies. According to this model, assertives are bivalent and performatives are trivalent speech acts (2000: 89). The latter involve an agent, an object, and a recipient. The tetravalent actantial model of the factive verbs, which involve four actants, then “allows opening up the traditional speaker/hearer schema by acknowledging action from a distance, which could be called ‘tele-action’” (Cooren 2006: 2). The utterance “John asked Mary to send his report to Bill” is an example of a tetravalent teleactive speech act. To telecommunicate, thus means to delegate a speech act to another agent who communicates it on behalf of the delegating agent: one discursive agent mobilizes “other agents that serve as representative or delegates” of the delegating agent (Cooren 2000: 6). In view of the fundamental differences between Peirce’s pragmaticist semiotics and Greimas’s structural discourse semiotics, it is remarkable that both schools of semiotics, for very different reasons, agree in their criticism of the speaker-centered approaches to speech acts. It is equally remarkable that, quite independently of each other, both schools of semiotics postulate models of semiotic agency in verbal interaction radically distinct from those proposed by the classics of linguistic pragmatics. Nevertheless, the fundamental difference between the two semiotic approaches in this latter respect must not be ignored. Whereas Greimas’s model attributes agency to actants in discourse, Peirce attributes agency and autonomy to the sign itself.

Notes 1. Cf. Posner (1991) and Nöth (2000: 433–438). 2. See Parret (1983), Posner (1997), Nadin (1987, 1993), Deledalle, ed. (1989), Driel, ed. (1991), Hilpinen (1995, 1998), Pankow (1995), Pietarinen (2004, 2005, 2006), Rellstab (2008), and Andersen (2009). 3. Cf. Liszka (1996: 3–15), Pietarinen (2004), and Houser (2009: 91). 4. CP refers to Peirce (1931–58). 5. Cf. Fisch (1986: 339), Liszka (1996: 109), and Bergman (2007). 6. EP 2 refers to Peirce (1998). 7. Oehler (1990, 1995), Colapietro (1995), Bergman (2000, 2004), Santaella and Nöth (2004), and Pietarinen (2006). 8. See also Prieto (1966, 1975; cf. Nöth 2000: 228). 9. See 2.2 and Burkhardt (1990). 10. MS refers to Peirce (1979). 11. Cf. Allwood (1978: 4) and Levinson (1983: 103). 12. For Peirce’s distinction between presciss (from prescind) and precise see Spinks (1991: 19–23) and CP 5.449, 1905. Roughly, prescission is giving “attention to one element and neglect of the other” (CP 1.549, 1867).

Semiotic foundations of pragmatics 13. 14. 15. 16. 17. 18. 19.

195

W 1 refers to Peirce (1981). Cf. Oehler (1995: 268) and Johansen (1993). Cf. Santaella (1988, 1990, 1994), Joswick (1996), and Pape (1996b). Brock (1975, 1981a, b), Martens (1981), and Thibaud (1997). Cf. Brock (1981b) and Hilpinen (1998). For Greimas, see Nöth (2000: 112–19). Sbisà and Fabbri (1980), Sbisà (1983, 1984, 1985, 1987, 1994, 2002), and Cooren (2000, 2006).

References Allwood, Jens 1978 On the Analysis of Communicative Action (Gothenburg Papers in Theoretical Linguistics 38). Gothenburg: University, Department of Linguistics. Andersen, Christiane 2009 Semiotics. In: Jef Verschueren and Jan-Ola Östman (eds.), Key Notions for Pragmatics, 217–228. Amsterdam: Benjamins. Austin, John 1962 How to Do Things with Words. London: Oxford University Press. Bar-Hillel, Yehoshua 1954 Indexical expressions. Mind 63: 359–74. Bergman, Mats 2000 Reflections on the role of the communicative sign in semeiotic. Transactions of the Charles S. Peirce Society 36.2: 225–254. Bergman, Mats 2003 Peirce’s derivations of the interpretant. Semiotica 144: 1–17. Bergman, Mats 2004 Fields of Signification. Explorations in Charles S. Peirce’s Theory of Signs. (Philosophical Studies from the University of Helsinki 6), Helsinki, PhD dissertation. Bergman, Mats 2007 The secret of rendering signs effective: the import of C. S. Peirce’s semiotic rhetoric. The Public Journal of Semiotics 1.2: 2–11. Brock, Jarrett 1975 Peirce’s conception of semiotics. Semiotica 14: 124–141. Brock, Jarrett 1979 Principal themes in Peirce’s logic of vagueness. Peirce Studies (Lubbock, Texas) 1: 41–50. Brock, Jarrett 1981a An introduction to Peirce’s theory of speech acts. Transactions of the Charles S. Peirce Society 17: 319–326. Brock, Jarrett 1981b Peirce and Searle on assertion. In: Kenneth Ketner, Joseph M. Ransdell, Carolyn Eisele, Max H. Fisch and Charles S. Hardwick (eds.), Proceedings of the C.S. Peirce Bicentennial International Congress. 281–287. Lubbock, TX: Texas Tech Press (Graduate Studies Texas Tech University 23).

196

Winfried Nöth

Bühler, Karl 1933 Die Axiomatik der Sprachwissenschaft. Frankfurt am Main: Klostermann. Burkhardt, Armin, (ed.) 1990 Speech Acts, Meaning and Intentions. Berlin: Walter de Gruyter. Buyssens, Eric 1943 La communication et l’articulation linguistique. Bruxelles / Paris: Presses Universitaires (revised edition 1964). Carnap, Rudolf 1975 Introduction to Semantics and Formalization of Logic, 2 vols. in 1. Cambridge: Harvard University Press [1942/43]. Colapietro, Vincent 1992 Purpose, power, and agency. The Monist 75: 423–444. Colapietro, Vincent 1995 Immediacy, opposition, and mediation: Peirce on irreducible aspects of the communicative process. In: Leonore Langsdorf and Andrew R. Smith (eds.), Recovering Pragmatism’s Voice: The Classical Tradition, Rorty, and the Philosophy of Communication, 23–48. Albany, NY: State University Press. Cooren, François 2000 The Organizing Property of Communication. Amsterdam: Benjamins. Cooren, François 2006 Between semiotics and pragmatics: Opening language studies to textual agency. Paper presented at the annual meeting of the International Communication Association, Dresden. . Davidson, Donald 1980 Essays on Actions and Events. Oxford: Oxford University Press. Deacon, Terrence 1997 The Symbolic Species. New York, NY: Norton. Deledalle, Gérard (ed.) 1989 Semiotics and Pragmatics. Amsterdam: Benjamins. Derrida, Jacques 1988 Limited Inc. Evanston, IL: Northwestern University Press. Driel, Herbert van (ed.) 1991 Het semiotisch pragmatisme van Charles S. Peirce. Amsterdam: Benjamins. Eisele, Carolyn 1987 Peirce’s pragmatics. In: Herbert Stachowiak (ed.), Der Aufstieg des pragmatischen Denkens im 19. und 20. Jahrhundert (Pragmatik: Handbuch des pragmatischen Denkens, vol. 2.), 83–97 Hamburg: Meiner. Fisch, Max H. 1986 Peirce, Semeiotic, and Pragmatism. Ed. Kenneth L. Ketner and Christian J. W. Kloesel. Bloomington: Indiana University Press. Gorlé, Dinda 1993 Evolving through time: Peirce’s pragmatic maxims. Semiosis 71/72: 3–13. Goudge, Thomas A. 1965 Peirce’s index. Transactions of the Charles S. Peirce Society 1: 52–70.

Semiotic foundations of pragmatics

197

Greimas, Algirdas J. and Joseph Courtés 1979 Sémiotique: Dictionnaire raisonné de la théorie du langage. Paris: Hachette. Trad. 1982. Semiotics and Language: An Analytical Dictionary. Bloomington, IN: Indiana University Press. Grice, Paul 1957 Meaning. Philosophical Review 64: 377–388. Grice, Paul 1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press. Habermas, Jürgen 1981 Theorie des kommunikativen Handelns, 2 vols. Frankfurt am Main: Suhrkamp. Habermas, Jürgen 1995 Peirce and communication. In: Kenneth Laine Ketner (ed.), Peirce and Contemporary Thought. Philosophical Inquiries, 243–266. New York: Fordham University Press. Hervey, Sándor G. J. 1982 Semiotic Perspectives. London: Allen and Unwin. Hilpinen, Risto 1995 Peirce on language and reference. In: Kenneth Laine Ketner (ed.), Peirce and Contemporary Thought. Philosophical Inquiries, 272–303. New York: Fordham University Press. Hilpinen, Risto 1998 Pragmatics and pragmatism: On C.S. Peirce’s pragmatic theory of meaning. In: Paul Weingartner, Gerhard Schurz, and Georg Dorn (eds.), Die Rolle der Pragmatik in der Gegenwartsphilosophie – The Role of Pragmatism in Philosophy, 151–164 (=Akten des 20. Internationalen Wittgenstein-Symposiums. Schriftenreihe der Wittgenstein-Gesellschaft, vol. 26) Wien: Hölder-Pichler-Tempsky. Houser, Nathan 1998 Introduction. In: Nathan Houser and Christian J. W. Kloesel (eds.), The Essential Peirce. Vol. 1, xvii-xxxviii. Bloomington: Indiana University Press. Houser, Nathan 2009 Peirce, phenomenology and semiotics. In: Paul Cobley (ed.), The Routledge Companion to Semiotics, 89–100. London: Routledge. Jakobson, Roman 1960 Linguistics and poetics. In: Thomas A. Sebeok (ed.), Style in Language, 350–377. Cambridge, MA: MIT Press. Johansen, Jørgen Dines 1993 Dialogic Semiosis. Bloomington: Indiana University Press. Johansen, Jørgen Dines 2002 Literary Discourse: A Semiotic-Pragmatic Approach to Literature. Toronto: University of Toronto Press. Joswick, Hugh 1996 The object of semeiotic. In: Vincent M. Colapietro and Thomas M. Olshewsky (eds.), Peirce’s Doctrine of Signs, 93–102. Berlin: de Gruyter. Kappner, Stefan 2004 Intentionalität aus semiotischer Sicht: Peirceanische Perspektiven. Berlin: de Gruyter.

198

Winfried Nöth

Lakoff, George and Mark Johnson 1980 Metaphors We Live By. Chicago: University Press. Leech, Geoffrey N. 1983 Principles of Pragmatics. London: Longman. Levinson, Stephen C. 1983 Pragmatics. Cambridge: University Press. Liszka, J. Jakób 1996 A General Introduction to the Semeiotic of Charles S. Peirce. Bloomington: Indiana University Press. Liszka, J. Jakób 2009 Rethinking the pragmatic theory of meaning. Cognitio (São Paulo) 10.1: 61–80. Luhmann, Niklas 1984 Soziale Systeme. Frankfurt am Main: Suhrkamp. Martens, Ekkehard 1981 C. S. Peirce on speech acts. In: Kenneth Ketner, Joseph M. Ransdell, Carolyn Eisele, Max H. Fisch and Charles S. Hardwick (eds.), Proceedings of the C.S. Peirce Bicentennial International Congress, 289–292. Lubbock, TX: Texas Tech Press (=Graduate Studies Texas Tech University 23). Montague, Richard 1970 Pragmatics and intensional logic. Synthese 22: 68–94. Morris, Charles W. 1938 Foundations of the Theory of Signs (Foundations of the Unity of Science: Towards an International Encyclopedia of Unified Science, vol. 1.2.) Chicago, IL: University Press [1970]. Morris, Charles W. 1946 Signs, Language, and Behavior. New York: Braziller. (Quoted from the reprint in C. W. Morris. 1971. 73–398.) Morris, Charles W. 1964 Signification and Significance: A Study of the Relations of Signs and Values. Cambridge, MA: MIT Press. Morris, Charles W. 1971 Writings on the General Theory of Signs. The Hague: Mouton. Murphy, John P. 1990 Pragmatism from Peirce to Davidson. Boulder, CO: Westview Press. Nadin, Mihai 1987 Pragmatics in the semiotic frame. In: Herbert Stachowiak (ed.), Der Aufstieg des pragmatischen Denkens im 19. und 20. Jahrhundert, 148–170. (Pragmatik: Handbuch des pragmatischen Denkens, vol. 2.) Hamburg: Meiner. Nadin, Mihai 1993 Semiotics in action: The pragmatic level. In: Herbert Stachowiak (ed.), Spachphilosophie, Sprachpragmatik und formative Pragmatik, 219–250. (Pragmatik: Handbuch des pragmatischen Denkens, vol. 4.) Hamburg: Meiner. Nöth, Winfried 2000 Handbuch der Semiotik, 2nd rev. ed. Stuttgart: Metzler. Nöth, Winfried 2002a Charles Sanders Peirce, pathfinder in linguistics. Interdisciplinary Journal for Germanic Linguistics and Semiotic Analysis 7.1: 1–14.

Semiotic foundations of pragmatics

199

Nöth, Winfried 2002b Wörter als Zeichen: Einige semiotische Aspekte der Sprache. In: Jürgen Dittmann and Claudia Schmidt (eds.), Über Wörter, 9–32. Freiburg: Rombach. Nöth, Winfried 2006 Representation and reference according to Peirce. Caderno. 9a Jornada do Centro de Estudos Peirceanos / IIIrd Advanced Seminar on Peirce’s Philosophy and Semiotics, 71–83. (São Paulo, PUC-SP), São Paulo. Nöth, Winfried 2008 Natural signs and the natural ingredients of cultural signs. In: Martin Döring, Hermine Penz and Wilhelm Trampe (eds.), Language, Signs and Nature: Ecolinguistic Dimensions of Environmental Discourse, 21–37. Tübingen: Stauffenburg. Nöth, Winfried 2009 On the instrumentality and semiotic agency of signs, tools, and intelligent machines. Cybernetics and Human Knowing 16.1–2: 11–36. Nöth, Winfried and Lucia Santaella 2009 Die Relevanz der Peirceschen Semiotik des Vagen für die Theorie der Kommunikation. In: Siegfried J. Schmidt (ed.), Die (Un)möglichkeit der Kommunikation. Frankfurt am Main: Peter Lang (Special Issue SPIEL: Siegener Periodicum zur Internationalen Empirischen Literaturwissenschaft 26.1 [2007], 73–96). Oehler, Klaus 1990 Peirce oder Habermas? In: Elisabeth Walther and Udo Bayer (eds.), zeichen von zeichen für zeichen: festschrift für max bense, 42–47. Baden-Baden: Agis. Oehler, Klaus 1995 A reponse to Habermas. In: Kenneth Laine Ketner (ed.), Peirce and Contemporary Thought, 267–271. New York: Fordham University Press. Pankow, Christiane 1995 Semiotics. In: Jef Verschueren, Jan-Ola Östman and Jan Blommaert (eds.), Handbook of Pragmatics Manual, 469–476. Amsterdam: Benjamins. Pape, Helmut 1996a Charles Sanders Peirce. In: Tilman Borsche (ed.), Klassiker der Sprachphilosophie, 307–324. München: Beck. Pape, Helmut 1996b Object and final cause in Peirce’s semeiotic. In: Vincent M. Colapietro and Thomas M. Olshewsky (eds.), Peirce’s Doctrine of Signs, 103–117. Berlin: de Gruyter. Parmentier, Richard J. 1985 Signs’ place in medias res: Peirce’s concept of semiotic mediation. In: Elizabeth Mertz and Richard J. Parmentier (eds.), Semiotic Mediation, 23–48. Orlando, FL: Academic Press. Parret, Herman 1983 Semiotics and Pragmatics. Amsterdam: Benjamins. Parret, Herman (ed.), 1994 Pretending to Communicate. Berlin: de Gruyter.

200

Winfried Nöth

Peirce, Charles Sanders 1931–58 Collected Papers, Vols. 1–6. ed. Charles Hartshorne and Paul Weiss, Vols. 7–8, ed. Arthur W. Burks. Cambridge, MA: Harvard University Press (quoted as CP). Peirce, Charles Sanders 1979 The Charles S. Peirce Papers, 30 reels, 3rd microfilm edition. Cambridge, MA: The Houghton Library, Harvard University, Microreproduction Service (quoted as MS). Peirce, Charles Sanders 1981 Writings of Charles S. Peirce. Vol. 1: 1857–1866. Bloomington, IN: Indiana Univ. Press (quoted as W 1). Peirce, Charles Sanders 1998 The Essential Peirce. Selected Philosophical Writings. Vol. 2, 1893–1913). Ed. by the Peirce Edition Project. Bloomington, IN: Indiana University Press (quoted as EP 2). Pietarinen, Ahti-Veikko 2004 Grice in the wake of Peirce. Pragmatics and Cognition 12.2: 295–315. Pietarinen, Ahti-Veikko 2005 Relevance theory through pragmatic theories of meaning. In: Bruno G. Bara, Lawrence Barsalou and Monica Bucciarelli (eds.), Proceedings of the XXVII Annual Conference of the Cognitive Science Society, 1767–1772. Mahwah, NJ: Lawrence Erlbaum, Pietarinen, Ahti-Veikko 2006 Signs of Logic. Peircean Themes on the Philosophy of Language, Games, and Communication. Dordrecht: Springer. Plato 1953 The Dialogues of Plato, transl. Benjamin Jowett, 4th ed. (1st ed. 1871). Oxford: Clarendon. Also: http://www.gutenberg.org/dirs/1/6/1/1616/1616.txt. Posner, Roland 1985 La syntactique, sa relation à la sémantique et la pragmatique, à la morphologie et la syntaxe, et à la syntagmatique et la paradigmatique. In: Herman Parret and Hans-Georg Ruprecht (eds.), Exigences et perspectives de la sémiotique. Recueil d’hommages pour A.J. Greimas, vol. 1, 73–96. Amsterdam: Benjamins. Posner, Roland 1991 Research in pragmatics after Morris. Dedalus (Lisbon) 1: 115–156. Prieto, Luis J. 1966 Messages et signaux. Paris: Presses Universitaires. Prieto, Luis J. 1975 Etudes de linguistique et de sémiologie générales. Genève: Droz. Ransdell, Joseph 1977 Some leading ideas of Peirce’s semiotic. Semiotica 19: 157–178. Rellstab, Daniel 2008 Peirce for linguistic pragmatics. Transactions of the Charles Sanders Peirce Society 44.2: 312–346. Santaella, Lucia 1988 Charles Sanders Peirce’s object (of the sign). Versus 49: 53–58.

Semiotic foundations of pragmatics

201

Santaella, Lucia 1990 An introduction to Peirce’s concept of object. Cruzeiro semiótico 13: 75–82. Santaella, Lucia 1994 Why things are objects of the sign. Cruzeiro semiótico 18: 37–46. Santaella, Lucia 1999 Methodeutics, the liveliest branch of semiotics. Semiotica 124: 377–395. Santaella, Lucia and Winfried Nöth 2004 Comunicação e semiótica. São Paulo: Hacker. Savan, David 1994 C. S. Peirce and American semiotics. In: Michael Shapiro and Michael C. Haley (eds.), The Peirce Seminar Papers 2, 179–208. Providence, RI: Berghahn. Sbisà, Marina 1983 Actes de langage et (acte d’)énonciation. Langages 18.70: 99–106. Sbisà, Marina 1984 On illocutionary types. Journal of Pragmatics 8: 93–112. Sbisà, Marina 1985 Manipulation et sanction dans la dynamique des actes de langage. In: Herman Parret and Hans-Georg Ruprecht (eds.), Exigences et perspectives de la sémiotique. Recueil d’hommages pour A.J. Greimas, Vol. 1, 529–538. Amsterdam: Benjamins. Sbisà, Marina 1987 Speech acts and context change. In: Thomas T. Ballmer and Wolfgang Wildgen (eds.), Process Linguistics, 254–280. Tübingen: Niemeyer. Sbisà, Marina 1994 Language and dialogue in the framework of the analytical philosophy of ordinary language. In: Edda Weigand (ed.), Concepts of Dialogue, 159–170. Tübingen: Niemeyer. Sbisà, Marina 2002 Cognition and narrativity in speech act sequences. In: Anita Fetzer and Christiane Meierkord (eds.), Rethinking Sequentiality, 71–97. Amsterdam: Benjamins. Sbisà, Marina and Paolo Fabbri 1980 Models ( ? ) for a pragmatic analysis. Journal of Pragmatics 4: 301–319. Searle, John R. 1969 Speech Acts: An Essay in the Philosophy of Language. Cambridge: University Press. Searle, John R. 1983 Intentionality. Cambridge: University Press. Shannon, Claude E. and Warren Weaver 1949 The Mathematical Theory of Communication. Urbana: University of Illinois Press. Short, Thomas L. 2007 Peirce’s Theory of Signs. Cambridge: University Press. Sperber, Dan and Deirdre Wilson 1986 Relevance: Communication and Cognition. Oxford: Blackwell.

202

Winfried Nöth

Spinks, Cary W. 1991 Peirce and Triadomania. Berlin: de Gruyter. Stetter, Christian 1989 Some remarks concerning the theory of speech acts and the idea of “pragmatics”. In: Gerard Deledalle (ed.), Semiotics and Pragmatics, 155–165. Amsterdam: Benjamins. Strawson, Peter 1971 Intention and convention in speech acts. In: John R. Searle (ed.), The Philosophy of Language, 23–38. Oxford: University Press. Tesnière, Lucien 1959 Eléments de syntaxe structurale. Paris: Klincksieck. Thibaud, Pierre 1997 Between saying and doing: Peirce’s propositional space. Transactions of the Charles Sanders Peirce Society 33: 270–327. Ventola, Eija 1987 The Structure of Social Interaction: A Systemic Approach to the Semiotics of Service Encounters. London: Frances Pinter. Watzlawick, Paul, Janet H. Beavin, and Don D. Jackson 1967 Pragmatics of Human Communication. New York: Norton. Wittgenstein, Ludwig 1953 Philosophische Untersuchungen – Philosophical Investigations. Oxford: Blackwell.

7.

Pragmatics in modern philosophy of language Nikola Kompa and Georg Meggle

Introduction According to what might be called the ‘traditional paradigm’ in the philosophy of language, linguistic interpretation is a linear and serial process. In particular, pragmatic interpretation (such as the working out of Gricean implicatures) is claimed to commence only after the semantic interpretation of a given sentence has been worked out. This traditional paradigm has come under attack from various directions. Some philosophers contend, for instance, that there is no such thing as a purely semantic content or a semantically expressed proposition (or truth-condition) but that pragmatic inferences are necessary at every level of linguistic interpretation: this claim will be discussed in the first part (chapter 1) of this article. One may even go one step further and maintain that semantic notions need to be spelled out pragmatically in the first place, that linguistic interpretation requires comprehension of a particular type of action and that semantics is thus simply a part of pragmatics: this will be explored in the second part (chapter 2) of our article.

1.

Pragmatic inferences and semantic rules

1.1.

Two approaches to language and meaning

In 1938, Charles Morris distinguished three different “dimensions of semiosis” in his theory of signs:1 One may study the relations of signs to the objects to which the signs are applicable. This relation will be called the semantical dimension of semiosis, symbolized by the sign ‘DSEM’; the study of this dimension will be called semantics. Or the subject of study may be the relation of signs to interpreters. This relation will be called the pragmatical dimension of semiosis, symbolized by the sign ‘DP’; the study of this dimension will be named pragmatics. One important relation of signs has not yet been introduced: the formal relations of signs to one another. […] This third dimension will be called the syntactical dimension of semiosis, symbolized by the sign ‘DSYN’, and the study of this dimension will be named syntactics. (Morris 1938: 21–22)

Syntax studies the relations of signs to other signs, semantics is concerned with the relation between signs and objects, and pragmatics takes the relation between signs and their interpreters as its subject matter. To partition the domain this way has been common ever since. Morris’s characterization, however, leaves the three areas severely underdetermined. For example, pragmatics is commonly supposed

204

Nikola Kompa and Georg Meggle

to cover such diverse phenomena as speech acts, deixis, presuppositions or implicatures, their unifying feature being only that they have something to do with the relation between signs and interpreters (and that they cannot easily be accommodated within a semantic theory). Also, the characterization may be not as clear-cut as Morris and others hoped it to be. In particular (and as a consequence), where exactly to draw the line between semantics and pragmatics is a hotly contested issue in current philosophical and linguistic debate.2 Moreover, the difficulties one encounters when trying to sharpen the distinction between semantics and pragmatics point to a deeper problem about how to theoretically approach language in general. It is common in the philosophy of language to contrast two different approaches to questions concerning language, meaning and interpretation. Robert Brandom, for example, puts the difference thus: some philosophers view language as a kind of practice or activity, others prefer a semantic Tarskian approach to language (cf. Brandom 2002: 40–41). The former focus on how language is used; they study language as a kind of practice, as something speakers do in order to achieve certain aims. The latter, on the other hand, strive to develop a theory of content that is closely tied to Alfred Tarski’s truth-theoretical considerations. Underlying the two different approaches to language are different views of what a language is, which David Lewis aptly characterizes thus: What is a language? Something which assigns meanings to certain strings of types of sounds or of marks. It could therefore be a function, a set of ordered pairs of strings and meanings. […] (Lewis 1975/1983: 163)

Those who favor this view of what a language is think of language as an abstract system that systematically pairs certain strings of sounds or marks with certain other entities, called meanings. The mapping can be done, it seems, without taking language users into account. In the contrasting view, on the other hand, the language user takes center stage: What is a language? A social phenomenon which is part of the natural history of human beings; a sphere of human action, wherein people utter strings of vocal sounds, or inscribe strings of marks, and wherein people respond by thought or action to the sounds or marks which they observe to have been so produced. (Lewis 1975/1983: 164)

Which of these two views of language will prove theoretically fruitful is the point at issue between semantically and pragmatically oriented accounts of language and linguistic interpretation. Very roughly, the “pragmaticist” will advocate a usagebased approach to language. S/he will, accordingly, think of language as a human activity, a social phenomenon that helps us interact and communicate with our fellow human beings. His/her opponent, the “semanticist”, takes a language to be a formal system that can be studied fairly independently of how human beings actually use language. His/her aim is to come up with a formal semantic theory that allows us to assign every sentence of language L its meaning in L. More specifically, s/he will try to pair syntactically analyzed sentences (of a language) with

Pragmatics in modern philosophy of language

205

their truth-conditions, as s/he typically takes the meaning of a sentence to be its truth-condition (cf., e.g., Portner and Partee 2002: 3). Semantic interpretation was commonly taken to be independent of pragmatics. Some even claimed that semantics comes first (cf. Manor 2001: 67f). Pragmatic interpretation (such as the derivation of Gricean implicatures) comes in only after the semantic interpretation of the expressions in question has been worked out. But new work on the context sensitivity of natural languages questions not only the ‘semantics first’-view but the idea of ‘semantic independence’ as such. More specifically, in recent years the project of a formal semantic theory has come under attack from what came to be called contextualist approaches to language and meaning. Very roughly, the ‘contextualist’ can be characterized as someone who takes context sensitivity to be a ubiquitous feature of natural languages, a feature much more pervasive than deixis or anaphora and much less well-behaved than implicatures, for example. S/he favors a pragmatically oriented, usage-based approach to meaning, arguing that linguistic (even truth-conditional) interpretation involves pragmatic inferences all the way down. Her/his opponent, on the other hand, advocates a formal semantic approach to meaning according to which truth-conditional interpretation is a purely syntactic-semantic endeavor. S/he would, presumably, agree with the following characterization: […] in general, formal theories can be characterized as fundamentally syntax-driven theories, which claim that it is possible to deliver an account of the propositional or truth-conditional content of a sentence in natural language simply via formal operations over the syntactic features of that sentence, that is, over the lexical items it contains and their mode of composition. (Borg 2004b: 3)

The pragmatically oriented contextualist and his/her semantically oriented opponent negotiate what was traditionally taken to be key semantic notions: the notion of what is said in the utterance of a sentence, the notion of the proposition expressed by the utterance, the notion of a sentence’s truth-conditions, and – at the most basic level – the notion of meaning itself. Semanticists try to save one or another of these notions from pragmatic intrusion. 1.2.

Formal semantics

Let us first try to get a better grasp of what form a formal semantic theory might take before we consider, in a second step, the objections the contextualist raises against this kind of approach. Very roughly, formal semantic theories try to develop, for as large a fragment of natural language as possible, a formal representation that contains all syntactically and semantically relevant information. In a first step, the logical form of each sentence in the relevant fragment of the language in question has to be worked out. In a second step, the elements of logical form have to be interpreted, i.e., each element has to be assigned a semantic value in ac-

206

Nikola Kompa and Georg Meggle

cordance with the respective interpretation rules. More specifically, the idea is that there are interpretation rules that assign semantic values to the basic expressions in the language. Composition rules then tell us which value to assign to complex expressions, the semantic value of a complex expression being a function of the semantic values of its parts. In this manner we work our way up until, after finitely many steps, we are able to assign semantic values (i.e., truth-conditions) to whole sentences. Compositionality obviously plays an important role here, for if language were not compositional, no recursive procedure for assigning semantic values to expressions of arbitrary complexity would be available. Yet such a recursive procedure is exactly what the semanticist is looking for: a procedure that yields for every well-formed sentence (of finite length) its meaning, i.e., the proposition expressed or its truth-conditions, after finitely many steps. And to interpret an expression is to follow such a recursive procedure. Here is a somewhat lengthy quote from Jason Stanley, a (moderate) proponent of formal semantics, to illustrate the point: Successful interpretation involves assigning denotations to the constituents of the logical form, and combining them in accord with composition rules that do not vary with extra-linguistic context. The denotations that successful interpreters will assign to constituents of a logical form will be constrained by the linguistic conventions governing those elements. […] What results from a successful application of this first stage of interpretation is a unique proposition, a fully truth-evaluable entity. (Stanley 2002: 149–150)

One of the attractive features of a formal semantic approach is that it seems to provide a good explanation of the productivity of language. Obviously, competent speakers are able to produce novel sentences and to understand sentences they have not encountered before. This fact stands in need of explanation. The formal semanticist seems able to provide an explanation by pointing out the recursive character of linguistic interpretation. Yet what semantic value do interpretation rules assign to the basic expressions of a language? What is the semantic value, i.e. the meaning, of a linguistic expression? It is the answer to this question on which philosophical theories of meaning differ. According to the reference theory of meaning it is the expression’s reference, often called its extension: what the expression is used to talk about or refer to. Unfortunately, the reference theory of meaning faces a couple of serious and well-known objections, some of which already attracted Gottlob Frege’s attention and brought him to distinguish between an expression’s reference (what he called “Bedeutung”) and its sense (“Sinn”, cf. Frege 1891, 1892a and 1892b). Frege’s insights, in turn, loomed large in the development of so called intensional semantics. Rudolf Carnap, who attended some of Frege’s lectures in Jena, strived to “find a suitable method for the semantic analysis of meaning” (Carnap 21956: 2) and proposed the concept of extensions and intentions of an expression as a suitable tool. This distinction between an expression’s extension and intension soon caught on.

Pragmatics in modern philosophy of language

207

Very roughly, the idea is this: It may not be adequate to equate an expression’s meaning with its actual extension because the latter may depend on the way the world is, which may be subject to change. Yet isn’t the task of an expression’s meaning to pursue this kind of dependence? In other words, isn’t it the case that meaning determines the expression’s extension in any possible world? Such functions from possible worlds to extensions are called intensions. The intension of an expression is supposed to determine its extension (i.e. its truth-value, in the case of whole sentences) in any possible world. In other words, intensions take possible worlds as arguments and provide extensions as truth-values. Hence, intensions seem to be better candidates than extensions for what we intuitively think of as an expression’s meaning. One of the most influential later developments in intensional semantics is Montague Grammar. It was originally developed by Robert Montague who elaborated on ideas introduced by Rudolf Carnap, Alfred Tarski and Alonzo Curch and subsequently modified and expanded by other logicians, philosophers and linguists (cf. Montague 1970a, 1970b, 1973, and also Partee 1973, 1975; Porter and Partee 2002). In the course of these considerations it became more and more obvious that the extension of an expression not only depends on the way the world is but also, in some cases at least, on quite specific features of the context in which the expression is used. The most prominent examples are indexical expressions such as “I”, “here” or “now”. What these expressions refer to depends on the context of utterance; similarly for demonstrative pronouns such as “this” or “that”. Insight into this double-dependence of extension on context and world respectively has led to the development of what came to be called two-dimensional semantics. According to a renowned view held by David Kaplan (1989), for example, the overall import of a linguistic expression breaks down into two components: its character and its content. Both can be construed as functions, the character of an expression being a function from the context of utterance into the expression’s content, the content being a function from possible circumstances of evaluation (very roughly, something like a possible world or an index, terminology differs here) into the extension of the expression in the respective circumstances.3 This opens up a new way of accommodating context sensitive expressions within a formal semantic theory: just associate with each context sensitive expression a function or rule (a character) that determines, for any possible context of its use, the expression’s content (its extension, in the case of indexical or demonstrative expressions) in that context. Thought of in this way, linguistic (truth-conditional) interpretation is (context-sensitivity notwithstanding) still a purely syntactic-semantic enterprise. No pragmatic processes seem to be required; a result that nicely fits in with the formal semantic project as sketched above.

208 1.3.

Nikola Kompa and Georg Meggle

Context-sensitivity and pragmatic inferences

Still, as mentioned above, the formal semantic project came under attack when it became obvious that there is much more context-sensitivity in natural language than the formal semanticist is willing to acknowledge. Moreover, it is far from clear that she could accommodate all the context-sensitivity within her formal semantic approach even if she wanted to. The main problem is that pragmatic inferences seem to be required after all in order to truth-conditionally interpret utterances containing context-sensitive expressions and to assign these expressions a ‘semantic’ value in context. Yet the nature of pragmatic inferences does not square well with the allegedly deductive, algorithmic nature of semantic interpretation (cf. Borg 2004a: 218–219). 1.3.1.

The non-monotonic nature of pragmatic inferences

The need for pragmatic inferences in linguistic interpretation has long been acknowledged in philosophical and linguistic theorizing. Ever since Paul Grice developed his theory of conventional and conversational implicatures the need for pragmatic inferences had become obvious. According to Grice (1989), what a speaker means by an utterance divides up into what she says and what she implicates. What she says is (subject to certain qualifications) closely tied to the conventional meaning of the words used. The implicatures (at least the conversational ones), on the other hand, have to be pragmatically worked out. Usually, in generating an implicature, a speaker exploits certain maxims her hearers expect her to comply with, maxims to the effect that she should try to be relevant, succinct, truthful and informative. By violating or flouting one of these maxims, a speaker can communicate something other or even more than what she explicitly said. To use one of Grice’s own examples, if you ask me whether our common friend Smith has a girlfriend at the moment, and I reply by saying that he has been paying a lot of visits to Zurich lately, you might reason as follows. “What she said was not directly relevant for my question. But then, I have no reasons to assume that she is not cooperative. So she must have meant to communicate something other than what she said. What could that be? Maybe she meant to suggest that Smith has a girlfriend in Zurich.” The conclusion is defeasible, though. It might need revision in the light of new information. And this line of reasoning is an example of a pragmatic inference. Since Grice, it has been common to think of pragmatic inferences as being restricted to cases such as these. More specifically, pragmatic inferences are commonly taken to come into play in linguistic interpretation only after the semantic, truth-conditional content of the sentences has been worked out. Yet that is exactly what the contextualist denies. She claims that pragmatic inferences are required all the way down. And the problem is that the defeasible nature of prag-

Pragmatics in modern philosophy of language

209

matic inferences clashes with the deductive nature of semantic interpretation. This is problematic for various reasons: Firstly, because in these cases it is not enough to just follow any allegedly semantic rule or algorithm, because there is no semantic rule or algorithm to follow. (Though that is not to say that there might not be any syntactic rule, i.e. a rule of grammar). Secondly, a different conception of context is called for than the one invoked in syntactic-semantic interpretation if our aim is to accommodate all different forms of context-sensitivity. According to a very common conception of context (e.g. Lewis 1980; Kaplan 1989), an utterance context is taken to comprise a neat list of contextual features such as a world, a speaker, a time and a place of utterance, and thus to determine at least the values of indexical expressions.4 Yet such a conception of context will not do for present purposes. For one thing, because the idea of context determining anything stands in dire need of explanation. More importantly, what is needed is a conception of context according to which a context is seen as something that provides information: context has to provide all the information that speaker and interpreters have to draw on in order for communication and interpretation to be successful. That seems to be the idea behind Kent Bach’s notion of wide context: Wide context does not literally determine anything. It is the body of mutually evident information that the speaker exploits to make his communicative intention evident and that his audience relies upon, taking him to intend them to do so, to identify that intention. (Bach and Bezuidenhout 2002: 285)

Bach emphasizes that wide context only enables the hearer to figure out the semantic value of a given expression but does not determine it. Content is ascertained on the basis of information provided by wide context. This epistemic notion of context as providing speaker and hearer with information about how to interpret a given utterance seems to be better suited for modelling all different forms of context-sensitivity.5 Thirdly, and most importantly, pragmatic inferences often take the form of an inference to the best explanation or rather to the best interpretation.6 As a result, they are liable to be defeated by new information. And there is no limit to the amount of information that may turn out to be relevant (cf. Borg 2007: 356 and Borg 2004b: 79; Recanati 2004: 54). More specifically, pragmatic inferences are defeasible, non-monotonic, in the following sense: It may be the case that although conclusion C ‘follows’ (not in the deductive sense, of course) from a set P of premises, it no longer follows from a set M, such that M ⊃ P, i.e. such that P is a proper subset of M. In other words, adding new premises might turn a ‘good’ inference into a ‘bad’ one. (Defeasible reasoning is, of course, not confined to cases of linguistic interpretation. There is, by now, a huge literature on this mode of reasoning; see for an overview Koons 2009). But if that is so, then there may not be an algo-

210

Nikola Kompa and Georg Meggle

rithmic procedure for pragmatic interpretation as there is no guarantee that the task can be completed after finitely many steps. Inferences concerning speaker intentions are examples of pragmatic inferences. Here is an example of Borg’s to illustrate the defeasible nature of intentional interpretation: So, imagine that you see Sally filling a glass of water from the tap. Then you might reason as follows: ‘Sally is getting a glass of water form the tap. The best explanation for this action is that Sally is thirsty and wants a drink; therefore Sally is thirsty and wants a drink’. Clearly this is a non-demonstrative piece of reasoning and it is susceptible to the influence of an open-ended range of contextual factors. For instance, say that you know that Sally has just come in with Sourav and that Sourav is wearing running gear and looks out of breath, then the best explanation for Sally’s action might be that Sourav is thirsty and Sally is getting a drink for him. Or imagine that Sally has just glanced at her potted plant, then the best explanation might be that she wants to water her plant. (Borg 2004b: 78–79)

Interpreting a person’s actions in the way described clearly requires pragmatic inferences. And the interpretation of linguistic utterances is similar in that respect to the interpretation of actions. For many linguistic expressions are such that in order to figure out what has been said, speaker intentions – at least to the extent that speakers manage to make them manifest – need to be taken into account. Actually, even indexical and demonstrative expressions seem to require pragmatic inferences for their interpretation; what place or time exactly a speaker refers to with an indexical expression seems to depend, partially at least, on which region in space or what stretch of time he intended to refer to. (But let us put these problems aside for the time being.) Yet intentional interpretation is not the only case where pragmatic inferences come into play. In interpreting utterances, hearers not only draw on contextually provided information about (manifest) speaker intentions; other contextual information has to be taken into account, too. 1.3.2.

Context-sensitivity – some examples

For the purpose of illustration, let us consider a few more examples of context-sensitivity in natural language. Suppose a speaker says The suitcase is too heavy. at one time while trying to put a suitcase into the trunk of a car, at another time while packing his hand luggage for a flight to Paris (for a similar example, cf. Bach 1994: 268). Or suppose a speaker says: Everyone came to the party. (Stanley 2005: 223–224) at one time while discussing whether all the ‘important’ people in his department came to the party, at another time when commenting on how crowded the party was. As opposed to sentences such as “triangles have three sides” or “bachelors are unmarried”, our example sentences are context-sensitive in that they can be used to

Pragmatics in modern philosophy of language

211

say different things relative to different contexts of utterance; and their truth-conditions differ accordingly. Here is another example. Jill can’t continue. (Borg 2004b: 228) This sentence can be used to say that Jill cannot continue school, or that she cannot continue dance classes, or university education, and so on. What it is used to say on a particular occasion of utterance usually depends on context and can, accordingly, be figured out by taking context into account. Moreover, whether what is thereby said is true or false will depend on context, too. Similarly if a speaker says Tom is tall. The speaker might have used the sentence to say that Tom is tall for a fifth grader or that he is tall for an NBA player, etc. Moreover, his utterance may well be true in the first case but false in the second.7 Accordingly, the contextualist will claim that in order to figure out what has been said (not what has been implicated in the Gricean sense!) on a particular occasion of their use, most natural language sentences require pragmatic inferences on the part of the interpreter. Moreover, the contextual clues the interpreter has to draw on include not only information about the speaker’s intentions but also information about the purpose or point of the conversation, the participants’ shared background assumptions and interests, etc. Even more interesting from a contextualist point of view are still subtler forms of context sensitivity. Many natural-language expressions can be used to depict slightly different actions or entities in different contexts of their use (owing to the fact that language is used for classification, and classification itself is context-sensitive). Accordingly, they can be interpreted differently, again depending on the purpose or point of the conversation, the participants’ shared background assumptions, etc. That is just to say that there is ample polysemy in natural language, where “[p]olysemy is understood here in a broad sense as variation in the construal of a word on different occasions of use” (Croft and Cruse 2004: 109). Compare the following utterances: (a) (b) (c) (d) (e) (f) (g) (h)

Jane opened the window Bill opened his mouth. Sally opened her book to page 56. Mike opened his briefcase. Pat opened the curtains. The child opened the package. The carpenter opened the wall. The surgeon opened the wound. (Carston 2002: 325)

To open a wound is not the same thing as to open a book. And a wall is usually opened not in the way one opens one’s mouth. The word “open” can be used to depict a variety of different actions; and an interpreter has to contextually work out, i.e., pragmatically infer the appropriate understanding. There may be default inter-

212

Nikola Kompa and Georg Meggle

pretations, though, depending on the object of the verb. But these preferred interpretations are not inevitable. Also, one might think that in all these cases there is something like a semantic core meaning. In the above example, “manipulate such that the inside/interior/thing underneath can be seen” might be a promising candidate. But then the question arises of whether “open” is supposed to be synonymous with “manipulate such that the inside/interior/thing underneath can be seen”. If it is not supposed to be synonymous, the meaning of “open” differs from its core meaning, a puzzling result. Yet if it is supposed to be synonymous then one might wonder what the point of exact synonyms could possibly be in natural language? (And let us put aside the problem that giving a synonym is not the same as giving the meaning of an expression, for one might know that “a” and “b” are synonymous, without knowing the meaning of either “a” or “b”.) Moreover, new work on language acquisition seems to show that children, when learning a language, operate on the assumption that every difference in form (of linguistic expressions) marks a difference in meaning (cf., e.g., Clark 1993: 64; Tomasello 2003: 72–75ff; Bloom 2000: 65). Otherwise language acquisition would be a mystery. So the idea of exact synonyms and, consequently, of core meanings in the case of polysemy is rather problematic. Whether a sentence containing a polysemic expression is correctly applicable to a particular situation depends on the participants’ interests and purposes, their background assumptions and interests, etc. Given such and such purposes, interests, etc, it may be correctly applied, given different purposes, interests, etc., it may be not correctly applied. Yet couldn’t we specify further and try to make some of the purposes, interests or common assumptions explicit? The problem is, as Anne Bezuidenhout has pointed out, that [t]here is no sentence that we can produce that can settle all questions about how some original sentence is to be understood, since language doesn’t function that way. It is not self-interpreting. (Bezuidenhout 2002: 113)

In performing pragmatic inferences, one may have to draw on contextual information concerning the circumstances of utterance, speaker intentions (to the extent that they are manifest), the participants’ interests and background assumptions, the purpose or point of the conversation, and so on. More generally, we interpret someone’s utterance against the background of a shared system of knowledge and a common view of “how nature works and how our culture works” (Searle 1980: 226–227). This idea has been elaborated, among others, by Charles Travis (e.g., 1985, 1996, 1997), Julius Moravcsik (1998), and Anne Bezuidenhout (e.g, 2002), to name just a few. Moravcsik, for example, takes polysemy to affect almost all natural language expressions (cf. Moravcsik 1998: 33).8 Here is an example of his to illustrate the point:

Pragmatics in modern philosophy of language

213

The verb ‘walk’ admits of polysemy. On the one hand, there is a common meaning core, locomotion with legs in appropriate position. But one has to add: ‘appropriate distance, covered in an appropriate time.’ For what counts as a walk for a toddler’s first attempts does not count as a walk for a normal, healthy adult, and the walk of a recovering patient in a hospital is still a different matter. What counts as a walk depends on different senses of the word. (Moravcsik 1998: 35–36)

How the verb “walk” is to be understood, i.e., what counts as a walk, depends on context, more specifically on the various ways we can and do interact with the world, on our purposes, intentions, etc.: “Our understanding of words is influenced by the variety of ways in which we relate to the world.” (Moravcsik 1998: 91) And there is no limit to the different applications a given word may enjoy; there is no way of telling in advance what the different senses of a word are that we need it to have in order to be able to talk about everything we want to talk about (cf. Moravcsik 1998: 37). There are as many different uses to which we may put the expressions of our language as there are different purposes to which we may interact with the world. The meanings of our words provide rough and ready guidelines for their own future applications, in the light of new interests, purposes, assumptions etc. Yet, as Moravcsik emphasizes, we cannot predict or exactly specify all past and future human interactions with nature that have or will affect our uses of the words of our language. To echo John Austin, the question is: ‘Why do we call different things by the same name?’ (Austin 31979a: 69) A sensible answer seems to be: It allows us to use a finite vocabulary to talk about or describe an (in principle) infinite array of situations. Jonathan Cohen puts the point thus: Rather, it is an enormous convenience that the same word can often be uttered in one or other of several different though related senses. Instead of having to learn a very much larger number of words, each with fixed and context-independent meanings, we can learn a relatively small number of words with variable meanings and then exploit their verbal or situational contexts of utterance in order to disambiguate their actual occurrences. (Cohen 1985: 132)

Our expressions are flexible enough to fit all the various situations we encounter almost perfectly. Jon Barwise and John Perry speak of the ‘efficiency of language’ (Barwise and Perry 1983: 5; cf. also Goodman 21976: 80). The expressions of our language answer to our human needs, they serve our practical purposes. They can be contextually adjusted, if need be. And need there is as we never encounter exactly the same situation twice over. And when we encounter a situation we have not encountered before, our expressions allow us to conceptually structure it by using familiar vocabulary. Sometimes, though, we have to ‘unduly’ stretch the meaning of a given expression so that it will become applicable to the case at hand. That yields polysemy or metaphor, depending on the extent to which the new situation differs from old ones. Yet metaphor, just as polysemy, seems to defy formal treatment. Both phenomena seem to require pragmatic inferences for their interpretation.

214

Nikola Kompa and Georg Meggle

Metaphor in particular is commonly used to conceptualize abstract and mental phenomena, often with the help of vocabulary familiar from the realm of sense experience. Metaphor is, as Nelson Goodman nicely put it, “a matter of teaching an old word new tricks.” (Goodman 21976: 69) That is why we are feeling blue, or complain about someone being cold, and so on. We also use metaphors in the explanation of human behavior. We say that he broke down under pressure thereby exploiting the metaphor the mind is a brittle object (cf. Lakoff and Johnson 1980: 28). We say that she is feeling up, that her spirits rose, or that she sank into a coma. With the help of these ‘orientational metaphors’ (as Lakoff and Johnson 1980: ch. 4 call them), we conceptualize mental phenomena by giving them a spatial orientation. But abstract phenomena too are commonly conceptualized with the help of metaphorical expressions. We construe theories and arguments as buildings: The theory needs more support, it lacks a foundation; the argument collapsed, it was not solid enough (cf. Lakoff and Johnson 1980: ch. 10). Or we think of love as if it were a journey: look how far we have come; yet now we are at a crossroad; from here on we have to go separate ways as this isn’t going anywhere; we are in a deadend street. We commonly use expressions from the concrete realm of sense experience to conceptualize abstract and mental phenomena (cf. also Keller and Kirschbaum 2003: 36 and 99). So if one were to say, as formal semanticists with their fondness for a clear-cut distinction between literal and figurative meaning would be inclined to do, that metaphorical utterances are always strictly speaking false, that would render most of our talk about abstract or mental phenomena false. Polysemy and metaphor are central features of natural languages without which we would not be able to structure the multifariousness of our experience and to conceptually copy with mental and abstract phenomena. Some philosophers take these and similar considerations to show that word meanings (or concepts) are not ready-made entities but have to be constructed in context. According to Robyn Carston, for instance, often in linguistic interpretation “an ad hoc concept is constructed and functions as a constituent of what is explicitly communicated”. (Carston 2002: 357) And, presumably, pragmatic inferences are required in the construction of these ad hoc concepts. But one might even go one step further and try to do without something like linguistic meaning (as traditionally understood) altogether (cf., e.g., Recanati 2004: 146ff; Croft and Cruse 2004: 97ff). The idea would be that although words encode semantic information, the information itself does not amount to something like a full-fledged meaning. Rather, it provides the starting point for pragmatic processes, which result in a contextually specified interpretation of the expressions in question. “Linguistic meaning” would provide information that constrains interpretation, while pragmatic processes are required to yield the contextually relevant contribution the expression makes on a particular occasion of its use. According to this conception of linguistic interpretation, the notion of linguistic meaning itself needs to be pragmatically spelled out.

Pragmatics in modern philosophy of language

1.4.

215

Conclusion

The preceding considerations cast doubt on the idea of a functional, systematic pairing of sentences with meanings (truth-conditions) as propagated by formal semantic approaches to language. This way of approaching language seems to ignore the highly context-sensitive nature of linguistic interpretation. Many natural language expressions require for their proper interpretation that context be taken into account. This kind of context-sensitivity rests on a conception of context according to which context provides information about the participants’ shared interests, intentions, purposes, background assumptions and so on. The information is extracted by pragmatically inferring the contextually appropriate interpretation of a given context-sensitive utterance. Pragmatic inferences are indispensable in figuring out what a given expression contributes to the truth-conditions of a sentence containing it (i.e., to the proposition expressed). Accordingly, the value assigned to the expression hardly deserves to be called a (purely) semantic value. And the process of interpretation performed hardly deserves to be called (purely) semantic interpretation. Rather, pragmatic processes seem to intervene at (almost) all stages of linguistic interpretation. And maybe even the notion of linguistic meaning itself is a pragmatic notion.

2

Pragmatic semantics

2.1.

The center of pragmatic semantics

Just cancel the second word (“maybe”) of the last sentence in the preceding paragraph, and you have already reached the very center of Pragmatic Semantics (PS). Whereas traditional semantics (including formal truth-conditional semantics as its most developed form) starts with linguistic expressions and then asks how to define the various kinds of their respective linguistic meanings, PS starts just the other way round, namely with the understanding (i.e., the knowing of the meanings) of actions. From the point of view of PS, semantics is a special part of pragmatics (Action Theory), in the ideal case therefore a part of Action Logic. Throughout, we will have to make use of the following two-fold distinction Type Token

Action (3) (4)

Action-Product (1) (2)

or of its already language-related specification, respectively: Utterance Type Token

Utterance-Product Expression (form) (Concrete) Expression

216

Nikola Kompa and Georg Meggle

Whereas traditional semantics starts from the right side, (1) above, PS starts from and is primarily focussed on the left one. Correspondingly, “pragmatics” is covering quite differing programs. Whereas, as exemplified in chapter 1 above, according to traditional semantics it may mean anything which ‘goes beyond’ field (1), the program of “Pragmatic Semantics” as sketched here is to define in the end even (1)-meanings themselves in terms of action-theoretical concepts. In PS “actions” (and their meanings) come first. For this semantic paradigm, there even need not be any expressions to start from. Actions, think of gestures, for example, do not need to be product-producing actions (utterances of expressions) in order to have meanings. Thus, in this approach “expressions” themselves may be recognized or introduced as what they are: constructions on the basis of (some continuity in) the meanings of the corresponding actions. Thus, PS is primarily a Semantics of Actions. 2.2.

Rule-based versus intention-based pragmatic semantics

In Pragmatic Semantics we also find the division between nominalists and Platonists, depending on whether type-meanings (meanings of action-types, field (3) in the schema above) can themselves be defined in terms of token-meanings ((4) in our schema) or not. Since type-meanings (of the PS-relevant sort) are social facts (i.e., state of affairs which, relative to population P, are facts iff it is common believe in P that they are facts), whereas the respective meanings of action-tokens are just meanings for some particular subject, this (Platonist versus nominalist) division is a special case of the more general division within Social Philosophy between collectivist versus individualist approaches, according to whether or not meanings as social facts are taken to be conceptually inexplicable with reference to individualistic terms. This distinction is very often identified with a distinction between the two possible readings of Wittgenstein’s basic equation of “meaning = use”, 9 which one may refer to as the rule-based reading versus the intention-based reading. Whereas according to the rule-based reading we have, in order to understand (i.e. to know the meaning of) a sign A, just to know the rules we have to follow in order to use A correctly, the intention-based reading is referring to what in a particular situation A is being used for. Which alternative will be the most fruitful one? This has been the most debated question in the whole field of pragmatic semantics up to now (see 2.5–2.7 below). 2.3.

Intention-based semantics

This overview will start from Intention-Based Semantics, which is definitely following a strictly nominalistic / individualistic strategy by embedding PS into gen-

Pragmatics in modern philosophy of language

217

eral individualistic Action Theory including Rational Decision Theory as its most important part (see Jeffrey 1983). The positions of the main collectivist alternatives can then easily be located by referring to their denials of – and their alternatives to – the respective aspects of Intention-Based Semantics. For Intention-Based Semantics, as sketched by Herbert P. Grice (1957) and further developed mainly by Stephen Schiffer (1972), Jonathan Bennett (1976), David Lewis (1969), Franz von Kutschera (1975) and Georg Meggle (2010), human languages are primarily used for interpersonal communication. Accordingly, linguistic meanings as social facts are primarily defined in terms of the communicative regularities or conventions holding in the respective social groups (populations). But as the meanings of communicative action tokens may and often do diverge from the meanings of the exemplified action type, “communicative actions” have first to be defined more generally, i.e. on the token level. Thus, the Intention-Based Semantics-program (even when restricted to the action side, i.e., not yet including some expressions as their action-products) has to involve two steps: (i) explication of general concepts of communicative acts (as tokens); and (ii) explication of (meanings of) communicative action types by means of (i). Whereas step (ii) is at the very center of Intention-Based Semantics, (i) is its necessary preliminary stage defining communicative acts in terms of their respective communicative intentions.10 2.3.1.

General communication theory

Communicative acts are instrumental acts, with the most basic types of communicative acts being either directives or informatives, depending on whether S’s (the speaker’s) primary communicative aim is to get H (the hearer) either to do or to believe something. But notice that this way of conveniently speaking of “speaker” and “hearer” does not imply that the respective communicative acts are restricted to acoustic utterances. The differentia specifica of communicative acts versus instrumental acts in general is that communicative acts are acts which, at least from the point of view of S, will be successful if and only if they will be understood by their addressee H. This entails the so called reflexivity-condition of communication: For an act to be a communicative one, its actor S has to intend it to be understood by H (as being a communicative act). A General Communication Theory has to explicate these postulates in a non-circular way. For this explication, one best starts from the more general communicative acts: Instrumental acts are acts done for some end (or, to put it differently, done with specific intentions). More precisely (but already strongly adapted to the special cases of ‘decisions under certainty’, i.e., cases where the deciding person expects the results of her actions with a maximal degree of probability), person X is doing f with the intention of bringing it about that the state of affairs A will be the case: (1) X does f, (2) X wants that A, and (3) X believes (expects) that A

218

Nikola Kompa and Georg Meggle

will be brought about as and only as a consequence of her doing f. Instrumental acts will thus be successful iff the belief (3), let us call it the expectation of success, turns out to be true, i.e., if A will in fact come about in the way expected by X, i.e., as and only as a consequence of X’s doing f. And you understand this act iff you know its ends, i.e., iff you know the intentions with which f is or was being done by X.11 As there are no restrictions as to what ends one may try to achieve by means of communication, we have to distinguish between the intentions with which S is performing her communicative act (i.e. her communication related intentions) in general versus her act’s communicative intentions in particular, this latter term referring only to those intentions that are necessary for an act to be a (respective) communicative act. Unlike instrumental acts, communicative acts do not only involve the expectation of success, but also the expectation of understanding (i.e., the expectation that H will understand S’s communicative act), both expectations connected in the differentia specifica way already mentioned above: that S expects that success will be brought about as and only as a consequence of understanding. Thus, for the communicative act to be successful this expectation must turn out to be true, i.e. communicative success must in fact come about in the way expected by S, i.e., as and only as a consequence of H’s understanding. The most crucial and contested question to be settled is: What do we have to know in order to understand S’s communicative act? To answer this question, one has to notice that the Reflexivity-Condition (RC) above already implies what we might call the absolute openness of communication. Whatever condition has to be satisfied in order for an act to be a communicative one (let’s say condition ), it follows from RC that this very condition  itself has to be intended to be recognized by H as being satisfied as well, which is itself a necessary condition for the act to be a communicative one, etc. Logically, this openness implied by RC will have no upper end. For such an unrestricted (and in this sense ‘absolute’) openness let’s write openness* for short. From RC we get straight to O*: O*

Communicative intentions must be open* ones.

On the grounds of this condition, Grice’s first proposal that from S’s point of view only H’s recognition of S’s primary communicative intention is necessary, can immediately be seen to be inadequate. Grice’s communications are not open enough. But Grice’s Basic Model (BM) can easily be improved:12 BM

S’s doing is an attempt at communication, addressed to H, and of the content that H is to do r, iff (i) S intends, by doing the act, to make H do r; and (ii) S believes (expects) that she will achieve this aim iff she is letting H openly* know that S has this intention (i)

Pragmatics in modern philosophy of language

219

Communication is thus regarded as proposed by Grice plus the factor of openness*. This kind of openness*, however, relates only to communicative intentions, i.e., to those intentions which are necessary conditions for an act to be a communicative act. Sincerity is not a necessary condition for an act to be a communicative one; communicative openness (being a case of openness*) is therefore absolutely compatible with the insincerity of the communicative act in question (including downright lies). This concept of communicative acts (as defined in BM) is a very broad one. Though it involves the expectations of understanding and of success, it implies nothing about the reasons these expectations may or will have to be based on in order to be (rationally to be expected by S to be) satisfied. This is fine as far as communicative acts (and their subjective meanings) in general are concerned. But things change when we come to the special cases of communication by means of (types of) actions with some intersubjective communicative meaning. 2.3.2.

Meanings of action types

Meaning and understanding (i.e. knowledge and recognition of meaning) are two sides of one and the same coin. This holds both for subjective and intersubjective meanings. (Inter-) subjective meanings of (the type f of) an action are relative to the (class of) persons who understand the actions (of type f) as having these (inter-) subjective meanings. In addition, meanings of both kinds are relative to (types of) situations. The fact, that some gesture (of type) f, let’s say “thumbs up”, as used in the population P in situations of type , means “p” (e.g.: “you are fine”) may be equated roughly with the following regular meaning condition of f in P relative to : Whenever (or when in sufficiently many cases) in a -situation a S (as a member of P) addresses H (as another member of P), making a gesture (of type) f, it holds that by doing f, S is trying to communicate to H that p (“You are fine”). Since (type-) meanings (as social facts) have to be known to be (intersubjective) meanings, the above regular-meaning-definiens will have to be postulated as being common knowledge in P (regular meaning in a narrow sense of f in P relative to ).13 Regular meanings of f (in P for ) are, in -situations and for all members of P, the best possible reasons for the respective S’s expectation of understanding. Conventional meanings result from the likewise rationally founded expectation of success: H will believe what S is trying to make her believe by this very action f (namely that p) iff H will understand S’s communicative action f. Normally this will hold only if (S expects that) H believes that, as far as p is concerned, neither S will be deceptive to her (H) nor that S himself will be at fault. And these expectations will be stable (within P and relative to -situations) in the long run only if they will always (or in most or at least in sufficiently many cases) turn out to be satisfied. And this again will be true only if both S and H will

220

Nikola Kompa and Georg Meggle

be following a joint strategy, which in turn will be stable only if this strategy is based on joint interests, as for example, that H is to believe p iff it is in fact the case that p. Iff, in P and relative to -situations, the communicative expectations of success are based in this way, then the regular meanings of the respective communicative actions will also be conventional ones. This explication is roughly the same as David Lewis’s explication of signalling-conventions (in Lewis 1969). Now, and this is the decisive step from actions to expressions as their products, the meanings of (complete or unstructured) signs or expressions (you may think of the gesture “thumbs up” again) can be equated with the conventional meanings of the communicative acts of producing these expressions. 2.3.3.

Linguistic meaning

Unfortunately, when turning to languages and linguistic meanings of structured signs, Lewis opted out of the whole Intention-Based Semantics-program by putting his proposals on the shoulders of traditional realistic semantics. How this predicament may be avoided in relation to a simple language (of first order predicate logic) is shown in Kutschera (1975), who provides us with an “illustration of the schema according to which a grammar can be constructed”. For a discussion of such a pragmatic reconstruction see Bennett (1976) and Meggle (2010), the former trying to couch it in more behavioristic terms, the latter in more formal terms. Following this approach, the meanings of so called realistic semantics (even in its most developed form of intensional semantics) can be shown to be successfully explicable in terms of Intention-Based Semantics. But, as in many other places, explicability does not imply reduction. 2.4.

Classical speech act theory

Though rightly looked at as being a paradigm case of pragmatics, (classical) speech act theory was definitely not (intended to be) a special case of PS. This was clear from the start because John L. Austin (1962), in his basic distinction between locutionary, illocutionary and perlocutionary (aspects of) acts, was referring to the use of a sentence in its literal meaning “in the sense of Frege”. From Austin’s point of view, speech act theory should only be interested in the illocutionary functions of speech acts. This view is shared by his most influential interpreter, John R. Searle. In this sense, Searle’s slogan that “a theory of language is part of a theory of action” (Searle 1969: 17) can only be misleading. As to the class of illocutionary acts, Austin did not give us an explicit definition, only an open list of examples. But there are two elements in his characterizations of illocutions which obviously contradict the explication of communicative acts (in the wide sense of directives and informatives) as sketched in 2.3.1 above. According to Austin, illocutionary acts are (i) essentially conventional acts; and, as

Pragmatics in modern philosophy of language

221

such, (ii) they have to exclude any ‘perlocutionarity’.14 Whereas (i) may be weakened (cf. Bach and Harnish 1979) by distinguishing between essentially conventional illocutionary acts (including Austin’s favorites of marrying or baptizing someone) on the one side, and ‘mere’ communicative actions on the other side, the ‘non-perlocutionarity’ of illocutions remained the central dogma of classical speech-act-theory (even with regard to communicative acts). There is some truth in this dogma, but this truth can be easily reduced to the distinction between communication related intentions in general (i.e. perlocutionary intentions in a broad sense) versus communicative intentions in particular (i.e. perlocutionary intentions in the narrow sense) (cf. 2.3.1 above). With regard to the broad sense of “perlocutionary”, the non-perlocutionary dogma is (by definition) true; but in this sense it is (as implied in 2.3.1) no objection to the presented definition of (the communicative variant of) illocution at all. 2.5.

Speech act based PS

It was William P. Alston (1964) who first tried to build a PS on speech act theory. His basic idea was that two (unstructured) expressions  and can be said to have the same meaning iff they have the same illocutionary potential, i.e., could be used to perform the same illocutionary acts, where “the same illocutionary acts” was meant to refer not only to acts of the same illocutionary types (assertions, directives, questions etc.) but also to these acts as having also the same contents (assertions that p, directives to bring about q, questions as to whether it is the case etc.). In other words: Illocutionary potential in Alston’s sense is already a content-determined illocutionary act potential (CDIAP). And though his basic idea was presented by him only in the form of an identity criterion for the meanings of  and , it was clear that Alston wanted to identify and thus also define the meanings of expressions by means of their CDIAPs. That was the essential difference: Whereas in Austin’s theory meanings of expressions are presupposed as being given, Alston’s idea was to define these expression meanings themselves by means of the meanings of illocutionary actions. “If this is the line along which meaning should be analyzed, then the concept of an illocutionary act is the most fundamental concept in semantics and, hence in the philosophy of language” (Alston 1964: 37 [emphasis mine]). If we substitute “a communicative act” for “an illocutionary act”, Alston’s speech act approach would seem to coincide with the Intention-Bases Semantics as sketched in 2.3 above. But notice that even Alston in 1964 subscribed to the two dogmas of classical speech act theory. However, nearly 40 years later, he has extended his basic idea of meanings as CDIAPs into a full blown PS of its own (Alston 2000). Beside Alston, the most detailed and best developed rule-oriented PS is Eike von Savigny’s theory (1983). Whereas the intention-based PS as sketched above starts with the communicative meaning of concrete utterances (action-

222

Nikola Kompa and Georg Meggle

tokens, i.e. field (4) of the schema in 2.1 above), Savigny’s rule-based PS starts with the conventional meaning of utterance types (action-types, i.e.field (3) in the same schema). According to this approach, differences in meaning are correlated with differences in the conventional commitments of the persons involved in the respective utterance-situation. And these commitments in turn are defined by means of rules as explicated by Hart (1961) with reference to regularities, sanctions and acceptance of these sanctions. Finally, sentence meanings are to be introduced as theoretical constructs in order to systematically explain the salient structures to be found in these regularities. To systematically compare the IB- versus these and other Rule-Based Approaches to PS with regards to their various pros and cons is one of the biggest desiderata in the whole field of modern philosophy of language. 2.6.

Variety of language games

As far as Wittgenstein’s “meaning = use” equation is concerned (see 2.2 above), the usual distinction between the rule-reading versus the intention-reading of it should not be the last word. At least not in the sense of an either-or distinction because there is some chance that both readings need not be really exclusive. First, as noticed in 2.6, the two readings may refer to different explication-levels, the intention-reading to the token-one, the rule-reading to the type-one. Next, as noticed in 2.4, only a special class of perlocutionary intentions is relevant for the definition of communicative/illocutionary acts: communicative intentions (see 2.3.1). As these intentions are necessary conditions for some action to be a communicative action (a directive, for example), they may be looked at as being part of “the internal relation”, which, according to Wittgenstein (§ 21), is constitutive of the kind of meaning his various language games are meant to be exemplifications of. According to this reconstruction of Wittgenstein’s meaning as use theory,15 the relevant intention-reading is part of the rule-reading itself. But there are many other language games (many other illocutions) whose internal relations do not include any reference to perlocutionary intentions. One might therefore say that meanings (illocutions) do not per se entail perlocutions. But from this it does not follow that perlocutions (as part of meanings) are per se excluded. There is more than just one kind of meaning. 2.7.

Evaluations

This difference is of the utmost importance for any evaluation of the various PSapproaches. What kinds of illocution are the most important ones? The IB-semanticists (as sketched in 2.3 above) take it that interpersonal communication is at the very basis. But they do not deny, of course, that languages may have many other functions as

Pragmatics in modern philosophy of language

223

well. Some neo-Griceans, interpreting Grice’s original starting-point of “speaker’s meaning” as covering much more than communicative intentions, do accordingly extend this basis and include, e.g., the intention to express one’s thoughts or feelings (Davis 2003). Others, like Austin, think it best to start with fully institutionalized or even ritualized kinds of performances. Finally, there are scholars like Brandom (1994), who take the whole PS-project (of starting with primitive acts or intentions and then constructing something like linguistic meanings out of them) to be on the wrong track because they claim that language has to be presupposed before attributing even these most primitive elements. Before we postulate the one PS which embraces all sorts of things language may be used for, we should keep in mind that we just cannot come up with the complete theory of language and still hope for its consistency.

Notes 1. For an overview of the development of semiotics, see e.g., Scholz 2003, for semiotics and pragmatics Nöth, this volume, and for a general historiographic overview of pragmatics Koyama, this volume. 2. For an account of the relationship between semantics and pragmatics cf. Saeed, this volume. 3. Kaplan also speaks of the character of indexical expressions as of a rule that “fully determine[s] the referent for each context”. (1989: 491) 4. For this specific conception of context cf. Lewis “A context is a location – time, place, and possible world – where a sentence is said.” (1980: 21), and Kaplan, who adds the speaker or acting participant as a further constituent, argues that an index representing a context is a 4-tuple , consisting of a world (w), an actor (x), a place (p) and a time (t) such that x is in w at t in p (Kaplan 1989: 509). 5. Stalnaker’s notion of context might be promising here: “So I propose to identify a context (at a particular point in a discourse) with the body of information that is presumed, at that point, to be common to the participants in the discourse”. (1998: 98) 6. For more on inferences to the best explanation in general, see, e.g., Lipton (1991). 7. Some context-sensitive sentences may have preferred or default completions, though. In the absence of any special stage setting, we will interpret the sentence: “Sam’s daughter got married and became pregnant”, for example, as meaning that Sam’s daughter first married and then became pregnant. For more on the notion of a preferred or default interpretation see Levinson (2000). 8. See also Searle (1978: 220–221) and (1980: 221). 9. Which may be treated as something like the cornerstone of the whole of PS. 10. For formal explications of step (i) see Meggle (1997) and for step (ii) Meggle (2010). 11. Following Max Weber one could also say that Action Understanding in this sense is to know the action’s subjective meaning / “subjektiver Sinn”. 12. With (BM) being already a reconstructed version of Grice’s own version (1957): “S meant something by x” is (roughly) equivalent to “S intended the utterance of x to produce some effect in an audience by means of the recognition of this intention”.

224

Nikola Kompa and Georg Meggle

13. Where Common Knowledge in P that p may be easily defined this way: (i) All members of P know that p; (ii) All members of P know that (i) – etc., with this “etc” to be defined in formal terms as usual by recursion. 14. See Austin (1962: ch. 9), where perlocutionary effects are by definition restricted to effects going beyond (conventional) illocutionary effects. 15. Very forcefully proposed for the first time by Feyerabend (1955) and heavily relying on Wittgenstein’s famous language game, cf. § 2 in his Philosophical Investigations.

References Alston, William P. 1964 Philosophy of Language. Englewood Cliffs, NJ: Prentice-Hall. Alston, William P. 2000 Illocutionary Acts and Sentence Meaning. Ithaca and London: Cornell University Press. Austin, John L. 1962 How to do things with Words. Oxford: Oxford University Press. Austin, John L. 31979a The Meaning of a Word. In: John Austin, Philosophical Papers, 55–75. Oxford: Oxford University Press. Bach, Kent and Robert M. Harnish 1979 Linguistic Communication and Speech Acts. Cambridge, MA: Harvard University Press. Bach, Kent 1994 Semantic slack. In: Savas L. Tsohatzidis (ed.), Foundations of speech act theory: Philosophical and linguistic perspectives, 267–291. London/New York: Routledge. Bach, Kent and Anne Bezuidenhout 2002 Distinguishing Semantics and Pragmatics. In: Joseph Keim Campbell, Michael O’Rourke and David Shier (eds.), Meaning and Truth: Investigations in Philosophical Semantics, 284–309. New York/London: Seven Bridges Press. Barwise, Jon and John Perry 1983 Situations and Attitudes. Cambridge, MA: MIT Press. Bennett, Jonathan 1976 Linguistic Behaviour. Cambridge: Cambridge University Press. Bezuidenhout, Anne 2002 Truth-Conditional Pragmatics. Philosophical Perspectives 16: 105–134. Bloom, Paul 2000 How Children Learn the Meanings of Words. Cambridge, MA: MIT Press. Borg, Emma 2004a Formal Semantics and Intentional States. Analysis 64: 215–223. Borg, Emma 2004b Minimal Semantics. Oxford: Oxford University Press. Borg, Emma 2007 Minimalism versus Contextualism in Semantics. In: Gerhard Preyer and Georg Peter (eds.), Context-Sensitivity and Semantic Minimalism: New Essays on Semantics and Pragmatics, 339–359. Oxford: Oxford University Press.

Pragmatics in modern philosophy of language

225

Brandom, Robert B. 1994 Making it explicit. Reasoning, Representing and Discursive Commitment. Cambridge, MA: Harvard University Press. Brandom, Robert, B. 2002 Pragmatics and Pragmatisms. In: James Conant and Urszula M. Zeglen (eds.), Hilary Putnam: Pragmatism and Realism, 40–59. London: Routledge. Clark, Eve V. 1993 The lexicon in acquisition. Cambridge Textbooks in Linguistics, Cambridge: Cambridge University Press. Carnap, Rudolf 1956 Meaning and Necessity: A Study in Semantics and Modal Logic. 2nd ed. Chicago: The University of Chicago Press. Carston, Robyn 2002 Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell. Cohen, Jonathan L. 1985 A Problem about Ambiguity in Truth-Theoretical Semantics, Analysis, Vol. 45/3: 129–134. Croft, William and Cruse, Alan D. 2004 Cognitive Linguistics. Cambridge Textbooks in Linguistics, Cambridge: Cambridge University Press. Davis, Wayne A. 2003 Meaning, Expression, and Thought. Cambridge: Cambridge University Press. Feyerabend, Paul 1955 Wittgenstein’s Philosophical Investigations. Philosophical Review: 449–483. Frege, Gottlob 1891 Funktion und Begriff. Lecture held 1891 at a meeting of the “Jenaischen Gesellschaft für Medizin und Naturwissenschaft”. Reprinted in Gottlob Frege 1986, 6th ed., 18–39. Frege, Gottlob 1892a Über Sinn und Bedeutung. Zeitschrift für Philosophie und philosophische Kritik 100, 25–50. Reprinted in Gottlob Frege 1986, 6th ed., 40–65. Frege, Gottlob 1892b Über Begriff und Gegenstand. Vierteljahrsschrift für wissenschaftliche Philosophie XVI, 192–205. Reprinted in Gottlob Frege 1986,6th ed., 66–80. Frege Gottlob 1986 Funktion, Begriff, Bedeutung. Edited by Günther Patzig, 6th ed. Vandenhoeck und Ruprecht, Göttingen. Goodman, Nelson 1976 Languages of Art. An Approach to a Theory of Symbols. 2nd ed. Indianapolis: Hackett. Grice, H. Paul 1957 Meaning. The Philosophical Review 66: 377–388. Reprinted in Grice 1989. Grice, H. Paul 1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press. Hart, Herbert L.A. 1961 The Concept of Law. Oxford: Clarendon Press.

226

Nikola Kompa and Georg Meggle

Jeffrey, Richard C. 1983 The Logic of Decision. 2nd ed. Chicago: University of Chicago Press. Kaplan, David 1989 Demonstratives. An Essay on the Semantics, Logic, Metaphysics, and Epistemology of Demonstratives and Other Indexicals. In: Joseph Almog, John Perry, and Howard Wettstein (eds.), Themes from Kaplan, 481–563. Oxford: Oxford University Press. Keller, Rudi and Ilja Kirschbaum 2003 Bedeutungswandel: Eine Einführung. Berlin: de Gruyter. Koons, Robert 2009 Defeasible Reasoning. In: Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2009 Edition). . Kutschera, Franz von 1975 Conventions of Language and Intensional Semantics. Theoretical Linguistics Vol 2/3: 255–283. Kutschera, Franz von 1976 Einführung in die intensionale Semantik. Berlin: de Gruyter. Lakoff, George and Mark Johnson 1980 Metaphors We Live By. Chicago: University of Chicago Press. Levinson, Stephen C. 2000 Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Lewis, David 1969 Convention. Cambridge, MA: Harvard University Press. Lewis, David 1975 Languages and Language. In: Keith Gunderson (ed.), Minnesota Studies in the Philosophy of Science, Vol. VII, 3–35. University of Minnesota Press. Reprinted in David Lewis 1983: Philosophical Papers. Volume 1, 163–188. Oxford: Oxford University Press. Lewis, David 1980 Index, Context, and Content, in Stig Kangar and Sven Öhman (eds.), Philosophy and Grammar, 79–100. Dordrecht: Reidel. Lipton, Peter 1991 Inference to the Best Explanation. London: Routledge. Meggle, Georg 1997 Grundbegriffe der Kommunikation. Berlin: de Gruyter. Meggle, Georg 2010 Handlungstheoretische Semantik. (Habilitation thesis 1984). Berlin: de Gruyter. Montague, Richard 1970 English as a Formal Language. In: Bruno Vicentini et al. (eds.), Linguaggi nella società e nella tecnica, 189–224. Mailand: Edizioni di Comunità. Reprinted in Montague 1974, 188–221. Montague, Richard 1970b Universal grammar. Theoria 36: 373–398. Reprinted in Montague 1974, 222–246.

Pragmatics in modern philosophy of language

227

Montague, Richard 1973 The proper treatment of quantification in ordinary English. In: Jaakko J. Hintikka, Julius M. Moravcsik and Patrick Suppes (eds.), Approaches to natural Language, 221–242. Dordrecht: Reidel. Reprinted in Montague 1974, 17–34. Montague, Richard 1974 Formal Philosophy: Selected Papers of Richard Montague. Edited by Richmond H. Thomason. New Haven/London: Yale University Press. Moravcsik, Julius M. 1998 Meaning, Creativity, and the Partial Inscrutabilty of the Human Mind. CSLI Publications, Stanford. Manor, Ruth 2001 On the Overlap of Pragmatics and Semantics. Synthese 128: 63–73. Morris, Charles 1938 Foundations of the Theory of Signs. In: Otto Neurath, Rudolf Carnap and Charles Morris (eds.), International Encyclopedia of Unified Science. Vol. 1, Nr.2, 77–138. Chicago: University of Chicago Press. Reprinted in Charles Morris, 1971, Writings on the General Theory of Signs, 17–71. Mouton: The Hague/Paris, quotations are to Morris 1971. Partee, Barbara 1973 Some transformational extensions of Montague grammar. Journal of Philosophical Logic 2: 509–534. Partee, Barbara 1975 Montague grammar and transformational grammar. Linguistic Inquiry 6/2: 203–300. Portner, Paul and Barbara H. Partee 2002 Introduction. In: Paul Portner and Barbara H. Partee (eds.), Formal Semantics: The Essential Reading, 1–16. Oxford: Blackwell. Recanati, François 2004 Literal Meaning. Cambridge; Cambridge University Press. Savigny, Eike von 1983 Zum Begriff der Sprache. Konvention, Bedeutung, Zeichen. Stuttgart: Reclam. Schiffer, Stephen 1972 Meaning. Oxford: Oxford University Press. Scholz, Oliver R. 2003 Semiotik und Hermeneutik. In: Roland Posner, Klaus Robering and Thomas A. Sebeok (eds.), Semiotik/Semiotics: Ein Handbuch zu den zeichentheoretischen Grundlagen von Natur und Kultur / A Handbook on the Sign-Theoretic Foundations of Nature and Culture, Vol. 3, 2511–2561. Berlin: de Gruyter. Searle, John R. 1969 Speech Acts. Cambridge: Cambridge University Press. Searle, John R. 1978 Literal meaning. Erkenntnis 13: 207–24. Searle, John R. 1980 Background of Meaning. In: John Searle, Ferenc Kiefer, Manfred Bierwisch (eds.), Speech Act Theory and Pragmatics, 221–232. Dordrecht: Reidel.

228

Nikola Kompa and Georg Meggle

Stalnaker, Robert C. 1998 On the Representation of Context. Journal of Logic, Language, and Information 7. Reprinted in Stalnaker 1999 Context and Content. 96–113. Oxford: Oxford University Press. Stanley, Jason 2002 Making it Articulated. Mind and Language 17: 149–168. Stanley, Jason 2005 Semantics in Context. In: Gerhard Preyer and Georg Peter (eds.), Contextualism in Philosophy: Knowledge, Meaning, and Truth, 221–253. Oxford: Oxford University Press, Tomasello, Michael 2003 Constructing a Language: A Usage-Based Theory of Language Acquisition. Cambridge, MA: Harvard University Press. Travis, Charles 1985 On what is strictly speaking true. Canadian Journal of Philosophy 15: 187–229. Travis, Charles 1996 Meaning’s Role in Truth. Mind 105: 451–466. Travis, Charles 1997 Pragmatics. In: Bob Hale and Crispin Wright (eds.), A Companion to the Philosophy of Language, 87–107. Oxford: Blackwell. Wittgenstein, Ludwig 1953 Philosophische Untersuchungen. Ed. by G.E.M. Anscombe and Rush Rhees. Oxford: Basil Blackwell.

8.

Foundations of pragmatics in functional linguistics Saskia Daalder and Andreas Musolff

1.

Introduction

The relatively young discipline of linguistic pragmatics has had forbears in several older branches of linguistics. This chapter focuses on important “prefigurations” of parts of linguistic pragmatics from the late nineteenth century to around 1970 which were informed by a “functional” view of language. We view “functional linguistics” as the study of language that starts its research with a focus on the functions of language in social life: the effects of language use, differentiated as to types of communicators, types of contexts and types of language uses. This includes the analysis of the primarily intended goals of the speech participants in their speaking and reacting as well as the study of the long-time effects of language use in the life of the individual and in the history of the language. One could name quite a few functionally inclined linguists of that period whose theories are to some extent congenial to modern pragmatics, starting with the theorists Dwight Whitney (1827–1894), Michel Bréal (1832–1915), Georg von der Gabelentz (1840–1893) and Henry Sweet (1845–1912) and ending with, e.g., André Martinet (1908–1999) and Michael Halliday (*1925). However, in this article we will confine ourselves to results of functional linguistics that have actually been transmitted to linguistic pragmatics and may be counted among its foundations. Such results are found foremost in the works of four theorists: the German psychologist Karl Bühler (1879–1963), the Czech scholar of English Vilém Mathesius (1882–1945), the Russian general linguist and scholar of Slavonic languages Roman Jakobson (1896–1982) and the British general linguist Michael Halliday. Of these, we will discuss the relevant theories of the first three. Bühler proposed an influential model of three fundamental “language functions”, which Jakobson came to extend to the model with six such functions that has had a strong impact on pragmatics, the ethnography of communication and social semiotics.1 Furthermore, both Bühler and Jakobson analysed the functioning of the deictic elements of language, emphasising their role in the “situated” nature of language use. Mathesius, who was more specifically concerned with the comparison of living languages as to their possibilities of expression, provided the foundations of the functional analysis of utterances in terms of the notions “theme” and “rheme” (“topic”–“comment”), which has become one of the stock items of pragmatic analysis. For their ideas in this domain, both Bühler and Mathesius were indebted to the work of the older and less well-known psychologically oriented German lin-

230

Saskia Daalder and Andreas Musolff

guist Philipp Wegener (1848–1916), who held views that also show a strong affinity to those of modern linguistic pragmatics. In our sections 2–5, we will in turn discuss the contributions by Wegener, Bühler, Mathesius and Jakobson. In the concluding section 6, we consider the questions of whether and how their insights into the functionality of language can still be of value for linguistic pragmatics. In our view, these insights still provide valid reasons to integrate the description of elements of the situational context into linguistic analysis itself and thus to question the wisdom of completely separating research into the system of language from pragmatics, the study of its use.

2.

Philipp Wegener

Philipp Wegener (1848–1916), a classics teacher, grammar school headmaster and psychological linguist from Northern Germany, developed in a relatively small body of work a remarkably modern, communication-oriented analysis of language.2 It resulted from his opposition against a particular claim of contemporary historical linguistics. The Neogrammarian theorists had acknowledged the relevance of psychological factors for processes such as the borrowing of linguistic elements by one language or dialect from another, but not for so-called “internal”, primarily phonetic language change, which was thought to be caused purely by physiological mechanisms. Motivated by his study of developments in the local dialects, Wegener wished to demonstrate the concrete psychological and communicative background also of “internal” changes, and this forms the first topic in his 1885 publication Investigations into the Fundamental Questions on the Use of Language (Untersuchungen über die Grundfragen des Sprachlebens).3 Wegener uses the key concept of “automatisation”: changed phonetic realisations may appear in hindsight not to fulfil any special goals, but they are in fact the results of conscious choices by speakers, made at one point in time, of alternative possibilities with special communicative effects, which subsequently have become automatised through habit. A communicative basis can also be seen in the ontogeny of language. In time, the child learns to use sound patterns with meanings of a general nature, but this develops only after a stage in which a particular vocal expression is a purposive action, essentially connected to the child’s momentary needs and meant to appeal to the carers for help (Wegener [1885] 1991: 4–19). Wegener goes on to claim that the general purpose of speaking is indeed to influence the actions, volition or thoughts of another person in a way the speaker thinks worthwhile (Wegener 1991: 64–67, 180). He has no use for the well-known subjectivist picture of the individual speaker who, in speaking, merely “mirrors” his own thoughts; such a view would make the idea of communication as accommodation between different minds all but void (Wegener 1902: 402). Focusing rather on the role of the hearer, Wegener emphasises that the hearer is the true locus

Foundations of pragmatics in functional linguistics

231

of meaning, because he constructs meaning first and foremost not from linguistic cues but according to his knowledge of the world and of the purposes of social action, helped further by his inferential abilities. Within this dialogical theory of meaning, the function of the words and grammatical elements the speaker presents is to help organise the hearer’s expectations (Wegener 1991: 157–158, 180–182). The importance of the asymmetrical interaction between speaker and hearer is highlighted in the detailed analysis of many grammatical constructions that Wegener provides in his book. As the general functional structure of utterances, Wegener proposes a bipartition consisting of “exposition” (“Exposition”) and “predicate” (“Prädicat”).4 The “predicate” carries the core function of the utterance, it indicates the new and important material for the situation of utterance, and whatever effect is achieved in the hearer is connected primarily to this part. The “exposition” is everything in the utterance that the speaker has deemed necessary for the hearer to situate the information contained in the predicate in the way it is meant to be situated (Wegener 1911; 1991: 19–22). Elements of the situation itself may provide enough information to let the predicate do its work, as Wegener explains. Any aspects of the hearer’s momentary attention, such as a context of actions that have just been carried out (or are in the process of being carried out), a topic from the social or institutional situation at hand or an element from the general cultural situation, can be sufficient for him to understand an unaccompanied predicate “Well done” as uttered by the speaker. Language thus need not always be “congruous” to what the mind understands (Wegener 1991: 21–27).5 However, an “exposition” of linguistic material is often added to help achieve clarity as to the “situated” fact in relation to which the predicate has to be understood. Note that according to Wegener, the exposition need not always precede the predicate; there may be many reasons for a speaker to choose the reverse ordering. Intonation, but also grammar, can fulfil the task of indicating to the hearer just what part of the utterance has which role. Wegener (1991: 47–60) also relates important phenomena of word semantics to his communicative bipartition of utterances.6 It is not the same for a word to appear in the exposition of an utterance or to appear in the predicate. The words of the exposition are used by the hearer to determine the real-world referent, and their meanings should therefore be relatively “congruous” to the relevant thought. By contrast, words chosen for the predicate may well have a variable, only loosely-fitting relation to their real-world referent and often have a metaphorical character. For his precise understanding, the hearer can rely on inferences that are drawn from elements of the situation and from expositional parts of the utterance. When such inferences become automatised through repeated use, the word meaning loses its flexibility, increases in its “congruousness” to thought and thereby becomes appropriate to fulfil a role also in the expositional parts of utterances. To give a modern example: it is often said, figuratively, that “the family is the cornerstone of society”. The phrase has become so well-known that the original predicate part (“the

232

Saskia Daalder and Andreas Musolff

cornerstone of society”) can now also be used in a role of exposition and still be understood to have the family as referent: “the cornerstone of society is in trouble”. But this does not yet hold unrestrictedly for “cornerstone” on its own; of course, this could well be a further development. The process as sketched amounts to semantic change and, more specifically, the well-attested, continuous “fading” (“abblassen”) of metaphors (Wegener 1991: 51–57). It is an interesting fact that in this picture, pragmatic analysis can explain linguistic phenomena that are usually considered to be semantic or grammatical in nature. Pragmatics could thus be conceived of as an overarching discipline encompassing linguistics – the full potential of such a conception does not seem to be exhausted by modern pragmaticists.7 Although Wegener’s approach was appreciated and taken into account by such functionalist theorists as the important Polish anthropologist Bronisław Malinowski (1884–1942) and the British linguists Alan Gardiner (1879–1963) and Michael Halliday,8 it never gained real notoriety in general linguistics and was not used by the founders of modern linguistic pragmatics. Nor did the “revival” of Wegener’s pioneering ideas in works by the German communication theorists Gerold Ungeheuer (1930–1982) and Johann Juchem (1939–2003)9 succeed in making them known beyond a rather small circle of pragmaticists. However, some of Wegener’s proposals did reach modern pragmatics through an “intermediary”. Karl Bühler helped spreading Wegener’s analysis of the interweaving, in each utterance, of the speaker’s expression of thoughts, the intended effect on the hearer and the representation of a part of reality. This tripartite analysis had a major influence on the development of Bühler’s theory of the functioning of language, as can be seen already in one of his earliest linguistic publications (Bühler 1909b: 119–123).10 It eventually led to Bühler’s model of three language functions, the “organon model”, to be discussed in section 3.1, which in turn was further differentiated by Roman Jakobson and became part of linguistic pragmatics in that extended version (see section 5.1).

3.

Karl Bühler

3.1.

Bühler’s “axiomatic” view of language functions

In 1934, Karl Bühler (1879–1963), Ordinarius professor of psychology at the University of Vienna, published a monograph entitled Theory of Language. The Representational Function of Language (Sprachtheorie. Die Darstellungsfunktion der Sprache). Its introductory First Part deals with “The Principles of Language Studies”, which include four basic “axioms”. The first of these axioms, the “organon-model”, is introduced as the “maxim of the three semantic functions” (Leitsatz von den drei Sinnfunktionen) of all linguistic structures (Bühler 1990: 28; 1934:

Foundations of pragmatics in functional linguistics

233

22).11 Bühler cites Plato’s definition in Kratylos of language as an organum (‘means’/‘tool’) for a speaker to “inform the other of something about the things” (Bühler 1990: 30; 1934: 24) to motivate the term “organon-model”. It was this 1934 version of the model of language “functions” that inspired a number of later scholars to develop the notion of “Functional Linguistics” with a strong emphasis on pragmatic as well as sociolinguistic aspects, and it is quoted as such in textbooks to this day.12 Before we look at its systematic theoretical status and further adaptation, however, we need to analyse its development in Bühler’s own work. 1934 did not mark the starting point of Bühler’s work on “meaning functions” by any means; rather, it was its climax and almost also its ending due to the interruption of Bühler’s work in 1938 when Nazi Germany annexed Austria.13 Building on Edmund Husserl’s (1859–1938), Anton Marty’s (1847–1914) and Wilhelm Wundt’s (1832–1920) theories of meaning,14 as well as on his own work on “thought psychology” in the context of the Würzburg school around Oswald Külpe (1862–1915),15 Bühler had published an article on the “Critical assessment of new sentence theories” (Kritische Musterung der neueren Theorien des Satzes) as early as 1918, in which he proposed a triadic model of language “functions” (Leistungen). It is based on the three main foundations of any meaning-constitutive interaction, i.e., 1) the speaker whose feelings and attitudes are given “expression” (Kundgabe), 2) the hearer for whom it provides a “stimulus” that elicits reactions (Auslösung), 3) objects or states of affairs that form the referents for representation (Darstellung) (Bühler 1918: 16). After moving to the Chair of Psychology at Vienna University in 1922, Bühler went on to develop this triadic model further in two major psychological studies. In The Crisis of Psychology (Die Krise der Psychologie) of 1927, the model serves as the basis for a critical assessment of contemporary psychology. The two basic semantic functions of “expression” (Kundgabe, which, as Bühler emphasises, also includes its receptive counterpart, Kundnahme) and “guidance” (which also involves mutuality) are present even in animal communication; the ability to produce “symbolic representations” (of objects/states of affairs), on the other hand, is characteristic of human communication and helps to adapt the first two functions to a higher level, so that all three of them are only fully activated in human communication (Bühler 1927: 32–33, 37, 50–51). The three functions model thus forms a framework for conceiving of psychology as the integrated study of “subjective experience”, social “behaviour” and “structures of objective sense” (Bühler 1927: 29, 57–62). The enormous conceptual range of this “function”-concept is evident also in the further monograph on Theory of expression (Ausdruckstheorie) of 1933, where Bühler refers to it as the basis to analyse the relationship between miming, gesture and language (Bühler 1933a: iii–iv).

234

Saskia Daalder and Andreas Musolff

With specific respect to linguistics, Bühler explicitly claimed “axiomatic” status for the functions model in his two articles “The whole of language theory, its system and its parts” (Das Ganze der Sprachtheorie, ihr Aufbau und ihre Teile) and “The axiomatics of the language sciences” (Die Axiomatik der Sprachwissenschaften), published in 1932 and 1933, respectively. He now called the functions “expression” (Ausdruck), “appeal” (Appell), and “representation” (Darstellung) (Bühler 1932: 106; 1933a: 74–90) and incorporated them into a set of four principles, which prefigure the “axiomatic” introductory part of Sprachtheorie (although they appear in different order).16 Compared with the earlier versions of the three functions-model, both articles share with the Sprachtheorie the detailed discussion and acknowledgement of Ferdinand de Saussure’s (1857–1913) Cours de linguistique générale (published posthumously in 1916; German translation published in 1931). Saussure’s famous lectures had motivated Bühler to reconceptualise language theory from a “semiological” (Saussure), or as Bühler calls it, “sematological” viewpoint (1990: 33, 42; 1934: 27, 34–35). However, Bühler’s interpretation of the Cours differed in significant ways from contemporary “structuralist” readings.17 Whilst acknowledging the epoch-making impact of the Cours (Bühler 1990: 7–8; 1934: 7), he criticises its presentation of the “speech circuit” as a relapse into 19th century “psycho-physics” and he insists that linguistic communication comprises different types of signs (1990: 31–34, 42–43; 1934: 25–28, 34–37). Most important in our context is his discussion of Saussure’s distinction of a linguistique de la langue from a linguistique de la parole. Bühler puts the emphasis not on the langueaspect, but instead praises Saussure for having shown “what would have to be discovered in order to be really able to initiate a linguistique de la parole” (1990: 8; 1934: 7). This foregrounding of the speech/action aspect of language was at odds with the predominant reading of the Cours as an endorsement of a system-oriented “core linguistics”. It thus provides a first indication of Bühler’s interest in the pragmatic orientation as an essential methodological perspective for linguistic research. Besides the “pragmatically” interpreted Cours, Bühler cites three further works as coming closest to providing a functionalist grounding of linguistics as an urgently needed “complement to the old grammar” (1990: 27; 1934: 23): a) Philipp Wegener’s Investigations into the Fundamental Questions of the Use of Language ([1885] 1991) (discussed in our section 2) had taught Bühler about the listener’s role in the social action dimension of speech;18 b) Karl Brugmann’s (1849–1919) essay on Indo-European pronoun systems (cf. 1904) had provided him with a theory of demonstrative deixis as a framework to analyse the “speech situation”; c) Alan Gardiner’s Theory of Speech and Language (1932), which owed its publication partly to Bühler’s encouragement.19 Bühler praised it as the “most interesting attempt” to carry out a project similar to his own but adds the reservation that Gardiner strives exclusively for a “situational theory of language”, whereas

Foundations of pragmatics in functional linguistics

235

he, Bühler, wants to account also for types of language use that are “removed from the situation” (Bühler 1990: 28; 1934: 34–35).20 Supported by these three authorities (as well as by the parole-oriented reading of Saussure’s Cours), the “organon”-model appears in the Sprachtheorie as the first of four axioms; it postulates three “semantic functions” (semantische Funktionen) for every instance of a “linguistic sign” (Sprachzeichen): Es ist Symbol kraft seiner Zuordnung zu Gegenständen und Sachverhalten, Symptom (Anzeichen, Indicium) kraft seiner Abhängigkeit vom Sender, dessen Innerlichkeit es ausdrückt, und Signal kraft seines Appells an den Hörer, dessen äußeres und inneres Verhalten es steuert wie andere Verkehrszeichen. (Bühler 1934: 28). It is a symbol by virtue of its co-ordination to objects and states of affairs, a symptom (Anzeichen, indicium: index) by virtue of its dependence on the sender, whose inner state it expresses, and a signal by virtue of its appeal to the hearer, whose inner and outer behaviour it directs as do other communicative signs. (Bühler 1990: 35).

It is in this version that the “organon”-model has entered pragmatic-functionalist textbooks as a well-known historical reference (see above). However, its presentation in the Sprachtheorie as the first axiom (possibly meant to underline its preeminence) is less convincing than its introduction in the Axiomatics essay from a year earlier. In that essay it appears as the last axiom and forms the conclusion of the previous argumentation.21 In the Sprachtheorie, the model is followed first by illustrative comments on various styles and registers in which either of the three functional dimensions appears dominant (e.g. poetry, rhetoric, scientific discourse) and a hint that its main thesis “will be verified when all three books [on language] that the organon model requires have been written” (Bühler 1990: 39; 1934: 33). This hint underlines the point of the subtitle of Sprachtheorie, “The Representational Function of Language” (Die Darstellungsfunktion der Sprache): this work does not contain the complete “Theory of Language” whose necessity is implied by the organon model; it still focuses mainly on the one function that had been the object of traditional language philosophy and semantics. This discrepancy, although openly announced, is regrettable from a pragmatic point of view, for historical reasons (as the three promised books were never completed: The Crisis of Psychology and Expression Theory contain valuable linguistic insights but have a much wider, psychologically defined remit) as well as in theoretical respects. As a largely programmatic formulation, the organon model lacks an explication by Bühler that would have clarified its methodological implications. In the Sprachtheorie itself, the model is only intermittently and cursorily referred to and does not provide a systematic structuring principle of the book. The following remarks therefore highlight aspects of the Sprachtheorie that can be most closely linked to the three functions model, without claiming to give a systematic overview of Bühler’s opus magnum. Within the Axiomatics-part of the Sprachtheorie, the second axiom is the “signcharacter of language”, which Bühler takes over not just from Saussure but from

236

Saskia Daalder and Andreas Musolff

the sign-theoretical tradition going back to the medieval scholastic debates on representation and which he relates in a bold move to the contemporary phonemephone distinction. Bühler’s formula for the principle of the semantic character of signs is “abstractive relevance” (Bühler 1990: 40–56; 1934: 33–48). Building on the organon- and sign-axioms, he attempts in axiom III to achieve an integration of functional and systemic language aspects by way of a cross-classification based on the parole-langue and ergon-energeia distinctions by Saussure and Humboldt. This yields four perspectives on language as an object of scientific investigation, i.e. language as (1) “speech action” (Sprechhandlung), (2) “language work” (Sprachwerk), (3) “speech act” (Sprechakt) and “language structure” (Sprachgebilde) (Bühler 1990: 57–80; 1934: 48–69). This schema has been criticised for being terminologically and methodologically inconsistent.22 Its English translation in the 1990 edition adds further problems as it suggests “false friends” such as Sprechakt and speech act. Bühler’s understanding of this aspect is far removed from Austinian, Searlean or Gricean concepts; rather, it rests on the philosophical notion of “sense-conferring acts” (“sinnverleihende Akte”), which Bühler had adopted from Husserl. Bühler praises Husserl’s Logische Untersuchungen for highlighting sense-conferring acts as a necessary object of language theory but he criticises the implication of an exclusively subject-centred perspective on meaning constitution, against which he insists on the inter-subjective side of all use of signs (1990: 73–74; 1934: 63–65; see also Nerlich and Clarke 1996: 194–198, 211–215). The most interesting aspect of Bühler’s schema from a pragmatic perspective is the category of the “speech action”, which encapsulates the insight that “all concrete speech is in vital union with the rest of a person’s meaningful behaviour; it is among actions and is itself an action” (Bühler 1990: 61; 1934: 52). In the discussion in the Sprachtheorie Bühler gives only a relatively brief illustration by way of famous quotations that embody a whole situation (e.g. Caesar’s “alea jacta est”) and refers to developmental psychology; he also highlights the “field”-character of every speech action and the importance of the act(ion) history, which includes the language acquisition process (1990: 65–66; 1934: 56). In the Axiomatics essay, however, Bühler gives a more detailed explanation, drawing strongly on the action theory developed by the fellow Würzburg School psychologist Abraham Anton Grünbaum (1885–1932), who had given a paper on “Language as action” (Sprache als Handlung) at a workshop organised by Bühler at a psychology congress in Hamburg in 1931 (Bühler 1933a: 50–51). For Grünbaum, every “action” is a dynamic system of organism and environment (System: Organismus – Umwelt); its constitutive aspects are the agentive organism’s need (Bedürfnis) and intention (Intention), the active completion (Erledigung) of the action and the agent’s reaction to the completion (Erledigungsreaktion) (Grünbaum 1932: 165–173). For the application to the speech action, Bühler re-labels the “reaction to the completion” as “fulfilment” (Erfüllung) and describes it further as consisting of the speaker’s con-

Foundations of pragmatics in functional linguistics

237

scious realisation that the process of meaning constitution has been concluded (Bühler 1933a: 50–51).23 In the Sprachtheorie, Grünbaum’s theory is not explicitly mentioned but Bühler uses Grünbaum’s terminology in some parts and refers for further reading to his own Axiomatics essay as well as to disputes among the schools of Gestalt theory, where he takes a polemical stance against the supposed “monism” of the Berlin School of Gestalt psychology around Max Wertheimer, Wolfgang Köhler and Kurt Koffka (Bühler 1990: 66; 1934: 56). In this allusion-laden discussion the contours of Bühler’s concept of speech as action are sometimes barely discernible. There is just one hint that the topic is taken up again in chapter 10, which deals with the “integration of speaking into other meaningful behaviour”, and more specifically into different types of “surrounding fields” (Umfelder) (Bühler 1990: 61; 1934: 52). This distance or detour of six chapters is motivated by Bühler’s more immediate focus on the concept of “language fields”, which overlaps with the notion of “surrounding fields”, thus leading to a potentially confusing double-terminology. Within the Axiomatics-part of Sprachtheorie, the field concept comes into focus in the fourth axiom, which deals with the syntax-semantics interface under the heading “Word and Sentence: the S[ymbol]-F[ield]-system of the type language” (Bühler 1990: 81–90; 1934: 69–78), but the concept also dominates the remainder of the monograph. It serves Bühler to gain an overview of language phenomena that ranges from phonology and morphology via syntax and case theory to text theory and stylistics. The principle unifying this theoretical and methodological tour de force is the distinction between, on the one hand, those aspects of linguistic signs that receive their precise meaning in the “deictic field” (Zeigfeld), i.e. the spatio-temporal surroundings, the relationship of speaker and listener, their shared knowledge and action goals, and on the other hand, those aspects of linguistic signs that are fully integrated in syntagms and are detachable from the immediate situation, thus forming a “symbol field” (Symbolfeld).24 As a programmatic outline, this “two-field theory” provides, as Garvin (1994) has pointed out, an alternative to any rigid “syntax-semantics-pragmatics” distinction and has proved to be a constant source of fresh insights. In terms of a theoretical focus, however, Bühler’s wide-ranging discussion covers too many data and theoretical aspects and the important insights from the “organon model” and the “speech action” concept are taken up only intermittently in the “field”-theoretical chapters. In the following section, we highlight these pragmatically relevant theory aspects without claiming to provide a comprehensive reconstruction of Bühler’s general field-theoretical framework.

238 3.2.

Saskia Daalder and Andreas Musolff

Linguistic deixis and action

The two main concepts in the Sprachtheorie that are of relevance for linguistic pragmatics are those of the “deictic field” and the “surrounding fields”, i.e. the context in which speech is meaningful as social action. Terminologically, both categories are evidently strongly integrated in Bühler’s overarching “field”-theory, but conceptually they also depend on the three-functions model. In both cases, the multi-functionality of language and the fundamental importance of the speakerhearer relationship, which derive from the “organon model”, are constitutive of Bühler’s argumentation. The origin of linguistic indexicality/deixis lies in the I-now-here-position of the speaker, which serves as the starting point for spatial, temporal and social orientation and co-ordination between the communication partners (Bühler 1990: 117–136; 1934: 102–120). The interlocutors’ shared horizon can then be extended through the quasi-figurative use of the co-ordinates of the demonstratio ad oculos (“visual [or other sensory] demonstration”) for the purpose of “imaginationoriented deixis”. When referring to objects that are not immediately “given” in the discourse situation, speaker and hearer must either project them linguistically into that situational context (e.g. by putting a problem “before” themselves) or, vice versa, project their own situation co-ordinates onto an imagined context, or use a mixture of both. Such “imagination-oriented deixis” is present already in everyday language use and marks the first step in the gradual emancipation of language from the immediate situational context (Bühler 1990: 158–166; 1934: 140–148). The highest developmental “stage” of deictic language use is reached when indexical language signs are used anaphorically and cataphorically as “joints of speech”, e.g. as relative pronouns, text-deictic prepositions, prepositional adverbs or conjunctions (Bühler 1990: 438–452; 1934: 385–397). Here, the symbol-field is reconstructed along deictic co-ordinates; at the same time, deictic signs are used to “demonstrate” relationships within a representational context.25 The second main pragmatically interesting discussion in the Sprachtheorie, which Bühler had already flagged up in the context of the Axiomatics section when dealing with the “speech action”, is chapter 10. Its title provides an additional typology of “surrounding” fields: “The sympractical, the symphysical and the synsemantic field of Language Signs” (1990: 175; 1934: 154). Again, as in the case of the deixis-symbolising continuum, Bühler conceptualises the field-series as a cline of abstraction. Linguistic signs can be tied closely to a specific situation or specific objects (= the sympractical and symphysical fields) and so remain situation-bound. The more strongly they are integrated into the linguistic “con-text” (= the synsemantic field) however, the more “liberated” they are from the specific circumstances of the speech situation. In some genres, e.g. algebraic formula systems, the liberation from concrete, situation-related content is almost taken to the maximum (cf. chapters 11, 25 of Sprachtheorie). The overlap between the “synsemantic” field

Foundations of pragmatics in functional linguistics

239

and the symbol field (Bühler 1990: 179–189; 1934: 159–168) blurs the boundaries between the two “field-theoretical” concepts to some extent,26 which may be indicative of Bühler’s difficulty to combine his focus on functional aspects of language with the perceived necessity of integrating these insights into the predominant contemporary tendencies of explaining language mainly as a system of representation. In the context of the “surrounding fields”, Bühler comes closest to a pragmatic interpretation of language data in the passages dealing with the “sympractical field”, which presuppose his three-functions axiom and Grünbaum’s “action field” theory (see above). The sympractical field includes what would be called in modern terminology the “situational context” (as opposed to “co-text”), which communication partners need to know (and share) in order to be able to disambiguate situation-bound utterances such as short commands, requests, greetings etc. Such utterances had been viewed traditionally as deficient, on account of not being syntactically “well-formed” sentences (Bühler 1990: 176–179, 187–189; 1934: 156–159, 166–167).27 Through examples such as that of a patron in a coffee house asking for a particular type of coffee by way of uttering something like “an espresso, please”28 and of a passenger in a tram car choosing between types of tickets, Bühler explains that an “ellipsis” interpretation is mistaken. In such cases, the speaker does not have to produce semantically explicit or grammatically wellformed utterances, because the interlocutor, on account of shared knowledge of the situation (the Umfeld), understands motives and plans for the action (in modern terminology: the action “schema” or “script”) and uses the language sign only as a diacritic. The situational context provides a sufficient “fulfilment” of all meaning functions in the action field: the utterance is not elliptical but rather, in Bühler’s terminology, “empractical” (1990: 177–179; 1934: 155–158). Generalizing from Bühler’s hints, one can argue that language use is empractical in principle, insofar as it is always aimed at resolving ambiguities of intersubjective co-operation and based on some degree of shared background knowledge (or at least, background assumptions). Whilst Bühler never formulated any communicative “co-operation” or “relevance” principles in the Sprachtheorie, his notion of the sympractical field as the basic action-structure in which language functions operate can be viewed at least as compatible with more recent pragmatic and relevance-theoretical approaches.29 4.

Vilém Mathesius and the Prague Linguistic Circle

4.1.

Introduction

From 1930 onward, Bühler had scattered connections with members of the Prague Linguistic Circle, the avant-garde group of linguists that propagated a view of language as a “functional system” with “structural laws” (Prague Linguistic Circle

240

Saskia Daalder and Andreas Musolff

[1929] 1983: 77–79). Founded by the Czech linguist and scholar of English Vilém Mathesius (1882–1945) in 1926, the Circle united under Mathesius’s presidency some of his Czechoslovak colleagues and linguists of other nationalities who lived in Prague or nearby, notably a number of Russian and Ukrainian émigrés. Whereas the “functional” point of view was the original contribution of Mathesius himself, Roman Jakobson, the Russian vice-president of the Circle, added a “structuralist”, much more system-oriented approach. Not surprisingly, Mathesius’s and Jakobson’s legacies to pragmatics vary. The following section will detail Mathesius’s functionalist findings that are relevant for pragmatics; in section 5, we will turn to foundations of pragmatics deriving from Jakobson’s work. 4.2.

Mathesius’s functional view of language

Since his earliest linguistic writings in the first decade of the century, Mathesius had argued for a view of language as a means to an end.30 In the years of his university education in Prague, his general methodological outlook was formed by the call for “concreteness” from the Czech philosopher Tomásˇ Masaryk (1850–1937), who coupled this appeal with the desirability of democratising scientific and philosophical education (Masaryk 1887). Science had to give priority to the study of “close and accessible phenomena”, including their comparative analysis; this would prepare for historical research at a later stage.31 Accordingly, Mathesius sought his linguistic inspirations in the works of opponents of Neogrammarian dogmatism, such as Sweet, Otto Jespersen (1860–1943), von der Gabelentz and Wegener (Mathesius [1936] 1966: 137), and he became an advocate of synchronic language study. The ideal of “concreteness” took shape in the emphasis which Mathesius laid on the inherent synchronic variability of linguistic phenomena among individuals of a language community as well as in the speech of a single individual; he spoke of the “static oscillation” in speech and called this aspect of language its “potentiality” (Mathesius [1911] 1983a: 3). He stressed that the seeming homogeneity of language is a product of the method of analysis and that without the phonetic phenomenon of “synchronic oscillation”, languages could not change (Mathesius [1911] 1983a: 4, 32). These claims sound very modern and they were indeed approvingly cited in the influential paper on variability and language change by Weinreich, Labov and Herzog (1968: 167–169), although these authors also criticised Mathesius for not having pointed to the sociolinguistic regularities that are inherent in variation. In his ideas on “synchronic oscillation” in the domain of semantics, Mathesius was at first strongly influenced by the stylistics of the Geneva linguist Charles Bally (1865–1947) as well as by the “idealistic” approach of the German scholar of Romance languages Karl Vossler (1872–1949). This stance implied an emphatic interest in the expression of the speaking individual rather than in the social char-

Foundations of pragmatics in functional linguistics

241

acter of language, which he saw as having been overstated by nineteenth-century linguistics (Mathesius [1911] 1983a: 29–31, 35–36; [1926] 1983b: 52–53). In his later work, Mathesius redressed the balance to some extent by way of an acknowledgement of the dominance, at least for language change, of the communicative function of language vis-à-vis the “original” expressive function (Mathesius [1926] 1983b: 54–55). However, on account of his continued interest in stylistics he kept a predilection for the expressive function and defined his “functional” point of view as one that “takes the meaning or function as its starting point and tries to find out by which means it is expressed” (Mathesius [1926] 1983b: 56–58; also [1929] 1983c: 122–125). Mathesius thus came to neglect some of the necessary focus on the hearer as the key player in the constitution of meaning, something which Wegener and Bühler had already reached in their work. This has to be viewed as a general restriction of Praguian functional linguistics in its foundational role for pragmatics. Mathesius’s emphasis on the expressive function is understandable in the light of his long-standing interest in the comparative study of word order patterns, which was first kindled by the work of the French scholar of Ancient Greek Henri Weil (1818–1909) (Weil [1844] 1887). Confronted with differences in word order between Czech and English, Mathesius saw languages of different “character”. Where a speaker of Czech usually filled the slot of the grammatical subject with the agent of the action expressed by the verb, an English speaker appeared to prefer the “theme” (“Thema”) of the utterance for this role. The Czech speaker therefore was relatively free to order his sentence parts according to their contextual status of “theme” or “nucleus” (“rheme”). For the English speaker to have a similar possibility readily at hand, the language had developed grammatical means for alternative choices of the grammatical subject, i.e. several types of passive constructions, which could be observed to make out a considerable proportion of English verbal constructions (Mathesius 1929: 202–203; [1929] 1983c: 126–128, 133– 134; 1930). The “character” of a language was thought to explain the links between such descriptive facts of different types. Mathesius’s “characterology” (or: language typology) thus relied on what he called the “actual articulation”32 of utterances, which was to be distinguished from its logico-grammatical articulation. The notion of “actual articulation” pointed to the inherent “situatedness” of the structure in a specified context, and to its “concrete” quality; in short: it was a pragmatic structuring. Specifically, it consisted of a “basis” or “theme” on the one hand and a “nucleus”, “enunciation”, or “rheme” on the other.33 In a formulation from Mathesius (1975a: 81): “The element about which something is stated may be said to be the basis of the utterance or the theme, and what is stated about the basis is the nucleus of the utterance or the rheme.” To an extent, Mathesius’s proposal of a “functional sentence perspective” resembles Wegener’s ideas on the bipartition of utterances in “exposition” and “predicate” that we discussed in section 2. Still, we note the less pronounced com-

242

Saskia Daalder and Andreas Musolff

municative emphasis with Mathesius: the hearer is absent in his formulation, and even the speaker is no more than implied. Furthermore, Mathesius’s insight in the role of the situation is much less developed. In any case, there appear to be no references by Mathesius to Wegener’s theories on this subject, at least not from the period in which Mathesius developed his notion of “functional sentence structure” (the late 1920s). We have to decide that the relation is probably not a direct one. It is certain however that Mathesius’s notions found favour in many quarters of (pragmatic) linguistics and have become, under some of their original names or as “topic” and “comment”, classics of pragmatic analysis (see, e.g., Lyons 1977: 501–512; Levinson 1983: 88–89).34

5.

Roman Jakobson

5.1.

Jakobson’s model of linguistic functions

The Russian general linguist and scholar of Slavonic languages Roman Jakobson (1896–1982) was intrigued by the seemingly autonomous functionality (“teleology”, “self-movement”) of language change all his life (Liberman 1987). For a long time, Jakobson hoped to be able to account for it with the help of a Hegelian logic of “system change” (Toman 1995: 171–176). But in the milieu of the Prague Linguistic Circle, he gradually came to recognise that the stages of a change in progress had in fact to be conceived as different simultaneous facts, or styles, within the language system. Especially after Jakobson’s emigration to the United States in 1941, his teleological point of view shifted to a focus on the less mysterious functions in a “means-ends model” of a synchronic “system of systems” (Jakobson 1963: 107). In his new homeland, Jakobson found the study of language to have been, as he remarked at a conference of anthropologists and linguists in 1952, “strongly bulwarked by the impressive achievement of two conjoined disciplines – the mathematical theory of communication and information theory” (Jakobson 1953: 12). This led him to discuss the factors of the communication process that were treated, e.g., in the well-known publication by Shannon and Weaver (1949). In his paper, Jakobson enumerates the four factors connected to the message in the information theorists’ terminology (sender, receiver, topic of the message, code) and points to the variability of their relative importance for a message. Apart from being associated with the factor of topic (information), a message may also be “oriented” toward the sender or the receiver, leading to a “function” of the message different from a purely informative one, or, as the case may be, a hierarchical bundle of several functions (Jakobson 1953: 13). These three basic function varieties can be considered equivalent to Bühler’s (1990, 1934) functions of representation, expression and appeal, in the three-functions model that Jakobson subsequently

Foundations of pragmatics in functional linguistics

243

refers to as the “traditional model of language” (Jakobson 1960: 355; [1976] 1985: 115).35 It should be kept in mind however that Jakobson views the language functions as systemic facts throughout (facts of a “system of systems”). This distinguishes his theory from Bühler’s approach (recall Bühler’s championing of parole) and even more from Wegener’s; indeed, we are not aware of any references to Wegener in Jakobson’s work. Jakobson had come to linguistics via the study of Russian verse and of poetics in general (cf. Rudy 1978; Toman 1995; Waugh and Rudy 1998: 2262–2265). Together with his compatriot, the literary scholar Jurij Tynjanov (1894–1943), he had made a case for a parallel development of language, literature and other cultural systems by structural necessity; there was to be no separation of the interconnected codes of language and literature (Tynjanov and Jakobson ([1928] 1972). In a section of the 1929 “Theses” of the Prague Linguistic Circle, Jakobson and the Czech literary scholar Jan Mukarˇovsky´ (1891–1975)36 had sought to contrast poetic language with “communicative” language. They had stated that “poetic expression is directed towards the way of expression itself” and that the linguistic means of different levels, which in communicative speech tend to become automatised, in poetic utterances tend to become foregrounded themselves (Prague Linguistic Circle [1929] 1983: 94; cf. also 89). It is this idea that Jakobson takes up when he, in his paper from 1953, adds a fourth, “poetic” function of language to Bühler’s set of three. For a message that is partly loosened from its referential meaning by being oriented towards “its own self”, towards the “palpability” of its signs, the poetic function of language is at work, be it as the predominant function of a literary fragment or as a subordinate, additional function of an otherwise “prosaic” message (Jakobson 1953: 14). Jakobson mentions also a fifth possibility: an orientation of the message toward the factor of code, interesting because here the code (i.e., the langue) is not just an abstraction but interacts with language use. The orientational “set” (explained as “Einstellung”) for the code is witnessed in non-specialist discourse about such linguistic subjects as clarification of code elements, code switching, bilingualism, areal diffusion, language variation and language change (Jakobson 1953: 14–18). This function is described with more clarity and termed the “metalingual” function in a publication based on Jakobson’s presidential address at a meeting of the Linguistic Society of America in 1956 (cf. Jakobson 1985: 116–121). A sixth function is mentioned in that address as well: Jakobson analyses anthropologist Malinowski’s “phatic communion” function37 as being attached to messages with a set for the factor of “contact”, or “channel”, as when interlocutors, without much information transfer, check or revive the physical communication channel, or keep it going (Jakobson 1985: 115).38 Jakobson (1960: 353–357) sums up the resulting model of six language functions (with some terminological changes): the referential (denotative, cognitive) function, the emotive (expressive) function, the conative39 function, the poetic

244

Saskia Daalder and Andreas Musolff

function, the metalingual (or “glossing”) function and the phatic function. They are explained by the orientation of the message toward one of the six constitutive factors of speech events, to wit the context referred to (the referent), the addresser, the addressee, the message, the code, and the contact channel. It is in this version that Jakobson’s model of language functions has been widely cited by linguists of several persuasions, including theoreticians of linguistic pragmatics (see, e.g., Hymes [1962] 1968: 99, 110–124; Hymes 1983: 334; Lyons 1977: 52–56; Levinson 1983: 41; Schiffrin 1994: 32–33; Duranti 1997: 284–287). It is clear, however, that the model is focused on language as an undivided phenomenon, in which, just as in Bühler’s approach, the many facets of a pragmatic nature remain “inclusive”, and therefore partly implicit. In order to gain an unimpeded view of pragmatic facts, a systematic linguistic pragmatics had still to be drawn out. 5.2.

Jakobson’s treatment of deixis

Among the further aspects of Jakobson’s functionalism that are relevant for the development of linguistic pragmatics, his view of deictic phenomena is the most important. In a publication of 1957, which goes back to a paper delivered in Geneva in 1950 (cf. Jakobson 1950), he treats the differentia specifica of indexical items such as the personal pronouns I and you (Jakobson [1957] 1971). Like Bühler, Jakobson points to the fact that their precise meaning in context needs to take into account aspects of the situation, always unique, in which they happen to be used; adopting Otto Jespersen’s (1860–1943) (1922: 123) term, he calls them “shifters” (Jakobson 1971: 131). Jakobson’s main interest however is focused on the question of the status of these shifters as elements of the linguistic code, and he insists that their symbolic meaning (for I, “the addresser of the pronoun instance”, and for you, “the addressee”) should not be neglected either. Using Charles Sanders Peirce’s (1839–1914) classification of sign modes and specifically following its interpretation by the American mathematician and philosopher Arthur Burks (1915–2008), Jakobson argues in favour of according them the status of “indexical symbols”.40 First and second person pronouns in use are associated with the object they represent by an “existential” relation (i.e., pronoun token and object exist in the same situation), therefore they indeed function as “indices”, just like acts of pointing with the index finger would. However, on account of their “conventional” relation with the role of speaker or hearer, which in different language codes holds for different forms (Latin ego, German ich, Russian ja), they partake just as well of the symbolic mode of signs (Jakobson 1971: 132).41 Whereas the sober-minded Jespersen (1922: 123), in reaction to a remark by the German idealist philosopher of subjectivity Johann Gottlieb Fichte (1762–1814), had belittled attempts to derive deep philosophical conclusions from the “little linguistic trick” involved in using the pronoun I correctly, Jakobson goes on to develop its theoretical significance further. Considering the constitutive fac-

Foundations of pragmatics in functional linguistics

245

tors of speech situations, he notices for the shifters an “overlapping” of the factors Code and Message: their general meaning, which is part of the Code, contains a reference to the very Message of which the shifter in an instance of its use happens to form a part (Jakobson 1971: 132). Jakobson explains this possibility by asserting that both Code and Message can in principle refer as well as be an object of reference; in the specific case of the shifters, the Code has the function of referring, while the Message is being referred to.42 This explanation takes up a similar remark by the French linguist Émile Benveniste (1902–1976) in an article about pronouns that was published a year earlier in the Festschrift For Roman Jakobson (Benveniste [1956] 1966a: 252). Benveniste, likewise anxious to find an aspect of the Saussurean langue (the Code) within the typical discourse (Message) reality of first and second person pronouns, goes further than Jakobson in postulating that it is precisely the use of elements symbolizing the roles of utterer-conceptualiser (énonciateur) and “allocutee” of a discourse event which leads to language users’ consciousness of their role as “subjects”. Those symbolic features of the language system then get parcelled out into psychological reality. Benveniste also mentions that the process is strengthened by various verbal forms that align with the perspective of the utterer of the discourse (Benveniste 1966a: 254–255; also [1958] 1966b: 258–263). In his 1957 article, Jakobson indeed uses the theory of the shifters to set up a new classification of verbal categories, in particular those of Russian. Like first and second person pronouns, verbal shifter categories (several forms in the paradigms of person, mood, tense and evidentiality) have an encoded meaning in which the speech event (the message) and/or one of its participants play a decisive role (Jakobson 1971: 133–147).43 The relevance of the idea of the interweaving of Code (langue) and Message (instance of discourse) to linguistic pragmatics becomes evident when the phenomenon of the shifters is described as a fact about speech events rather than the linguistic code. It can be said then that by means of various indexical elements of the language, speech events often encode, or “index”, important components of the community’s social life. It was the American anthropologist Dell Hymes (1927–2009) who realised that on this basis, Jakobson’s speech functions could be extended and seen to be differently organised for different communities, opening up “an additional realm of structure” (Hymes 1983: 335). This was to result in the ethnography of speaking and more broadly, the ethnography of communication (Hymes [1962] 1968: 110–129; 1964; 1972). 6.

Conclusion

Philipp Wegener, Karl Bühler, Vilém Mathesius and Roman Jakobson – all of them characteristically “modernist” theorists – were fully aware of the pervasive relevance of language in countless domains of human life and culture. They projected

246

Saskia Daalder and Andreas Musolff

the pragmatic aspects which they had identified, such as the multifunctionality of language, the functional structure of the sentence and the nature of deixis, onto a rich picture of language that did not separate semantic from pragmatic phenomena and often contained no fixed answer to the question as to what is language “itself” and what is its use. Seen from a present-day perspective, their works may appear to be lacking in systematic treatment of pragmatic facts. By the same token, however, they have had – and continue to have – potential not for just one but many branches of linguistic pragmatics. Wegener’s ideas for the linguistic analysis of utterances together with their embedding situations may still be used to further loosen the grip of the received wisdom about the structure of linguistic utterances as phenomenal entities per se. Similarly, Bühler’s and Jakobson’s theories of the functions of language, the associated models of the communication situation and the deictic signs/ “shifters” that derive their specific meaning from it offer the chance to review the “immunity” of the theory of the language system to results of the analysis of language use.

Notes 1. Outside the discipline of linguistics, Bühler’s model has been taken up by Jürgen Habermas, who in his theory of communicative action (cf. 1981) uses the three functional dimensions as reference points for a model of social communication. According to Habermas, every speaker, when making a meaningful utterance, intends it to be understood by the hearer as legitimate (within the social context), truthful (i.e. as an expression of the speaker’s own convictions) and true (in relation to objective states of affairs). Even though these assumptions are often disappointed in actual communication, they act as regulating norms for meaningful social exchange (Habermas 1981, vol. 1: 372–377, 412). See further the article by Cooke in this volume. 2. For information on Wegener and his work, see especially Knobloch (1991); also Juchem (1984, 1986), Knobloch (1988: 292–297), Nerlich (1990: 153–192; 1992: 81–87), Graffi (1991: 87–89), Elffers (1993), Nerlich and Clarke (1996: 177–183), Morpurgo Davies (1998: 318–320), Grimm-Vogel (1998), Tenchini (2008). 3. An English translation of Wegener’s 1885 book appeared as part of Abse (1971); note that Knobloch (1991: xliv) expresses a caveat about details of the translation. 4. In his publication from 1885, Wegener uses the term “logical predicate”; in later works he switches to “psychological predicate” and “statement” (“Mitteilung”). 5. Knobloch (1991: xvi, xxxii) mentions Lazarus ([1856–1857] 1884) as a possible source of Wegener’s ideas in this area. About Moritz Lazarus (1824–1903), see Knobloch (1988: 268–273, 411–422) and Nerlich and Clarke (1996: 165–168). 6. For details, see Hülzer (1987), Knobloch (1988: 296–297; 1991: xviii-xix), Nerlich (1990: 181–185; 1992: 86–87). 7. This is the opinion also of Knobloch (1988: 297) and Nerlich (1992: 87). 8. See the references to Wegener’s work in Malinowski ([1923] 1946: 297), Gardiner (1932: passim) and Halliday (1978: 109), particularly in relation to the notion of “situation”.

Foundations of pragmatics in functional linguistics

247

9. See Ungeheuer (1987, 1990) and Juchem (1984, 1986). 10. For details see Knobloch (1991: xxxix-xli). 11. Quotations from Bühler’s works are referenced, where possible, to both the original German publications and their English translation, i.e., in the case of the Sprachtheorie, 1934 for the German volume, 1990 for its translation. 12. See e.g. Schlieben-Lange (1979: 13); Levinson (1983: 41, 61); Brown and Yule (1983: 1): Coupland (2007: 12). 13. After the Sprachtheorie (1934), Bühler published several other linguistic articles up until 1938 (Bühler 1935, 1936a, 1936b, 1938). Immediately after the occupation of Vienna by German troops, Bühler was imprisoned by the Gestapo. His wife and academic collaborator, Charlotte Bühler, managed to effect his release and to organise their emigration to the USA, but the continuation of their careers proved difficult: they only gained posts as visiting or assistant professors and did not return to Europe after the war (see Garvin 1966, Ch. Bühler 1984, Musolff 1997). The 1960 article “Von den Sinnfunktionen der Sprachgebilde” (Of the Semantic functions of linguistic structures) is an excerpt from Sprachtheorie and of dubious editorial status (e.g. it cites the “Organonmodell” as “Organmodell”). 14 Husserl (1900–1901); Wundt (1900); Marty (1908). Bühler reviewed Marty’s work with specific emphasis on the functional typology, see Bühler (1909a: 964–946). 15. For Bühler’s experiments on sentence comprehension in the context of “Thought Psychology”, see Bühler (1908), Wundt (1908), Wettersten (1988). 16. For the changes in the “functions”-terminology and the status and ordering of the axioms, see Wunderlich (1969); Kamp (1977: passim); Beck (1980: 169–192); Busse (1975: 213–215, 222–229); Camhy (1980a: 92–94); Swiggers (1981: 54–55); Innis (1982: 3–10); Eschbach (1984b: 91–93); Henzler (1987: 354–356); Knobloch (1988: 424–430); Graumann (1988); Musolff (1990: 26–31); Vonk (1992: passim); Elffers (2005; 2008: 21–24). 17. For the contemporary context of the Saussure-reception in relation to Bühler see Scheerer (1980: 36–40), Koerner (1984), Musolff (1990: 40–43) and Ehlich (1996). 18. Details of what Bühler had learnt from Wegener can be found in Bühler (1909b: 119–123). 19. See Gardiner (1932: viii). 20. For detailed discussion of these and other pragmatic-functional influences cf. Graumann (1984), Herrmann (1984), Knobloch and Schallenberger (1993), Musolff (1993) and Nerlich and Clarke (1996: 182–183, 226–236, 341–343). 21. See Bühler (1933b: 90); in Ströker’s edition (1976: 116–117). 22. See Laziczius (1939); Lohmann (1942–1943); Camhy (1980b); Ortner (1986). 23. For detailed discussion of Grünbaum’s influence on Bühler see Knobloch and Schallenberger (1993: 85–90). 24. For Bühler’s field-theoretical approach (as opposed to the contemporary notion of “semantic fields”) cf. Wolf (1932); Heger (1984); Musolff (1990: 61–120); Garvin (1994). 25. For Bühler’s influence on modern pragmatically oriented studies of deixis see Hörmann (1978: 394–424); Weissenborn (1988); Conte et al. (1989); Innis (1992: 556–557); Ehlich (1996: 959). 26. Garvin (1994: 60–64) interprets the relationship as a superordination of the more general “surrounding field” over the two subcategories of “symbolic” and “deictic field”; he

248

27. 28. 29.

30.

31. 32.

33.

34. 35.

36. 37. 38.

39. 40.

Saskia Daalder and Andreas Musolff

bases this reading on an excellent comparison with the contemporary gestalt-psychological theories, which Bühler himself highlighted (1990: 175–176, 1934: 154–155) as well as with early semiotics (Charles Morris). However, in the Sprachtheorie this classification is not explicated nor consistently followed – it constitutes a further theoretical (and highly plausible) development, the implications of which are still to be worked out in detail. For critical assessments see Klein (1984), Ortner (1983). Living in Vienna at the time, Bühler of course chose an appropriate local variety (einen schwarzen ‘a black one’). But compare Nerlich and Clarke’s (1996: 236–237) critical comment that Bühler, despite being one of the most important figures among language researchers engaged in “pragmatics avant la lettre”, “never crossed the last hurdle on his way to a fully developed theory of linguistic pragmatics”, especially on account of his “[clinging] too much to his three functions of language, overlooking to some extent the polyfunctionality of speech acts, rediscovered by Wittgenstein”. Toman (1995: 77–79). About Mathesius’s work, see further Trnka ([1946] 1966), Vachek (1982), Danesˇ (1987, 2003), Fronek (1988), Graffi (1991: 179–183, 249–250), Toman (1995: 71–86, 97–101 and passim), Lesˇka (1995), Nekula (1999). See Toman (1995: 84–86, 97–101), also Fronek (1988). This notion features in the original title of the Czech article Mathesius ([1939] 1975b); see the translator’s remark in Mathesius (1975b: 479). It came to be translated as “information-bearing structure of the sentence”, shortened to “information structure” and alternating with “functional sentence perspective” (“FSP”). The terminology differs; see Mathesius ([1929] 1983c: 126–128; 1929: 202; 1930: 58). The term “rheme” came into use only later, through the writings of the Brno scholar of English Jan Firbas (1921–2000). See the explanation by Vachek in Mathesius (1975a: 185 n71), further Vachek (1982: 123–124) and Danesˇ (1987: 23–26; 2003: 40–41). The Prague version of the information structure of the utterance has remained a distinctive one. See a short description in Hajicˇová (1994). Jakobson referred to Bühler’s three functions already in 1932 (Jakobson [1932] 1966: 27–28). In its “Theses” of again some years earlier, the Prague Linguistic Circle had distinguished, under the heading “On the functions of langugage”, “intellectual speech” from two types of “emotional speech” and had mentioned the different “hierarchy” of functions for each utterance (Prague Linguistic Circle [1929] 1983: 88–89). Jakobson probably wrote (part of) this section (nr. 3a) of the “Theses” (cf. Toman 1995: 289 n13). For the attribution to Jakobson and Mukarˇovsky´, see again Toman (1995: 289 n13). See Malinowski ([1923] 1946: 315–316). Cf. Nekula and Ehlers (1996: 190, 192) on the at first marginal place of the “phatic” function of language in Jakobson’s model. On Jakobson’s model of language functions, see Holenstein (1976: 153–164); Waugh (1976: 25–26); Waugh and Monville-Burston (1990: 15–16); Nekula and Ehlers (1996); Nerlich and Clarke (1996: 284–286); Waugh and Rudy (1998: 2260–2261). “Conative”: with the notion of “striving” to effect some act by the addressee; cf. Jakobson ([1932] 1966: 22, 27). See Burks (1949). Among the signs having the property of indexicality, Burks views the “indexical symbol” as the most fundamental type; a “pure” index cannot in effect point to a well-defined object and is therefore a less fundamental type of sign (Burks 1949: 678–688).

Foundations of pragmatics in functional linguistics

249

41. Jakobson’s (1971: 132) statement to the effect that Bühler acknowledged the indexical side of personal pronouns only seems not fully justified. It is when Bühler introduces the indexical field (Zeigfeld) that he contrasts indexical and appellative words as to the “field” in which, specifically, their meaning “fulfilment” (Bedeutungspräzision, Bedeutungserfüllung, see above, 2.2) is located: in the indexical field and the symbol-field of an utterance, respectively (Bühler 1990: 91–94; 1934: 78–81). This does not imply that personal pronouns would not have a symbolic meaning aspect to begin with, such as “the sender of the word in question” (Sender des aktuellen Wortes) (Bühler 1990: 91; 1934: 78); indeed, the notion of meaning fulfilment presupposes the opposite idea. 42. Considering the three other possibilities of combining the choices of Code/Message and referring/referred-to item, Jakobson theorises further in the following way. When the functions are turned round and it is rather the Message that refers and the Code that is referred to, we have to do with the elucidation of words and sentences (what would be named the “metalingual” function of language a few years later); when a Message refers to a Message, the result is the well-known phenomenon of reported speech; and lastly, with “Code referring to Code” we could describe the general nature of the meaning of proper names, Jakobson holds (Jakobson 1971: 130–131). 43. On Jakobson’s notion of “shifters”, see Holenstein (1976: 158, 162–163); Waugh (1976: 24–25, 32, 48, 52); Waugh and Monville-Burston (1990: 17–19); Nerlich and Clarke (1996: 282–284).

References Abse, D. Wilfred 1971 Speech and Reason: Language Disorder in Mental Disease. With a Translation of “The Life of Speech” by Philipp Wegener. Charlottesville, VA: University of Virginia Press. Beck, Götz 1980 Sprechakte und Sprachfunktionen: Untersuchungen zur Handlungsstruktur der Sprache und ihren Grenzen. Tübingen: Niemeyer. Benveniste, Émile 1966a La nature des pronoms. In: Émile Benveniste, Problèmes de linguistique générale, tome 1, 251–257. Paris: Gallimard. First published The Hague: Mouton [1956]. Benveniste, Émile 1966b De la subjectivité dans le langage. In: Émile Benveniste, Problèmes de linguistique générale, tome 1, 258–266. Paris: Gallimard. First published Paris: PUF [1958]. Brown, Gillian and George Yule 1983 Discourse Analysis. Cambridge: Cambridge University Press. Brugmann, Karl 1904 Die Demonstrativpronomina der indogermanischen Sprachen: Eine bedeutungsgeschichtliche Untersuchung. In: Berichte über die Verhandlungen der Königlichen sächsischen Gesellschaft der Wissenschaften zu Leipzig, Philologisch-Historische Classe, No. 6: 3–151.

250

Saskia Daalder and Andreas Musolff

Bühler, Charlotte 1984 Karl Bühler: Eine biographische Skizze. Transl. by Achim Eschbach. In: Achim Eschbach (ed.), Bühler-Studien, Vol. 1, 25–30. Frankfurt am Main: Suhrkamp. Bühler, Karl 1908 Tatsachen und Probleme zu einer Psychologie der Denkvorgänge: II) Über Gedankenzusammenhänge. Archiv für die gesamte Psychologie 12: 1–23; III) Über Gedankenerinnerungen. Archiv für die gesamte Psychologie 12: 24–92. Bühler, Karl 1909a Review of Anton Marty, Untersuchungen zur Grundlegung der allgemeinen Grammatik und Sprachphilosophie (1908). Göttingische gelehrte Anzeigen 171: 947–979. Bühler, Karl 1909b Über das Sprachverständnis vom Standpunkt der Normalpsychologie aus. In: Bericht über den 3. Kongreß für experimentelle Psychologie (Frankfurt 1908), 94–130. Leipzig: Wilhelm Engelmann. Bühler, Karl 1918 Kritische Musterung der neuern Theorien des Satzes. Indogermanisches Jahrbuch 6: 1–20. Bühler, Karl 1927 Die Krise der Psychologie. Jena: J.A. Barth. Bühler, Karl 1931 Phonetik und Phonologie. Travaux du Cercle Linguistique de Prague 4: 22–53. Bühler, Karl 1932 Das Ganze der Sprachtheorie, ihr Aufbau und ihre Teile. In: Gustav Kafka (ed.), Bericht über den 12. Kongreß der Deutschen Gesellschaft für Psychologie (Hamburg 1931), 95–122. Jena: G. Fischer. Bühler, Karl 1933a Ausdruckstheorie: Das System an der Geschichte aufgezeigt. Jena: G. Fischer. Bühler, Karl 1933b Die Axiomatik der Sprachwissenschaften. Kant-Studien 38: 19–90. (Reprinted 1969 and 1976 under the same title in a truncated version without the introductory chapter, edited by Elisabeth Ströker, 25–117). Frankfurt am Main: Klostermann. Bühler, Karl 1934 Sprachtheorie: Die Darstellungsfunktion der Sprache. Jena: G. Fischer. Bühler, Karl 1935 Forschungen zur Sprachtheorie: Einleitung. Archiv für die gesamte Psychologie 94: 401–412. Bühler, Karl 1936a Psychologie der Phoneme. In: Daniel Jones and Dennis B. Fry (eds.), Proceedings of the Second International Congress of Phonetic Sciences, 162–169. Cambridge: Cambridge University Press. Bühler, Karl 1936b Das Strukturmodell der Sprache. Travaux du Cercle Linguistique de Prague 6: 3–12.

Foundations of pragmatics in functional linguistics

251

Bühler, Karl 1938 Der dritte Hauptsatz der Sprachtheorie: Anschauung und Begriff im Sprechverkehr. In: Henri Piéron and Emile Meyerson (eds.), 11e Congrès International de Psychologie (Paris 1937). Paris: Rapports et Comptes Rendus, 196–203. Bühler, Karl 1960 Von den Sinnfunktionen der Sprachgebilde. In: Richard Wisser (ed.), Sinn und Sein: Ein philosophisches Symposion; 95–112. Tübingen: Niemeyer: Bühler, Karl 1990 Theory of Language: The Representational Function of Language. Translation by Donald Fraser Goodwin. Amsterdam: Benjamins. Burks, Arthur W. 1949 Icon, index, and symbol. Philosophy and Phenomenological Research 9: 673–689. Busse, Winfried 1975 Funktionen und Funktion der Sprache. In: Brigitte Schlieben-Lange (ed.), Sprachtheorie, 207–240. Hamburg: Hoffmann and Campe: Camhy, Daniela G. 1980a Karl Bühlers Sprachtheorie. PhD. Diss. Graz: University of Graz. Camhy, Daniela G. 1980b Sprechereignis als Handlung: Bühlers Beitrag zu handlungstheoretischen Ansätzen in der Sprachphilosophie. In: Rudolf Haller and Wolfgang Grassl (eds.), Sprache, Logik und Philosophie, 401–405. Wien: Hölder-Pichler-Temsky. Conte, Maria Elisabeth, János Sánder Petöfi and Emel Sözer (eds.) 1989 Text and Discourse Connectedness. Amsterdam: Benjamins. Coupland, Nikolas 2007 Style: Language, Variation and Identity. Cambridge: Cambridge University Press. Danesˇ, Frantisˇek 1987 On Prague School functionalism in linguistics. In: René Dirven and Vilém Fried (eds.), Functionalism in Linguistics, 3–38. (Linguistic and Literary Studies in Eastern Europe 20.). Amsterdam: Benjamins. Danesˇ, Frantisˇek 2003 Vilém Mathesius’ Konzeption der funktionalen Linguistik. In: Marek Nekula (ed.), Prager Strukturalismus: Methodologische Grundlagen, 31–47. (Slavica: Monographien, Hand-, Lehr- und Wörterbücher 3.). Heidelberg: Universitätsverlag Winter. Duranti, Alessandro 1997 Linguistic Anthropology. (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press. Ehlich, Konrad 1996 Sprache als System versus Sprache als Handlung. In: Marcelo Dascal, Dietfried Gerhardus, Kuno Lorenz, Georg Meggle (eds.), Sprachphilosophie: Ein internationales Handbuch zeitgenössischer Forschung, vol. 1, 952–963. (Handbücher zur Sprach- und Kommunikationswissenschaft 7.1). Berlin/New York: de Gruyter.

252

Saskia Daalder and Andreas Musolff

Elffers, Els 1993 Philipp Wegener as proto-speech act theorist. Linguistics in the Netherlands 1993: 49–59. Amsterdam: Benjamins. Elffers, Els 2005 Reichling and de Groot: Two Dutch reactions to Bühler’s organon-model. Historiographia Linguistica 32: 87–116. Elffers, Els 2008 Interjections and the language functions debate. The Henry Sweet Society Bulletin 50: 17–29. Eschbach, Achim (ed.) 1984a Bühler-Studien. 2 vols. Frankfurt am Main: Suhrkamp. Eschbach, Achim 1984b Karl Bühlers Axiomatik und das Axiomensystem der Zeichentheorie. In: Carl F. Graumann and Theo Herrmann (eds.), Karl Bühlers Axiomatik: Fünfzig Jahre “Axiomatik der Sprachwissenschaften”, 53–104. Frankfurt am Main: Klostermann. Fronek, Josef 1988 T.G. Masaryk and Prague School linguistic structuralism. In: Josef Novák (ed.), On Masaryk: Texts in English and German, 81–96. (Studien zur österreichischen Philosophie, 13). Amsterdam: Rodopi. Gabelentz, Georg von der 1875 Weiteres zur vergleichenden Syntax: Wort- und Satzstellung. Zeitschrift für Völkerpsychologie und Sprachwissenschaft 8: 129–165, 300–338. Gardiner, Sir Alan Henderson 1932 The Theory of Speech and Language. Oxford: Clarendon Press. (2nd ed. with additions 1951.) Garvin, Paul L. 1966 Karl Bühler’s contributions to the theory of linguistics. The Journal of General Psychology 75: 212–215. Garvin, Paul L. 1994 Karl Bühler’s Field Theory in the Light of the Current Interest in Pragmatics. In: Sveˇtla Cˇmejrková and Frantisˇek Stícha (eds.), The Syntax of Sentence and Text: A Festschrift for Frantisˇek Danesˇ, 59–66. Amsterdam: Benjamins. Graffi, Giorgio 1991 200 Years of Syntax: A Critical Survey. (Studies in the History of the Language Sciences 98.) Amsterdam: Benjamins. Graumann, Carl F. 1984 Wundt – Mead – Bühler: Zur Sozialität und Sprachlichkeit menschlichen Handelns. In: Carl F. Graumann and Theo Herrmann (eds.), Karl Bühlers Axiomatik: Fünfzig Jahre “Axiomatik der Sprachwissenschaften”, 217–247. Frankfurt am Main: Klostermann. Graumann, Carl F. 1988 Aspektmodell und Organonmodell: Die Problematik des Verhältnisses zwischen Sprachwissenschaft und Psychologie bei Karl Bühler. In: Achim Eschbach (ed.), Karl Bühler’s Theory of Language, 107–124. Amsterdam: Benjamins. Grimm-Vogel, Irmingard Hildburg 1998 Philipp Wegener 1848–1916: Wesen – Wirken – Wege. Dissertation Rheinischen Friedrich-Wilhelms-Universität zu Bonn.

Foundations of pragmatics in functional linguistics

253

Grünbaum, Abraham Anton 1932 Sprache als Handlung. In: Gustav Kafka (ed.), Bericht über den 12. Kongreß der Deutschen Gesellschaft für Psychologie (Hamburg 1931), 164–176. Jena: G. Fischer. Habermas, Jürgen 1981 Theorie des kommunikativen Handelns. 2 vols. Frankfurt am Main: Suhrkamp. Hajicˇová, Eva 1994 Topic/Focus and related research. In: Philip A. Luelsdorff (ed.), The Prague School of Structural and Functional Linguistics: A Short Introduction, 245–275. (Linguistic and Literary Studies in Eastern Europe 41.). Amsterdam: Benjamins. Halliday, Michael A.K. 1978 Language as Social Semiotic: The Social Interpretation of Language and Meaning. London: Edward Arnold. Heger, Klaus 1984 Zeigfeld und Symbolfeld. In: Carl F. Graumann and Theo Herrmann (eds.), Karl Bühlers Axiomatik: Fünfzig Jahre “Axiomatik der Sprachwissenschaften”, 105–116. Frankfurt am Main: Klostermann. Henzler, Stefan 1987 Die Funktionen der Sprache im Gebrauch: Zu Konsistenz und Funktionalismus von Karl Bühlers Sprachtheorie. In: Werner Abraham and Ritva Århammar (eds.), Linguistik in Deutschland, 353–365. Tübingen: Niemeyer. Herrmann, Theo 1984 Sprachstrukturen und ihre aktionale Realisation: Psychologische Erwägungen zu Karl Bühlers “Axiomatik der Sprachwissenschaften”. In: Carl F. Graumann and Theo Herrmann (eds.), Karl Bühlers Axiomatik: Fünfzig Jahre “Axiomatik der Sprachwissenschaften”, 143–172. Frankfurt am Main: Klostermann. Holenstein, Elmar 1976 Roman Jakobson’s Approach to Language: Phenomenological Structuralism. Transl. by Catherine Schelbert and Tarcisius Schelbert. Bloomington: Indiana University Press. Husserl, Edmund 1900–1901 Logische Untersuchungen. Vol. I: Prologomena zur reinen Logik. Vol. II/1: Untersuchungen zur Phänomenologie und Theorie der Erkenntnis. Vol. II/2: Elemente einer phänomenologischen Aufklärung der Erkenntnis. Halle: Niemeyer. Hörmann, Hans 1978 Meinen und Verstehen: Grundzüge einer psychologischen Semantik. Frankfurt am Main: Suhrkamp. Hülzer, Heike 1987 Die Metapher: Kommunikationssemantische Überlegungen zu einer rhetorischen Kategorie. Münster: Nodus. Hymes, Dell 1964 Introduction: Toward ethnographies of communication. In: John J. Gumperz and Dell Hymes (eds.), The Ethnography of Communication, 1–34. (Special issue of American Anthropologist 66, 6, part 2.)

254

Saskia Daalder and Andreas Musolff

Hymes, Dell H. 1968 The ethnography of speaking. In: Joshua A. Fishman (ed.), Readings in the Sociology of Language, 99–138. The Hague: Mouton. First published Washington, D.C. [1962]. Hymes, Dell 1972 Models of the interaction of language and social life. In: John J. Gumperz and Dell H. Hymes (eds.), Directions in Sociolinguistics: The Ethnography of Communication, 35–71. Oxford: Blackwell. Hymes, Dell 1983 The pre-war Prague school and post-war American anthropological linguistics. In: Dell Hymes, Essays in the history of linguistic anthropology, 331–345. Amsterdam: Benjamins. Innis, Robert E. 1982 Karl Bühler: Semiotic Foundations of Language Theory. New York/London: Plenum Press. Innis, Robert E. 1992 Karl Bühler (1879–1963). In: Marcelo Dascal, Dietfried Gerhardus, Kuno Lorenz, Georg Meggle (eds.), Sprachphilosophie: Ein internationales Handbuch zeitgenössischer Forschung. Vol. 1, 550–562. (Handbücher zur Sprach- und Kommunikationswissenschaft 7.1.) Berlin/New York: de Gruyter. Jakobson, Roman 1950 Les catégories verbales. [Summary]. Cahiers Ferdinand de Saussure 9: 6. Jakobson, Roman 1953 [From the point of view of linguistics.] In: Results of the Conference of Anthropologists and Linguists, 11–21. (Indiana University Publications in Anthropology and Linguistics; International Journal of American Linguistics, Memoir 8.) Baltimore: Waverly Press. Jakobson, Roman 1960 Closing statement: Linguistics and poetics. In: Thomas A. Sebeok (ed.), Style in Language, 350–377. Cambridge, MA: MIT Press. Jakobson, Roman 1963 Efforts toward a means-ends model of language in inter-war continental linguistics. In: Christine Mohrmann, F. Norman and Alf Sommerfelt (eds), Trends in modern linguistics, 104–108. Utrecht/Antwerp: Spectrum. Jakobson, Roman 1966 Zur Struktur des russischen Verbums. In: Eric P. Hamp, Fred W. Householder and Robert Austerlitz (eds.), Readings in Linguistics II, 1–5. Chicago/London: The University of Chicago Press. First published Prague: Prague Linguistic Circle [1932]. Jakobson, Roman 1971 Shifters, verbal categories, and the Russian verb. In: Roman Jakobson, Selected Writings, vol. 2: Word and Language, 130–147. The Hague/Paris: Mouton. First published Cambridge, MA: Harvard University [1957]. Jakobson, Roman 1985 Metalanguage as a linguistic problem. In: Roman Jakobson, Selected Writings, vol. 7: Contributions to Comparative Mythology: Studies in Linguistics and Philology, 1972–1982, ed. by Stephen Rudy, 113–121. Berlin: Mouton de Gruyter. First published [1976].

Foundations of pragmatics in functional linguistics

255

Jespersen, Otto 1922 Language: Its Nature, Development and Origin. London: George Allen and Unwin; New York: Holt. Juchem, Johann G. 1984 Die Konstruktion des Sprechens: Kommunikationssemantische Betrachtungen zu Philipp Wegener. Zeitschrift für Sprachwissenschaft 3: 3–18. Juchem, Johann G. 1986 Wegener und Wundt. Kodikas/Code 9: 155–166. Juchem, Johann G. 1987 Einleitung. In: Gerold Ungeheuer, Kommunikationstheoretische Schriften, vol. 1: Sprechen, mitteilen, verstehen. Ed. by Johannes G. Juchem, 1–17. (Aachener Studien zur Semiotik und Kommunikationsforschung 14.). Aachen: Rader Verlag. Kamp, Rudolf 1977 Axiomatische Sprachtheorie: Wissenschaftstheoretische Untersuchungen zum Konstitutionsproblem der Einzelwissenschaften am Beispiel der Sprachwissenschaftstheorie Karl Bühlers. Berlin: Duncker and Humblot. Klein, Wolfgang 1984 Bühler Ellipse. In: Carl F. Graumann and Theo Herrmann (eds.), Karl Bühlers Axiomatik: Fünfzig Jahre “Axiomatik der Sprachwissenschaften”, 117–141. Frankfurt am Main: Klostermann. Knobloch, Clemens 1988 Geschichte der psychologischen Sprachauffassung in Deutschland von 1850 bis 1920. (Reihe Germanistische Linguistik 86.) Tübingen: Niemeyer. Knobloch, Clemens 1991 Introduction. In: Philipp Wegener, Untersuchungen über die Grundfragen des Sprachlebens. New edition by Konrad Koerner, with an introduction by Clemens Knobloch, xi–li. (Classics in Psycholinguistics 5.) Amsterdam: Benjamins. Knobloch, Clemens and Stefan Schallenberger 1993 Sprechhandlung und Sprachbedeutung in der Sprachpsychologie um 1930. Histoire – Épistémologie – Langage 15/1: 81–109. Koerner, E. F. Konrad 1984 Karl Bühler’s theory of language and Ferdinand de Saussure’s “Cours”. Lingua 62: 3–24. Lazarus, Moritz 1884 Geist und Sprache: Eine psychologische Monographie. 3rd ed. Berlin: Dümmler. First published Berlin [1856–1857]. Laziczius, Julius von 1939 Das sog. dritte Axiom der Sprachwissenschaft. Acta Linguistica 1: 162–167. Lesˇka, Oldrˇich 1995 Vilém Mathesius (1882–1945). [In Czech.] Slovo a Slovesnost 56: 81–97. Levinson, Stephen C. 1983 Pragmatics. Cambridge: Cambridge University Press. Liberman, Anatoly 1987 Roman Jakobson and his contemporaries on change in language and literature (the teleological criterion). In: Krystyna Pomorska, El˙zbieta Chodakowska,

256

Saskia Daalder and Andreas Musolff

Hugh McLean and Brent Vine (eds.), Language, Poetry and Poetics: The Generation of the 1890s: Jakobson, Trubetzkoy, Majakovskij, 143–155. Berlin: Mouton de Gruyter. Lohmann, Johannes 1942–1943 Karl Bühlers “Drittes Axiom”. Acta Linguistica 3: 5–16. Lyons, John 1977 Semantics. 2 vols. Cambridge: Cambridge University Press. Malinowski, Bronislaw 1946 The problem of meaning in primitive languages. In: Charles K. Ogden and Ivor A. Richards, The Meaning of Meaning: A Study of the Influence of Language upon Thought and of the Science of Symbolism. 8th ed, 296–336. London: Kegan Paul. First published London: Kegan Paul [1923]. Marty, Anton 1908 Untersuchungen zur Grundlegung der allgemeinen Grammatik und Sprachphilosophie. Vol. 1. Halle: Niemeyer. Masaryk, Tomásˇ Garrigue 1887 Versuch einer concreten Logik: Classification und Organisation der Wissenschaften. Wien: C. Konegen. Mathesius, Vilém 1929 Zur Satzperspektive im modernen Englisch. Archiv für das Studium der neueren Sprachen und Literaturen 155: 202–210. Mathesius, Vilém 1930 On linguistic characterology with illustrations from Modern English. In: Actes du premier Congrès International de Linguistes à la Haye, du 10–15 avril 1928, 56–63. Leiden: Sijthoff. Mathesius, Vilém 1966 Ten years of the Prague Linguistic Circle. Transl. by Josef Vachek. In: Josef Vachek (ed.), The Linguistic School of Prague: An Introduction to its Theory and Practice, 137–151. (Indiana University Studies in the History and Theory of Linguistics.) Bloomington/London: Indiana University Press. First published in Czech, Prague [1936]. Mathesius, Vilém 1975a A Functional Analysis of Present Day English on a General Linguistic Basis. Ed. by Josef Vachek, transl. by L. Dusˇková. (Janua Linguarum, Series Practica 208). The Hague/Paris: Mouton; Prague: Academia. First published in Czech, Prague [1961]. Mathesius, Vilém 1975b On information-bearing structure of the sentence. Transl. by Tsuneko Olga Yokoyama. Harvard Studies in Syntax and Semantics 1: 467–480. First published in Czech, Prague [1939]. Mathesius, Vilém 1983a On the potentiality of the phenomena of language. Transl. by Josef Vachek. In: Josef Vachek and Libusˇe Dusˇková (eds.), Praguiana: Some basic and less known aspects of the Prague Linguistic School, 3–43. (Linguistic and Literary Studies in Eastern Europe 12.). Amsterdam: Benjamins. First published in Czech, Prague [1911].

Foundations of pragmatics in functional linguistics

257

Mathesius, Vilém 1983b New currents and tendencies in linguistic research. In: Josef Vachek and Libusˇe Dusˇková (eds.), Praguiana: Some basic and less known aspects of the Prague Linguistic School, 45–63. (Linguistic and Literary Studies in Eastern Europe 12.). Amsterdam: Benjamins. First published Prague [1926]. Mathesius, Vilém 1983c Functional linguistics. Transl. by Libusˇe Dusˇková. In: Josef Vachek and Libusˇe Dusˇková (eds.), Praguiana: Some basic and less known aspects of the Prague Linguistic School, 121–142. (Linguistic and Literary Studies in Eastern Europe 12.). Amsterdam: Benjamins. First published in Czech, Prague [1929]. Morpurgo Davies, Anna 1998 History of Linguistics. Vol. 4: Nineteenth-Century Linguistics. Ed. by Giulio Lepschy. (Longman Linguistics Library.) London/New York: Longman. First published in Italian, Bologna [1992]. Musolff, Andreas 1990 Kommunikative Kreativität: Karl Bühlers Zweifelderlehre als Ansatz zu einer Theorie innovativen Sprachgebrauchs. Aachen: Alano. Musolff, Andreas 1993 Karl Bühler’s and Alan Gardiner’s concepts of metaphor in the context of their theories of speech and language. Beiträge zur Geschichte der Sprachwissenschaft 3: 255–272. Musolff, Andreas 1997 Karl Bühler. In: Jef Verschueren, Jan-Ola Östman, Jan Blommaert, Chris Bulcaen (eds.), Handbook of Pragmatics 1997, 1–15. Amsterdam: Benjamins. Nekula, Marek 1999 Vilém Mathesius. In: Jef Verschueren, Jan-Ola Östman, Jan Blommaert, Chris Bulcaen (eds.), Handbook of Pragmatics 1999, 1–14. Amsterdam: Benjamins. Nekula, Marek, and Klaas-Hinrich Ehlers 1996 Jakobsons Bilanz des Strukturalismus: Ein Vortrag, der jahrzehntelang nicht gedruckt werden konnte. Beiträge zur Geschichte der Sprachwissenschaft 6: 187–199. Nerlich, Brigitte 1990 Change in Language: Whitney, Bréal, and Wegener. (Routledge History of Linguistic Thought Series.) London/New York: Routledge. Nerlich, Brigitte 1992 Semantic Theories in Europe 1830–1930: From Etymology to Contextuality. (Studies in the History of the Language Sciences 59.) Amsterdam: Benjamins. Nerlich, Brigitte and David D. Clarke 1996 Language, Action, and Context: The Early History of Pragmatics in Europe and America, 1780–1930. Amsterdam: Benjamins. Ortner, Hanspeter 1983 Karl Bühlers Stellung in der Ellipsenforschung. Zeitschrift für Germanistische Linguistik 11: 149–165. Ortner, Hanspeter 1986 Bühlers Vierfelderschema (Das dritte Axiom der “Sprachtheorie”) – Grundgedanken und Rezeption. Kodikas/Code, Ars Semeiotica 9: 211–226.

258

Saskia Daalder and Andreas Musolff

Prague Linguistic Circle 1983 Theses presented to the first Congress of Slavists held in Prague in 1929. Transl. from Czech by Josef Vachek. In: Josef Vachek and Libusˇe Dusˇková (eds.), Praguiana: Some basic and less known Aspects of the Prague Linguistic School, 77–120. (Linguistic and Literary Studies in Eastern Europe 12.). Amsterdam: Benjamins. First distributed as loose sheets in Czech, Prague [1929]. Rudy, Stephen 1978 Jakobsonian Poetics of the Moscow and Prague Periods. Doctoral dissertation Yale University. Saussure, Ferdinand de 1931 Grundfragen der allgemeinen Sprachwissenschaft. Berlin: W. de Gruyter. Transl. by Hermann Lommel of: Cours de linguistique générale, ed. by Charles Bally and Albert Sechehaye, 1916. Scheerer, Thomas M. 1980 Ferdinand de Saussure: Rezeption und Kritik. (Erträge der Forschung 133.) Darmstadt: Wissenschaftliche Buchgesellschaft. Schiffrin, Deborah 1994 Approaches to Discourse. Oxford: Blackwell. Schlieben-Lange, Brigitte 1979 Linguistische Pragmatik. Stuttgart: W. Kohlhammer. Shannon, Claude E. and Warren Weaver 1949 The Mathematical Theory of Communication. Urbana: University of Illinois Press. Swiggers, Pierre 1981 Two dimensions in Karl Bühler’s sign theory. Kodikas/Code, Ars Semeiotica 4: 53–56. Tenchini, Maria Paola 2008 Aspetti funzionali e pragmatici nel pensiero di Philipp Wegener. Con la traduzione antologica di Philipp Wegener, “Untersuchungen über die Grundfragen des Sprachlebens”. (Quaderni del Centro di Linguistica dell’Università Cattolica 11.) Brescia: Editrice La Scuola. Toman, Jindrˇich 1995 The Magic of a Common Language: Jakobson, Mathesius, Trubetzkoy, and the Prague Linguistic Circle. Cambridge, MA: MIT Press. Trnka, Bohumil 1966 Vilém Mathesius. Transl. by Vladimir Honsa. In: Thomas A. Sebeok (ed.), Portraits of Linguists: A Biographical Source Book for the History of Western Linguistics, 1746–1963, vol. 2, 474–489. Bloomington: Indiana University Press. First published in Czech, Prague [1946]. Tynjanov, Jurij, and Roman Jakobson 1972 Probleme der Literatur- und Sprachforschung. Translated by Rolf Fieguth und Inge Paulmann. In: Texte der russischen Formalisten, vol. 2: 386–391. First published [1928]. Ungeheuer, Gerold 1987 Kommunikationstheoretische Schriften, vol. 1: Sprechen, mitteilen, verstehen. Ed. by Johannes G. Juchem. Aachen: Rader Verlag.

Foundations of pragmatics in functional linguistics

259

Ungeheuer, Gerold 1990 Kommunikationstheoretische Schriften, vol. 2: Symbolische Erkenntnis und Kommunikation. Ed. by H. Walter Schmitz. Aachen: Alano Verlag. Vachek, Josef 1982 Vilém Mathesius’ living heritage to world linguistics. Philologica Pragensia 25:121–127. Vonk, Frank 1992 Gestaltprinzip und abstraktive Relevanz: Eine wissenschaftshistorische Untersuchung zur Sprachaxiomatik Karl Bühlers. Münster: Nodus. Waugh, Linda R. 1976 Roman Jakobson’s Science of Language. Lisse: Peter de Ridder Press. Waugh, Linda R., and Monique Monville-Burston 1990 Introduction: the life, work, and influence of Roman Jakobson. In: Roman Jakobson, On Language. Ed. by Linda R. Waugh and Monique MonvilleBurston, 1–45. Cambridge, MA: Harvard University Press. Waugh, Linda R., and Stephen Rudy 1998 Jakobson and structuralism. In: Roland Posner, Klaus Robering and Thomas A. Sebeok (eds.), Semiotik: Ein Handbuch zu den zeichentheoretischen Grundlagen von Natur und Kultur, vol. 2, 2256–2271. (Handbücher zur Sprach- und Kommunikationswissenschaft 13.2.). Berlin/New York: de Gruyter. Wegener, Philipp 1902 Review of Berthold Delbrück, Grundfragen der Sprachforschung (Strassburg: Trübner, 1901). Literarisches Centralblatt 12 (22. März 1902): 401–410. Wegener, Philipp 1911 Exposition und Mitteilung: Ein Beitrag zu den Grundfragen des Sprachlebens. In: Festschrift zur Feier des 350 jährigen Bestehens des Gymnasiums zu Greifswald, Bd. 2: Wissenschaftliche Aufsätze, 1–21. Greifswald: Kunike. Wegener, Philipp 1991 Untersuchungen über die Grundfragen des Sprachlebens. New edition by Konrad Koerner, with an introduction by Clemens Knobloch. (Classics in Psycholinguistics 5.) Amsterdam: Benjamins. First published Halle: Niemeyer, [1885]. Weil, Henri 1887 The Order of Words in the Ancient Languages compared with that of the Modern Languages. Transl. by Charles W. Super. Boston: Ginn. First published in French, Paris: Joubert [1844]. Weinreich, Uriel, William Labov and Marvin I. Herzog 1968 Empirical foundations for a theory of language change. In: Winfred P. Lehmann and Yakov Malkiel (eds.), Directions for Historical Linguistics: A Symposium, 95–195. Austin/London: University of Texas Press. Weissenborn, Jürgen 1988 Von der demonstratio ad oculos zur Deixis am Phantasma: Die Entwicklung der lokalen Referenz bei Kindern. In: Achim Eschbach (ed.), Karl Bühler’s Theory of Language, 257–276. Amsterdam: Benjamins. Wettersten, John 1988 Külpe, Bühler, Popper. In: Achim Eschbach (ed.), Karl Bühler’s Theory of Language, 327–347. Amsterdam: Benjamins.

260

Saskia Daalder and Andreas Musolff

Wolf, Käthe 1932 Darstellungsfelder in der Sprache. In: Gustav Kafka (ed.), Bericht über den 12. Kongreß der Deutschen Gesellschaft für Psychologie (Hamburg 1931), 449–453. Jena: G. Fischer. Wunderlich, Dieter 1969 Karl Bühlers Grundprinzipien der Sprachtheorie. Muttersprache 79: 52–62. Wundt, Wilhem 1900 Die Sprache. 2 Parts (= Vol. 1 of Völkerpsychologie). Leipzig: Engelmann. Wundt, Wilhem 1908 Kritische Nachlese zur Ausfragemethode. Archiv für die gesamte Psychologie 11: 445–459.

9.

Foundations: ethnomethodology and Erving Goffman Christine Domke and Werner Holly

1.

Genesis and foundations of ethnomethodology

1.1.

Harold Garfinkel’s perspective on people’s everyday life activities

The most crucial assumption of ethnomethodology is that social reality exists only through the everyday activities of a given society’s members. Ethnomethodologists thus focus on the question of how social order arises in everyday life and is recognized as such. Answering this question entails identifying the actual practices used by members of a society to create order. The goal of ethnomethodological studies is to identify and reconstruct the methods used by individuals in everyday life to produce meaningful and differentiable activities and to indicate how they interpret the activities of others. Ethnomethodology was founded in the 1960s by Harold Garfinkel, who received his doctorate in 1952 for a dissertation entitled “The perception of the other: A study in social order” under the direction of the sociologist Talcott Parsons (cf. Heritage [1984] 1992; Schegloff 1992). While Garfinkel viewed Parsons’ studies (alongside those of Alfred Schütz and Edmund Husserl) as crucial in the development of his own perspective on “the world of everyday activities” (Garfinkel [1967] 1996: ix), he also insisted on distancing himself from his former mentor’s theoretical assumptions about the crucial question of how social order arises. In Parsons’ structural functionalism, social order is made possible by the orientation of the individual to intersubjectively valid norms. According to Parsons, these cultural values and rules serve as a framework for the actions of the individual, who perceives and interprets social reality on the basis of these shared norms. Garfinkel, however, rejected the notion that the individual is geared to given norms and asked instead how he goes about interpreting situations. Garfinkel thus shifted the focal point away from collectively held norms as constitutive of social order to identify the continuous interpretation and attribution of meaning as central to the genesis of social order. Hence in his own analysis of social order, Garfinkel concentrated on the assessments made by ordinary people and the activities they produce which can be perceived by others (Heritage 1992: 7–36). Garfinkel understood ethnomethodology in contrast to the then prevalent assumption of an “objective reality.” In the preface to his book “Studies in Ethnomethodology” (1996), he remarks that the objective reality of social facts should not be seen as given, but rather as the result of an ongoing accomplishment of the members of a society. Social order thus arises out of

262

Christine Domke and Werner Holly

the interactive activities of the participants. These activities also serve as “accounts” to explain how the activities of the participants should be understood. On the basis on these reflections, everyday life became a legitimate subject of sociological analyses. Moreover, the new focus directed attention at examining how the order of everyday life, including the world of work, is organized. For as Garfinkel (1996: 35–75) pointed out in the late 1960s, although “the familiar common sense world of everyday life” occupied a central place as a topic of sociological inquiry, most scholarly literature included little data and few methods which could be used to investigate the features of “familiar scenes”. His main concern was thus to affirm and to (re)discover common sense activities in a society’s everyday life as a relevant object of study. 1.2.

Early ethnomethodological investigations

One of the first ethnomethodological studies (see e.g. Turner 1974b; Sudnow 1972) examines the activities which jurors – as non-professional adjudicators – engage in so that others will recognize them as jurors and take their actions seriously (Garfinkel 1974: 16). Using tape recordings and interviews with jurors as a basis, this study focuses on “becoming a juror” (Garfinkel 1996: 110). This process includes, for example, using “rules of the official line” instead of the “rules of daily life” (Garfinkel 1996: 110–113). When describing their actions, the jurors present their decisions in correspondence with the official rules as “right,” as based on law and evidence, not personal preferences or interests. Moreover, they justify their decisions by stressing the “integrative features” of the deliberations, thereby neglecting the “anomic” or random ones (Garfinkel 1996: 113). These examples demonstrate how a continuous perspective, that of the jurors, is successively generated. Another early study about “suicidal deaths” examines the question of how the involved parties, in this case coroners, arrive at the diagnosis “suicide” in cases of “sudden, unnatural death” (Garfinkel 1996: 11). This work analyzes the process of inquiry through which investigators establish “what really happened” as they provide an account of how a particular person died (Garfinkel 1996: 15). Ethnomethodological studies are thus concerned with how activities of everyday life are organized and how this organization is accomplished through people’s everyday life activities. The approach as well as the name ethnomethodology can be traced back to the study on jurors (see above) which Garfinkel carried out together with Saul Mendlovitz in Chicago during the 1950s before going on to the University of California at Los Angeles (UCLA) in 1954. When he wrote up the study about the tapes and interviews, he was looking for a suitable name for the “doing methodology” as observed in the jurors (Garfinkel 1974: 16). The prefix “ethno-” is intended to refer to a society’s body of general knowledge that serves the individual as a point of orien-

Foundations: ethnomethodology and Erving Goffman

263

tation. Hence the term ethnomethodology designates the procedures and practices, that is, ethno-methods, which are valid for the members within a particular society and which produce structure. While the approach of ethnomethodology is analogous to, for example, ethnobotany, it is not intended to stand for a specific knowledge domain such as, for example, ethno-physics (Garfinkel 1974: 16–17). The genesis of the term highlights one of the main thoughts behind this sociological research approach as laid out in Garfinkels 1967 volume “Studies in Ethnomethodology”: the search for the everyday methods through which a society’s members structure their activities and make them comprehensible and visible as such should be oriented towards situations and conditions of everyday life (1996: vii). This direct orientation to the accomplishment of everyday activities was achieved with the help of tape recordings and ethnographical observations from the field. Basing their work on common sense activities and daily situations, Garfinkel and other ethnomethodologists of his day vehemently denounced all forms of standardization of ethnomethodological research and stressed what they were not intending to do: neither theories, nor advice, nor recipes, but rather the practices and features of everyday life would be the focus of their investigations (Garfinkel 1996: viii; see also Turner 1974a). This avoidance of solely theoretical assumptions to focus on everyday activities represents one of Garfinkel’s central ideas, which can be linked to classical American pragmatism: The relevance of real lived experiences, habitual practices, and language with its indexically and socially constitutive characteristics had already been established in the work of Peirce, Mead, Dewey, and others. Seen from this perspective, the questions arises as to whether certain elements of Garfinkel and ethnomethodology can be seen as the fulfillment of the “unfinished business” of pragmatism (Emirbayer and Maynard 2006). In “Studies in Ethnomethodology” (1996) Garfinkel presents various studies from the preceding twelve years, thereby demonstrating the possibilities for his form of analysis of everyday activities. In addition to introductory articles and the study of the everyday practices of jurors mentioned above, the volume also includes two of Garfinkel’s best-known areas of investigation: one about a psychiatric clinic and the methods used there to identify patients for treatment, and another known as “Agnes.” “Agnes” is a pseudonym for a young woman who was born and raised as a male person (Garfinkel 1996: 120). At the age of 17, Agnes assumed a female identity by altering her appearance and dressing as a woman; three years later she underwent a sex-change operation. Garfinkel’s study uses Agnes’ story to illustrate that, contrary to commonly held notions about the fundamental difference between the sexes, the perception of “normal sexuality” is actually the result of a constant situational production which, in Agnes’ case, was constituted by the daily “management of herself as natural female” (1996: 184).

264 1.3.

Christine Domke and Werner Holly

Basic assumptions of ethnomethodology: making activities accountable

While the establishment of a full-fledged research program was not a concern of the first ethnomethodologists (Turner 1974a), several of their assumptions can be seen as central. As mentioned above, ethnomethodology assumes that the members of a society provide accounts of their own actions. The main focus of ethnomethodological studies is to show how people make their activities observable, reportable and visible so that social settings become accountable for their participants. Another basic thought behind ethnomethology is that the actions themselves indicate how they should be understood as they are being carried out. Because they constitute both a part of the situation in progress and a means of interpreting it, these human actions can be characterized by a basic reflexivity. Hence the context in which an activity should be understood is locally produced in situ by the activities themselves. This reflective character of the accounts makes it possible to assign sense and meaning in spite of the indexicality of the expressions and the vagueness of language. Another central concept in the work of Garfinkel known as indexicality emphasizes the contextuality of language. Because indexical expressions are so common and – in spite of the efforts of many scholars who are ironically mentioned by Garfinkel – irreplaceable, their properties provide important clues for understanding everyday actions. Concerned only with “objective expressions” or context-free language, the “exact sciences” describe indexical, that is, contextual expressions as simply “awkward” (Garfinkel 1996: 5) in spite of their frequency of occurrence and their utility; from this perspective, indexical expressions appear to be “irremediable nuisances” (Garfinkel 1996: 7). Garfinkel, however, is most interested in exploring the rational properties (1996: 9–11) of indexical expressions and thus the following central question: How is it possible to gain understanding in spite of the recognizable and “irremediable” contexuality of language and human action? In order to illustrate this contextuality and reflexivity of language and human activity, Garfinkel had his students conduct experiments in familiar scenes of everyday affairs known as “breaching experiments.” He wanted his students to examine ordinary conversations and episodes to test the degree to which everyday events are characterized by “seen but unnoticed” features (Garfinkel 1996: 36–49). Garfinkel borrowed the term “seen but unnoticed” from the phenomenologist Alfred Schütz, whose work about the world of everyday life influenced Garfinkel profoundly (Heritage 1992: 37–74). Schütz referred to the “background expectancies” which the members of a society use to understand and interpret human actions as the “attitude of daily life.” Garfinkel’s aim was thus to find out how commonplace scenes become familiar for the members of a society. The experiments were designed to discover the mutual expectations and assumptions which people share as a kind of basis. Garfinkel’s method was to examine what happens when people depart from the usual everyday order by breaking social rules or norms, or

Foundations: ethnomethodology and Erving Goffman

265

in his words: “to start with familiar scenes and ask what can be done to make trouble” (Garfinkel 1996: 37). Garfinkel’s students were asked to engage someone in an ordinary conversation and to then respond to indexical phrases which are normally not further scrutinized by asking the person to clarify their remarks. For example when asked “How are you?”, they would respond: “How am I in regard to what? My health, my finances, my school work […]?” (Garfinkel 1996: 44). Unaware of the experiment, most people reacted to this unexpected change in the normal course of conversation by showing confusion or annoyance or by changing the subject. It also became clear that one explanation tended to result in another, indicating that indexicality cannot be suspended or “healed.” Reflexivity, which is, as mentioned above, constitutive of all actions, normally prevents these kinds of deviations in everyday life and assures that the participants in a conversation alternately receive hints to help them deal with vagueness as a matter of course. 2.

Conversation Analysis

2.1.

Harvey Sacks’ “Lectures on Conversation”

These basic assumptions about the production of order through everyday activities have also been fundamental for conversation analysis (CA), which grew out of ethnomethodology from the mid to late 1960s to became an independent area of research (see e.g. Sudnow 1972; Turner 1974b). Harvey Sacks, who, like Garfinkel, was at UCLA from 1963, is known as the founder of CA. Through his work, he paved the way for a growing interest in everyday conversations as an object of study. Garfinkel and Sacks were engaged in close scientific exchange since they first met in a seminar by Talcott Parsons at Harvard in 1954 (Schegloff 1992: xiii; see also Bergmann 2000a). Known as the cornerstone of CA, “Lectures on Conversation” is composed of lectures which Sacks held between 1964–1972. Following his early death in 1975, Sacks’ former student Gail Jefferson edited the lectures (Sacks 1989, 1992 Vol. I & II). Sacks used recorded conversations (including telephone recordings from an emergency psychiatric hospital) as a basis for intensely analysing the structural characteristics of verbal interactions. In the course of his work, he introduced and advanced the position that with respect to both verbal and nonverbal activities, the generation of social order in everyday life is a legitimate subject of sociology (Sacks 1972). Topics covered in his extensive analyses include lies to be expected in conversations (e.g. as an answer to the question “How are you?”, Sacks 1992, Part V: 8, 9), the techniques used to choose a conversation’s next speaker, and the procedures used by members to categorize someone or an event. These techniques are used automatically by the participants in a conversation and assure, as a rule, two things: that conversations proceed not chaotically, but according to a certain order, and that each party is able to assign meaning to the utterances of the other. These order providing mechanisms are the focus of CA.

266

Christine Domke and Werner Holly

Sacks dealt extensively with questions about the everyday use of categories of membership (Sacks 1992, inter alii Part I: 6, Part V: 11, 12). In his view, the procedures and devices which members use to categorize others constitute a “central machinery of social organization.” Sacks developed an “apparatus” called MIR device (“M” for membership, “I” for inference-rich, “R” for representative) (1992: 41) which includes three categories of questions: questions about the chosen category sets (e.g. race, sex, age, occupation); questions about the conclusions which are made possible by the categories used, e.g. “baby” and “mommy” in the wellknown example “The baby cried. The mommy picked it up” (see Sacks 1992, Part III: 1–2; Sacks [1972] 1974); and questions about the activities ascribed to particular members as category bound, e.g. “crying” to babies and “picking up” to mothers. 2.2.

The sequential organization of verbal interaction

Sacks thus took what Garfinkel (and others) discussed about the relevance of everyday activities for the production of social order and applied it to verbal interactions. CA empirically examines how order is generated in conversations. Proceeding from the basic principle of “ongoing accomplishment” (see above), the question of the successive generation of various ordering structures comes into focus. The processual creation of ordering structures during interactions is conceptualized as doing (see Sacks 1984 on “doing being ordinary,” Garfinkel’s analysis of “Agnes” from 1996 on “doing gender,” Ayaß 2004 on “doing interviewing”). A concentration on sequential organization constitutes a central and pivotal point of analyses in the work of Sacks and within CA. The ever-present question about the localisation of individual elements of conversation puts the focus on the direct context of single utterances. By analysing the sequence of utterances generated locally within the situation, it is possible to determine what an utterance refers to, what problem a particular turn solves, and which turn can be expected to follow. Sequential organization is the core of the methodical approach of CA, which tries to identify the processes which give order to conversation. In conjunction with Emanuel Schegloff, a former fellow student who later became a colleague, and Gail Jefferson, a former student, Sacks developed various research methods for studying sequential organization. These methods constitute the centrepiece of CA and prominently demonstrate how CA examines interactions within the context of their own order, which is produced turn by turn. These assumptions about the selfreferential and reflexive processes involved in the production of a conversation make it possible to describe the autonomous nature of interactions and to designate them as systems (see below). The first basic structure describes how people take turns in ordinary discourse according to a “turn-taking-machinery” which functions automatically during conversation. Analysing this machinery indicates how the “simplest systematic” for

Foundations: ethnomethodology and Erving Goffman

267

the organization of turn-taking in conversation functions and how the latent problem of choosing the next speaker at a “transition relevance place” can be solved, e.g. through “self-selection” or “other selection” (Sacks, Schegloff, and Jefferson [1974] 1978). The local, sequential organization of a conversation underscores Garfinkel’s assumptions about the reflexivity of all actions in everyday life: understanding in verbal interaction is possible because the utterance of one speaker reflects and makes accountable his understanding of what was said before. One of the central methods used by speakers to organize their conversation in the course of conversation are so-called “utterance pairs” or “adjacency pairs” (Schegloff and Sacks 1973; Schegloff 1972). Because some utterances provoke certain responses, they are said to operate according to a property called “conditional relevance” (inter alii Schegloff 1968), which compels the next speaker to respond in an expectable way in terms of both content and form (as a rule, questions are followed by answers and greetings are followed by return greetings). Understanding is made possible not only in the sequence of utterances and the structural order (question/ answer, accusation/defence), but also by virtue of the fact that utterances are retrospectively repaired. The sequential repair of utterances constitutes a special feature of conversation: an utterance can be readily “revised” or repaired with respect to formulations or details in a second turn. As shown in various studies (Schegloff, Jefferson, and Sacks 1977; Schegloff 1979; Jefferson 1974), repairs can be differentiated according to who initiates them and who actually carries them out. The analyses indicated that participants prefer self-initiated self-repair (as opposed to other-repair), which means that the speaker of the utterance requiring repair makes the correction without prompting from another. This idea has also been developed in further studies of conversation under the name of a general “preference organization” and can be described with reference to common adjacency pairs such as agreement/disagreement (inter alii Pomerantz 1984). In addition to these three central areas (turn-taking, adjacency pairs, repairs), a number of structural characteristics and specific tasks of verbal interaction (such as recipient design, laughter, opening and ending of conversation, stories) as well as conversation types (everyday conversations, consultations, trouble talks, e.g. Jefferson 1984) have been analyzed since the beginnings of CA (see Schenkein 1978; Psathas 1979; Atkinson and Heritage 1984; Hutchby and Wooffitt 1998). The initial work of Sacks, Sudnow, Schegloff, and Jefferson in Berkeley and at UCLA placed a sustained focus on the regularities of everyday speech (also in working life) as well as on the (ethno)-methods by means of which this verbal order is generated. 2.3.

The analysis of empirical data

Because “talk-in-interaction” (Schegloff 1987) had not been the subject of such minute analysis up until this point, another achievement of the first CAs was to develop and promote appropriate methods for archiving and preserving conver-

268

Christine Domke and Werner Holly

sations (Bergmann 1981, 1985). Using recordings of natural data (initially telephone conversations) as a basis, transcripts were made which attempted to record entire exchanges word for word. Hence word interruptions, reformulations and “ums” were not edited out, but rather became a part of the transcript. Both the recordings and the transcripts provide the data used in the analysis of authentic conversations (Hutchby and Woofitt 1998; ten Have 2007). Both then and now, these natural data stand in strong contrast to data which have been contrived or generated in experimental settings (inter alii Heritage 2001). With the help of sound recordings and, more recently, video recordings (on this topic see Heath and Luff 2000; Goodwin 2000), CA preserves interactions by recording them for future use. These data transcripts then form the basis for exploring the central issue of ordering structures. Seen from this perspective, reconstructive methods (e.g. narrative interviews, memorized observations) commonly used prior to CA (and even today) do not constitute data which have been recorded verbatim from an actual situation. Already tainted by the perspective of the one reporting it, data collected in this way cannot provide a basis for analyzing the genuine course of interactions (Bergmann 1985). The strictly empirical approach of CA starts “from the data themselves” (Schegloff and Sacks 1973: 291) in order to reconstruct the mechanisms und methods through which the participants of verbal interaction generate turn-by-turn order. This focus on the course of the interactions themselves disaffirms the practice of arbitrarily applying a methodological or normative apparatus which is always supposed to be universally valid. Just as ethnomethodologists reject theoretical claims and programmatic assertions (see Turner 1974b; Garfinkel 1996), there is a consensus within CA about the inappropriateness of these kinds of standardizations. In the view of CA, verbal interactions should be studied using methods that arise out of the analysis of the data: to use hypotheses or theoretical links as a starting point would be to narrow one’s perspective on the order of a given set of interactions. As it seeks to identify the specific features of a particular interaction using turn-by-turn analysis of the data, this position can also be understood as a kind of “analytic mentality” (Schenkein 1978: 6). This absolute concentration on the order of interaction as recorded by the transcript (which now includes nonverbal elements as well) emphasizes the special empiric understanding of CA and its perspective on interactions. The fact that the methods of the participants are concretely reconstructed “from the data themselves” not only makes it possible to study authentic processes by reconstructing the creation of order and the sequential structure of talk turn by turn. This method of analysis with its step by step procedure also highlights the autonomy of interaction. In other words, the particular empiric understanding points to certain assumptions about the object, the interaction itself. With each successive turn, the order of interaction shows which options emerge for the participants, what the interaction “requires,” and what interactively appears as relevant. From this perspective, self-

Foundations: ethnomethodology and Erving Goffman

269

referential and reflexive processes of interaction move to the centre of CA. This focus on the autonomy of interaction makes it possible to connect CA with more recent system theory (Luhmann 1984, 1997), which presupposes the self-referential and autopoietic organization of a system (see Hausendorf 1992; Schneider 1997; Domke 2006). These interconnections correspond to CA’s fundamental interest in interaction and open up empirical links to system theory’s strictly theoretically based thoughts on the formation of systems. One of the leading achievements and characteristics of CA is the way in which it – unlike mechanistic communication models – points out the internal dynamics of interactions. However, this focus on interaction as an autonomous ‘system’ also comprises a great challenge: the context of the interaction and the participants must be taken into account appropriately and methodically examined. It is in these points that CA differs from Erving Goffman’s perspective on interactions.

3.

Erving Goffman’s Interaction Order

Alongside Garfinkel’s ethnomethodology, Erving Goffman’s interaction analysis paved the way for numerous studies on the subject of everyday interactions and analyses of how social order is generated by these interactions. Goffman taught at the University of California at Berkeley from 1957–1968 before taking a position at the University of Pennsylvania. His first book, The Presentation of Self in Everyday Life was published in 1959 and is considered to be one of his best-known works. Goffman had a great influence on Harvey Sacks, who studied with Goffman in Berkeley before going on to UCLA. In his introductions to Sacks’ lectures, Schegloff emphasises that both sociologists influenced each other, an aspect which needs to be explored more closely (Schegloff 1992: xxii). The special nature of their relationship is underlined by Goffman’s reported answer to the question of whether Sacks studied under him: “What do you mean; I was his student!” (Schegloff 1992: xxiii). In spite of numerous ties and points of common ground (Drew and Wootton 1988; Widmer 1991) Goffman’s work can neither be subsumed under the name of ethnomethodology nor that of CA, but rather occupies an independent area of empirical social research with its own form of analysing everyday activities. However, this does not mean that a “Goffman School” or methodology actually developed out of his studies. It was primarily by pointing out the crucial role and function of the interaction order as an institution “in its own right” (Goffman 1983b) that Goffman came to have such an important influence on conversation analysts like Sacks and Schegloff (Heritage 2001; see also Bergmann 1991, 2000b). According to Goffman, the features of “traffic rules” for pedestrians can be studied in a “crowded kitchen” as well as anywhere else, and distinguishing between public and private settings is not crucial for understanding social interactions (Goffman 1983b). In his view, solely an understanding of face-to-face inter-

270

Christine Domke and Werner Holly

action as an independent domain with independent structures is necessary for the analysis of everyday situations in the presence of others. However, this does not mean in reverse that it can be effective to analyse the structure and rules of conversations out of context. On the contrary: in his studies on interaction order (initially Interaction Ritual, 1967), Goffman focuses on the rituals and processes (such as deference and demeanor, see also 1967) which participants adhere to in their everyday presentations of self in order to create “face” and engage in “moral enterprises” (Drew and Wootton 1988: 6–7). From this perspective, the importance of interactions for the formation of socially relevant structures (such as face, role) becomes clear. In order to differentiate social situations, Goffman uses, among others, the term encounter to describe participants engaged in copresence-based, mutually ratifying interaction, and the term gathering for unfocused situations (such as waiting for a train) (Goffman 1964, 1983b). Goffman’s analyses work under the assumption that the individual self is created and shaped in every everyday face-to-face interactions. In his early essays “On Cooling the Mark Out” (Goffman 1952) and “On Face-work” (Goffman 1955), Goffman introduces the concept of face-work as an essential process of presenting the self and interacting with others. He developed the concept further in later studies. For example in Stigma (1963b), he explores how persons who deviate from “normality” (such as the handicapped or homosexuals) try to come to terms with their threatened identity in everyday life. The exclusion of the individual in public places, e.g. via the claiming of “territories” (such as seats or phone boxes) is the focus of Behaviour in Public Places (Goffman 1963a). Goffman uses central concepts from the theater to describe the production of self (Goffman 1959). Like actors in a play, people constantly engage in “impression management” in order to embody a certain role for their “audience.” This everyday dramaturgy includes various methods and attributes such as the front stage, where the performers play their roles in full view of the audience, and the backstage, a less accessible area which is not available for public viewing. The key points outlined above indicate a fundamental difference between Goffman’s ideas and the core of CA. In CA, the focus is on the sequential structure of interaction and the empirical analysis of interaction structures. This approach brings the autonomy of interaction to the fore, which allows for an understanding of interactions as systems as described in 2.3. Goffman, by contrast, is more interested in the individual and the effects of situations on his activities. The pertinent concepts of face and ritual put the focus on the individual in relation to his surroundings. For Goffman, interaction within the structures of these surroundings is less relevant than the individual participant in his emerging social identity. Hence, by looking at the pertinence of interaction for both CA and Goffman, significant differences can be identified. In light of the mutual influence between CA and Goffman, these differences have prompted numerous attempts to explain and distinguish the two approaches (see Bergmann 1991).

Foundations: ethnomethodology and Erving Goffman

271

Goffman addresses linguistic systems and dialogic structures more concretely (see Bergmann 1991) in his later works. The essays collected in Forms of Talk (Goffman 1981) identify the various possibilities for differentiating social situations and encounters and discuss the role of the participants. Of particular relevance are his thoughts on participation status and participation framework (Goffman 1981: 137), concepts which stand in contrast to the rigid ascriptions of a speaker/hearer model (see Auer 1999: 155–163; Bergmann 1991: 314–317). In one of the essays entitled “Footing,” which first appeared in 1979 in the magazine Semiotica, Goffman describes a range of different types of speakers and hearers (Goffman 1981: 124–157). The participant who is engaged in talking is at that moment merely the “sounding box in use” (Goffman 1981: 144); he or she functions as an animator. The animator and the author of what is said are not necessarily identical and must be differentiated. A third function is that of the principal, the one who is responsible for what is said. Taken together, these ascriptions reveal what Goffman refers to as the production format of an utterance (Goffman 1981: 145). Hearers can also be differentiated into different modes (Goffman 1981: 131–137). There are two kinds of ratified hearers: those who are directly addressed by the speaker and those who are unaddressed. Non-ratified participants may be listening intentionally, which Goffman calls eavesdropping, or unintentionally, which he refers to as overhearing. All of these elements constitute the footing of an interaction. Displayed interactively, footing is characterized by constant change. The footing of an interaction also acts as a kind of orientation aid similar to the more general frame of events, a concept which Goffman developed in Frame Analysis (Goffman 1974) with reference to the earlier work of Bateson. For Goffman, frames involves definitions of situations and cognitive structures which guide participants and indicate what might be expected of them. Frame analysis (e.g. Goffman 1974) includes Goffman’s analyses of response cries, a term which was also the title of an article (Goffman 1978) later published in Forms of Talk. Speakers produce exclamatory utterances like “Oops!” or “Brr!” in everyday life as “self talk” without directly addressing these interjections to a particular person. Using his distinction between social situation and encounter (see above), Goffman describes these response cries as “creatures of social situation, not states of talk” (Goffman 1981: 121) and thus identifies them as conventionalised expressions and not as “primitive, unsocialised” (Goffman 1981: 120. The open state of talk is yet another of Goffman’s important designations which is useful for describing certain communicative situations. An open state designates a conversation in which the participants have “the right but not the obligation” to talk, as is the case when individuals are involved in a joint physical task or a common meal (Goffman 1981: 134–135; also 1983a).

272 4.

Christine Domke and Werner Holly

Recent developments

Taking the fundamental studies described here as a starting point, the analysis of verbal interaction has developed further in many respects. In the following, a few of the most prominent out of a large number of works will be mentioned, starting with the “studies of work” which grew out of ethnomethodology. In the introduction to a volume of studies (1986) which he edited, Garfinkel emphasises that the analysis of work processes (e.g. within a jazz ensemble studied by Harvey Sacks) makes it clear that “there exists a locally produced order of work’s things” (Garfinkel 1986: vii) which has largely been ignored. The specific features of individual occupational activities have become the focus of studies on everyday practices in the working world including both verbal and nonverbal activities (see Bergmann 2000b). So-called “workplace studies,” which concentrate primarily on the relationship between “society and technology,” have also been influenced by studies of work (Knoblauch and Heath 1999). This area of research focuses on computer supported work environments and the consequences of the mechanisation of various processes (inter alii Heath and Luff 2000). The next step will be to extend the methods used in these studies to carry out large-scale document and video analyses; this endeavour will provide adequate material for future works of conversation analysis. Other important studies which were motivated by conversation analysis analyse the characteristics of work realms or specific forms of interactions by focusing on the verbal component (see Drew and Heritage 1992 on “Talk at work”; Clayman and Heritage 2002 on “News Interview”). In more recent studies, this concentration on speech has given way to a more holistic approach (Dausendschön-Gay and Krafft 2002), which tries to incorporate facial expressions, gestures and the environment of interactions into each analysis. Conversation analytical work with video recordings has led to a discussion of classical CA focus areas (such as turntaking, conversational openings) and elaborated the interplay and necessary coordination between speech, gesture and objects for various areas including medicine, everyday conversations and the media (inter alii Goodwin 2000; Ayaß 2004; Kissmann 2009; Schmitt 2007; Streeck 2009).

5.

Influence of conversation analysis on linguistic pragmatics

Conversation analysis is one of the two cornerstones of pragmalinguistics. In retrospect, it appears to be the less prevalent of the two, but in the long run it may be the more sustainable. In simplistic terms, one could say that linguistic pragmatics has essentially followed two major lines in its reorientation. On the one hand, it has taken its impulse from the philosophy of language, incorporating and integrating not only ordinary language philosophy along with the speech act theory of Austin and Searle, but also the concept of meaning and understanding as explicated by

Foundations: ethnomethodology and Erving Goffman

273

their attributed intellectual “father” Ludwig Wittgenstein and further developed by Grice. On the other hand, linguistic pragmatics has been influenced by sociological impulses which were mediated chiefly through conversation analysis and its historical backdrop, including names ranging from Harold Garfinkel and Alfred Schütz to Max Weber. In addition, the microsociological studies of Erving Goffman with their direct relevance to every kind of interaction have impacted the field to a lesser degree. No linguist working in the field of pragmatics has been able to avoid coming to grips with speech act theory and its string of tradition because “speech act” is the central concept that has defined not only the whole linguistic subdiscipline, but also the refounding of the entire field of linguistics. The study of conversation analysis (and especially Goffman), however, has been less obligatory for two reasons. First, because it applies “only” to verbal interaction, CA was thought to represent a limited subject area. Secondly, CA was in competition with other approaches which also attempted to analyze spoken interaction or which did not make a rigorous distinction between (spoken) interactions and (written) texts. Another possible explanation for the resistance to CA may be the challenges it posed to adherents of speech act theory from the philosophy of language who were accustomed to working with constructed and isolated example sentences. The consistent empirical approach to collecting data and diligently transcribing them required a fundamentally different kind of methodology than was required by grammar-oriented research. Finally, many were put off by what they saw as a kind of “unorthodox leftist movement,” a reputation which CA earned with its “wild” 1960s California image, its antitheoretical mistrust of academic omniscience, and its uniquely different perspective: the adherents of CA were searching for the “methods” of the members of society themselves. Others, however, were attracted by this unconventionality. For example, Bergmann (1981: 38) admires “das Unsystematische und Unabgeschlossene, das Ungeschützte und Riskant-Artistische, das Fragmentarische, Labyrinthhafte und Antiakademische” [the unsystematic and inconclusive, the unprotected and risky-artistic, the fragmentary, labyrinthian and anti-academic]. In any case, it cannot be denied that the actual subject matter of CA was paradoxically just what the more theoretical and less disputed branch of pragmatics actually had in mind without analyzing it empirically: “ordinary language.” While CA researchers repeatedly denied having a primary or special interest in language and conversation (Schegloff and Sacks 1973, cited in Coulthard 1977: 52; see also Sacks’ comment as quoted in Auer 1999: 137), they nonetheless were compelled to – and wanted to – attract interest by designating themselves as conversation analysts, especially in light of the new pragmatic orientation of the field of linguistics and the relevance and central importance of their subject matter, as highlighted here by Levinson: “It is not hard to see why one should look to conversation for insight into pragmatic phenomena, for conversation is clearly the prototypical kind of language usage […]” (1983: 284).

274

Christine Domke and Werner Holly

The gradual triumph of CA, particularly in the field of linguistics, was thus inevitable to a certain extent. The appearance of important studies by sociological researchers in linguistic journals (Sacks, Schegloff, and Jefferson 1978; Schegloff, Jefferson, and Sacks 1977 in Language) or in a socioliguistic context (e.g. Sacks 1974, first in Gumperz and Hymes 1972) was no accident. The proximity and kinship to newly developing sociolinguistic approaches (e.g. the ethnography of speaking) created a situation in which the various communication-based approaches benefited each other in an area of overlap between sociology, anthropology, and linguistics; this interconnectedness also strengthened interest in works on conversation analysis. Recognition of CA expanded beyond a small circle of interested scholars to the entire (Anglo-Saxon) field of linguistics after the publication of the two chapters “The ethnography of speaking” and “Conversation analysis” (following the chapter entitled “Speech acts”) in An Introduction to Discourse Analysis by Malcolm Coulthard (1977). Coulthard founded the “Birmingham School” together with John Sinclair on the basis of their study of classroom interaction (Sinclair and Coulthard 1975). Within the field of German linguistics, two research structures in Bielefeld and Freiburg were primarily responsible for the spread of studies in conversation analysis. As early as 1973, researchers in Bielefeld published two volumes of texts by American sociologists, including studies by Garfinkel, Sacks und Psathas on ethnomethodology and by Hymes on the ethnography of speaking (Arbeitsgruppe Bielefelder Soziologen 1973). The work of the group in Bielefeld also led to the publication of a highly regarded article by the linguist Werner Kallmeyer and the sociologist Fritz Schütze on “Conversation Analysis” (Kallmeyer and Schütze 1976). The Freiburg project “Dialogstrukturen” [dialogue structures] from 1974 under the direction of Hugo Steger investigated spoken language, thereby incorporating CA research in its investigation of spoken language (Berens et al. 1976; for details see Schwitalla 2001a). Thanks to this early engagement in the field, the 1980 Annual Conference of the Institute for German Language on the topic of dialogue research included an overview article on ethnomethodological conversation analysis (Bergmann 1981). This topic was also covered in later publications, including German guides to dialogue analysis and conversation linguistics (Bergmann 1994, 2001) as well as an introduction to linguistic interaction by Auer (1999) on the basis of several “classic” authors in the field. The research paradigm “conversation analysis” was effectively linguistically canonized in the 1980s with the publication of the chapter on conversational structure in Pragmatics by Stephen Levinson (1983), the internationally best-known book on the topic. Levinson deals extensively with the methods and results of CA in this chapter, and similar introductions published later follow his example (e.g. Green 1989; Mey 1993; Grundy 1995; Meibauer 1999). Levinson’s work also served as a model for other writers because he discusses CA in contrast to other approaches, which are combined under the label “discourse analysis” (DA).

Foundations: ethnomethodology and Erving Goffman

6.

275

Conversation analysis vs. other approaches in the linguistics of spoken interaction

While Levinson advocates CA, others have taken a position between reservation and rejection. As a whole, the field of discourse analysis is quite diverse, especially since it sometimes includes analyses of both written texts (not considered here) and larger units such as discourse (in the sense of Foucault). This discussion, however, will continue to focus on spoken language. Alongside the previously mentioned approaches (Dell Hymes’ ethnography of speaking, Sinclair and Coulthard’s Birmingham School), John Gumperz developed his interactional sociolinguistics; Levinson also cites Van Dijk (1972) and Labov and Fanshel (1977) as typical representatives of discourse analysis; Bublitz (1991) and Ventola (2001) provide a general overview of conversation and discourse analysis, and Hausendorf discusses the situation in German-speaking regions (2001). Some approaches are based more strongly on speech acts: in Germany, for example “dialogue analysis” and “dialogue grammar” (e.g. Hundsnurscher 1995, 2001; Hindelang 1994). The volume Handbuch der Dialoganalyse [Handbook of Dialogue Analysis] (Fritz and Hundsnurscher 1994) includes a discussion of the “Geneva model of discourse analysis” (Moeschler 1994, 2001) and an action theoretically based concept of “linguistic communication analysis” (Gloning 1994), which is associated with Heringer’s “practical semantics.” Konrad Ehlich and Jochen Rehbein take their own separate direction with their “functional pragmatics” in “Diskursanalyse” (Rehbein 2001). Yet another example of DA is the “critical discourse analysis” (“Kritische Diskursanalyse”) of the “Wiener Schule” under Ruth Wodak (e.g. Wodak et al. 1990), an approach which is based on Fairclough’s “Critical Discourse Analysis” (CDA) (e.g. Fairclough 1995). Writing from the perspective of CA, Levinson (1983: 286–294) targets text grammatical or speech act theoretical approaches of DA with his criticisms of “methods and theoretical tools” (Levinson 1983: 294). Claiming that DA sometimes falls back on constructed examples, he criticizes a lack of empirical focus as well as a fundamentally intuitive approach which ultimately cannot tolerate the possibility of falsifiability. He also takes issue with DA’s acceptance of the familiar theoretical principles of well-formedness (see also Kohrt 1986: 74–75), a viewpoint which sees discourses as nothing more than strings of interlinked sentences. In the eyes of DA’s critics, this “oversimplifying grasp” (Kohrt 1986: 80) can easily lead to the construction of rigid or even normative structures (Bublitz 1991: 271). According to Levinson, the practice of assigning individual utterances to one particular speech act and then constructing structurally definable sequence mechanisms (as in sentences) is bound to fail because it cannot do justice to the complex reality of the unpredictable dynamic interactions which are themselves embedded in complex contexts (Levinson 1983: 288–294). Proponents of CA defend their method by pointing out that it is strictly empirical and inductive, with “as little ap-

276

Christine Domke and Werner Holly

peal as possible to intuitive judgements” (Levinson 1983: 287), and that it does not immediately categorize individual cases, but rather attempts to compile functionally similar cases from the data in order to prove the validity of an analysis (Bergmann 2001: 923). The most effective argument for CA is definitely its indisputable success in the recognition of the structures which organize speech. However, CA has also had its critics. For example, Levinson (1983: 287) makes reference to Labov and Fanshel (1977: 25) and Coulthard and Brazil (1979) as he points out: “DA theorists can accuse CA practitioners of being inexplicit, or worse, plain muddled, about the theories and conceptual categories they are actually employing in analysis.” It is indeed questionable whether all existing order can be deduced solely from the object of investigation. In other words, it is doubtful that utterances are each reflexive and indexical enough to be understood without being categorized by the interpreter “from the outside.” In fact, it is commonly recognized that by drawing on his or her “analytical mentality,” the ethnomethodologist is forced to draw repeatedly on the intuitive understanding which he or she possesses as a competent member of a linguistic community (Bergmann 1981: 23). At the same time, this demonstrative refusal to draw on additional outside knowledge in an attempt to immunize CA against possible allegations of interpretive arbitrariness has not only led several critics to label the field as “anti-psychological” and “individual-psychological” (Redder 1990: 7), but it has also caused some to cross the line to (unjustifiably) accuse CA of “secret positivism” (Flader and Trotha 1988). A further criticism cites an alleged failure to take social background into account (e.g. Fairclough 1989: 12; see also Bublitz 1991: 271), but this assessment is not particularly convincing in view of CA’s sociological and sociocritical origins. Moreover, more recently, a growing number of studies in CA have focused on institutional and work contexts as well as political communication. On the whole, the field of linguistic pragmatics has not really seen an intense dispute over theoretical and methodological issues surrounding CA; the predominant pattern has been rather a more or less reflected and more or less selective appropriation of methods and results. Evidence of these integrative tendencies can be found in many textbooks and introductions, but they arise above all out of the experiences and needs of concrete research. Hence at the end of the 1990s, Hausendorf (2001: 977) came to the following conclusion with respect to the situation in the German-speaking realm: “To the extent to which the methodological praxis of conversation analysis consolidated itself and the findings and conclusions of empirically sophisticated – i.e. complex and long-term – projects were presented in the course of the 1990s, the methodological struggle between the various approaches has faded away.”

Foundations: ethnomethodology and Erving Goffman

7.

277

Influence of Erving Goffman on linguistic pragmatics

While CA quickly attained prominence within linguistic pragmatics on account of its direct relevance for the study of spoken language, the linguistic importance of Goffman’s research initially appeared to be limited to marginal questions. However, his work has been widely recognized and discussed within many other disciplines because of its broad spectrum of topics (self-portrayal, asylums, interaction rituals, role distance, gender advertising, frame analysis), so that gradually, his studies have also entered the field of linguistics indirectly via other routes. At least four strands of Goffman’s work have been adopted by pragmatics in one form or another over the course of time. First, Goffman’s concept of face was further developed by Brown and Levinson ([1978] 1987) into a politeness model with universal applicability. This model was adopted in numerous studies on the topic of “language and politeness,” but it was also vehemently criticized for its weak empirical basis of single utterances as well as for the questionable way in which the model was universalized (for summary see Fraser 2001; Bublitz 2009: 270). However, the complexity of Goffman’s concept (see Tracy 1990) comprises at its core something that goes beyond mere aspects of politeness to encompass the ritual negotiation of identities and relationships between interaction partners. This second Goffmanian strand was received by conversation analysis quite early in a few instances (e.g. Holly 1979, 2001), but it has taken more time to develop the concept fully for linguistics with respect to routines (Coulmas 1981), rituals (e.g. Rauch 1992; Werlen 2001), conflict settlement (e.g. Schank and Schwitalla 1987; Schwitalla 2001b) and further issues. The third important strand constitutes Goffman’s frame analysis, which has been applied more or less prominently in many pragmalinguistic studies (e.g. Tannen 1984). Finally, the fourth strand involves Goffman’s explicit investigations of conversations (“Forms of Talk”), in particular the deconstruction of participant roles. This deconstructive aspect of Goffman’s work was adopted by Levinson (1988) and formed the basis of a new conceptualization of the idealized and simplistic “speaker-hearer-constellation” (see Schwitalla 2001c), e.g. to analyse multiple addressing (Kühn 1995) or testimonials in a television commercial (Holly 2007). Moreover Goffman’s analysis of different intensities of focus on talk generated the useful concept of “open state of talk,” which is connected to Bühler’s observations on action-accompanying talk and is also essential for the analysis of certain uses of speech (see e.g. on the reception of television texts Holly, Püschel, and Bergmann 2001). The fact that four articles in the Handbook of Linguistics of Text and Conversation focus primarily on Goffmanian concepts (Werlen 2001; Schwitalla 2001b, 2001c; Holly 2001) is an index of how widely Goffman’s research has been received in the field of pragmalinguistics. Nonetheless, his importance for the analysis of spoken language continues to be underestimated in pragmalinguistics, probably because of his independence as a scholar,

278

Christine Domke and Werner Holly

which fostered a mentality against the formation of schools, and his conceptual flexibility.

8.

The convergence of various approaches

Although ethnomethodological conversation analysis constitutes a markedly independent and distinctive school of thought in the study of spoken language, there is a gradual tendency toward increasing convergence with other related approaches which are anchored more firmly in the fields of anthropology, sociology or linguistics; Heritage (2001) combines the three under the label “ethno-sciences.” These ethno-sciences share in common not only work with empirical data, but also a series of core ideas which he characterizes in the following: This research embodies the core notions that (i) communicative meaning is inherently contextual in character, (ii) social context is unavoidably dynamic and is managed through the participants’ actions, (iii) the specific contextual significance of actions is ‘structurally’ achieved by means of rules and practices of conduct which are systematically related and organized as systems, (iv) the contextual significance of action also involves a web of inferences which are inescapable, very often involve personal, moral and social accountability, and thus connect interaction with culture, social structure and personality, and (v) all this is managed through the integrated significance of talk, paralanguage and body movement. (Heritage 2001: 916)

With these kinds of common convictions as a basis, infighting, which also has seen adherents of CA criticize a less than rigorous handling of methodological questions (Streeck 1983), could relax. At the same time, this convergence is being promoted in the field of linguistics by an increasing return to the original competencies, resulting in a common focus on linkages between grammatical and interactive structures (Uhmann 1997; Ochs, Schegloff, and Thompson 1997), in particular with reference to prosodic-intonational structures as foreseen in the concept of “interactional linguistics” by Elisabeth Couper-Kuhlen und Margret Selting. This approach sees itself as an “interface between linguistics und interaction theory (above all functional linguistics, conversation analysis, contextualization theory, anthropological linguistics)” (Selting and Couper-Kuhlen 2000: 76). Other efforts concentrate more closely on the connection between grammar and interaction (e.g. Deppermann, Fiehler, and Spranz-Fogasy 2006), so that the discipline of conversation linguistics reencounters many older orientations in the investigation of spoken language (see Schwitalla 2001a), be it on the basis of Ehlich und Rehbein’s “Functional Pragmatics” which was originally intended to focus on grammar as well, or on the basis of “construction grammar” (Günthner and Imo 2006; Deppermann 2007).

Foundations: ethnomethodology and Erving Goffman

279

References Arbeitsgruppe Bielefelder Soziologen (ed.) 1973 Alltagswissen, Interaktion und gesellschaftliche Wirklichkeit. 2 vols. Reinbek: Rowohlt. Atkinson, John Maxwell and John Heritage (eds.) 1984 Structures of Social Action: Studies in Conversation Analysis. Cambridge: Cambridge University Press. Auer, Peter 1999 Sprachliche Interaktion. Eine Einführung anhand von 22 Klassikern. Tübingen: Niemeyer. Ayaß, Ruth 2004 Konversationsanalytische Medienforschung. Medien und Kommunikationswissenschaft, 52 (1): 5–29. Berens, Franz-Josef, Karl-Heinz Jäger, Gerd Schank and Johannes Schwitalla 1976 Projekt Dialogstrukturen. Ein Arbeitsbericht. München: Hueber. Bergmann, Jörg R. 1981 Ethnomethodologische Konversationsanalyse. In: Peter Schröder and Hugo Steger (eds.), Dialogforschung. (Jahrbuch 1980 des Instituts für deutsche Sprache), 9–53. Düsseldorf: Schwann. Bergmann, Jörg R. 1985 Flüchtigkeit und methodische Fixierung sozialer Wirklichkeit. Aufzeichnungen als Daten der interpretativen Soziologie. In: Wolfgang Bonß and Heinz Hartmann (eds.), Entzauberte Wissenschaft. Zur Relativität und Geltung soziologischer Forschung, 299–321. Göttingen: Schwartz. Bergmann, Jörg R. 1991 Über Erving Goffmans Soziologie des Gesprächs und seine ambivalente Beziehung zur Konversationsanalyse. In: Robert Hettlage and Karl Lenz (eds.), Erving Goffman – ein soziologischer Klassiker der zweiten Generation, 301– 326. Bern/Stuttgart: Haupt. Bergmann, Jörg R. 1994 Ethnomethodologische Konversationsanalyse. In: Gerd Fritz and Franz Hundsnurscher (eds.), Handbuch der Dialoganalyse, 3–16. Tübingen: Niemeyer. Bergmann, Jörg R. 2000a Harold Garfinkel und Harvey Sacks. In: Uwe Flick, Ernst von Kardorff and Ines Steinke (eds.), Qualitative Forschung. Ein Handbuch, 51–62. Reinbek: Rowohlt. Bergmann, Jörg R. 2000b Ethnomethodologie. In: Uwe Flick, Ernst von Kardoff and Ines Steinke (eds.), Qualitativeforschung. Ein Handbuch, 118–135. Reinbek: Rowohlt. Bergmann, Jörg R. 2001 Das Konzept der Konversationsanalyse. In: Brinker et al. (eds.), 919–927. Brinker, Klaus, Gerd Antos, Wolfgang Heinemann and Sven Frederik Sager (eds.) 2001 Text- und Gesprächslinguistik. 2nd vol. (Handbücher zur Sprach- und Kommunikationswissenschaft 16.) Berlin/New York: de Gruyter.

280

Christine Domke and Werner Holly

Brown, Penelope and Stephen C. Levinson 1987 Politeness. Some Universals in Language Usage. Cambridge: Cambridge University Press. First published in: Esther N. Goody (ed.), Questions and Politeness, 56–289. Cambridge: Cambridge University Press 1978. Bublitz, Wolfram 1991 Discourse Analysis: The State of the Art. In: Claus Uhlig and Rüdiger Zimmermann (eds.), Anglistentag 1990 Marburg. Proceedings, 259–284. Tübingen: Niemeyer. Bublitz, Wolfram 2009 Englische Pragmatik. Eine Einführung. 2nd ed. Berlin: Erich Schmidt Verlag. Clayman, Steve and John Heritage 2002 The News Interview. Journalists and Public Figures on the Air. Cambridge: Cambridge University Press. Coulmas, Florian 1981 Routine im Gespräch. Zur pragmatischen Fundierung der Idiomatik. Wiesbaden: Athenaion. Coulthard, Malcolm 1977 An Introduction to Discourse Analysis. London: Longman. Coulthard, Malcolm and David Brazil 1979 Exchange Structure. Birmingham: Birmingham University. Dausendschön-Gay, Ulrich and Ulrich Krafft 2002 Text und Körpergesten: Beobachtung zur holistischen Organisation der Kommunikation. Psychologie und Sozialwissenschaft. Zeitschrift für Qualitative Forschung. Themenheft: Sprechen vom Körper – Sprechen mit dem Körper 4: 30–60. Deppermann, Arnulf 2007 Grammatik und Semantik aus gesprächsanalytischer Sicht. Berlin/New York: de Gruyter. Deppermann, Arnulf, Reinhard Fiehler and Thomas Spranz-Fogasy (eds.) 2006 Grammatik und Interaktion. Untersuchungen zum Zusammenhang zwischen grammatischen Strukturen und Interaktionsprozessen. Radolfzell: Verlag für Gesprächsforschung. Domke, Christine 2006 Besprechungen als organisationale Entscheidungskommunikation. Berlin/ New York: de Gruyter. Drew, Paul and John Heritage (eds.) 1992 Talk at work: interaction in institutional settings. Cambridge: Cambridge University Press. Drew, Paul and Anthony Wootton (eds.) 1988 Erving Goffman. Exploring the interaction order. Cambridge: Polity Press. Reprinted Cambridge: Polity Press [1996]. Emirbayer, Mustafa and Douglas Maynard 2006 Pragmatism and Ethnomethodology. Paper presented at the annual meeting of the American Sociological Association, Montreal Convention Center. Montreal, Canada: Online . 2009-05-24 from http://www.allacademic.com/ meta/p95497_index.html.

Foundations: ethnomethodology and Erving Goffman

281

Fairclough, Norman 1989 Language and Power. London: Addison Wesley. Fairclough, Norman 1995 Critical Discourse Analysis: The critical study of language. London: Longman. Flader, Dieter and Thilo von Trotha 1988 Über den geheimen Positivismus und andere Eigentümlichkeiten der ethnomethodologischen Konversationsanalyse. Zeitschrift für Sprachwissenschaft 7: 92–115. Fraser, Bruce 2001 The form and function of politeness in conversation. In: Brinker et al. (eds.), 1406–1425. Fritz, Gerd and Franz Hundsnurscher (eds.) 1994 Handbuch der Dialoganalyse. Tübingen: Niemeyer. Garfinkel, Harold 1952 The Perception of the Other: A Study in Social Order. Ph.D. Dissertation. Boston: Harvard University. Garfinkel, Harold 1974 The Origins of the Term ‘Etnomethodology’. In: Turner, Roy (ed.), Ethnomethodology. Selected Readings, 15–19. Harmondsworth: Penguin Education. Garfinkel, Harold (ed.) 1986 Ethnomethodological studies of work. London: Routledge and Kegan Paul. Garfinkel, Harold 1996 Studies in ethnomethodology. Cambridge: Polity Press. First published Englewood Cliffs, NJ: Prentice Hall [1967]. Gloning, Thomas 1994 Praktische Semantik und Linguistische Kommunikationsanalyse. In: Fritz et al. (eds.), 113–129. Goffman, Erving 1952 On Cooling the Mark Out: Some Aspects of Adaption to Failure. Psychiatry 15 (4), 451–463. Reprinted in: Charles Lemert and Ann Branaman (eds.), The Goffman Reader, 4–20. Oxford: Blackwell [1997]. Goffman, Erving 1955 On Face-work: An Analysis of Ritual Elements in Social Interaction. Psychiatry 18 (3), 213–231. Goffman, Erving 1959 The Presentation of Self in Everyday Life. Garden City, NY: Doubleday, Anchor Books. Goffman, Erving 1963a Behaviour in Public Places: Notes on the Social Organization of Gatherings. New York: The Free Press. Goffman, Erving 1963b Stigma: Notes on the Management of Spoiled Identity. Engelwood Cliffs, NJ: Prentice Hall. Goffman, Erving 1964 The Neglected Situation. American Anthropologist 66: 133–136. Reprinted in Charles Lemert and Ann Branaman (eds.), The Goffman Reader, 229–233. Oxford: Blackwell [1997].

282

Christine Domke and Werner Holly

Goffman, Erving 1967 Interactional Ritual: Essays on Face-to-Face Behaviour. Garden City, NY: Doubleday, Anchor Books. Goffman, Erving 1974 Frame Analyses. An Essay on the Organization of Experience. London: Harper and Row. Goffman, Erving 1978 Response Cries. Language 54 (4): 787–815. Reprinted in Erving Goffman, Forms of talk, 78–122. Philadelphia: University of Pennsylvania [1981]. Goffman, Erving 1979 Footing. Semiotica 25 (1/2): 1–29. Reprinted in Erving Goffman, Forms of talk, 124–159. Philadelphia: University of Pennsylvania [1981]. Goffman, Erving 1981 Forms of talk. Philadelphia: University of Pennsylvania Press. Goffman, Erving 1983a Felicity’s Condition. American Journal of Sociology 89 (1): 1–51. Reprinted in Charles Lemert and Ann Branaman (eds.), The Goffman Reader, 167–192. Oxford: Blackwell [1997]. Goffman, Erving 1983b The Interaction Order. American Sociological Review 48: 1–17. Reprinted in Charles Lemert and Ann Branaman (eds.), The Goffman Reader, 233–261. Oxford: Blackwell [1997]. Goodwin, Charles 2000 Practices of seeing: visual analysis – an ethnomethodological approach. In: Theo van Leeuwen and Carey Jewitt (eds.), Handbook of visual analysis, 157–182. London: Sage. Green, Georgia M. 1989 Pragmatics and Natural Language Understanding. Hillsdale, NJ: Lawrence Erlbaum. Grundy, Peter 1995 Doing Pragmatics. London: Arnold. Gumperz, John and Dell Hymes (eds.) 1972 Directions in Sociolinguistics: The Ethnography of Communication. New York/London: Holt, Rinehart and Winston. Günthner, Susanne and Wolfgang Imo (eds.) 2006 Konstruktionen in der Interaktion. Berlin/New York: de Gruyter. Hausendorf, Heiko 1992 Gespräch als System. Linguistische Aspekte einer Soziologie der Interaktion. Opladen: Westdeutscher Verlag. Hausendorf, Heiko 2001 Gesprächsanalyse im deutschsprachigen Raum. In: Brinker et al. (eds.), 971– 979. Have, Paul ten 2007 Doing Conversation Analysis. A Practical Guide. 2nd ed. Los Angeles: Sage. Heath, Christian and Paul Luff 2000 Technology in action. Cambridge: Cambridge University Press.

Foundations: ethnomethodology and Erving Goffman

283

Heritage, John 1992 Garfinkel and ethnomethodology. Cambridge: Polity Press. First published Cambridge: Polity Press [1984]. Heritage, John 2001 Ethno-sciences and their significance for conversation linguistics. In: Brinker et al. (eds), 908–919. Hindelang, Götz 1994 Sprechakttheoretische Dialoganalyse. In: Fritz et al. (eds.), 95–112. Holly, Werner 1979 Imagearbeit in Gesprächen. Zur linguistischen Beschreibung des Beziehungsaspekts. Tübingen: Niemeyer Holly, Werner 2001 Beziehungsmanagement und Imagearbeit. In: Brinker et al. (eds.), 1382–1393. Holly, Werner 2007 Audiovisuelle Hermeneutik. Am Beispiel des TV-Spots der Kampagne “Du bist Deutschland”. In: Fritz Hermanns and Werner Holly (eds.), Linguistische Hermeneutik, 389–428. Tübingen: Niemeyer. Holly, Werner, Ulrich Püschel and Jörg Bergmann (eds.) 2001 Der sprechende Zuschauer. Wie wir uns Fernsehen kommunikativ aneignen. Wiesbaden: Westdeutscher Verlag. Hundsnurscher, Franz 1995 Some Remarks on the Development of Dialogue Analysis. In: Franz Hundsnurscher and Edda Weigand (eds.), Future Perspectives of Dialogue Analysis, 79–93. Tübingen, Niemeyer. Hundsnurscher, Franz 2001 Das Konzept der Dialoggrammatik. In: Brinker et al. (eds.), 945–952. Hutchby, Ian and Robin Wooffitt 1998 Conversation Analysis: Principles, Practices and Applications. Cambridge: Polity Press. Jefferson, Gail 1974 Error corrections as an international resource. Language in Society 2: 181–199. Jefferson, Gail 1984 On the organization of laughter in talk about troubles. In: John Maxwell Atkinson and John Hertitage (eds.), Structures of social action: studies in conversation analysis, 346–369. Cambridge: Cambridge University Press. Kallmeyer, Werner and Fritz Schütze 1976 Konversationsanalyse. Studium Linguistik 1: 1–28. Kissmann, Ulrike Tikvah (ed.) 2009 Video Interaction analysis. Methods and Methodology. Frankfurt am Main: Lang Knoblauch, Hubert and Christian Heath 1999 Technologie, Interaktion und Organisation: Die Workplace Studies. Schweizerische Zeitschrift für Soziologie 25 (2): 163–181. Kohrt, Manfred 1986 Dialoggrammatik und/oder Konversationsanalyse. In: Franz Hundsnurscher and Edda Weigand (eds.), Dialoganalyse. Referate der 1. Arbeitstagung Münster 1986, 69–82. Tübingen: Niemeyer.

284

Christine Domke and Werner Holly

Kühn, Peter 1995 Mehrfachadressierung. Untersuchungen zur adressatenspezifischen Polyvalenz sprachlichen Handelns. Tübingen: Niemeyer. Labov, William and David Fanshel 1977 Therapeutic Discourse: Psychotherapy as Conversation. New York: Academic Press. Lemert, Charles and Ann Branaman (eds.) 1997 The Goffman reader. Ed. with introductory essays by Charles Lemert and Ann Branaman. Oxford: Blackwell. Levinson, Stephen C. 1983 Pragmatics. Cambridge: Cambridge University Press. Levinson, Stephen C. 1988 Putting Linguistics on a Proper Footing. In: Paul Drew and Anthony Wootton (eds.), Erving Goffman. Exploring the interaction order, 161–227. Cambridge: Polity Press. Luhmann, Niklas 1984 Soziale Systeme. Grundriss einer allgemeinen Theorie. Frankfurt am Main: Suhrkamp. Luhmann, Niklas 1997 Die Gesellschaft der Gesellschaft, Vols 1 and 2. Frankfurt am Main: Suhrkamp. Meibauer, Jörg 1999 Pragmatik. Eine Einführung. Tübingen: Stauffenburg. Mey, Jacob L. 1993 Pragmatics. An Introduction. Oxford: Blackwell. Moeschler, Jacques 1994 Das Genfer Modell der Gesprächsanalyse. In: Fritz et al. (eds.), 69–94. Moeschler, Jacques 2001 The Geneva School. In: Brinker et al. (eds.), 952–957. Ochs, Elinor, Emanuel A. Schegloff and Sandra Annear Thompson (eds.) 1997 Interaction and grammar. Cambridge: Cambridge University Press. Pomerantz, Anita 1984 Agreeing and disagreeing with assessments: some features of preferred/dispreffered turn shapes. In: John Maxwell Atkinson and John Heritage (eds.), Structures of Social Action: Studies in Conversation Analysis, 57–101. Cambridge: Cambridge University Press. Psathas, George (ed.) 1979 Everyday language: studies in ethnomethodology. New York: Irvington. Rauch, Elisabeth 1992 Sprachrituale in institutionellen und institutionalisierten Text- und Gesprächssorten. Frankfurt am Main: Lang. Redder, Angelika 1990 Grammatiktheorie und sprachliches Handeln: “denn” und “da”. Tübingen: Niemeyer. Rehbein, Jochen 2001 Das Konzept der Diskursanalyse. In: Brinker et al. (eds.), 927–945.

Foundations: ethnomethodology and Erving Goffman

285

Sacks, Harvey 1972 An Initial Investigation of the Usability of Conversational Data for Doing Sociology. In: David Sudnow (ed.), Studies in social interaction, 31–74. New York: The Free Press. Sacks, Harvey 1974 [1972] On the Analysability of Stories by Children. In: Roy Turner (ed.), Ethnomethodology. Selected Readings, 216–232. Harmondsworth: Penguin Education. First published in John Gumperz and Dell Hymes (eds.), Directions in sociolinguistics: The ethnography of communication. New York: Holt, Rinehart, & Winston. Sacks, Harvey 1984 On doing ‘being ordinary’. In: John Maxwell Atkinson and John Heritage (eds.), Structures of Social Action: Studies in Conversation Analysis, 413–429. Cambridge: Cambridge University Press. Sacks, Harvey 1989 Lectures 1964–1965. Ed. by Gail Jefferson. Dordrecht: Kluwer Academic Publishers. Sacks, Harvey 1992 Lectures on Conversation. Vols I and II. Ed. by Gail Jefferson. With an introduction by Emanuel A. Schegloff. Oxford: Basil Blackwell. Sacks, Harvey, Emanuel A. Schegloff and Gail Jefferson 1978 A simplest systematics for the conversation of turn taking for conversation. In: Jim N. Schenkein (ed.), Studies in the organization of conversational interaction, 7–55. New York: Academic Press. First published Language 50: 696–735 [1974]. Schank, Gerd and Johannes Schwitalla (eds.) 1987 Konflikte in Gesprächen. Tübingen: Narr. Schegloff, Emanuel A. 1968 [1972] Sequencing in conversational openings. American Anthropologist 70: 1075–1095. Reprinted in John Gumperz and Dell Hymes (eds.), Directions in sociolinguistics: the ethnography of communication, 346–380. New York: Rinehart & Winston. Schegloff, Emanuel A. 1972 Notes on a Conversational Practice. Formulating Place. In: David Sudnow (ed.), Studies in social interaction, 75–119. New York: The Free Press. Schegloff, Emanuel A. 1979 The Relevance of Repair to Syntax-for-conversation. In: Talmy Givón (ed.), Syntax and Semantics 12: Discourse and Syntax, 261–286. New York: Acadamic Press. Schegloff, Emanuel A. 1987 Some sources of misunderstanding in talk-in-interaction. Linguistics 25: 201–218. Schegloff, Emanuel A. 1992 Introduction. In: Harvey Sacks: Lectures on Conversation. Vols I and II. Ed. by Gail Jefferson, ix–xii. Oxford: Basil Blackwell. Schegloff, Emanuel A., Gail Jefferson and Harvey Sacks 1977 The Preference for Self-Correction in the Organization of Repair in Conversation. Language 53 (2): 361–382.

286

Christine Domke and Werner Holly

Schegloff, Emanuel A. and Harvey Sacks 1973 Opening up closings. Semiotica 8: 289–327. Schenkein, Jim N. (ed.) 1978 Studies in the organization of conversational interaction. New York: Academic Press. Schmitt, Reinhold (ed.) 2007 Koordination. Analysen zur multimodalen Interaktion. Tübingen: Narr. Schneider, Wolfgang Ludwig 1997 Die Analyse von Struktursicherungsoperationen als Kooperationsfeld von Konversationsanalyse, objektiver Hermeneutik und Systemtheorie. In: Tilmann Sutter (ed.), Beobachtung verstehen, Verstehen beobachten. Perspektiven einer konstruktivistischen Hermeneutik, 164–227. Opladen: Westdeutscher Verlag, Schwitalla, Johannes 2001a Gesprochene-Sprache-Forschung und ihre Entwicklung zu einer Gesprächsanalyse. In: Brinker et al. (eds.), 896–903. Schwitalla, Johannes 2001b Konflikte und Verfahren ihrer Bearbeitung. In: Brinker et al. (eds.), 1374– 1382. Schwitalla, Johannes 2001c Beteiligungsrollen im Gespräch. In: Brinker et al. (eds.), 1355–1361. Selting, Margret and Elizabeth Couper-Kuhlen 2000 Argumente für die Entwicklung einer interaktionalen Linguistik. Gesprächsforschung – Online-Zeitschrift zur verbalen Interaktion 1: 76–95 Sinclair, John and Malcolm Coulthard 1975 Towards an analysis of discourse. London: Oxford University Press. Streeck, Jürgen 1983 Konversationsanalyse. Ein Reparaturversuch. Zeitschrift für Sprachwissenschaft 2: 72–104. Streeck, Jürgen 2009 Gesturecraft. The manufacture of meaning. Amsterdam: Benjamis. Sudnow, David (ed.) 1972 Studies in social interaction. New York: The Free Press. Tannen, Deborah 1984 Conversational Style: Analyzing Talk Among Friends. Norwood, NJ: Ablex. Tracy, Karen 1990 The many faces of facework. In: Howard Giles and W. Peter Robinson (eds.), Handbook of language and social psychology, 209–226. Chichester: John Wiley and Sons. Turner, Roy 1974a Introduction. In: Roy Turner (ed.), Ethnomethodology. Selected Readings, 7–12. Harmondsworth: Penguin Education. Turner, Roy (ed.) 1974b Ethnomethodology. Selected Readings. Harmondsworth: Penguin Education. Uhmann, Susanne 1997 Grammatische Regeln und konversationelle Strategien. Fallstudien aus Syntax und Phonologie. Tübingen: Niemeyer.

Foundations: ethnomethodology and Erving Goffman

287

Van Dijk, Teun 1972 Some Aspects of Text Grammars. The Hague: Mouton. Ventola, Eija 2001 Discourse studies in the English-speaking countries. In: Brinker et al. (eds.), 979–994. Werlen, Iwar 2001 Rituelle Muster in Gesprächen. In: Brinker et al. (eds.), 1263–1278. Widmer, Jean 1991 Goffman und die Ethnomethodologie. In: Robert Hettlage and Karl Lenz (eds.), Erving Goffman – ein soziologischer Klassiker der zweiten Generation, 211–242. Bern/Stuttgart: Haupt. Wodak, Ruth, Johanna Pelikan, Peter Nowak, Helmut Gruber, Rudolf de Cillia and Richard Mitten 1990 “Wir sind alle unschuldige Täter!”. Diskurshistorische Studien zum Nachkriegsantisemitismus. Frankfurt am Main: Suhrkamp.

10.

Pragmatics in Habermas’ Critical Social Theory Maeve Cooke

1.

Introduction

Sociology, as the scientific or systematic study of society, emerged as an independent academic discipline only towards the end of the nineteenth century. From the beginning it has been marked by a dispute as to the method of inquiry appropriate to it. A central question is whether its methodology should model that of the natural sciences or that of humanistic and cultural studies. The first position is often attributed to Auguste Comte and nineteenth and twentieth century positivism; it holds that social phenomena can be subjected to observable natural laws. The second position is associated with Wilhelm Dilthey and nineteenth and twentieth century historicism and hermeneutics; it asserts an essential difference between the methodology of the natural sciences and humanistic and cultural studies, emphasizing the interpretative component of social scientific inquiry. The first position adopts an external perspective on the production of social order, viewing it as a quasi-natural process that can be observed in its empirical regularities and explained by way of nomological hypotheses. The second position adopts an internal perspective on the production of social order, viewing it as a meaningfully structured reality for social agents that should be interpreted in terms of intentional actions. To be sure, the lines of division are by no means clear cut. Max Weber, for example, defines sociology as a science that attempts the interpretative understanding of social action in order thereby to arrive at a causal explanation of its course and effects (Weber 1969). In line with this, he holds that rational interpretation, in the sense of objective or impartial appraisal, is possible, when the agent and her interpreter equally accept the standards of judgement as valid (Habermas 1988). Moreover, Weber is ambivalent on the question of whether the methods of the social sciences should model those of the natural sciences. In this regard we find striking differences in attitude and emphasis in his writings (McCarthy 1978: 141–145). We cannot explore this debate further here. For our present purposes, it is sufficient to note that, today, there is widespread, if not universal, agreement that the methods of the natural sciences cannot simply be mapped onto sociology. Furthermore, among those who accept this, a significant number agree that the difference is due in large part to the intentional component of human action.1 By this is meant, first, that the concept of society is internally linked with the concept of human action and, second, that in the case of such action it is always in principle

290

Maeve Cooke

possible to ask what the agent(s) intended – meant – when carrying it out. The implication here is that a description of a particular human action that failed to consider the intentions of the agent would lack an essential ingredient.2 This implies in turn that intention, understood as subjective meaning, is a sociological primitive: an irreducible element of sociological analysis. It should also be noted that even if we accept the reference to subjective meaning as constituting an essential methodological difference between the natural sciences and sociology, a number of important questions remain open, for example, the questions of whether causal explanations of social phenomena are possible and the degree to which objectivity is achievable in sociology. Thus, accepting the irreducibility of subjective meaning does not commit the sociologist to a purely internal perspective on social phenomena and require her to abstain from empirical-analytic analyses and causal explanations. However, it presents her with the challenge of developing a theoretical framework, and corresponding methodology, that negotiates the tensions between the external and internal perspectives. In the early 1970s, as part of an endeavour to meet this challenge, Jürgen Habermas, a philosopher and sociologist in the Frankfurt School tradition of critical social theory, turned to pragmatics.3 The importance of pragmatics in contemporary sociology is largely due to Habermas’ contribution to the study of society. This contribution is so significant that no sociologist today can ignore his social theory and, by extension, the programme of pragmatics underpinning it. While Habermas owes a significant debt to his colleague Karl-Otto Apel, who introduced him to pragmatics in the late 1960s, Apel’s work has been more influential in philosophy than social theory (Apel 1980). Furthermore, unlike Habermas, Apel did not develop his reflections on pragmatics into a systematic programme. For these reasons, we will take Habermas as our reference point in the remainder of our discussion. In turning to pragmatics, Habermas’ concerns were not only methodological. He sought, in addition, to provide an answer to central questions of sociology, such as how we should conceptualize social action, how social order is possible and how social order is reproduced. A further, equally important, aim was to provide a basis for the rational critique of society without appealing to metaphysical notions of truth, justice or the good. The Theory of Communicative Action, first published in two volumes in 1981, attempts to meet these aims. In the following, we will examine the role of pragmatics, first, from the point of view of method, where we will consider the conceptual strategy he deems most appropriate for sociological inquiry; then, from the point of view of the interconnected questions of social action, social order and social reproduction; finally, from the point of view of social critique, where we will look at the role pragmatics plays in critique of social pathologies and in developing a postmetaphysical conception of reason, which is the normative reference point for this critique. In the concluding section we will look briefly at some challenges to Habermas’ appeal to pragmatics.

Pragmatics in Habermas' critical social theory

2.

291

Method

In the course of the 1970s Habermas develops a theory of pragmatics that, as indicated, is supposed to help him answer central sociological questions relating to social action, social order and social reproduction as well as provide a normative basis for his critical theory of society. In a preliminary essay based on lectures delivered at Princeton University in 1971, he focuses specifically on methodology, identifying a number of metatheoretical issues as centrally important in the endeavour to develop a conceptual strategy appropriate for sociology.4 In each case, in line with his concern to avoid a simple decision for either explanation or interpretation, Habermas adopts a position that seeks to retain the most valuable elements in the apparently conflicting approaches. For example, in his reflections on the debate between methodological individualism and holism, he does not straightforwardly endorse either of these approaches but attributes to them complementary strengths and weaknesses that make each useful for different purposes and in different contexts (Habermas 1971: 13–14). Most relevant in the present instance is the methodological decision he makes in favour of subjective meaning as a sociological primitive (Habermas 1971: 3–10). This decision leads him to view society as a continuous process of generating a meaningfully structured reality, whereby meanings are produced according to an underlying system of abstract rules and have a sociologically relevant, intentional component. Having generally endorsed theories that take subjective meaning as an irreducible element of sociological analysis, Habermas now makes a distinction within this set between constitutive and communicative theories.5 It is at this point that pragmatics enters the scene. In his account, constitutive theories base the process of the generation of meaning on a transcendental subject. Some, taking their lead from Husserl’s account of the constitution of the everyday world of lived experience (the “lifeworld”), conceive of this transcendental subject as an intelligible ego modelled on the empirical individual subject. Others, taking their lead from Hegel and Marx, conceive of it as a species-subject that constitutes itself in history. In each case, they encounter the problem of how to move from the level of experience to the level of society. More precisely, they run up against the problem of how adequately to conceptualize intersubjective meanings within a subject-object model of cognition and action. In models of this kind, knowledge and action are conceived monologically, as the result of a relationship between a subject and an object that is distinct from it. However, if meaning is conceived of essentially as the product of a relationship between a subject and some distinct object, commonly shared, intersubjective meanings appear to have a secondary status. From a social-theoretical point of view this gives rise to considerable difficulties. One difficulty is that it makes it hard to explain how society, as a process of generating intersubjective meanings, can be granted the primordial status it has in phenomenological accounts. This weakness of constitutive theories, together with other, related ones, prompts Ha-

292

Maeve Cooke

bermas to look to communicative theories, which construe meaning as intersubjectively generated from the beginning. In this preliminary essay, Habermas finds himself in the unwelcome situation of being unable to refer to an existing body of work that would support the kind of communicative theory he is seeking; he admits that the theoretical approach that he deems methodologically most appropriate for social scientific inquiry has not yet been adequately developed. However, he is clear as to the direction it is necessary to take. What is required in his view is a sociological theory that would start from the premise that everyday communication, as a process of generating shared meanings, is the basis for social action (and socialization); the use of ordinary language would be seen as constituting the intersubjective relations within which action takes place (as well as the personality structures of the subjects of action). His name for this project is “universal pragmatics.” Habermas’ programmatic essay “What is Universal Pragmatics?” first appeared in German in 1976. The adjective “universal” indicates the difference between his reconstructive approach to the study of linguistic behaviour and empirically oriented analyses. In his account, empirical-pragmatic approaches set themselves the task of describing the speech acts typical of a certain milieu, which may then be analyzed from sociological, ethnological and psychological points of view. Habermas favours a broader perspective. He aligns himself with general pragmatic theory, which is concerned not with specific instances of language use, but rather with formally reconstructing the rule system that underlies the ability of a subject to utter sentences in any relevant situation whatsoever (Habermas 1998: 54). Accordingly, he characterizes the task of universal pragmatics as being to identify and reconstruct universal conditions of possible Verständigung, which we can translate here as mutual understanding6 (Habermas 1998: 21). Soon after the publication of this essay, however, he distanced himself from the name “universal pragmatics.” In a footnote to the English translation, published three years later in 1979, he writes that he is no longer happy with his original terminology, suggesting that his project instead be named “formal pragmatics”. In opting for the word “formal” he wishes to emphasize the continuity between his approach and formal semantics, notwithstanding his focus on the use of language in speech acts or utterances as opposed to the semanticists’ concern with the properties of isolated sentences.7 His choice of the word “formal” also makes sense in the context of his adoption of the procedure of rational reconstruction which, as indicated, is the method he attributes to his universally oriented, pragmatic approach. By “rational reconstruction” he means the systematic reconstruction of the intuitive knowledge of competent subjects (Habermas 1998: 29–41). Reconstructions relate to pretheoretical knowledge of a general sort, to universal capabilities and not merely to the competences of particular groups or individuals. Thus, what begins as an explication of meaning ultimately aims at the reconstruction of species competences. Habermas sees it as the great merit of Noam Chomsky to have developed this idea

Pragmatics in Habermas' critical social theory

293

in the case of grammatical theory (Chomsky 1969). He also commends Jean Piaget and Lawrence Kohlberg for, respectively, their theories of cognitive and moral development, which, like Chomsky’s, rationally reconstruct the pretheoretical and implicit knowledge and competences of speaking and acting subjects in general. In distinguishing his formal-pragmatic approach from an empirical-pragmatic one, Habermas does not want to deny the importance of empirical validation. As indicated, reconstructive theories belong to the category of general theory in the sense that they seek to reconstruct the rule system that underlies the ability of a subject to generate well-formed sentences in any language whatsoever. At the same time, they belong to the category of empirical theories. Their explications of general structures and universal conditions have a merely hypothetical status and must be subjected to the usual methods of testing. Thus, Chomsky’s and Habermas’ reconstructions of linguistic competence are open to checking against speaker’s intuitions, scattered across as broad a sociological spectrum as possible (Habermas 1984:138). As is appropriate for scientific investigations of this sort, the checking procedure is circular: it moves in a circle between theory formation and a more precise rendering of the object domain (Habermas 1998: 40). From this it is clear that Habermas places reconstruction, as an explicative exercise, on the side of interpretation rather than explanation, thereby connecting it with an internal rather than external perspective on social phenomena. The point of reconstructions is neither to describe behaviour nor to systematize their findings in law-like explanations, but rather to bring to consciousness implicit knowledge through a maieutic method of interrogation, which involves the choice of suitable examples and counterexamples, the use of contrast and similarity relations, translation, paraphrase and so on (Habermas 1998: 40). This is not to deny that the procedures employed in constructing and testing hypotheses, in appraising competing reconstructive proposals and in gathering and selecting data, are not in many ways like the procedures used in the nomological sciences (Habermas 1998: 46). Like such sciences, reconstructive theories usually operate at a high level of abstraction, seeking to explicate basic conceptual frameworks and to formalize assumptions built into everyday behaviour. The key difference, to recall our earlier discussion, is that reconstructive social theories take subjective meaning as an irreducible element of social scientific analysis whereas theories that adopt a purely external perspective do not. We have seen that Habermas turns to pragmatics in order to find a method of sociological analysis that takes subjective meaning as a sociological primitive, while avoiding the weaknesses of constitutive models which attribute the process of generating a meaningfully structured reality to some sort of transcendental subject. His aim is to develop a communicative model that replaces the subject-object paradigm underlying constitutive theories with an intersubjective paradigm in which the generation of shared meanings is seen as fundamental in the constitution of social reality. He sees the use theory of meaning developed by Ludwig Wittgen-

294

Maeve Cooke

stein in his later work as making an important contribution towards developing such a model (Wittgenstein 2001). But for reasons that will become clearer in the next section, he finds speech act theory as developed by Austin and by Searle (in his early work) more fruitful for his purposes. Speech act theory shares with reconstructive language analysis a concern to explicate the rules that a competent speaker must master in order to form grammatical sentences and to utter them in an acceptable way. However, by contrast with Chomskian linguistics, instead of starting from the postulate of an implicit ability to produce sentences, speech act theory starts from the assumption of an implicit ability to employ sentences in speech acts (utterances). Thus, it is concerned with a competent speaker’s ability to communicate as opposed to produce a grammatical sentence. Accordingly, a general theory of speech acts explicates the system of rules that competent speakers master when they fulfil the conditions for the successful employment of sentences in utterances, no matter to which particular languages the sentences may belong and in which random contexts the sentences may be uttered (Habermas 1998: 47). The importance of speech act theory in the development of Habermas’ conceptual strategy is twofold. First, by virtue of its focus on utterances as communicative actions, it enables him to show how the meaning of speech acts is produced by way of an action involving speaker and hearer(s); this provides support for his shift in paradigm from a subject-object based, monological model of meaning production to an intersubjective one. Second, it enables him to show how the production of meaning is internally connected with a number of different kinds of claims to validity. The connection between meaning and validity in a number of dimensions emerges from speech act theory’s analysis of the successful employment of sentences in utterances. Habermas contends that for an utterance or speech act to be successful it must count as true for the speaker and hearers, in the sense of representing something in the world, must count as normatively right, in the sense of conforming to socially recognized expectations and must count as truthful, in the sense of expressing something intended by the speaker. As we shall now see, the connection between meaning and validity claims is of crucial importance in Habermas’ accounts of social action, social order and social reproduction and in his proposal for a postmetaphysical conception of reason that permits social critique.

3.

Social action, social order and social reproduction

Formal pragmatics forms the basis for the kind of communicative theory that Habermas thinks necessary in order to avoid the weakness of constitutive social theories and those that share their subjectivist orientation. In his view, such theories not only have difficulties adequately conceptualizing meanings shared in common; they are also not well suited to the task of adequately conceptualizing social action,

Pragmatics in Habermas' critical social theory

295

social order and social reproduction. This is because they tend to privilege instrumental action, understood as action undertaken as a means to achieving an end that is specified in advance of the undertaking. Accounts of social order and reproduction that take instrumental action as the fundamental mode of social action run up against the problem of how to explain long-term social cohesion, when the preservation and reproduction of the social order in question does not evidently equally enable everyone to obtain their particular ends (cf. Habermas 1998: 234). A central aim of Habermas’ theory of communicative action is to show that instrumental action is not the primary mechanism in social cohesion and social reproduction. He draws on formal-pragmatic analyses of everyday language in order to achieve this aim. As the background to our discussion in this section and the next, it is important to note that in The Theory of Communicative Action Habermas offers an account of the developmental path of modern societies in terms of two categorially distinct modes of societal integration. In what he calls the “lifeworld”, co-ordination of action takes place primarily by way of communicative action, which relies on intentionally acting human subjects who establish validity-based relationships with one another. In the “system”, by contrast, co-ordination of action takes place primarily by way of the functional interconnection of action consequences, bypassing the intentions of the agents concerned. For Habermas, rationalization processes, by which he means the development of the internal logic of a particular mode of societal integration, are possible and desirable in each case. The rationalization of the system refers to increasing complexity and increasing capacity to take on steering functions such as material reproduction and administration. By contrast, the rationalization of the lifeworld is a matter of communicative rationalization: the development of the internal logic of communicative action. It takes place in the domains of cultural tradition, social integration and socialization and entails the increasing independence of processes of justification from traditional normative contexts of validity, together with increasing reliance on action oriented to achieving understanding (Verständigung). Communicative action is a form of social interaction in which the plans of various agents are co-ordinated through an exchange of speech acts. Drawing on the speech act theory of Austin and the early Searle, Habermas maintains that the success of a speech act is not just a matter of comprehending the sentence embedded in the utterance, but of responding in the appropriate manner to what the speaker does in performing the action. In addition to commending these speech act theorists for emphasizing the action component of linguistic understanding, he praises them for directing attention to modes of language use other than the assertoric and descriptive modes. In this respect, too, they take their lead from the later Wittgenstein. However, Habermas finds speech act theory more promising than Wittgenstein’s use theory of language, due to its assertion of a connection between meaning and validity in a context-transcending sense. For Habermas, truth and truth-analogous

296

Maeve Cooke

concepts such as moral rightness have a context-transcending aspect in the sense that they assert a validity that transcends particular places and times. This is one reason why he refers to them as universal validity claims.8 The connection between meaning and universal validity disappears from view in Wittgenstein’s theory, since validity is held to be relative to particular, historically specific, language games. To be sure, Habermas finds Austin’s and Searle’s approaches to universal validity claims unsatisfactory. He criticizes Austin for blurring the distinction between clear cut universal validity claims such as truth and normative rightness and a host of other evaluative criteria (Habermas 1998: 290–291). He admonishes Searle for acknowledging only the representational function, together with the claim to truth corresponding to it, as the only clear cut universal validity claim (Habermas 1998: 291–292). Against Austin and Searle Habermas seeks to demonstrate that everyday language use is connected with three clear cut universal validity claims. This is one of the two main respects in which Habermas moves beyond speech act theory. The other is his emphasis on the relationship of obligation into which the speaker and hearer(s) enter with the performance of a speech act. Taking his initial lead from Austin, he finds it instructive to consider what a speaker does in saying something. He begins by discussing Austin’s distinction between locutions and illocutions (Austin 1962). Initially Austin reserved the term “locution” for the propositional content of what is said and the term “illocution” for the performative act of uttering sentences with a propositional content. He later reformulated the distinction as one between constatives and performatives, defining constatives as speech acts in which we concentrate on the truth or falsity of what is said and performatives as speech acts in which we concentrate on the illocutionary force of the utterance (Habermas 1998: 66–81). Habermas finds this way of distinguishing between constative and performative speech acts unhelpful, proposing instead that we distinguish between utterances on the basis of the kinds of validity claims they raise. In his reformulated version of Austin’s classificatory schema, constatives involve the cognitive use of language and raise claims to truth (examples are assertions and descriptions); here we thematize the content of the utterance as a statement about what is or could be the case. Regulatives involve the interactive use of language and raise claims to normative rightness (examples are promises, requests and warnings); here we thematize the relationship of obligation into which speaker and hearer enter with the speech act. Habermas adds a third category, expressives, to these revisions of Austin’s schema; as the name indicates, these involve the expressive use of language and raise claims to truthfulness (examples are confessions and disclosures); here we thematize the truthfulness or sincerity of the speaker’s intentions. He proposes, further, that we understand the illocutionary force of the utterance – what the speaker does in performing any speech act – as a rationally binding force (bindende Kraft), in the twin sense of binding and compelling (Habermas 1998: 81–88). This rationally binding force is held to be operative in all three modes of language use: the constative, regulative and the expressive.

Pragmatics in Habermas' critical social theory

297

This, then, is Habermas’ answer to the question of what we do in performing a speech act: we enter into a relationship of rationally-based obligation with the hearers. The obligation is rationally-based since it arises from the speaker’s undertaking to support what she says with reasons. He conceptualizes this in terms of a validity claim raised by the speaker. The speaker claims that what she says is true or that she is normatively right to say it or that she is sincere in saying it, thus raising a claim to validity that calls for a “yes” or “no” response from the hearer. The success of the interaction depends on the speaker’s ability to supply reasons to support the validity of the claim, if challenged, and on the hearer’s ability to provide reasons in support of her “yes” or “no”. For Habermas, therefore, the illocutionary force of an utterance is not simply an aspect of its meaning conveyed by a performative verb; it is a co-ordinating power that is based on reciprocal obligations to provide reasons. This is what he means by the rationally binding force of linguistic interaction. Habermas holds, furthermore, that every speech act raises all three claims simultaneously. More precisely, he holds that with every speech act, the speaker raises one of these three kinds of validity claims directly, while the other two are raised indirectly. Thus, for example, if a speaker states she has no money to pay the rent, she is raising a claim to truth that calls for a “yes” or “no” response from the hearer(s) – or for an abstention on grounds of insufficient knowledge of the facts. At the same time she raises a claim that her utterance is appropriate in the context in question (“normatively right”) and that she is sincere in uttering it. In this case, a claim to truth is raised directly and claims to normative rightness and sincerity are raised indirectly. If a speaker requests a hearer to lend her some money, she raises a claim to normative rightness directly and claims to truth and truthfulness indirectly (Habermas has clarified that indirect claims to truth should be understood as claims that certain existential presuppositions are met, 1998: 146). A speaker who confesses that she has spent all her money raises a claim to truthfulness directly and claims to normative rightness and truth indirectly. The thesis that every speech act raises three claims to validity simultaneously is important for Habermas’ attempt to show that an orientation to a plural – multi-dimensional – conception of validity is built into everyday language use. As we shall see, this plural conception of validity is the basis for his non-logocentric conception of communicative rationality, which does not privilege propositional truth but conceives of reason as an interplay of truth, moral rightness and truthfulness/authenticity. The thesis of three simultaneous validity claims is also important from the point of view of social integration and reproduction. Its purpose is to show that the three validity claims raised with every speech act amount to three structural components of speech acts that correspond to the three functions of the communicative use of language: (a) representing states of affairs (or something in the world that confronts the speaker), (b) entering into a relationship with a hearer and (c) giving expression to the speaker’s intentions or experiences. Referring to the communi-

298

Maeve Cooke

cation-theoretic approach expounded by the German psychologist Karl Bühler (Bühler 1934), Habermas posits these as three mutually irreducible but internally connected linguistic functions.9 His thesis is that these linguistic functions, carried out with every speech act, serve to maintain and reproduce the lifeworld in its three domains of cultural tradition, social integration and socialization. For this, however, the thesis of three simultaneous validity claims is not sufficient. Even if successful (see Cooke 1994: 90–91), it would show merely that the communicative use of language fulfils functions of representation, of establishing a relationship with the hearer and of expression; it would not show that the communicative use of language is of fundamental importance in maintaining and reproducing the lifeworld. Showing the latter is central to Habermas’ critical project. We will recall his claim that the colonization of the communicatively regulated domains of the lifeworld by the functional rationality of the system is a deeply troubling pathology of modernity. If he were unable to show that communicative action is indispensable for the purposes of establishing, maintaining and reproducing social order, he would lack grounds for referring to the shrinking of the communicatively regulated domains of the lifeworld as a pathology and for adopting the metaphor of colonization to describe the expansion of functional rationality into the communicatively regulated domains. In order to meet this challenge, Habermas seeks to demonstrate the primacy of communicative action over instrumental (and, in particular, strategic) action. Once again, he employs a formal-pragmatic strategy. This entails showing the primacy of the communicative use of language over other modes, specifically the strategic one. Just as the communicative use of language amounts to communicative action, the strategic use of language amounts to strategic action. Strategic actions are instrumental insofar as one agent uses another as a means of achieving his respective goals; he treats other persons as though they were objects or entities in the physical world. In contrast to communicative action, in which the success of the interaction is not at the disposal of the individual agent but depends on all parties involved, strategically acting agents are not dependent on the co-operation of others (Habermas 1998: 220–227). While Habermas acknowledges that strategic action, like communicative action, is a mode of action within the lifeworld, he insists that it is a merely secondary mode of social integration. His position, in other words, is that communicatively regulated interactions must predominate in the lifeworld if this is to maintain and reproduce itself. He seeks to show the priority of communicative action by demonstrating that the strategic use of language is parasitic on the communicative one.10 In The Theory of Communicative Action Habermas attempts to show that the use of language oriented to reached understanding (Verständigung) is the primary or original mode of language use by drawing on Austin’s distinction between locutions and perlocutions (Habermas 1998: 122–129) In Austin’s account, the locutionary act is the act of saying something, the illocutionary act is the action per-

Pragmatics in Habermas' critical social theory

299

formed by saying something and the perlocutionary act is the effect the speaker produces upon the hearer. Initially, Habermas took this to mean that perlocutionary acts are an indication of the integration of speech acts into contexts of strategic action (Habermas 1998: 126). This interpretation neglects the importance of distinguishing between different types of perlocutions. It is important, above all, to distinguish between the contingent, often unintended effects on the hearer(s) of particular speech acts and the intended, but concealed effects of particular speech acts. An example of the former is the pleasure a speaker gives to his hearer’s husband by lending him money for the rent (for the act might have given him displeasure); an example of the latter is the speaker’s intention to use his loan of money to blackmail the hearer. Habermas has subsequently modified his position, clarifying that only perlocutionary effects of the latter kind are relevant for his argument (Habermas 1998: 329–333). He sees them as examples of latently strategic action. As illustrated by the case where the speaker intends to blackmail the hearer by lending her money, the success of latently strategic action depends on the speaker’s success in hiding his intentions from the hearer; he must successfully pretend that he is using language with an orientation towards mutual understanding (Verständigung), as opposed to strategically, using the interaction with the hearer as a means to furthering his own particular ends. The latently strategic mode of language use is by definition parasitic on the communicative use (Cooke 1994: 24).11 Thus, demonstrating its secondary character will not be sufficient for Habermas’ purposes. The more interesting and difficult case is the manifestly strategic use of language. In this mode of language use, the presupposition of an orientation towards reaching understanding is overtly suspended (Habermas 1998: 225–226). A speaker who backs up his demand that a hearer lend him some money with the threat of sanction (for example, by threatening to harm her daughter) illustrates such language use. In order to make good his claim that the communicative use of language is the primary mode of language use, Habermas has to show that this manifestly strategic use of language, too, is parasitic on the communicative one. This part of Habermas’ argument is underdeveloped (Cooke 1994: 23–26). In this section we have considered the importance of formal pragmatics in Habermas’ accounts of social action, social order and social reproduction. We have seen that its role is threefold: to demonstrate that everyday linguistic communication has an in-built rationally binding force, to show that this binding force is operative in three dimensions corresponding to the three linguistic functions of representation, establishing a relationship with the hearer(s), and expression, and to demonstrate that the communicative use of language is the primary mechanism in establishing, maintaining and reproducing the social order. In the next section, we will look at the importance of formal pragmatics for Habermas’ project of social critique.

300 4.

Maeve Cooke

Social critique

Habermas is first and foremost a critical social theorist. His sociological concerns are intimately bound up with his concern to diagnose and criticize the pathologies of modernity. We have already mentioned the colonization of the lifeworld as one of these pathologies. Habermas, drawing on Weber, sometimes refers to this as the modern loss of freedom (Habermas 1987: 301–302). The second principal pathology he diagnoses is a loss of meaning (here, too, he draws on Weber’s analysis of modernity). By this he means the cultural impoverishment of the lifeworld that results from the differentiation of the cultural value spheres of science, law/morality, and art under conditions of modernity; these value spheres become fields for the specialized treatment of questions of truth, normative rightness and beauty/authenticity by experts. While these specialized discourses promote a growth in cognitive-instrumental, moral-practical and aesthetic-expressive knowledge, thus giving rise to significant learning processes, they are also a potential source of pathology. This is because with progressive modernization, the gap between the expert cultures and the general public becomes ever wider. Learning processes within the expert cultures do not automatically flow back into the communicative practices of the lifeworld, contributing to their renewal and regeneration, but remain cut off from them. Lacking revitalizing impulses from the specialist cultures (or, indeed, from other sources12), everyday communicative practices threaten to become impoverished. His critique of these pathologies is guided by a corresponding normative vision. At the core of this vision is a conceptual triad that has defined his work from his earliest major publication The Structural Transformation of the Public Sphere (Habermas 1989): the concepts of the public sphere, discourse and reason. The interplay of these concepts yields a utopian vision of a communicatively rationalized lifeworld (Habermas 1991: 64–69). In the projected society, cultural traditions would be reproduced through processes of intersubjective evaluation of validity claims, legitimate orders would be dependent on critical and open argumentative practices for justifying laws/norms and for making political decisions, and individual identities would be self-regulated through processes of critical reflection. We must assume, furthermore, that in such a communicatively rationalized lifeworld there would be ongoing processes of semantic renewal in the form of continuous flows of knowledge from the specialized, expert cultures into the communicative practices of everyday life.13 This utopian projection of a rationalized lifeworld is in turn part of a vision of society in which the development of systems rationality and communicative rationality would be evenly balanced: systems rationality would no longer encroach into the communicatively regulated domains of the lifeworld. Formal pragmatics contributes to Habermas’ critical project in two main respects. His formal-pragmatic investigations are intended, first, to provide a normative reference point for the critique of the two principal pathologies of modern

Pragmatics in Habermas' critical social theory

301

societies: the colonization of the communicatively regulated domains of the lifeworld by systems rationality and its cultural impoverishment due to the emergence of specialized discourses among experts in the areas of science, law/morality and art. They are intended, second, to provide a normative reference point for criticism of the conduct and outcomes of actual practices of deliberation in everyday life. In each case, the normative reference point is given the name “communicative rationality”. As we have seen, Habermas’ thesis of the colonization of the lifeworld makes the claim that the uneven developmental path of modern societies is pathological. What is required, instead, is an evenly balanced pattern of societal rationalization. Habermas expresses the idea of a non-pathological, evenly balanced process of societal rationalization in terms of the harmonious, free interplay of the three cultural value spheres of science, morality/law and art, in which no one sphere would predominate at the expense of the other (Habermas 1984: 240). His formal-pragmatic investigations are used as a normative underpinning for this idea. The relevant formal-pragmatic finding is that everyday linguistic communication fulfils three mutually irreducible functions: representation, establishing a relationship between speaker and hearer(s), and expression; as we have seen, this is the conclusion Habermas draws from his thesis that the speaker raises three validity claims simultaneously with every speech act. He asserts, in addition, that these three linguistic functions can be correlated with the kind of knowledge defining each of the three cultural value spheres: scientific knowledge, moral-practical knowledge and aesthetic-expressive knowledge. He then makes a further correlation between the three cultural value spheres and the three moments of a once substantively unified concept of reason, which have become separated from one another under conditions of modernity;14 the image of harmonious, free interplay is extended to these three moments of reason, evoking an idea of reason whose unity is defined purely formally as opposed to substantively, and that is non-logocentric, expanding the traditional philosophical focus on theoretical reasoning and truth to include other dimensions of validity (such as justice and beauty) on an equal footing. If the colonization thesis suggests as its opposite evenly balanced processes of functional rationalization and communicative rationalization, the thesis of cultural impoverishment suggests as its opposite the continuous feedback back into everyday communicative practices of the knowledge produced by experts in their various fields. However, whereas Habermas evidently wishes to use his formal-pragmatic analyses to provide empirically based, normative support for the idea of evenly balanced processes of societal rationalization, he gives no indication of any formal-pragmatic finding that would lend support to the idea of the free flow of knowledge between expert cultures and the lifeworld. Nor is it easy to see which part of his analyses could help him in this regard. In addition to these two principal pathologies he diagnoses and attributes to deficiencies in modern processes of societal rationalization, Habermas identifies a

302

Maeve Cooke

number of further potential pathologies (Habermas 1987: 142–144); by contrast with the pathology of colonization, which threatens it from the outside, these potential disturbances arise within the lifeworld (in this they are similar to the pathology of cultural impoverishment). In addition to the loss of meaning that results from the encapsulation of specialist cultures, he mentions the anomie that results from disturbances in the process of social integration and the mental illnesses (psychopathologies) that result from disturbances in the process of individual socialization. At least initially, Habermas held that formal pragmatics could play a role in the critique of these disturbances. In The Theory of Communicative Action and in the preliminary studies that led up to it, he suggested that formal-pragmatic analyses could help to establish a normative basis for criticizing the phenomena in question as pathologies. The specific strategy he envisaged was a formal-pragmatic account of systematically distorted communication (Habermas 1974). Despite occasional remarks in his writings up to the middle of the 1980s that indicate that he had not yet abandoned this strategy, he has not since then made any serious attempt to develop such an account. His key intuition seemed to be that systematically distorted communication violates what formal pragmatics shows to be the internal organization of speech, by undermining the three linguistic functions of representation, establishing a relationship between speaker and hearer(s) and expression. The violation, which takes the form of disconnecting meaning from validity, speaking from acting and/or meaning from intention, is due to the overwhelming pressure exerted on the internal organization of speech by its external organization; by the external organization of speech he means the regulation of the normative context in which the communicative interaction takes place: for example, the determination of who is allowed to take part in which discussion, who can initiate topics, who can bring the discussion to a close, who can contribute and in which order, how the topics are ordered, and so on. To be sure, in order to develop a convincing account of systematically distorted communication, Habermas would have had to embed it within his critical account of societal rationalization as colonization (at the time of publication of his 1974 essay, he had not yet provided such an account); specifically, he would have had to show how the infiltration of functional rationality into the communicatively regulated domains of social life affects the external organization of speech, contributing to a violation of its internal organization. Formal-pragmatic arguments would not be helpful here. Nor would they be able to ground the normative force of the thesis implicit in Habermas’ account of systematically distorted communication: the thesis that the regulation of the context in which speech takes place should conform to demanding normative standards of inclusivity, fairness and openness. In short, even a well-developed formal-pragmatic account of systematically distorted communication would show only what the internal organization of speech and its disruption would look like; on its own it would not be sufficient to provide the sought-for normative basis for critique of disturbances such as anomie and psychopathologies (Cooke 1994: 148–150).

Pragmatics in Habermas' critical social theory

303

As indicated, the normative reference point for Habermas’ diagnosis and critique of social pathologies is the concept of communicative rationality. Habermas sets himself the task of elaborating a concept of reason that responds to two (closely connected) challenges facing contemporary critical social theorists: maintaining a dialectics of immanence and transcendence and maintaining concepts of validity that transcend times and spaces while acknowledging the historical situatedness of all actual conceptions of context-transcending validity. The first is specific to theorists in the Frankfurt School tradition who take seriously the Hegelian thesis that reason is embedded within history itself; the second arises for any social theorists today who seek to uphold an idea of reason while taking on board the post-Enlightenment suspicion of transcendental ideas such as truth, justice and the good. Formal pragmatics plays a role in meeting each of these challenges. The term “Left-Hegelian” is sometimes used to characterize modes of critical social theory that, taking their lead from Marx and Lukács, are concerned to re-articulate Hegelian speculative thinking in a materialist way. Most critical social theorists in the Frankfurt School tradition fall into this category. The aspect of Hegel’s thinking most relevant here is his emphasis on reason’s simultaneous immanence and transcendence. Hegel endeavours to show a) that reason is immanent to human history in the sense that the latter is a process in which reason actualizes itself by way of the progressive overcoming of contradictions and b) that reason always transcends its particular historical self-actualizations.15 The attempt to maintain a dialectical relationship between immanence and transcendence is taken up explicitly by Theodor W. Adorno, Max Horkheimer, Herbert Marcuse and other first generation critical social theorists (see, for example, Adorno 1981) and continues to play a central role today in the writings of Axel Honneth, the most prominent member of the third generation of critical social theorists (Honneth 1994, 2003). Habermas, the most prominent member of the second generation is less inclined to describe his project in explicitly Left-Hegelian terms; nonetheless, a concern to sustain a dialectics of immanence and transcendence is certainly also one of its most important elements (see, for example, Habermas 1996:1–25). In Habermas’ work, this amounts to the elaboration of a conception of reason that is at once context-transcending, in the sense of transcending times and places, thus extending to human beings in general, and immanent, in the sense of being anchored in practices within everyday historical practices, specifically in everyday practices of using language communicatively. Habermas’ theory of communicative action is of crucial importance to his attempt to elaborate a conception of reason that is at once immanent and transcendent. We have seen that he uses a formal-pragmatic analysis of everyday language use as the basis for an account of social action, order and reproduction that has a normative as well as an empirical content. It is empirical in the sense that it uses empirical evidence to reconstruct the relevant structural features of everyday lan-

304

Maeve Cooke

guage use; it is normative in the sense that the resulting account of language is used as the basis for a normative idea of a communicatively rationalized lifeworld that co-exists harmoniously with a well-functioning system. An analogous interplay of the empirical and normative characterizes the concept of communicative rationality. This concept expresses the rational potential implicit in – immanent to – everyday communicative action. As a practice of raising and responding to validity claims, communicative action is conceptually tied to more or less rudimentary practices of argumentation. Habermas maintains that the validity claims raised in communicative action have a context-transcending, critical force that derives from idealizations built into the concept of argumentation (see Cooke 2006: 47–53). One set of idealizations refers to the procedure of argumentation: to the normative promise contained in the general understanding of how argumentation should be conducted. Examples here are the idealizing presuppositions that all participants are motivated by a concern to find the single right answer, that no force is exerted except that of the better argument, that all relevant arguments are considered, that no one should be excluded from the deliberation on grounds such as race, class and gender and that everyone’s voice should be given an equal hearing. Another set refers to the outcome of argumentation: to the normative promise contained in the general understanding of the aim of argumentation. An example here is the idealizing presupposition that a discursively reached agreement warrants the truth of propositions and constitutes the justice of norms. In both cases, the tension between the normative promise contained in these idealizations and what actually happens in everyday communicative practices provides a basis for criticism: in the one case, they permit criticism of the ways in which the outcomes of argumentation are reached; in the other case, they permit criticism of the outcomes from the point of view of an idea of context-transcending validity (e.g. truth or justice). The critical power of communicative action resides primarily in the tension between an idealized notion of argumentation and what happens in actual empirical practices. It should be noted that this critical power is not restricted to a particular socio-cultural context. Since it is grounded in universal features of language use, it expresses a critical perspective with context-transcending force, in the sense that its validity would have to be accepted by everyone, everywhere, irrespective of sociocultural context. It should be noted, too, that the critical power of the concept of communicative rationality refers to an idealized practice of communication that has the status of a methodological fiction. Habermas warns against essentialist misunderstandings of the “ideal speech situation”, emphasising that it is merely an idealizing projection and best understood as enabling a fruitful thought experiment (Habermas 1996: 322–323). In this way, his formal-pragmatic analyses of language enable Habermas to assert an idea of reason that is at once immanent to history, in the sense that it is built into everyday practices of linguistic communication, and transcendent, in the sense that it provides a reference point for critique of all actually existing communicative practices.

Pragmatics in Habermas' critical social theory

305

As part of his engagement with the post-Enlightenment critique of reason, from the Romantics, through Nietzsche and historicism, up to contemporary post-structuralism, Habermas emphasizes the postmetaphysical character of the concept of communicative rationality. The role played by formal pragmatics here is very similar to its contribution to sustaining a dialectics of immanence and transcendence. By “postmetaphysical” he means, principally, that the concept of communicative rationality breaks with traditional views of the substantive unity of reason; refrains from specifying the content of truth, justice and the good life; moves from a subject-object model of cognition to an intersubjective one; gives up the traditional fixation on truth and theoretical reason; and renounces all dependency on “otherworldly” projections such as God or the Good. Our discussion in the foregoing has cast light on how his formal-pragmatic investigations help him to achieve these aims. The reference to reason built into everyday language use is a formal as opposed to substantive conception of reason, for it amounts only to a procedure for determining what is true or right and leaves open the question of the respective content of these concepts; it is non-subjectivist in the sense that it starts from interaction in communicative practices rather than an individual subject’s relation to an object independent of it; it is non-logocentric in the sense that it does not assert the primacy of theoretical or propositional truth but comprises a plurality of dimensions of validity that interact with each other and relate to each other on an equal footing; it is historically rooted in the sense that it is contained within actual practices of using language; and it is “innerworldly” in the sense that the ideal reference point to which it appeals is not some “otherworldly” idea of God or the Good but an idealized projection of a speech situation in which the idealizing presuppositions relating to the conduct of intersubjective deliberation would be realized in practice and in which the outcomes of such deliberation would contribute constructively to validity in a context-transcending sense. In this section we have considered the ways in which formal pragmatics enables Habermas to develop a concept of communicative rationality that is the indispensable normative reference point for his critical social theory, both with regard to the critique of social pathologies and with regard to criticism of the shortcomings and distortions of actual practices of intersubjective deliberation. In my final remarks I will outline the principal lines of objection that have been raised against Habermas’ use of formal pragmatics as the normative foundation for his critical social theory.

5.

Objections

There are three principal lines of objection to Habermas’ use of formal pragmatics as the underlining for his critical project. The first casts doubt on the ability of formal pragmatics to provide a foundation for social critique that is sufficiently robust

306

Maeve Cooke

for his purposes. The second criticizes the specific shape of the conception of reason that emerges from his formal-pragmatic analyses. The third claims that his appeal to formal pragmatics, and the vision of a communicatively rationalized lifeworld that he derives from these analyses, amounts to a foreshortened idea of human flourishing. 1) The first line of objection challenges the ability of formal pragmatics to provide an empirically based, normative foundation for social critique; this objection can be further divided into three sub-objections: a) an objection to the universality of the idealizations built into everyday communicative practices; b) an objection to the idea of the harmonious, free interplay of the three cultural value spheres/moments of reason as a standard for criticizing the colonization of the lifeworld; c) an objection that Habermas does not show the primacy of the communicative use of language over the strategic one. a) The criticism here is that the concept of communicative action that Habermas extracts from his formal pragmatic analyses is too weak for the critical purposes he attributes to it (Benhabib 1992; Cooke 1994; Cooke 2006). Some of the idealizations he identifies as necessary presuppositions of the communicative use of language are not a feature of language use in general; they orient communicative practices only in certain socio-cultural contexts as a result of specific historical developments; in fact, they seem to be features of language use only under conditions of modernity, in societies in which there has been a desacralization of knowledge, a secularization of authority and an internalization of the principle of universal moral respect. Since the critical force of the concept of communicative rationality depends to a considerable degree on idealizations relating to social inclusiveness and equality, it is particularly worrying that they appear to be features only of certain kinds of modern society. This presents Habermas with the following dilemma: on the one hand, if he wants to maintain a formal-pragmatic basis for his critical perspective, he will have to acknowledge the socio-cultural specificity of the resulting concept of communicative rationality, significantly restricting the scope of its critical power; on the other hand, if he wants to offer a critical perspective that is not restricted to particular socio-cultural contexts but rather universal in scope, he will have to supplement his formal-pragmatic analyses with arguments of other kinds (Cooke 2006: 57–61). b) This objection queries the usefulness of the metaphor of the harmonious, free interplay of the three moments of reason for the purposes of critique of the colonization of the lifeworld by functionalist rationality. It points out that Habermas’ critique of colonization appeals to a normative idea of evenly balanced, non-selective processes of societal rationalization that calls for a balanced relationship between communicative rationality and functional rationality. The difficulty is that the balance evoked by the metaphor of three harmoniously interrelating moments of reason is not a balance between communicative

Pragmatics in Habermas' critical social theory

307

rationality and functional rationality, but rather a balance between cognitiveinstrumental, moral-practical and aesthetic-expressive rationality. For the metaphor to work, we would have to be able to equate functional rationality with cognitive-instrumental rationality. Habermas, however, insists that they are not equivalent: whereas instrumental rationality is the reasoning employed by social agents to achieve particular, intended ends, functional rationality bypasses the intentions of the agent (see Cooke 1994: 144–146). c) The nub of this objection is that in order to assert the primacy of the communicative use of language Habermas needs to show that manifestly strategic action, too, is parasitic on the communicative use. The objection is not just that Habermas shows only the parasitic nature of latently strategic action; it is the stronger point that it is impossible in principle to show which use of language is most fundamental without appealing to controversial philosophical presuppositions concerning the rational use of language; indeed it is impossible to decide whether there is a fundamental use of language at all (Apel 1998). In other words, formal-pragmatic arguments will never suffice to establish the primacy of any mode of language use. A related objection is that even successful demonstration of the primacy of the communicative use of language would leave open a number of important questions; these include questions such as how much communicative action is necessary for the maintenance of social order and social reproduction, the extent to which the erosion of communicative action can count as “colonization” and the degree of reflexivity, inclusiveness and openness communicative action has to display in order for social reproduction to count as successful (Cooke 1994: 24–28). 2) The second line of objection relates to the specific shape of the concept of communicative rationality that emerges from Habermas’ formal-pragmatic analyses. Critics have drawn attention, in particular, to his sharp separation between the use of language oriented to mutual understanding and the use of language to disclose new perspectives (Cooke 1994: 74–84). This produces an impoverished account of communicative reason that excludes modes of language use that bring us to see the world in a new way from the category of validity-oriented and, hence, rational communication; ultimately, it leads to an inability to explain the power of critical social thinking to engage the imagination as well as the reason of its addressees, thereby motivating them to work towards social change for the better (Cooke 2006). 3) The final line of objection is directed at the very project of formal pragmatics. One of the strongest criticisms of Habermas in this regard comes from within the Frankfurt School tradition of critical social theory. Honneth, described above as a third generation Frankfurt School critical theorist, commends Habermas for the paradigm shift he introduces into this tradition of thinking through replacing a subject-object model of cognition and action by an intersubjective one. However, he discerns certain biases in Habermas’ critical social theory, which he traces

308

Maeve Cooke

back to its reliance on formal pragmatics. The key point is that by reducing practical reason to a conception of rationality that is extracted from an analysis of everyday linguistic practices, Habermas adopts a critical perspective on society that is blind to forms of human suffering, and corresponding forms of human flourishing, that cannot be conceptualized as failures of linguistic communication. His critical perspective has no room, for example, for critique of violations of bodily integrity, and the corresponding idea of human flourishing as involving bodily selfconfidence, or for critique of the disaffection and alienation that result from lack of recognition of the value of one’s work, and corresponding idea of human flourishing as involving self-esteem based on one’s achievements. For Honneth, this leads to a regrettable narrowing of the scope of social criticism, permitting only normative deficits of certain kinds to come into view. The objection, in sum, is that Habermas’ theory fails to exploit the full potentials of an intersubjective framework, leading to a form of critical social theory that is too thin and too weak to provide a basis for critique of the principal pathologies of contemporary capitalism. His alternative proposal is for an intersubjective model that does not prioritize the linguistic aspects of human flourishing but takes account of its bodily-affective and creatively-productive ones as well. Habermas has responded to some but not all of these points of criticism. Moreover, in those places where he appears to have acknowledged the force of the criticisms, his responses have often been indirect. They manifest themselves in a gradual moving away from the strong reliance on formal pragmatics that informed his theory of communicative action towards more critical-hermeneutic and historicist lines of argument (Cooke 2006: 71). To be sure, Habermas continues to affirm the importance of formal-pragmatic arguments in his critical project (Habermas 1996: 109). We may take this as evidence of his unwavering commitment to a conception of reason that, though historically situated, transcends times and places. Historicist and hermeneutic approaches, although productive in many ways, are unable to account for context-transcending validity. Were Habermas to dispense with formal pragmatics, or even significantly to reduce his project’s reliance on it, he would have to find some other way of accounting for the context-transcending power of rational critique.

Notes 1. Important exceptions include sociologists who embrace systems theory as inspired by Niklas Luhmann and structuralism as inspired by Ferdinand de Saussure. Both systems theory and structuralism conceive of the deep structures that generate society as subjectless systems of rules. This also holds for poststructuralists such as Michel Foucault. 2. As we shall see, this position by no means implies the need for intentionalist accounts of linguistic behaviour, which privilege the speaker’s intentions and neglect both the importance of what is said and the intersubjective relationship established in every com-

Pragmatics in Habermas' critical social theory

3.

4.

5.

6.

7.

8.

9. 10.

11.

12. 13.

309

municative use of language. Nor does it imply that the speaker’s account of her own intention is incontrovertible. Habermas views the tension as potentially positive insofar as it serves as a reminder that both perspectives have their strengths and weaknesses. Although, as we shall see, he makes a methodological decision in favour of the internal perspective, his theory of communicative action acknowledges the importance of the external perspective, most obviously when it offers a causal explanation of the pathologies of modern societies in terms of the colonization of the communicatively regulated domains of social life by the functional rationality of the economic and administrative systems. These questions are further expounded in his longer study, The Logic of the Social Sciences (Habermas 1988), an earlier version of which appeared as an article in the Philosophische Rundschau in 1967. I have summarized Habermas’ discussion here. In the essay in question, Habermas distinguishes between three kinds of generative theory of society because he includes theories that do not take subjective meaning as a sociological primitive. Thus, in addition to the two I mention, he identifies ‘system theories that lack a subject’ (Habermas 1971: 18). See note 4 above. Verständigung (n.): “reaching understanding”, “mutual understanding” or “communication”. The corresponding verb is sich verständigen. The term is ambiguous. It embraces linguistic comprehension (Verstehen) but goes beyond this to refer to the process of reaching understanding, in the sense of reaching an agreement with another person or persons. When Habermas refers to “formal semantics” he is thinking of the philosophical theories of meaning that take their inspiration from Gottlob Frege and Betrand Russell and in the second half of the twentieth century were associated with thinkers such as Donald Davidson and Michael Dummett; Habermas mentions Peter Strawson, but his place in this tradition is not straightforwardly evident (Habermas 1998: 99, note 62). Habermas sometimes describes such claims as universal in the sense of transcending times and spaces (e.g. Habermas 1994: 42). However, when he refers to universal validity claims on other occasions, he seems to mean universal in the sense of present in every speech act (e.g. Habermas 1998: 89). On Bühler (and Jakobson) cf. Daalder and Musolff this volume. Habermas also asserts the primacy of the communicative use of language over what he calls the indirect use of language such as the figurative and the symbolic (Habermas 1998: 166). Since his argument to this effect is cursory, I will not discuss it in the following. His neglect of the figurative and the symbolic uses of language, and demotion of them to a subordinate status, is symptomatic of his overly sharp distinction between reason and imagination (see Cooke 1994: 74–84 and Cooke 2009). This is true irrespective of whether the intention to deceive or manipulate is conscious or unconscious. Habermas distinguishes between conscious and unconscious latently strategic action. Unconscious latently strategic action is correlated with what he calls “systematically distorted communication” (Habermas 1998: 169). I cannot adequately discuss the latter topic in the present context; I return to it briefly in the next section. See note 13 below. For twenty years subsequent to the publication of The Theory of Communicative Action, Habermas had little to say about the problem of cultural impoverishment; in his most recent writings, however, he directly addresses the related question of semantic renewal,

310

Maeve Cooke

drawing attention to the potential semantic productivity of the beliefs and traditions of the major world religions (Habermas 2008). 14. Habermas sees Immanuel Kant as the first philosopher to take seriously the differentiation of reason into its moments (Habermas 1990: 1–20, esp. 2–3). 15. I leave open the question of whether Hegel thought this process of self-actualization would ever be concluded.

References Adorno, Theodor W. 1981 Cultural Criticism and Society. In: Theodor W. Adorno Prisms. Trans. by Samuel and Shierry Weber. Cambridge, MA: MIT Press. Apel, Karl-Otto 1998 Openly Strategic Uses of Language: A Transcendental-Pragmatic Perspective. In: Peter Dews (ed.), Habermas: A Critical Reader, 272–290. Oxford: Blackwell. Apel, Karl-Otto 1980 Towards a Transformation of Philosophy. Trans. Glin Adey and David Frisby. London: Routledge and Kegan Paul. Austin, John L. 1962 How to Do Things with Words. Oxford: Oxford University Press. Benhabib, Seyla 1992 Situating the Self. New York: Routledge. Bühler, Karl 1934 Sprachtheorie. Jena: Fischer. Chomsky, Noam 1969 Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Comte, August 1893 The Positive Philosophy of Auguste Comte. Trans. H. Martineau, 3rd edition. London: Kegan Paul, Trench, Trübner and co. Benhabib, Seyla 1992 Situating the Self. New York: Routledge. Cooke, Maeve 1994 Language and Reason. A Study of Habermas’s Pragmatics. Cambridge, MA: MIT Press. Cooke, Maeve 2006 Re-Presenting the Good Society. Cambridge, MA: MIT Press. Cooke, Maeve 2009 Zur Rationalität der Gesellschaftskritik. In: Rahel Jaeggi and Tilo Wesche (eds.), Was ist Kritik?, 117–133. Frankfurt/Main: Suhrkamp. Dilthey, Wilhelm 1996 Hermeneutics and the Study of History. Edited by Rudolf Makkreel and Frithjof Rodi. Princeton, NJ: Princeton University Press. Fraser, Nancy and Axel Honneth 2003 Redistribution or Recognition? A Political-Philosophical Exchange. London: Verso.

Pragmatics in Habermas' critical social theory

311

Habermas, Jürgen 1971 Reflections on the Linguistic Foundation of Sociology: The Christian Gauss Lectures (Princeton University, February-March 1971). In: Habermas 2001, 1–104. Habermas, Jürgen 1974 Reflections on Communicative Pathology. In: Habermas 2001, 129–170. Habermas, Jürgen 1984 The Theory of Communicative Action, vol. 1. Trans. Thomas McCarthy. Boston: Beacon Press. Habermas, Jürgen 1986 A Reply. In: Axel Honneth and Hans Joas (eds.), 214–264. Habermas, Jürgen 1987 The Theory of Communicative Action, vol. 2. Trans. Thomas McCarthy. Boston: Beacon Press. Habermas, Jürgen 1988 On the Logic of the Social Sciences. Trans. Shierry Weber Nicholsen and Jerry A. Stark. Cambridge, MA: MIT Press. Habermas, Jürgen 1989 The Structural Transformation of the Public Sphere. Trans. Thomas Burger with the assistance of Frederick Lawrence. Cambridge, MA: MIT Press. Habermas, Jürgen 1990 Moral Consciousness and Communicative Action. Trans. Christian Lenhardt and Shierry Weber Nicholsen. Cambridge, MA: MIT Press. Habermas, Jürgen 1991 The New Conservatism: Cultural Criticism and the Historians’ Debate. Trans. Shierry Weber Nicholsen, Cambridge, MA: MIT Press. Habermas, Jürgen 1992 Postmetaphysical Thinking. Philosophical Essays. Trans. William Mark Hohengarten, Cambridge, MA: MIT Press. Habermas, Jürgen 1996 Between Facts and Norms. Trans. William Rehg, MIT Press, Cambridge, MA: MIT Press. Habermas, Jürgen 1998 On the Pragmatics of Communication. Trans. and edited by Maeve Cooke. Cambridge, MA: MIT Press. Habermas, Jürgen 2001 On the Pragmatics of Social Interaction. Preliminary Studies in the Theory of Communicative Action. Edited by Barbara Fultner. Cambridge, MA: MIT Press. Habermas, Jürgen 2008 Between Naturalism and Religion. Trans. Ciaran Cronin. Cambridge, MA: MIT Press. Honneth, Axel 1994 The Social Dynamics of Disrespect: On the Location of Critical Theory Today. Constellations 1(2): 255–269. Honneth, Axel 2003 The Point of Recognition. A Rejoinder to the Rejoinder. In: Nancy Fraser and Axel Honneth, 237–247.

312

Maeve Cooke

Honneth, Axel and Joas, Hans (eds.) 1986 Communicative Action. Cambridge: Polity. McCarthy, Thomas 1978 The Critical Theory of Jürgen Habermas. Cambridge: MA: MIT Press. Weber, Max 1969 Basic Concepts in Sociology. Trans. H. P. Secher. Westport, Conn: Greenwood Press. Wittgenstein, Ludwig 2001 Philosophical Investigations. Trans. Gertrude E. M. Anscombe. Oxford: Blackwell.

Part III Key topics in pragmatic description

11.

Deixis and indexicality William F. Hanks

1.

Introduction

The study of deixis is central to pragmatics, because deictic systems define points of intersection between linguistic structure and the social settings in which speech takes place. It is standard in contemporary English language sources to distinguish Spatial (here, there), Temporal (now, then, tense), Person (pronominals), Discourse (coreference or reference to prior talk) and Social Deixis (honorification and any indicators of social identity or status relations among participants and contexts).1 Any one of these functional foci may be formally marked in utterances morphologically, lexically, or by construction type. Deictic phenomena inevitably interact with gesture (conventionalized or not) and with many other parts of the grammar, including evidentiality, status, modality, aspect, noun classification, possession, spatial and temporal description. This chapter focuses on those forms whose primary function is to individuate objects of reference (including events, material things, talk itself or any individuated concept). Referential deixis is found in all human languages and includes at least demonstratives, person markers, locative, directional and temporal markers, but excludes much “social deixis” such as honorification (where social status is indexed but usually not singled out for comment) and standard sociolinguistic markers (where social factors are indexed but not singled out for comment). By starting from the distinctive structure of referential deixis, one can better characterize its relation to grammar and to other varieties of indexicality. The main grammatical question is how deictic functions are encoded in languages. The challenge for pragmatics is to determine the constant capacity of deictic types to contribute to token acts of situated referring, and the relation between such referring and other aspects of utterance context. Although they appear in various grammatical categories, deictics are all part of a single field, which I will call the deictic field (Bühler 1934). The deictic field is a psycho-social construct composed of: (1) the roles of communicative agents relative to the participant frameworks they occupy, that is, the relations between Speaker (Spr), Addressee (Adr), and Others, as defined by the language and communicative practices of its speakers; (2) the positions occupied by objects of reference, that is, the kind of object and its role in the activity in which the deictic utterance is produced; and (3) the multiple dimensions whereby participants have cognitive access to objects, such as perceptibility, relative proximity, memory, anticipation and prior talk. When a Spr performs an act of deictic reference, as in ‘that’s John right over there (punctual point)’, ‘we came from over there (vague

316

William F. Hanks

point)’, or ‘here’s your coffee (extending in hand)’, she produces a social relation between herself, her interlocutor and the object. The deictic field is a template for the elementary social relations on which conversation rests. The remainder of the paper is organized into two parts, theoretical approaches and typological approaches.

2.

Theoretical approaches to deixis

Deictics are what Sacks (1992) called “indicator terms” and what linguists and philosophers have called variously “indexicals”,2 “shifters”3 or, as in the present paper, “deictics” (Fillmore 1997, Hanks 1983, 1990). In his theory of semiotics (cf. Nöth this volume), CS Peirce (1955) distinguished indexes from icons and symbols, according to the relation between the perceptible sign vehicle (or representamen) and its object (that for which it stands). Icons relate to their objects by way of similarity, symbols by way of sheer convention, and indexicals by way of “dynamical contiguity.” As Peirce (1955:107) put it, the index stands for “dynamical (including spatial) connection [of the deictic form] both with the individual object, on the one hand, and with the senses or memory of the person for whom it serves as sign, on the other.” Certain indexes are in a causal relation to their object (smoke indexes fire, a weather vane indexes wind direction) and others are conventionalized (‘this’ indexes objects close at hand, ‘there’ places at a remove from the location of utterance, road signs index road conditions). In the prototypical cases, the index does more than stand for its object, it directs an addressee’s attention to it, and this directive force is especially salient in certain types of deixis (e.g. presentatives, such as ‘Here it is (take it!)’, ‘There he goes! (look!)’). Such expressions do more than individuate an object; they direct the addressee to it, and the proper uptake is usually a gesture or a shift in attention focus. Morris (1946) simplified the Peircean framework, distinguishing characterizing signs (icons and symbols, both of which characterize their object) from indices (which are devoid of descriptive information about their object) and universal signs (anyone, anywhere, etc., which are applicable to any item in a class). From this perspective, what is noteworthy about deictics is that they contribute to individuated referring without in any way describing their objects. As Sacks (1992: 520) put it, such terms, which he called “indicators,” are “altogether ‘abstract.’ […] capable of invoking the sheer fact of the setting,” without in any way formulating it. The definition of deixis in terms of individuation without meaning is common in philosophical literature (Searle 1969:80, Nunberg 1993, Bar Hillel 1954), but has less prominence in the linguistic literature, where cross-language comparison and detailed description reveals many conventionalized features in deictic systems (see below). While Morris’s framework played a pivotal role in the development of

Deixis and indexicality

317

linguistics (cf. Nöth this volume), his definition of indexicality is too oversimplified to describe natural language deixis. By contrast, Peirce insisted that his three sign types designate relations which typically combine in actual signs. The conventionality of any language guarantees that the pairing of form with meaning in the language is conventional, no less so for deictics than for other kinds of expression. Moreover, this conventionality encompasses deictic features such as punctate vs. regional space, centripetal vs. centrifugal movement relative to utterance context, kinds and degrees of perceptual access to the objects denoted by demonstratives, the presence or absence of gender, animacy and noun class in demonstratives, and any special presentative forms in a language (see French voici, voilà, Yucatec Maya hé’ebe’ ‘there it is (listen!)’.The result is that actual deictic forms and arrays always combine sheer indexical functions with other kinds of information (indexical, iconic or symbolic). If one maintains the more austere philosophical definition of deixis as pure indexicality, we are led to conclude that natural language systems are always impure blends between indexical and non-indexical functions. A more useful and realistic alternative is to conclude that deixis in natural language requires a more robust Peircean definition of the phenomenon, and challenges researchers to explain why certain bundles of functions cluster around indexicality in language after language. This point is all the more clear when we consider that deictics are typically organized in closed class paradigms, where obviously indexical functions are bundled with a wide array of other information. Natural language deictics are what Bloomfield (1933) called category representatives, that is, they bear features for morphosyntactic category and function as syntactic heads. Some but not all deictics may be used either exophorically, for reference to extra-linguistic context, or endophorically, for anaphoric co-reference with prior talk (A: ‘I prefer this one’, B: ‘Oh, I like that too’) or cataphoric co-reference with subsequent talk (‘the two statements have this in common, they are terse’) (see Halliday and Hasan 1976). While there are borderline cases where this distinction is difficult to draw, there are also regular patterns in which deictic types are specialized for one or the other. Anaphora, for instance, is often associated with forms which when used exophorically indicate remoteness of the object, while spatial adverbial paradigms often include one or more forms that can be used only exophorically. They also differ as to whether they may be lexically expanded, as in ‘here’ vs. ‘here in my room,’ ‘this’ vs. ‘this blue pencil,’ and so forth. It is common to describe the simple forms as pro-locative, pronominal, and so forth, and the expanded forms as adlocative, adnominal and so forth. Moreover, deictics in many languages may be combined in distinctive n-tuples of coreferential or mutually reinforcing forms (‘this here book’, Yucatec leti’ e hé’ela’ < ‘the one the here.it.is’ = ‘this here one’, té’e waya’ < ‘right.there here’ = ‘right here’ [point within a region]). It is also common to find distinctions in spatial deixis according to the regional scope of the place referred to, as in Yucatec té’ela’ ‘right here’ (punctual) vs. waye’

318

William F. Hanks

‘here’ (regional). By contrast, English ‘here’ may denote a point on the speaker’s nose, the room, house, neighborhood or country in which it is uttered, depending on context, and similar extensions apply to ‘there, now, then,’ and others. In short, when one examines the pragmatics of actual deictic systems, many features arise that are elided in the austere definition of indexicality as semantic emptiness. In his classic work Sprachtheorie, Bühler (1934) initiated what would become another line of thought concerning deixis in language. Bühler distinguished three dimensions in the elementary speech event, the Speaker, the Addressee and the Object talked about. Intersecting all three poles is the division between what he called the Zeigfeld, the demonstrative field in which objects can be shown, and the Symbolfeld, the discourse field to which description of the object, and other discourse belongs. This relatively simple schema already generates six values (three in each field) and provides the elementary framework for a functional analysis of deixis: deixis and demonstration belong to the field of exophoric reference to objects available in the immediate perceptual field. Anaphora, reference to prior talk and to remembered objects are treated as blends of the Zeigfeld and Symbolfeld. Jakobson (1960, 1971) expanded and refashioned Bühler’s distinctions into the division between the speech event (Es) in which talk occurs and the narrated event (En) talked about. Jakobson eschews the term “deixis,” but treats all referential indexicals under the rubric of “shifters” (a term coined by Jespersen 1965: 290). Shifters are linguistic categories whose distinctive feature is that they individuate a referent (in the En) in relation to the situation of speech (Es). Thus if Ben tells Katherine ‘I saw him right here yesterday’, every item but the verb ‘see’ is a shifter because each one establishes a relation between the utterance and the event narrated: ‘I’ equates the current speaker with the seer, the past tense of the verb places the En in a time prior to the moment of utterance (reinforced by ‘yesterday,’ the day prior to the day of utterance) and ‘him’ indicates that the person seen is neither the current Spr nor Adr. Jespersen adopted the label “shifter” in light of the distinctive behavior of these grammatical categories in reported speech. If Katherine subsequently reports Ben’s utterance to Madeleine, she says ‘Ben told me he saw him there yesterday’. In reported speech the shifters “shift” to accommodate the new speech event frame of the quoted speech, while maintaining sameness of reference to the event of seeing and its participants. Inversely, in direct quotation, the linguistic forms remain constant, but the reference is dislocated. If Katherine directly quotes Ben, she says ‘Ben said “I saw him here yesterday”’. Notice that despite the ‘I, here’ in this utterance, Katherine does not therein refer to herself or to the location in which she reports the speech. In effect, the introductory phrase ‘Ben said’ projects a prior Es in which the utterance was produced, and it is that prior event that serves as the ground of reference for the deictics. Hence there is no shift in the forms, but the speech event relative to which they are interpeted has been transposed. Silverstein (1976) follows Jakobson’s shifters model and expands it to a wide array of indexical effects, only some of which contribute to reference. Essen-

Deixis and indexicality

319

tially the same set of distinctions underlies the still more elaborate analysis of the speech event in the ethnography of speaking.4 Phenomenologists were centrally concerned with indexicality (Husserl 1978, Schutz 1970, Merleau-Ponty 1967) as evidence of the unavoidable situatedness of speech in the immediate experiential settings of its production. While phenomenology has little impact in English language linguistics, it is fundamental to ethnomethodology (Garfinkel 1967, 1972) in its claim that all meaning is (indexically) produced in the “ad hoc” processes of interaction, and similarly to the approaches to interaction pioneered by Cicourel (1964, 1974, 1985, 1992), Schegloff (1972, 1982), Gumperz (1982) and Goodwin (1981, 2003). From a psycholinguistic perspective, Rommetveit (1974) combines phenomenology with linguistics to argue that linguistic production relies not on abstract semantics, but on the experiential worlds of speakers. Conversation analysis as developed by Sacks, Schegloff and others carries forth the basic premise that speech is pervasively indexical and grounded in the interaction between co-participants. Developing the structural linguistics of Meillet (1926), Martinet (1960) and the Prague School, Benveniste (1965, 1974) incorporated phenomenology in his landmark papers on subjectivity in language. This range of approaches to indexicality raises a number of issues. There is widespread agreement that indexicality designates the context dependency between utterances and speech contexts, and that deictic types are specialized in individuating single objects or groups of objects in relation to the participants in the event of speaking. We can schematize this as in Figure 1, where the ‘Indexical ground’ is the zero-point, or pivot relative to which the deictic points to the object denoted. Thus for example, in an utterance like ‘it’s sunny here’, the locative ‘here’ projects a relation of relative immediacy to or inclusion of the Speaker’s location at the moment of utterance. Relation Indexical ground

Object

Figure 1. Relational structure of referential deixis

This three-way distinction raises questions on which there is considerable debate in the literature. First, what kinds of relational features are encoded in deictic types? The spatialist view holds that relative contiguity (this = proximal, that = non-proximal and so forth) is fundamental, and other pragmatic effects are incidental to deixis or derived by metaphorical extension (Anderson and Keenan1985) or contextual inference (Levinson 1983). For careful assessment of these alternatives, argued with data from Lao, see Enfield (2003, 2009). By combining a spatialist definition with Gricean implicature (Grice 1967, 1978) and conversational organization, the enriched values of deictic tokens can be “calculated” as implica-

320

William F. Hanks

tures. By contrast, recent non-spatialist views reject the a priori primacy of space, and substitute for it something closer to accessibility, that is, how the participants have or gain access to the object.5 Spatial contiguity sometimes provides a support for accessibility, but is only one possible parameter alongside others, including perception (via any of the senses), prior talk and memory, salience (cognitive or perceptual), evaluative stance, relative familiarity with the object, and in some cases ownership of the object by a participant. Deictics may be relatively lean in terms of descriptive information, but the pragmatics of deixis nonetheless responds systematically to the features of actual objects. Where the spatialist approach presumes that space is basic across all contexts, the accessibility approach looks to situational relevance, the interactive move made by the deictic utterance, the social definition of the object and other features of context (Hanks 2005a). The second major question bears on how one defines the indexical ground. Traditional treatments assume that the zero-point of deixis is the speaker, as in ‘“here” means a place close to me’. One of the earliest statements of what can be called the egocentric approach is Russell’s description of “egocentric particulars” (1940), which cover the range of deictics (compare Evans 1982: Chapter 6, Gale 1968). Most traditional grammars and typologies (eg. Anderson and Keenan 1985) assume egocentricity, and Benveniste’s (1966, 1974) anchoring of indexical reference in subjectivity makes the same move (cf. Lyons 1982: 121). With the exception of Schutz, whose emphasis was intersubjectivity, most phenomenology also proceeds from the ‘ego-subject’ (albeit with different inflections). Egocentricity is consistent with the commonsense idea that the solitary speaker is the one who produces her utterance, and the minimal requirement for utterance production is a speaker. The challenge for this position is to explain the many cases in which a deictic responds not to speaker accessibility, but to the addressee’s access to the object, or to the relative symmetry of access between the co-participants. Moreover, while the idealization of the solitary speaker has been productive for certain purposes, Conversation Analysis played a leading role in showing it to be patently inadequate as a basis for pragmatics.6 There are various alternatives to egocentricity. Reichenbach (1947) described indexicals under the heading of token-reflexivity, by-passing the problem of the Speaker in favor of the utterance itself as zero-point. The utterance “here” denotes a place immediate to the utterance token ‘here,’ just as “I” denotes the individual who produces the token ‘I’. This removes the problem of the zero-point from the realm of interaction, stating in effect that these forms denote in relation to instances of themselves. Close description of actual deictic usage indicates that elements such as common ground,7 collaboration between participants (Goodwin 1981, 2003), and the relative symmetry of participant access (whether speaker and addressee have roughly the same or significantly different access to the object in situ; Hanks 1990, 1992) all impact on the use and understanding of deictic tokens. Any of these factors may be conventionalized in the semantics of deictic types.

Deixis and indexicality

321

This suggests that the zero-point of indexicality is best conceived as the participation framework in which talk is produced. Just as space may be a key kind of access for some uses, the speaking ego may be the ground for some uses. But neither space nor egocentricity is a priori privileged. These different approaches to deixis imply different background pictures of utterance context. The egocentric spatial picture holds that deictic acts occur when the Spr, the Adr and the Object are physically copresent and mutually perceptible. The zero-point of the deictic utterance is the Spr’s body and the utterance orients the Adr’s attention to a specific sector of the broader sphere of spatial proximity, not unlike a pointing gesture (which may indeed be part of the utterance). To understand a deictic is not to “interpret” it but simply to recognize by observation what it singles out in the physical situation of utterance. The space of dialogue may be interperspectival, but it is the Spr who produces the utterance and his or her body that serves as the indexical anchor-point: a self-contained individual, oriented in physical space and endowed with a sphere of contiguity and sensory access. This picture is dominant in much of the English language literature on deixis. It is found in most modern linguistic descriptions, especially typological and psycholinguistic work, for which it seems to offer a controlled basis of cross-linguistic comparison.8 Despite its appeal, however, this picture omits critical features of deictic practice, including the mutual orientation of interactants, all non-perceptual modes of access such as background knowledge, memory or anticipation, and all extra-physical aspects of social settings. Non-spatial aspects of deictic speech are either ignored, treated as metaphorical extensions from space, or derived from other, non-deictic principles. There is a second, widely attested background picture for deixis, according to which the core of context is social interaction.9 In this view, deictic utterances involve at least the following factors. (i) The reciprocal perceptual fields of the parties, especially reciprocal vision in the face-to-face (Schutz 1970, Goodwin 1981, 1994, 2000, Kendon 1992). (ii) The broader “reciprocity of perspectives” whereby each party assumes that the other has a perspective and that if (s)he adopted that perspective, the world would look to her as it does to the other (Schutz 1970). This is precisely reflected in the alternation of “I, you, here, there” among Spr and Adr: each party to talk is an ‘I,’ but only to himself, and each one is a ‘you,’ but only to the other, my ‘here’ is your ‘there’ and so forth. (iii) Participants’ bodies are expressive through gestures, as well as receptive through the senses, and gesture is a key aspect of deixis.10 (iv) Any interaction has sequential organization, perhaps the best studied feature of conversation. Sequential organization is embodied in turn taking, adjacency, contiguity and ‘proximateness’.11 (v) The interaction has a motivational structure: the Spr speaks in response to something and in order to achieve something else; the Adr responds in the light of the aim of the first utterance, in order to achieve his or her own aim, which then serves as the ‘because motive’ of the next turn, and so on back and forth (Schutz 1970).

322

William F. Hanks

(vi) Any utterance has a “relevancy structure” according to which certain things matter and others don’t.12 (vii) Utterance meaning must be “negotiated” or worked out by the co-engaged parties. It is not given in advance, nor is it fixed by the intentions of the Spr (Clark, Schreuder and Buttrick 1983). For deixis, this means that, for instance, the reading of a deictic token as spatial proximity, perceptibility, cognitive salience, egocentric or altercentric can vary by context and must be worked out by the participants. (viii) Participants display to one another their sense of the current situation, the relevancy structure they are assuming, and their current relation to their own speech. Deixis is a primary resource for such display precisely because it points into the situation and thereby “positions” the Spr.

In the Interactionist picture, there is an emergent space of interaction, but it is not the kind of egocentric, physical space presumed by the Spatialist view (Kendon 1992), and to study it, we must attend to sequential organization of talk, to situational variation and to the micro-ethnography of everyday usage. In the course of deictic practice, interactants must jointly establish and display the relevance of spatial perceptual, discursive or other contextual frames. Moreover, in much ordinary referring, sequential placement of a deictic utterance in talk combines with spatial location of the object in the course of identifying the denotatum. Ordinary practice therefore blends endophoric with exophoric functions. For the purpose of deictic analysis, situational relevance has two main sources: (1) local relevance turns on the immediate circumstances of utterance, including sequential placement of the utterance, the location of the object and the positions of the interactants in the interactive field, and (2) global relevance turns on the embedding of the deictic field in a larger social field. The broader social setting can define the boundaries of “here” for the purposes of talk, and factors such as social identity or hierarchy and activity spaces may license extensions, transpositions, or they may constrain usage. For instance, if a foreman on a work site says ‘I am installing the electric outlet here’, the utterance can be fully normative even if the foreman does none of the actual work. The hierarchical role of the foreman licenses first person singular usage for tasks of which he is the leader but not literally the one who does the work. Similarly, spatial references in the extended households of Yucatec speakers respond systematically to the social organization of space, so that “here” may refer to the entire household or exclusively to the portion occupied by the Spr. Shamans and doctors who use imaging technologies (divining crystals, x-ray images and the like) routinely make deferred ostensions in which pointing to the image is pragmatically equivalent to pointing to its bodily counterpart. In cases like these, the activity space in which the deictic practice takes place provides specific affordances and relevancies that help motivate practice. The Spatialist and Interactionist picures of utterance context can be combined so long as we treat them as complementary: spatialism is a claim about the re-

Deixis and indexicality

323

lational features of deixis, and interactionism is a claim about the indexical ground. We can grant interaction as the ordinary context of utterances, while still claiming that the semantics of deixis is egocentric and spatial. In this case, something akin to Gricean implicature is invoked to derive contextual effects from what is literally spatial reference. At the same time, conventionalized non-spatial features, such as perceptual access, are taken to be “non-deictic.” Recent work by a variety of scholars has combined grammar and interaction, thereby opening the possibility of a blended account of deixis.13 While some combination of linguistic and interactional analysis is undoubtedly necessary to describe the pragmatics of deixis, it is not without cost. If the interactionist grants that verbal deixis is fundamentally spatial in its semantics, then interaction becomes an epiphenomenon of grammar; it is merely how people ordinarily use linguistic forms whose meanings are fixed apart from context. This is surely contradictory to the work of ethnomethodologists such as Garfinkel (1967), who took the strong position that all meaning is produced in the course of ordinary talk, and it also contradicts the more recent position that talk is inherently situated. Similarly, the spatialist position is reduced to an idealization if one grants that neither egocentricity nor spatial relations are necessarily salient in ordinary interaction. Without recourse to the counter-evidence of actual usage, and protected by the vagueness of Gricean implicature, it is hard to see what evidence could be adduced either for or against spatialism. What the two background pictures have in common is that they take the conditioning circumstances of pragmatics to be local to the utterance: in the spatial view, it is the field of proximity to the speaker, and for the interactionist, it is the immediate sequential context, augmented by those features of context to which participants are oriented (Schegloff 1987). Linguistic anthropologists have questioned the localism of both approaches, in an effort to understand the how broader social values and definition of the occasions of talk systematically affect deictic practice (Hanks 1990, 2005a, Enfield 2003, Goodwin 2006, Silverstein 1976). Cross-cultural descriptions of talk suggest that there is no value-neutral way to define the “dynamical contiguity” that is at the core of deixis. Starting from the linguistic structure of deictic systems, to the local circumstances of deictic acts, this third approach proceeds to the social fields in which the deictic field is embedded (Hanks 2005a, 2006a). This includes the identities of participants, the genres of practice of which deixis is a part, the social definition of place and time and the values attaching to objects of reference. The resulting framework is less elegant than formal accounts, but more empirically precise to the varieties of deictic practice that we find in actual talk.

324 3.

William F. Hanks

Typological approaches

Deictic systems can be compared along a number of axes according to their grammatical and functional dimensions. The first point of comparison is an inventory of the grammatical categories for which the forms are marked: Noun (pronouns, articles, demonstratives, e.g. I, the, this, that), Adverbs of space (here, there), time (now, then), manner (like this, like that), Presentative forms (French voici, voilà, Maya hé’ela’ ‘here it is (giving)’, hé’elo’ ‘there it is (pointing)’, hé’ebe’ ‘there it is (sniffing, cocking ear)’, Verbs (especially motion verbs like ‘come, go’, see Fillmore 1997), and various elements that attach to major categories (directionals, as in Mayan verb stems; clitics and utterance or phrase initial or final particles). Prototypical deictics are category representatives that can substitute for, or be expanded by lexical descriptions (‘this’ vs. ‘this red pen’). In the former case, the deictics is pro-nominal, and in the latter it is ad-nominal. In some languages the pronominal forms and the adnominal qualifying usages correspond to different series, as in Mulau, a Daic language, and Japanese (Diessel 1999: 59). Diessel (1999: 58) reports that in twenty-four of the eighty-five languages he sampled, pronominal and adnominal usages are marked by different categories, and in languages such as Turkish and Lezgian, in which the same deictic roots are used in both functions, the pronominal forms inflect for case, whereas the adnominal ones do not.14 The grammatical classification of deixis raises various questions, of which two deserve mention here. The first is whether all or part of certain category-specific subsystems are derived from forms in other categories. For instance, adverbial deictics in some languages are derived from nominal ones, as in Quechua (kay ‘this’ > kay-pi ‘here’, chay ‘that [medial]’ > chay-pi ‘there’, haqay ‘that [dist]’ > haqay-pi ‘there [dist]’ (Calvo Perez 1999, Lasater 2009). Temporal deictics are often derived from spatial ones, as in Yucatec tol-akhéak-o’ ‘back then [remote past]’ < tol-o’ ‘out there [remote]’, or from non-deictic lexemes (cf. ‘today, yesterday, aujourd’hui, Maya behé’ela’e’ ‘today (lit. ‘road.here.it.is’). The second question is the degree to which the different subsystems in any language are similar or different in structure. Do we find the same series of oppositions in different subsystems, as in French ceci, cela (nominal), ici, là (spatial) voici, voilà (presentative) or English ‘this, that, here, there, now, then’? The apparent proportionality of deictics across catgories has encouraged many scholars to propose paraphrase relations among deictics, such as ‘this’ is the one ‘here’, ‘that’ is the one ‘there’, ‘here is the place occupied by I’, and so forth.15 Both the derivational relations between subsystems and the relative proportionality of oppositions go to the larger question of the degree to which deixis in any language is a single system unified by shared forms, regular formations or recurrent arrays of distinctions. Comparative research has demonstrated that languages differ on this axis.

Deixis and indexicality

325

A closely related axis is the degree of compositionality of deictic forms. Languages like English and Hausa have monomorphemic deictics. nân nan cân can

Prox to Spr (immediate field) Distal to SPr/Prox to Adr Prox to Spr (non-immediate field) Distal to Spr (Hill 1982, Jaggar and Buba 1994)

By contrast languages such as Yucatec show bimorphemic stems in all categories, with a number of more complex stems among the temporal adverbs (Hanks 1990): lel-a’ té’el-a’ hé’el-a’ bey-a’ lel-o’ té’el-o’ hé’el-o’ bey-o’ wal-ak-héak-a’ ‘at this time of day in the past’ tol-ak-héak-o’ ‘at that time in the remote past’

In Quechua, nominal demonstratives mark a three way contrast between Prox, Medial, Dist with monomorphemic forms, and the spatial adverbs are derived from the demonstratives by suffixation of [-pi], as in: kay chay haqay

‘this’ ‘that’ ‘that’

kay-pi ‘here’ chay-pi ‘there’ haqay-pi ‘over there’

Similarly, Mandarin demonstratives are monomorphemic, whereas spatial adverbs are derived by suffixation (Tao 1999, Yang 2006). zhè zà

‘this, hi accessibility’ ‘that, low accessibility’

zhèr, zhèli nàr, nàli

‘here’ ‘there’

At the high end of compositionality are languages like Nungubuyu, Santali and Malagasy, where deictic words may be composed of five or more morphemes, each of which contributes to the meaning of the whole. For Nungubuyu, Heath (1980) reports that four basic roots distinguish Proximal, Accessible (to Adr), Distal and Anaphoric in both demonstrative and adverbial categories, with the adverbs derived by suffixation. These are cross-cut by + ≠ ø Concrete (precise vs. vague), + ≠ ø Kinetic (trajectory vs. location), + ≠ – Centric (where +Centric is subdivided into Centripetal ≠ Centrifugal, and –Centric is used for transverse relative to the Spr’s field of vision). The nominal forms are further subdivided by + ≠ ø Absolute (which Heath glosses ‘definiteness’), gender (masc, fem), number (sg, dual, pl) for human referents, and an additional five noun classes for non-humans. Each of these distinctions is marked by morphological exponents, resulting in relatively compositional forms. In Santali all deictics are composed of minimally three morphemes, but potentially more (Zide 1972). There are three bases (proximal, medial, distal) that combine with three regions (immediate, medial, remote) to yield nine different spatial values. The resulting forms may be further marked as Lateral (for objects at the edge of the perceptual field, like Nungubuyu non-Centric), Intensive (for referen-

326

William F. Hanks

tial precision, like Nungubuyu Concrete), Visible ≠ Invisible (two distinct morphemes), Animate ≠ Inanimate, and a three way number distinction. Denny (1982: 371–372) reports that in Inuktitut, there are twelve deictic roots, distinguishing ‘Out of field’ ≠ ‘In-field’ and within In-field, Away (non-prox) ≠ At (prox), and within Away, Bounded (Exterior ≠ Interior) ≠ Vertical (Inferior≠Superior) ≠ Horizontal. The In-field categories are further divided by a binary division between Restricted (equidimensional, precise) ≠ Extended (elongate, mobile, regional). Further modifications (by inflection or derivation) include grammatical category (nominal, adverbial, predicative), transposition of the indexical ground from Spr to Adr or Other, case, number, ‘existence in a place’ and motion yield over six hundred and eighty-five deictic words. In general, the greater the degree of compositionality, the greater the functional transparency of the composed forms and the more likely it is that some deictic morphemes (root or affix) will recur across subsystems, resulting in a greater degree of proportionality (although non-compositional languages, like English, can be highly proportional). Grammatical categories, derivational relations, proportionality of oppositions and compositionality jointly make up the paradigmatic “skeleton” of deixis in any language and provide a basis for more detailed comparison of functional structures in different subsystems and across languages. Are there recurrent gaps in the paradigms, markedness asymmetries, or domains in which functions are hyperdifferentiated? Individual subsystems in any language and across languages differ widely in the relative delicacy of the distinctions they mark. Many linguists have sought to organize this variability in typologies based on the number of opposed terms given deictic systems display. Frei (1944) first formulated this approach, noting that the most commonly attested systems are two- or three-term systems, often augmented by an additional term that is neutral as to distance. Anderson and Keenan (1985) build on Frei’s schema, as does Fillmore (1997) and more recent typological work done by scholars at the Max-Planck Institute for Psycholinguistics. Among languages with four or more terms, Anderson and Keenan (1985) cite Malagasy, Quileute and Sre, to which we can add Yucatec Maya (Hanks 1990), Tlingit (Boas 1917, Swanton 1911), Crow (Graczyk 1986) and Santali (Zide 1972). There are several challenges facing this kind of typology. First, how to distinguish between multi-term systems in which each term indicates a degree of spatial remove vs. ones in which an erstwhile medial is actually neutral or marks a shift in the indexical ground. The former are treated as distance-centric, the latter as “person-centric” (Anderson and Keenan 1985). A second challenge is to decide on which forms belong in the inventory. In highly compositional systems or ones in which multiple deictics are combined in relatively fixed phrases, the line between “basic” deictics, on which the typology is focused, and non-basic ones can be difficult to draw (a difficulty pointed out by Frei 1944: 114, 120). A language that ap-

Deixis and indexicality

327

pears two-term from a restrictive definition of the paradigm may in fact systematically mark three or more distinctions at the level of conventionalized phrases. Inversely, a system with three forms may resolve to two binary distinctions. Hence the number of formal distinctions need not correspond to the number of functional distinctions. Finally, the number of terms in one subsystem of a language need not be identical to the number in another. In Maya for instance, the Manner deictics show a binary split, plus a neutral term, whereas the Spatial deictics make a fiveway distinction and the nominal deictics a four-way split (Hanks 2005a).16 Diessel (1999: 51) presents a useful summary of what he treats as semantic, syntactic and pragmatic features of deixis. Under “deictic semantics” he lists Distance (prox, medial, dist), Visibility (vs. Invisibility), Elevation (up, down relative to origo), Geography (uphill, downhill, upriver, downriver) and Movement (toward, away from, across the visual field of the Spr). Under “quality semantics” he cites Ontology (location, object/person), Animacy, Humanness, Sex, Number and Boundedness (bound, unbound). The difference is that the deictic features apply to the relation between the ground or origo and the denotatum, whereas qualitative features characterize the denotatum itself. Grammatical categories, case marking and agreement are grouped under Syntax, while Pragmatic features consist of Use (exophoric, anaphoric, discourse deictic, recognitional) and Reference (emphasis, contrast, precision). While this way of arraying the features is heuristically useful, it leaves out features of the indexical ground (origo as speaker, addressee, common ground or other) and draws an artifical distinction between semantics and pragmatics (since conventionalized indexicality is both “semantic,” because encoded, and thoroughly “pragmatic”, because linked to usage). Furthermore, by splitting “deictic” from “use” features, the typology obscures the fact that spatial location, vision, endophoric and recognitional features denote alternative modes of access (perceptual and cognitive) to the referent. For our purposes, it is more useful to classify features in terms of the three-part relation between the indexical ground, the relational features and the object denoted (see Figure 1).

4.

The relational values

The relational values of deictics are widely discussed in the literature. They include Relative Immediacy (in space or time), Interiority (inside, outside, lateral), Location vs. Trajectory, Perception (visual or other), and several varieties of Cognitive Access (anaphora, reference to discourse, relative salience, perspective [prospective, retrospective] and recognitional). Temporal deixis poses numerous problems beyond the scope of this essay, including tense, aspect, mode and the status of quasi-deictic shifters like ‘today, tomorrow, yesterday, ago’. Like pronoun systems, temporal shifters are functionally, and sometimes formally related to other kinds of deixis, but sufficiently different to merit separate treatment.

328 4.1.

William F. Hanks

Spatial immediacy

Frei (1944) noted that most deictic systems distinguish two or three degrees of relative proximity between the ground and the object. Examples are English (here ≠ there), Vietnamese (Close vs. Far; Diessel 1999: 36), West Greenlandic (Prox, Medial, Dist; Fortescue 1984 cited by Diessel 1999: 46) and Turkish (Prox, Med, Dist; see Diessel 1999: 60). Among three-term systems, Anderson and Keenan (1985: 282–285) and Diessel (1999: 50) distinguish two kinds: (1) distanceoriented systems, in which the medial term marks mid-range from Spr and the Distal term marks remote from both Spr and Adr, and (2) person-oriented systems in which the medial marks Prox or Identifiable to Adr. The relation between deictic distinctions and participants was addressed by Brugmann (1904), Boas (1911), Bühler (1934), Frei (1944), Kurylowicz (1972) and Fillmore (1982: 47–48). In the current framework, the so-called person oriented systems illustrate not distinct relational features, but distinct indexical grounds, a point to which we return below. According to Anderson and Keenan (1985: 286–289), numerous languages (including Malagasy, Santali, Sre, Quileute, Tlingit) distinguish four or more degrees of distance from ground. Both Fillmore (1982: 48–49) and Diessel (1999: 40–42) argue that no language distinguishes more than three degrees of distance, and such expanded systems always involve some other dimensions (such as perception, Inside ≠ Outside or a shift in the origo). Debates over the status of the medial term and hyper-differentiated distance values both illustrate the difficulty of isolating purely spatial distinctions in the deictic field. In addition to degrees of remove from the indexical origo, some languages distinguish values on a dimension that can be called “interiority”, that is, whether the referent is inside, outside or lateral to the indexical origo. West Greenlandic for instance makes a three-way spatial distinction in the simple demonstratives, but also has a special deictic root for reference to objects outside the utterance space, as well as four separate roots that combine distance with verticality: medial.up, medial.down, distal.up, distal.down relative to origo (Fortescue 1984: 259–262). Denny (1982: 361) reports a five-way distinction for the distal locative ‘there’ in Inuktitut: ‘up.there, down.there, in.there, out.there, over.there (horizontal)’. In the spatial deictics in Yucatec Maya, there is no verticality, but regional forms (as opposed to punctual ones) distinguish Inside (inclusive of origo) from Outside (exclusive of origo)17. In a number of languages, including Nungubuyu, Inuktitut, Santali and Kiowa, there are specialized deictics for reference to objects in motion, for which the origo serves as source space for centrifugual motion, the goal space for centripedal, or the ground for lateral motion (transverse or along the edge). Heath (1980: 152) describes the Nungubuyu forms as “Kinetic”, and shows the three way distinction between [-ala-] ‘approaching’, [-ali-] ‘going away’ and [-waj-] ‘across’. In Inuktitut, deictics may combine with a motion infix [-qa-] plus case marking to distinguish

Deixis and indexicality

329

source, goal and path readings (Denny 1982: 373). For Crow, Graczyk (1986) cites the postposition [-ss(ee)] which marks centripedal trajectory. Zide’s (1972: 267–268) description of Santali deixis includes the infix [-Vh-], which marks that the referent is off to the side (lateral) of the origo space. Diessel (1999: 46) cites a similar distinction in Kiowa between centrifugal and centripedal deixis. In other languages, such as English and many Mayan languages, directionality is encoded in the verb complex, either lexically or by way of affixes (see Fillmore 1975 on English ‘come, go’). Dixon (1972: 48) reports an uphill, downhill distinction in Dyirbal deictics, and Matisoff (1973: 110–111) sites a similar one in Lahu. 4.2.

Perceptual access to the object

In theory, perceptual access to the referent may play a role in any exophoric deixis, and vision is a key source in spatial deixis. While noting the importance of visual access in some deixis, Fillmore (1982: 51) adopted a restrictive definition in which such perceptual conditioning is not deictic per se. By contrast, Bühler (1934) starts from the premise that sound is critical, because to identify the referent of ‘here’ one need only follow the speaking voice to its source. In some languages, such features are conventionalized: certain deictics or bound morphemes indicate the presence or absence of perceptual access to the object. Despite the variation in perceptual coding, it appears that in all languages, if perceptual access is encoded, it is access for the Spr (and not the Adr or some Other) that is in play (Diessel 1999: 41–42). In other words, no language specifically marks that the Object is perceptually available to the Adr, even though there are many cases in which spatial contiguity is computed relative to the Adr (see next section). Diessel cites seven languages in his sample in which Visibility is overtly marked, and there are many more cases. In some languages, both Visible and Invisible are coded in separate roots or stems. For example, Quileute shows a root distinction between Visible and Invisible, where Visible is further subdivided into Prox (Spr), Prox (Adr), Medial (Spr + Adr) and Distant (Andrade 1933). Chinook makes a basic distinction between Visible (also used for Present) and Invisible (also used for Past) (Boas 1911: 617–623). Kwakwa’la shows a similar distinction (Anderson and Keenan 1985: 290). Crow (Graczyk 1986) shows a slight variant in the four-way distinction between Immediate, Medial, Distal (+ Vis), Distal (-Vis). These oppositions between presence and absence of visual access are all equipollent in that both terms are overtly marked. By contrast, many languages show privative oppositions, in which one term is marked and the other unmarked. Santali appears to show a three-way split between stems unmarked for perception, ones marked for “lateral visual” and ones explicitly marked for unseen objects (Garvey et.al. 1964: 115, Zide 1972: 267). West Greenlandic has a special root [-im] for Invisible objects (Fortescue 1984: 259– 262, Diessel 1999: 46), as do the Malagasy locative deictics (Anderson and Kee-

330

William F. Hanks

nan 1985: 293). Diessel (1999: 42) cites Ute to illustrate what he considers the most common alignment: Prox ≠ Distal ≠ Invisible. It appears that in any language with a privative opposition on the parameter of vision, it is the Invisible form that gets special marking. No language has a special form marking Visible unless it also has one for Invisible. The implication is that visual access is the normal state of affairs, and the use of simple proximal or distal forms in a language with an overt Invisible, pragmatically implies that the object is within sight. This is consistent with the fact that vision plays a central role in locating objects in space. The close association between visual and spatial identification is reflected in the widespread tendency among linguists to overlook the role of other senses in identifying objects of deictic reference. With the exception of Bühler (1934) and Hanks (1983, 1990, 1993a, 2005a) none of the works cited in this chapter describes auditory, olfactory, tactile or gustatory access of deictic referents. While such distinctions are evidently rarely encoded in deictics, they frequently part of the phenomenal conditions in which deictic acts occur. But in Yucaec Maya presentative deictics, there is a three way division between Tactual (for objects within touch of the Spr), Visual (objects visible to both participants) vs. Peripheral sensory (objects perceived by hearing, smell or taste) (Hanks 1993a, 2007b).18 4.3.

Cognitive access

Deictics in any language may individuate objects that are cognitively accessible, but neither perceivable nor locatable in the current utterance space. This includes for instance, cognitive salience, as in the Hausa distinction between High and Low focus deictics (Anderson and Keenan 1985: 289, cf. Jaggar and Buba 1994). It also includes oppositions between Prospective (‘here we go’, ‘I’m going this way’) and Retrospective (‘so that’s my story [resumptive conclusion of narrative]’, ‘I came that way’). Prospect and retrospect are closely related to future and past respectively, and when deployed in Discourse, they merge with cataphora and anaphora. In many languages, such as English, non-proximal forms are used for anaphora (recall) and proximal ones for cataphora (anticipation), but in some, there are special anaphoric deictics. Heath (1980) cites a special Anaphoric base in Nungubuyu and Hanks (1990) shows that Yucatec spatial deictics have a form unmarked for distance or interiority, but used for reference to places known (either anaphoric or recognitional). The so called recognitional usage, identified by Schegloff (1972) and cited by Diessel (1999), denotes usages of names, descriptions and deictics in which the Spr assumes that the Adr will be able to identify the referent without further specification. Essentially, this is a species of cognitive access based on common ground. Hanks (1990) and Diessel (1999: 103, 113) distinguish anaphora proper (coreference with a prior expression in the discourse) from reference to prior discourse (as in English “That’s an outrageous claim,” said in response position). In

Deixis and indexicality

331

any spatialist account, these uses are treated either as secondary, non-deictic uses (as in Diessel 1999), or as metaphorical extensions from space (as in Anderson and Keenan 1985). In an interaction-based approach in which the core function of deixis is to signal participant access and direct the Adr’s attention to referents, they demonstrate that cognitive access functions alongside spatial and perceptual access. Terminological differences aside, any language in which indexical reference were limited to the exophoric field of utterance would be too constrained to meet the demands of ordinary talk. This is obvious in any language with dedicated forms for cognitive access, but is a factor of the pragmatics of deixis in any language. 4.4.

Features of the indexical origo

Among the highly significant contributions made by Bühler (1934) was his discussion of what we have called deictic transpositions. These are uses in which the origo of deictic reference is shifted so that forms normally understood in relation to the Spr are grounded on the Adr, some other object, or a narrative space. An easy example is provided by talk to young children, in which a caregiver “speaks for the child” e.g. a mother saying to a grimacing infant ‘Oh Mommy, I don’t like that’. As in quoted speech, the transposed deictics are grounded in the origo of another, in this case the addressee. A narrator can project herself into a narrative space, as when one retells an event from the past, saying ‘So here I am walking down the street and this guy comes up to me and asks me my name’. The use of the proximal deictic and the verb “come” are grounded in the narrative, not in the current situation of speech. A related transposition takes place whenever a distant place or object is imaginatively projected into the current indexical ground of utterance by using unexpected proximal forms: ‘So this is Paris, this is Berlin and this is Strasbourg (all on a table-top)’. Similarly, even very small utterance spaces can be treated as if they were large, thus licensing use of normally remote forms for objects actually within reach: ‘you’ve left your queen way over there’ (in reference to a chess piece on a board between the interlocutors). What these examples have in common is that deictic usage in effect manipulates the indexical ground of utterance. In the examples cited so far, these are purely pragmatic effects with little or no grammatical conventionalization. By contrast, many languages have special deictic forms whose function is to signal shifts in the indexical ground. Anaphora and recognitional uses can be seen in this light if we consider that they function only when both Spr and Adr share cognitive access to the object. Thus anaphoric forms are symmetric in that they presuppose that both participants can recover the antecedent. From a pragmatic perspective, symmetry of access is the prototypical condition for deixis. In the face-to-face, what is close to one party is also close to the other, and their perceptual fields are convergent (overlapping or reciprocal). Key aspects of cognitive access either rest on common ground or are co-constructed. By contrast, actual ego-

332

William F. Hanks

centric or altercentric uses are asymmetric because the origo attaches selectively to one party but not the other. Whenever participants are spatially, perceptually or cognitively separated, the ground is de facto asymmetric. Languages have various ways of encoding different types of origo, as is clear in the so-called person-centered systems. According to Frei (1944), Bisaya marks a six-way distinction between Distal and Proximal relative to three distinct origos: Spr, Adr and Spr+Adr. Anderson and Keenan (1985:287) cite demonstratives in Bemba, which mark a five-way distinction: Immediate (Spr), more Immediate to Spr than to Adr, Immediate (Spr + Adr), Immediate (Adr) and Distal (Spr + Adr). These two languages appear to be maximal in that they distinguish both Proximal and Distal relative to all three origos. Bemba is evidently unique in conventionally distinguishing degrees of relative symmetry of access to the object. Chinook (Boas 1911) and Kwakwa’la (Anderson and Keenan 1985: 290) are similar in that they lack a distal category, but mark Proximal relative to three distinct grounds: Spr, Adr and Other. Japanese (Diessel 1999: 59), Ewondo (Diessel 1999: 40), Spe (Anderson and Keenan 1985: 287) and Quileute (Andrade 1933) are nearly identical in distinguishing Prox (Spr) vs. Prox (Adr) vs Medial or Distal (Spr+Adr), and Diessel cites several other languages with a similar paradigm. Inuktitut (Denny 1982: 362) illustrates yet another solution, with a special affix marking transpositions from the Speaker-ground to Other (including Adr or third). There are several reasons that it is challenging to actually demonstrate these different values for the indexical ground. Detailed observation of ordinary usage is requisite, and most of the published literature simply fails to provide the necessary evidence. In any language lacking dedicated symmetric forms (for which the common ground is the origo), one must infer or guess at how the attested forms are used when access to the object is roughly the same for both participants. In languages with a three-way split among Spr≠Adr≠Other, either the Other category covers common access, or the use of either egocentric or altercentric forms pragmatically implies common ground. Similarly, in languages like Japanese, the question is how one denotes an object proximal to both parties, since shared access is apparently signalled in the distal form only. Heath (1980) does not mention the ground for either Proximal or Distal deictics in Nungubuyu, whereas separate forms mark Prox (Adr) and Anaphoric (which implies shared cognitive access). Presumably the Prox and Distal categories may be either egocentric (Spr) or symmetric (Spr+Adr). If we take the published descriptions at face value, the following generalizations emerge: (1) No language has a distinctive category for the Spr-origo, unless it also has a form for Adr-origo. In the absence of an Adr-origo form, the attested forms are simply unmarked and likely to be used for egocentric, altercentric and shared ground. Adr-centered deixis is more pragmatically marked than Spr-centeric deixis, and more likely to be overtly signalled. (2) Medial and Distal categories tend strongly to relate to symmetric origo (Spr+Adr). Even if Speaker-cen-

Deixis and indexicality

333

tric and Addressee-centric forms are distinguished in the Proximal deictics, Distal forms are likely to be symmetric (Spr+Adr). (3) In languages lacking a dedicated Anaphoric form, the non-Proximal forms are most likely to subserve this function (because they share the indexical symmetry of Distal categories). (4) Immediate forms mark higher salience than their non-Immediate counterparts, and are more likely to be used for new information, high-focus and prospective cognitive access. There are many more subdivisions among deictics in the published literature, and long-term fieldwork in any language demonstrates the complexity of decitic practices, in which all of the foregoing factors may be simultaneously in play. In numerous languages for instance, there are special forms to distinguish precise or punctate from vague or regional reference. Compare Punctual in Yucatec Maya (Hanks 1990), Concrete in Nungubuyu (Heath 1980), Precise in Tsimshian (Boas 1911, Dunn 1979), Intensive in Santali (Zide 1972) and Restriced in Inuktitut (Denny 1982: 372). Insofar as these categories serve to mark the relative precision with which the object is individuated, as in ‘this very one vs. this one’, they nicely display one of the functional requisites of deictic reference, that is, the resolution of reference. Such resolution relies on the co-operation of spatial, perceptual and cognitive access, along with the range of indexical origos to which they are anchored.

Language key to items in bibliography N ATIVE N ORTH A MERICA Tsimshian Tlingit Chinook Wasco-Wishram Bella Coola Crow Mohawk Oneida Inuktitut Aleut Quileute Greenlandic M ESOAMERICA Tlapanec Yucatec Toba Quechua

AND

A RCTIC Boas 1911b, Dunn 1979 Boas 1917, Swanton 1911 Boas 1911a Silverstein 1978 Davis and Saunders 1975, 1976 Graczyk 1986 Bonvillain 1981 Abbott 1981 Denny 1978, 1982 Bergsland 1951 Andrade (1933) Fortescue 1984

AND

L ATIN A MERICA Wichmann 1993 Hanks 1983–2009 (various) Manelis Klein 1981 Calvo Perez 1999

334

William F. Hanks

I NDO -E UROPEAN French Germanic Russian A FRICA Hausa Swahili Malagasy

Morel and Danon-Boileau, eds. 1992 Brugmann 1904, Bühler 1934 Friedrich 1979

Hill 1982, Jaggar and Buba 1994 Leonard 1982, 1985 Cousins 1894, Garvey 1964, Dez 1980, Anderson and Keenan 1985, Hanks 1986

A SIA (S OUTH , S OUTHEAST ), P APUA N EW G UINEA Mandarin Tao 1999 Japanese Kuno 1973 Santali (Munda) Zide 1972 Belhare (Sino-Tibet) Bickel 2000 Lao Enfield 2003, n.d. Lahu Matisoff (1973) Jahai Burenhult 2003, n.d. Yélî Dnye Levinson n.d. (to appear) A USTRALIA Nunggubuyu Ngandi Mara Guugu Yimithirr Dyirbal Wardaman Kayardild

Heath 1980 Heath 1978 Heath 1981 Haviland 1993 Dixon 1972 Merlan 1994 Evans 1995

Notes 1. Cf. Fillmore 1971, Silverstein 1976, Hanks 1984, 1990, 2005a, Diessel 1999, Enfield 2003. 2. Cf. Morris 1946, Peirce 1955, Eco et al. 1988, Husserl 1978, Benveniste 1974, cf. Searle 1969. 3. Cf. Jespersen 1965 [1924], Jakobson 1971 [1957], Silverstein 1976. 4. Cf. Hymes 1974, Gumperz and Hymes 1964, Sherzer 1983, 1991. 5. Cf. Hanks 1990, 2005a, Himmelmann 1996, Janssen 2000, 2002, Leonard 1985, Mondana 2005. 6. See Garfinkel 1967, Schegloff 1972, Enfield 2009, Sidnell 2010. 7. See Clark and Marshall 1981, Clark, Schreuder and Buttrick 1983, Enfield 2006.

Deixis and indexicality

335

8. Cf. Frei 1944, Anderson and Keenan 1985, Diessel 1999: 158, Lyons 1977, Senft and Smits 2000: 65–80, Levinson 2003. 9. See Sacks 1992, Goodwin 1981, Schegloff 1984, Heritage 1984, Sidnell 1997, 2005. 10. Cf. Haviland 1993, 2001, 2003, Goodwin 1994, 2000, Kita 2003, Enfield 2009. 11. Cf. Sacks, Schegloff and Jefferson 1974, Schegloff 1987, 1992, Sidnell 2005, 2010, Stivers 2007. 12. Cf. Schutz 1970, Goodwin 1994, Goodwin and Heritage 1990, Sacks 1992, Schegloff 1972, 1992. 13. See Haviland 1996, 2003, Goodwin 1981, 2003, Sidnell 1997, 2005 and Ochs, Schegloff and Thompson 1996. 14. See Diessel 2005 for similar generalizations over a larger sample of languages. 15. Compare Russell 1940, Reichenbach 1947, Frei 1944, Kurylowicz 1972, Lyons 1982, Fillmore 1997. 16. For recent discussions of deictic typology, see Himmelmann 1996, Fillmore 1997 and Diessel 1999. 17. Cf. Hanks 1990, where I note that Maya and Inuktitut illustrate interiority as well as does Greenlandic; Santali indicates lateral to origo, which I suggest is related to interiority. There is something like a family of effects around interiority, inclusiveness, edge-ness (at boundary of origo). My use of “interiority,” perhaps too vague, is meant to cover this range. 18. From the viewpoint of ordinary usage and native metalinguistic commentaries, it is clear that Yucatec speakers distinguish a three-way split between Tactual, Visual and Peripheral Sensory access, although the comparative evidence suggests that the Peripheral Sensory category may be better analyzed as [+ Perceived, +Invisible], similar to the systems in other languages. In any event it is clear that vision is the most commonly marked modality and it is the absence of visual access that usually gets special marking.

References Abbott, Clifford 1981 Here and there in Oneida. International Journal of American Linguistics 47(1): 50–57. Anderson, Stephen and Edward Keenan 1985 Deixis. In: Timothy Shopen (ed.), Language Typology and Syntactic Description, vol. 3: Grammatical Categories and the Lexicon, 259–308. Cambridge: Cambridge University Press. Andrade, Manuel J. 1933 Quileute. Handbook of American Indian Languages 3, 151–292. New York: Columbia University Press. Bar Hillel, Yehoshua 1954 Indexical expressions. Mind LXIII: 359–379. Benveniste, Emile 1965 Language and Human Experience. Diogenes 51: 1–12. Benveniste, Emile 1966 De la subjectivité dans le langage. In: Problèmes de Linguistique Générale, vol. 1, 258–266. Paris: Gallimard.

336

William F. Hanks

Benveniste, Emile 1974 Le langage et l’expérience humaine. In: Problèmes de Linguistique Générale, vol. 2, 67–78. Paris: Gallimard. Benveniste, Emile 1974 Problèmes de Linguistique Générale, vol. 2. Paris: Gallimard. Bergsland, Knut 1951 Kleinschmidt Centennial IV: Aleut demonstratives and the Aleut-Eskimo relationship. International Journal of American Linguistics 17(3): 167–179. Bickel, Balthazar 2000 Deictic transposition and referential practice in Belhare. Journal of Linguistic Anthropology 10(2): 224–247. Bloomfield, Leonard 1933 Language. New York: Holt. Boas, Franz 1911a Chinook. In: Handbook of American Indian Languages. Smithsonian Institution, Bureau of American Ethnology, Bulletin 40, ed. by Franz Boas. 2 vols, I: 559–678. Washington: Government Printing Office. Boas, Franz 1911b Tsimshian. In: Handbook of American Indian Languages. Smithsonian Institution, Bureau of American Ethnology, Bulletin 40, ed. by Franz Boas. 2 vols. I: 287–414. Washington: Government Printing Office. Boas, Franz 1917 Grammatical Notes on the Language of the Tlingit Indians. Vol. VIII, No. 1. Philadelphia: Anthropological Publications of the University Museum. Bonvillain, Nancy 1981 Locative semantics in Mohawk: Time and space. International Journal of American Linguistics 47(1): 58–65. Brugmann, Karl 1904 Demonstrativpronomina in den indogermanischen Sprachen. Leipzig: Teubner. Bühler, Karl 1934 Sprachtheorie. Jena: Springer. Burenhult, Niclas 2003 Attention, accessibility and the addressee: The case of the Jahai demonstrative ton. Pragmatics 13(3): 363–379. Burenhult, Niclas To appear Distance undone: new light on the Jahai multi-term demonstrative system. In: Michael Dunn and Sérgio Meira (eds.). Demonstratives in cross-linguistic perspective. Calvo Perez, Julio 1999 Pragmatica y gramatica del Quechua Cuzqueño. Cuzco: Centro de Estudios Regionales Andinos Bartolome de las Casas. Chapman, Mackenzie 2009 A brief introduction to demonstratives in Mandarin Chinese. (Unpublished paper, Summer Institute of Linguistics of the Linguistic Society of America, 7pp).

Deixis and indexicality

337

Cicourel, Aaron 1964 Method and Measurement in Sociology. New York: Free Press. Cicourel, Aaron 1974 Cognitive Sociology: Language and meaning in social interaction. New York: Free Press. Cicourel, Aaron 1985 Text and discourse. Annual Reviews of Anthropology 14:159–185. Cicourel, Aaron 1992 The interpenetration of Communicative contexts. In: Alessandro Duranti and Charles Goodwin (eds.), Rethinking Context: Language as an interactive phenomenon, 291–310. Cambridge: Cambridge University Press. Clark, Herbert H. and Catherine Marshall 1981 Definite reference and mutual knowledge. In: Aravind K. Joshi, Bonnie L. Webber and Ivan A. Sag (eds.), Elements of Discourse Understanding, 10–63. Cambridge: Cambridge University Press. Clark, Herbert H., Robert Schreuder and Samuel Buttrick 1983 Common ground and the understanding of demonstrative reference. Journal of Verbal Learning and Verbal Behavior 22: 245–258. Cousins, William E. 1894 A Concise Introduction to the Study of the Malagasy Language as Spoken in Imerina. 3rd ed. Antananarivo: The Press of I.M.S. Davis, Philip W. and Ross Saunders 1975 Bella Coola Nominal Deixis. Language 51(4): 845–858. Davis, Philip W. and Ross Saunders 1976 Bella Coola Deictic Roots. International Journal of American Linguistics 42(4): 319–330. Denny, J. Peter 1978 Locating the universals in lexical systems for spatial deixis. In: Donka Farkas, Wesley M. Jacobsen, and Koral W. Todrys (eds.), Papers from the Parasession on the Lexicon, 71–84. Chicago: Chicago Linguistic Society. Denny, J. Peter 1982 Semantics of the Inuktitut (Eskimo) spatial deictics. International Journal of American Linguistics 48(4): 359–384. Dez, Jacques 1980 Structures de la Language Malgache, Éléments de grammaire à l’usage des francophones. Paris: Publications Orientalistes de France. Diessel, Holgar 1999 Demonstratives: Form, function and grammaticalization. Amsterdam: Benjamins. Diessel, Holger 2005 Distance contrasts in demonstratives. In: Martin Haspelmath, Matthew Dryer, David Gil, and Bernard Comrie (eds.), World Atlas of Language Structures, 170–173. Oxford: Oxford University Press. Dixon, Robert M.W. 1972 The Dyirbal Language of Northern Queensland. Cambridge: Cambridge University Press.

338

William F. Hanks

Dixon, Robert M.W. 1980 The Languages of Australia. Cambridge: Cambridge University Press. Dunn, John A. 1979 A Reference Grammar for the Coast Tsimshian Language. Canadian Ethnology Service, Paper 55. Ottawa: National Museum of Man, Mercury Series. Eco, Umberto, Marco Santambrogio and Patrizia Violi 1988 Meaning and mental representations. Bloomington, IN: Indiana University Press. Enfield, Nicholas J. 2003 Demonstratives in space and interaction: Data from Lao speakers and Implications for semantic analysis. Language 79(1): 82–117. Enfield, Nicholas J. 2006 Social consequences of common ground. In: Nicholas Enfield and Stephen Levinson (eds.), Roots of Human Sociality, Culture, Cognition and Interaction, 399–430. Oxford: Berg. Enfield, Nicholas J. 2009 The Anatomy of Meaning, speech, gesture and composite utterances. Cambridge: Cambridge University Press. Enfield, Nicholas J. and Tanya Stivers (eds.) 2007 Person Reference in Interaction, Linguistic, Cultural and Social Perspectives. Cambridge: Cambridge University Press. Evans, Gareth 1982 The Varieties of Reference. New York: Oxford University Press. Evans, Nicholas 1995 A grammar of Kayardild. Berlin: Mouton de Gruyter. Evans, Nicholas 2003 Bininj Gun-wok: a pan-dialectal grammar of Mayali, Kunwinjku and Kune. 2 vols., xxix + 746. Canberra: Pacific Linguistics. Fillmore, Charles J. 1971 Santa Cruz Lectures on Deixis. Bloomington, IN: Indiana University Linguistics Club. Fillmore, Charles 1982 Towards a linguistic framework for spatial deixis. In: Robert Jarvella and Wolfgang Klein (eds.), Speech, Place and Action, 31–59. Chichester: John Wiley. Fillmore, Charles 1997 Lectures on Deixis. Stanford: CSLI publications. Fortescue, Michael 1984 West Greenlandic. London: Croom Helm. Frei, Henri 1944 Systèmes de déictiques. Acta Linguistica 4: 111–129. Friedrich, Paul 1979 Language, Context and the Imagination. Stanford: Stanford University Press. Gale, Richard M. 1968 Is it now now? Mind 73: 97–105. Gale, Richard M. 1968 Indexical expressions, egocentric particulars and token-reflexive words. In: P. Edwards (ed.), Encyclopedia of Philosophy. Vol IV, 151–155. New York: Collier Macmillan.

Deixis and indexicality

339

Garfinkel, Harold 1967 Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall. Garfinkel, Harold 1972 Studies in the routine grounds of everyday activities. In: David Sudnow (ed.), Studies in Social Interaction, 1–30. New York: Free Press. Garvey, Catherine 1964 Malagasy Introductory Course. Washington DC: Center for Applied Linguistics. Goodwin, Charles 1981 Conversational Organization: Interaction between speakers and hearers. New York: Academic Press. Goodwin, Charles 1994 Professional Vision. American Anthropologist 96(3): 606–633. Goodwin, Charles 2000 Practices of Seeing, Visual Analysis: An Ethnomethodological Approach. In: Theo van Leeuwen and Carey Jewitt (eds.), Handbook of Visual Analysis, 157–82. London: Sage. Goodwin, Charles 2003 Pointing as situated practice. In: S. Kita (ed.), Pointing: Where language, culture and cognition meet, 217–242. Hillsdale, NJ: Erlbaum. Goodwin, Charles 2006 Human sociality as mutual orientation in a rich interactive environment: Multimodal utterances and pointing in aphasia. In: Nicholas Enfield and Stephen Levinson (eds.), Roots of Human Sociality, Culture, Cognition and Interaction, 97–125. Oxford: Berg. Goodwin, Charles and John Heritage 1990 Conversation Analysis. Annual Review of Anthropology 19: 283–307. Graczyk, Randolph 1986 Crow deixis. Unpublished Ms., Department of Linguistics, University of Chicago. Green, Keith (ed.) 1995 New Essays in Deixis. Amsterdam: Rodopi. Grice, H. Paul 1967 Logic and conversation. William James Lectures, Cambridge, MA: Harvard University. Grice, H. Paul 1978 Further notes on logic and conversation. In: Peter Cole (ed.), Syntax and Semantics 9: Pragmatics, 113–128. New York: Academic Press. Gumperz, John 1982 Discourse Strategies. Cambridge: Cambridge University Press. Gumperz, John and Dell Hymes (eds.) 1964 The Ethnography of Communication. American Anthropologist, Special Publication 66, Number 6, Part 2. Halliday, Michael A. K. and Ruqaiya Hasan 1976 Cohesion in English. London: Longman. Hanks, William F. 1983 Deixis and the Organization of Interactive Context. Unpublished PhD thesis. University of Chicago, Department of Linguistics, Department of Anthropology.

340

William F. Hanks

Hanks, William F. 1984 The evidential core of deixis in Yucatec Maya. In: Papers from the Fifteenth Regional Meeting of the Chicago Linguistic Society, 154–173. Chicago: Chicago Linguistic Society. Hanks, William F. 1986 Markedness and category interactions in the Malagasy deictic system. University of Chicago, Department of Linguistics Working Papers. 20pp. Hanks, William F. 1990 Referential Practice: Language and lived space among the Maya. Chicago: University of Chicago Press. Hanks, William F. 1992 The indexical ground of deictic reference. In: Alessandro Duranti and Charles Goodwin (eds.), Rethinking Context: Language as an interactive phenomenon, 43–77. Cambridge: Cambridge University Press. [Reprinted from Papers from the Parasession on Language in Context. Chicago Linguistic Society, 1989]. Hanks, William F. 1993a Metalanguage and pragmatics of deixis. In: John Lucy (ed.), Reflexive language: Reported speech and metapragmatics, 127–157. Cambridge: Cambridge University Press. Hanks, William F. 1993b Notes on semantics in linguistic practice. In: Craig Calhoun and Moise Postone (eds.), Towards a reflexive sociology: The social theory of Pierre Bourdieu, 139–155. Oxford: Basil Blackwell. Hanks, William F. 1995 Language and Communicative Practices. Critical Essays in Anthropology. Edited by Maurice Bloch, Pierre Bourdieu, and John L. Comaroff, Boulder: Westview Press. Hanks, William F. 1996 Language form and communicative practices. In: John Gumperz and Stephen Levinson (eds.), Rethinking Linguistic Relativity, 232–270. Cambridge: Cambridge University Press. Hanks, William F. 1999 Intertexts: Writings on Language, Utterance and Context. Denver: Rowman and Littlefield. Hanks, William F. 2001 Indexicality. In: Alessandro Duranti (ed.), Key terms in language and culture, 119–121. Oxford: Blackwell. Hanks, William F. 2005a. Explorations in the deictic field. Current Anthropology 46(2): 191–220. Hanks, William F. 2005b Pierre Bourdieu and the Practices of Language. Annual Review of Anthropology 34: 67–83. Hanks, William F. 2006a Context. In: Encyclopedia of Language and Linguistics, 2nd edition. Ed. by Keith Brown. Oxford: Elsevier. Hanks, William F. 2006b Deixis. In: Encyclopedia of Literature. London: Routledge.

Deixis and indexicality

341

Hanks, William F. 2006c Spatial frames of reference in language and thought. Review article of Space in Language and Cognition, by Stephen Levinson. Language in Society 35(2): 285–296. Hanks, William F. 2007 Construal of persons in Maya conversation. In: Tanya Stivers and Nicholas Enfield (eds.), Person Reference in Interaction: Linguistic, cultural and social perspectives, 149–171. Cambridge: Cambridge University Press. Hanks, William F. 2007b The evidential core of deixis in Yucatec Maya. In: Zlatka Guerncheva (ed.), L’énonciation médiatisée. Tome II, 311–334. Paris-Louvain: Peeters. Hanks, William F. 2008 Língua como prática social: das relações entre língua, cultura e sociedade a partir de Bourdieu e Bakhtin. Cortez Editora: São Paulo, Brasil. Hanks, William F. 2009 Fieldwork on Deixis. Journal of Pragmatics 41: 10–24. Special issue on Emancipatory Pragmatics. Ed. by William Hanks, Sachiko Ide, Yasuhiro Katagiri. Haviland, John B. 1993 Anchoring, iconicity and orientation in Gugu Yimithirr pointing gestures. Journal of Linguistic Anthropology 3(1): 3–45. Haviland, John B. 1996 Projections, transpositions and relativity. In: John Gumperz and Stephen Levinson (eds.), Rethinking Linguistic Relativity, 271–323. Cambridge: Cambridge University Press. Haviland, John B. 2001 Gesture. In: Alessandro Duranti (ed.), Key Terms in Language and Culture, 83–86. Malden, MA: Blackwell. Haviland, John B. 2003 How to point in Zinacantán. In: Sotaro Kita (ed.), Pointing: Where language, culture and cognition meet, 139–170. Mahwah, NJ: Erlbaum. Heath, Jeffrey 1978 Ngandi Grammar, Texts and Dictionary. Australia Institute of Aboriginal Studies, Canberra. Atlantic Highlands, NJ: Humanities Press. Heath, Jeffrey 1980 Nunggubuyu deixis, anaphora and culture. In: Jody Kreiman and Almerindo Ojeda (eds.), Papers from the Parasession on Pronouns and Anaphora, 151–165. Chicago: Chicago Linguistic Society. Heath, Jeffrey 1981 Basic Materials in Mara: Grammar, Texts and Dictionary. (Pacific Linguistics, Series C, No. 60.) Department of Linguistics, Research School of Pacific Studies. The Australian National University. Heath, Jeffrey 2004 Person. In: Geert Booij, Christian Lehmann, Joachim Mugdan and Stavros Skopeteas (eds.), Morphologie: Ein internationales Handbuch zur Flexion und Wortbildung / Morphology: An International Handbook on Inflection and Word-Formation. Vol. 2, 998–1015. Berlin: de Gruyter.

342

William F. Hanks

Heritage, John 1984 A change-of-state token and aspects of its sequential placement. In: Maxwell Atkinson and John Heritage (eds.), Structures of social action: Studies in conversation analysis, 225–246. Cambridge: Cambridge University Press. Hill, Clifford 1982 Up/down, front/back, left/right: A contrastive study of Hausa and English. In: Jürgen Weissenborn and Wolfgang Klein (eds.), Here and There: Cross-linguistic Studies on Deixis and Demonstration, 13–42. Amsterdam: Benjamins. Himmelmann, Nicholas P. 1996 Demonstratives in narrative discourse: A taxonomy of universal uses. In: Barbara Fox (ed.), Studies in Anaphora. 205–254. Amsterdam: Benjamins. Husserl, Edmund 1978 The origin of geometry. In: Thomas Luckmann (ed.), Phenomenology and Sociology: Selected Readings, 42–70. New York: Penguin. Hymes, Dell 1974 Foundations in Sociolinguistics: An Ethnographic Approach. Philadelphia: University of Pennsylvania Press. Jaggar, Phillip and Malami Buba 1994 The space and time adverbials NAN/CAN in Hausa: cracking the deictic code. Language Sciences 16: 387–421. Jakobson, Roman 1960 Closing statement on linguistics and poetics. In: Thomas Sebeok, (ed.), Style in Language, 350–377. Cambridge, MA: MIT Press. Jakobson, Roman 1971 [1957] Shifters, verbal categories and the Russian verb. In: Roman Jakobson: Selected Writings. Vol. 2: Word and Language, 130–147. The Hague: Mouton. Janssen, Theodorus 2000 Deixis and reference. In: Geert Booij, Christian Lehmann, Joachim Mugdan and Stavros Skopeteas (eds.), Morphologie: Ein internationales Handbuch zur Flexion und Wortbildung / Morphology: An International Handbook on Inflection and Word-Formation, 983–998. Berlin: de Gruyter. Janssen, Theodorus 2002 Deictic principles of pronominal, demonstrative and tense deictics. In: Frank Brisard (ed.), Grounding: The epistemic footing of deixis and reference, 151–196. Berlin: de Gruyter. Jespersen, Otto 1965 [1924] The Philosophy of Grammar. New York: WW Norton. Jungbluth, Konstanze 2005 Pragmatik der Demostrativpronomina in den iberoromanischen Sprachen. Tübingen: Niemeyer. Kendon, Adam 1992 The negotiation of context in face-to face interaction. In: Alessandro Duranti and Charles Goodwin (eds.), Rethinking context, 323–334. Cambridge: Cambridge University Press. Kita, Sotaro (ed.) 2003 Pointing: Where language, culture, and cognition meet. Mahwah, NJ: Erlbaum.

Deixis and indexicality

343

Kita, Sotaro and James Essegbey 2001 Pointing left in Ghana: How a taboo on the use of the left hand influences gestural practice. Gesture 1(1): 73–95. Kuno, Susumo 1973 The Structure of the Japanese Language. Cambridge, MA: MIT Press. Kurylowicz, Jerzy 1972 Universaux linguistiques. In: Luigi Heilman (ed.), Proceedings of the Eleventh Congress of Linguists, 47–73. Bologna. Lasater, Ann 2009 Notes on Quechua deixis. (Unpublished Ms., Summer Institute of Linguistics of the Linguistic Society of America, Berkeley, 9 pp.) Leonard, Robert 1982 The semantic system of deixis in Swahili. PhD thesis, Columbia University. Leonard, Robert 1985 Swahili demonstratives: Evaluating the validity of competing semantic hypotheses. Studies in African Linguistics 16(3): 281–293. Levinson, Stephen 1983 Pragmatics. Cambridge: Cambridge University Press. Levinson, Stephen 2003 Space in language and cognition: Explorations in cognitive diversity. Cambridge: Cambridge University Press. Levinson, Stephen To appear Demonstratives in Yélî Dnye, the language of Rossel Island Papua New Guinea. In: Michael Dunn and Sérgio Meira (eds.), Demonstratives in crosslinguistic perspective. 29 pp. Levinson, Stephen et al. n.d Demonstratives: Towards a semantic typology. MPI Nijmegen working paper (Table, 1 page). Linell, Per 2001 Approaching Dialogue, talk, interaction and contexts in dialogical perspectives. Philadelphia: Benjamins. Lyons, John 1977 Semantics. Cambridge: Cambridge University Press. Lyons, John 1982 Deixis and subjectivity: Loquor ergo sum? In: Robert Jarvella and Wolfgang Klein (eds.), Speech, Place and Action: Studies in Deixis and Related Topics, 101–124. New York: Wiley. Manelis Klein, Harriet E. 1981 Location and Direction in Toba: Verbal morphology. International Journal of American Linguistics 47(3): 227–235. Martinet, André 1960 Elements de Linguistique Générale. Paris: A. Colin. Matisoff, James 1973 The Grammar of Lahu. Berkeley: University of California Press. Meillet, Antoine 1926 Linguistique historique et Linguistique Générale. Paris: Champion.

344

William F. Hanks

Merlan, Francesca 1994 A Grammar of Wardaman: A Language of the Northern Territory of Australia. Berlin: de Gruyter. Merleau-Ponty, Maurice 1967 [1945] The Phenomenology of Perception. Paris: Gallimard. Mondana, Lorenza 2005 La constitution de l’origo déictique comme travail interactionnel des participants: une approche praxéologique de la référence spatiale. Intellectica 41–42: 75–100. Morel, Mary-Annick and Laurent Danon-Boileau (eds.) 1992 La Déixis, Colloque en Sorbonne (8–9 juin 1990). Paris: Presses Universitaires de France. Morris, Charles 1946 Signs, language, and behavior. New York: George Braziller. Nunberg, Geoffrey 1993 Indexicality and deixis. Linguistics and Philosophy 16(1): 1–43. Ochs, Elinor, Emanuel Schegloff and Sandra Thompson (eds.) 1996 Interaction and Grammar. Cambridge: Cambridge University Press. Ozyürek, Asli and Kita Sotaro n.d. Joint attention and distance in the semantics of Turkish and Japanese demonstrative systems. (Ms) Peirce, Charles S. 1955 Philosophical writings of Peirce. Ed. by Justus Buchler. New York: Dover Publications. Reichenbach, Hans 1947 Elements of Symbolic Logic. New York: Macmillan. Rommetveit, Ragnar 1974 On Message Structure. London: Wiley. Russell, Bertrand 1940 An Inquiry into Meaning and Truth. London: Allen and Unwin. Sacks, Harvey 1992 Lectures on conversation. 2 volumes. Ed. by Gail Jefferson. Oxford: Blackwell. Sacks, Harvey, Emanuel Schegloff and Gail Jefferson 1974 A Simplest Systematics for the Organization of Turn-Taking in Conversation. Language 50(4): 696–735. Sadock, Jerrold 1978 On testing for conversational implicature. In: Peter Cole (ed.), Syntax and Semantics 9: Pragmatics, 281–298. New York: Academic Press. Schegloff, Emanuel 1972 Note on a conversational practice: Formulating place. In: David Sudnow (ed.), Studies in Social Interaction, 75–119. New York: Free Press. Schegloff, Emanuel 1982 Discourse as an interactional achievement. In: Deborah Tannen (ed.), Analyzing Discourse: Text and Talk. Georgetown University Round Table on Languages and Linguistics 1981, 71–94. Washington: Georgetown University Press.

Deixis and indexicality

345

Schegloff, Emanuel 1984 On some questions and ambiguities in conversation. In: J. Maxwell Atkinson and John Heritage (eds.), Structures of social action: Studies in conversation analysis. 28–52. Cambridge: Cambridge University Press. Schegloff, Emanuel 1987 Between micro and macro: Context and other connections. In: Jeffrey Alexander, Bernhard Giesen, Richard Münch and Neil J. Smelser (eds.), The MicroMacro Link, 207–234. Berkeley: University of California Press. Schegloff, Emanuel 1992 In another context. In: Alessandro Duranti and Charles Goodwin (eds.), Rethinking context, 191–288. Cambridge: Cambridge University Press. Schutz, Alfred 1970 Phenomenology and Social Relations. Ed. by Helmut R. Wagner. Chicago: University of Chicago Press. Searle, John R. 1969 Speech Acts. Cambridge: University Press. Senft, Gunter and Roel Smits (eds.) 2000 Annual report 2000. Nijmegen: Max Planck Institute for Psycholinguistics. Sherzer, Joel 1983 Kuna Ways of Speaking. Austin: University of Texas Press. Sherzer, Joel 1991 Verbal Art in San Blas. Cambridge: Cambridge University Press. Sidnell, Jack 1997 Organizing Social and Spatial Location: Elicitations in Indo-Guyanese Village Talk. Journal of Linguistic Anthropology 7(2): 143–165. Sidnell, Jack 2005 Talk and Practical Epistemology: The Social Life of Knowledge in a Caribbean Community. Amsterdam: Benjamins. Sidnell, Jack 2010 Conversation Analysis, Comparative Perspectives. Cambridge: Cambridge University Press. Silverstein, Michael 1976 Shifters, linguistic categories, and cultural description. In: Keith Basso and Henry Selby (eds.), Meaning in Anthropology, 11–55. Albuquerque, NM: University of New Mexico Press. Silverstein, Michael 1978 Deixis and deducibility in a Wasco-Wishram passive of evidence. In: Proceedings of the Fourth Annual Meeting of the Berkeley Linguistics Society, 238–253. Berkeley, CA: Berkeley Linguistics Society. Stivers, Tanya 2007 Alternative recognitionals in person reference. In: Tanya Stivers and Nicholas Enfield (eds.), Person Reference in Interaction: Linguistic, cultural and social perspectives, 73–96. Cambridge: Cambridge University Press. Swanton, John 1911 Tlingit. In: Handbook of American Indian Languages. Smithsonian Institution, Bureau of American Ethnology, Bulletin 40, ed. by Franz Boas. 2 vols., I: 159–204. Washington: Government Printing Office.

346

William F. Hanks

Tao, Hongyin 1999 The grammar of demonstratives in Mandarin conversational discourse: A case study. Journal of Chinese Linguistics 27(1): 69–103 Weissenborn, Jürgen and Wolfgang Klein (eds.) 1982 Here and There: Crosslinguistic Studies on Deixis and Demonstration. Amsterdam: Benjamins. Wichmann, Søren 1993 Spatial deixis in Azoyú Tlapanec. Sprachtypologie und Universalienforschung (STUF), Berlin, 46(2): 113–132. Wilkins, David 1999 Demonstrative questionnaire: “THIS” and “THAT” in comparative perspective. In: David Wilkins (ed.), Manual for the 1999 field session. Nijmegen: Max-Planck Institute for Psycholinguistics. Yang, Yuling 2006 A study on the asymmetry of individual zhe and individual na in contexts [in Chinese]. Chinese Teaching in the World 4: 33–41. Zide, Norman 1972 A Munda demonstrative system: Santali. In: Jacqueline M.C. Thomas and Lucien Bernot (eds.), Langues et Techniques, Nature et Sociétés I. Approche Linguistique, 267–274. Paris: Editions Klincksieck.

12.

Reference and anaphora Monika Schwarz-Friesel and Manfred Consten

Introduction This chapter will provide an overview of pragmatic research on how people use and comprehend referring expressions. A referring expression is typically defined as one which is produced in order to identify or re-identify (in case of anaphora) an object or a set of objects for a listener or reader in a relevant domain of text or discourse. Thus, reference and anaphora are concepts inherently pragmatic in nature. Without considering pragmatic notions (such as focus of information, world knowledge, contextual influence) their description in linguistic theory would remain insufficient. We will focus on past and current research on reference and anaphora. One line of this research is to delineate the mental processing of referring expressions with the focus on text comprehension.

1.

Notions of reference

Reference (-theory) is the area of scientific research which deals with the ancient problem of how language relates to the world. Reference is a term that has undergone major changes during the last five decades of research. In this section, the development from a static notion of reference as a relation independent from language users (1.1) to a dynamic notion of reference as a cognitive process (1.2) is outlined. 1.1.

Static notions of reference

In the tradition of philosophical logics and formal semantics, reference has been defined as a static relationship between expressions and the things they denote. The basis for this notion is an objective view of the world. For many philosophers of language, the abstract relationship between words and things is primary, irrespective of intentions, contextual influence and conceptual knowledge. In this view, the function of language is to depict reality; each word stands for a certain class of objects. Linguistics concentrated on the denotation of different kinds of referential expressions. The distinction between sense and reference, formulated by Frege a century ago (1892), had been in the focus of semantic research for quite some time. A semantic theory had to specify for each expression in a language what semantic information that expression conveyed. In the tradition of

348

Monika Schwarz-Friesel and Manfred Consten

Russell (1905), who has defined referring as asserting the existence and the uniqueness of an object, the referential potential of definite descriptions and proper names became a field of intensive linguistic research (see Thrane 1980). Apart from the question whether reference is regarded as a static relation or as a process, it has been discussed which kinds of noun phrase use are referential at all: Donnellan (1966) excluded “attributive use” of definite NPs from his notion of reference. Thus, in his well-known example (1) the NP Smith’s murderer is used referentially only in case the speaker knows who Smith’s murderer is and intends to identify this individual for the hearer. Involving hearers’ roles in reference, Donnellan approaches a dynamic notion of the term (see 1.2). Still, it crucially depends on truth conditionality, and there is no notion of reference to mental concepts like W HOEVER HAS MURDERED S MITH let alone hypothetical referents. (1) Smith’s murderer must be insane. The main topics in reference theory were naming, truth and extension. By focussing on the semantic features of definite noun phrases and reference to real physical objects, however, the actual use of referential expressions was largely neglected. According to formal semantics, reference is not related to an intersubjective, mental or perceptual world, but to a formal model that functions as a substitute for reality (Dowty 1979). In model-theoretic semantics, the intension of an expression is a function from possible worlds to truth values, and an extension (i.e. the set of possible referents) is the truth value for a particular world. Thus, reference theory in modern linguistics has been dominated by a formal and not an empirical approach. A reference theory of truth conditions, however, disregards the mental apparatus of language users. This semantic, static notion of reference is still (albeit implicitly) prevailing in some contemporary formal semantic approaches like Discourse Representation Theory.1 1.2.

Dynamic notions of reference

Dynamic notions of reference crucially arose in the framework of pragmatic approaches like the speech act theory; however, some semanticists made attempts to integrate procedural aspects of reference: Strawson (1950) claimed (at least implicitly) that speakers have to account for the recipient’s capability of identifying a referent and that an act of reference has to be accomplished by the recipient. Reference as a dynamic notion is thus a collaborative activity of speakers and hearers who activate extra-linguistic mental representations by using language. It is based on a common ground, i.e. knowledge shared by speakers and recipients (Clark and Marshall 1981; Clark, Schreuder, and Buttrick 1983). Linguists have realized that it is a mistake to think that the referring expression itself can guarantee uniqueness in discourse. We assign referential distinctness not simply on the basis of lexical or grammatical features such as definiteness but also on some vaguer notions of

Reference and anaphora

349

plausibility and assumptions about the speaker’s knowledge of the distinctive properties of the domain (Stenning 1978). In the 1970s interest became centered on the pragmatic aspects of reference. Reference, it has been argued, is not an abstract semantic relation, but part of a speech act. This is reflected in Donnellan’s (1978) terms of “speaker’s reference” versus “semantic reference” (the latter term being equivalent to “meaning”). With Searle (1969), a reference act is part of a propositional act, which consists of the reference act and a predication about the referent, propositional act and illocutionary act being the core aspects of a speech act. However, pragmatics did not from the outset pay equal attention to the speaker’s and the recipient’s roles in reference acts: reference is merely defined by the speaker’s intention to make the recipient recognize a referent (Donnellan 1966, 1978); accordingly, a reference act is felicitous in case the referent exists, regardless of the hearer’s reception. Similarly, Searle (1969) defines “successful reference” from the speaker’s point of view, but he accounts for recipients who, in this model, will ask for more specification until a state of “fully consummated reference” is achieved. In the late 1970s, interest shifted towards a more cognitive based explanation of reference. Emphasis was now placed on the mental representations and operations involved in referential communication. How recipients identify referents became part of the broader question of how recipients extract the novel information conveyed by an utterance and incorporate this information into memory (Clark and Sengul 1979). “How do we process referential expressions in on-line understanding?” became one of the main questions. Important work has been done in psycholinguistics and cognitive psychology dealing with the constructive nature of referential interpretation. Bransford, Barclay, and Franks (1972) and others have demonstrated that memory representations of a sentence combine information provided by the meaning of the sentence with information drawn from real-world knowledge. Empirical studies have shown that language understanding is a constructive process. In on-line-processing studies it has been shown that recipients instantaneously use pragmatic information (real world knowledge as well as cotextual information) in resolving referential ambiguities and in interpreting anaphors (Marslen-Wilson and Tyler 1982; McDonald and MacWhinney 1995). Accordingly, the structure of semantic memory and the representation of concepts as part of world knowledge has become the focus especially of neuropsychological studies relevant for reference (Damasio et al. 1996; Caramazza and Mahon 2006; Caramazza and Mahon 2009). In keeping with this mentalistic framework, the concept of world has changed as well: It is no longer presumed that reference implies the existence of an object in the physical world (see 1.1.). In reference theory, the traditional realistic approach to perception is being challenged. We cannot make statements about the world independent of our perception and cognition. We can only talk about the world we live in, i.e., the world structured by our perceptual and cognitive mechanisms. Ref-

350

Monika Schwarz-Friesel and Manfred Consten

erence is concerned with the projected world which is a mental construct of our mind/brain. Thus, the notion of world is explained as a mental phenomenon (Jackendoff 1983, 2002; Schwarz 1996). Reference research has also benefited a lot from the development of cognitive science, an interdisciplinary approach to the study of mental faculties which has so strongly been influencing recent research. According to cognitive science, the mind is too complex to be seen clearly from the perspective of a single discipline. Thus, the scientific understanding of cognition requires a synthesis of psychology, linguistics, AI, philosophy, anthropology (and most recently neuroscience), the result of which is an increased interdisciplinary awareness (Schwarz 2008; for methodological reflections and applications Derry, Gernsbacher, and Schunn 2005). Thus, interesting questions concern the nature of our mental knowledge, how we represent this knowledge in memory, and how we put this knowledge to use. The following working hypothesis holds true for cognitive science: The human mind uses mental representations to store knowledge about the world. Human beings are viewed as language processors that make use of this body of stored knowledge to operate on temporary structures. To sum up, the area of reference theory has radically changed. The main achievement is that reference is no longer seen as a static relation, but as the result of a complex mental process which involves all types of knowledge (essentially perceptual, conceptual and linguistic knowledge). A modern theory of pragmatic/cognitive reference must take into account a series of connected topics which until recently have been dealt with in separate fields: pragmatics, language understanding, perceptual processing, and conceptual categorization. The rules and principles which underlie referential communication cannot be explained adequately without the knowledge of how we store and access both meanings and concepts in memory.2 Thus, cognitive reference theory deals with three main questions: What is the knowledge basis of referentiality in long term memory? How do we activate this knowledge in the process of referring? How does this activation interact with more general procedures of cognition (perception and conceptual categorization)? 1.3.

Text, mind, and world: basic assumptions on referential representation and underspecification

Current reference theory does not merely describe the function of referential expressions but attempts to explain the production and comprehension of referential items as part of natural language communication in text and discourse grounded in pragmatic and cognitive principles. To fully understand a text, a recipient must integrate information in one sentence with information in another. Explaining how people do so is one of the main tasks of modern text linguistics. On the grammatical surface of the text, particular expressions (discourse markers) or grammatical

Reference and anaphora

351

patterns serve as explicit cohesive ties to relate different parts of the text.3 The text, as a consequence, is perceived as a coherent structure, rather than as an assembly of disjointed referential information. One of the central assumptions in modern text linguistics, however, is that the explicit grammatical structure of the text is to be regarded as just one part or level of textual coherence. Accordingly, recent developments in text research suggest that the implicit conceptual relations in text are of vital importance and deserve particular attention.4 At the same time, many researchers have realized that no serious analysis of coherence can be carried out without appealing to its cognitive aspects.5 Completing the propositional form by cognitive operations in order to get to the underlying referential representation becomes necessary quite frequently and often involves some kind of semantic-thematic decomposition as can be seen in (2): (2) She dug a hole into the frozen ground. In order to get to the complete conceptual representation, some slot-filling or referent-recreating operation involving WITH AN INSTRUMENT, TYPICALLY A SHOVEL has to be performed.6 In (3) it can be seen that these default elaborations play an important role in establishing and accepting coherence: (3) She opened the door reluctantly. (a) When she turned the key …/ (b) ?? When she turned the spoon …/ (c) ?? When she turned the shovel … The continuation the key in (a) is accepted as coherent without any cognitive effort, whereas in (b) and (c) the recipient would have to make an extra inference to establish some kind of (strange or unusual) continuity relation. Referential underspecification is an essential feature of the relation between verbal expressions and the conceptual representations they are intended to express. Therefore, underspecification and implicitness in text should by no means be regarded as some kind of textual deviance.7 On the contrary, systematic underspecification is to be regarded as default. It is the result of applying the principle of relevance to the production process by leaving out superfluous information (see Sperber and Wilson 1986/21995). In accordance with some principle of cognitive economy, underspecification is achieved by selecting only the really relevant information and thereby avoiding redundancy (Schwarz 2000a: 83; Sanders and Spooren 2001: 4). In order to elaborate the linguistically encoded meaning of a text (which is based on the grammatical and lexical text structure), recipients automatically construct a mental text-world model; i.e. recipients enrich the text base by incorporating both information from the text and information activated through conceptual instantiation and inferential processing.

352 1.4.

Monika Schwarz-Friesel and Manfred Consten

Text-world models and reference construction

The text-world model represents a referential constellation of states-of-affairs in working and in episodic memory which is (in most cases) more complex and elaborated than the semantic text basis, i.e. a propositional level derivable from the text surface. Accordingly, we have to distinguish three levels: the text surface comprising grammatical structure and cohesive means, the semantic level, which is a level of lexical meaning and compositionally derivable from the text, and the text-world model, which is not a text level but a mental level of referential structures (Schwarz 2001: 18). From a procedural perspective, semantic enrichment and conceptual elaboration (as part of text-world model building) occur spontaneously and have to be considered as part of our textual competence. The recipients’ elaborative capacity is based on their memory structures and the ability to activate parts of these knowledge structures or to draw inferences. Hence, text comprehension implies both decoding the linguistically encoded meaning of the text and at the same time constructing a mental text-world model which is a cognitive conceptualization of the world depicted in the text. The term text-world model refers to the assumption that recipients create a mental model of the world described in a specific text and store it in episodic memory (Schwarz 2001). The general idea behind this is that verbal expressions serve as mental processing instructions for the recipient (Givon 1992, 1995; Fauconnier 1994). As already pointed out, the text-world model of a text is based on the information of the text itself but elaborated by the recipients’ cognitive processing. It is the result of applying knowledge representations to the text base. Building up a text-world model is an automatic process which always takes place in the comprehension process.8 Once a text-world model has been developed during reading, all subsequently encoded information will be integrated into the configurations of the model. These configurations or mental files, which are stored in working memory, are constantly being updated as the reading process progresses. For instance, the conceptual file of the referent introduced in the first sentence of (4) includes the information (YOUNG MAN IN C OLOGNE ) in short-term memory. (4) Köln. Ein junger Mann wurde gestern Abend Opfer eines brutalen Überfalls. Als der 25jährige Vizedirektor des Kaufparks in Nippes die Tür öffnete … Joachim H. sah plötzlich eine Kanone vor seinen Augen … Der Vater von 2 kleinen Kindern … (Newspaper article, Kölner Stadtanzeiger, 11 March 2004) ‘Cologne. A young man was brutally assaulted yesterday evening. When the 25-year-old assistant director of the shopping center in Nippes opened the door … Joachim H. saw a gun right before his eyes … The father of 2 little kids …’ With each sentence new information is added to the conceptual knot (which functions as a file address in working memory), and the end result of the referential

Reference and anaphora

353

movement is a file which includes the following information: (Y OUNG M AN J OACHIM H. IS 25 YEARS OLD , HAS 2 LITTLE CHILDREN , WORKS IN C OLOGNE AS ASSISTANT DIRECTOR ). Building up coherence and resolving anaphoric expressions in text comprehension are two processes which go hand in hand. Focussing on the role of the text-world model and the leading principle of conceptual plausibility in understanding and constructing coherent representations of texts, the interface between verbal structures and memory representations has been illustrated. Coherence has been described as conceptual continuity, in the sense that the states-of-affairs expressed by parts of the text can be connected and integrated into the text-world model automatically and without effort on the base of our linguistic and conceptual competence. The maintenance of continuity can best be described as a process of mapping referential representations onto conceptual representations evoked by prior text. 1.5.

Accessibility of referents

One of the most important aspects of explaining referential coherence in text involves the division between old (or given) and new information. Different cognitive accounts concerning the information distribution and thematic development have been proposed in recent years.9 “Old” or “given” information corresponds to entities assumed to be accessible in text, “new” information corresponds to entities to be assumed inaccessible to recipients. Considering the resolution of (definite) referential expressions which connect sentences and serve as signals for topic continuity leads to the question of how recipients access and recover the referents of these expressions in text comprehension and what information has to be available in order to establish continuity. The notion of accessibility is of particular interest here. This notion refers to the degree of activation of information in long- or short-term memory (see Givon 2005 on “grounding”). It is assumed that highly accessible mental entities (those which are most active in consciousness) require less processing effort to be retrieved than do entities of low accessibility.10 Ariel (1990) mentions four factors that affect accessibility: recency of mention (the more recent the last mention of an entity, the more accessible it will be); saliency (either physical or textual salience); competition (salience of an entity compared to other entities of the same type that may also be present in the context); unity (whether an antecedent is within the same paragraph or point of view as an anaphor). One of the main functions of referring expressions is to indicate the degree of accessibility of the mental representations of their referents in a given context (Ariel 1990: 16; see von Heusinger 2000, 2003 for formal aspects of accessibility theory). A problem with these approaches arises from the one-to-one correlation that is assumed to hold between grammatical categories of referring expressions and the

354

Monika Schwarz-Friesel and Manfred Consten

accessibility of referents (such as zero-pronouns indicating highly accessible referents, i.e. in case of anaphoric chains, versus demonstratives indicating the lowest degree of accessibility, i.e. in case of deixis.) Firstly, such a matching seems to be empirically inadequate (i.e. for demonstrative anaphors expressing the speaker’s emotive attitude to the referent, see Averintseva-Klisch and Consten 2007). Secondly, a mere relationship between grammatical forms and referent status, regardless of language user’s perspectives, runs the risk of being a circular explanation (Consten 2004; Consten and Loll 2009): from the speaker’s point of view, properties of referents in a given context seem to be a precondition for using certain referential means. Speakers choose grammatical features (such as definite vs. indefinite, demonstrative vs. non-demonstrative, lexical vs. pronominal NP) according to the status of the referent in discourse (such as identifiability, first-mention vs. resumption). From the recipient’s point of view, grammatical features of the utterance evoke expectations about the referent in question. Grammatical features of referential means are the base for mental operations of reactivating referents retrievable from memory or establishing new referents. Linking accessibility and givenness within the account of our text-world model proposal, referential movement in text comprehension can best be described as a sequence of the following three transitional states: Activation, re-activation and deactivation.11 Activation of referents takes place when a new mental file (formally represented as a knot in a network) is opened. The referent is activated and stored as a conceptual label. Hence, incoming information about the same referent may be filed under this label. This process is to be characterized as re-activation, since the already established knot in the text-world model is once again activated in working memory. If a new referent is mentioned in the text, the referent in current focus is deactivated while at the same time a file for the new referent is opened and stored as an additional conceptual knot in the text-world model (Schwarz 2000b; Schwarz-Friesel 2007a). In this regard, “given” means that the referent is mentally accessible in text-world model, “new” means that this is not the case. The distributive flow of information in text comprehension, then, is to be described as an interplay of retaining continuity and progressive development. Traditionally, definite NPs serving as anaphors are regarded as the main grammatical means to indicate continuity by referring to accessible, given entities. In the following sections, however, we shall present a number of definite anaphoric NPs whose referents are seen as accessible even though they would procedurally be viewed as constituting new instead of given information.

2.

Anaphora and coreference

Anaphorical functions are closely related to the grammatical feature ‘definiteness’. Definiteness is characterized as a grammatical instruction to the recipient to access

Reference and anaphora

355

or construct the referents needed in order to establish a coherent representation.12 Definiteness itself does not spell out the precise connections that have to be constructed in the progress of resolving any individual NP. A unified theory of anaphora has to specify the conditions under which the various uses of anaphors can be resolved and how the resolution depends on the meaning of the antecedents/anchors.13 In order to provide a full account of how several kinds of anaphors are processed, a cognitive theory of reference resolution must include both representational and procedural factors of cognitive processing and pragmatic elaboration. 2.1.

Topic continuity and antecedent-search-models

The traditional view in text linguistics has been that anaphors are used to continue a pre-established reference in text by pointing back to a specific antecedent, usually an NP. Thereby, they sustain the current focus of attention and – at least with longer anaphoric chains – make their referent discourse topic of a text segment. Thus, the main anaphoric function is topic continuity. “The anaphoric procedure is a linguistic instrument for having the hearer continue (sustain) a previously established focus” (Ehlich 1982: 325).14 In this respect, “anaphoric elements tell the recipient where to search for the topic in the existing storage structure of episodic memory” (Givon 1992: 5). Hence, the classical and prototypical anaphoric case can be seen in example (5): (5) A man (referential expression 11) entered the room. Obviously, he/this man (referential expression 21) was drunk. The referential relationship between anaphor and antecedent is based on coreference (indicated by the subscript1), and the link between the expressions denoting coreference can be described in terms of grammatical congruence and semantic compatibility. In this respect, anaphors have been described as “retrieval cues” or “echoes” of their antecedents (Hintzman 1987: 412; Gernsbacher 1989: 100). According to this view, anaphora resolution is seen as some kind of search procedure which aims at discovering the best matching antecedent in textual structure. This mental checking in resolving anaphors in text comprehension always leads to the conclusion that the referential expression 2 refers to the same extra linguistic item as the referential expression 1. In on-line processing, the recipient must decide in each case of reference whether there is referential identity or not. Considering (6), one can conclude that the referent of R2 in (a) is identical with the referent of R1, whereas in (b) this is not the case (which becomes evident to the recipient only when the relative clause in R2 is processed). (6a) Peter G. (R11) has beaten his wife. The man (R21) is quite unscrupulous. (6b) Peter G.(R11) has beaten his wife. The man who lives across the street (R22) observed this.

356

Monika Schwarz-Friesel and Manfred Consten

Anaphors which establish topic continuity by merely reactivating the information knot of their antecedent involve pronouns, NPs with the same head noun, hyperonyms, synonyms and certain general expressions. The analysis of anaphors, however, calls for a more flexible description. Establishing the relationship of identity (R EFERENT OF R2 IS IDENTICAL WITH REFERENT OF R1) which is necessary for successful anaphoric processing very often demands more than just checking semantic compatibility. Reference chains in natural language texts reveal a lot more types of anaphora which do not fall into the category “echo of their antecedent.” The establishment of coreference and continuity relies much more on both the context and constructive mental activity of the recipient than on the grammatical devices in the text. Hence, establishing coherence in text is best to be described as the interaction of linguistic indicators and cognitive processes. The text itself gives only some information potential which offers readers certain (necessary) clues to the mental construction of conceptual representations of what is communicated. 2.2.

Progressive anaphors

We are going to discuss a number of anaphors with explicit antecedents which not only reactivate old files in text-world model constellations, but activate at the same time new information about the referent in focus. It should be noted that these types of anaphors occur quite frequently in natural language texts and are by no means to be evaluated as some kind of textual deviance. To illustrate these “progressive” or “moving” anaphors, consider example (7) which depicts specifying anaphors.15 The G RANDDAUGHTER -referent, who functions as the protagonist, once it is introduced with a lexical NP, is referred to with personal- or zero pronouns, whereas the G RANDMOTHER -referent, as the main theme of the story, is more closely specified with lexical anaphors. (7) Rentnerin erwürgt Hamburg – Eine Frau aus Hamburg-Bergedorf hatte ihre Großmutter seit drei Tagen nicht gesehen und machte sich Sorgen. Als sie die gehbehinderte alte Dame besuchen wollte, entdeckte sie die 81-Jährige tot in ihrer Wohnung. (beginning of a newspaper article, Express, 30 June 2006) ‘Pensioner [fem.]1 strangled Hamburg – A woman from Hamburg-Bergedorf2 had not met her2 grandmother1 for three days and Ø2 worried about her1. When she2 was going to visit the physically handicapped lady1, she2 found the 81-year-old1 dead in her1 flat.‘ In cases like (7), the anaphoric expressions are much more than a mere semantic echo of their antecedent. Several new information chunks are filed successively in the same conceptual node. Hence, continued elaboration through activation processes of the current open file is the case. In on-line comprehension, specifying an-

Reference and anaphora

357

aphors include both re-activation and activation in working memory. In cognitive terms, their informational status is both given and new, since the text-world model constellation changes (see Schwarz 2000b). Further, anaphors (with explicit antecedents) may not only introduce new information about a referent, they may even trigger new referents into the text-world model. In accordance with our conception of coherence, felicity of anaphors depends crucially on the plausibility criterion which is dependent both on long-term memory knowledge and on the text-world model at hand. In (8) we see a “combining anaphor” (“Kombinationsanapher” in German, Schwarz 2000b: 122) combing entities into a new referent with different ontological status.16 (8) Give the powder1 into a glass of water2 and drink down the dilution1+2. Here, a new referent (DILUTION ) is mentally constructed on the basis of old information. Resolving this anaphor means activating a new file for a new referent in the text-world model. If the lexical anaphor was replaced by a pronoun, the result would be awkward since lexical meaning has to be provided in order to specify the new referent. 2.3.

Complex anaphors: condensing and evaluating abstract referents

Referent-creating mechanisms are a major part in the interpretation of so-called complex anaphors as well, as in (9). (9) Unbestritten ist, daß die Zahl der Arbeitsplätze in der Industrie geringer wird, während gleichzeitig das Gewicht des Dienstleistungssektors zunimmt. a) Dieser Prozeß ist auch noch längst nicht abgeschlossen. (Newspaper article cit. from TiGer Korpus 18138–1813917; b) and c) are constructed versions) b) Diese Tatsache prägt die europäische Volkswirtschaft seit Mitte des 20. Jahrhunderts. c) Diese Behauptung ignoriert allerdings Statistiken, die noch Raum für Zweifel lassen. [‘It is indisputable that [the number of jobs is decreasing in the industrial sector with the importance of the service sector growing at the same time.] process ] state a) This process process has not finished yet. b) This fact fact has shaped European economy since the middle of the 20th century. c) This claim proposition, however, ignores statistics still leaving doubts.‘ In these examples, it is not an NP functioning as antecedent, but a whole clause. There are heterogeneous terms for these cases, see Halliday and Hasan (1976) for “extended reference” and “reference to fact”, Fraurud (1992) and Dahl and Hellmann (1995) for “situational anaphora”, Schmid (2000) for “shell nouns” and

358

Monika Schwarz-Friesel and Manfred Consten

Asher (1993) for “abstract object anaphora”, Schwarz (2000a) and Consten, Knees, and Schwarz-Friesel (2007) for “complex anaphora”. Furthermore, the phenomenon is discussed as “discourse deixis” (Webber 1991), see Lenz (1997) and Consten (2004) for differentiation between anaphora and discourse deixis. Referents of the complex anaphoric NPs are not individual entities (although coded by an NP) but second-order or third-order entities (see Lyons 1989 for aspects of semantic ascent) like events, processes, states, facts, and propositions (in (9) indicated by the indices p = process, s = state, f = fact). In (9), the referential structure denoted by the preceding that-clause is picked up as a mere process (i.e., concerning its temporal structure, see (a)) or with regard to its truth-conditional factual state (b), or the whole sentence including the main clause it is indisputable is regarded as the antecedent (thus, the respective referent is the STATE OF BEING INDISPUTABLE ) (see (c)). This referent becomes subject to a kind of metalinguistic statement and, as a result, is ontologically changed into an abstract object unbound to a truth value. With (b) and (c), a higher level in the semantic hierarchy of the text is produced. Functioning as a paraphrase of a preceding part in the text, a hypostatization takes place denoted by the anaphoric expression. On the basis of old information, a new referent is created in the dynamics of the text. Further, complex anaphors do not only include a categorization (e.g. the incident, this assumption) but often evaluate the referent by their lexical content (e.g. this tragedy, this outrageous imputation), see Schwarz-Friesel (2007b: 215–216, 2008: 199–201). Hence, two main functions can be made out: Complex anaphors give a classification (in the sense (X ← Y); THIS EVENT IS A DISASTER ) and a new conceptual file node (DISASTER ) is established in the text-world model. In cognitive terms of the given-new-distinction, both re-activation and activation take place at the same time (for further discussion of complex anaphors see Consten, Knees, and Schwarz-Friesel 2007, 2009; Consten and Knees 2008). So far, we have seen that in many cases the processing of anaphoric devices is not only constrained by their lexical content, but by the activation of conceptual knowledge representations and the performance of referent-creating operations. Besides, the examples discussed show that the resolution of many direct anaphors involves much more than a simple search-and-match-procedure. Their full interpretation requires a more complex cognitive process. Resolving anaphors may involve specification, combination, classification, and/or evaluation mechanisms. From a procedural perspective, many anaphors combine given and new information for the recipient and therefore can be characterized as “given-and-newentities” in the text.

Reference and anaphora

2.4.

359

Reference and inference: indirect anaphora – referring to mental representations

Definite NPs may introduce new referents into the text-world model and at the same time signal topic continuity, which can be most clearly illustrated with the phenomenon of indirect anaphora which by no means is a marginal use of definite NPs. A study by Fraurud (1990, 1992) has shown that in a corpus of written Swedish, over 60 % of all definite NPs occur as first mentions of a referent but are, in fact, necessarily related to the preceding text (see Schwarz 2000a and Consten 2004 for a detailed account of indirect anaphora as a form of domain-bound reference). The passage in (10) is perfectly coherent although there is no textual antecedent for the definite NPs the rails, the conductor and the engine. (10) We arrived late since our train stood still at the rails for one hour. The conductor said, “There are problems with the engine.” (constructed example) Indirect anaphora fall under the heading of anaphora because the interpretation of the definite NPs depends on the existence of an adequate textual element in the preceding text – adopting the terminology of Fraurud (1990, 1992), we will call such elements “anchors” and the process of linking indirect anaphors to these expressions “anchoring”18 (see (10) where train functions as anchor). Further, indirect anaphors establish referential continuity just like direct anaphora. In text comprehension, the resolution of indirect anaphors confronts the cognitive system of the recipient with the problem of identifying and localizing the referent to which the definite description refers in the text. The definite article signals accessibility of the referent, but in the text structure no such referent can be found. Thus, we are concerned with the accessibility of implicit referents in the textworld. The prevailing assumption is that there is an associative relationship between trigger and anaphor which helps to establish some link between the two expressions. Thus, these referential instances are often called “associative anaphors” (e.g. Hawkins 1978; Heim 1991). However, since only a small number of indirect reference relations can be explained by the activation of associations, we prefer the more general term “indirect anaphor” (see Erkü and Gundel 1987).19 According to most theories, the relation between anchor and indirect anaphor can best be explained in terms of associability, and the anchoring always involves some kind of inferential linking or bridging (see Clark 1977; Fraurud 1992; Garrod and Sanford 1994; Cornish 1999). According to another popular view, indirect anaphors are explained as frameevoked entities in implicit focus. In this respect, some authors distinguish direct and indirect anaphors with the help of explicit and implicit focus.20 Explicit focus is the current focus of attention, the highly activated part of workspace where incoming information is temporarily held, that is short-term memory which contains representations of active referents explicitly introduced. Implicit focus is the rep-

360

Monika Schwarz-Friesel and Manfred Consten

resentation of entities evoked implicitly as they form part of a frame retrieved from memory. Whereas the referents of direct anaphora are accessible in explicit focus, the referents of indirect anaphora have to be accessible in implicit focus, that is as default values of a frame or script. However, not all indirect anaphors in a text refer to particular parts of a frame. Besides, we have seen that the resolution of many direct anaphora demands more than a search-and-match in short-term memory. Here, too, mental configurations which exceed the explicit information in the text are built up (by the reader). Therefore, we will argue for a more complex cognitive approach. Taking a procedural perspective, we want to demonstrate that there are different types of indirect anaphora. A distinction will be drawn between indirect anaphora which are based on the activation of default knowledge in long-term memory and indirect anaphora which require a more refined cognitive process for their resolution. Further, we are going to explain why all kinds of indirect anaphora have to be seen as “given-and-new-entities”. Different types of indirect anaphora and their resolution can be gradually distinguished. The criteria established for distinguishing these types are based on the linguistic expressions used as anchor and on the different knowledge structures involved in interpreting the mental relationship between anchor and indirect anaphor. These conditions of anchoring do not sufficiently cover the whole range of the complex phenomenon but show some fundamental procedures of interpreting referring expressions which pragmatically depend on the preceding text. The examples (which are taken from natural language texts in German) concern the most frequent and prototypical types of textual indirect anaphora. (11) Zu den Neuheiten gehört eine elektronische Promille-Wegfahrsperre für Berufskraftfahrer […]. Mit der Zündung eines Autos verbunden, gibt die Barriere den Motor erst dann frei, wenn der Fahrer zuvor einen Alkoholtest per Atemprobe abgegeben hat. (Newspaper article cit. from TiGer Korpus, 6115–61116, modified) ‘One of the novelties is an electronic “blood alcohol immobilizer” for professional drivers. Connected with the ignition circuit of a car, the barrier deblocks the engine only when the driver has absolved a field sobriety test by breath check.’ This type of indirect anaphor is established by meronymic relation between the anchor a car and the indirect anaphors the engine and, in a broader sense of meronymy, the driver. There is a part-whole-relationship between the referents ENGINE and CAR, and an affiliation-relationship between the DRIVER-referent and the CARreferent. Example (12) shows an indirect anaphor based on the thematic roles of the verb park in the preceding sentence: the thematic structure of parking involves an agent (that is addressed directly by you) and an object that is parked. Unlike (11), the

Reference and anaphora

361

close anchoring allows for an pronominal indirect anaphor referring to the expected CAR-referent. (12) Ich wäre wunschlos glücklich, wenn Sie nicht immer auf dem Lehrerparkplatz parken würden. Das nächste mal lasse ich ihn abschleppen. (Beverly Hills 90210, German version of the TV-soap, cit. from Consten 2004: 93) ‘I’d be completely happy if you didn’t always park on the teacher’s parking lot. Next time I will have it towed away.’ In example (13), a script-based type is seen. The anchoring of the indirect anaphor is based on the activation of the appropriate default values in long-term memory. (13) In jenen letzten Augusttagen des Jahres 1961 erwacht in einem kleinen Krankenhauszimmer das Mädchen Rita Seidel … Die Krankenschwester tritt an das Bett … als morgens der Arzt … (Christa Wolf, 1973, Der geteilte Himmel, München: Deutscher Taschenbuchverlag. p.9). ‘In those last days of August 1961, a girl named Rita Seidel awakens in a hospital room. The nurse comes to the bed … the doctor …’ Establishing a coherent link between the sentences in (13) is to treat the nurse and the doctor as fulfilling typical roles or default values of the script H OSPITAL in long-term memory which is activated in a top-down-process by the anchor-expression hospital room. Thus, resolving indirect anaphora involves an elaboration which extends the explicit representation of the text. Comprehension depends on the extension of the current representation, taking into account more than textual information. It is a common view that interpretation of all indirect anaphors is based on some kind of inference or linking. This claim, however, neglects the difference between indirect anaphors which are based on a cognitive process involving the activation of semantic roles or default values and indirect anaphors which are truly based on an inference process. The term inference should not be used for all kinds of memory processes involved in text comprehension in the sense of “any piece of information that is not explicitly stated in the text” (McKoon and Ratcliff 1992: 440), but only for problem-solving activities, that is the constructive use of world knowledge. Otherwise, it is impossible to describe cognitive processes more precisely and in explicit detail. Lexical decomposition, slot-filling and spreading activation represent types of semantic activation processes which should not be equated with using knowledge in inferential problem-solving activities. An indirect anaphor which is resolved not merely by activating a semantic role or default value, but by drawing an inference is exemplified in (14), which is a quite macabre example:

362

Monika Schwarz-Friesel and Manfred Consten

(14) Das Urteil ist somit rechtskräftig und wurde bereits vollstreckt. Die Urne kann im Krematorium Dresden in Empfang genommen werden. (From a letter the East German secret service addressed to a detainee’s wife who had enquired about her missed husband. Published at a historical exhibition at Erfurt, 2000, cit. from Consten 2004: 98) ‘Thus, the judgement is non-appealable and has been enforced already. You can collect the funeral urn at the Dresden crematorium.’ The recipient did not yet know that her husband had been sentenced to death, and this fact is not mentioned in the preceding text of (14) either. Thus, she had to draw the inference THE ADJUDGEMENT WAS A DEATH SENTENCE in order to integrate the definite NP the funeral urn coherently. The process of inferential anchoring does not necessarily involve the activation of one specific frame or script. However, a lot of inferences are based on script knowledge (see Schwarz 2008: 198f). Besides, inferential processing does occur within the configuration (and constraints) of the text-world model, that is, referents have to fulfill requirements of conceptual plausibility. The examples so far discussed show that within the group of indirect anaphors different types can be made out which rely both on different knowledge structures and processing operations for their anchoring (Schwarz 2000a and 2007a; Consten 2004). Indirect anaphors which rely on the activation of semantic or conceptual knowledge are resolved within the immediate mental scope of the anchor expression comprising the prototypical information of the cognitive domain. Resolving inferable indirect anaphors involves the expansion of this scope. In general, anchoring conditions of referents for indirect anaphors can be formulated in the following way: The referent of an indirect anaphor must be either an identifiable part of the implicit semantic structure of the preceding sentences, or it must be a default value of a specific frame or script activated by the anchor expression, or it must be inferable on the basis of cognitive plausibility determined by prototypical world knowledge. The definite NP denoting an indirect anaphor refers to an entity being introduced into the text for the very first time (for the function of definiteness, see the beginning of 2.). In this respect, indirect anaphors carry new information into the text-world model. Even more, a new node for the new referent has to be established in an activation process in working memory. On the other hand, at the semantic level of the text the definite article signals that (given) information is easily accessible and, thus, demands a re-activation process (for indefinite indirect anaphors, see Schwarz 2000a: 67). Seen as grammatical means with processing instructions, indirect anaphors signal accessibility of the discourse referent while a new referent is introduced to the text-world model. From a given-new-perspective, thus, indirect anaphors combine both referential continuity and elaborative development in information processing. An acti-

Reference and anaphora

363

vation process is taking place by establishing a new file in short-term memory, while at the same time we have re-activation in long-term memory. More precisely, the cognitive domain of the anchor expression has been in the status of semi-activation (for aspects of activation levels and the status of semi-activation in longterm memory representations of the anchors, see Schwarz 2000a: 136–139). Hence, re-activation of the semi-activated domain accounts for the ease of processing and the acceptability of indirect anaphors as continuity markers. In this respect, indirect anaphors serve as means to establish coherence in the same way as direct anaphors (Schwarz-Friesel 2007a). To sum up: The anchoring of indirect anaphors is neither solely determined by the principle of associability nor by the activation of a script. It turns out that both associability and inferrability are restricted by the conceptual constraint of plausibility. A strong semantic link between anchor and indirect anaphor does not per se guarantee proper and successful anchoring. Rather, the plausibility within the specific text-world model governs the anchoring process. The immediate resolution of indirect anaphors comes about through an interaction between the information in the anchor and the anaphor, the mental accessibility of the implied referent within a specific context and the focus state of the prior and current textual representation. From a perspective of referential movement, indirect anaphors have to be regarded as “given-and-new-entities.” Combining both easy accessibility and the establishment of new nodes in text-world models, they serve as “progressive continuity markers.” To establish coherence, a mental representation for the textual referent must be either available in the conceptual text representation or inferable from memory at the time the anaphoric item is processed. Given this similarity between direct and indirect anaphors, the interpretation of direct and indirect anaphors seems to obey similar rules (see Schwarz 2000a; Consten 2003, 2004; Schwarz-Friesel 2007a). In both cases, the semantics of these expressions only provides a minimal framework, which needs to be complemented by different cognitive mechanisms at the conceptual level of the text-world model. In this respect, explaining anaphora means at the same time finding an explanation for the interaction between language structures and cognitive processes. Thus, anaphoric reference exemplifies prototypically the pragmatic dimension of reference.

Notes 1. DRT; fundamentals by Kamp (1981) and Heim (1988); for an important application to the theory of anaphora see Asher 1993. 2. In some theories of cognitive semantics a distinction is drawn between concepts (i.e. mental representations that store all kinds of knowledge about entities of the world) and meanings (i.e. concepts bound to linguistic forms). There is strong empirical evidence from aphasia and lg. acquisition for such a distinction (see Schwarz 2008: 59–74).

364

Monika Schwarz-Friesel and Manfred Consten

3. See Andersen and Fretheim (2000); Ariel (1998); Fraser (1999); Lenk (1998a and 1998b); Spooren and Risselada (1998). 4. See Gernsbacher and Givon (1995); Bublitz (1999); Spooren (1999); Sanders and Spooren (2001). 5. See Hobbs (1979); Rickheit and Habel (1995); Sanford and Moxey (1995); Cornish (1999); Schwarz (2000a); Zelinsky-Wibbelt (2003). 6. See Singer (1994); van den Broek (1994) for a detailed discussion of the procedural aspects of these elaborations; Recanati (2004) for problems that the necessity of conceptual elaborations entails for a notion of literal meaning. 7. This is implied in many accounts, i.e. with respect to underspecified anaphors Reinhart (1980); Prince (1981); Heim (1988); Ariel (1990); see Schwarz (2000a: 78). 8. For slightly different conceptions of text-worlds or mental models, see de Beaugrande and Dressler (1981); Van Dijk and Kintsch (1983); Johnson-Laird (1983); Garnham and Oakhill (1992); Givon (1995); Cornish (1999). 9. See Prince (1981); Grosz (1981); Givon (1992); Gordon, Grosz, and Gilliom (1993); Chafe (1994); Lambrecht (1994); Schwarz (2000a, b); Hasselgård et al. (2002) for language specific phenomena at the grammar-discourse-interface. 10. See Gundel et al. (1993); Ariel (1988); Walker, Joshi, and Prince (1998); Byron, BrownSchmidt, and Tanenhaus (2008). 11. See Givon (1992) and Chafe (1994); similar states have been postulated by centering theory, i.e. continuing, retaining and shifting; see Grosz (1981); Gordon, Grosz, and Gilliom (1993); Maes (1997); Walker, Joshi, and Prince (1998); Kibble and Power (1999). 12. For a survey of definite reference, see Consten (2004), for critical remarks on the interpretation of definite determiners, see Consten and Loll (2009). 13. See 2.4 for our notion of anchor. 14. See also Clark and Haviland 1977; Murphy 1985; Carter 1987; Bosch 1988; Givon 1992. 15. They are also called “labelling anaphors”, Francis (1986: 37); “Spezifikationsanaphern” in German, see Schwarz (2000a: 62–64.) 16. For similar examples, see Brown and Yule (1983); Stenning (1992); Freitas and Lopes (1996). 17. The ‘TiGer Korpus’ is a German newspaper corpus released in 2003, see www.ims.uni-stuttgart.de/projekte/TIGER. The numbers indicate the sentence numbers in the original corpus. 18. Hawkins (1978) calls them “trigger”, Cornish (1999) “antecedent trigger”. 19. Indirect anaphora has also been called “bridging reference” or “inferrables”. We prefer the term indirect anaphora since it includes all types of anaphora that occur as firstmention definites in text and need some kind of anchoring but do not necessarily depend on a bridging inference or rely on an associative relation; see Schwarz (2000: 49–50). 20. Sanford and Garrod (1981, 1994a, b); Garrod (1995), for critical remarks, see Cornish (2005) and Cornish et al. (2005).

Reference and anaphora

365

References Andersen, Gisle and Thorstein Fretheim (eds.) 2000 Pragmatic Markers and Propositional Attitude. Amsterdam: Benjamins. Ariel, Mira 1988 Referring and accessibility. Journal of Linguistics 24: 65–87. Ariel, Mira 1990 Accessing noun-phrase antecedents. London: Routledge. Ariel, Mira 1998 Discourse Markers and Form-Function Correlations. In: Andreas H. Jucker and Yael Ziv (eds.), Discourse Markers: Descriptions and Theory, 223–259. Amsterdam: Benjamins. Ariel, Mira 2001 Accessibility theory: an overview. In: Ted Sanders, Joost Schilperoord and Wilbert Spooren (eds.), Text Representation: linguistic and psycholinguistic processes, 29–87. Amsterdam: Benjamins. Asher, Nicholas 1993 Reference to Abstract Objects in Discourse. Dordrecht: Kluver. Averintseva-Klisch, Maria and Manfred Consten 2007 The role of discourse topic and proximity for demonstratives in German and Russian. In: Bergljot Behrens, Cathrine Fabricius-Hansen, Hilde Hasselgård and Stig Johansson (eds.), Language in Contrast Volume 7/Issue 2, 219–240. Amsterdam: Benjamins. Beaugrande, Robert-Alain de and Wolfgang U. Dressler 1981 Einführung in die Textlinguistik. Tübingen: Niemeyer. English translation: Robert-Alain de Beaugrande and Wolfgang U. Dressler 1994 Introduction to Text Linguistics. London: Longman. Bosch, Peter 1988 Representing and accessing focussed referents. Language and Cognitive Processes 3: 207–231. Bransford, John D., Richard R. Barclay and Jefferey J.Franks 1972 Sentence memory: a constructive versus interpretive approach. Cognitive Psychology 3: 193–209 Broek, Paul van den 1994 Comprehension and memory of narrative texts. In: Morton Ann Gernsbacher (ed.), Handbook of Psycholinguistics, 539–588. San Diego: Academic Press. Brown, Gillian and George Yule 1983 Discourse analysis. Cambridge: Cambridge University Press. Bublitz, Wolfram 1999 Introduction: Views of Coherence. In: Wolfram Bublitz, Uta Lenk and Eija Ventola (eds.), Coherence in Spoken and Written Discourse, 1–7. Amsterdam: Benjamins. Byron, Donna K., Sarah Brown-Schmidt and Michael K. Tanenhaus 2008 The overlapping distribution of personal and demonstrative pronouns. In: Gundel, Jeanette K. and Nancy Hedberg, (eds.), Reference – Interdisciplinary Perspectives, 143–175. Oxford: Oxford University Press.

366

Monika Schwarz-Friesel and Manfred Consten

Caramazza, Alfonso and Bradford Mahon 2006 The organization of conceptual knowledge in the brain: the future’s past and some future directions. Cognitive Neuropsychology 23: 13–38. Carter, David M. 1987 Interpreting Anaphors in Natural Language Texts. Chichester: Ellis Horwood. Chafe, Wallace L. 1994 Discourse, consciousness, and time. The flow and displacement of conscious experience in speaking and writing. Chicago: Chicago University Press. Clark, Herbert H. 1977 Bridging. In: Philip Nicholas Johnson-Laird and Peter C. Wason (eds.), Thinking. Readings in cognitive science, 411–420. Cambridge: Cambridge University Press. Clark, Herbert H. and Susan E. Haviland 1977 Comprehension and the given-new contract. In: Roy O. Freedle (ed.), Discourse production and comprehension, 1–40. Hillsdale, NJ: Lawrence Erlbaum. Clark, Herbert H. and Catherine Marshall 1981 Definite reference and mutual knowledge. In: Aravind K. Joshi, Bonnie L. Webber and Ivan A. Sag (eds.), Elements of discourse understanding, 10–63. New York: Cambridge University Press. Clark, Herbert H., Robert Schreuder and Samuel Buttrick 1983 Common ground and the understanding of definite reference. Journal of Verbal Learning and Verbal Behavior 22: 245–258. Clark, Herbert H. and C. J. Sengul 1979 In search of referents for nouns and pronouns. Memory and Cognition 7: 35–41. Consten, Manfred 2003 Towards a unified model of domain-bound reference. In: Friedrich Lenz (ed.), Deictic Conceptualisation of Space, Time and Person, 223–248. Amsterdam: Benjamins. Consten, Manfred 2004 Anaphorisch oder deiktisch? Zu einem integrativen Modell domänengebundener Referenz. Tübingen: Niemeyer. Consten, Manfred and Mareile Knees 2008 Complex Anaphors in Discourse. In: Anton Benz and Peter Kühnlein (eds.), Constraints in Discourse, 181–199. Amsterdam: Benjamins. Consten, Manfred, Mareile Knees and Monika Schwarz-Friesel 2007 The Function of Complex Anaphors in Texts. In: Monika Schwarz-Friesel, Manfred Consten and Mareile Knees (eds.), Anaphors in Texts, 81–102. Amsterdam: Benjamins. Consten, Manfred, Mareile Knees and Monika Schwarz-Friesel 2009 Complex Anaphors. The Impact of Ontology, Cotext and Conceptual Knowledge. In: Jordan Zlatev, Marlene Johansson Falck, Carita Lundmark and Mats Andrén (eds.), Studies in Language and Cognition, 285–302. Newcastle: Cambridge Scholars Publishing. Consten, Manfred and Annegret Loll 2009 Indirekte Anaphern – ein Zirkularitäts-problem zwischen Grammatik und Pragmatik. In: Attila Peteri (ed.), An der Grenze zwischen Grammatik und Pragmatik, 325–332. Frankfurt am Main: Peter Lang.

Reference and anaphora

367

Cornish, Francis 1999 Anaphora, Discourse, and Understanding. Oxford: Clarendon. Cornish, Francis 2005 Degrees of indirectness: two types of implicit referents and their retrieval via unaccented pronouns. In: António Branco, Tony McEnery and Ruslan Mitkov (eds.), Anaphora Processing: Linguistic, cognitive and computational modelling, 199–220. Amsterdam: Benjamins. Cornish, Francis, Alan Garnham, H. Wind Cowles, Marion Fossard and Virginie André 2005 Indirect anaphora in English and French: A cross-linguistic study of pronoun resolution. Journal of Memory and Language 52 (3): 363–376. Dahl, Östen and Christina Hellmann 1995 What happens when we use an anaphor? Paper. Stockholm: University, Department of Linguistics. Damasio, Hanna, Thomas J. Grabowski, Daniel Tranel, Richard D. Hichwa and Antonio R. Damasio 1996 A neural basis for lexical retrieval. Nature 380: 499–505. Derry, Sharon J., Morton Ann Gernsbacher and Christian D. Schunn (eds.) 2005 Interdisciplinary Collaboration – An emerging cognitive science. London: Routledge. Donnellan, Keith S. 1966 Reference and definite descriptions. Philosophical Review 75: 281–304. Donnellan, Keith S. 1978 Speaker reference, descriptions and anaphora. In: Peter Cole (ed.), Pragmatics, 47–68. New York: Academic Press. Dowty, David R. 1979 Word Meaning and Montague Grammar: the Semantics of Verbs and Times in Generative Semantics and in Montague’s PTQ. Dordrecht: Reidel. Ehlich, Konrad 1982 Anaphora and deixis: Same, similar or different? In: Robert J. Jarvella and Wolfgang Klein (eds.), Speech, Place and Action: Studies in Deixis and Related Topics, 315–339. Chichester: Wiley. Erkü, Feride and Jeanette K. Gundel 1987 The pragmatics of indirect anaphors. In Jef Verschueren and Marcella Bertucelli-Papi (eds.), The pragmatic perspective, 533–545. Amsterdam: Benjamins. Fauconnier, Gilles 1994 Mental spaces: Aspects of meaning construction in natural language. 2nd ed. Cambridge: Cambridge University Press. Francis, Gill 1986 Anaphoric Nouns. Birmingham: University of Birmingham. Fraser, Bruce 1999 What are Discourse Markers? Journal of Pragmatics 31: 931–952. Fraurud, Kari 1992 Processing noun phrases in natural discourse. Stockholm: University, Department of Linguistics. Fraurud, Kari 1990 Definiteness and the processing of NPs in natural discourse. Journal of Semantics 7(4): 395–434.

368

Monika Schwarz-Friesel and Manfred Consten

Frege, Gottlob 1892 Über Sinn und Bedeutung. Zeitschrift für Philosophie und philosophische Kritik C: 25–50. Freitas, Sergio and José G. Lopes 1996 Solving the reference to mixable entities. Paper read at the Workshop “Indirect Anaphors” in Lancaster. Garnham, Alan and Jane Oakhill (eds.) 1992 Discourse representation and text processing. A special issue of Language and Cognitiv processes. Hove: Lawrence Erlbaum. Garrod, Simon 1995 Distinguishing between Explicit and Implicit Focus during Text Comprehension. In: Gert Rickheit and Christopher Habel (eds.), Focus and Coherence in Discourse Processing, 3–17. Berlin/New York: de Gruyter. Garrod, Simon C. and Anthony. J. Sanford 1994 Resolving sentences in a discourse context: How discourse representation affects language understanding. In: Morton Ann Gernsbacher (ed.), Handbook of psycholinguistics, 675–698. San Diego: Academic Press. Gernsbacher, Morton Ann 1989 Mechanisms that improve referential access. Cognition 32: 99–156. Gernsbacher, Morton Ann and Talmy Givon (eds.) 1995 Coherence in spontaneous text. Amsterdam: Benjamins. Givon, Talmy 1992 The Grammar of Referential Coherence as Mental Processing Instructions. Linguistics 30: 5–55. Givon, Talmy 1995 Coherence in text vs. coherence in mind. In: Morton Ann Gernsbacher and Talmy Givon (eds.), Coherence in Spontaneous Text, 59–115. Amsterdam: Benjamins. Givon, Talmy 2005 Context as Other Minds. The Pragmatics of Sociality, Cognition and Communication. Amsterdam: Benjamins. Gordon, Peter C., Barbara J. Grosz and Laura A. Gilliom 1993 Pronouns, names, and the centering of attention in Discourse. Cognitive Science 17: 311–347. Grosz, Barbara 1981 Focusing and description in natural language dialogues. In: Aravind K. Joshi, Bonnie L. Webber and Ivan A. Sag (eds.), Elements of discourse understanding, 84–105. Cambridge: Cambridge University Press. Gundel, Jeanette K., Nancy Hedberg and Ron Zacharski 1993 Cognitive Status and the Form of Referring Expressions in Discourse. Language 69(2): 274–307. Halliday, Michael A. K. and Ruqaiya Hasan 1976 Cohesion in English. London: Longman. Hasselgård, Hilde, Stig Johansson, Bergljot Behrens, Cathrine Fabricius-Hansen (eds.) 2002 Information Structure in a Cross-linguistic Perspective. Amsterdam: Rodopi. Hawkins, John 1978 Definiteness and Indefiniteness: A Study in Reference and Grammaticality Prediction. London: Groom Helm.

Reference and anaphora

369

Heim, Irene 1991 Artikel und Definitheit. In: Arnim von Stechow and Dieter Wunderlich (eds.), Semantik. Ein internationales Handbuch der zeitgenössischen Forschung, 487–535. Berlin/New York: de Gruyter. Heim, Irene 1988 The semantics of definite and indefinite noun phrases. New York: Garland. Heusinger, Klaus von 2000 Accessibility, discourse anaphora, and descriptive content. In: Paul Baker (ed.), UCREL Technical Papers Vol. 12: Special Issue Proceedings of the Third International Conference on Discourse Anaphora and Anaphor Resolution (DAARC2000. 173–180. Lancaster: Lancaster University). Heusinger, Klaus von 2003 The double dynamics of definite descriptions. In: Jaroslav Peregrin, (ed.), Meaning in the Dynamic Turn, 150–168. Amsterdam: Elsevier. Hintzman, Douglas L. 1987 Schema abstraction in a multiple-trace model. Psychological Review 93: 411–428. Hobbs, Jerry R. 1979 Coherence and coreference. Cognitive Science 3: 67–90. Jackendoff, Ray 1983 Semantics and Cognition. Cambridge, MA: MIT-Press. Jackendoff, Ray 2002 Foundations of language. Oxford: Oxford University Press. Johnson-Laird, Philip Nicholas 1983 Mental Models. Towards a Cognitive Science of Language, Inference, and Consciousness. Cambridge, MA: Harvard University Press. Kamp, Hans 1981 A theory of truth and semantic representation. In: Jeroen Groenendijk, Theo Janssen and Martin Stokkoff, M. (eds.), Formal methods in the Study of language, 277–322. Amsterdam: Mathematical Theory. Kibble, Rodger and Richard Power 1999 Using centering theory to plan coherent texts. Proceedings of the 12th Amsterdam Colloquium, 187–192. ftp://ftp.itri.bton.ac.uk/pub/reports/ITRI-99–19.ps.gz Lambrecht, Knud 1994 Information structure and sentence form: topic, focus and the mental representation of discourse referents. Cambridge: Cambridge University Press. Lenk, Uta 1998a. Marking Discourse Coherence. Functions of Discourse Markers in Spoken English. Tübingen: Narr. Lenk, Uta 1998b Discourse Markers and Global Coherence in Conversation. Journal of Pragmatics 30: 245–257. Lenz, Friedrich 1997 Diskursdeixis im Englischen. Tübingen: Niemeyer. Lyons, John 1989 Semantic Ascent: A Neglected Aspect of Syntactic Typology. In: Doug Arnold, Claire Grover and Louisa Sadler (eds.), Essays on grammatical theory and universal grammer, 153–186. Oxford: Clarendon Press.

370

Monika Schwarz-Friesel and Manfred Consten

Maes, Alfons 1997 Referent ontology and Centering in Discourse. Journal of Semantics, 14: 207–235. Mahon, Bradford and Alfonso Caramazza 2009 Concepts and Categories: A Cognitive Neuropsychological Perspective. Annual Review of Psychology 60: 27–51. Marslen-Wilson, William and Lorraine Komisarjevsky Tyler 1982 The resolution of discourse-anaphors. Text 2: 263–291. McDonald, Janet L. and Brian MacWhinney 1995 The time course of pronoun resolution: Effects of implicit verb causality and gender. Journal of Memory and Language 34: 543–566. McKoon, Gail and Roger Ratcliff 1992 Inference during meaning. Psychological review 99: 440–466. Murphy, Gregory L. 1985 Processes of Understanding Anaphora. Journal of Memory and Language 24: 290–303. Prince, Ellen F. 1981 Toward a taxonomy of Given-New-information. In: Peter Cole (ed.), Radical Pragmatics, 223–255. New York: Academic Press. Recanati, François 2004 Literal Meaning. Cambridge: Cambridge University Press. Reinhart, Tanya 1980 Conditions of coherence. Poetics Today 1: 161–180. Rickheit, Gert and Christopher Habel (eds.) 1995 Focus and Coherence in Discourse Processing. Berlin/New York: de Gruyter. Russell, Bertrand 1905 On Denoting. Mind 30: 479–493. Sanders, Ted and Wilbert Spooren 2001 Text representation as an interface between language and its users. In: Ted Sanders, Joost Schiperoord and Wilbert Spooren (eds.), Text Representation: Linguistic and psycholinguistic aspects, 1–25. Amsterdam: Benjamins. Sanford, Anthony J. and Simon C. Garrod 1994a Resolving sentences in a discourse context. In: M. A. Gernsbacher (ed.), Handbook of Psycholinguistics, 675–698. San Diego: Academic Press. Sanford, Anthony J. and Simon C. Garrod 1994b Selective processing in text understanding. In: M. A. Gernsbacher (ed.), Handbook of Psycholinguistics, 699–719. San Diego: Academic Press. Sanford, Anthony J. and Simon C. Garrod 1981 Understanding Written Language: Explorations in Comprehension Beyond the Sentence. Chichester: Wiley. Sanford, Anthony J. and Linda M. Moxey 1995 Aspects of coherence in written language: A psychological perspective. In: Morton A. Gernsbacher and Talmy Givon (eds.), Coherence in Spontaneous Text, 161–187. Amsterdam: Benjamins. Schmid, Hans-Jörg 2000 English abstract nouns as conceptual shells. From corpus to cognition. Berlin: Mouton de Gruyter.

Reference and anaphora

371

Schwarz, Monika 1996 Reference as Process. In: Richard A. Geiger (ed.), Reference in a Multidisciplinary Perspective, 461–470. Hildesheim: Olms. Schwarz, Monika 2000a Indirekte Anaphern in Texten. Studien zur domänengebundenen Referenz und Kohärenz im Deutschen. Tübingen: Niemeyer. Schwarz, Monika 2000b Textuelle Progression durch Anaphern – Aspekte einer prozeduralen ThemaRhema-Analyse. Linguistische Arbeitsberichte 74: 111–126. Schwarz, Monika 2001 Establishing Coherence in Text. Conceptual Continuity Within the Domain of Text-world Models. In: Logos and Language 15–23. Schwarz, Monika 2008 Einführung in die Kognitive Linguistik. 3. Aufl.Tübingen: Francke. Schwarz-Friesel, Monika 2007a Indirect anaphora in text: a cognitive account. In: Monika Schwarz-Friesel, Manfred Consten and Mareile Knees (eds.), Anaphors in Text, 3–20. Amsterdam: Benjamins. Schwarz-Friesel, Monika 2007b Sprache und Emotion. Tübingen: Francke. Schwarz-Friesel, Monika, Manfred Consten and Mareile Knees (eds.) 2007 Anaphors in Text. Amsterdam: Benjamins. Searle, John R. 1969 Speech Acts. Cambridge: Cambridge University Press. Singer, Murray 1994 Discourse inference processes. In: Morton Ann Gernsbacher (ed.), Handbook of Psycholinguistics, 479–515. San Diego: Academic Press. Sperber, Dan and Deirdre Wilson 1995 [1986]Relevance: Communication and Cognition.2nd ed. Oxford: Basil Blackwell. Spooren, Wilbert 1999 Textlinguistik. In: Ralf Pörings and Ulrich Schmitz (eds.). Sprache und Sprachwissenschaft, 187–210. Narr: Tübingen. Spooren, Wilbert and Rodie Risselada 1998 Introduction: Discourse Markers and Coherence Relations. Journal of Pragmatics 30: 131–133. Stenning, Keith D. 1978 Anaphora as an approach to pragmatics. In: Morris Halle, Joan Bresnan, and George A. Miller (eds.), Linguistic Theory and Psychological Reality, 162– 200. Cambridge, MA: MIT Press. Stenning, Keith D. 1992 Representations of individuals and the processing of reference change. Quaterly Journal of Experimental Psychology. Human Experimental Psychology 45A 3: 373–398. Strawson, Peter F. 1950 On Referring. Mind 59: 320–344. Thrane, Torben 1980 Referential-Semantic Analysis. Cambridge: Cambridge University Press.

372

Monika Schwarz-Friesel and Manfred Consten

Van Dijk, Teun A. and Walter Kintsch 1983 Strategies of discourse comprehension. New York: Academic Press. Walker, Marilyn A., Aravind K. Joshi and Ellen F. Prince (eds.) 1998 Centering Theory in Discourse. Oxford: Clarendon Press. Webber, Bonnie L. 1991 Structure and ostension in the interpretation of discourse deixis. Language and Cognitive Processes 6: 107–135. Zelinsky-Wibbelt, Cornelia (ed.) 2003 Text, context, concepts. Berlin: de Gruyter.

13.

Speech acts Elena Collavin

1.

Introduction

Speech act theory originated from the works of the Oxford philosopher of ordinary language John Langshaw Austin (1911–1960). Austin first presented the main tenets of his theory in the lectures he gave at Oxford in the years 1952–1954 under the title “Words and deeds”, and subsequently in the William James Lectures he delivered at Harvard University in 1955. However, Austin pointed out that those ideas were formed as early as 1939 (1962:iii). After Austin’s premature death the notes of his William James Lectures were edited, supplemented with earlier notes and tape-recorded lectures, and published in 1962 under the title “How to do Things with Words” (Austin 1962). In 1969 one of Austin’s pupils, the American philosopher John R. Searle, published his own version of the theory under the title “Speech acts – An Essay in the Philosophy of Language”, followed by several more works dedicated to speech acts. Searle’s systematization and development of Austin’s ideas has been very influential, to the point that Searle’s interpretation of the theory is at times taken as the definitive view of speech acts. Two main ideas are at the core of speech act theory: the first holds that the meaning of an utterance is distinct from the function that the utterance performs (what we shall call the force of an utterance); the second is that all utterances amount to the execution of an act. Both these ideas well predate Austin’s theory (for a detailed overview see Sbisà 2009a). However, his is the first account that incorporates both in a radically innovative philosophical explanation of linguistic communication. In a nutshell, speech act theory advances the fundamental claim that speech is a form of action rather than a device for describing the world. The theory compels us to see communication not simply as the passing of information between a speaker and a hearer, but rather as the consequential and mutual acting of participants upon each other. The relevance of the theory extends to the analysis of linguistic communication in its many forms, and more fundamentally to the understanding of human communication tout court. Discussion of the theory soon crossed the borders of language philosophy. As a consequence the literature on the subject is vast, and some references to speech acts are a staple in artificial intelligence, discourse analysis, second language acquisition, interpreting, education, and social psychology.

374 2.

Elena Collavin

Origins

During the nineteen thirties Logical Positivism was a mainstream philosophical movement. Born of polemic with metaphysical theories, the approach charged philosophy with an apparently modest task: to elucidate, clarify and expose nonsense. According to this disciplinary frame, to understand a sentence amounts to knowing what its truth conditions are. Only statements that are potentially empirically verifiable make sense. For instance one understands the sentence (1) Giovanni owns a red car if one knows under which conditions (1) would be true. Within this framework the meaning of the sentence can be described as its truth conditions. In order to calculate the truth value of a sentence one must be able to unequivocally identify its referent, in this case, for example, Giovanni. As a consequence much debate at the time was devoted to statements with non existing referents, like (2) The unicorn owns a red car According to some authors, such a statement would be nonsensical since in our world, no referent could ever be found for ‘the unicorn’, and hence no truth value could ever be bestowed upon it; according to others, (2) is simply false. By either account a large number of perfectly common statements are rendered uninteresting for the theory. Only certain kinds of descriptive sentences of ordinary language can be dealt with by the theory. This theoretical approach is chiefly concerned with the capacity of human language to depict or represent nature, very much like natural sciences aim at doing in formalized ways. Although language is obviously used to do many other things aside from describing, Logical Positivism only concerned itself with the referential, descriptive faculty of language, other uses being deemed marginal. Ludwig Wittgenstein, whose Tractatus (1922) had been a main source of inspiration for Logical Positivism, famously changed his mind and went on to propose a radically different theory of linguistic meaning, which bears affinity with Austin’s. In a nutshell, in his late works Wittgenstein identifies linguistic meaning with whatever participants display they believe it to be, in conformity with the type of social activity they are engaged in. Meaning in such a view becomes an actionoriented business of strictly local, interpersonal conventions, and is de facto detached from the linguistic forms of communication supporting it. While Wittgenstein believes that linguistic forms obey rules, contrary to Austin he does not focus his interest on what these rule are. Austin knew Wittgenstein’s work. Wittgenstein’s Philosophical Investigations were discussed during some of the ‘Saturday morning’ weekly meetings with colleagues that Austin held over the years in various colleges (Warnock 1969:14). Austin’s proposal espouses Wittgenstein’s action-oriented view of linguistic meaning but strives to regiment it by painstakingly trying to pin down, categorize, and

Speech acts

375

peel apart the layers of actions performed linguistically on the one hand, and on the other to correlate form and force of utterances. First, he starts with a rebuttal of the received view of linguistic meaning.

3.

Austin on performative utterances

The pars destruens of Austin’s argument starts from “a kind of utterance which looks like a statement, and grammatically, I suppose, would be classed as a statement, which is not nonsensical, and yet is not true or false” (Austin 1979:235). Here are some examples: (3) (4) (5) (6) (7)

I promise I will come tomorrow I apologize I will kill you I award the plaintiff one million dollars in compensation I bet you ten dollars that tomorrow it will not rain in San Diego

Coining “a new word and an ugly word” Austin calls these utterances performatives to indicate that issuing the utterance is performing an action (Austin 1962: 6). In the appropriate circumstances with (3) the speaker takes on an obligation to come tomorrow, with (4) the speaker issues an apology, with (5) the speaker issues a threat (that is a promise to do something dreadful to the interlocutor), with (6) the speaker attributes an entitlement to the plaintiff, and with (7) the speaker proposes to bet ten dollars that tomorrow it will not rain in San Diego. Noticeably, the truth value of the above utterances does not seem to have much role to play. For instance, to the utterance (8) A: I promise I will come tomorrow it would not be in order to respond (9) B: That’s true but rather something to the tune of “ok”, or “I do not believe you”, or even “you always say that” which all acknowledge the fact that the speaker was making a (perhaps unreliable) promise. Explicit performatives share a similar grammatical structure characterized by the use of the present tense indicative and a first person singular personal pronoun as subject. Such performatives are used to simultaneously name and perform the action denoted by the verb. The following two utterances (10) Susan bets you $10 that tomorrow it will not rain in San Diego (11) I awarded the plaintiff $1,000,000 in compensation

376

Elena Collavin

are descriptive rather than performative due to the use of a third person subject and (10) and the past tense (11); accordingly, uttering (10) does not result in Susan betting $10, and uttering (11) does not award any sum of money. The type of verb involved also plays a part, as in (12) I own a red car which shares the grammatical structure of an explicit performative but lacks a verbum dicendi, which is another prerequisite for a verb being used as a performative. Unlike performatives, (12) can thus be assessed on the truth value scale. Austin calls these latter types of utterances “constatives” because contrary to performatives they are used to describe rather than to do something. While the linguistic form of utterances seems to be an essential aspect of their being performatives, in addition, a number of extra linguistic requirements must be fulfilled for a speaker to perform the intended action. These conditions are of a social and psychological nature: (A.I) There must exist an accepted conventional procedure having a certain conventional effect, that procedure to include the uttering of certain words by certain persons in certain circumstances, and further, (A. 2) the particular persons and circumstances in a given case must be appropriate for the invocation of the particular procedure invoked. (B.1) The procedure must be executed by all participants both correctly and (B.2) completely. (C.1) Where, as often, the procedure is designed for use by persons having certain thoughts or feelings, or for the inauguration of certain consequential conduct on the part of any participant, then a person participating in it and so invoking the procedure must in fact have those thoughts or feelings, and the participants must intend to so conduct themselves, and further (C.2) must actually so conduct themselves subsequently. (Austin 1962: 14–15)

When (5) is uttered by an actor on stage it does not result in the (real) speaker threatening the (real) addressee, and (6) generates an entitlement for the plaintiff only if it is pronounced during the appropriate judicial setting by the judge in charge of deciding the case. The same judge uttering the very same words while talking with her husband at the restaurant does not create any entitlement. Austin suggests a way of testing if a verb is performative by putting “hereby” before the verb conjugated in the first person singular of the indicative present. As expected “I hereby own a red car” is not an acceptable (and thus not a performative) utterance. Performatives are successful or unsuccessful, felicitous or infelicitous, rather than true or false. Austin distinguishes two main ways in which a performative utterance can fail. Violations to the conditions under A and B produce misfires; the act is performed purposefully but is void for lack of the required social convention,

Speech acts

377

persons, and circumstances. Violations of the conditions under C produce abuses; the act is professed but it is hollow because of lack of the appropriate psychological state or consequent behavior. In order to be executed some performatives may need uptake on the part of the interlocutor. An individual person can issue an order or permission on his own, but cannot make a bet without the cooperation of the interlocutor. When it comes to the psychological state of an individual, Austin makes clear that while a promise uttered with no intention to honor it is problematic, for all social purposes the promise has been issued and remains valid, “our word is our bond” (Austin 1962: 10). There are plenty of cases in which it is unclear if a faulty or incomplete performance of a performative utterance results in an infelicity. While administering the constitutionally prescribed Presidential Oath of Office to President Obama, Chief Justice John Roberts botched it by inverting the word order of the oath. As a result the Presidential Swearing In was not perfectly executed word by word. It was unlikely that the imperfect recitation of the oath witnessed by millions of citizens would have had legal bearing for the Presidential Office; nevertheless, later in the day the oath was administered a second time in closed chambers, and apparently perfectly executed (Toobin 2009).

4.

Austin’s theory of the total speech act in the total situation

Austin claims that there is a large number of performative verbs in English which, when put in explicit performative form, effect an action. Here are some: to christen, to bequeath, to promise, to acquit, to sentence, to resign, to excommunicate, to vote for, to bet. However, Austin had other things in mind than sieving the dictionary for special classes of verbs. His rhetorical edifice compels us to first see the difference between two distinct classes of verbs and utterances, only to show that the distinction does not hold. After having first created the constative/performative dichotomy, he ultimately erodes it and argues that all utterances are in fact used to perform speech acts. His argument is twofold. Firstly, there is the matter of linguistic form. We have seen that performatives are a special class of utterances with a particular linguistic form, but this is not always the case. A sign on the wall of a room that says: (13) “No smoking” or the utterance (14) “Smoking is not allowed” are unequivocally performatives but do not have the explicit first person present indicative performative form. Furthermore, (15) “I most definitely will come tomorrow”

378

Elena Collavin

amounts to assuming a commitment, but could not be made more explicit by adding “hereby”. The same action, moreover, can be performed by different linguistic means. Consider: (16) (17) (18) (19) (20) (21)

I compliment you on your wonderful performance! Wonderful performance! You were marvelous! I have never heard a more beautifully executed performance in my life Glinka himself would have loved it Bravo!

All amount to complimenting the performer. Clearly, utterances can be performative without having the form of a performative. Austin suggests that explicit performatives are just more unequivocal than implicit performatives in signaling the action that is being carried out. Austin’s argument implies that underlying a multiplicity of linguistic forms some aspect of the action performed (what we call the force of the utterance) in (16) to (21) is the same and can be made explicit using the performative form. The verb used might not be a good force indicator after all; rather, the mood of the verb, the intonation, adverbs, connecting particles, and overall the circumstances of the utterance all provide indications of what kind of action is performed (Austin 1962: 73–77). Linguistic form, that is, can be deceiving, and many utterances that do not look like explicit performatives can be performatives. Hence the domain of performatives grows, while that of constatives shrinks. Secondly, in Austin’s original distinction, using a performative utterance is an act, i.e. is doing something, whereas using a constative utterance is just stating something but not doing something (Austin 1979: 247). Statements are supposed to be either true or false; performative utterances either felicitous or infelicitous, which is unsatisfactory. First, the very concept of truth is problematic. Consider the utterance (22) France is hexagonal which illustrates the difficulties that can arise if we try to use the concept of truth in the domain of everyday language. If with the statement we meant that France is of a perfect hexagonal shape than the statement would be false, but in many contexts we would agree that the description is perfectly acceptable, or true. The truth value of many descriptive statements is not easily calculable. Furthermore, statements are liable to be assessed not only for their correspondence or failure to correspond with facts; they are also liable to the same infelicities that can affect performative utterances. Let us consider what is wrong with the following sentences: (23) ‘The cat is on the mat but I do not believe it is’ and

Speech acts

379

(24) ‘All John’s children are bald but John hasn’t got any children’ (Austin 1979: 248) In these examples of Moore’s paradox, Austin points out that there is “avowal of insincerity” that makes the statements nonsensical or rather infelicitous for violation of the conditions under C (cf. above). Frege proposed that to assert a proposition is to present it as true. (But see Alston 2007: 11). With a shift from the domain of propositions to that of utterances we can say that asserting something is expected to be associated with the belief that what is asserted is true. Besides, the same criteria that determine the felicity of a performative (appropriate persons and circumstances, the existence of social conventions, correctness of the execution, subsequent conduct, and the speaker’s psychological state (feelings, thoughts and intentions)) all come into play as much in uttering statements as in uttering an order or a promise. Furthermore, performatives can also be assessed for their correspondence to fact. Both the following utterances (25) I killed them and its explicit performative version (26) I confess that I killed them constitute an admission of responsibility, but at the same time both proffer a statement that is either true or false. We see then that stating something is performing an act just as much as giving an order or giving a warning; and we see, on the other hand, that, when we give an order or a warning or a piece of advice, there is a question about how this is related to fact which is not perhaps so very different from the kind of question that arises when we discuss how a statement is related to fact. Well, this seems to mean that in its original form our distinction between the performative and the statement is considerably weakened, and indeed breaks down. (Austin 1979: 251)

Austin rejects the constative/performative dichotomy in favor of a general theory of Speech acts, arguing that all utterances (as actual occurrences of linguistic behavior) are the performance of an action. Utterances do not correspond to well formed grammatical sentences, or to conversational turns (Sacks, Schegloff, and Jefferson 1974). Rather, the speech act can be isolated in utterances as something different and distinct from the meaning of the uttered sentence, something that Austin, using Frege’s terminology, calls the force of the utterance. […] besides the question that has been very much studied in the past as to what a certain utterance means, there is a further question distinct from this as to what was the force, as we may call it, of the utterance. (Austin 1979: 251) What we need besides the old doctrine about meanings is a new doctrine about all the possible forces of utterances […] (Austin 1979: 251)

380

Elena Collavin

For instance: (27) Is the door shut? and (28) Shut the door have the same propositional content but differ in the action that the utterance performs, hence in their force, the first utterance being a question, the second an order. From now on the unit of interest for the theory is the speech act, intended as a verbally performed action. If the business of a speech act is for the speaker to do something with the words he or she utters, a theory of speech acts must explain in which respect saying something may amount to doing something. After all, “‘doing something’ is a very vague expression” (Austin 1962: 91). According to Austin saying something amounts to simultaneously performing three types of acts: i) A Locutionary act, which is the uttering of a sentence with sense and reference. The locutionary act can be broken down into a phonetic act (the act of uttering certain noises), a phatic act (the act of uttering sounds that have sense and reference), and a rhetic act (the act of uttering noises that belong to a certain language vocabulary and grammar). A speech act is always the utterance of a phone, a pheme and a rheme. ii) An Illocutionary act, which is the action performed by virtue of the force associated with a given linguistic expression. It is the performance of an act in saying something, as opposed to the performance of an act of saying something (Austin 1962: 99). This level of action depends on the social conventions that allow us verbally to carry out clearly recognizable actions. The illocutionary force of an utterance can be isolated by asking in which sense we were using a given utterance. When we describe metapragmatically an utterance as an order, a request, a suggestion, a statement, a promise, a threat, we refer to its illocutionary force. For instance, we could say: “By saying ‘the soup you cooked was very poor’ she was criticizing him”. We can say so fairly unequivocally because we usually can recognize a criticism (and an order, a promise, and so on) when we hear or see one. Some verbs (in English, for instance, ‘order’, ‘request’, ‘promise’) are inherently illocutionary: when put in the first person present indicative they provide an explicit performative, whereas others cannot be used in the same way (‘I convince you’ or ‘I offend you’ do not work the same way as an explicit performative). iii) A Perlocutionary act, which is the production of a consequence by the utterance. Contrary to what occurs at the illocutionary level, perlocutions are not directly achieved by the conventional force of an utterance. They can be intentional or unintentional, and they might involve unexpected effects, different from those of an illocutionary act. Also, perlocutions occur at a further level, as

Speech acts

381

the interlocutor’s actual reaction to the speech act. A perlocutionary effect of the utterance “the soup you cooked was very poor” might be that one’s interlocutor is chagrined, gets angry, pours the soup down the sink, or all three. The distinction between illocution and perlocution is not clear cut, and it is not at all apparent that the perlocutionary level can be fully considered as a single level of action embodied in a given utterance (Austin 1962: 108–115). In order to identify the difference it might help to distinguish between the effects that are “built in” to given expression by virtue of its illocutionary force, and those that are not. For instance a (felicitous) order just by virtue of its illocutionary force will generate an obligation in the person who receives it, but how the individual acts upon the obligation (complying, refusing, questioning, being offended by it, and so on) belongs to the realm of perlocutions. The locutionary and the perlocutionary levels of the speech act have generated much debate (Searle, 1969; Searle, 1973, Warnock, 1969, Davis, 1980) but the illocutionary level is the nucleus of the Austinian theory and its later developments. According to Austin the illocutionary act is connected with the production of effects in the following sense: (i) The force and the content of the utterance must be understood. “I cannot say to have warned an audience unless it hears what I say and takes what I say in a certain sense” (Austin 1962: 115). (ii) The illocutionary act needs to take effect (see the above example of how giving an order generates an obligation). (iii) In many instances there needs to be the manifested cooperation of the audience for the illocutionary act to come into being. The necessary uptake encompasses an interlocutor’s understanding of the content and force of an utterance, but it can also require a specific second part (or response) to the utterance that ratifies the first part. While one may protest or give an order on her own, one definitely cannot bet $100 that tomorrow it will not rain in San Diego without someone taking up her bet. Austin’s theory is rich with suggestions and loose ends, a number of which remain unresolved. Scholars continue to debate how to distinguish clearly the three levels of the speech act, how to classify speech acts, how to assess the role of force indicators, and how to apply the values of truth and falsity to speech acts.

5.

Grice’s theory of non natural meaning

After Austin, the development of speech act theory was greatly influenced by Paul Grice’s notion of non natural meaning (Grice 1957; for an overview of Grice’s theory cf. Huang, this volume). In his work, Austin did not focus on the role that the

382

Elena Collavin

speaker’s intentions play in the production of speech acts. While he recognized that the appropriate inner state of the speaker may be fundamental for carrying out felicitous acts of a certain kind, Austin was chiefly concerned with how particular linguistic expressions, social conventions, the interlocutor’s uptake, and allocation of responsibility to the speaker are involved in carrying out verbally performed actions (Sbisà 2009a). Grice’s theory of speaker’s meaning revolutionized the study of language, including speech act theory, by bringing to the fore the communicative intentions of speakers. Grice distinguishes between the meaning of certain natural events and the intentional communicative acts of humans, or “natural signs or symptoms” on the one hand, versus “signals” (Clark 1996: 126) on the other. The former have a natural relationship with what they stand for; they are what Peirce calls indexes (Peirce 1989). The latter instead rely in part at least on the speaker’s intentions to signal something. For instance, the characteristic red spots on a child’s face are natural signs that the child has measles, and have what Grice calls a natural meaning (thus, “the red spots on her face mean she has measles”). However, the doctor’s telling gesture after having examined the child, or her declaring that the child has measles manifests the doctor’s intention to communicate what she believes to be the case (thus “the doctor’s gesture meant that the child has measles”). The latter case in Gricean terms is an instance of “non natural meaning”. Linguistic meaning is always of a non natural kind, even if our talking is likely also to provide a wealth of information that belongs to the realm of indexes (indicating our gender, age, geographic origin, social background, and so forth). Grice further distinguishes between speaker’s meaning and sentence meaning. This is based on the notion that understanding someone amounts to understanding what they intend to communicate to us. In a modified formulation that takes into account some later arguments (proposed by, among others, Strawson, Searle, and Sperber and Wilson), speaker’s meaning is defined as follows: In presenting s to audience A, a speaker S means for A that p if and only if: (i) S intends in presenting s to A that A recognize that p in part by recognizing that i. (Clark 1996: 130)

where i is the speaker’s intention. Speaker’s meaning is a type of reflexive intention, i contains reference to itself. A key part of the meaning resides in the interlocutor’s understanding of the effect that the speaker intends to achieve. A communicative act occurs through signaling and recognizing, and a key part of the process involves understanding the communicative intentions of the speaker. Communication is a joint action, as becomes apparent in cases where the literal meaning of a communicative act appears to be lacking, unclear, incomplete, false or absurd, but communication still succeeds. Indeed, Grice later proposed a general theory of meaning based precisely on what he calls the Cooperative Principle, which alongside the notion of conversational implicature provides a universal explanation of why, with the cooperation of one’s interlocutor and with mutual infer-

Speech acts

383

ential calculations, it is possible to communicate with the most eccentric linguistic and non linguistic means.

6.

Searle’s theory of speech acts

Searle (1969) criticizes Grice’s account of meaning because it relies almost exclusively on the speaker’s intentions. A given meaning can be produced by any sentence “given that the circumstances make possible the appropriate intentions” (Searle 1969: 45). Searle tries to strike a balance between intentionality and the role of conventions in the formation of meaning. In the performance of an illocutionary act in the literal utterance of a sentence, the speaker intends to produce a certain effect by means of getting the hearer to recognize his intention to produce that effect; and furthermore, if he is using words literally, he intends this recognition to be achieved in virtue of the fact that the rules for using the expressions he utters associate the expression with the production of that effect. It is this combination of elements which we shall need to express in our analysis of the illocutionary act. (Searle 1969: 45)

Assuming that the illocutionary force is signaled by indicators of force (that Searle calls Illocutionary Force Indicating Devices) the challenge is to explain exactly how the two – force and force indicators – are linked. For Searle the force indicators are constitutive of the force. Rather than revealing it they create it. Whereas for Austin felicity conditions needed to be fulfilled for the speech act to be carried out successfully, according to Searle felicity conditions are constitutive of the act itself. He distinguishes two elements in the syntactic structure of the sentence, which he calls the propositional indicator, and the illocutionary force indicator. (Searle 1969: 30) This is his version of the distinction between the illocutionary act and the propositional content of the illocutionary act. Searle rejects Austin’s threefold distinction within the locutionary level of the speech act (phonetic, phatic, rhetic) in order to locate the locutionary act at the level of enunciation (the “utterance act”). Others have also criticized Austin’s tripartite distinction claiming that it is hard to keep separate the locutionary act from its constituent levels (Warnock 1969). While some scholars have described Searle’s work as largely a systematization of Austin’s (Levinson 1983: 238), according to Sbisà (2009a) Searle’s theory sharply diverges from Austin’s. For Searle the speech act coincides with the illocutionary act, a conclusion that stems from the fact that Searle views speech acts as physical and psychological gestures of the speaker, whereas for Austin the action corresponds with the (necessarily interpersonal) ascription of responsibility to the agent for the production of a given illocutionary effect. Searle’s illocutionary act/speech act is divided between the illocutionary force and the propositional content, schematized by: F(p), where F describes the illocu-

384

Elena Collavin

tionary force indicators and p is the proposition. The propositional content determines the conditions of satisfaction, truth conditions in the case of belief, obedience in the case of a promise. Searle proceeds by analyzing the conditions sufficient for the successful performance of illocutionary acts. A proposition is what is asserted in the act of asserting, what is stated in the act of stating […] an assertion is a […] commitment to the truth of a proposition (Searle 1969: 29) The illocutionary force indicator shows how the proposition is to be taken, or to put it another way, what illocutionary force the utterance is to have; that is, what illocutionary act the speaker is performing in the utterance of the sentence. (Searle 1969: 30)

Some conditions are common to most illocutionary acts, while others are specific to some types of acts. Austin proposed a classification based on a list of performative verbs, whereas Searle’s is based on the felicity condition of illocutionary acts: all have a propositional content which specifies which propositional content the speaker has to express, preparatory conditions which are contextual but linked to the inner states of the speaker (what he thinks, what he believes, and so on), sincerity conditions and the essential condition which specifies which kind of illocutionary act a speaker intends to carry out. In his later work, Searle focuses on three main criteria for the classification of speech acts: the purpose of the type of act (what he calls the illocutionary point), the direction of fit of the type of act (from words to world, or from world to words) and the expressed psychological state of the speaker (Searle 1975). In either case, in Searle’s writing the illocutionary effect that results from successfully carrying out an illocutionary act depends on the interlocutor’s understanding of the force and propositional content of the act, and coincides with the interlocutor recognizing the speaker’s intentions. Sbisà (2009a, 2009b) notices that Searle’s theory is problematic precisely because it makes the illocutionary effect of the act coincide with the recognition of the speaker’s intentions by the hearer. Rather than looking at the specific illocutionary effects of an act, Searle’s analysis focuses on the types of intentions of the speaker, thus constructing action as the psycho-physical gesture of the speaker. Sbisà (2009b) contends that contrary to Searle, to Austin illocutionary acts are conventional in the sense that they have conventional effects and underlie the interpersonal rather than the mental aspect of uptake of the illocutionary force. On the role of intentionality Searle explains that he sees language as the natural extension of non linguistic biological capacities, and linguistic meaning as an extension of fundamental pre linguistic intentionality of a biologically nature like hunger and sexual desire. Perceptions and beliefs are also forms of intentionality because to hold beliefs is to hold something that determines conditions of satisfaction, and to hold the capacity to recognize if they are fulfilled or not (Searle 2007: 21). Indeed, to Searle linguistic action is in essence a mental act of the speaker. What gets communicated in speech acts are intentional

Speech acts

385

states (Searle 2007: 28). Thus, the analysis focuses on how propositional content, conditions of satisfaction, psychological mode and direction of fit function in the manifestation of intentions. For instance, beliefs aim to represent how things are, thus they have a mind-to-world direction of fit, while desires have a world-to-mind direction of fit. Searle (2002: 7) conceives of the structure of mind and the structure of speech acts as parallel because the structure of mental states and the structure of intentional states are similar. While he concedes that at times our verbal actions do not have the principal scope to transmit our mental states, he underlines that even when we carry out speech acts that have not as principal role the function of expressing feelings, we still express a mental state in the form of a sincerity condition (Searle 2002: 7).

7.

Searle on indirect speech acts

Searle points out that a speaker’s utterance meaning and the sentence meaning frequently diverge (Searle 1975: 59). This is the case most obviously with irony, metaphorical uses, and insinuation. However, even in most ordinary circumstances we can often see no correspondence between the three major sentence types in English (declarative, imperative and interrogative) and the forces that prototypically correspond to them (stating, ordering and requesting). Consider (29) “I want you to do it” (Searle 1975: 59) which is a request but has the declarative form of a statement. This is a very common occurrence. In English, for instance, imperatives are routinely used to invite, as in “take a seat” and “come in”. Utterances that have the illocutionary force indicators for one kind of illocutionary act can then be used to perform another kind of illocutionary act. This is a serious problem for the theory, which is based on the conventional nature of the illocutionary force, and on the existence of predictable indicators of force. A way to treat utterances whose force differs from what their force indicators suggest is to assume that they have a literal force, which is associated by rule, and an indirect force, which is inferred by virtue or further information available to an interlocutor. Searle offers an explanation for the case of directives based on his own felicity conditions. Let us consider the utterance “can you pass the salt?” Given a Speaker S, a hearer H, and an action A (passing the salt) the felicity conditions for such illocutionary act are: Preparatory condition: H is able to perform A Sincerity condition: S wants H to do A Propositional content condition: S predicates a future act A of H Essential condition: Counts as an attempt by S to get H to do A (Searle 1975: 71)

386

Elena Collavin

The request is indirect because the speaker seeks to ascertain if the interlocutor is capable of performing the action while in fact the speaker is actually requesting the hearer to pass the salt. The ability of the hearer to pass the salt is a preparatory condition for requesting to pass the salt. Still, this does not explain why it is the case that only certain requests are successfully made by asking if the interlocutor can perform the wanted action. As a preparatory condition for any request the speaker must hold the belief that the interlocutor can carry out the request, so why is it the case that sometimes the speaker asks? The answer may be found in the rituals of social interaction and the obligations of politeness (Brown and Levinson 1987), a dimension that is not integrated in the notion of illocutionary force.

8.

Bach and Harnish’s theory of speech acts

Although by far the most influential, Searle’s approach is not the only one to emphasize the role of speaker’s intention for explaining illocution. With the aim of offering a philosophically inspired explanation of linguistic communication, Bach and Harnish (1979) propose a different but equally Gricean-inspired, speaker-intention-focused classification of speech acts. They distinguish between conventions and rules, the conventions being actions that when performed in the appropriate circumstances count as something else, while rules are socially expected forms of behavior (Bach and Harnish 1979: 121). Based on their categorical division between conventions and rules the authors draw a distinction between communicative and conventional illocutionary acts: “Communicative illocutionary acts succeed by means of recognition of intention, whereas conventional ones succeed by satisfying a convention” (Bach and Harnish, 1979: 110). Conventional illocutionary acts are endemic to particular institutions. For an utterance to be a conventional illocutionary act not only must it be the utterance that the convention requires, it must be issued by the right person under the right circumstances (Bach and Harnish 1979: 110). Examples of conventional illocutionary acts are christening, billing, sentencing, disqualifying. Conventional illocutionary acts all belong to the class of what they call effectives and verdictives. These types of acts are considered not essentially communicative because they are not implemented by recognition of intention but by convention. Communicative illocutionary acts instead are based not on conventions but on the recognition of the intentions of the speaker (the “illocutionary effect”), and their classification is based on the illocutionary intents of the speaker. Illocutionary intents correspond to types of attitudes expressed by the speaker. The recognition of intentions occurs when the hearer takes the speaker’s utterance as reason to believe that the speaker has the expressed attitude. Although Bach and Harnish assign a different role to conventionality in the discussion of illocutionary force, they side with Searle in limiting the illocutionary force to the recognition of intention of the speaker, and leaving

Speech acts

387

the conventional effects of illocutionary acts within the realm of perlocutions (Sbisa 2009b).

9.

A plan based theory of speech acts

Philip Cohen and Raymond Perrault (1979) propose to describe speech acts as planning operators. While espousing Searle’s recognition of an intention based notion of speech act, they envision a formalization that would allow, for any communicative act, to isolate which intentions are involved and on which bases the speaker expects and intends those intentions to be recognized (Cohen and Perrault 1979: 178). They propose a theory that formally models the possible intentions underlying speech acts. Intentions are conceived as plans and a link is shown between plans and non linguistic behavior. This approach is focused on the individual who in the pursuit of a goal, plans and enacts a sequence of actions, the communication of desires, intentions, and emotional states, and which might include facilitating or making sure that others recognize these plans. Actions are defined in terms of operators that are broken down in preconditions, effects and bodies, the latter being the means by which effects are achieved (Cohen and Perrault 1979: 178). Operators are representations; they are evaluated on the basis of one’s representation of the world (Cohen and Perrault call the speaker “problem-solver”). The idea is that part of one’s model of the world consists of a symbolic description of other people’s models of the world (something akin to “I know what you know” or more modestly “I believe that I know what you believe”). Speech acts act as operators on the models that speaker and hearer maintain of each other. This is not much different from Grice’s model for calculating conversational implicatures (Grice 1975), which occur when what is said is different from what is meant, so that the accurate comprehension of the literal meaning of what is being said would not constitute understanding. Implicatures are pragmatic inferences that require the hearer to make a mental calculation in order to decide what the speaker really meant to say, for if taken literally what was said would not be fitting in the current conversation. Grice bases his model for calculating implicatures on the general Cooperative Principle, a descriptive rule stating that participants’ contributions to a conversation are relevant to the task at hand. The principle acknowledges that in order to comprehend non literal meaning we need to believe that our interlocutor is being cooperative, and as a consequence that his or her contribution must be relevant to the task we are involved in. A successful performance of a speech act produces some modification of the speaker’s model of the hearer, and of the hearer’s model of the speaker (Cohen and Perrault 1979: 179). The question is which modifications? Modeling their proposal on Chomsky’s notion of the ideal competent speaker, the authors focus on a competence theory of speech acts that

388

Elena Collavin

enables to state the conditions under which a competent participant can make the determination necessary to classify one’s utterances as amounting to the performance of a given speech act. The model has been embodied in a computer program, and detailed for two kinds of speech acts: making requests and giving information. But even though Cohen and Perrault’s theory has been influential on artificial intelligence studies applying speech acts theory, it suffers from the same flaws that plague Searle’s model: The speech act occurs within the speaker’s mind as the expression of his or her intentions, leaving no room for a constitutive contribution on the part of the interlocutor, whose role is only to decipher the speaker’s intentions.

10.

Speech acts and the analysis of discourse

To apply speech act theory to the analysis of discourse means to try and isolate the actions that are being performed in ongoing interaction or in a text. No matter how speech acts are conceived, the analysis will consist in mapping discourse onto categories of speech acts. As Levinson (1983: 278) points out, this could be either trivial, if we were to wrongly assume that the force of an utterance is indicated literally by linguistic devices, or otherwise it could involve the attempt to predict the force of a part of discourse in real circumstances. This proves to be largely an impossible task because the function of a given utterance is so essentially dependent on the overall circumstances in which it occurs that there may be little left to do for any theory of speech acts. Consider the following example given by Levinson, which is quoted from Sinclair and Coulthard’s ethnographic work in the classroom: (30) Teacher: What are you laughing at? Child: Nothing (Sinclair and Coulthard 1975: 30) It is the second part of the exchange as much as the scholastic context and the social roles of the participants that provide the cue that the teacher’s utterance is an order. The following example is taken from a 1993 televised political debate in Italy: (31) Casini: I want to ask you this question. Do you want to solve the problems of the south of Italy, or is it the case that you do not give a damn about the problems of the south of Italy? The closed question, which seems to syntactically anticipate either “yes” or “no” as an answer, is most likely a challenge and an accusation, which as it has been shown on the bases of corpus studies, is usually met with a denial (Schegloff, Jefferson and Sacks 1977). Further, an assertion like ‘The door is open’ in the appro-

Speech acts

389

priate circumstances can function as a declarative, a request, a criticism, a question, a complaint, and possibly other kinds of actions as well. Because ultimately it is not possible to associate form with function in the analysis of speech acts, the usefulness of the theory to explain how actions are performed in discourse is limited. As Levinson points out, failing the hypothesis of the literal force, we are left with an infinite array of language games (Wittgenstein 1953: 10–11) within which utterances acquire their meaning (and their force, if we wish to keep the two separated) by virtue of local, situated agreements among participants. Within a theory of speech acts it must be possible to unpack an expression and reveal the underlying illocutionary force of it, and assume that the original expression and the explicit one will display the same force. However, in the analysis of interaction we often find that it is not possible to univocally isolate one particular illocutionary force as it has been categorized within the theory. Further, some of the very reasons why it is interesting to study how people discursively act with one another might be seriously impeded if we are to assume that in the analysis of the speech act other elements, such as those indexing social relationship in terms of closeness and hierarchy, are irrelevant for the force of the utterance. For instance, the utterances (32) Can you pass the salt? (33) Would you be so kind as to pass the salt? (34) Salt please. cannot really be described as fully equivalent in terms of their illocutionary force precisely because they are not equivalent in how they index social relations between participants. Arguably, this aspect, while it does contribute to the forcefulness and the compulsory strength of a request, is not clearly accounted for in terms of illocutionary force. This is even more apparent in languages that use grammatical markers to distinguish formal versus informal forms of addressing. In Italian for instance (35) Mi passi il sale? ‘Do you [+informal] pass me the salt?’ and (36) Mi potrebbe passare il sale? ‘Could you [+formal] pass me the salt?’ have their illocutionary force at least partially embodied in the very grammatical markings that reveal the relationship between participants: the intimate or informal relationship triggers a more forceful and direct request in (35), while the formal relationship in (36) triggers a weaker and less imposing request. Within the theory there is no clear way to account for distinguishing between an illocutionary act of request made between strangers and one made between intimate companions, but to ignore the socially loaded deictic aspects of the exchange is to leave out elements that might actually weight on the force of the action performed. Further,

390

Elena Collavin

Levinson has shown that it is hard to associate a certain set of Searlian felicity conditions to actual elements of speech exchanges: questions, for instance, as a category are too variable to be tied by a given set of felicity conditions, and even apologies do not all have the same requirements of felicity conditions (Levinson 1983: 281; Levinson 1979). Responding to the objections moved by Levinson and Schegloff to the applicability of speech act theory to the analysis of discourse, Rees (1991) defends the specific contribution that the theory can provide to the analysis of verbal interaction, which she suggests be used alongside Conversation Analysis. Van Rees argues that already in Searle’s work there is a recognition that a single utterance might have multiple forces: (37) “It’s really quite late” can be an objection, a suggestion, or even a request (Searle 1969: 70) (and a load of further acts), and in interaction the illocutionary force might reside in supra- and sub-sentential units which do not coincide with a sentence at all (Rees 1991: 34). What counts in the attribution of illocutionary force to a part of discourse then is not a simple, literal association between utterance and force but, as she claims, the possibility to calculate in a replicable, non arbitrary way the force of a given stretch of talk. This compromising approach leaves unresolved the key issue of linguistic form in the matter of the calculability of the illocutionary force. After all, it just shifts the problem in the realm of the competence of the speaker, but still profitably uses the notions of performative and illocutionary force in the analysis of discourse, and has won over many scholars interested in the study of verbal interaction. While the difficulty to find a rule to relate form and function in a wide array of real occurrences of speech acts remains unresolved, key concepts from the theory have been widely adopted within the domain of discourse analysis.

11.

Beyond speech act theory: Action and the body

The speech act approach has largely been abandoned by those dedicated to the ethnographic study of linguistic action. Critics note that in both Austin’s and Searle’s formulation, the unit of analysis is the surgically delimited, well-formed, grammatical sentence, and that despite Austin’s early intuitions the developments of the theory devote no attention to discourse level phenomena which cannot be analyzed without taking social interaction (Goffman 1983) into account. Echoing Austin, Agha points out that the utterance of a performative locution is only a fragment of the performative act; it is merely ‘the leading incident’ (to use Austin’s phrase) in a larger order of happenings whose shape must be specified by further semiotic accompaniments in order for the utterance to have the effect in question (Agha 2007: 58).

Speech acts

391

In Austin’s formulation the total speech act is to be understood in the total situation, a claim that recognizes the fundamental role of participants other than the speaker, and of social arrangements alongside syntactic rules in carrying out speech acts. The division echoes the dichotomy between the meaning of a sentence on the one hand and the role of context on the other, where context supplements, clarifies or modifies the force of the verbal action. A distinction between language as an autonomous and primordial code and additional information provided by “extra linguistic” cues is rejected by those who study cognition in action and who advocate an ethnographic approach, which considers “different kinds of phenomena instantiated in diverse media” (Goodwin 2000: 1490) to understand action as jointly achieved by participants through talk, body movement and material affordances. Goodwin argues that a primordial site for the study of cognition and human action is an actual situation where several people talk while engaged in a task which involves their bodies in space, as well as tools and other physical objects around which they are working. Goodwin’s approach relies on the detailed transcription and analysis of video recorded data of naturally occurring multi party interactions. The data can be girls enthralled in a game of hopscotch, archeologists intent on unearthing an artifact, or surgeons performing a laparoscopic operation. Methodologically it is a far cry from the study of linguistic action which starts with a set of verbs culled from a language’s lexicon, and proceeds with the analysis of hypothetical utterances pronounced in imagined contexts the confines of which can be drawn at leisure. While studying linguistic action, this view pays attention to the multiplicity of semiotic resources available to participants and argues that the construction of action through talk within a situated interaction is accomplished through the temporally unfolding juxtaposition of quite different kinds of semiotic resources […] [T]hrough this process the human body is made publicly visible as the site for a range of structurally different kinds of display implicated in the constitution of the actions of the moment (Goodwin 2000: 1490).

Rather than being treated as distinct domains, language and material structure in the environment are “integrated components of a common process for the social production of meaning and action” (Goodwin 2000: 1490). This perspective proposes a radically different “geography of cognition” (Goodwin 2000: 1490), which is inspired by Erving Goffman’s performance oriented approach to complex interactional situations (Goffman 1959, 1983) by works of the Vygotskian circle and by ethnographic studies of workplace practices.1 Further, Goodwin’s approach pays close attention to the linguistic nuances of what is said, to the synchronic details of the interaction as they have been described within Conversation Analysis (Sacks, Schegloff and Jefferson 1974; Sacks 1995), and to the organization of face to face interaction. The human body is the locus for the occurrence of publicly visible displays that are implicated in the cooperative construction of actions and meanings. In these studies action is rarely achieved by

392

Elena Collavin

one individual, and meaning is a public, social phenomenon that does not somehow reside in the mind of a speaker. In his ethnographic work on the interactional life of a man who suffered a stroke and was left with the use of only two words, Goodwin shows how in such an extreme communicative situation meanings are co-constructed, and actions are the ongoing joint product of the work of participants, always subject to renegotiation and redirection. While apparently distant from some hypothetical model of interaction between competent participants, this aphasic man’s situation is shown to be exemplary of how actions are interactionally performed, and co-constituted through the cooperation of participants. Predictably, Goodwin rejects Searle’s intention centered notion of speech acts. Searle, like most speech acts theorists, focuses almost exclusively on linguistic properties of an utterance, everything else being relegated to perhaps necessary but largely intractable context. A fundamental flaw that Goodwin pinpoints is the role given to participants in the performance of a speech act. In Searle’s theory “the hearer exists only in the speaker’s imagination”; the role of the hearer is to understand the force of the utterance, not to co-construct the actions that are accomplished with talk (Goodwin 2000: 1491). In Goodwin’s work the analysis of action moves away from speaker’s intentions and returns to something closer to Austin’s notion of “uptake”. Social action is linked reflexively to its interpretation (Goodwin 2000: 1491). It can be accomplished only when both speaker and addressee can systematically recognize what is being done not utterance by utterance but within the evolving scenarios that are being projected into from moment to moment. The mutual understanding of what can be the “relevant next” move allows coordinated chains of which are both publicly available and jointly produced. “Speech acts” cannot rely on the mental life of a single participant (the speaker); rather, the notion relies on the sequential analysis of talk in interaction including such intersubjective mechanism as turn taking and repair (Goodwin 2000: 1492) where meanings are publicly negotiated. In this context action is “understood as encompassing this interactively organized process of public recognition of meaningful events reflexively linked to the ongoing production of these same events through the use of appropriate semiotic resources within an unfolding temporal horizon” (Goodwin 2000: 1492). In 1983 Levinson foresaw that unless a correlation could be found between form and force in everyday language use it would be hard to maintain a level of illocutionary force distinct from other action oriented facets of an utterance (Levinson 1983: 282) including social deixis, perlocution, and inferentially available meanings attachable to the utterance. So far, no version of the theory has offered such a correlation. Nonetheless, the main tenets of Austin’s and Searle’s work stand as a groundwork that any general theory of language use needs to confront.

Speech acts

393

Notes 1. In particular Ed Hutchins’ notion of “cognition in the wild”, which he developed during his long lasting research of collaborative cognition in the domain of piloting military and commercial airplanes and freight ships (Hutchins 1995; Hutchins and Palen 1997).

References Agha, Asif 2007 Language and social relations. Cambridge: Cambridge University Press. Alston, William P. 2007 Illocutionary acts and truth. In: Dirk Greimann and Geo Siegwart (eds.), Truth and speech acts, 10–30. New York: Routledge, Austin, John L. 1962 How to do things with words. London: Oxford University Press. Austin, John L. 1979 [1961] Philosophical papers. Ed. by James O. Urmson and Geoffrey J. Warnock, 3rd ed. Oxford: Oxford University Press. [First published by Clarendon Press]. Bach, Kent and Robert M. Harnish 1979 Linguistic communication and speech acts. Cambridge, MA: MIT Press. Brown, Penelope and Stephen Levinson 1987 Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Clark, Herbert H. 1996 Using language. Cambridge: Cambridge University Press. Cohen, Philip R. and C. Raymond Perrault 1979 Elements of a plan-based theory of speech acts. Cognitive Science 6: 177–212. Davis, Steven 1980 Perlocutions. In: John R. Searle, Manfred Bierwisch and Ferenc Kiefer (eds.), Speech act theory and pragmatics, 37–55. Dordrecht: Reidel. Frege, Gottlob 1918 Der Gedanke. Eine logische Untersuchung. Beiträge zur Philosophie des deutschen Idealismus 1(2): 58–77. [Translated in Gottlob Frege, 1988. Ricerche logiche, 43–74. Milano, Guerini,]. Goffman, Erving 1959 The presentation of self in everyday life. Garden City, NY: Doubleday, Anchor Books. Goffman, Erving 1983 The interaction order. American Sociological Review 48: 1–47. Goodwin, Charles 2000 Action and embodiment within situated human interaction. Journal of Pragmatics 32: 1489–1522. Grice, H. Paul 1957 Meaning. Philosophical Review 66: 377–388. [Reprinted in Peter F. Strawson (ed.), 1971. Philosophical logic. 39–48. Oxford: Oxford University Press].

394

Elena Collavin

Grice, H. Paul 1975 Logic and conversation. In: Peter Cole and Jerrold Morgan (eds.), Syntax and semantics 3: Speech acts, 41–58. London: Academic Press. Hutchins, Edwin 1995 Cognition in the wild. Cambridge, MA: MIT Press. Hutchins, Edwin and Leysia Palen 1997 Constructing meaning from space, gesture and speech. In: Lauren Resnick, Roger Saljio, Clotilde Pontecorvo and Barbara Burge (eds.), Discourse, tools and reasoning: Essays on situated cognition, 23–40. Berlin: Springer. Levinson, Stephen 1979 Activity types and language. Linguistics 17: 356–399. Levinson, Stephen 1983 Pragmatics. Cambridge: Cambridge University Press. Peirce, Charles Sanders 1989 What is a sign? In: Peirce Edition Project, The essential Peirce: selected philosophical writings, Volume 2 (1893–1913), 4–10. Bloomington, Indiana: University Press. Rees, M. Agnes van 1991 The adequacy of speech act theory for explaining conversational phenomena: A response to some conversation analytical critics. Journal of Pragmatics 17: 31–47. Sacks, Harvey 1995 Lectures on conversation: Volumes I and II. Ed. by Gail Jefferson, with an introduction by Emanuel A. Schegloff. Oxford: Blackwell. Sacks, Harvey, Emanuel A. Schegloff and Gail Jefferson 1974 A simplest systematics for the organization of turn-taking for conversation. Language 50: 696–735. Sbisà, Marina 2009a Speech act theory. In: Jef Verschueren and Jan-Ola Östman (eds.), Key Notions for Pragmatics, 229–244. Amsterdam: Benjamins. Sbisà, Marina 2009b Uptake and conventionality in illocution. Lodz Papers in Pragmatics 5: 33–52. Schegloff, Emanuel A. 1968 Sequencing in conversational openings. American Anthropologist 70: 1075– 1095. Schegloff, Emanuel. A., Gail Jefferson and Harvey Sacks 1977 The preference for self correction in the organization of repair in conversation. Language 53: 361–382. Searle, John R. 1969 Speech acts. Cambridge: Cambridge University Press. Searle, John R. 1973 Austin on locutionary and illocutionary acts. In: Isaiah Berlin, Lynd W. Forguson, David F. Pears, George Pitcher, John R. Searle, Peter F. Strawson and Geoffrey J. Warnock (eds.), Essays on J.L. Austin, 141–159. Oxford: Oxford University Press. Searle, John R. 1975 Indirect Speech acts. In: Peter Cole and Jerrold Morgan (eds.), Syntax and semantics 3: Speech acts, 59–82. London: Academic Press.

Speech acts

395

Searle, John R. 1979 Expression and meaning. Cambridge: Cambridge University Press. Searle, John R. 2002 Speech acts, minds, and social reality. In: Günther Grewendorf and Georg Meggle (eds.), Speech acts, minds, and social reality, 3–17. Berlin: Springer. Searle, John R. 2007 Illocutionary acts and the concept of truth. In: Dirk Greimann and Geo Siegwart (eds.), Truth and speech acts, 31–40. New York: Routledge. Sinclair, John and Malcolm Coulthard 1975 Towards an analysis of discourse: the English used by teachers and pupils. London: Oxford University Press. Toobin, Jeffrey 2009 Supreme Flub, The New Yorker, Blog Post, January 20, 2009. Available online at: http://www.newyorker.com/online/blogs/tny/2009/01/supreme-flub.html Warnock, Geoffrey J. 1969 John Langshaw Austin, a biographical sketch. In: Kuang Tih Fann (ed.), Symposium on J.L. Austin, 3–21. London: Routledge and Kegan Paul. Warnock, Geoffrey J. 1973 Some types of performative utterances. In: Isaiah Berlin, Lynd W. Forguson, David F. Pears, George Pitcher, John R. Searle, Peter F. Strawson and Geoffrey J. Warnock (eds.), Essays on J.L. Austin, 69–89. Oxford: Oxford University Press. Wittgenstein, Ludwig 1922 Tractatus logico-philosophicus. London: Routledge and Kegan Paul. Wittgenstein, Ludwig 1953 Philosophical Investigations. Oxford: Blackwell.

14.

Types of inference: entailment, presupposition, and implicature1 Yan Huang This article is dedicated to Professor Nigel Vincent, my mentor at Cambridge, on the occasion of his sixty-fifth birthday.

Introduction As one of the basic forms of reasoning, inference can in general be defined as a process of accepting a statement or proposition (called the conclusion) on the basis of the (possibly provisional) acceptance of one or more other statements or propositions (called the premises). Construed thus, it includes deduction, induction, and abduction. It also includes entailment,2 presupposition, and implicature. The aim of this article is to assess entailment, presupposition, and implicature – the three fundamental types of inference that are recognised in semantics and pragmatics (see e.g. Huang 2007, 2009g on the scope of pragmatics). Of the three categories of inference, entailment is essentially semantic in nature, though the notion of pragmatic entailment will be discussed. Presupposition straddles the semanticspragmatics boundary, resulting in semantic and pragmatic presupposition. Finally, with regard to the two breeds of implicature, while conversational implicature is pragmatic in nature, conventional implicature can be categorised either as semantic or pragmatic, depending on how the distinction between semantics and pragmatics is drawn. Concepts parallel to conversational implicature such as explicature and impliciture will also be considered. Finally, a brief comparison among entailment, presupposition, conversational implicature, and conventional implicature in terms of four parameters: defeasibility, negation, conventionality and semantics versus pragmatics will be made.

1.

Entailment

The term ‘entailment’ is derived from formal logic.3 It refers to a semantic relation that can be defined in terms of truth in (1). (1) Entailment A proposition (or sentence expressing a proposition) p entails a proposition (or sentence expressing a proposition) q if and only if the truth of p guarantees the truth of q.

398

Yan Huang

What (1) basically says is this: if p entails q, then q can be taken as following logically and inescapably from p. If p is true, then q is also true. By contrast, if p is false, nothing is said about the truth value of q. Defined thus, entailment represents a truth-functional relationship in the sense that its function is to predict the truth value of a proposition from what is known of the truth value of another. By way of illustration, let us consider (2). (I use ‘|–’ to stand for ‘entail’.) (2) a. All of the university’s professors are hard-working. b. |– Some of the university’s professors are hard-working. In (2), if the university in question has professors, then (2a) entails (2b), but not vice versa. The reason is that if it is true that all of the university’s professors are hard-working, then there is no way to avoid the conclusion that it is also true that some of the university’s professors are hard-working. Notice that entailment proceeds from our knowledge of language. It depends on the constituents of relevant sentences rather than context. Consequently, an entailment is not defeasible, that is, it cannot evaporate in any linguistic or non-linguistic context. This is why it must be seen as semantic in nature (see e.g. Huang 2007: 16–17). However, on Fauconnier’s (1975) and Israel’s (2004) view, there is also pragmatic entailment. By pragmatic entailment is meant an inference that is on the one hand default, and on the other defeasible. Two examples are given in (3) and (4). (3) a. b. (4) a. b.

John can run 100m in 9.9 seconds. |– John can run a slower 100m. John can’t run 100m in 9.9 seconds. |– John can’t run a faster 100m.

According to Fauconnier and Israel, the opposing, scalar inferences of (3b) and (4b) are pragmatic entailments. They are entailments because the inferences are default, i.e. automatically valid; they are pragmatic because the inferences can be defeated. But how the inferences can be cancelled seems to me to be unclear. One type of pragmatic entailment of this sort is scalar entailment stemming from what Levinson (2000) called a Hirschberg-scale – a contextually given ad hoc scale. The relation between the ordered elements in such a scale does not need to be that of semantic entailment. The scale can be based on any partially ordered contrast sets in a contextually salient way. Examples include destinations ordered in terms of their distance, autographs ordered in terms of their authors’ famousness, and problems ordered in terms of their complexity (see e.g. Huang 2009d for further discussion). Semantic entailment plays a prominent role in semantics and especially in formal semantics. It is explored to define many of the other fundamental semantic relations. For example, two sentences may be said to have a propositional equivalence, to be paraphrases of each other, or to be synonymous if and only if they have exactly the same set of entailments, as in (5). Next, two sentences may be con-

Types of inference: entailment, presupposition, and implicature

399

sidered to be contradictories if each entails the negation of the other so that whenever one is true the other must be false, as in (6). In a similar vein, a single sentence may be said to be a contradiction if it contains contradictory entailments, as in (7). While contradiction is a semantic relation in which one member of the pair must be true and the other false, contrariety is a semantic relation in which only one term may be true, though both terms may be simultaneously false. Consequently, two sentences S1 and S2 may be taken to have contrary propositions if and only if S1 entails the negation of S2 and the negation of S2 does not entail S1. This is exemplified in (8). Furthermore, semantic entailment is also used to define notions like analyticity and ambiguity. (5) (Propositional equivalence) a. The little hut was hidden by the trees. b. The little hut was concealed by the trees. (6) (Contradictory) a. No one likes dark tourism. b. At least someone likes dark tourism. (7) (Contradiction) ?John isn’t married, but his wife is a feminist. (8) (Contrariety) a. This skirt is blue. b. This skirt is red. In formal semantics, semantic entailment is employed to advance a theory of generalised quantifiers. In this theory, which is developed out of the work of Ladusaw (1979), the concept of directional entailingness (also called monotonicity) is invoked. Directional entailingness provides an account of patterns of semantic entailment between sets and subsets. Upward entailing or entailment (also called monotone increasing, usually represented as MONF) involves entailment from a subset to a set. In other words, the direction of entailment is from more specific to less specific. By contrast, downward entailing or entailment (also called monotone decreasing, normally represented as MONG) is concerned with entailment from a set to a subset. Another way of saying it is that the direction of downward entailment is from less specific to more specific. Now, contrast (9) and (10). (9) a. b. (10) a. b.

Every woman is cooking. Every woman is roasting beef. No woman is cooking. No woman is roasting beef.

In (9), (9a) does not entail (9b), but (9b) does entail (9a). Consequently, the quantifier every triggers an upward entailment here.4 By contrast, in (10) the reverse is true. Therefore, the quantifier no involves a downward entailment. The distinction between upward and downward entailment is applied particularly to the study of

400

Yan Huang

negative polarity items (NPIs). It predicts correctly that NPIs like any and ever can be licensed in the scope of a downward entailment operator but not in that of an upward entailment one (see e.g. Saeed 2009). The contrast is shown in (11). (11) a. ?Every woman is ever roasting beef. b. No woman is ever roasting beef. But there has also been evidence that downward entailment is neither necessary nor sufficient for the licensing of NPIs (e.g. Giannakidou 1998, Horn 2009, see also Atlas 2007). More recently, Chierchia, Crain and their associates (e.g. Chierchia 2004, Crain and Pietroski 2002) have used the upward/downward entailment distinction to mount a challenge to what Horn (2006) dubbed a ‘Golden Age Pure Pragmatics’ or GAPP-style analysis of scalar implicatures. According to Chierchia, Crain, and their associates, while a standard upper bounding Q-scalar implicature, that is, an implicature from ‘ … p …’ to ‘ … at most p …’, does arise from positive Horn-scales, it is quite weak and even blocked in negative Horn-scales and other downward entailing environments. On the basis of this claim, Chierchia, Crain, and their associates argued that Q-scalar implicatures must be computed compositionally. Consequently, they fall under compositional semantics, hence part of grammar. But as pointed out by Levinson (2000) and Horn (2006, 2009), the alleged blockage of Q-scalar implicatures is due to the fact that a Horn-scale is reversed under negation and other downward entailing operators, as in (12a), and consequently, a different Q-scalar implicature is derived from the inverse scale, as in (12c) (see also Huang 2009b). (I use ‘+>’ to signify ‘conversationally implicate’.) If this is the case, then Chierchia and Crain’s argument may not be maintained (see also Sauerland 2004, Russell 2006, and Geurts 2009 for various (neo)-Gricean analyses of Q-scalar implicature in downward entailment contexts). (12) a. b. The earthquake didn’t kill many of the villagers. c. +> The earthquake killed some of the villagers. There are, however, serious problems at the very heart of the (semantic) entailment based approach to meaning. One such problem, which is well-known, is concerned with the failure of a semantic entailment in a referentially opaque context. Consider first (13). (13) a. Barack Obama is a popular politician. b. The first African-American President of the United States of America is a popular politician. At the time of writing (April, 2009), the proper name Barack Obama in (13a) and the definite description the first African-American President of the United States of America in (13b) have the same extension, i.e. they refer to the same person. Given

Types of inference: entailment, presupposition, and implicature

401

Leibniz’s Law, which is a law of intersubstitutability salva veritate (Latin for ‘with the truth unchanged’), it states that the substitution of expressions with the same extension does not affect the truth-conditions of the sentences. Consequently, if (13a) expresses a true proposition, then so does (13b). In other words, (13a) and (13b) are synonymous because they mutually entail each other. However, in a referentially opaque context, truth cannot be preserved when the co-referential expressions are substituted for each other. This is the case in (14). (14) a. I want to dine with Barack Obama. b. I want to dine with the first African-American President of the United States of America. Here, (14a) and (14b) have different truth conditions. This is because the expression the first African-American President of the United States of America is given an intensional reading, that is, the reading that the speaker wants to dine with whoever happens to be the first African-American President of the United States of American and does not care and may not have known who that person is. The same can be said of (15). (15) a. John believes that Barack Obama is a popular politician. b. John believes that the first African-American President of the United States of America is a popular politician. If John does not know or believe that Barack Obama is the first African-American President of the United States of America, then (15b) means that John has a belief about someone who is the first African-American President of the United States of America without knowing or caring about who the actual holder of the office is. Consequently, such a belief of John’s is said to be de dicto – a belief about ‘what is said’. The same is true of Frege’s classical example, usually referred to as the Morning Star Paradox (see e.g. Cann 1993 for discussion about further problems of the (semantic) entailment based analysis). Verbs that create a referentially opaque context are commonly referred to as intensional verbs or verbs of propositional attitude. These include believe, know, doubt, want, hope, fear, look, seem, and seek. Other words that establish a referential opaque context include adjectives like alleged, prepositions like about, connectives like because, and modal words of various categories like must, may, probably, obviously, and permissible.

2.

Presupposition

2.1.

Defining presupposition

Presupposition is defined as a proposition or inference whose truth is taken for granted in the utterance of a sentence.5 Its main function is to act as a precondition

402

Yan Huang

of some sort for the appropriate use of that sentence. This background assumption will remain in force when the sentence that contains it is negated. The German mathematician and logician Gottlob Frege is generally recognised as the first scholar in modern times who (re)introduced the philosophical study of presupposition (see especially Frege 1892), though the notion of presupposition may go back at least as far as the mediaeval philosopher Petrus Hispanus (see e.g. Huang 2007). Presupposition is usually engendered by the use of particular lexical items and/or linguistic constructions. Lexical items and linguistic constructions that give rise to presuppositions are called presupposition triggers. While presupposition has in general been regarded as ‘a heterogeneous collection of quite distinct and different phenomena’ (Levinson 1983: 217) since Karttunen (1973), a list of its representative examples may be given in (16) – (23) (see e.g. Levinson 1983, Atlas 2005, Huang 2007). (I use the symbol ‘>>’ to stand for ‘presuppose’. The positive and negative versions of the examples are separated by ‘/’, and the lexical presupposition triggers are italicised.) (16) Definite descriptions The king of France is/isn’t bald. >> There is a king of France. (17) Factive predicates a. Epistemic or cognitive factives. John knows/doesn’t know that smoking is a dangerous pastime. >> Smoking is a dangerous pastime. b. Emotive factives John regrets/doesn’t regret that he has gone on the stage. >> John has gone on the stage. (18) Aspectual/change of state predicates Mary has/hasn’t stopped beating her boyfriend. >> Mary has been beating her boyfriend. (19) Iteratives John returned/didn’t return to Berlin. >> John was in Berlin before. (20) Implicative predicates John managed/didn’t manage to give up binge drinking. >> John tried to give up binge drinking. (21) Temporal clauses After she left school, Jane worked/didn’t work as a secretary. >> Jane left school. (22) Cleft sentences a. Cleft It was/wasn’t the porter who called a taxi for John. >> Someone called a taxi for John.

Types of inference: entailment, presupposition, and implicature

403

b. Pseudo-cleft What the porter called/didn’t call was a taxi. >> The porter called something. (23) Counterfactual conditionals If an ant was as big as a human being, it could/couldn’t run five times faster than an Olympic sprinter. >> An ant is not as big as a human being. 2.2.

Properties of presupposition

Presuppositions exhibit two main properties: (i) constancy under negation, and (ii) defeasibility. In addition, certain cases of defeasibility give rise to what is known as the projection problem of presupposition. 2.2.1.

Constancy under negation

By constancy under negation is meant that a presupposition generated by the use of a lexical item or a syntactic structure remains the same when the sentence containing that lexical item or syntactic structure is negated (e.g. Strawson 1952). Using constancy under negation as diagnostic, presuppositions can be more formally defined in (24). (24) An utterance of a sentence S presupposes a proposition p if and only if a. if S is true, then p is true; b. if S is false, then p is (still) true. What (24) basically says is this: for S to presuppose p, whenever S is true, p is also true, and whenever S is false, p is still true. There are, however, problems with constancy under negation. On the one hand, constancy under negation may not be necessary. For example, there is a class of sentences which are hard, if not impossible to negate, yet they bear presuppositions, as in (25). On the other hand, constancy under negation may not be sufficient. This is illustrated by (26). Although (26) satisfies constancy under negation, the inference is standardly analysed as a felicity condition on performing the speech act of requesting (e.g. Levinson 1983). This is because if the window is open, there is no need for one to perform the speech act of requesting someone else, for example, to open it. (25) Long live the king of France! >> There is a king of France (26) Do/don’t open the window. ?>> The window is not open

404 2.2.2.

Yan Huang

Defeasibility

Presuppositions are deafeasible. They are cancelled if they are inconsistent with (i) background assumptions, (ii) conversational implicatures, and (iii) certain discourse contexts. Furthermore, they can also disappear in certain intrasentential contexts, some of which give rise to the projection problem of presupposition. Defeasibility has in general been taken as the second most important property of presupposition. In the first place, presuppositions can disappear in the face of inconsistency with background assumptions or real-world knowledge. Contrast (27) and (28) (I use ‘~ >>’ to stand for ‘does not presuppose’.) (27) John got an assistant professorship before he finished his Ph.D. >> John finished his Ph.D. (28) John died before he finished his Ph.D. ~ >> John finished his Ph.D. While (27) presupposes that John finished his Ph.D. by virtue of the temporal clause, (28) does not carry that presupposition. This is because the putative presupposition conflicts with our real-world knowledge. Consequently, the unwanted presupposition vanishes. Secondly, presuppositions can be cancelled by inconsistent conversational implicatures. By way of illustration, consider (29). (29) If Susan has danced on the table, her mother will be unhappy that she has done so. +> perhaps Susan has danced on the table, perhaps she hasn’t ~ >> Susan has danced on the table The use of the factive predicate unhappy in (29) should give rise to the potential presupposition that Susan has danced on the table. However, there is also a Q-clausal conversational implicature due to the use of the conditional in (29), namely, perhaps Susan has not danced on the table. In the face of the contradictory conversational implicature, the putative presupposition is defeated. Thirdly, presuppositions are contextually cancellable, that is, they can evaporate if they run contrary to what the immediate discourse context tells us. Witness first (30), where the putative presupposition is defeated by the inconsistent proposition that has already been established in the immediate discourse background. (30) There is no king of France. Therefore, the king of France isn’t bald. ~ >> There is a king of France In (30), the second sentence should presuppose that there is a king of France. The reason it does not is because such a putative presupposition is inconsistent with the preceding proposition that has already been established in the immediate discourse

Types of inference: entailment, presupposition, and implicature

405

background. As a result, the unwanted presupposition fails to survive. Next, presuppositions can be suspended by so-called reduction arguments – arguments that proceed by eliminating each of the possibilities – in a discourse. Finally, presuppositions can disappear in a discourse where evidence for their truth is being weighed and rejected. In the fourth place, presuppositions can be blocked in certain intrasentential contexts, that is, they can be defeated by using other clauses in the same complex sentence to increment the local, intrasentential context. Three sub-cases are of particular interest. The first is that a presupposition of a sentence can be overtly denied in a co-ordinate clause without any apparent contradiction. This is exemplified in (31). (31) John doesn’t regret snoring atrociously because in fact he never did so! ~ >> John snored atrociously Notice that in many cases, outright denial of presupposition is not possible with positive sentences. (32)*John regrets snoring atrociously because in fact he never did so! There is thus, at least in these cases, an asymmetry between negative and positive sentences with regard to defeasibility. This asymmetry has led to an entailment analysis for positive sentences, namely, the argument that what is allegedly presupposed in these sentences is actually what is entailed (e.g. Atlas 2005). Since semantic entailments cannot be overtly denied without producing semantic anomaly, the anomaly displayed in (32) is entirely expected. A second point to note is that the negation involved in overt denial of presupposition in (31) is generally taken as a metalinguistic negation (see e.g. Horn 2006). Next, a presupposition of a sentence can be explicitly suspended in an if clause that follows. Witness (33). (33) Mary doesn’t regret being a pet sitter, if she actually ever was one. ~ >> Mary was a pet sitter Finally, presuppositions can disappear with certain verbs of saying such as say, mention, tell, ask and announce, and certain verbs of propositional attitude such as believe, think, imagine, dream and want (see e.g. Huang 2007: 72–73). 2.2.3.

Theoretical issues and analyses

In general, there are three main theoretical issues in accounting for presupposition. The first and oldest one is concerned with presupposition failure. If a statement presupposes something which does not exist, what are the consequences for that statement? As already mentioned, a much quoted example of statements of this kind is the positive version of (16). Given that there is no king of France, the question boils down to this: what is the truth status of the sentence? Russell’s (1905)

406

Yan Huang

view was that since the sentence asserts both that there is a king of France and that he is bald, if the king of France does not exist, the sentence is false. This analysis was widely accepted and remained practically unchallenged for almost half a century. Then in 1950, Strawson published a seminal paper (Strawson 1950) in which he presented an alternative account. According to this analysis, Russell failed to distinguish sentences from the use of sentences. In using the sentence, a speaker does not assert but merely presupposes that there is a king of France. If there is no king of France, then there is a presupposition failure. Consequently, the sentence is neither true nor false: there is simply a truth-value gap. Strawson’s account has led to the notion of semantic presupposition. Secondly, there is the question of the origin of presupposition, namely, where presuppositions come from. The standard, though imperfect answer to this question is that presuppositions are engendered by certain lexical and structural triggers. Thirdly and finally, the most important issue is concerned with the projection problem of presupposition – a special case of the Fregean compositionality, namely, to state and explain the presuppositions of complex sentences (as ‘wholes’) in terms of the presuppositions of their component simple sentences (as ‘parts’). The projection problem manifests itself in two opposite directions. On the one hand, the presuppositions of a component sentence may fail to be projected onto, and hence inherited by, the whole complex sentence. This is the case for e.g. (31) and (33). In addition, certain complex or compound sentences formed with if … then and either … or also belong to this category. On the other hand, the presuppositions of a component sentence may be preserved when that constituent sentence becomes part of a more complex sentence. This is what happens when the sentence containing a presupposition trigger is under negation, as in (16), when it is embedded in modal operators such as it’s possible that, it’s likely that, ought to, may be and should, and when it is embedded in the conditional and the disjunction formed with if … then and either … or. Three formal semantic-pragmatic models are particularly influential in tackling the projection problem. The first is what I called the filtering-satisfaction (or Karttunen-Stalnaker) analysis in Huang (2007). The central idea of this approach is that a presupposition is not cancellable. Rather it has to be satisfied (i.e. entailed) in its local context, which is conceived mainly as a set of propositions. The local context is constructed in a dynamic way and developing online. This makes it possible for an unwanted presupposition to be filtered out during the derivation of a sentence in a bottom-up manner (see e.g. Karttunen 1973, Stalnaker 1974). Stalnaker’s work has resulted in the development of the concept of pragmatic presupposition. Next, in contrast to the filtering-satisfaction analysis, underlying Gazdar’s (1979) cancellation analysis is the crucial assumption that a presupposition is cancellable. On this theory, what a presupposition trigger engenders is merely a potential presupposition. A potential presupposition will become an actual one, unless it is defeated. With respect to the projection problem, what the cancellation analysis pre-

Types of inference: entailment, presupposition, and implicature

407

dicts is that each and every presupposition of the embedded clause will become an actual presupposition of the complex sentence, unless it is nullified by certain linguistic and non-linguistic factors. Finally, in an attempt to combine the strengths of both the filtering-satisfaction and cancellation theories, Heim (1983) developed a particular version of the filtering-satisfaction model, couched in her dynamic semantic theory of context change. In Huang (2007), I dubbed this model the accommodation analysis. Central to this approach is Heim’s belief that the meaning of an expression, including the presupposition of a sentence, is its context change potential. The context change potential of a sentence is the intersection of any context set in a common ground together with the proposition expressed by the sentence. By common ground is meant ‘a background of beliefs or assumptions which are shared by the speaker and his audience and which are recognised by them to be so shared’ (Stalnaker 1974). Accommodation, then, provides a mechanism to increment the discourse context set with new, non-controversial assumptions. Of these analyses, Karttunen’s and Gazdar’s are representatives of the twocomponent model view of presupposition, according to which, sentences including those containing presuppositions have two types of content: semantic content and presuppositional content. The latter is pragmatic in character. By contrast, Stalnaker and Heim take the one-component model view, operating within the general framework of a theory of pragmatic presupposition. On this account, presuppositions are treated as conditions on whether a sentence can be admitted into a context (see e.g. Portner 2005).

3.

Implicature

The concept of implicature (both conversational and conventional) has its origin in the work of H. P. Grice, though some proto-Gricean ideas can be traced back at least to the first century B.C. rhetorician Dionysius and the fourth century rhetoricians Servius and Donatus (e.g. de Jonge 2001). The ideas were later reiterated by the nineteenth century English philosophers John Stuart Mill and Augustus De Morgan (e.g. Horn 2006). 3.1.

Conversational implicature

A conversational implicature is any meaning implied or expressed by, and inferred or understood from, the utterance of a sentence which is meant without being part of what is strictly said6 (e.g. Huang 2009a, b, f). It is derived via Grice’s (1975, 1989) cooperative principle and its attendant maxims of conversation. For example, when one utters the sentence in (34a), he or she (ceteris paribus) conversationally implicates (34b).

408

Yan Huang

(34) a. Some of the tourists are admiring the view. b. +> Not many/most/all of the tourists are admiring the view Conversational implicatures are characterised by a number of distinctive properties (Grice 1975, 1989, Levinson 2000, Huang 2007: 32–35, 2009a, b). In the first place, there is the property of defeasibility – conversational implicatures can simply vanish in certain linguistic or non-linguistic contexts. How? They are cancelled if they are inconsistent with (i) semantic entailments, (ii) background or ontological assumptions, (iii) contexts, or (iv) priority conversational implicatures. As an illustrating example, consider (35) and (36). (The symbol ‘~ +>’ is used to stand for ‘do not conversationally implicate’.) (35) John and Mary bought a villa in Auckland. +> John and Mary bought a villa in Auckland together (36) The Americans and the Russians tested an atom bomb in 1962. ~ +> The Americans and the Russians tested an atom bomb in 1962 together The utterance in (36) has the potential mirror maxim, ‘togetherness’ conversational implicature, as indicated in (35). However, this potential conversational implicature runs contrary to our background or ontological assumptions. Given our knowledge about history, it was impossible for the USA and the USSR to test an atom bomb together in 1962. Consequently, the potential ‘togetherness’ conversational implicature is defeated by inconsistent real-world knowledge. Notice that defeasibility is a necessary but not a sufficient condition for conversational implicature (Horn 2007). A second property exhibited by conversational implicatures is non-detachability – any linguistic expression with the same semantic content tends to carry the same conversational implicature. (A principled exception is those conversational implicatures that arise via the maxim of Manner.) This is because conversational implicatures are attached to the semantic content, rather than the linguistic form, of what is said. Therefore, they cannot be detached from an utterance simply by replacing the relevant linguistic expressions with their synonyms. This is illustrated in (37), which indicates that the use of any linguistic expression that is synonymous with almost will give rise to the same conversational implicature. (37) Gordon almost/nearly lost /came close to losing his job. +> Gordon did not (quite) lose his job Thirdly, we have the property of calculability – conversational implicatures can transparently be derived via the cooperative principle and its component maxims. A fourth property is non-conventionality – conversational implicatures, though dependent on the saying of what is coded, are non-coded in nature. In other words, they rely on the saying of what is said but they are not part of what is said. A fifth

Types of inference: entailment, presupposition, and implicature

409

property is reinforceability – conversational implicatures can be made explicit without producing too much of a sense of redundancy. This is because conversational implicatures are not part of the conventional import of an utterance. For instance, the conversational implicature in (34) is made explicit in (38). But (38) is not judged to be semantically redundant. (38) Some, but not many/most/all, of the tourists are admiring the view. The sixth property of conversational implicatures is universality – conversational implicatures tend to be universal, because they are motivated rather than arbitrary. For example, if a language has ‘all’ and ‘some’, the use of the semantically weaker ‘some’ will universally carry the conversational implicature ‘not all’. In Huang (2007: 35), data are drawn from Modern Standard Arabic, Catalan, Chinese, Modern Greek, Kashmiri and Malagasy to illustrate this point. Finally, there is the property of indeterminacy – some conversational implicatures may be indeterminate. They can be taken as conveying an open-ended range of implicaures relating to matters in hand. Suppose one says (39). He or she may generate a range of indeterminate conversational implicatures. (39) Our new professor is a machine. +> Our new professor is cold, or/and +> Our new professor is efficient, or/and +> Our new professor is a workaholic, or/and +> … A conversational implicature can be engendered in two distinct ways. On the one hand, it can arise from strictly observing the maxims of conversation. In Huang (2007: 27), I called conversational implicatures thus induced conversational implicaturesO. This is the case for e.g. (34), (35) and (37) above. On the other hand, a conversational implicature can be generated by way of a speaker’s ostentatiously flouting the maxims. In Huang (2007: 29), I dubbed conversational implicatures thus engendered conversational implicaturesF. This is the case with the generation of the conversational implicature in (39), which deliberately exploits Grice’s maxim of Quality. There is thus the first Gricean dichotomy between conversational implicatureO and conversational implicatureF. Grice’s achievement here was to have provided a unified analysis of both types. A second Gricean dichotomy, independent of the first, is between those conversational implicatures which arise without requiring any particular contextual conditions and those which do require such conditions. Grice called the first kind generalised conversational implicatures (GCIs) and the second kind particularised conversational implicatures (PCIs). By way of illustration, consider the two conversational implicatures in Mary’s utterance in (40).

410

Yan Huang

(40) John: How did yesterday’s guest lecture go? Mary: Some of the faculty left before it ended. +> (a) Not many/most/all of the faculty left before the lecture ended +> (b) The lecture didn’t go well The conversational implicature in (40a) has a very general currency. Any utterance of the form ‘Some x are Y’ will have the default interpretation ‘Not many/most/all x are Y’. This interpretation will go through without needing any particular context, hence (40a) is a GCI. By contrast, the conversational implicature in (40b) depends crucially on context of some kind. Mary’s reply points to a possible connection, namely, if some of the faculty left a lecture before it ended, the lecture may not have gone well. Without such a specific connection, we will not have the relevant conversational implicature, thus (40b) is a PCI. The theoretical importance of this Gricean dichotomy has recently been subject to heated debates. Hirschberg (1991), Welker (1994) and Carston (2002), for example, doubted whether such a distinction can be maintained. On the other hand, Levinson (2000) put forward a rigorous defence of it. Since its inception, Grice’s classical theory of conversational implicature has revolutionised pragmatic theorising, nourishing numerous neo- and post-Gricean variants and reformulations. Horn developed a bipartite model calling it Manichaean pragmatics. In this account, there are two genera of conversational implicatures: namely Q[uantity]- and R[elation]-implicatures (see e.g. Horn 2009). Arguing for a clear separation between pragmatic principles governing an utterance’s surface form and pragmatic principles governing its informational content, Levinson (2000) proposed that the original Gricean program (the maxim of Quality apart) be reduced to a tripartite model with three genera of conversational implicatures: what he dubbed the Q-, I[nformativeness]- and M[anner] implicatures. Q-implicatures can then be divided into three species: what I termed (i) Q-scalar , (ii) Q-clausal and (iii) Q-alternate implicatures in Huang (2007: 42–44). Furthermore, within Q-alternate implicatures, we have two subtypes: Q-ordered alternate and Q-unordered alternate implicatures such as . Notice that Q-ordered alternate implicatures can arise either from a non-entailment scale which is given by the lexicon without requiring any specific context (e.g. ) or from what Levinson (2000) called a Hirschberg-scale or what Horn (2007, 2009) dubbed a rank order, following in part work by Lehrer (1974). As already mentioned above, a Hirschberg-scale is essentially a nonce scale, that is, a contextually given ad hoc scale. Such a scale can be based on any partially ordered contrast sets in a contextually salient way. Examples include , , and (see e.g. Huang 2009d for further discussion). On the other hand, in relevance theory (Sperber and Wilson 1995), the majority of classical and neo-Gricean GCIs is refashioned as an explicature – a proposition that is an infer-

Types of inference: entailment, presupposition, and implicature

411

ential development of one of the linguistically given incomplete conceptual representations or logical forms of the sentence uttered (but see e.g. Burton-Roberts 2007 for the comment that ‘development’ in the definition is not defined in relevance theory). Explicature corresponds roughly to the intuitive notion of what is said, though the Gricean notion of what is said is abandoned in relevance theory. Conversational implicatures in the relevance-theoretic sense, called r-implicatures in Huang (2007: 195), are largely PCIs in the classical and neo-Gricean sense. In a similar vein, Recanati (2004), a radical contextualist in the current divide between contextualism and (semantic) minimalism in the philosophy of language, reduced some cases of conversational implicature to what he called the ‘pragmatically enriched said’. Also, Bach (2004) argued that certain aspects of speaker meaning are neither part of what is said nor of what is conversationally implicated. Consequently, he proposed a third category of communicative content, intermediate between what is said and what is conversationally implicated, and dubbed the vehicle of such a content ‘conversational impliciture’, because it is what is implicit in what is said. In other words, on Bach’s view, some cases of conversational implicature in the classical and neo-Gricean sense are in fact (conversational) implicitures (see also Bach 2006, Huang 2009e and Garret and Harnish 2007). At this point, it is useful to discuss some of the current debates on conversational implicature. In the first place, the treatment of cardinals as scalar expressions engendering Q-scalar implicatures is highly controversial (see e.g. Carston 2002, Bultinick 2005, Hurewitz et al. 2006, Horn 2009 for arguments against, and Levinson 2000 for arguments for the scalar analysis). Secondly, there is the issue of the epistemic strength of Q-scalar implicatures. It is concerned with the question of what it is a speaker Q-scalar implicates against. Two neo-Gricean pragmatic positions can be identified here: the weak epistemic one represented by e.g. Hirschberg (1991), Sauerland (2004), Geurts (2009), and Horn (2009), and the strong epistemic one advocated by e.g. Gazdar (1979) and Levinson (2000: 77–79). In the third place, a heated debate has been going on for the last two decades or so, focusing on the nature of Gricean and neo-Gricean GCIs in general and Q-scalar implicatures in particular. One view is that GCIs in general and Q-scalar implicatures in particular convey default meanings, sans a conscious inferential process and irrespective of a particular context (e.g. see Levinson 2000 for strong defaultism and Horn 2009 for weak defaultism). Another, relevance-theoretic position is that they are essentially inferred contextually (e.g. Sperber and Wilson 1995). In addition, a more recent, third view holds that the derivation of Q-scalar implicatures relies heavily on structural factors (cf. Section 2 above and Chierchia 2004). Furthermore, all the three views have recently been subject to studies in experimental pragmatics. While much of the relevance-theoretically oriented experimental work favours the contextual inference approach (e.g. Noveck and Sperber 2007, Noveck and Reboul 2008), there is also evidence in support of the default inference theory (e.g. Grodner et al. 2007) and the structural inference view (e.g. Panizza and Chier-

412

Yan Huang

chia 2008). Next, in recent years, there has also been an intense debate on (i) whether or not pragmatically enriched or inferred content can enter or ‘intrude’ upon the conventional, truth-conditional content of what is said, and (ii) if the answer to (i) is positive, then what the pragmatic intrusion under consideration is. Concerning the first question, a dividing line can generally be drawn between pragmaticists like Levinson (2000), Recanati (2004) and relevance theorists (Sperber and Wilson 1995, Carston 2002), who argue for pragmatic intrusion (but see e.g. Horn 2009 for reservations), and semanticists like Cappelen and Lepore (2005) and King and Stanley (2005), who argue against it. Next, regarding the second question, two current positions can roughly be identified. The first is that pragmatic intrusion is of a special kind, which differs from conversational implicature. Within this camp, three lines of argument are of particular interest. According to Sperber and Wilson (1995), the pragmatic content is an explicature. Secondly, there is the position taken by Recanati (2004) that it is a pragmatically enriched part of what is said. A third argument is due to Bach (2004), who took the view that the pragmatic inference under consideration is his impliciture. On the other hand, the second position is represented by Levinson (2000). Within the neo-Gricean framework, Levinson argued that these so-called explicatures/pragmatically enriched said/implicitures are in fact the same beast as conversational implicatures. See also my neo-Gricean and revised neo-Gricean analyses of anaphora in e.g. Huang (1991, 2000a, b, 2004, 2007 and 2009b), which in effect argue that the presemantic, pragmatic inference involved in the interpretation of anaphora is a conversational implicature. Finally, it is worth mentioning that there have been various attempts to integrate the classical and neo-Gricean pragmatic theories of conversational implicature with other current linguistic theories. These linguistic theories include decision theory (Merin 1999), bidirectional optimality theory (Blutner 2004) and game theory (Benz et al. 2006). 3.2.

Conventional implicature

We turn next to the second category of implicature postulated by Grice, namely, conventional implicature. (In fact, Frege’s 1892, 1918–1919 analysis of the Andeutung relation is a direct precursor of Grice’s concept of conventional implicature. See e.g. Bach 1999, Feng 2006, Horn 2007). An Andeutung or conventional implicature is a non-truth-conditional meaning which is not derivable from general considerations of cooperation and rationality, but arises solely from the conventional features attached to particular lexical items and/or linguistic constructions involved. A few standard examples are given in (41)-(43). (I use ‘+>>’ to stand for ‘conventionally implicate’.)

Types of inference: entailment, presupposition, and implicature

413

(41) p therefore q +>> q follows from p Lihua is a Chinese; she, therefore, knows how to use chopsticks. (42) p but q +>> p contrasts with q John is poor but he is honest. (43) Even p +>> contrary to expectation Even a child can scoff all the small bananas. In (41), the conventional implicature triggered by the use of therefore is that being a Chinese provides some good reason for knowing how to use chopsticks. In (42), there is a conventional implicature of contrast between the proposition contained in p and that contained in q. Finally in (43), even, being epistemic in nature, conventionally implicates some sort of unexpectedness, surprise or unlikeliness (Farncescotti 1995). Other representative lexical items that are considered to engender conventional implicatures in English include actually, also, anyway, barely, besides, however, manage to, moreover, on the other hand, only, still, so, though, too and yet. What, then, are the essential properties of conventional implicature? From a traditional point of view, conventional implicatures are considered to have the following properties (Grice 1989, Levinson 1983, Huang 2007, 2009c). Firstly, conventional implicatures are not derived from Grice’s cooperative principle and its component maxims, but are attached by convention to particular lexical items and/or linguistic constructions. They are therefore an arbitrary part of meaning, and must be learned ad hoc. Secondly, conventional implicatures are not calculable via any natural procedure, but are rather given by convention, thus they must be stipulated. Thirdly, conventional implicatures are not defeasible, that is, they cannot be cancelled. Fourthly, conventional implicatures are detachable, because they depend on the particular lexical expressions and/or linguistic constructions used. Finally, conventional implicatures do not tend to be universal. It should be pointed out that unlike the concept of conversational implicature, the notion of conventional implicature is not taken to be very coherent. Even Grice himself (1989: 46) warned that ‘the nature of conventional implicature needs to be examined before any free use of it, for explanatory purposes, can be indulged in’. Horn (2004: 6) has gone a step further by claiming that ‘the role played by conventional implicature within the general theory of meaning is increasingly shaky’. Since its inception, conventional implicature has been subject to numerous attempts to reduce it to semantic entailment, conversational implicature, and presupposition (Levinson 1983), and more recently, to part of what is said (Bach 1999), part of tacit performatives (Rieber 1997), vehicles for performing second-order speech acts (Bach 1999), and procedural meaning in relevance theory (Blakemore 2004). But recently, Potts (2005) has made a brave attempt to resurrect the concept of conventional implicature. He ‘retain[ed] Grice’s brand name but alter[ed] the product’ (Horn 2007) by focusing on expressives like epithets, attributive adjectives

414

Yan Huang

and honorifics and supplements like non-restrictive relatives, parentheticals and appositives rather than lexical items such as but, therefore and even. He isolated four essential properties of conventional implicature. The first of these properties is conventionality – conventional implicatures are part of the conventional meaning of the expressions involved. The second property is commitment – conventional implicatures are commitments, and thus engenders entailments. The third property is speaker orientation – the commitments are made by the speaker of an utterance. The final property is independence – conventional implicatures are logically and compositionally independent of what is said (see also von Heusinger and Turner 2006). Taking the view that conventional implicature is semantic in nature, Potts developed a logic of the notion by modelling it with a type-driven multi-dimensional semantic translation language (see Feng 2006, Horn 2007 for criticisms of this analysis). Feng (2006) presented another development of Grice’s notion of conventional implicature. The properties extracted by him for conventional implicature are (i) non-truth-conditionality, (ii) speaker orientation, (iii) infallibility, (iv) occurrency, (v) dependency, and (vi) context-sensitivity. He further argued that properties (i) – (iv) are intimately associated with subjectivity. Finally, contrary to Potts’s view, Feng maintained that conventional implicature has both a semantic and pragmatic character. This is why a conventional implicature is so named by Grice. ‘[I]t is so named because it involves both linguistic and contextual information. It is conventional because it is associated with the conventional linguistic meaning of a certain expression … It is implicated rather than said because its full content requires contextual information, and does not affect the truth conditions of the utterance’ (Feng 2006: 184). This view is also echoed in Horn (2007: 50), who said that ‘[conventional implicature] is semantic insofar as it involves an aspect of the conventional meaning of a given expression rather than being computable from general principles of rational behavior or communicative competence, but it is pragmatic insofar as it involves considerations of appropriateness rather than truth of the sentence in which it appears.’ Whether belonging to semantics or balancing on the edge between semantics and pragmatics, Potts’s, Feng’s and Horn’s recent works have shown that the Fregeo-Gricean concept of conventional implicature is, after all, not that incoherent.

4.

A brief comparison

In this final section, I shall make a brief comparison among semantic entailment, presupposition, conversational implicature and conventional implicature. I shall compare them in terms of four parameters: (i) defeasiblility, (ii) negation, (iii) conventionality, and (iv) semantics versus pragmatics. First, defeasibility. While semantic entailment and conventional implicature are not defeasible, most cases of presupposition and all cases of conversational im-

Types of inference: entailment, presupposition, and implicature

415

plicature are defeasible. In the study of inference, a hierarchy is set up to rank various types of inference as defeasible, nonmonotonic, default, probabilistic, and plausibilistic, with defeasible inference at the top. An inference is defeasible when it has the possibility of error. Looked at in this way, nonmontonic inference is a major type of defeasible inference. It is currently reasonable but its reasonability can be cancelled upon the admittance of new information into common ground. Default inference is then a particular kind of nonmonotonic inference. It is characterised by what computer scientists and artificial intelligence (AI) workers call ‘negation as failure’. An example might be that from the ‘failure’ of the airport departures board to list a late-night flight to London Heathrow, one may infer that such a flight has been cancelled. In this sense, both presupposition and conversational implicature are a case of presumptive reasoning like default inference. Probabilistic inference is stronger than plausibilistic inference (see e.g. Woods 2009). Next, we turn to negation. While presupposition can survive negation, entailment cannot. As for conversational and conventional implicatures, their behaviour under negation is rather complex and needs to be further studied. Finally, let us consider the question of conventionality and the related question of where in the linguistic terrain the four types of inference are situated. Clearly, semantic entailment is the most conventional and falls in the province of semantics. Presupposition is less conventional than conventional implicature but more conventional than conversational implicature. Therefore, it straddles the boundary between semantics and pragmatics, but more on the side of pragmatics. Of the three types of implicature, although GCIs are more conventional than PCIs, both types are the pursuit of pragmatics. Finally, conventional implicature, which is more conventional than both GCIs and PCIs, can be considered either as semantic or pragmatic, depending on how the boundary between semantics and pragmatics is placed. If semantics is taken to be concerned with those aspects of meaning that affect truth conditions, then the investigation of conventional implicature falls on the pragmatic side of the divide rather than on the semantic side, since, as noted above, it does not make any contribution to truth conditions. On the other hand, if pragmatics is conceived of as dealing with those inferences that are non-conventional, hence cancellable, then conventional implicature falls within the province of semantics but outside that of pragmatics, since it cannot be defeated.

Notes 1. I wish to thank the anonymous reviewers and Wolfram Bublitz and Neal Norrick, the two editors of this volume for their stimulating comments on an earlier version of this chapter. The work reported on here was partially supported by a grant from the University of Auckland Faculty Research Development Funds, which is hereby gratefully acknowledged.

416

Yan Huang

2. On Saeed’s (2009) view, entailment is not an inference. However, Cummings (2005: 83–84) argued convincingly that such a view is too narrow. 3. The phrase ‘p entails q’ appeared first in Moore (1922). The term ‘entailment’, according to the second edition of the Oxford English Dictionary, was first used by L. S. Stebbing in 1933, who wrote: ‘Professor Moore’s analysis of the distinction between material implication and entailment makes it possible’. 4. Notice that whereas every is upward entailing with respect to its right, VP argument, as in (9), it is downward entailing in its left, NP argument (e.g. Saeed 2009). For example, Every woman is cooking unilaterally entails Every old woman is cooking. In terms of monotone increasing/decreasing, the former is called right monotone increasing, and the latter is labelled left monotone decreasing. 5 The term ‘presupposition’ seems to be introduced into English by Strawson (1952). 6. Note that Saul (2002) was of the view that Grice’s main goal is to develop a theory of speaker-meaning. Following Saul and biting the bullet, Horn (2009) now holds that conversational implicature is a component of speaker-meaning rather than a pragmatic inference. By contrast, Levinson (2000), Atlas (2005) and many others are still treating conversational implicature as a pragmatic inference. My definition is applicable to both sides.

References Atlas, Jay D. 2005 Logic, Meaning, and Conversation. Oxford: Oxford University Press. Atlas, Jay D. 2007 On a pragmatic explanation of negative polarity licensing. In: Noel BurtonRoberts (ed.), 10–23. Bach, Kent 1999 The myth of conventional implicature. Linguistics and Philosophy, 22: 327–66. Bach, Kent 2004 Pragmatics and the philosophy of language. In: Laurence R. Horn and Gregory Ward (eds.), 463–87. Bach, Kent 2006 The top ten misconceptions about implicature. In: Betty. J. Birner and Gregory Ward (eds.), Drawing the Boundaries of Meaning, 21–30. Amsterdam: Benjamins. Benz, Anton, Gerhard Jäger and Robert van Rooy 2006 Game Theory and Pragmatics. London: Palgrave Macmillan. Blakemore, Diane 2004 Discourse markers. In: Laurence R. Horn and Gregory Ward (eds.), 221–240. Blutner, Reinhard and Henk Zeevat (eds.) 2004 Optimality Theory and Pragmatics. London: Palgrave Macmillan. Bultinck, Bert 2005 Numerous Meaning: The Meaning of English Cardinals and the Legacy of Paul Grice. Oxford: Elsevier.

Types of inference: entailment, presupposition, and implicature

417

Burton-Roberts, Noel 2007 Varieties of semantics and encoding: Negation, narrowing/loosening and numericals. In: Noel Burton-Roberts (ed.), 90–114. Burton-Roberts, Noel (ed.) 2007 Pragmatics. London: Palgrave Macmillan. Cann, Ronnie 1993 Formal Semantics. Cambridge: Cambridge University Press. Cappelen, Herman and Ernie Lepore 2005 Insensitive Semantics. Oxford: Blackwell. Carston, Robyn 2002 Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell. Chierchia, Gennaro 2004 Scalar implicatures, polarity phenomena, and the syntax/pragmatics interface. In: Adriana Belletti (ed.), Structures and Beyond, 39–103. Oxford: Oxford University Press. Crain, Stephen and Paul Pietroski 2002 Why language acquisition is a snap? The Linguistic Review 19: 163–183. Cummings, Louise 2005 Pragmatics. Edinburgh: Edinburgh University Press. Cummings, Louise (ed.) 2009 The Pragmatics Encyclopaedia. London: Routledge. de Jonge, Casper C. 2001 Natura artis magistra: ancient rhetoricians, grammarians, and philosophers on natural word order. In: Ton van der Wouden and Hans Broekhuis (eds.), Linguistics in the Netherlands 2001, 159–166. Amsterdam: Benjamins. Farncescotti, Robert Michael 1995 “Even”: the conventional implicature approach reconsidered. Linguistics and Philosophy 18: 153–73. Fauconnier, Gilles 1975 Pragmatic scales and logical structure. Linguistic Inquiry 6: 353–375. Feng, Guangwu 2006 A theory of conventional implicature and pragmatic markers in Chinese. Ph.D. dissertation. University of Reading. Frege, Gottlob 1892 Über Sinn und Bedeutung. Zeitschrift für Philosophie und Philosophisch Kritik 100: 25–50. [Trans. On sense and reference. In: Peter Geach and Max Black (eds.), 1952. Translations from the Philosophical Writings of Gottlob Frege. Oxford: Blackwell.] Frege, Gottlob 1918–1919 The Thought: A Logical Inquiry. Translation in Mind 55: 289–311 [(1956)]. Garret, Merrill and Robert M. Harnish 2007 Experimental pragmatics: Testing for implicature. Pragmatics and Cognition 15: 65–90. Gazdar, Gerald 1979 Pragmatics: Implicature, Presupposition and Logical Form. London: Academic Press.

418

Yan Huang

Geurts, Bart 2009 Scalar implicature and local pragmatics. Mind and Language 24: 51–79. Giannakidou, Anastasia 1998 Polarity Sensitivity as (Non-)Veridical Dependency. Amsterdam: Benjamins. Grice, H. Paul 1975 Logic and conversation. In: Peter Cole and Jerrold Morgan (eds.), Syntax and Semantics 3: Speech Acts, 41–58. London: Academic Press. Grice, H. Paul 1989 Studies in the Way of Words, Cambridge, MA: Harvard University Press. Grodner, Daniel, Natalie Klein, Kathleen M. Carbary and Michael Tanenhaus 2007 Experimental evidence for rapid interpretation of pragmatic some. Paper presented at XPRAG, Berlin. Heim, Irene 1983 On the projection problem for presuppositions. Proceedings of WCCFL 2: 114–25. Heusinger, Klaus von and Ken Turner 2006 (By way of an) introduction: a first dialogue on the semantics-pragmatics interface. In: Klaus von Heusinger and Ken Turner (eds.), Where Semantics Meets Pragmatics, 1–18. Oxford: Elsevier. Hirschberg, Julia 1991 A Theory of Scalar Implicature. New York: Garland. Horn, Laurence R. 2004 Implicature. In: Laurence R. Horn and Gregory Ward (eds.), 3–28. Horn, Laurence R. 2006 The border wars: a neo-Gricean perspective. In: Klaus von Heusinger and KenTurner (eds.), Where Semantics Meets Pragmatics, 21–48. Oxford: Elsevier. Horn, Laurence R. 2007 Toward a Fregean pragmatics: Voraussetzung, Nebengedanke, Andeutung. In: Istvan Kecskes and Laurence R. Horn (eds.), Explorations in Pragmatics: Linguistic, Cognitive and Intercultural Aspects, 39–69. Berlin: Mouton de Gruyter. Horn, Laurence R. 2009 WJ-40: Implicature, truth, and meaning. International Review of Pragmatics 1: 3–34. Horn, Laurence R. and Gregory Ward (eds.) 2004 The Handbook of Pragmatics. Oxford: Blackwell. Huang, Yan 1991 A neo-Gricean pragmatic theory of anaphora. Journal of Linguistics 27: 301–335. Huang, Yan 1994/2007The Syntax and Pragmatics of Anaphora. Cambridge: Cambridge University Press. Huang, Yan 2000a Anaphora: A Cross-linguistic Study. Oxford: Oxford University Press. Huang, Yan 2000b Discourse anaphora: Four theoretical models. Journal of Pragmatics 32: 151–176.

Types of inference: entailment, presupposition, and implicature

419

Huang, Yan 2003 On neo-Gricean pragmatics. International Journal of Pragmatics 13: 87– 110. Huang, Yan 2004 Anaphora and the pragmatics-syntax interface. In: Laurence R. Horn and Gregory Ward (eds.), 288–314. Huang, Yan 2007 Pragmatics. Oxford: Oxford University Press. Huang, Yan 2009a Neo-Gricean pragmatics and the lexicon. International Review of Pragmatics 1: 118–153. Huang, Yan 2009b Neo-Gricean pragmatic theory of conversational implicature. In: Bernd Heine and Heiko Narrog (eds.), The Oxford Handbook of Linguistic Analysis, 669–696. Oxford: Oxford University Press. Huang, Yan 2009c Implicature. In: Louise Cummings (ed.), 234–38. Huang, Yan 2009d Scalar implicature. In: Louise Cummings (ed.), 441–444. Huang, Yan 2009e Impliciture. In: Louise Cummings (ed.), 238–40. Huang, Yan 2009f What is said. In: Louise Cummings (ed.), 520–522. Huang, Yan 2009g Pragmatics. In: Louise Cummings (ed.), 370–375. Hurewitz, Felicia, Anna Papafragou, Lila Gleitman and Roger Gelman 2006 Asymmetries in the acquisition of numbers and quantifiers. Language Learning and Development 2: 77–96. Israel, Michael 2004 The pragmatics of polarity. In: Laurence R. Horn and Gregory Ward (eds.), 701–723. Karttunen, Lauri 1973 Presuppositions of compound sentences. Linguistic Inquiry 4: 169–193. King, Jeffrey and Jason Stanley 2005 Semantics, pragmatics, and the role of semantic content. In: Zoltan G. Szabó (ed.), Semantics vs. Pragmatics, 111–164. Oxford: Clarendon Press. Ladusaw, William A. 1979 Polarity sensitivity as inherent scope relations. Ph.D. dissertation. University of Texas at Austin. Lehrer, Adrienne 1974 Semantic Fields and Lexical Structure. Amsterdam: Benjamins. Levinson, Stephen C. 1983 Pragmatics. Cambridge: Cambridge University Press. Levinson, Stephen C. 2000 Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: The MIT Press.

420

Yan Huang

Merin, Arthur 1999 Information, relevance, and social decision making: Some principles and results of decision-theoretic semantics. In: Lawrence S. Moss, Jonathan Ginzburg and Maarten de Rijke (eds.), Logic, Language, and Computation, volume 2, 179–221. Stanford: CSLI. Moore, George Edward 1922 Philosophical Studies. London: Routledge and Kegan Paul. Noveck, Ira and Anne Reboul 2008 Experimental pragmatics: A Gricean turn in the study of language. Trends in Cognitive Sciences 12: 425–31. Noveck, Ira and Dan Sperber 2007 The why and how of experimental pragmatics: The case of ‘scalar inference’. In: Noel Burton-Roberts (ed.), 184–212. Panizza, Daniele and Gennaro Chierchia 2008 Two experiments on the interpretation of numerals. Paper presented at Workshop on Experimental Pragmatics/Semantics, the 30th Annual Convention of the German Society of Linguistics. Portner, Paul H. 2005 What is Meaning? Oxford: Blackwell. Potts, Christopher 2005 The Logic of Conventional Implicatures. Oxford: Oxford University Press. Recanati, François 2004 Literal Meaning. Cambridge: Cambridge University Press. Rieber, Stephen 1997 Conventional implicatures as tacit performatives. Linguistics and Philosophy 20: 51–72. Russell, Benjamin 2006 Against grammatical computation of scalar implicatures. Journal of Semantics 23: 361–82. Russell, Bertrand 1905 On denoting. Mind 14: 479–93. Saeed, John I. 2009 Semantics. 3rd edition. Oxford: Blackwell. Sauerland, Uli 2004 Scalar implicatures in complex sentences. Linguistics and Philosophy 27: 367–391. Saul, Jennifer 2002 What is said and psychological reality: Grice’s project and relevance theorists’ criticisms. Linguistics and Philosophy 25: 347–372. Sperber, Dan and Deidre Wilson 1995 Relevance: Communication and Cognition. Oxford: Blackwell. Stalnaker, Robert C. 1974 Pragmatic presupposition. In: Milton Munitz and Peter Unger (eds.), Semantics and Philosophy. New York: New York University Press, 197–214. Strawson, Peter F. 1950 On referring. Mind 59: 320–344.

Types of inference: entailment, presupposition, and implicature

421

Strawson, Peter F. 1952 Introduction to Logical Theory. London: Methuen. Welker, Katherine 1994 Plans in the common ground: Toward a generative account of conversational implicature. Ph.D. dissertation. Ohio State University. Woods, John 2009 Inference. In: Louise Cummings (ed.), 218–220.

Part IV The place of pragmatics in the description of discourse

15.

Pragmatics and grammar Arnulf Deppermann

1.

Introduction

In this article, “grammar” is used as an overarching term, encompassing morphosyntax as well as the syntax of the sentential structures (phrase structure and topological structure). Grammar thus includes syntactic categories, functions, and rules, grammatical constructions, and their realization in terms of forms and patterns of linearization. As grammatical constructions might be more or less lexically fixed and more or less restricted to, e.g., semantic classes of verbs, the boundary between lexicon and grammar is at best fuzzy, and may be better conceived of as a continuum (Croft 2008). “Pragmatics” here is also used in a wide sense, referring to the use of language in context. In this way, it encompasses cognitive preconditions and functions of language, such as attention, memory, conceptualization, categorization, and inference, as well as interactional uses, such as transfer of knowledge, turn-taking, speech acts, and rhetorical functions, displays of emotion and epistemic stance, and indexing social identities and relationships. In order to set the scene, here are some basic questions about the relationship between grammar and pragmatics, which will be taken up in this article: a) The role of pragmatics for grammar – Which pragmatic factors have an impact on grammar? Candidate factors include cognitive, interactional, situational, cultural, and bodily factors. – Which role do pragmatic factors play in defining grammatical structures? Can or should grammatical structure be defined on a pure formal basis or does it ultimately rest on functional categories? – How is pragmatics involved in parsing and interpreting grammatical constructions? – Which grammatical phenomena are governed by pragmatic factors both in terms of their diachronic emergence and regarding their situated use? – How does usage shape grammar over time? – Which role do pragmatic uses and contexts play in the acquisition of knowledge about grammar? b) The role of grammar for pragmatics – How are grammatical structures adapted to pragmatic preferences and functions? – How are they adapted to their interactive and situational context? – How is grammatical structure used as a flexible tool for speakers’ pragmatic concerns?

426

Arnulf Deppermann

– Do grammatical structures as such have pragmatic meaning? – How are grammatical constructions coordinated with other orders of embodied action? When asking questions about the relationship between grammar and pragmatics, we are interested in a subset of issues of the more general question of how language form relates to language function (see Newmeyer 1998).

2.

Theoretical positions: Formal vs. functional approaches to grammar

No linguist would deny that the choice and at least some aspects of the interpretation of grammatical structures are determined by pragmatic considerations. However, it is a disputable matter how deeply grammar and pragmatics are intertwined, and consequently, if grammatical analysis needs to attend to pragmatic factors or if it should rather treat grammar as a self-contained module which is working and needs to be analyzed by its own self-sufficient principles. The latter would be the Chomskian view, coined the “autonomy of syntax”-thesis (Chomsky 1965). According to this view, pragmatic factors neither explain the basic parameters and principles of grammar (e.g., universal organizing principles such as recursivity and head-specifier-structures), nor do grammatical features arise from pragmatic functions and pressures: “In general, it is not the case that language is readily usable or ‘designed for use’” (Chomsky and Lasnik 1993: 509). Grammar (which, in this view, is language proper) is considered as a formal, algebraic apparatus for generating possible, well-formed linguistic structures and for deriving hypotheses about linguistic structure from perceived speech. While this generativist stance is sometimes dubbed as a “formalist approach”, functionalists of various kinds (see below) assign pragmatics an important role in explaining grammatical structure. Basically, they view language as a means for the communication of thoughts and for social interaction. They hold that pragmatic functions (such as information-transfer, turntaking, etc.), principles (such as iconicity, categorical organization around prototypes, etc.), and pressures (having to do with processing time, attention span, etc.) motivate, constrain, and sometimes even fully explain the becoming, use, and interpretation of grammatical structures. There is, however, a broad variation between different functionalist approaches in terms of which grammatical phenomena and which pragmatic factors they focus on and how they conceive of the impact pragmatics is said to have on grammar (see below and Newmeyer 1998). 2.1.

Pragmatic approaches to grammar

In what follows, we will shortly review approaches which see grammar rooted in pragmatic motivations.

Pragmatics and grammar

2.1.1.

427

Functional Grammar and Functional Discourse Grammar

The Dutch theory of Functional grammar (FG; Dik 1997) starts from the assumption that language is structured according to its function as a means of communication of knowledge. It assigns pragmatics the hierarchically dominant role in determining the semantics and the grammar of clauses. In essence, however, pragmatics in this approach boils down to the information status of speaker and addressee in the ongoing discourse, which is said to determine the information structure of the clause in terms of choosing referring expressions, presuppositions, and organizing focus vs. background. Expression rules determine how pragmatic and semantic structures are mapped onto word order and morphosyntax. FG is a nonderivative, cognitivist theory, which requires typological, psychological, and pragmatic adequacy as standards for modelling linguistic structures in formalisms, which provide for testable predictions (Dik 1997). Functional Discourse Grammar (FDG) (Hengeveld and Mackenzie 2008) revises FG so as to become a discourse grammar, centering around the discourse act, which may be larger or smaller than a sentence. Discourse acts are dominated by speakers’ intentions (interpersonal level). In a top-down fashion, they govern the representational (semantics), the structural (morphosyntax) and the phonological level, which are each formally characterized in strict separation. In contrast to the other approaches discussed here, (D)FG formalizes its assumptions in a predicate calculus. 2.1.2.

Systemic-Functional Grammar

Since the 1960s, Systemic-Functional Grammar (SFG) has developed out of the Firthian British functional linguistics. Its leading figure, M.A.K. Halliday (1985), assumes that grammar is a system for the creation of meaning, which is structured according to three metafunctions: the interpersonal (enacting relationships between speaker and addressee), the ideational (construing experience), and the textual (organizing interpersonal and ideational meaning in the unfolding text or interaction). Grammars are taken to consist of systems of options, which are available for realizing the metafunctions. Examples are the systems ‘mood’ for realizing the interpersonal metafunction, ‘transitivity’ (referring to types of processes and the participants involved) for realizing the ideational metafunction, and ‘theme’ for realizing the textual metafunction. Systems are specific for ranks in the constituency structure (clause, phrase, word, morpheme). An important place within SFG is assigned to the study of registers, i.e., varieties of language which are tied to certain types of situation (e.g., different degrees of formality, languages for special purposes). SFG assumes that the context of the situation is to be categorized according to three domains, which roughly correspond to the metafunctions: tenor (social/interpersonal relationships), field (processes talked about), and

428

Arnulf Deppermann

mode (medium/channel). Some scholars add genre (relating to the context of culture) as a fourth important domain (Eggins 1994: Ch. 3).

2.1.3.

Functional Pragmatics

Functional Pragmatics is a German approach which sees language as a means designed for the transfer of knowledge and the achievement of mutual understanding (Ehlich 1991; Redder 2008). Building on Speech Act theory and Karl Bühler’s theory of linguistic fields (Bühler 1934; Daalder and Musolff this volume), Ehlich (1991) distinguishes five linguistic procedures which are used to label five corresponding functional fields with respect to a specific addressee. Each procedure is realized by specific linguistic means. Field

Procedure

Linguistic means

Symbol

symbolic

lexemes of content words

Pointing

deictic

local, temporal, personal deictics; tense markers

Guiding

expeditive

interjections; intonation; imperative and vocative mood

Operational

operational

anaphora; determiners; sub-/conjunctions; prepositions, accent

Painting

expressive

imitation; exclamative intonation

Figure 1. Fields, procedures, and linguistic means in Functional Pragmatics (adapted from Hoffmann 2001)

2.1.4.

Cognitive Linguistics and Construction Grammar

Most approaches of cognitive linguistics, such as Cognitive Linguistics (Lakoff 1987), Construction Grammar (Goldberg 1995; 2006), Cognitive Grammar (Langacker 1987; 2008), and Radical Construction Grammar (Croft 2001), consider grammar to be a semiotic system. Grammatical constructions, which are seen as the basic, irreducible building blocks of linguistic knowledge (see, e.g., Croft 2001; Goldberg 1995), are defined as form-meaning-pairings (Croft 2008). According to Langacker’s symbolic thesis, “Grammar (or syntax) does not constitute an autonomous formal level of representation. Instead, grammar is symbolic in nature, consisting in the conventional symbolization of semantic structure” (Langacker 1987: 2). The meaning-pole of grammatical constructions includes not only semantics in a narrow sense, but also discursive, interactional, textual, genre-related, rhetorical, and other pragmatic “meanings”.1 Besides, following Fillmore (1985), Lakoff (1987), and others, Cognitive Linguistics denies that there is a clear boundary between semantics and pragmatics, claiming that all linguistic knowl-

Pragmatics and grammar

429

edge is in fact encyclopaedic and based on experience. It is only very recently that cognitive linguistic approaches have turned to the analysis of text-corpora. When dealing with pragmatic aspects of the interpretation of grammatical constructions, they mainly focus on information structure2 and implicature (Kay 2004), while discursive phenomena are largely disregarded (but see Langacker 2001). Cognitive linguists hold that grammar is basically organized in the same way as the lexicon, “linguistic units” (Langacker 2000) or “constructions” (Croft 2008) being the overarching term. All linguistic structures are subject to processes of categorization (Langacker 2000). Just like lexical items, grammatical constructions are structured in terms of categories which exhibit a prototypical structure, resulting in radial and sometimes fuzzy categories, where the individual members need not share a common set of necessary and sufficient features, but may be related to one another by family resemblances. Grammatical constructions extend over time according to principles of metonymical and metaphorical extension, which supply for motivated options of the change of categorical structures, which, however, cannot be predicted. 2.1.5.

Usage-based approaches

The usage-based approach (Barlow and Kemmer 2000; Bybee 2006, 2007) judges frequency of usage to be the key factor for the becoming of grammatical constructions. Exemplar models predict that frequency of use is stored and determines the cognitive entrenchment of linguistic units because it correlates with the probability of storing a linguistic structure as an unanalyzed unit. Usage-based approaches resonate with connectionist models (MacWhinney and Bates 1989; Bod, Hay and Jannedy 2003), which rely on neurobiological principles. They predict that high frequency of reception and production of some linguistic structure strengthens its activation pattern and facilitates automatic processing. Simultaneously, the occurrence of linguistic structures leads to an inhibition of activation and de-automatization of competing structures (e.g., prepositional genitive with of vs. morphosyntactic genitive with -s). In this view, grammaticality (and conventionality in general) is a matter of degree and preference, depending on weighted strengths, which are associated with supposed neural activation patterns and which are subject to constant change. Synchronically, usage provides for prototype effects of grammatical structures, frequent structures becoming templates for categories and extensions organized around them. Diachronically, it is seen as a major cause of language change and, in particular, of grammaticalization (see 4.8). Such findings are in line with the position of Emergent Grammar advocated by Hopper (1998, 2004; Auer and Pfänder 2010). He rejects the assumption of any innate grammatical capacity and he denies that there was anything like the idea of a linguistic system. In his view, language exists only as an ever changing process of linguistic usage. It is informed by prior linguistic experiences, which provide the grounds for the routine

430

Arnulf Deppermann

use of (often schematically organized) “fragments”, which are creatively used as resources and adapted to the situated contingencies of local linguistic practice. In Hopper’s view, sense-making does not rest on rules for well-formed linguistic structures, but emerges from hermeneutic interpretation, which is in need of interactional ratification. 2.1.6.

Interactional Linguistics

Interactional Linguistics (IL) is strongly based on the methodology of Conversation Analysis (see Domke and Holly this volume). It regards the sequential organization of talk in oral interaction as the primary habitat of language.3 IL assumes reciprocal influences between linguistic and interactional structures: Linguistic structures are functionally adapted to discursive tasks, while patterns of discourse are sensitive to the linguistic options which individual languages afford and which speakers choose. IL insists that a theory of the grammar of spoken language must elucidate the principles of the online-production and -reception of talk-in-interaction (Auer 2009). Most important in this respect is the mechanism of projection of possible/required grammatical continuations and possible/expectable next actions given the production of an ongoing turn so far (Auer 2005 and 2009). IL studies put special emphasis on the interplay of grammar and prosody with respect to action formation (see also Couper-Kuhlen this volume). In contrast to the other approaches discussed here, IL is not a linguistic theory with a specific model of grammar and a definite descriptive apparatus, but rather a methodological approach. IL refuses abstract theorizing and insists on the analysis of collections of linguistic phenomena in talk-in-interaction based on audio-/video-recordings and transcripts. During the last years, however, researchers in IL increasingly adopt concepts from construction grammar as descriptive framework4 and positions from usage-based approaches and emergent grammar (Auer and Pfänder 2010).

3.

The relevance of pragmatics for grammar

In what follows, the impact of pragmatic tasks, processes, and conditions on grammatical structure both in terms of its situated use and of properties of the linguistic system will be discussed. Pragmatic key factors include temporality (online production and reception), interaction, cognition, action, the temporal and spatial arrangement of the participants, the modes and materiality of embodied and mediated communication and frequency of usage. For grammatical phenomena relating to the pragmatics-semantics interface, like the interpretation of connectives and scalar terms and the role of explicature, impliciture, enrichment, conventional and generalized implicatures in relationship to propositional content and “what is said”, see Saeed (this volume).

Pragmatics and grammar

3.1.

431

Processing constraints and the grammatical packaging of information

The communication of thoughts and the performance of actions by linguistic practice in real time is constrained by speakers’ and hearers’ capacities of cognitive processing. The temporal scope of attention and working memory delimits the complexity of syntactic and propositional structures which can be processed as units of planning and understanding. According to Chafe (1994: 109), there is a “one new idea constraint”, meaning that “neither the speaker nor the listener is able to handle more than one new idea at a time”. This constraint guides the packaging of information in prosodic and syntactic units. Preferentially, speakers formulate one new information per intonation unit, which has its (one) focus accent on the item which expresses the new information. The intonation unit almost never comprises more than one clause. As a consequence, syntactic structures with multiple recursions, which force speaker and hearer to keep track of two or even more syntactic and semantic structures simultaneously, are strongly dispreferred because of the processing difficulties they cause. An important syntactic repercussion of the ‘one new idea constraint’ can also be seen in the constraints of Preferred Argument Structure (Du Bois 2003), which have been proved to hold for a large variety of typologically unrelated languages (Du Bois, Kumpf and Ashby 2003). Du Bois (2003) claims that speakers avoid more than one lexical core argument (which encodes the new information). For transitive constructions, this amounts to the constraint to avoid lexical subjects (and to express them pronominally), while direct and indirect objects can be lexicalized. The constraints of preferred argument structure explain, e.g., preferences for cleft-constructions which can be observed in many languages, such as French (cf. Lambrecht 1987). Instead of producing a transitive clause with both lexical subject and object, speakers prefer a cleft-construction, which package the information in two clauses with only one lexical argument each. (1) Transitive clause Beaucoup de gens n’aiment plus cette façon de parler. ‘Many people do not like this way of talking anymore.’ (2) il y a-cleft Il y a beaucoup de gens qui n’aiment plus cette façon de parler. ‘There are a lot people who do not like this way of talking anymore.’ Such cleft-constructions fulfil the same pragmatic function as left dislocation and free topics: They are used to introduce a referent in a segment of its own, which, in addition, is often produced as a separate intonation unit apart. Only after this, the referent is topicalized and predicated upon. (3) Left dislocation aber ich mein ne gewisse neugier (.) die muss ja bleiben ‘but I mean a bit of curiosity (.) it must remain’

432

Arnulf Deppermann

(4) Free topic die polnische Wurst da (.) ist das so wie Salami? ‘the-FEM Polish sausage over there (.) is it-NEUTR like Salami?’ Grammatical constructions as in (3) and (4) give evidence that in contrast to what Speech Act Theory holds (cf. Searle 1969), an elementary speech act need not consist of reference plus predication. Rather, establishing a referent often is a fullblown action by itself. It is performed by clausal and phrasal constructions which separate it from the following predication and serve as a starting point for it, which first is grounded independently. Moreover, referential act and predicative act each occupy an intonation unit of their own. In a similar vein, speakers prefer right-branching constructions over leftbranching constructions because they allow for a cumulative processing of structure and meaning, which correlates with the temporal production of phrases. Leftbranching constructions, however, require language users to put elements “on hold” in the working memory, before being able to assign them to their syntactic heads, which are produced later. The need to keep track of several syntactic and semantic representations in parallel also strongly disfavors central embedding, let alone multiple recursion (such as multiple center-embedded relative clauses). 3.2.

Structuring information with respect to relevance and the recipient’s knowledge

Text and talk is designed with respect to what speakers/writers assume to be common ground with their addressees. New information, which is either unknown or not salient to the addressee at the moment, is introduced into the discourse by relating it to given information, which is judged to be salient, i.e., activated and retrievable with low processing effort for the hearer. In order to structure and encode information accordingly, speakers keep track of the cognitive availability of referents (but also other denotata, such as states, actions, and events). Given vs. new is not a simple dichotomy. Rather, there are different degrees of accessibility of referents (e.g., Chafe 1994), which depend crucially on their identifiability and their state of prior activation. Denotata may be accessible because they are situationally salient (via perception), established as parts of world knowledge (like the prime minister), have been introduced in prior discourse, or can be inferred from it via bridging assumptions which, e.g., relate parts to wholes (Clark 1977). (5) Bridging assumption (whole f part) The couple moved on to the altar. The bride wore white. f Bridging assumption about co-reference: The bride is part of the couple mentioned before.

Pragmatics and grammar

433

The assumed familiarity of referents for the recipient thus can range from (brand-) new over inferrable to (exophorically or anaphorically) evoked items (cf. Prince 1981). Accessibility of referents depends on several factors, such as its topical vs. non-topical status, recency of last mention of the referent, or the degree of cohesion of the syntactic unit which contains the prior mention to the ongoing unit (Ariel 1990). Ariel (2008: 44–52) proposes a ‘referential marking scale’. It orders referential expressions according to the (assumed) degree of accessibility of the referent for the recipient, which in turn depends on the degree of its activation in the recipient’s memory that can be assumed when producing the referential expression (see also Gundel, Hedberg and Zacharsky 1993). While new elements are introduced by indefinite NPs, accessible and given elements are referred to by definite NPs or (if situationally salient or evoked) by pronouns. Starting with the most marked expression to be used for the least accessible referents, the referential marking scale for person reference looks like this in more detail (Ariel 2008:44): full name > long definite description > short definite description > last name > first name > distal demonstrative > proximate demonstrative > stressed pronoun > unstressed pronoun > cliticized pronoun > verbal person inflections > zero

This scale shows how the two pragmatic preferences for person reference, namely minimization and use of (recipient designed) recognitionals (Sacks and Schegloff 1979), translate into preferences for the use of grammatical resources. Their relevance can also be seen in cases of repair like (6), where more minimized forms are used first and get replaced by (grammatically and semantically) more elaborate forms if they fail to engender the recipient’s recognition. (6) Preference for minimization and use of recognitionals (example from Sacks and Schegloff 1979: 19) A: I was the only other than the tch Fords? Uh Mrs. Holmes Ford? You know uh the the the cellist? B: Oh yes. She is the cellist. Apart from morphosyntactic coding, word order is used as a syntactic device to index the assumed familiarity of referents. Halliday (1985: 275) claims that there is a “natural” ordering principle: “the Given typically precedes the New”, i.e., the speaker organizes the message as a trajectory starting from the current state of consciousness of the recipient to an updated state (see also Chafe 1994). Therefore, given information is expressed as theme in the prototypical case. In Germanic and Romance languages, the theme is the first constituent of a sentence, whereas in Japanese it is marked by the post-positioned theme marker -wa. New information is presented in the rheme, i.e., the remainder of the sentence (Halliday 1985: Ch.3). In addition to the assumed familiarity of referents, a second important dimension of information structure is the organization of focus vs. background (Büring 2007;

434

Arnulf Deppermann

Krifka 2007). ‘Focus’ refers to the part of an intonation unit which is framed as being most relevant. Some of the means speakers can use to mark focus are the same as for marking accessibility: word order and specific topological constructions like (pseudo-)clefts and passives. In addition, most important is the placement of the so-called focus accent (Uhmann 1991; Büring 2007), which marks the focal word (but can project to higher-order nodes such as phrases or even the whole clause). “All other things being equal a Speaker will choose the Theme from within what is Given and locate the focus, the climax of the New, somewhere within the Rheme” (Halliday 1985: 278). While there is a high correlation between focus, new information and rheme, they need not coincide (Gundel and Fretheim 2004). The focus does not necessarily need to be put on new information, but it can, e.g., be contrastive, i.e., marking an element which stands in contrast to a prior turn or the recipient’s probable expectations. The focus can also comprise the whole sentence, as in the case of offering requested information to open questions like (7) Focus on whole sentence (= requested answer) A: What’s up? B: The computer doesn’t work. or as in newspaper headlines (8) Focus = whole headline (The New York Times, November 27, 2009) Pilots Were ‘Distracted,’ Transcripts Show In elliptical responses, only the focus elements are realized. (9) Adjacency ellipsis: Question-answer sequence A: wieviel Lieder können wir spielen? was meint ihr? ‘How many songs can we play? What do you think?’ M: Z::irka fuffzehn; ‘About fifteen.’ In spoken language, the focus is often placed in thematic position, particularly, if it is contrastive. (10) Focus on theme A: WANN ist er gekommen? ‘when did he come?’ B: GEStern war er=s erste mal da. ‘yesterday he was there for the first time.’ Another grammatical construction for focus marking by assigning it to the theme is preposing a constituent (off he went; cf. Ward and Birner 2004). Focus preposing needs to be distinguished from topicalization of a referent, which does not carry a focus accent and contains neither focus nor new information.

Pragmatics and grammar

435

As stated above, focus marking can be independent of accessibility status: (11) Contrastive accent Das war nicht ICH. Das warst DU. ‘It wasn’t ME. It was YOU.’ In fact, focus marking and marking of accessibility are not determined by identifiability or activation status of referents, but they can be used as rhetorical devices to introduce new information by simply presupposing it as hearer-old (see also Edwards 1997:117–122): (12) Coding givenness as a rhetorical device (advertisement by Volkswagen) Just when you thought our legendary hatchback couldn’t get any better, along comes the stunning new Golf. In sum, signalling familiarity of referents and assigning focus vs. background results from an interaction between grammatical forms (such as definiteness, pronominalization, deictics, etc.), word order, and prosodic cues (accentuation), which at times can be very complex. 3.3.

Projecting turn-construction and turn-taking

In spoken language, the clause is not only the preferred format for packaging ideas (see 4.1), it is also used as the primary construction format for turns and actions (Thompson and Couper-Kuhlen 2005). Participants use clausal structures for indexing and recognizing transition relevance places at which turn-transition might be due (Selting 2000; Sacks, Schegloff and Jefferson 1974). A basic mechanism to do this is temporal projection (Auer 2005 and 2009), which rests on the incremental production of linguistic structures in time. Linguistic structures produced in the turn so far open up syntactic gestalts, i.e., probable or required continuations, which have to be closed by producing the required missing elements. E.g., the apodosis of a conditional projects a protasis (and this is: a clause-combination) to follow, verbs project required arguments by virtue of their valence frame, determiners project noun phrases, etc. Projection depends on typological features of the language. E.g., in German the verbal brace (Verbklammer) requires speakers to finish clauses with an infinitival form (infinitive or participle) of the main verb (unless it is used in present tense without a modal). Listeners can use this feature to generate expectations about when the turn is possibly completed. SVO-languages like English permit an early projection of turn trajectories because the verb makes specific choice of thematic roles expectable. An OV-language like Japanese, however, does not allow for early projection, because constituents other than V do not project the number of arguments and because Japanese allows for a much higher degree of ellipsis. Still, it marks endings more clearly in terms of syntax than SVO-languages because of the regular positioning of the verb in sentence-final position and be-

436

Arnulf Deppermann

cause of the additional use of sentence-final particles (Thompson and CouperKuhlen 2005). Projectability of turn-endings makes itself felt not only in smooth turn-transition, but also in cases of overlap, which tend to occur more often if elements are added after a possible clausal gestalt has been completed and thus a possible transition relevance place has been reached (Jefferson 1973; Schegloff 1996). The collaborative production of sentences is another evidence for participants’ orientation to the exigencies of grammatical structure in terms of projectable trajectories of clauses (Hayashi 2003; Lerner 1991, 1996, 2004): After the first speaker has produced the first clause of a complex sentence, second speakers manage to deliver the second clause which was type-projected (e.g., a then-clause after an if-clause has been completed) or they may add phrases or words at the places where their occurrence is due. Although clauses have more or less determinate endings, which may signal turn-transition, they can be expanded by additional elements, so-called “expansions” or “increments”, which are syntactically related to the syntactic structure constructed so far.5 Such turn-continuations occur regularly if a projected next speaker does not take the turn when a transition relevance place is reached. They address problems of understanding and affiliation with the turn so far, which might account for an actual lack of response or an anticipated difficulty in responding. They serve to repair the respective deficits of the turn and to renew the occasion for turn-transition. Whether clauses project turn-transition and whether expansions are a continuation of an ongoing turn-constructional unit or a new unit does not only depend on syntax, but also on the prosodic contour of the clause and on the degree of prosodic integration of the continuation. Transition relevance places are not only determined by syntax, but they emerge from an interplay of syntactic, prosodic, semantic, and, at times, also pragmatic and genre-related considerations (as, e.g., narrative structure). 3.4.

Social action formats

Grammar is a primary resource for performing recognizable social actions. Speech act theory conceives of the relationship between speech acts and grammar in terms of illocutionary force indicating devices, which index and sometimes even encode types of speech acts. Most important indicators are sentence type (declarative, imperative, interrogative), word-order, modal verbs and modal particles, performative formulae, punctuation, and sentence-final pitch contours (Searle and Vanderveken 1985; Liedtke 1998). Empirical research into the relationship between action and grammar has refined this picture quite a lot: Instead of relying on theoretically derived speech act taxonomies and trying to find linguistic correlates for very broadly defined (universal) types of speech acts, corpus-based research shows that linguistic action needs to be understood in terms of much more specific emic types of actions to which speakers and recipients orient as such. At times,

Pragmatics and grammar

437

they are tied to very specific sequential contexts, genres, and speech communities. These actions are realized by idiomatic grammatical constructions, which have become closely associated with them by routine usage in specific pragmatic contexts and genres, leading to the conventionalization of pragmatic meaning (Feilke 1996 and 2004). Examples of ‘social action formats’ (Fox 2007) are constructions with German nicht verstehen (‘do not understand’; Deppermann 2010), which are used for building reproaches: (13) Social action format: [nicht verstehen (können) COMP] indexes reproach B1: ich versteh auch andereseits NICHT wenn die‘on the other hand I don’t understand when the’ klägerin zum arzt nach BRUCHtal MUSS JA? ‘plaintiff (=A1) needs to go to the doctor in Bruchtal’ dass sie dann in de innenSTADT im maxim ‘that she then goes to Maxim’s in the city’ KAFfee trinkt ‘to have a cup of coffee’ (1.8) A1: wieso is des verBOte? ‘why is that forbidden?’ Deontic infinitives in German are another example. Deontic infinitives are not syntactically integrated, i.e., they are “free infinitives”, which are neither governed by a matrix sentence nor are they part of a composite verbal phrase made up by an auxiliary and an infinitive (cf. Fries 1983; Deppermann 2006 and 2007). With deontic infinitives, the speaker expresses a normative stance towards some kind of (future) action denoted by the verb. They are employed for a variety of deontic actions, which may vary in terms of the degree of obligation, such as commands, requests, recommendations, and suggestions, but also the expression of a wish (optative) or a complaint. (14) Social action format: Deontic infinitive used for suggestion A: also der Ablauf is grob (--) ‘so the procedure will be something like this’ Anmoderation; (.) oder lAbern; ‘anchorage or talk-INF’ M: vorstellen auf jeden fall ‘introduce-INF in any case’ (15) Social action format: Deontic infinitive used for complaint noch drei wochen arbeiten! ‘still three weeks of work-INF’ arbeiten arbeiten immer nur arbeiten ‘work-INF work-INF always only work-INF’

438

Arnulf Deppermann

An English example is the use of the grammatical construction [PRO COPULA (Grading particles: really/so/quite/pretty) PRED-ADJ] for assessments, in particular, for compliments (Goodwin and Goodwin 1987): (16) Social action format: [PRO COPULA (Grading particle) PRED-ADJ] used for assessments This is so nice. That’s quite good. Another example is the use of post-positioned causal adverbial clauses (starting with because) as justifications dealing with actual or anticipated disaffiliation on the recipient’s part (Ford 1993): (17) Social action format: Post-positioned causal adverbial clause dealing with a missing recipient’s reaction (example from Ford, Fox and Thompson 2003: 126) A: well do you think it’s umm (0.2) ahm (0.2) stress? (.) Cause a lot of back I know back pain (0.2) comes from stress. Actions are rather contextually indexed than invariably coded by grammatical constructions: Sequential, genre-related, and other contextual conditions, lexical choice of the verb, and prosody determine which precise action is performed by a deontic infinitive; causal adverbial phrases may also be used to secure understanding or to refute a statement, which an opponent had asserted before. Grammar only very rarely stands in a context-free, one-to-one relationship to some type of action. 3.5.

Displaying stance and marking discursive coherence

Texts and interactions do not simply consist in the exchange of information and the performance of actions. Speakers use various grammatical means to display their epistemic and evaluative stance with regard to information and action, to convey how they relate to recipients’ assumed understandings and to connect their utterance with prior and subsequent clauses and turns. Various kinds of discourse particles (see Fischer 2006) such as discourse markers (like you know, well, and in some of its uses; Schiffrin 1987), response particles (like uhum, yes, okay; Gardner 2001; Sorjonen 2001), and modal particles (like just, germ. wohl, eben, halt, ruhig; Diewald 2006a) are grammatical means which are specialized for displaying how some action relates to prior (or presupposed) context in terms of coherence, epistemic stance, or affiliation with the partner. With the exception of modal particles, discourse particles are not syntactically integrated. Mostly, they occupy turn-initial positions in the so-called pre-front field (Auer 1996b), or, in the case of response particles, they sometimes make up the whole turn. Pre-front field elements can also be adverbials, address terms, and routine formulae.

Pragmatics and grammar

439

(18) Pre-front field: Connective + response particle also gut das haben sie verstanden ‘so okay you got this’ (19) Pre-front field: Address term + discourse marker herr meier ich mein wir haben uns ja gestern schon drüber unterhalten ‘mister meier I mean we talked about this already yesterday’ Pre-front field elements serve as pivotal elements which retrospectively tie the upcoming action to prior context and prospectively project the type of action to come in various respects. Projector constructions comprise another set of prefatory grammatical structures which serve for anticipatory framing of upcoming discursive action (Günthner 2008; Hopper and Thompson 2008). Projector constructions can be realized in various forms, which are in part language specific: as pseudo-clefts (wh-clefts; Günthner and Hopper 2010; Pekarek-Doehler 2010), it-extrapositions (Prince 1978), [N be that]-constructions (Günthner 2008; Schmid 2001), il y a-clefts, or the je veux dire que-construction (Pekarek-Doehler 2010). (20) German pseudo-cleft (from Günthner and Hopper 2010) was so die sachn in natUrkostlädn angeht, ‘as far as the things in health food shops are concerned’ die sind ja dann meistens ‘they are mostly’ erfahrungsgemäß auch relativ teuer ‘also quite expensive’ (21) English pseudo-cleft (from Günthner and Hopper 2010) and so what happens is, your mic runs into i:t, and then it plugs into your receiver (22) French pseudo-cleft (from Pekarek-Doehler 2010) ce que je voulais dire ‘what I wanted to say’ c’est vrai qu’ il y a quand-même tout ce phénomène, ‘it’s true that there is nevertheless this whole phenomenon’ Examples (20)-(22) show that the two parts of the pseudo-clefts need not be linked by a copula and a subordinating construction (what p is that q). Moreover, the second part often does not consist of a clearly delimited clause, but rather of some more vaguely defined longer stretch of talk. The projector construction thus is a very interesting case of the intersection of grammar and pragmatics (cf. Hopper and Thompson 2008). The first parts (the wh-clauses etc.) do not project some spe-

440

Arnulf Deppermann

cific clausal structure to follow, but they rather serve as flexible fragments for holding the turn by projecting a multi-unit turn to follow. Depending on the metacommunicative framing or the announcement made by the wh-clause, it projects an argumentation, a complex statement or a small story. 3.6.

Presupposing context

Ellipsis provides the most striking evidence for the claim that syntax is tied to the here and now of talk (Bühler 1934) and that it is “positionally sensitive” (Schegloff 1996). According to traditional definitions, ellipsis applies to grammatical structures, which are acceptable in context (i.e., they are not anakoluths), but which are “phonologically reduced” with respect to the exigencies of syntactic completeness and of their semantic interpretation in situ. Ellipsis results from economical motivation (cf. Grice’s second maxim of quantity, see Grice 1975). It requires that elided elements can be recovered from context (Halliday 1985: Ch.9.3). Context can offer the missing elements in basically two ways: Ellipsis may be anaphorical or situational. a) Anaphorical ellipsis copies presupposed elements from prior text or talk (save adaptations of speaker-listener deixis). Adjacent ellipsis is a most obvious case. Typically, a second turn is restricted to the production of the (requested, problematic etc.) focus element, while the prior turn is presupposed as background: (23) Adjacent ellipsis: Answer IN is sexualität für sie total WEG? ‘is sexuality completely off for you?’ IE als unterhAltungsthema (.) JA. ‘as a topic for talk (.) yes’ (24) Adjacent ellipsis: Follow-up question IE: das finde ich falsch ‘that is wrong in my view’ IN: warum? ‘why’ However, phonological reduction is not the full story. There are in cases where there is no prior clausal structure which can simply be copied. Consider cases like (25) and (26): (25) Unattached NP increment (example from Ford, Fox and Thompson 2002) She had it [a baby] yesterday. Ten pounds.

Pragmatics and grammar

441

(26) Free NP CH: der abwasch ist FERtig (-) darf ich jetzt FERNsehen? the dishes are cleaned (-) may I watch TV now? PA: ERST hausaufgaben DANN fernsehen. first homework then TV While the subject can be copied from the preceding turn constructional unit in (26), there is no verbal antecedent to be presupposed for the elliptical structure. Instead, it has to be constructed on the basis of the representation of the conceptual content of the discourse so far and by resorting to encyclopaedic knowledge and sequential expectations with regard to next actions. E.g., in (26), the request for permission makes permission vs. refusal conditionally relevant and assigns normative power to the respondent. In addition, the construction [first NP/VP then NP/VP] is an elliptical social action format, which codes a normative order of actions. Thus contextual resources of interpretation can combine with the semantics and pragmatics which are tied to elliptical grammatical constructions as such. Ellipsis may not only be motivated by striving for economy, but it may also have rhetorical functions. Deontic infinitives (cf. (14)-(15)) are a case in point: By neither encoding subject (i.e., an agent) nor mood, the construction neither specifies who has to carry out the action nor the degree of obligation assigned to the agent. In some contexts, the lack of mood-coding and the omission of arguments (agents, but also often patient, location, instrument, etc.) is a politeness strategy (for requests and commands), while in others (like advertisements, see (27)) it allows the recipient to define for him/herself which degree of normative obligation suits and if s/he feels addressed by the deontic infinitive at all. (27) Deontic infinitives in advertisement Jetzt mitspielen und gewinnen! ‘Join-INF the game and win-INF now!’ In addition, these omissions help to profile (Langacker 2008: 66–70) the action as the most important element of the situational frame evoked by the infinitive clause. These factors seem to account for the fact that in many contexts deontic infinitives are preferred over imperatives and modal constructions, which require the speaker to encode agents and the degree of normative obligation. Ellipsis can reach back far beyond an immediate prior interactional context. Raymond (2004) shows how stand-alone so in American conversation is used to prompt an action from a recipient as an upshot from the ongoing interactional sequence which had been due already (much) earlier in the interaction. The syntactic incompleteness of so is a case of aposiopesis. It invites the recipient’s collaboration by indexing that the recipient is supposed to know on behalf of the prior interactional sequence which kind of action is warrantably expected from him/her. A similar case is sentence-final but, which indexes a “hanging” objection which can be recovered from context (Mulder and Thompson 2008).

442

Arnulf Deppermann

b) When situational ellipsis is used, it is the non-verbal context with respect to the speaker’s origo (Bühler 1934), which supplies the missing information. (28) Situational ellipsis: pragmatic routines (customer to waiter) The bill, please. (29) Situational ellipsis: pragmatic routine and local specification (announcement at a railway station platform while train is approaching) Zurückbleiben! ‘Keep away!’ (28) and (29) are examples for routinized elliptical constructions which are tied to specific social and (in the second case) spatial configurations which they require for their interpretation. Note that there is a reflexive relationship between the construction and the context: While the situational context is necessary for their precise situated interpretation, the formulae themselves organize contexts by indexing features which are (possibly) relevant. 3.7.

Multimodal coordination

Situational ellipsis mostly does not simply rest on some spatiotemporal context which is taken for granted. Rather, the production of grammatical structure is coordinated in a fine-grained way with other modes of communication (Kress and van Leeuwen 2001; Norris 2004) which make contexts selectively salient and which cooperate in determining meaning. In oral interaction, the meanings of spatial reference and situational ellipsis are established by a complex interplay of linguistic structure, pointing, available visual clues, background knowledge, and larger activity frames, which provide for putative relevance and possible semiotic importance of spatial structures (Goodwin 2003). E.g., pointing gestures need to precede deictic terms used for spatial reference within or shortly before the same turn in order to draw the recipient’s attention to the objects designated and thus to constrain the situated meaning of the terms used. An example for the multimodal embedding of ellipsis would be a director saying much too slow while watching a film take on the monitor at the film set. Her visual orientation to the screen, knowledge about some prior interaction with the camera woman, and the routine practice of the director reviewing film takes in order to decide if they are acceptable will be required in order to understand that the turn is taken to mean that the camera moves “much too slow” and that the take needs to be redone. As in this example, coordination of gaze between speaker and hearer, visually available context, and gesture mostly provide for the “thematic” referential aspects of communication, whereas linguistic communication is restricted to “rhematic” predications expressing some stance or request for action. However, gesture can also be used to elaborating a predication, as in the case of modal in-

Pragmatics and grammar

443

dexicals like germ. so and engl. like (Streeck 2002). They need to be accompanied by an illustrative gesture in order to clarify which semantic aspect (like ‘size’, ‘form’, ‘direction of the referent’, ‘kind of movement’) is indexed by the modal expression. 3.8.

Grammatical variation relating to genre and medium

There are huge differences between spoken and written varieties in virtually all areas of grammar, e.g., regarding the use of (complex) clause patterns (such as extraposition, subjunction), tense and aspect marking, valency, or discourse markers (Biber, Johansson and Leech 1999; Bublitz, Lenk and Ventola 1999). A closer look, however, reveals that there is no simple dichotomy between spoken and written language. Grammar is sensitive to fine-grained differences of register (e.g. Biber 2006), genre and medium. During the last decades, micro grammars related to genres of computer mediated communication (CMC) have developed.6 In genres and technical forms of communication with a high degree of interactivity like chats and SMS in particular, new grammatical conventions have emerged for the display of emotions (multiplication of vowels and diacritics, expressive use of capital letters, emoticons), the mimicry of nonverbal communication (root inflectives known from comics such as germ. heul, grins), the adaptation of writing to oral interaction (eye dialect, dialectal and substandard spellings: boyz, wuddya; use of discourse markers and non-standard grammar), and we-codes, which are intransparent to out-groups (e.g. by using acronyms like lol, CTN, 4U; Schlobinski 2006). In CMC in general, standards of linguistic correctness have become liberalized, giving way to innovative playful, expressive, and economical uses. In print and internet media, texts often combine with the visual design of pictures (Kress and van Leeuwen 2006; Stöckl 2004) and other semiotic modes such as sounds, animations, and movies (e.g., in vlogs). On websites, e.g., large parts of texts are elliptical, consisting of “free nominals” which are not integrated into any syntactic structure. However, their meaning is supplied by the visual line-up they are embedded in and by the users’ pragmatic knowledge about how to use them on behalf of there being links to other pages, forms to be filled out, etc. (Schmitz 2010). 3.9.

Pragmatic factors of grammaticalization

Language-internal and social factors have already been discussed for a long time as determinants for linguistic change (see Aitchison 2001). Over the last three decades, inferential, discursive, and usage-based processes have increasingly become recognized as major factors promoting linguistic change. A core area of change is grammaticalization. The term refers to the emergence of new grammatical forms (from old ones); more specifically, grammaticalization involves loss of lexical

444

Arnulf Deppermann

meaning (semantic bleaching), gain of functional meaning (e.g., word-class conversion), loss of phonological substance, restricted syntactic distribution or morphological incorporation, and may involve the development of a discursive pattern into a grammatical convention. Grammaticalization has been found to be a probably unidirectional process, which runs along clines of increasing grammaticalization (Haspelmath 1999; Hopper and Traugott 2003: Ch.5). One such cline, which can be attested for several languages, is the following diachronic path: content word f function word (analytic) syntax (e.g., PRO) f agglutinative morphology (clitization) f flexive morphology (e.g., subject marker) f zero. Grammaticalization can be motivated by various strategies and preferences of language users: – Articulatory economy leads to phenomena like reduction/attrition, clitization, assimilation/co-articulation (Lass 1987). – Processing economy can be seen at work in heavy NP-shifts (Wasow 2002), which correspond to the “Gesetz der wachsenden Glieder” (law of increasing constituents, Behaghel 1909). – The enrichment of expressive power by innovations. – The desire to communicate unambiguously leads to, e.g, the decay of homonyms. – Iconicity (Haiman 1985) is a cognitive motivation which builds on principles of naturalness (e.g., constructing superlatives by reduplication: chef chef; again and again) – The desire to accrue social prestige, to gain the status of a member and to identify with a social group leads to the use of linguistic practices of prestigious ingroups (Labov 2001; LePage and Tabouret-Keller 1985). Innovations in youth language provide paramount examples. To be sure, linguistic change in general and grammaticalization in particular do not only involve change of forms, but also semantic and pragmatic change. A common mechanism of change is metaphorical extension, which works by analogical transfer of concepts from a source to a target domain. Metaphoric change can be observed in many languages for deictics, where a metaphorical transfer of expressions for body parts and bodily orientation to spatial reference in general comes about (Heine, Claudi and Hünnemeyer 1991; Heine 1997). A well attested path of metaphorical change in the meaning of prepositions and connectives is spatial f temporal f causal f epistemic/speech act/discursive/textual (Sweetser 1990; Haspelmath 1997). The general dynamics of metaphorical extension consists in using anthropocentric, visually based, concrete concepts for more abstract, discursive, and subjective entities. The diachronic change of, e.g., connectives (like while, since, after (all)) thus shows increasing subjectification and abstraction in terms of being used for displaying speaker’s attitudes, epistemic perspectives and expectations (Sweetser 1990).

Pragmatics and grammar

445

While metaphor operates on the paradigmatic level (supplanting one meaning with another), metonymically based change rests on reanalysis of syntagmatic relationships, which in addition often become contracted over time (Hopper and Traugott 2003: Ch.4). In metonymic change, a lexical item or (parts of) a construction adopt meanings from their collocational contexts. Metonymic change is more gradual than the more abrupt and drastic metaphoric change, because what develops is rather informative strengthening or enrichment (i.e., an implicature is added; Ariel 2008:163) and successive transformation. A much discussed example of metonymic change is the development of to be going to (e.g., Hopper and Traugott 2003: 88–93; Bybee 2006): Originally, it denoted ‘physical movement to some place’; later it became a purpose-construction, referring to the intention of an actor; nowadays, it has become a generalized future construction, which may also be used with inanimate, non-intentional subjects. Morphologically, there is a reanalysis of going to, which is no longer interpreted as a progressive form of a full verb, but as a tense marker, and there is phonetic reduction (gonna), which attests to the unit status of the construction. This reduction is absent in the progressive use (I’m going to a party vs. * I’m gonna a party). In many cases, implicatures (or invited inferences, see Traugott 2004) are at the beginning of grammaticalization. The starting point are conversational implicatures which heavily depend on specific supportive contexts. Gradually, they become generalized implicatures, i.e., prototypical meanings (Traugott 2004). Grammaticalization has come about when the new meaning is firmly established, so that there are (at least) two meanings, i.e., the item/structure has become ambiguous. The next step then is reorganization and differentiation of the two meanings in terms of the development of specialized distributions and collocational potentials, differences in potentials for TMA-marking, etc. (Diewald 2006b). While cognitive motivations and interpretations can be seen as important driving forces for grammaticalization, cognitive factors as such are never directly responsible for language change. They can leave their fingerprint on language structure only via their discursive realization. Moreover, discourse is not only a medium for the linguistic reflection of cognitive preferences. There are genuine discursive motivations for the development of grammatical structures. “Probably all successful grammaticalization paths had previously led to salient discourse patterns, themselves only necessary (but not sufficient) conditions for conventionalization” (Ariel 2008:150). Except for language planning, coining of terminology, and some innovations promoted by the media and by subcultural actors, people only rarely intend to promote language change. And even if there is some definite intention of innovation, its success fully depends on how it is taken up and spread by recipients, both in terms of the interpretation of the innovation and the frequency of use. At least in the early stages of innovation, intentional strategies of people pursuing communicative goals often are responsible for language change, which, however, as such mostly happens as an unintended by-product of

446

Arnulf Deppermann

these strategies. Keller (2004) terms such changes “invisible hand-processes”. Sometimes, the long-term consequences of the factual collaboration of intentional strategies of a community of linguistic actors can even lead to a result which is quite the opposite of what was originally intended. An example is the dynamics of diachronic change of terms for women in German (Keller 2004): The collective adherence to a strategy to select distinctive terms for addressing and talking about women eventually leads to a pejorization of these terms (Weib, Weibsbild, Frauenzimmer): Frequent use makes them lose their distinctive potential, thus speakers strive to replace them by more marked and highly valued terms, which get increasingly used as a consequence and thus also lose their distinctive value, etc. (The same process can be observed in an accelerated pace with innovations in youth language, e.g., regarding evaluative predicates). So, in the long run politeness and distinction strategies have the inverse impact on the terms used. Other discursive preferences are conservatism (in a more macro view) and, in terms of a micro view of dialogical sequences, priming and resonance (Pickering and Garrod 2004; Brennan and Clark 1996). They account for the preference to follow precedents, which thus can become canonical. These strategies of repetition are most promising in terms of securing understanding and building on intersubjectivity achieved before (Feilke 1996). They provide for a systematic interactive motivation of the grammaticalization of options which initially may simply have been chosen arbitrarily among a set of possible options (see Clark and WilkesGibbs 1986). Patterns of use lead to the gradual emergence of new grammatical patterns from non-innovative, grammatical usage. The emergence of discourse particles (well, but, though) from connectives and other discourse markers (like y’know, verstehste ‘do you understand’, voilà ) from mental verbs are examples of gradual change. Precursor for the development of discourse markers from mental verbs is the emergence of the construction [matrix sentence with mental verb + dependent main clause] (Auer 1998), i.e., a semantically dependent clause follows which does not exhibit any marker of subordination. The matrix sentence can increasingly change into a semantically bleached discourse marker, which then, in later phases, may fully grammaticalize in terms of adopting a specific distribution (often syntactically unembedded in turn-initial or turn-final position) and retaining only “light” lexicosemantic content (if any) (Thompson 2002; Imo 2007; Karkkäinen 2003). Dialogical patterns can also be a source of grammaticalization. A case in point are clause combinations which arise from interactional sequences and transform them into grammatical formats to be used by individual speakers in single turns (Givón 1979). An example is the emergence of conditionals from a questionanswer-response sequence of three turns:

Pragmatics and grammar

447

(30) Grammaticalization of interactional sequences in complex sentences: Conditionals A: Yes/No-Question (Will you help me?) B: Agreement (Yes.) A: Consequence (Then I’ll give you two dollars.) f A: Protasis (If you’ll help me), Apodosis (then I’ll give you two dollars.) Discursive and cognitive factors are often intertwined in processes of grammaticalization. The category ‘noun’ exemplifies how cognitive concerns are mediated by discursive ones. Cognitive grammar claims that the grammatical category ‘noun’ has a schematic semantics, which consists in conceptualizing referents as “things” which are defined as “any product of grouping and reification” (Langacker 2008:105). The analysis of the realization of nouns in discourse, however, shows that prototypical representatives of nouns, which are count nouns and exhibit all grammatical features which are regarded as typical of nouns (determiners, gender-, number-, and case-marking, etc.) introduce discourse topics which are “manipulable”, i.e., available for subsequent co-reference (Hopper and Thompson 1984). Grammatical categoriality thus is achieved by discursive patterns, not by mental conceptualization alone. Discursive considerations can even override cognitive motivations. In the overwhelming majority of cases, causal adverbial phrases (with because, weil, par-ce que etc.) follow their nucleus assertion, although adherence to an iconicity principle would predict that they precede because causes are prior to effects (Ariel 2008: 175 f.). However, it is neither “the way the world is” (temporal order) nor cognitive preference for iconic representation, but the discursive function of dealing with the recipient’s actual or anticipated disaffiliation by supplying justifications, causes, etc. (see Ford 1993), which motivates the syntactic distribution of causal adverbial phrases. According to usage-based approaches to linguistic change, frequency of use of linguistic forms is the driving force in grammaticalization (Bybee and Hopper 2001; Bybee 2006, 2007). Neither morphological nor semantic innovation needs to be the starting point for change, but it is the preferential, salient discourse pattern (Ariel 2008), which, via routinization, finally becomes entrenched as a cognitive unit (cf. Langacker 2000). An example of this is the cline of PRO becoming clitic, then morphosyntactic marking and finally zero. Another case in point is the increasing phonetic reduction of the French negative pre-verbal particle ne, so that only the post-verbal pas is retained as marker of sentential negation. Entrenchment has effects of its own, phonological reduction (leading to clitization, assimilation/ fusion, and elision; see above) being the most prominent one. Frequency has also two other effects (Bybee 2006): conservation, and autonomy. The conservation effect is evidenced by the resistance of irregular forms to change because of their high token frequencies. A case in point are irregular English past tense forms like

448

Arnulf Deppermann

was, came, had. They survive, while verbs with low token frequency successively succumb to the high type-frequency of the regular, schematically derived past tense morphology with -ed inflection (ex. wept f weeped). Autonomy (or emancipation) means the re-analysis of old and the emergence of new constructions, which become used as unanalyzed units. They change and sometimes even cut across syntactic categories, as could be seen in the case of gonna or the emergence of discourse markers. Autonomy and unit status are also evidenced by increasing speed of production and comprehension, prosodic segmentation from the discursive context and syntactic parsing as a unit. However, it is important to note that although frequency may play a major role in (the stabilization of) grammaticalization, this is not necessarily its cause. Rather, we need to ask why some pattern becomes frequent. Motivations like economy (against redundancy of coding and high predictability/inferrability in context) and others (see above) may cause or at least support high frequency of use. Also, frequency does not automatically lead to grammaticalization, because not all frequent patterns change, and the effect of frequency is mediated by factors like genre, speakers’ self-monitoring, linguistic ideologies, etc. 3.10.

Acquiring grammatical constructions

Usage-based approaches show that the process of acquiring increasingly abstract constructions rests on interactional experience and general cognitive abilities (Tomasello 2003; 2006). According to Tomasello (1999, 2003; Tomasello et al. 2005), anthropological, social and cognitive prerequisites are the driving forces for the child’s construction of a language: the ascription of communicative (Gricean) intentions to interlocutors, the coordination of joint attention (Bruner 1983, Clark 2001), intersubjective cooperation (Tomasello 2008), and cognitive capabilities like categorization, inference and induction of schemata, statistical comparison and analogy, which are used for the identification of patterns of experience, action, and linguistic practice. In contrast to Chomsky’s assumption about the ‘poverty of the stimulus’, which is said to account for the need for a specialized language acquisition device, social interaction in terms of situated routines and modelling and correction by care-givers in child-directed speech have also been shown to be of major importance for language-learning (Cameron-Faulkner, Lieven and Tomasello 2003). Tomasello (2006: 258) questions the necessity to assume a languagespecific capacity, as Universal Grammar does, because children “acquire the more regular and rule-based constructions in the same way they acquire the more arbitrary and idiosyncratic constructions”. Instead of applying abstract class-specific rules to words according to their membership in syntactic categories (cf. Pinker 1999), knowledge about abstract syntactic structures emerges from lexical strings by applying universal cognitive mechanisms of learning with the aid of the specific pragmatic structures of context. The first schematic constructions to develop are

Pragmatics and grammar

449

two-word pivot constructions, where a lexically fixed pivot is combined with various other items. The next step are item-based constructions and constructional islands (Tomasello 2003). These are constructional frames associated with specific lexical items according to the frequency of their occurrence. They contain categorical slots for arguments and first syntactic markers (like word-order, morphosyntactic markers, prepositions). Children start to use and understand frequent verbs in the context of various constructional frames, while less frequent verbs are mastered only in a smaller number of constructional contexts (e.g., cut is only used with direct object, while draw is also used with additional locative, instrumental or beneficient prepositional phrases). So, syntactic rules do not seem to be categorically available in abstraction from lexicalization and context, but they are acquired only incrementally by a process of increasing schematization. This process rests on the discovery of analogies and variational patterns by comparison (see also Clark 1992 for the heuristics of conventionality and contrast). It lays the base for inferring slot-filling categories, which are increasingly abstracted and generalized to fully schematic, abstract constructions via induction. The discovery of analogies does not merely rest on perceptual features and cognitive operations over formal categories, but more importantly on functionally based distributions, which cooccur with recurrent patterns of action (like ‘object transfer’ or ‘caused motion’), and interactional participation in action sequences (like ‘referring and naming’, ‘question-answer’, narratives). 4.

Future prospects

As the paradigm of “armchair linguistics” gives way to empirical research, scholars tend to conceive of pragmatics in a more comprehensive way, including areas of research which were formerly assigned to text-, socio-, psycho- or discourse linguistics. Theoretical notions and methods originating from these areas are increasingly integrated in pragmatic theorizing. In the last decade, a variety of new empirical methods has emerged. They provoke changes in the view of the relationship between pragmatics and grammar and lead to new questions. – The increasing availability of large machine-readable (still overwhelmingly written) corpora and the use of procedures of data-mining allows for the discovery of preferential patterns of co-occurrence and of contextual conditions of use by statistical analyses (Stefanowitsch and Gries 2008). Such usage-based studies strongly favor probabilistic and context-sensitive models of grammar instead of universalistic rule-based approaches. – Video recordings of embodied multimodal interaction are prerequisites for understanding the fine-grained sequential and simultaneous coordination between grammar, prosody, gesture, facial expression, and other body movements in real time. They also help to elucidate how the roots of linguistic rou-

450

Arnulf Deppermann

tines may lie in gestural routines, both ontogentically and phylogenetically (Tomasello 2008). – Psycholinguistic experiments using parameters like latency of reaction are used to test hypotheses about the psychological reality of alleged relationships between (syntactically based) semantic vs. inferred meanings in terms of cognitive processing (Meibauer and Steinbach 2009). – Researchers have started to use physiological and neurobiological measures like event-related potentials (Kutas and Federmeier 2007) and brain-imaging (Bornkessel-Schlesewsky and Friederici 2007) as evidence for assumptions about mental processing and cerebral localization of phenomena located on the grammar-pragmatics interface. – The process of the acquisition of grammatical structures based on some basic cognitive capabilities and discursive experience has started to be modelled by learning robots, e.g., in the approach of Fluid Construction Grammar (Steels and de Beule 2006). Changes of the subject matter of linguistic pragmatics, which are due to sociocultural and technological developments, also call for revised models and raise new questions. New internet and communication technologies, which are creatively adopted by users, lay the grounds for a seemingly endless rush of new genres, multimodal configurations and modes of interactivity, in which new grammatical opportunities and routines in new communities of practice emerge. Processes of migration and globalization lead to revisions and re-formations of speech communities, including the emergence of new linguistic varieties and inter-cultures of mixed codes which may become conventionalized.

Notes 1. See already the seminal paper by Fillmore, Kay and O’Connor (1988). 2. Cf. ditransitive vs. to-dative (Goldberg 1995 and 2006) and argument omission (Goldberg 2005). 3. Ochs, Schegloff and Thompson (1996), Selting and Couper-Kuhlen (2001), Barth-Weingarten (2008). 4. Cf. e.g. Ono and Thompson (1995), Deppermann (2006, 2007), Günthner (2006) and Imo (2007). 5. Auer (1996a and 2007), Schegloff (1996), Ford, Fox and Thompson (2002), CouperKuhlen and Ono (2007a). 6. Crystal 2006; see also vol. 9 of the Handbook of Pragmatics handbook series: Pragmatics of Computer-mediated Communication, ed. by Susan C. Herring, Dieter Stein and Tuija Virtanen.

Pragmatics and grammar

451

References Aitchison, Jean 2001 Language change: Progress or decay? 3rd ed. Cambridge: Cambridge University Press. Ariel, Mira 1990 Accessing noun phrase antecedents. London: Routledge. Ariel, Mira 2008 Pragmatics and grammar. Cambridge: Cambridge University Press. Auer, Peter 1996a On the prosody and syntax of turn-continuations. In: Couper-Kuhlen, Elizabeth and Selting, Margret (eds.), Prosody in conversation. Interactional studies, 57–100. Cambridge: Cambridge University Press. Auer, Peter 1996b The pre-front field in spoken German as a grammaticalization position. Pragmatics 6(3): 295–322. Auer, Peter 1998 Zwischen Parataxe und Hypotaxe: ‘abhängige Hauptsätze’ im Gesprochenen und Geschriebenen Deutsch. Zeitschrift für Germanistische Linguistik 26(3): 284–307. Auer, Peter 2005 Projection in interaction and projection in grammar. Text 25(1): 7–36. Auer, Peter 2007 Why are increments such elusive objects? Pragmatics 17(4): 647–658. Auer, Peter 2009 On-line syntax: Thoughts on the temporality of spoken language. Language Sciences 31: 1–13. Auer, Peter and Stefan Pfänder (eds.) 2010 Emergent constructions. Berlin: de Gruyter. Barlow, Michael and Suzanne Kemmer (eds.) 2000 Usage-based models of language. Stanford CA: Center for the Study of Language and Information. Barth-Weingarten 2008 Interactional linguistics. In: Gerd Antos, Eija Ventola and Tilo Weber (eds.), Handbook of Applied Linguistics. Vol. 2: Interpersonal Communication, 77–106. Berlin: de Gruyter. Barth-Weingarten, Dagmar and Couper-Kuhlen, Elizabeth 2002 On the development of final though: A case of grammaticalization? In: Ilse Wischer and Gabriele Diewald (eds.), New reflections on grammaticalization, 345–361. Amsterdam: Benjamins. Behaghel, Otto 1909 Beziehungen zwischen Umfang und Reihenfolge von Satzgliedern. Indogermanische Forschungen 25: 110–142. Biber, Douglas, Stig Johansson and Geoffrey Leech 1999 Longman grammar of spoken and written English. London: Longman. Biber, Douglas 2006 University Language. Amsterdam: Benjamins.

452

Arnulf Deppermann

Bod, Rens, Jennifer Hay and Stefanie Jannedy (eds.) 2003 Probabilistic linguistics. Cambridge, MA: MIT Press. Bornkessel-Schlesewsky, Ina and Angela D. Friederici 2007 Neuroimaging studies of sentence and discourse comprehension. In: Gareth Gaskell (ed.), The Oxford handbook of psycholinguistics, 407–424. Oxford: Oxford University Press. Brennan, Susan E. and Herbert H. Clark 1996 Conceptual pacts and lexical choice in conversation. Journal of experimental psychology: Learning, memory, and cognition 22: 1482–1493. Bruner, Jerome 1983 Child’s talk: Learning to use language. New York: Norton. Bublitz, Wolfram, Uta Lenk, Uta and Eija Ventola (eds.) 1999 Coherence in spoken and written discourse. Amsterdam: Benjamins. Bühler, Karl 1934 Sprachtheorie. Jena: Gustav Fischer. Büring, Daniel 2007 Semantics, intonation and information structure. In: Gillian Ramchand and Charles Reiss (eds.), The Oxford handbook of linguistic interfaces, 445–474. Oxford: Oxford University Press. Bybee, Joan 2006 From usage to grammar: The mind’s response to repetition. In: Language 82(4): 711–733. Bybee, Joan 2007 Frequency and the organization of language. Oxford: Oxford University Press. Bybee, Joan and Paul J. Hopper (eds.) 2001 Frequency and the emergence of linguistic structure. Amsterdam: Benjamins. Cameron-Faulkner, Thea, Elena Lieven and Michael Tomasello 2003 A construction based analysis of child directed speech. In: Cognitive Science 27: 843–873. Chafe, Wallace 1994 Discourse, consciousness, and time. Chicago IL: University of Chicago Press. Chomsky, Noam 1965 Aspects of the theory of syntax. Cambridge, MA: MIT Press. Chomsky, Noam and Howard Lasnik. 1993 The theory of principles and parameters. In: Jacobs, Joachim, von Stechow, Arnim, Sternefeld, Wolfgang and Theo Vennemann (eds.), Syntax. Vol. 1, 506–569. Berlin: de Gruyter. Clark, Eve V. 1992 Conventionality and contrast: Pragmatic principles with lexical consequences. In: Eva Feder Kittay and Adrienne Lehrer (eds.), Frames, fields, and contrasts: New essays in semantic and lexical organization, 171–188. Hillsdale, NJ: Lawrence Erlbaum. Clark, Eve 2001 Grounding an attention in the acquisition of language. In: M. Andronis, C. Ball, H. Elston & S. Neuvel (eds.), Papers from the 37th meeting of Chicago Linguistic Society. Vol. 1, 95–116. Chicago: Chicago Linguistic Society.

Pragmatics and grammar

453

Clark, Eve 2004 Pragmatics and language acquisition. In: Lawrence Horn and Gregory Ward (eds.), Handbook of pragmatics, 562–577. Oxford: Blackwell. Clark, Herbert H. 1977 Bridging. In: Philip Johnson-Laird and Peter C. Wasow (eds.), Thinking, 411–420. Cambridge: Cambridge University Press. Clark, Herbert H. and Deanna Wilkes-Gibbs 1986 Referring as a collaborative process. Cognition 22: 1–39. Couper-Kuhlen, Elizabeth and Tsuyoshi Ono 2007a Turn continuation in cross-linguistic perspective. Special Issue of Pragmatics 17(4). Couper-Kuhlen, Elizabeth and Tsuyoshi Ono 2007b ‘Incrementing’ in conversation. A comparison of practices in English, German, and Japanese. Pragmatics 17(4): 513–552. Croft, William 2001 Radical construction grammar. Oxford: Oxford University Press. Croft, William 2008 Construction Grammar. In: Dirk Geeraerts and Hubert Cuyckens (eds.), Handbook of Cognitive Linguistics, 463–508. Oxford: Oxford University Press. Croft, William and A. Cruse 2004 Cognitive linguistics. 225–327. Cambridge: Cambridge University Press. Crystal, David 2006 Language and the Internet. 2nd ed. Cambridge: Cambridge University. Deppermann, Arnulf 2006 Construction Grammar – Eine Grammatik für die Interaktion? In: Arnulf Deppermann, Reinhard Fiehler and Thomas Spranz-Fogasy (eds.), Grammatik und Interaktion, 43–66. Radolfzell: Verlag für Gesprächsforschung. Available at: www.verlag-gespraechsforschung.de/2006/deppermann.htm. Deppermann, Arnulf 2007 Grammatik und Semantik aus gesprächsanalytischer Sicht. Berlin: de Gruyter. Deppermann, Arnulf 2010 Constructions vs. lexical items as sources of complex meanings. A comparative study of constructions with German verstehen. In: Peter Auer and Stefan Pfänder (eds.), Emergent constructions. Berlin: de Gruyter. Diewald, Gabriele 2006a Discourse particles and modal particles as grammatical elements. Approaches to discourse particles. In: Kerstin Fischer (ed.), Approaches to discourse particles, 403–425. Amsterdam: Elsevier. Diewald, Gabriele 2006b Context types in grammaticalization as constructions. In: Constructions HSV 1-9/2006 (Special Volume). Available at http://www.constructions-online.de/ articles/specvol1/686. Dik, Simon C. 1997 The Theory of Functional Grammar. 2 vols. Berlin: de Gruyter. Du Bois, John 2003 Discourse and Grammar. In: Michael Tomasello (ed.), The new psychology of language. Vol. 2, 47–87. Mahwah, NJ: Lawrence Erlbaum.

454

Arnulf Deppermann

Du Bois, John, Lorraine E. Kumpf and William J. Ashby (eds.) 2003 Preferred argument structure. Amsterdam: Benjamins. Edwards, Derek 1997 Discourse and cognition. London: Sage. Eggins, Suzanne 1994 An introduction into Systemic Functional Linguistics. London: Continuum. Ehlich, Konrad 1991 Funktional-pragmatische Kommunikationsanalyse. In: Dieter Flader (ed.), Verbale Interaktion, 127–143. Stuttgart: Metzler. Fillmore, Charles J. 1985 Frames and the semantics of understanding. Quaderni di Semantica 6(2): 222–254. Fillmore, Charles J., Paul Kay and Mary Catherine O’Connor 1988 Regularity and idiomaticity in grammatical constructions: The case of let alone. Language 64: 501–538. Feilke, Helmuth 1996 Sprache als soziale Gestalt. Frankfurt/Main: Suhrkamp. Feilke, Helmuth 2004 Kontext – Zeichen – Kompetenz. In: Kathrin Steyer (ed.), Wortverbindungen – mehr oder weniger fest, 41–64. Berlin: de Gruyter. Fischer, Kerstin (ed.) 2006 Approaches to discourse particles. Amsterdam: Elsevier. Ford, Cecilia E. 1993 Grammar in interaction. Adverbial clauses in American English conversations. Cambridge: Cambridge University Press. Ford, Cecilia E., Barbara A. Fox and Sandra A. Thompson 2002 Constituency and the grammar of turn increments. In: Cecilia E. Ford, Barbara A. Fox and Sandra A. Thompson (eds.), The language of turn and sequence, 14–38. Oxford: Oxford University Press. Ford, Cecilia E., Barbara A. Fox and Sandra A. Thompson 2003 Social interaction and grammar. In: Michael Tomasello (ed.), The New Psychology of Language. Vol. 2, 119–143. Mahwah NJ: Lawrence Erlbaum. Fox, Barbara A. 2007 Principles shaping grammatical practices: an exploration. Discourse Studies 9(3): 299–318. Fries, Norbert 1983 Syntaktische und semantische Studien zum frei verwendeten Infinitiv und zu verwandten Erscheinungen im Deutschen. Tübingen: Narr. Gardner, Rod 2001 When listeners talk: Response tokens and listener stance. Amsterdam: Benjamins. Givón, Talmy 1979 From discourse to syntax. Grammar as a processing strategy. In: Talmy Givón (ed.), Syntax and semantics. Vol. 12: Discourse and syntax, 81–112. London: Academic. Goldberg, Adele 1995 Constructions. A construction grammar approach to argument structure. Chicago IL: University of Chicago Press.

Pragmatics and grammar

455

Goldberg, Adele 1998 Patterns of experience in patterns of language. In: Michael Tomasello (ed.), The new psychology of language, 203–220. Mahwah, NJ: Lawrence Erlbaum. Goldberg, Adele 2004 Pragmatics and argument structure. In: Lawrence Horn and Gregory Ward (eds.), Handbook of pragmatics, 427–441. Oxford: Blackwell. Goldberg, Adele 2005 Argument realization. The role of constructions, lexical semantics and discourse factors. In: Jan-Ola Östman and Mirjam Fried (eds.), Construction grammar, 17–43. Amsterdam: Benjamins. Goldberg, Adele 2006 Constructions at work: the nature of generalization in language. Oxford: Oxford University Press. Goodwin, Charles 2003 Pointing as situated practice. In: Sotaro Kita (ed.), Pointing: Where language, culture and cognition meet. 217–41. Mahwah, NJ: Lawrence Erlbaum. Goodwin, Charles and Marjorie Harness Goodwin 1987 Concurrent operations on talk: Notes on the interactive organization of assessments. IPRA Papers on Pragmatics 1(1): 1–54. Green, Georgia M. 2004 Some interactions of pragmatics and grammar. In: Lawrence Horn and Gregory Ward (eds.), Handbook of pragmatics, 407–426. Oxford: Blackwell. Grice, H. Paul 1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and semantics. Vol. 3, 41–58. New York: Academic Press. Gundel, Jeanette and Thorstein Fretheim 2004 Topic and focus. In: Lawrence Horn and Gregory Ward (eds.), Handbook of pragmatics, 175–196. Oxford: Blackwell. Gundel, Jeanette, Nancy Hedberg and Ron Zacharsky 1993 Givenness, implicature, and the form of referring expressions in discourse. Language 69: 274–307. Günthner, Susanne 2006 Von Konstruktionen zu kommunikativen Gattungen. Die Relevanz sedimentierter Muster für die Ausführung kommunikativer Aufgaben. Deutsche Sprache 1(2), 173–190. Günthner, Susanne 2008 Projektorkonstruktionen im Gespräch: Pseudoclefts, die Sache ist-Konstruktionen und Extrapositionen mit es. Gesprächsforschung 9: 86–114. Available at: http://www.gespraechsforschung-ozs.de/heft2008/ga-guenthner.pdf. Günthner, Susanne and Paul J. Hopper 2010 Zeitlichkeit & sprachliche Strukturen: Pseudoclefts im Englischen und Deutschen. Gesprächsforschung 11: 1–28. Available at: www.gespraechsforschung-ozs.de/heft2010/ga-guenthner.pdf Haiman, John (ed.) 1985 Iconicity in syntax. Amsterdam: John Benjamins. Halliday, Michael A.K. 1985 An Introduction into Functional Grammar. London: Edward Arnold.

456

Arnulf Deppermann

Haspelmath, Martin 1997 From space to time. Temporal adverbials in the world’s languages. München: Lincom. Haspelmath, Martin 1999 Why is grammaticalization irreversible? In: Language 37: 1043–1068. Hayashi, Makoto 2003 Joint utterance construction in Japanese conversation. Amsterdam: Benjamins. Heine, Bernd 1997 Cognitive foundations of grammar. Oxford: Oxford University Press. Heine, Bernd, Ulrike Claudi and Friederike Hünnemeyer 1991 Grammaticalization: A conceptual framework. Chicago: University of Chicago Press. Hengeveld, Kees and J. Lachlan Mackenzie 2008 Functional Discourse Grammar. Oxford: Oxford University Press. Hoffmann, Ludger 2001 Einleitung. In: Ludger Hoffmann (ed.), Funktionale Syntax, 1–17. Berlin: de Gruyter. Hopper, Paul J. 1998 Emergent grammar. In: Michael Tomasello (ed.), The new psychology of language, 155–175. Mahwah, NJ: Erlbaum. Hopper, Paul J. 2004 The openness of grammatical constructions. Papers from the 40th regional meeting of the Chicago Linguistic Society 40: 239–256. Hopper, Paul J. and Elizabeth Closs Traugott 2003 Grammaticalization. 2nd ed. Cambridge: Cambridge University Press. Hopper, Paul J. and Sandra A. Thompson 1984 The discourse basis for lexical categories in Universal Grammar. In: Language 60: 703–752. Hopper, Paul and Sandra A. Thompson 2008 Projectability and clause combining in interaction. In: Ritva Laury (ed.), Crosslinguistic studies of clause combining, 99–123. Amsterdam: Benjamins. Imo, Wolfgang 2007 Konstruktionsgrammatik und Gesprochene Sprache-Forschung. Tübingen: Niemeyer. Jefferson, Gail 1973 A case of precision timing in ordinary conversation: Overlapped tag-positioned address terms in closing sequences. Semiotica 9: 47–96. Karkkäinen, Elise 2003 Epistemic stance in English conversation. Amsterdam: Benjamins. Kay, Paul 2004 Pragmatic aspects of grammatical constructions. In: Lawrence Horn and Gregory Ward (eds.), Handbook of pragmatics, 675–700. Oxford: Blackwell. Keller, Rudi 2004 On language change. London: Routledge. Kress, Gunter and Theo van Leeuwen 2001 Multimodal discourse. London: Edward Arnold.

Pragmatics and grammar

457

Kress, Gunter and Theo van Leeuwen 2006 Reading images: The grammar of visual design. London: Routledge. Krifka, M. 2007 Basic notions of information structure. In: Caroline Féry, Gisbert Fanselow and Manfred Krifka (eds.), Interdisciplinary studies on information structure (ISIS) 6. Working Papers of the SFB 632, 13–56. Potsdam: Universitätsverlag. Kutas, Marta and Kara D. Federmeier 2007 Event-related brain potential (ERP) studies of sentence processing. In: Gareth Gaskell (ed.), The Oxford handbook of psycholinguistics, 385–406. Oxford: Oxford University Press. Labov, William 2001 Principles of linguistic change. Vol. 2. Oxford: Blackwell. Lakoff, George 1987 Women, fire and dangerous things. What categories reveal about the mind. Chicago IL: University of Chicago Press. Lambrecht, Knud 1987 On the status of SVO sentences in French discourse. In: Russell S. Tomlin (ed.), Coherence and grounding in discourse, 217–262. Amsterdam: Benjamins. Langacker, Ronald W. 1987 Foundations of cognitive grammar. Vol. 1: Theoretical prerequisites. Stanford CA: Stanford University Press. Langacker, Ronald W. 2000 A dynamic usage-based model. In: Michael Barlow and Suzanne Kemmer (eds.), Usage-based models of language, 1–60. Stanford, CA: Center for the Study of Language and Information. Langacker, Ronald W. 2001 Discourse in cognitive grammar. Cognitive Linguistics 12(2): 143–188. Langacker, Ronald W. 2008 Cognitive grammar. Oxford: Oxford University Press. Lass, Roger 1987 Historical linguistics and language change. Cambridge: Cambridge University Press. LePage, Robert B. and Andrée Tabouret-Keller 1985 Acts of identity: Creole-based approaches to language and ethnicity. Cambridge: Cambridge University Press. Lerner, Gene 1991 On the syntax of sentences in progress. Language in Society 30: 441–458. Lerner, Gene 1996 On the semi-permeable character of grammatical units in conversation: Conditional entry into the turn space of another speaker. In: Elinor Ochs, Emmanuel A. Schegloff and Sandra A. Thompson (eds.), Interaction and grammar, 238– 276. Cambridge: Cambridge University Press. Lerner, Gene 2004 The place of linguistic resources in the organization of talk-in-interaction: Grammar as action in prompting a speaker to elaborate. Research on language and social interaction 37(2): 151–184.

458

Arnulf Deppermann

Liedtke, Frank 1998 Grammatik der Illokution. Tübingen: Narr. MacWhinney, Brian and Elizabeth Bates 1989 The crosslinguistic study of sentence processing. Cambridge: Cambridge University Press. Meibauer, Jörg and Markus Steinbach (eds.) 2009 Experimental-pragmatics/semantics. Amsterdam: Benjamins. Mulder, Jean and Sandra A. Thompson 2008 The grammaticalization of but as a final particle in English conversation. In: Ritva Laury (ed.), Crosslinguistic studies of clause combining, 179–204. Amsterdam: Benjamins. Newmeyer, Frederick J. 1998 Language form and language function. Cambridge MA: MIT Press. Newmeyer, Frederick J. 2003 Grammar is grammar and usage is usage. Language 79: 682–707. Norris, Sigrid 2004 Analyzing multimodal interaction. London: Routledge. Ochs, Elinor, Emmanuel A. Schegloff, and Sandra A Thompson (eds.) 1996 Interaction and grammar. Cambridge: Cambridge University Press. Ono, Tsuyoshi and Sandra A. Thompson 1995 What can conversation tell us about syntax? In: Philip W. Davis (ed.), Alternative linguistics, 213–271. Amsterdam: Benjamins. Pekarek-Doehler, Simona 2010 Clause-combining and the sequencing of actions: projector fragments in French conversation. In: Ritva Laury and Ryoko Suzuki (eds.), Subordination in conversation. Crosslinguistic analysis of form-function mappings, Amsterdam: Benjamins. Pickering, Martin J. and Simon C. Garrod 2004 Towards a mechanistic psychology of dialogue. Behavioral and Brain Sciences 27: 169–226. Pinker, Steven 1999 Words and rules. New York: Morrow. Prince, Ellen F. 1978 A comparison of WH-clefts and IT-clefts in discourse. In: Language 54: 883–906. Prince, Ellen F. 1981 Toward a taxonomy of given/new information. In: Peter Cole (ed.), Radical pragmatics, 223–255. New York: Academic Press. Raymond, Geoffrey 2004 The stand-alone “so” in ordinary conversation. In: Research on Language and Social Interaction 37(2): 185–218. Redder, Angelika 2008 Functional pragmatics. In: Gerd Antos and Eija Ventola (eds.), Handbook of applied linguistics. Vol 2. Interpersonal communication, 133–178. Berlin: de Gruyter. Sacks, Harvey and Emmanuel A. Schegloff 1979 Two preferences in the organization of reference to persons in conversation and their interaction. In: George Psathas (ed.), Everyday language, 15–21. New York: Irvington.

Pragmatics and grammar

459

Sacks, Harvey, Emmanuel A. Schegloff and Gail Jefferson 1974 A simplest systematics for the organisation of turn-taking in conversation. Language 50(4): 696–735. Schegloff, Emmanuel A. 1996 Turn organization: One intersection of grammar and interaction. In: Elinor Ochs, Emmanuel A. Schegloff and Sandra A. Thompson (eds.), Interaction and grammar, 52–133. Cambridge: Cambridge University Press. Schiffrin, Deborah 1987 Discourse markers. Cambridge: Cambridge University Press. Schlobinski, Peter (ed.) 2006 Von *hdl* bis *cul8er*. Sprache und Kommunikation in den neuen Medien. Mannheim: Duden. Schmid, Hans-Joerg 2001 ‘Presupposition can be a bluff’: How abstract nouns can be used as presupposition triggers. Journal of Pragmatics 33(10): 1529–1555. Schmitz, Ulrich 2010 Schrift an Bild im World Wide Web. In: Arnulf Deppermann and Angelika Linke (eds.), Sprache intermedial, 383–418. Berlin: de Gruyter. Searle, John R. 1969 Speech acts. Cambridge: Cambridge University Press. Searle, John R. and Daniel Vanderveken 1985 Foundations of illocutionary logic. Cambridge: Cambridge University Press. Selting, Margret 2000 The construction of units in conversational talk. Language in Society 29: 477–517. Selting, Margret and Elizabeth Couper-Kuhlen (eds.) 2001 Studies in interactional linguistics. Amsterdam: Benjamins. Sorjonen, Marja-Leena 2001 Responding in conversation: A study of response particles in Finnish. Amsterdam: Benjamins. Steels, Luc and Joachim de Beule 2006 A (very) brief introduction to Fluid Construction Grammar. Proceedings of the 3rd International Workshop on Scalable Natural Language (ScaNaLu06): 73–80. Available at: http://www3.isrl.illinois.edu/~junwang4/langev/localcopy/pdf/Steels06Intro_FCG.pdf. Stefanowitsch, Anatol and Stefan Th. Gries 2008 Corpora and grammar. In: Anke Lüdeling and Merja Kytö (eds.), Corpus linguistics. Vol. 2, 933–951. Berlin, New York: de Gruyter. Stöckl, Hartmut 2004 Die Sprache im Bild – Das Bild in der Sprache. Berlin: de Gruyter. Streeck, Jürgen 2002 Grammars, words, and embodied meanings. On the evolution and uses of so and like. Journal of Communication 52(3): 581–596. Sweetser, Eve E. 1990 From etymology to pragmatics. Cambridge: Cambridge University Press. Thompson, Sandra A. 2002 “Object complements” and conversation. Towards a realistic account. Studies in Language 26(1): 124–164.

460

Arnulf Deppermann

Thompson, Sandra A. and Elizabeth Couper-Kuhlen 2005 The clause as a locus of grammar and interaction. Discourse Studies 7(4–5): 481–505. Tomasello, Michael 1999 The cultural origins of human cognition. New York: Harvard University Press. Tomasello, Michael 2003 Constructing a language: A usage-based theory of language acquisition, 94–281. New York: Harvard University Press. Tomasello, Michael 2006 Acquiring linguistic constructions. In: Deanna Kuhn and Robert S. Siegler (eds.), Handbook of child psychology. Vol. 2, 255–291. New York: Wiley. Tomasello, Michael 2008 Origins of human communication. Cambridge MA: MIT Press. Tomasello, Michael, Malinda Carpenter, Josep Call, Tanya Behne and Henrike Moll 2005 Understanding and sharing intentions: The origins of cultural cognition. Behavioral and Brain Sciences 28: 675–691. Traugott, Elizabeth Closs 2004 Historical pragmatics. In: Lawrence Horn and Gregory Ward (eds.), Handbook of pragmatics, 538–561. Oxford: Blackwell. Uhmann, Susanne 1991 Fokusphonologie. Eine Analyse deutscher Intonationskonturen im Rahmen der nicht-linearen Phonologie. Tübingen: Niemeyer. Ward, Gregory and Betty Birner 2004 Information structure and non-canonical syntax. In: Lawrence Horn and Gregory Ward (eds.), Handbook of pragmatics, 153–174. Oxford: Blackwell. Wasow, Tom 2002 Postverbal behavior. Stanford: CSLI.

16.

Pragmatics and semantics John Saeed

1.

Introduction

In this article we focus on some relatively recent ideas and discussion about the relationship between pragmatics and semantics. As a result we do not seek to characterize either enterprise comprehensively, or symmetrically, even if such were possible in this limited space. Hence our discussion omits consideration of major concerns of both: for pragmatics, for example, of the social and cultural dimensions of language use. In particular speech-act theorists are underrepresented, despite the importance to pragmatics of their concern with social institutions, conventions and language activities.1 Characterizing the relationship between pragmatics and semantics is made difficult by the differences in views of these fields in different linguistic theories. Moreover, other disciplines are keen to explore the same territory. Philosophers Lepore and Smith, for example, in prefacing a handbook of the philosophy of language, state aims that many linguists would recognize: […] the philosophy of language now focuses on its primary concern: the nature of natural language and the extraordinary capacity of human beings to use it to express their thoughts about the world and other subject matters. The way language works, how specific linguistic devices function to achieve their effects, how we come to know these properties of expressions, and how we exploit them in our talk: all this is pursued by contemporary philosophers of language. (2006: viii)

Looking the other way across the disciplinary fence, linguists may profess scepticism about the philosophical method, alleging a dependence on introspection, a concentration on a few European languages and scant attention to other linguistic levels such as phonology and syntax. However, the views of Frege, Russell, the early Wittgenstein, Quine, Davidson and other philosophers have been of fundamental importance in many linguistic theories of semantics, in particular in pursuit of the idea that languages are, or can be profitably viewed as, formal systems. As will be clear from other contributions to this handbook, the ideas of the later Wittgenstein, Grice, Austin, Strawson and Searle have been fundamental in the development of pragmatics. More generally, the influence of philosophy on linguistic semantics can be seen in the search for a role for propositions in the representation of sentence meaning.2 The idea that semantics is concerned with representations is widespread, even in functional approaches, along with the associated idea that pragmatics is then concerned with interpersonal action.3 Problems with determining how semantic repre-

462

John Saeed

sentations relate to propositions have been crucial to the developing views of what role pragmatics has in describing meaning. In the second half of the last century it became clear to many scholars that if propositions are bearers of determinate truth or falsity then the literal meanings of sentences cannot be equivalent to propositions. Typically when uttered, sentences don’t provide enough information: they leave a gap that has to be bridged by the listener. Increasingly the listener’s interpretive activity has been characterized as pragmatic.

2.

Context and propositional meaning

Various types of gap have been identified. Perhaps the clearest case is deixis, such as the italicized deictic expressions (or indexicals): You’re here now. Such expressions of course crucially require contextual information to be accessed, and thus their linguistic meanings do not seem to be part of the content of the utterances that contain them. They form part of what Gumperz (1992) calls “contextualization cues”, which invoke the relevant contextual assumptions in the participants. Some writers, following Kaplan (1989a, 1989b) have identified a split in the meaning of such expressions between a context independent element, termed ‘character’ by Kaplan, and the context dependent meaning, which he termed ‘content’. The character determines how the context will determine the content: in other words how the intended reference is identified and thus contributes to determining what proposition is expressed. A further distinction is then made between two types of indexical: the first are ‘pure’ indexicals like I and yesterday, for which the contextual determination is thought to be straightforward or ‘automatic’, in the terms of Perry (2001) and which are independent of speaker’s intentions. The others (termed true demonstratives by Kaplan 1989a), like he and that (used deictically rather than anaphorically), require inferential work for the addressees to be able to identify the referent. This latter set is called discretionary by Perry (2001). This difference might be significant for setting the pragmatics-semantics boundary since the automatic reference identification could be assumed to be a semantic process, akin to literal meaning, while the discretionary determination of reference could be held to be pragmatic, involving inferences about speaker intentions (see King and Stanley 2005 for discussion). One problem is that the automatic set seems to shrink on examination. Traditionally English here and now have been considered ‘pure’ indexicals. However though we might want to say that here includes the position of the speaker and now the time of utterance, in fact the determination of the zones these refer to seems to depend on context-dependent inferences, sensitive to the speaker’s intent. Moreover, linguistic studies of richer deictic systems than we find in English have revealed referential ‘grids’ that involve both literal and inferential elements of meaning, as in Yup’ik (Eskimo-Aleut; Alaska):

Pragmatics and semantics

463

Yup’ik demonstratives (Anderson and Keenan 1985: 295) Extended Restricted Obscured man’a una ‘this (near SP)’ tamana tauna ‘that (near ADR)’ imna ‘the aforementioned one’ ukna ‘the one approaching the speaker’ augna ingna amna ‘the one going away from the speaker’ agna ikna akemna ‘the one across there’ qaugna kiugna qamna ‘the one inland, inside, upriver’ qagna keggna qakemna ‘the one outside’ un’a kan’a camna ‘the one below, toward river’ unegna ugna cakemna ‘the one downriver, by the exit’ paugna pingna pamna ‘the one up there, away from river’ pagna pikna pakemna ‘the one up above’

This system incorporates a semantic classification of the objects to which the demonstratives refer: ‘extended’ forms are for either large expanses of land or water, or objects that are lengthy or moving; ‘restricted’ applies to objects that are stationary, or moving within a confined area, and fairly small in extent, relatively near, and visible; and ‘obscured’ describes objects that are farther away and not clearly in sight. At the same time the demonstratives crucially rely on the position of the speaker and so are deictic (see Anderson and Keenan 1985 for details). Hanks (1992) argues that in many languages deictic expressions, whose core function is to identify the referent through different relations to elements of the speech act, are typically fused with other information about the category of referents, the speaker’s attitude or stance and the intended speech act. Another type of gap between sentences and propositions is caused by the need for contextual disambiguation. Ambiguity, which may be structural as in (1), involve quantifier scope as in (2), or be lexical as in (3), results in a sentence fitting two or more propositions and thus requires the audience to determine which ‘fit’ is more appropriate for the context: 1. He tracked the man with binoculars. 2. All reviewers watched one movie. 3. She held the note. The ambiguity in (1) depends on whether the prepositional phrase with binoculars is interpreted as modifying the verb track or the noun phrase the man. Both interpretations are possible but describe different situations. In (2) the ambiguity derives from interpretations of the relative scope of the quantifiers all and one: if all is given wide scope over one, then there may be as many films as reviewers; if one has scope over all then they all want to see the one film. While there have been attempts to determine such scope ambiguities by semantic or logical rules, other writers (for example, Saba and Corriveau 2001: 276) suggest that quantifier scope

464

John Saeed

interpretation must involve a pragmatic level. One type of evidence is the influence of real world knowledge. Thus while the sentence in (4) has the predicted scope ambiguity (One and the same doctor read all the books / Each book was read by a possibly different doctor), this typically doesn’t arise in (5) because of what we know about the world: 4. A doctor read every book. 5. A doctor lives in every village. Lexical ambiguity as in (3), where the noun note can of course be a musical sound or a written message and consequently different senses of held are invoked, arises where the same phonological shape is shared by two or more distinct semantic units (or lexemes). A distinction is often made between homonymy, like pump ‘device for moving liquid or gas’ and pump ‘shoe’ where the semantic units are clearly distinct, and polysemy, where different but related senses are involved, like fault in ‘Whose fault was the banking crisis?’ and ‘The San Andreas Fault doesn’t go through San Francisco’. The decisions about relatedness are, however, not without difficulty (Lyons 1977: 550–69). The question of how the multiple senses are accessed and selected in context has been an important research issue in the psychology of language (see Simpson 1984, 1994 for discussion). It has also proved to be a practical problem for natural language processing, for example in machine translation, and the investigation of the relevant contextual information for lexical disambiguation has been an important research strand in computational linguists, as described in Agirre and Edmonds (2006). Remaining with quantifiers it has been widely noted (e.g. Recanati 1989, 1993b, Bach 1994, Büring 1997) that in normal use the domain of a quantifier is typically underdetermined and needs to be contextually enriched. So for the second sentence in (6a) below, Recanati (2000: 67–8) suggests the representation in (6b), where the leftmost square bracket identifies the situation or context to which the utterance relates: 6. a. Berkeley is a nice place. There are bookstores and coffee shops everywhere. b. [Berkeley] > Although in this case the domain is mentioned in the previous sentence, in commonly encountered examples like (7) below, the relevant domain is left for the audience to supply from context. 7. Are you going to the party? Everyone will be there. If quantifier restriction is freely supplied by contextual information then many writers have concluded that the semantic content of the utterance underspecifies the intended proposition.4 In a similar vein, several writers, including Sadock (1984), Horn (1992), and Carston (1998), have argued about the extent of pragmatic enrichment needed to

Pragmatics and semantics

465

explain uses of cardinal numbers, which may in different contexts for a cardinal number x have the intended interpretations, exactly x as in (8) below, at least x in (9), or at most x in (10): 8. He had six children but one died so now he has five. 9. She needs to reach thirty metres to make the javelin final. 10. Undergraduates may borrow six books from the library. Another type of underspecification is shown by gradable adjectives (Kennedy 2007), for example, like tall, expensive, and fast, which require contextual ‘fill in’ to be interpreted. This is clear in implicit comparison uses like: 11. Isabel is tall. 12. Restaurants in Dublin are expensive. Participants have to access contextually determined standards to interpret such examples: Isabel might be tall for a seven year old child, for example. Context update models (Kyburg and Morreau 2000, Barker 2002) provide a presuppositional account of this process. A related phenomenon is loose talk, argued by Lasersohn (1999) to be distinct from vagueness. This is where speakers employ a pragmatically licensed degree of approximation in their use of words, even where the terms are quite sharply defined. Thus a speaker might in casual conversations say (13) below: 13. I’ve a three am. flight tomorrow. even if the flight is actually at 3.05 am. If the context is right, the inaccuracy is treated as pragmatically irrelevant. As pragmatic research has developed over the past decades more examples of such under-determinacy have been identified. Typical examples include possessive constructions as in (14); understood time references in (15); and the strengthening of and in conjoined subjects as in (16): 14. The conversation turned to John’s book. 15. I’ve had lunch. 16. John and Mary lifted the piano. In (14) the relationship described by the possessive pronouns is contextually determined: the book he wrote, the one he owns, the one he was assigned to review, etc. If (15) is a reply to an invitation, it communicates that the speaker has eaten lunch within a contextually appropriate period of time (e.g. that day). In (16) there may be a stereotypical or default interpretation that John and Mary lifted the piano together (Levinson 2000). The increasing role of contextual inference in the determination of the truthdeterminate propositional meaning expressed by utterances has been called pragmatic intrusion (Levinson 2000). The growing emphasis on the context sensitivity

466

John Saeed

of language raises fundamental challenges to philosophical approaches which base theories of meaning on theories of truth. The varying extent to which philosophers permit the intrusion of contextual information into semantics can be viewed as grades of contextualism (Recanati 2004, 2005) as discussed in Preyer and Peter (eds. 2005, 2007). Some philosophical defences against contextualism have been proposed, for example by Borg (2004) and Cappelen and Lepore (2005). These authors defend formal semantic approaches by proposing versions of semantic minimalism, where context dependence is viewed as limited to cases triggered by a limited set of linguistic forms. In Cappelen and Lepore (2005) this set of forms is reminiscent of Kaplan’s (1989a, 1989b) indexicals, mentioned earlier. Their account ‘pushes’ contextual variation into what they call speech act pluralism. In this view, a sentence without one of the context-sensitive elements always expresses the same proposition, regardless of context. However when uttered in context such a sentence may be expanded to express infinitely many propositions, only some of which are relevant to the context, and it is these expanded propositions that motivate truth-value judgments. In other words semantic content and speech act content can be different and, although the former is viewed as a proposition, it is the latter that we have truth condition intuitions about. The cost of this, from a linguistics perspective, is a weakening of the role of intuitions about semantic content, which might risk the empirical underpinning of the study of meaning. Other philosophers of course are not troubled by a separation of semantic representations from truth conditions and propositions. The philosopher John Searle adopts a contextualist position: We know without hesitation what are the truth conditions of ‘The fly is on the ceiling’, but not of ‘The cat is on the ceiling’, and this difference is not a matter of meaning, but a matter of how our factual background information enables us to apply the meanings of sentences. In general, one can say that in most cases a sentence only determines a set of truth conditions relative to a set of assumptions that are not realized in the semantic content of the sentence. (Searle 1979: 86)

Searle argues that background assumptions have to be incorporated before a determinate proposition can be communicated. In particular, in this view, an addressee cannot respond to the intended speech act, say an order, until this is done. Noam Chomsky, whose view of a semantic representation concentrates on an individual’s mind-internal mapping between grammar and conceptual structure, shows even greater scepticism about a role for truth conditional denotation in semantic analysis: “the argument for a reference-based semantics (apart from an internalist syntactic version) seems to me weak. It is possible that natural language has only syntax and pragmatics” (1995: 26). Chomsky’s view is that the meanings of sentences do not specify truth conditions even when relativized to context. This is reminiscent of Strawson’s (1950) view that although sentences can be used to assert things that are true or false, sen-

Pragmatics and semantics

467

tences themselves are neither true nor false. Pietroski (2005: 256) characterizes Chomsky’s position as follows: Rather the meaning of S is a compositionally determined intrinsic property of S that constrains and guides without determining how S can be used to make true or false assertions in various conversational situations … making truth-evaluable assertions is one of the things we can do with sentences, in contexts, though uses of this kind are highly variable; while people refer to things, words don’t; and sentence use may not be a theoretically tractable phenomenon.

Chomsky (1995) effectively proposes a break in the link assumed by many philosophers between the truth-conditional semantics of artificial languages and accounts of meaning for natural languages. In doing this he takes a use-based theory deep into traditional semantic territory and makes sense of the dramatic suggestion in the quotation above.

3.

Neo-Gricean inferential pragmatics

Within contemporary linguistics, debate about the relation between pragmatics and semantics has followed from Paul Grice’s (1975, 1978, 1989) influential distinction between ‘what is said’ in an utterance and what a speaker means by an utterance.5 This is often characterised (see for example Carston 2004) as a distinction between ‘what is said’ as propositional, or truth-conditional, content, and what is intentionally implied by the speaker’s utterance. As noted by several writers (e.g. Szabó 2006: 386), Grice never quite seems to give us a tight definition of this notion of ‘what is said’, identifying it as “closely related to the conventional meaning of the words (the sentence) he has uttered” (1989: 25). This leaves it somewhat open as to how much resolution of context-sensitive expressions, such as indexicals, is necessary to reach ‘what is said’. The distinction in any case depends on his notion of conversational implicature as implications that are not part of the conventional meaning of the utterance, are context-dependent and, crucially, intended by the speaker, as in a typical (invented) example (17): 17. A: Would you like another drink? B: I’d better go home; I’ve an early start in the morning. Here B’s reply may be interpreted to implicate a refusal or simply ‘no’. Grice’s work suggests a distinction between semantics and pragmatics: semantics is concerned with the proposition expressed while pragmatics is concerned with the implicatures of the utterance. His work opens up the prospect of an inferential pragmatics that explores the gap between the linguistic content of the speaker’s utterance, when viewed analytically, on the one hand, and what listeners seem to understand as speaker’s meaning. Pragmatic processing relies on

468

John Saeed

the predictability of inferential behaviour, seen by Grice as depending on co-operative principles underlying communication (Grice 1975, 1978). His cooperative principle and maxims below predict that the listener assumes, and the speaker relies on this, that the speaker is speaking truthfully, relevantly, and appropriately, that is performing subtasks of assessing what the hearer knows, selecting the right degree of clarity, etc. Clearly this is not a claim that speakers always behave in this way; rather it is a claim that inferential strategies seem to rely on such assumptions. Grice’s Co-operative Principle (see Grice 1989:26) ‘Make your contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged.’ Grice’s maxims (adapted from Grice 1975: 46) i. The Maxim of Quality Try to make your contribution one that is true, i.e. a) do not say what you believe is false b) do not say that for which you lack adequate evidence. ii. The Maxim of Quantity Make your contribution as informative as is required for the current purposes of the exchange (i.e. not more or less informative). iii The Maxim of Relevance Make your contributions relevant. iv. The Maxim of Manner Be perspicuous, and specifically: a) avoid ambiguity b) avoid obscurity c) be brief d) be orderly. His approach moreover promises an account of pragmatic processing. Grice proposed that conversational implicatures must be capable of being worked out, i.e. a reasoning process must be identifiable, even if it is not deduction. For every implicature one can predict a process of calculation involving ‘what is said’ (literal meaning), the cooperative principle and context. A characteristic Gricean formulation is: Working out schema for conversational implicatures (a) The speaker (S) has said that p. (b) There is no reason to think that S is not observing the maxims. (c) S could not be doing this unless he thought that q. (d) S knows (and knows that the hearer (H) knows that he knows) that H can see that he thinks that the supposition that he thinks q is required.

Pragmatics and semantics

469

(e) S has done nothing to stop H from thinking that q. (f) S intends H to think, or is at least willing to allow H to think, that q. (g) And so, S has implicated that q. (adapted from Grice 1975: 50) Grice used the term non-natural meaning (meaningnn) for interactional communication between mutually aware participants to distinguish it from simpler forms of signification. He provided an initial characterization of this as below:

Figure 1: Grice’s Meaningnn (Levinson 2000: 13)

As suggested above, the first split in Figure 1 above might be taken to coincide with the Gricean division between semantics and pragmatics, though Grice never puts it in these terms, since ‘what is said’ is presumably the truth-conditional content. However the notion of conventional implicature causes problems here since it is conventionally marked or coded but inferentially determined. A typical example is English but, which as shown below consistently carries a meaning of unexpectedness or contrast: 18. a. He’s a politician but he’s honest. b. Joan is tall but Libby is short. c. But you know I don’t drink coffee! Despite this degree of conventionality, the meaning of but is excluded from ‘what is said’ because it does not contribute to the truth conditions of the proposition expressed. It also requires contextual inferences to determine the precise interpretation. Subsequent work in this general framework has described such behaviour without this notion of conventional implicature and, as we shall see below, more generally has suggested different borders between semantics and pragmatics. The pragmatic inferences called conversational implicatures are claimed to have specific features. They are described, for example, as cancellable, non-detachable and non-conventional. They are called cancellable because, for example, we can add in (19) below another proposition to cancel the implicature without causing anomaly:

470

John Saeed

19. a. John has three children, in fact ten. b. John has three children, maybe more. This contrasts with other types of inference like presupposition in (20), and entailment in (21), where anomaly results: 20. ?The King of France is bald and/but there’s no King of France. 21. ?The president was assassinated yesterday but he’s not dead. Non-detachability is the claim that they are attached to a semantic level rather than to surface linguistic form. Thus by changing the specific linguistic form, say via synonyms, there is no change in the implicatures. Levinson (1983: 116–117) claims that if we know a context where (22a) implicates (22b) then (23a-d) will also produce (22b):6 22. (a) John’s a genius. (b) John is an idiot. 23. (a) (b) (c) (d)

John’s a mental prodigy. John’s an exceptionally clever human being. John’s an enormous intellect. John’s a big brain.

Finally non-conventionality captures the fact that the inference is not encoded in the linguistic form of the utterance, as is clear from examples like (17) earlier. To complete the Gricean taxonomy: particularized conversational implicatures (PCIs) depend completely on specific context of utterance: they are once-off inferences. Generalized conversational implicatures (GCIs) are context dependent but range over contexts, like a default. They only fail if blocked by inconsistency with context. Carston (2004), following Levinson (2000), provides the following example to illustrate the difference: 24. A: Did the children’s summer camp go well? B: Some of them got stomach ’flu. GCI: Not all the children got stomach ’flu. PCI: The summer camp didn’t go as well as hoped. This Gricean conception of pragmatics has inspired an extensive literature on the nature of both the maxims and the implicature themselves (cf. Huang this volume). Horn (1984, 1989, 2004) for example argues for a reworking of the Gricean maxims into two basic pragmatic principles, briefly outlined below: Horn’s Q- and R-principles (Horn 1984, 1989) (a) The Q-principle Make your contribution sufficient; Say as much as you can (given the R-principle)

Pragmatics and semantics

471

(b) The R-principle Make your contribution necessary; Say no more than you must (given the Q-principle) Horn’s Q-principle conflates Grice’s Quantity maxim and parts 1 and 2 of Grice’s Manner maxims and can basically be paraphrased as ‘say the most that you are licensed to say’. Alternatively, taking the hearer’s point of view in mind, it could be paraphrased as ‘maximize the informational content of what you say’. On the other hand the R-principle is a speaker-orientated economy principle. The two speaker versus hearer economy principles are held to be in tension. The best known application of the Q-principle is the Q- or Horn scales of strength, schematized below: Q-scale/Horn scale (Huang 2007: 38) For to form a Q- or Horn scale, (i) A(S) entails A(W) for some arbitrary sentence frame A; (ii) S and W are equally lexicalized, of the same word class, and from the same register; and (iii) S and W are ‘about’ the same semantic relation, or from the same semantic field. Thus Horn scales are typically scales of alternates, ordered strong to weak, for example: 25. (a) (b) The Q-maxim operating on scales ensures that the use of a weaker alternate implies the negative of a stronger: 26. a. ‘Not all the listeners enjoyed the show’ implicates ‘some (i.e. not none) did’ b. ‘Their results are similar’ implicates ‘Their results are not identical’ A similar but distinct revision is attempted by Levinson (1987a, 1987b, 1991, 2000), who, seeking to distinguish semantic content from surface form in relation to economy, proposes three principles: the Q-, I-, and M-principles. The Q-principle, which corresponds to Grice’s Maxim of Quantity 1, is given in simplified form below: Levinson’s Q-principle (simplified; Levinson 2000, Huang 2007) Speaker: Do not say less than is required (bearing in mind the I-principle) Addressee: What is not said is not the case. This principle, like Horn’s Q-principle, gives rise to scales. In addition to the simple scales as in (25) above, Levinson identifies more complex cases of alternate sets. We can briefly mention two. The first are clausal, where the alternates are dis-

472

John Saeed

tinguished by the stronger entailing its subparts and the weaker not entailing its components, as in (27): 27. Example of Levinson Q-clausal: conditionals a. , b. ‘Since he’s here, he can play’ entails ‘He is here’; and ‘He can play’ c. ‘If he’s here, he can play’ does not entail ‘He is here’, nor ‘He can play’ In this example, ‘since p then q’ entails p, and entails q, as shown in (27b). However, ‘if p then q’ does not entail p, and does not entail q, as shown in (27c). Because of this relationship, by uttering ‘If he’s here, he can play’ the speaker, by excluding the stronger, implicates: ‘Maybe he’s here; maybe not; maybe he can play; maybe not’ i.e. uncertainty. The second type is simple alternates, where the alternates are a non-entailing contrast set. These may be ordered in strength () as in (28) or an unordered taxonomy, like the colour terms in (29): 28. Q-alternate: 29. Hence from (29) a speaker uttering ‘His car is blue’ (weakly) implicates ‘His car isn’t silver, etc.’ Levinson’s I- and M-Principles concern economy. His I-principle, which corresponds to Grice’s Maxim of Quantity 2, is in simplified form: Levinson’s I-principle (simplified; Levinson 2000, Huang 2007) Speaker: Do not say more than is required (bearing in mind the Q-principle) Addressee: What is generally said is stereotypically and specifically exemplified. This can be paraphrased as ‘the use of a semantically general expression implicates a semantically specific interpretation’. This principle is used to explain how speakers employ or assume defaults and stereotypes, like those in (30): 30. A cheese board = for cutting cheese not made of cheese A kitchen knife = used in kitchens not for cutting kitchens It is also used for the examples of strengthening implications mentioned earlier, for example the occasion when conjunctions like English and can imply temporal and causal sequence as in (31); when if is interpreted as ‘if and only if’ as in (32); and the strengthening of negation as when (33a) implicates (33b): 31. “It’s always the same at parties: either you get drunk and no one will talk to you or no one will talk to you and you get drunk” (Blakemore 1992: 80) 32. If you pass your examinations, I’ll buy you a car. 33. a. Susana doesn’t like the Irish weather. b. Susana dislikes the Irish weather.

Pragmatics and semantics

473

Levinson’s M-Principle relates to the form of utterances, and can be characterized as: Levinson’s M-principle (simplified; Levinson 2000, Huang 2007) Speaker: Do not use a marked expression without reason. Addressee: What is said in a marked way is not unmarked. This principle explains speakers’ use of unusual (marked) expressions to avoid defaults and stereotypes. It seeks to explain the pragmatic effects of negative asymmetry, where (34a) below will be interpreted differently from (34b); or where periphrasis carries distinctions of meaning as in (35): 34. a. I don’t dislike Henry. b. I like Henry. 35. a. Bob opened the door. b. Bob caused the door to open. Once again these principles are viewed as operating in tension; the interaction between them is a complicated and controversial issue: see Levinson (2000) and Huang (2007) for discussion. The work of neo-Gricean writers like Horn and Levinson has shed light on a wide range of linguistic expressions, seeking to determine the role of semantic encoding and pragmatic inference in their interpretation. An important claim of Levinson (2000) is that generalized conversational implicatures represent a form of conventional language use, characterized as default inferences, which can be over-ruled in context. In this view, these linguistically-triggered default inferences contribute to propositional content. This then changes the Gricean relation between semantics and pragmatics. Since pragmatic inferences contribute to propositional content, then clearly if there is purely semantic representation it is pre- or sub-propositional. Or, to put it another way: both semantic interpretation and pragmatic inference combine in the formation of propositional truth-bearing content.

4.

Relevance theory

This pragmatic intrusion, to use Levinson’s term, is even more marked in relevance theory (Sperber and Wilson 1995, Carston 2002), which is a more radical development from Gricean inferential pragmatics. Relevance theory is an approach to communication based on the view that people are predisposed to pay attention to phenomena in their environment when doing so is likely to bring about improvements in their belief system. In other words, we tend to pay attention to stimuli which we expect will turn out to be relevant to us. Some stimuli, which include

474

John Saeed

pointing gestures and utterances, are designed to create an expectation that paying attention to them will yield significant cognitive and other rewards. In other words, they are designed to create expectations of their own relevance. This type of stimuli is called in this theory ostensive stimuli and the kind of communication which involves the production and interpretation of ostensive stimuli is called ostensiveinferential communication. In this view a phenomenon is relevant to an individual (a) to the extent that the cognitive effects achieved when it is processed in context are large, and (b) to the extent that the processing effort required for achieving the effects is small (Sperber and Wilson 1995: 153). This definition of relevance provides the basis for two principles of human cognition and communication, known as the cognitive and the communicative principles of relevance: The Cognitive Principle of Relevance Human cognition tends to be geared to the maximization of relevance. (Sperber and Wilson 1995: 260) The Communicative Principle of Relevance Every act of ostensive communication communicates a presumption of its own optimal relevance. (Sperber and Wilson 1995: 260) The presumption can be described as follows: Presumption of Optimal Relevance (a) The ostensive stimulus is relevant enough for it to be worth the addressee’s while to process it. (b) The ostensive stimulus is the most relevant one compatible with the communicator’s abilities and preferences. (Sperber and Wilson 1995: 270) The influence of Grice is clear here, with its assumption that the cost, so to speak, of inference can only be motivated by a shared presumption of certain strategies on both the speaker’s and hearer’s behalves. Consequently, in this account, there are certain assumptions about economy, as described below by Carston (2002): Relevance-theoretic Comprehension Strategy (a) construct interpretations in order of accessibility (i.e. follow a path of least effort); (b) stop when your expectation of relevance is satisfied. (Carston 2002: 380) Of course expectations of relevance may vary across different situations of utterance and in practice adult hearers are expected to have developed strategies to account for speakers’ fallibility or deceptiveness; and other type of communicative

Pragmatics and semantics

475

noise. Giving a schematic view of the sub-tasks in the comprehension process, Wilson and Sperber summarize these as below: Sub-tasks in the comprehension process (Wilson and Sperber 2004: 615). a. Constructing an appropriate hypothesis about explicit content (in relevancetheoretic terms, explicatures) via decoding, disambiguation, reference resolution, and other pragmatic enrichment processes. b. Constructing an appropriate hypothesis about the intended contextual assumptions (in relevance-theoretic terms, implicated premises). c. Constructing an appropriate hypothesis about the intended contextual implications (in relevance-theoretic terms, implicated conclusions) From a relevance theory point of view, the processes of recovery of explicatures and implicatures are both equally inferential and interact freely. In distinction to Levinson’s (2000) position, relevance theorists argue that the pragmatic processes leading to explicatures (i.e. explicit propositional content) are not restricted to special linguistic constructions, nor are they implicatures. Like all pragmatic processes though, they are motivated by the same presumption of optimal relevance. As we can see, the tasks in constructing a proposition involve context-dependent inference. In this way relevance theory reflects the influence of context in the most basic of what has sometimes been seen as coded semantic meaning. Blakemore (2002) suggests that ostensive stimuli may contribute to the comprehension process in three main ways. First, they may focus the addressee’s attention in a particular direction (e.g. pointing gestures and acts of displaying in general). For example, a speaker might ask where the bus stop is and the addressee might point to its location. In this case, the communicator, by displaying an object to the addressee, suggests that it is relevant as evidence of his/her informative intention. Secondly, verbal ostensive stimuli typically encode conceptual representations which contribute to the truth-conditional content of the utterance directly. For example, instead of pointing the addressee in this scene might say ‘It’s across the road’. In this case, the words used present some evidence of what the speaker intends to communicate by virtue of the concepts that they encode. Thirdly, lexical items may encode procedures (i.e. constraints on or indications of the way pragmatic inference should proceed). Blakemore (2002) discusses the English examples in (36–38): 36. Tom can open the safe. He knows the combination. 37. Tom can open the safe. So he knows the combination. 38. Tom can open the safe. After all he knows the combination.

476

John Saeed

Example (36) expresses two propositions. While the explicit content is unambiguous, the implicatures that can be retrieved are not. How the implicit content can be constrained is exemplified in (37) and (38). In these examples, the discourse markers so and after all guide the computational route that needs to be taken by the hearer. In (37) the second statement is a conclusion drawn from the first. If someone can open the safe then he knows the combination. ‘So’ encodes a constraint on the way the utterance it introduces should be interpreted in context. In ‘So p’, p is relevant as a conclusion (which follows from some salient premises). In (38) the second proposition strengthens the existing assumption that Tom can open the safe. A pronoun like ‘he’ in these examples encodes some procedural information which helps the hearer to form a conceptual representation of the intended referent. So, there are two types of procedural constraints: constraints on the derivation of implicatures (e.g. ‘so’, ‘after all’, ‘however’) and constraints on pragmatic inferences involved in deriving the propositional content of the utterance (e.g. pronouns, tense markers). We can give a short example from Carston (2006) of how this approach applies to a hypothetical conversational exchange in (39): 39. A conversational exchange (Carston 2006: 39) Bob: How is Jane feeling after her first year at university? Sue: She didn’t get enough units and can’t continue. 40. A possible interpretation of Sue’s reply (Carston 2006: 39) a. [JANE 1 DIDN ’ T PASS ENOUGH UNIVERSITY COURSE CREDITS TO QUALIFY FOR ADMISSION TO SECOND YEAR STUDY ]p & AS A RESULT OF P JANE 1 CANNOT CONTINUE WITH UNIVERSITY STUDY. JANE IS NOT FEELING HAPPY. b. The interpretation in (40) contains two different types of propositional results. The first, (40a) is an expanded or explicit version of the linguistically encoded material. So although, as Carston notes, the word she doesn’t conventionally contain the meaning Jane or the word units contain an element of meaning university, it seems that the linguistic items can be a trigger to identify the fuller version, by a process of working out or inference. The second interpretation is somewhat different. It provides an answer to Bob’s question, but has to be seen as providing it indirectly since the proposition can’t be seen as directly connected to the linguistic content of Sue’s answer. Though both parts of the interpretation in (40) seem to depend on context they seem to do it in different ways. We can easily imagine a scenario where (40a) is the same, but where Jane is quite happy not continuing her studies. Perhaps she was forced to attend by her parents and this gives her the excuse to leave. Thus the indirect answer would be the opposite. Thus the pragmatic processes in (40a) involve expanding the linguistic content while those in (40b) involve identifying intended implications of the expanded linguistic content. Both however are seen as pragmatic processes.

Pragmatics and semantics

477

This relevance theory view of the underspecification of encoded meaning is clearly shown in their work on lexical meaning. The contextual variability of word meaning is viewed as access to lexical concepts governed by relevance-based contextual inferences: Quite generally the occurrence of a word in an utterance provides a piece of evidence, a pointer to a concept involved in the speaker’s meaning. It may so happen that the intended concept is the very one encoded by the word, which is therefore used in its strictly literal sense. However, we would argue that this is no more than a possibility, not a preferred or default interpretation. Any interpretation, whether literal or not, results from mutual adjustment of the explicit and implicit content of the utterance. (Sperber and Wilson 1997:15)

The contextual variability of word meaning is seen as a product of a number of inferential processes, including narrowing, as in Wilson and Carston’s (2006:409) example: 41. All politicians drink. Here the meaning of drink has been narrowed from the meaning ‘drink liquid’ to mean ‘drink alcohol’ and possibly further, to ‘drink alcohol in sufficient quantities to be worth commenting on.’ In relevance theory terms, drinking liquid is clearly necessary for staying alive and in most contexts the fact that a person drinks liquid has no obvious relevance. Thus in many contexts the expected interpretation of (42) below: 42. Gordon doesn’t drink. is that Gordon doesn’t drink alcohol. Other processes include broadening or approximation (Wilson and Sperber 2004) where the sense of word is extended, as the word square in: 43. His father has a square face too. In this example it is unlikely that the speaker intends the usual dictionary meaning of square. In this view, lexical narrowing and broadening are types of pragmatic enrichment. Relevance theorists treat such processes as general in lexical interpretation. Wilson and Carston (2007) quote Murphy’s (1997) experimental work on evidence for distinct word senses by respondents’ supplied antonyms that contribute to explicatures. For the English word fresh when used with different nouns these include those in Table 1:

478

John Saeed

Table 1.

Antonyms of fresh (Murphy 1997:237–9)

fresh

ANTONYMS

shirt

dirty

vegetables

rotten

fish

frozen

sheets

recently slept-in

water

dirty/salt

bread

stale

air

polluted

outlook

tired

assistant

experienced

idea

old

They argue that this is a typical example where a general concept, here FRESH , gets ‘tailored’ to individual contexts. Wilson and Carston (2007) argue that a whole range of what have been considered individual lexical processes, narrowing, broadening, hyperbole (as in 44 below), metaphor (45), category extension (46), etc. can be given a unified account as inferential processes constrained by the encyclopaedic entries for concepts and the expectations of relevance. In particular they suggest that this involves a process of ad hoc concept construction, a once-off contextual interpretation, conventionally shown as starred concepts, i.e. FRESH * for the last example. One important aspect of this approach is the claim that processes like broadening, hyperbole and metaphor are outcomes of essentially the same pragmatic processes, by different ‘balancing’ of encyclopaedic information, contextual assumptions and derived contextual implications. Thus, their example (47) below could be intended and interpreted as literal, approximation, hyperbole or metaphor, depending on the contextual features (Wilson and Carston 2007: 248–9): 44. 45. 46. 47.

5.

Nobody walks in Los Angeles. Manchester United clawed their way back into the game. New Orleans is the new Hollywood. The water is boiling.

Cognitive semantics

Aspects of this relevance theory approach to lexical pragmatics suggest parallels to that of cognitive semantics (Lakoff 1987, Langacker 1987, 1991, Talmy 2000 inter

Pragmatics and semantics

479

alia). In cognitive semantics we also see a rejection of the view that literal encoded meaning has a privileged position, and of the view that hyperbole, metonymy, metaphor, etc. are special rhetorical devices or tropes. We also see a search for common general principles of sense determination and extension in context; and an emphasis on the role of encyclopaedic knowledge in accessing concepts. However, cognitive semanticists use these features of lexical meaning as an argument against identifying a clearly-defined semantics/pragmatics distinction at all. Langacker (1987:154), for examples, states: “The distinction between semantics and pragmatics (or between linguistic and extralinguistic knowledge) is largely artefactual, and the only viable conception of linguistic semantics is one that avoids such false dichotomies and is consequently encyclopedic in nature.” Langacker argues that it is impossible to distinguish between lexical and encyclopedic meaning, and, in common with other cognitive linguists, views the principles involved in organizing concepts and lexically accessing them as part of more general cognitive processes, rather than specialized linguistic operations. However the similarities with the relevance theory position are clear. Cognitive semanticists argue that the conventional meaning associated with a particular linguistic unit acts as a trigger for a process of meaning construction which involves the determination of an appropriate interpretation for the context of the utterance. Fauconnier and Turner (2002), for example, discuss interpretation of the English word safe. They provide some examples of its range of meanings in the context of a child playing on a beach: 48. a. The child is safe. b. The beach is safe. c. The shovel is safe. As they note, in this context, the interpretation of (48a) is that the child will not come to any harm. On the other hand, (48b) is not likely in this context to mean that the beach will not come to harm, but rather that the beach is an environment in which the risk of the child coming to harm is small. Finally, (48c) is not likely to mean that the shovel will not come to harm, but that it will not cause harm to the child. Their point is that there is no single fixed property that safe assigns to the words child, beach and shovel. In order to understand what the speaker means, the addressee employs encyclopaedic knowledge of children, beaches and shovels, and knowledge relating to what it means to be safe. We can see a parallel with Wilson and Carton’s (2007) picture of participants constructing a meaning by contextual adjustment based on encylopaedic knowledge. Although there are individual differences among writers within this approach, in general cognitive semanticists place no emphasis on a distinction between semantics and pragmatics, seeking instead to identify a range of cognitively motivated representations and computations that straddle the traditional border. An important constituent of this approach is of course the rejection of truth conditions,

480

John Saeed

or correspondences to states of affairs, as a way of defining meaning. Lakoff (1988:121) adopts a notion of experiential cognition, contrasting it with its ‘objectivist’ counterpart: –

Where objectivist cognition views human thought as fundamentally disembodied, experientialist cognition sees human thought as essentially involving the kind of structured experience that comes from having human bodies, especially from innate human sensory-motor capacities. – Where objectivist cognition sees meaning in terms of a ‘correspondence theory’, as the association of symbols with external objects, experientialist cognition sees meaning as essentially involving an imaginative projection, using mechanisms of schematization, categorization, metaphor and metonymy to move away from what we experience in a structured way with our bodies to abstract cognitive models. – Where objectivist cognition sees thought processes as the manipulation of abstract symbols by a great many highly-structured algorithms, experientialist cognition posits a small number of general cognitive processes whose application to abstract highly-structured cognitive models constitutes reason. Jackendoff (1990) makes a similar point, though couching his view in more Chomskyan terms: Hence a truth-conditional semantics in the Tarskian or Davidsonian sense requires a theory of E-language, of language as an abstract artefact extrinsic to speakers … within a theory of I-semantics, a statement in the Tarskian vein like ‘Sentence S in Language L is true if and only if condition C is met’ is taken as shorthand for something like ‘Speakers of Language L treat Sentence S as true if and only if their construal of the world meets condition C’, and it is subject to similar caveats about attentional and processing limitations. This is the basis of the approach of Conceptual Semantics, in which a level of mental representation called conceptual structure is seen as the form in which speakers encode their construal of the world. (Jackendoff 1990: 12)

This stance is reflected in the cognitive semantic approach to metonymy and metaphor. Though they are distinct processes for cognitive semanticists they both rely on structured fields of real world knowledge, identified as conceptual or cognitive models. In this view metonymy is a referential strategy that is based on links within the same cognitive model while metaphor is a mapping across cognitive models. Metaphor is essentially a way of understanding one conceptual model by relating it to another, usually more concrete and familiar. There have been a number of proposals for such cognitive models including scripts (Schank and Abelson 1977), frames (Fillmore 1982, 1985), and idealized cognitive models (Lakoff 1987). The important fact for this literature is that these cognitive models are bodies of real world knowledge, often culturally specific. Metaphor is then simply a special type of the more general cognitive process of analogical mapping. Fauconnier (1997: 19) uses the example of the coining of the term computer virus for a specific

Pragmatics and semantics

481

type of harmful program, which he suggests is based on a conceptual model of biological viruses which is generalized or schematized away from the biological details: Biological virus schema (Fauconnier 1997: 19) x is present, but unwanted; it comes in, or is put in, from the outside; it does not naturally belong; – x is able to replicate; new tokens of x appear that have the same undesirable properties as the original x; – x disrupts the “standard” function of the system; – the system should be protected against x; this might be achieved if the system were such that x could not come into it, or if other elements were added to the system that would counteract the effects of x, or eject x, or destroy x –

This schema is transferred to the general aspects of the computer situation; it provides a way of characterizing the new domain. The schema is itself based on lower level conceptual schemas called image schemas, especially characterizing path and force dynamics: entry, resistance etc. This analogical mapping between a health schema and a computer domain licenses a whole system of lexical innovations so that files can be said to be infected and can be placed in quarantine. The conclusions cognitive linguists draw from such analyses include the following: (a) that the supposedly linguistic process is actually part of a more general cognitive process; and (b) that such processes are based on real-world experience and the resulting conceptual models. These two points may be used as arguments against what are characterized as assumptions of formal approaches to language: that language is different from other cognitive processes and that linguistic knowledge is separable from general knowledge. We can take as an example of the resulting blurring of traditional semantic and pragmatic processes Fauconnier’s (1994, 1997) work on reference. Fauconnier has in his theory of mental spaces sought to integrate traditional topics such as referential opacity, modality, presupposition projection, counterfactuals, time deixis, and indirect reference, together with metaphorical uses of language. Fauconnier suggests that speakers and hearers construct mental models of situations and it is through these mental models and operations upon them that reference is possible. These mental spaces in effect can be seen as a cognitive version of possible world semantics (Lewis 1973), with the difference that more linguistic behavior is brought under the analysis. For Fauconnier the creation of discourse consists of the creation and manipulation of series of mental spaces in which individuals and situations are identified, and between which connections are made, as negotiated by the participants. He identifies linguistic markers, which he calls spacebuilders, as invitations to the hearers to construct mental spaces and the relevant referential links:

482

John Saeed

49. spacebuilders (Fauconnier 1994): adverbials of location and time like in your new novel, in Peter’s painting, when she was a child, after we find the crash site, etc.; modals like possibly and really; connectives like if … then; etc. As well as these overt markers, spacebuilders, the use of mental spaces is also evident in the normal use of names, which can be shifted. As in the following example where the same name can be used for a real, historical person; a character in a film; and the actor playing the character: 50. (a) Truman Capote wrote In Cold Blood. (b) In the film Capote, Truman Capote develops a close relationship with Perry Smith, one of the killers. (c) I didn’t like the film: Capote was too handsome. (i.e. the actor Philip Hoffman) This shifting of referential expressions like names occurs naturally with deictic pronouns of course: 51. If I were you, I’d sell those shares now. Such connections can be built up by a number of devices, the creation of fictional characters, by representation in images, etc. Thus a speaker might say, looking at a television screen: 52. Ronaldo looks a bit out of focus. Here, in Fauconnier’s terminology, the name of the Ronaldo is the trigger and the target is the image. Real Ronaldo (a) and screen Ronaldo (b) are linked by the pragmatic function IMAGE , so a description of real Ronaldo (his name, da) can be used to identify his television image (b). In other words, we assume that a speaker can use a person’s name to refer to an image of that person. This is clearly related to metonymy: here an electronic pattern created within a machine is given the name of a person whose image it relays. This relates to the traditional view that metonymy derives from a notion of contiguity or connectedness. The approach extends naturally of course to more traditional cases: 53. a. The blue suit at the end of the bar is about to throw up. b. The trains are on strike. Fauconnier uses a very general statement to describe all of this referential behaviour: The Identification (ID) Principle (Fauconnier 1994:3) If two objects (in the most general sense), a and b, are linked by a pragmatic function F (b = F(a)), a description of a, da, may be used to identify its counterpart b.

Pragmatics and semantics

483

The important points here are that this is a pragmatic process and that any link is possible: the versions of connectedness can be very idiosyncratic and culturally specific, though some of course, like the use with images in paintings, photographs, film and video, quickly become conventionalised across cultures sharing the technology. We might note once again that one major implication of this is that it overtly recognises that reference depends on conceptual systems, including cultural knowledge, contextual knowledge etc. The cognitive processes freely access all kinds of information. Fauconnier has developed this notion of mental spaces further into conceptual integration theory (Fauconnier and Turner 2002). In this theory knowledge from multiple mental spaces can be blended to form a new space, forming an emergent meaning structure that is distinct from the information provided by each input space. This model seeks to represent speakers’ abilities to create blended metaphors and scenarios, as in the following newspaper example: 54. On Wednesday, he faces a fast-paced tango with Gordon Brown during the pre-Budget report debate.7 Here we see the blend integrating cultural knowledge about British politics and ballroom dancing. Finally, an important element in the cognitive linguistics rejection of a clear distinction between pragmatic and semantics is the notion of construal (Langacker 1987), which identifies the dynamic processes by which the speaker conceptualizes a situation for communicative purposes. This involves, for example, a speaker conceptualizing entities at various levels of specificity, selecting a perspective on a scene, relating it to background assumptions and expectations and profiling some entities as more prominent than others. In this view the speaker imposes a structure on a scene, which is conditioned by and in turn conditions the linguistic form, by selecting grammatical structures, lexical items, etc. This process is also in part governed by the communicative context, including the speaker’s communicative goals. This approach therefore dissolves any hard and fast distinction between representation and communicative acts since construal works at all levels of the grammar and lexicon and involves access to encylopaedic and contextual knowledge.

6.

Conclusion

From our discussion it is clear that we can conclude with no neat characterization of the distinction between pragmatic and semantics. There is a wide range of positions in the literature. We can however suggest some trends. More types of information and processes involved in the interpretation of utterances have come to be viewed as extra-linguistic, and therefore seen by scholars as pragmatic. For the

484

John Saeed

many scholars who relate sentence meaning to truth conditions or propositions, this can be seen as an encroachment, or intrusion to use Levinson’s term, of pragmatics into semantics. From this perspective, the task is to establish what information about meaning is encoded in the linguistic material uttered, that is, by the grammatical rules of the language. Or to put it another way: what is the nature of the underspecified semantic representations? Recanati (2005) has characterized this shift amongst philosophers of language as a movement from literalism to contextualism, with many intervening positions. Within linguistics some pragmaticists, such as neo-Griceans and relevance theorists, without jettisoning the importance of truth conditions to meaning, have increasingly seen the notion as orthogonal to the distinction between pragmatics and semantics. Relevance theorists in particular view the difference in terms of a computational distinction between linguistic decoding and inference. For others, like Chomsky (1995), who reject the relation to states of the world as integral to the semantic description of natural languages, the semantic representations are not of course underspecified. However here the question of how linguistic rules and representations map to conceptual structures is part of semantics while a correspondingly large role is given to pragmatics in the derivation of contextually appropriate interpretations. Finally, cognitive semanticists, as we have seen, see any clear distinction between pragmatics and semantics as an artificial construct deriving from a mistaken distinction between expressing information and managing communication.

Notes 1. But cf. Collavin this volume. 2. Here we leave aside those writers, for example Montague (1974), who propose a formal logical approach to meaning which treats natural languages like artificial logical languages and identifies truth relations between linguistic forms and extensions, with no intervening psychological entities. 3. See for example Stalnaker’s definition: “Pragmatics is the study of linguistic acts and the contexts in which they are performed” (1999: 43). 4. See Stanley and Szabó (2000) and Stanley (2002) for discussion of domain restriction. 5. For the following see also Huang this volume. 6. It is not certain though that this is a feature of Manner implicatures, where effects are produced by selecting longer or shorter equivalent descriptions. 7. ‘Vincent Cable: The dance-crazy politician aiming to match Gordon Brown step for step’, Marie Woolf, Independent Newspaper, Monday, 8 December 2003.

Pragmatics and semantics

485

References Agirre, Enoko and Philip Edmonds (eds.) 2006 Word Sense Disambiguation: Algorithms and Applications. Dordrecht: Springer. Almog, Joseph, John Perry, and Howard Wettstein (eds.) 1989 Themes from Kaplan. Oxford: Oxford University Press. Anderson, Stephen R. and Edward L. Keenan 1985 Deixis. In: Timothy Shopen (ed.), Language Typology and Syntactic Description. Vol. III: Grammatical Categories and the Lexicon, 259–308. Cambridge: Cambridge University Press. Atlas, Jay David 1979 How linguistics matters to philosophy: presupposition, truth, and meaning. In: David Dinneen and Choon-Kyo Oh (eds.) Syntax and Semantics 11: Presupposition, 265–281. New York: Academic Press. Atlas, Jay David 2005 Logic, Meaning and Conversation: Semantical Underdeterminacy, Implicature, and their Interface. Oxford: Oxford University Press. Austin, John Langshaw. 1975 How to Do Things with Words. 2nd edition. Oxford: Clarendon Press. (First published 1962). Bach, Kent 1994 Conversational impliciture. Mind and Language 9: 124–62. Barker, Chris 2002 The dynamics of vagueness. Linguistics and Philosophy 25.1: 1–36. Blakemore, Diane 1992 Understanding Utterances: an introduction to pragmatics. Oxford: Blackwell. Blakemore, Diane 2002 Linguistic Meaning and Relevance: The Semantics and Pragmatics of Discourse Markers. Cambridge: Cambridge University Press. Borg, Emma 2004 Minimal Semantics. Oxford: Oxford University Press. Büring, Daniel 1997 The Meaning of Topic and Focus. London: Routledge. Cappelen, Herman and Ernie Lepore 2005 Insensitive Semantics. Oxford: Blackwell. Carnap, Rudolf 1942 Introduction to Semantics. Cambridge, MA: Harvard University Press. Carston, Robyn 1998 Informativeness, relevance and scalar implicature. In: Robyn Carston and Sachico Uchida (eds.), Relevance Theory: Applications and Implications, 179–236. Amsterdam: Benjamins. Carston, Robyn 1999 The semantics/pragmatics distinction. In: Ken Turner (ed.), The semantics/ pragmatics interface from different points of view, 85–125. Oxford: Elsevier. Carston, Robyn 2002 Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell.

486

John Saeed

Carston, Robyn 2004 Truth-conditional content and conversational implicature. In: Claudia Bianchi (ed.), The Semantics/Pragmatics Distinction, 65–100. Stanford, CA: CSLI. Carston, Robyn 2006 Linguistic communication and the semantics/pragmatics distinction. UCL Working Papers in Linguistics 18: 37–69. Chomsky, Noam 1988 Language and Problems of Knowledge. The Managua Lectures. Cambridge, MA: MIT Press. Chomsky, Noam 1995 Language and nature. Mind 104.413: 1–61. Davidson, Donald 1984 Inquiries into Truth and Interpretation. Oxford: Clarendon. Davidson, Donald 1990 The structure and content of truth. Journal of Philosophy 87: 279–328. Duranti, Alessandro and Charles Goodwin (eds.) 1992 Rethinking Context. Cambridge: Cambridge University Press. Fauconnier, Gilles 1988 Quantification, roles and domains. In: Umberto Eco, Marco Santambrogio and Patrizia Violi (eds.) Meaning and Mental Representations, 61–80. Bloomington, IN: Indiana University Press. Fauconnier, Gilles 1994 Mental Spaces: aspects of meaning construction in natural language. 2nd edition. Cambridge: Cambridge University Press. Fauconnier, Gilles 1997 Mappings in Thought and Language. Cambridge: Cambridge University Press. Fauconnier, Gilles and Mark Turner 2002 The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Fillmore, Charles J. 1982 Frame semantics. In: The Linguistic Society of Korea (ed.), Linguistics in the Morning Calm, 111–138. Seoul: Hanshin. Fillmore, Charles J. 1985 Frames and the semantics of understanding. Quaderni di Semantica 6: 222–254. Frege, Gottlob 1980 Translations from the Philosophical Writings of Gottlob Frege. Edited by Peter Geach and Max Black. Oxford: Blackwell. Grice, Herbert Paul 1975 Logic and conversation. In: Peter Cole and Jerrold L. Morgan (eds.), Syntax and Semantics. Vol. 3: Speech Acts, 43–58. New York: Academic Press. Grice, Herbert Paul 1978 Further notes on logic and conversation. In: Peter Cole (eds.) Syntax and Semantics. Vol. 9: Pragmatics, 113–128. New York: Academic Press. Grice, Herbert Paul 1989 Studies in the Way of Words. Cambridge, MA: MIT Press.

Pragmatics and semantics

487

Gumperz, John 1992 Contextualization and understanding. In: Duranti and Goodwin (eds.), 229– 252. Hanks, William F. 1992 The indexical ground of deictic reference. In: Duranti and Goodwin (eds.), 43–76. Horn, Laurence R. 1984 Toward a new taxonomy for pragmatic inference: Q-based and R-based implicatures. In: Deborah Schiffrin (ed.), Meaning, Form and Use in Context: Linguistics Applications, 11–42. Washington, DC: Georgetown University Press. Horn, Laurence R. 1989 A natural history of negation. Chicago: University of Chicago Press. Horn, Laurence R. 1992 The said and the unsaid. In: Chris Barker and David Dowty (eds.), SALT II: Proceedings of the Second Conference on Semantics and Linguistic Theory. 163–192. (Ohio State Working Papers in Linguistics 40.) Columbus, OH: Ohio State University Linguistics Department. Horn, Laurence R. 2004 Implicature. In: Laurence R. Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 3–28. Oxford: Blackwell. Huang, Yan 2007 Pragmatics. Oxford: Oxford University Press. Jackendoff, Ray 1990 Semantic Structures. Cambridge, MA: MIT Press. Kaplan, David. 1989a Demonstratives. In: Joseph Almog, John Perry, and Howard Wettstein (eds.), Themes from Kaplan, 481–563. Oxford: Oxford University Press. Kaplan, David 1989b Afterthoughts. In: Joseph Almog, John Perry, and Howard Wettstein (eds.), Themes from Kaplan, 565–614. Oxford: Oxford University Press. Kennedy, Chris 2007 Vagueness and grammar: the semantics of relative and absolute gradable adjectives. Linguistics and Philosophy 30.1: 1–45. King, Jeffrey C. and Jason Stanley 2005 Semantics, pragmatics and the role of semantic content. In: Szabó (ed.), 111–164. Kyburg, Alice and Michael Morreau 2000 Fitting words: vague language in context. Linguistics and Philosophy 23: 577–597. Lakoff, George 1987 Women, Fire and Dangerous Things. Chicago: University of Chicago Press. Lakoff, George 1988 Cognitive semantics. In: Umberto Eco, Marco Santambrogio and Patrizia Violi (eds.), Meaning and Mental Representations, 119–154. Bloomington, IN: Indiana University Press. Langacker, Ronald W. 1987 Foundations of Cognitive Grammar. Vol. 1. Stanford, CA: Stanford University Press.

488

John Saeed

Langacker, Ronald W. 1991 Foundations of Cognitive Grammar. Vol. 2: Descriptive Applications. Stanford, CA: Stanford University Press. Langacker, Ronald W. 1999 Grammar and Conceptualization. Berlin: de Gruyter. Lasersohn, Peter N. 1999 Pragmatic halos. Language 75.3: 522–551. Lepore, Ernest and Barry C. Smith (eds.) 2006 The Oxford Handbook of Philosophy of Language. Oxford: Oxford University Press. Levinson, Stephen C. 1983 Pragmatics. Cambridge: Cambridge University Press. Levinson, Stephen C. 1987a Minimization and conversational inference. In: Jef Verschueren and Marcella Bertuccellii Papi (eds.), The Pragmatic Perspective, 61–129. Amsterdam: Benjamins. Levinson, Stephen C. 1987b Pragmatics and the grammar of anaphora. Journal of Linguistics 23: 379–434. Levinson, Stephen C. 1991 Pragmatic reduction of the binding conditions revisited. Journal of Linguistics 27: 107–61. Levinson, Stephen C. 2000 Presumptive Meanings: the Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Lewis, David K. 1973 Counterfactuals. Oxford: Blackwell. Lyons, John 1977 Semantics. 2 volumes. Cambridge: Cambridge University Press. Montague, Richard 1974 Formal Philosophy: selected papers of Richard Montague. Edited and with an introduction by Richmond H. Thomason. New Haven: Yale University Press. Murphy, Gregory 1997 Polysemy and the creation of novel word meanings. In: Thomas B. Ward, Steven M. Smith and Jyotsna Vaid (eds.), Creative Thought: An Investigation of Conceptual Structures and Processes, 235–265. Washington, DC: American Psychological Society. Ortony, Andrew 1979 Metaphor and Thought. Cambridge: Cambridge University Press. Peleg, Orna, Rachel Giora and Ofer Fein 2004 Contextual strength: the whens and hows of context effects. In: Ira A. Noveck and Dan Sperber, (eds.), Experimental pragmatics, 172–186. Basingstoke: Palgrave. Perry, John 2001 Reference and Reflexivity. Stanford, CA: CSLI Publications. Pietroski, Paul M. 2005 Meaning before truth. In: Preyer and Peter (eds.), 255–302.

Pragmatics and semantics

489

Preyer, Gerhard and Georg Peter (eds.) 2005 Contextualism in Philosophy: Knowledge, Meaning, and Truth. Oxford: Clarendon Press. Preyer, Gerhard and Georg Peter (eds.) 2007 Context-Sensitivity and Semantic Minimalism. Oxford: Oxford University Press. Recanati, François 1989 The pragmatics of what is said. Mind and Language 4: 295–329. Recanati, François 1993a Direct Reference: From Language to Thought. Oxford: Blackwell. Recanati, François 1993b Domains of discourse. Linguistics and Philosophy 19: 445–475. Recanati, François 2000 Oratio Obliqua, Oratio Recta: an Essay on Metarepresentation. Cambridge, MA: MIT Press. Recanati, François 2004 Literal meaning. Cambridge: Cambridge University Press. Recanati, François 2005 Literalism and contextualism: some varieties. In: Preyer and Peter (eds.), 171–196. Saba, Walid S. and Jean-Pierre Corriveau 2001 Plausible reasoning and the resolution of quantifier scope ambiguities. Studia Logica 67.2: 271–89. Sadock, Jerrold M. 1984 Whither radical pragmatics? In: Deborah Schiffrin (ed.), Meaning, Form, and Use in Context: Linguistic Applications. Georgetown University Roundtable, 139–149. Washington, DC: Georgetown University Press. Saeed, John Ibrahim 2009 Semantics. 3rd ed. Oxford and Malden, MA: Wiley-Blackwell. Schank, Roger C. and Robert P. Abelson 1977 Scripts, Plans, Goals and Understanding. Hillsdale, N.J.: Lawrence Erlbaum. Searle, John 1979 Metaphor. In: Ortony (ed.), 83–111. Simpson, Greg B. 1984 Lexical ambiguity and its role in models of word recognition. Psychological Bulletin 96: 316–340 Simpson, Greg B. 1994 Context and the processing of ambiguous words. In: Morton Ann Gernsbacher (ed.), Handbook of psycholinguistics, 359–374. San Diego: Academic Press. Sperber, Dan and Deirdre Wilson 1995 Relevance. 2nd ed. Oxford: Blackwell. Sperber, Dan and Deirdre Wilson 1997 The mapping between the mental and the public lexicon. UCL Working Papers in Linguistics 9: 1–20. (Republished in Peter Carruthers and Jill Boucher (eds.), 1998. Language and Thought, 184–200. Cambridge: Cambridge University Press).

490

John Saeed

Stalnaker, Robert 1999 Context and Content. Oxford: Oxford University Press. Stanley, Jason 2002 Nominal restriction In: Gerhard Preyer and Georg Peter (eds.), Logical Form and Language, 365–388. Oxford: Oxford University Press. Stanley, Jason and Zoltán Gendler Szabó 2000 On quantifier domain restriction. Mind and Language 15.2/3: 219–261. Strawson, Peter Fredrick. 1950 On referring. Mind 59. 235: 320–344. Szabó, Zoltán Gendler 2001 Adjectives in context. In: Istvan Kenesi and Robert M. Harnish (eds.), Perspectives on Semantics, Pragmatics and Discourse, 199–146. Amsterdam: Benjamins. Szabó, Zoltán Gendler (ed.) 2005 Semantics versus Pragmatics. Oxford: Clarendon Press. Szabó, Zoltán Gendler 2006 The Distinction between semantics and pragmatics. In: Lepore and Smith (eds.), 361–389. Talmy, Leonard 2000 Toward a Cognitive Semantics. 2 vols. Cambridge, MA: MIT Press. Wilson, Deirdre and Robyn Carston 2007 A unitary approach to lexical pragmatics: relevance, inference and ad hoc concepts. In: Noel Burton-Roberts (ed.), Pragmatics, 230–59. Basingstoke: Palgrave Macmillan. Wilson, Deirdre and Dan Sperber 2004 Relevance theory. In: Lawrence Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 607–632. Oxford: Blackwell.

17.

Pragmatics and prosody: prosody as social action Elizabeth Couper-Kuhlen

This chapter is an attempt to take stock of advances in the field of prosody, in particular those which have had a direct impact on pragmatics and the study of language use in social contexts. It begins by looking backward, to where the field of prosody began and at what has been achieved so far with respect to the pragmatic functions of prosody, and ends by looking forward from the present standpoint to the challenges that lie ahead. By “prosody” I refer primarily to those auditory but non-segmental dimensions of speech which have a variable relation to the words used (Crystal 1969, Couper-Kuhlen 1986).1

1.

Looking backward

Where and how did the field of linguistic prosody start? What have been the mileposts along the way to its present position? The history of prosody as a linguistic enterprise dates from approximately the middle of the twentieth century, with scholars such as J.R. Firth, David Abercrombie, David Crystal, Michael Halliday in the UK and Kenneth Pike, George Trager, Dwight Bolinger in the US playing a leading role in establishing its credentials as a discipline within the language sciences. Early students of linguistic prosody, particularly those working in the University College of London tradition, focused primarily on intonation and conceptualized its basic unit, the tone unit, as corresponding roughly to the sentence. Descriptive models of the tone unit were developed primarily through introspection and the examination of read-aloud laboratory speech. The tone unit, it was thought, could be mapped onto sentences according to syntactic category (cf. e.g. Armstrong and Ward 1926, O’Connor and Arnold 1961, Halliday 1967). Five features characterized this early view of intonation: (i) metaphor of “superposition” In the American tradition, pitch and stress were called suprasegmentals (Trager and Smith 1957). These were conceptualized as being superimposed on the segmentally based units of language (syllables, morphemes, words) when the latter were spoken or read aloud. In the British tradition, the stressed syllables were treated as a frame on which the melodic line was suspended.

492

Elizabeth Couper-Kuhlen

(ii) the structuralistic perspective Pitch and stress, working in tandem, were thought to create the basic unit of spoken language, the tone unit, which itself was divided into smaller structural parts: e.g., pre-head, head (or body), nucleus and tail (cf. e.g. Crystal 1969, O’Connor and Arnold 1961). Speech was viewed as exhaustively segmentable into such discrete structural units. (iii) the context-free approach Tunes were mapped onto sentences with different syntactic shapes in a contextfree manner. These sentences were consequently isolates, and when pitch and stress were superimposed on them, the resulting melodies were also isolates. Intonation was acknowledged to be relative, but little thought was given to the question of “relative to what?” There was only one voice producing little snippets of melody with no relation to prior tunes or to other voices. In short, there was no provision for harmony. (iv) syntactically based functions The type of melody used was thought to be determined first and foremost by syntactic sentence type (Armstrong and Ward 1926). For instance, declaratives and wh-interrogatives were said to take final falls in the neutral or default case, yes-no interrogatives to take final rises. Departures from these neutral contours were said to convey attitudinal meaning, declaratives with final rises to express doubt, yes-no interrogatives with final falls to express seriousness or urgency (O’Connor and Arnold 1961). (v) phonological approach to meaning Intonation was believed to create contrasts in closed systems of meaning, as could be shown through minimal pairs (Halliday 1967, Schubiger 1958). For instance, sentences such as She speaks English naturally, George has plans to write or Isn’t he sure of himself were said to be disambiguated by placement of unit boundaries, position of tonic stress and type of main accent, respectively. In the early days then, prosody was equivalent to intonation, and intonation (= pitch and stress) was thought to be superimposed on the occasion of speaking on particular syntactic sentence types, which themselves were conceptualized as single utterances, considered in isolation from other utterances. Today we have come to think quite differently about prosody.2 In part thanks to developments in the field of sociology, including Berger and Luckmann’s social construction of reality (1967), Sacks, Schegloff and Jefferson’s ethnomethodological conversation analysis (1974), John Gumperz’ contextualization theory (1982), and the work of many others whose names go unmentioned here, we have now come to think of language as a tool for communication and interaction, and consequently of prosody in terms of social action.

Pragmatics and prosody: prosody as social action

493

The social actions which prosody is implicated in are as varied as are the tasks which participants must address in communicating with one another. For the conduct of mundane conversation, that primordial site of communication, the following are among those involved: – – – – – – – –

Turn construction Action formation Floor management Multi-unit turn construction Turn continuation Sequence organization Repair Stance and alignment

In the following I attempt to summarize what is known at present about the contribution of prosody as a resource for each of these conversational tasks. 1.1.

Prosody as a resource for turn construction

Intonation is no longer conceptualized as something which is optionally overlaid or superimposed on a piece of decontextualized language. Instead intonation – and prosody more generally – is considered to be an integral part of language (Linell 2005: 63). Language products, we have now come to see, do not exist without prosody, i.e. without some kind of pitch, loudness and timing, even if these are realized “silently” in reading and writing. Since the natural habitat of language is interaction and it is in turns-at-talk that “bursts of language” are found (Schegloff 1996), we now think of intonation and prosody primordially as affording resources for the construction of turns-at-talk. Given that there are no such things as turns-at-talk without prosodic formatting, prosody is seen as one of the building blocks of turns. Prosodic resources are used alongside lexical and syntactic resources as practices for the construction of turns-at-talk. 1.2.

Prosody as a resource for action formation

Today it is no longer customary to think in terms of simple and direct equations between sentence type (e.g. interrogative vs. declarative) and final pitch movement (rise vs. fall). Instead, grammar and prosody are thought of as being deployed together to build turns which implement particular social actions in interaction. Prosody then always has some kind of supporting role to play in action formation. A number of studies have highlighted this function.3 For instance, in dialect survey interviews specific prosodic features on word repetitions have been shown to cue the difference between an understanding check of a vocabulary item and a “mulling over” to determine its familiarity (Kelly and Local 1989). In riddle guessing

494

Elizabeth Couper-Kuhlen

sequences on call-in radio, another set of prosodic features accompanying the repetition of a guess by the compère cues the difference between an understanding check of that guess and a negative, disparaging assessment of it (Couper-Kuhlen 1996). In child-adult picture labeling, the prosody with which care-takers repeat the child’s version of a label conveys the difference between an invitation to repair that label and a confirmation that it is correct (Tarplee 1996). Interestingly, studies like these testify to the fact that not only final pitch movement but also less wellknown prosodic dimensions such as phonation type (e.g. breathy vs. non-breathy voice), pitch register (absolute vs. relative) and timing (e.g. delayed vs. welltimed) operate in tandem in specific sequential environments for action formation. One demonstration of the power of prosodic and phonetic formatting for implementing particular actions will be found in recent work on assessments [or evaluation] (actions which evaluate some personally experienced object or situation). Assessing turns which are produced in second position, i.e. in response to a prior assessment, and with prosodic and phonetic downgrading, i.e. with narrower pitch span, lower pitch accents and reduced volume vis-à-vis the prior assessment, will be interpretable as projecting disagreement, rather than agreement, with the first assessment, even though their assessing term may be the same or lexically upgraded (Ogden 2006).4 As Ogden points out, this kind of phenomenon is revealing with respect to what prosodic skills a competent speaker of English is expected to have: viz., they must be able to match their own prosodic and phonetic production of a turn to that of another speaker and possibly manipulate it for the creation of interactional meaning. Importantly, the relation between prosody and action formation is not contextfree but dependent on (a) co-occurring lexico-syntactic forms and (b) location in a particular sequential position. For example, (a) topic proffers5 in English conversation will have final falling pitch if they are done with a wh-interrogative but final rising pitch if they are yes-no interrogatives (Couper-Kuhlen, forthcoming, b). (b) Turns which are repeated following other-initiation of repair are likely to be done with prosodic and phonetic upgrading (i.e. expanded pitch range, longer word durations, greater loudness and articulatory resettings) if they were originally well placed in the sequence but without this kind of upgrading if they were sequentially misplaced (Curl 2004). Thus, just as there are no simple equations between sentence type and final pitch, so there are no simple relations between prosodic shape and action type without a concomitant consideration of lexico-syntax and sequential position. 1.3.

Prosody as a resource for floor management

Acknowledging that language, and thus prosody as well, inhabits turns at talk, we have also come to recognize its role in signaling transition readiness. Here the studies which have been carried out so far indicate that prosody, in conjunction

Pragmatics and prosody: prosody as social action

495

with lexico-syntax, has an important role to play in turn-taking. Our judgment of whether someone has finished speaking or not depends to no little extent on whether a stretch of speech sounds prosodically complete or not. But the contribution of prosody to turn transition is sensitive to considerations of syntactic and pragmatic completeness as well (Ford, Fox and Thompson 1996, Ford and Thompson 1996, Selting 1996a, 2000, 2005). That is, a possibly complete prosodic contour will not signal transition readiness unless it is accompanied by possibly complete syntax and unless the activity underway has also reached a point of pragmatic completion. The role of prosody in turn-taking is now known to involve more than simply rising or falling intonation (Szczepek Reed 2004).6 We have learned, for instance, that different kinds of pausing, e.g. whether a speaker “trails off” into silence or produces silence by holding a glottal stop over time, can distinguish turn-yielding from turn-holding (Local and Kelly 1986). Although we by no means fully understand how turn transition works, we do now know that English dialects can differ significantly in terms of how syllable-based pitch, loudness and duration are configured at the ends of floor-yielding turns (Local, Kelly and Wells 1986, Wells and Peppé 1996, Wells and Macfarlane 1998). And we know that speakers of some languages rely much more heavily on prosodic features other than pitch in managing the floor in interaction. In Finnish, for instance, it appears to be non-modal voice quality (including creaky phonation) which does much of the work (Ogden 2004, Ogden and Routarinne 2005). Other languages, by contrast, appear to rely only minimally on prosody for managing the floor (Tanaka 2000, 2004). Prosody is also known to have a crucial role to play in cueing how overlapping talk is understood. Competitive incomings7, for instance, i.e. incomings which lay an immediate claim to speakership well before a current speaker’s point of possible completion, will be marked as such with high pitch and increased loudness (French and Local 1983). The current floor-holder will typically respond either with a prosodic “return of competition” (increased loudness, decreased pace) or with a “yield to competition” (decreased volume). By contrast, entries at non-completion points which do not lay claim to the floor, e.g. quips and asides, are lower in pitch and volume than prior talk and the speaker’s norm. Here the current floor-holder does not treat the incoming speech as turn-competitive but drops out immediately, only to resume (with a small delay) once it is finished (French and Local 1983). In sum, current research indicates that, in one way or another and to varying degrees, prosody is implicated in all aspects of turn-taking, including whether and how speakership transfers from one participant to another in interaction. In this sense it is central to the concerns of both interactants and analysts.

496 1.4.

Elizabeth Couper-Kuhlen

Prosody as a resource for multi-unit turn construction

Today it is also customary to look beyond the single intonation unit or phrase to consider how adjacent intonation phrases relate to one another. This means focussing on the prosodic “seams” between adjacent units in single-speaker turns-attalk. To do this, the beginnings of new intonation phrases must be related to the endings of preceding ones, patterns of rhythmic timing across units identified and attention paid to long-range articulatory and phonatory settings at unit boundaries.8 Registering this level of detail allows us to say that some neighboring intonation phrases cohere more closely than others. It is the presence of prosodic cohesion between adjacent intonation phrases in a speaker’s turn which is related to various practices for “multi-unit turn construction”, or building one turn out of several turn units. For instance, at points of possible turn completion there is the practice of abrupt joins: here what we find at the seam between two adjacent units is a local speeding-up, a sudden diminuendo and closer than expected temporal proximity between the two units. Interestingly, these features display that the actions involved are being treated as disjunct with one another (Local and Walker 2004). Similarly, there is the practice of parenthetic inserts: here it is shifts to low density of accentuation, faster tempo and lower overall pitch and amplitude compared to prior talk which signal that the current bit of talk is being presented as parenthetic (Uhmann 1992, Local 1992, Wichmann 2001, Mazeland 2007). Further, there is the practice of reported speech, where speakers will be heard to use clear prosodic breaks in pitch level, rhythm and loudness when switching from their own voice to that of another (Couper-Kuhlen 1999, Anward 2002, Holt and Clift 2007). Such prosodic shifts function like quotation marks in speech, although their placement with respect to the quotation itself is more flexible (Klewitz and Couper-Kuhlen 1999). Yet not only dis-junctive, also con-junctive relations between units within a speaker’s turn are conveyed by clusters of prosodic features linking neighboring intonation phrases. There is, for instance, the practice of rush-throughs. This involves speeding up the pace of talk towards the end of one turn-constructional unit, withholding outbreaths and other marks of finality, and moving quickly into the next turn-constructional unit, so as to present the two units in question as welded together quasi-seamlessly (Schegloff 1987). Or the practice of projecting, at the end of one unit, that another unit will follow. For instance, in conjunction with the particle no when it is used to begin a turn following a topic proffer, mid-to-high pitch range and increasing intensity project that the speaker will go on to produce another unit of related talk (Ford, Fox and Hellermann 2004). Prosodic conjunction can also be created through the practice of emphatic scansions, where the isochronous and isometric9 distribution of syllables across units in a turn-at-talk produces a “scanned” multi-unit turn with special rhetorical effects (Auer, CouperKuhlen and Müller1999). Further, there is the practice of conversational listing,

Pragmatics and prosody: prosody as social action

497

where in conjunction with lexico-syntactic parallelism, the use of recurrent “plateau” contours and other forms of prosodic stylization creates an enumerating multi-unit turn (Selting 2004, 2007). All of these practices involve particular prosodic features which, together with specific lexico-syntactic forms, link adjacent units and signal that disjunctive or conjunctive relationships hold between them. It is thus by comparing and contrasting prosodic formatting at the boundaries between adjacent units within a single speaker’s turn that we discover distinct combinations of pitch configuration, loudness and timing and specific patterns of phonatory and articulatory setting which together, and along with lexico-syntax, contribute to the construction of specific kinds of complex, multi-unit turns-at-talk. 1.5.

Prosody as a resource for turn continuation

Today we now also compare and contrast prosodic formatting across non-adjacent same-speaker turn units and turns. Obviously, there are challenges here due to the temporal discontinuities involved, but techniques have been developed to deal with them. By comparing the prosody at the provisional end of a speaker’s turn with that at the beginning of a subsequent continuation of that turn by the same speaker, “increments”10 can be described as prosodically fitted to their hosts if they are not significantly higher in pitch or louder in volume than the turn they are continuing and do not involve noticeable changes in articulatory setting or rate of articulation (Walker 2004). More generally, any type of same-speaker continuation following a possible turn completion can be said to be either prosodically integrated if it is tacked on to the prior intonation contour, or prosodically exposed if it has a significant pitch jump together with a marked change of tempo, loudness or rhythm (Auer 1996). Further, when a speaker’s turn is interrupted or discontinued and returned to again later, there is now a prosodically and phonetically grounded way to talk about continuing that turn vs. re-starting it: resumption of a curtailed turn is done by matching the pitch and loudness levels at drop-out, whereas recycled material is presented as a new contribution if its pitch and loudness levels are re-set (Local 1992). In addition to prosodic formatting across same-speaker turns, and just as importantly, prosodic formatting across turns by different speakers can now be compared and contrasted. This is no small feat, given the fact that the speakers involved may have significantly different pitch spans and loudness ranges as well as different habitual speech rates and rhythms. Yet normalization procedures have been developed for comparing pitch and loudness, accent timing, tempo and speech rate in cross-speaker talk (cf. e.g. Couper-Kuhlen 1996, Local 2005, Ogden 2006). As a result, meaningful statements can now be made about the prosody of a speaker’s unfinished (or finished) turn and the prosody of a collaborative completion (or continuation)11 of that turn by another speaker (Local 2005, Szczepek Reed 2006). Collaborative completions, for instance, have been found to be typically no faster

498

Elizabeth Couper-Kuhlen

or louder than the turn they are completing and to typically match the overall pitch of the prior talk. They are thus displayed, and accountably so, not only as lexicosyntactic but also as prosodic continuations of what a prior speaker has begun but not (yet) brought to completion (Local 2005). In sum, when strict adjacency between stretches of talk is lacking, it is often prosody which contributes to marking cohesive relations between them. Current research is revealing just how sensitive participants are to the prosodic details of continuing a turn-at-talk, be it one’s own or that of another. Along with lexico-syntactic considerations, it is these details which help convey whether the talk is being presented as part of a turn already begun or as something new. 1.6.

Prosody as a resource for sequence organization

We are now also beginning to appreciate the role of prosodic and phonetic resources in signaling how a new turn-at-talk relates to prior turns. For instance, in radio phone-in calls following an exchange of greetings and possibly how-areyou’s, callers can signal that a turn in first-topic slot is the reason-for-the-call by using extra high pitch on its first accented syllable (Couper-Kuhlen 2001).12 If it is merely an aside, it will lack these features. Once a sequence has reached a point of possible completion, a cluster of prosodic cues will conspire to convey whether the next turn is beginning a new sequence or continuing the ongoing one: turns which are new beginnings are typically louder and higher in pitch from the outset than prior same-speaker talk and are often accompanied by rhythmic breaks (Goldberg 1978, Couper-Kuhlen 1993, 2004). Turns which continue the prior line of topical or sequential development lack these features. New sequences which are begun after a prior one has been closed down are also often marked as phonetically discontinuous by the use of clicks and loud inbreaths (Wright 2007). Prosodic and phonetic features also contribute to signaling the possible end of a sequence. Unit-final “conjunctionals” such as but, and and so have been considered ambivalent in terms of whether the speaker is projecting further talk or not (Jefferson 1983), but a consideration of their prosodic and phonetic delivery resolves much of the ambivalence (Local and Kelly 1986). Stand-alone variants of so, for instance, will be heard as potentially sequence-closing if they are relatively low in pitch, have quiet volume and lack final glottal closure, but as projecting more on-topic talk if they are relatively high in pitch, loud and have final glottal closure (Local and Walker 2005). As Schegloff (1990) has pointed out, adjacent turns-at-talk can either “clump” together, i.e. form coherent courses of action, or not. Recent research shows that such “clumping” is often underlined by specific prosodic and phonetic devices.

Pragmatics and prosody: prosody as social action

1.7.

499

Prosody as a resource for repair

In addition to signaling structural dimensions of talk, prosody, along with lexicosyntactic resources, is now known to be crucial with respect to what Goffman has called the ritual requirements of talk. In self-repair, for instance, phonetic and glottal cut-offs will mark a potential trouble source in a speaker’s turn and initiate its repair (Jasperson 2002); rhythmic breaks and extreme spikes in pitch and loudness will accompany the upcoming self-repair itself (Cutler and Ladd 1983). As for other-initiation of repair, interactants use prosodic resources when they produce a noticeable silence instead of the conditionally relevant response and thereby implicitly convey that the prior turn may need some reworking. “Noticeable silences” are of course notoriously problematic for analysts: how much elapsed time between turns counts as a delay? Yet for participants they are less so. Recent studies suggest that participants orient to inter-turn silence in relation to the pacing of prior talk (Couper-Kuhlen 1993, Auer, Couper-Kuhlen and Müller 1999). Prosodic and phonetic resources are also deployed in other-initiated repair when a turn has been targeted as problematic via a generalized repair initiator such as what? or huh? and is repeated with prosodic and phonetic upgrading or downgrading (Curl 2004, 2005). In short, from what we know so far, it is fair to say that if we did not pay attention to prosodic and phonetic practices in talk we would be unable to account for much of repair-related behavior, including when repair is initiated, what is targeted as a trouble source, whether a piece of talk is carrying out a repair or not and how the responsibility for remedial action is being assigned. 1.8.

Prosody as a resource for displaying and negotiating stance and alignment

Finally, prosody is currently coming to be appreciated as a resource for the display and negotiation of stance and alignment. The most persuasive research findings here have situated the use of specific prosodic and phonetic practices within particular action sequences (Goodwin and Goodwin 2000). For example, in news deliveries: the valence with which a piece of news is delivered can be ascribed not only lexically but also prosodically. Good news will often be signaled with high pitch level, wide pitch range, animated contour and faster speech rate, while bad news will have low pitch level, narrow pitch range, stretched vowels, slower pace and often breathy or creaky voice quality (Freese and Maynard 1998). And if a piece of news is receipted with a freestanding news-receipt token such as oh, this turn will be heard as displaying surprise if the oh is lengthened and delivered with an extra-high and pointed (rising-falling) contour13 but as displaying sympathy or disappointment if it has low pitch, compressed pitch range, soft volume and weak and/or breathy voice (Reber 2008, Couper-Kuhlen 2009). Further, in children’s game disputes, when cries of hey!, no! and out! are produced with sig-

500

Elizabeth Couper-Kuhlen

nificant lengthening, extra high pitch and pronounced falling-rising contours, this will mark not only disagreement but an affective dimension of strong opposition (M. Goodwin, C. Goodwin and Yaeger-Droh 2002). As these and similar studies show, it is by examining prosodic and phonetic resources for the implementation of specific actions situated in particular sequences and activities that their distinctive contribution to the display and interpretation of emotive involvement can be appreciated. But surely one of the most fascinating tasks for which interactants deploy prosody is the negotiation of stance and alignment. In telephone openings, for instance, callers attend to subtle indications of mood and enthusiasm as displayed by the pitch level at which their co-participants deliver and respond to greetings and how are you’s. When pitch levels are adjusted to mutually accommodate, this leads to the joint negotiation of the tenor at which the conversation is to be conducted (Schegloff 1998). Similar online prosodic negotiation will be found with respect to rhythm and pacing in conversation openings and closings on the telephone (Couper-Kuhlen 1993, Auer, Couper-Kuhlen and Müller 1999). More generally speaking, in virtually any kind of second position, speakers adjust the timing of their incoming according to the pacing of the prior turn. The way they do this cues meaningful inferences such as ‘Interaction proceeding apace’ or ‘I have a problem with responding as expected’ (Auer, Couper-Kuhlen and Müller 1999). Recipients also orient in other respects to their interlocutor’s prosody and thereby convey alignment with the prior turn or speaker (Szczepek Reed 2006).14 This is the dimension of prosodic harmony which was missing in early treatments of intonation, and it is perhaps one of the most exciting aspects of interactive prosodic practice currently being investigated today. 1.9.

Summary

Looking back now at the work in prosody and interaction which has emerged over the recent years, it becomes apparent that the nature of the enterprise has subtly changed. For one, attention is no longer paid only to final, or nuclear pitch movement. Consideration is also given to onset level, pitch range and pitch register; to the timing of accented and unaccented syllables; to amplitude shifts and levels of loudness, to pausing and audible breathing; to phonation types and articulatory settings. Moreover, since the approach is an inductive one, further ways in which prosodic, phonetic and paralinguistic dimensions of speech are relevant to interaction are likely to be discovered in the future.15 Some significant advances have already been made. We have new ways of describing prosodic and phonetic relationships with categories such as prosodic matching and prosodic orientation, phonetic upgrading and downgrading, hypo- and hyper-articulation, etc. And we have new tools for analyzing prosodic and phonetic patterns, e.g. the speech analysis program P RAAT , which is not only robust enough to handle non-laboratory data and

Pragmatics and prosody: prosody as social action

501

relatively large chunks of speech at a time but is also freely available and constantly being updated.16 But we are still debating what kind of notation system is needed for the representation of prosodic and phonetic features in talk-in-interaction.17

2.

Looking forward

What does the future hold in store? What challenges still remain for research into prosody and interaction? Two recent developments are bringing about some reorientation among students of interactional prosody. They concern (i) the scope of the field and (ii) its status as a resource for social action in interaction. 2.1.

Scope of prosody

What should we call prosodic? With the increasing awareness that many phonatory, articulatory and voice-quality parameters form clusters along with traditional prosodic parameters for structuring conversation and conveying interactional meaning, the original narrow understanding of prosody as restricted to pitch, loudness and timing is gradually giving way. Just like the more traditional prosodic dimensions, phonetic and paralinguistic features also convey what belongs together and what does not (the disjunctive-conjunctive dimension of talk) as well as what matches and/or upgrades a prior turn-unit and what downgrades it (the affiliative-disaffiliative dimension of talk). For this reason, it is becoming increasingly artificial to restrict one’s attention to a mere subset of these phenomena. Local has described the situation this way: (…) in structuring their talk-in-interaction speakers systematically draw on bundles of phonetic features which cut across the traditional classification of phonetic parameters into “prosodic” on one hand versus articulatory (or “segmental”) on the other. This suggests that phonetic parameters are best treated as falling into functional bundles or clusters, irrespective of their “prosodic” or “segmental” characteristics, on the basis of how speakers deploy them to achieve particular interactional goals. (Local 2004:396)

So prosodists today are faced with a terminological dilemma: what to call the prosodic, alias paralinguistic, alias phonetic features which speakers routinely use for the management of interaction in all its structural/organizational and its ritual/ interactional facets? Should everything be called “phonetic”? This runs the risk of making the object of inquiry indistinguishable from that of linguistic phoneticians whose concern is the phonetic-phonemic interface. Should everything be called “prosodic”? This runs the risk of allowing articulatory dimensions to be forgotten along the way. One is reminded of the debate over sexism in language: are women made invisible through the use of linguistic forms such as everyone … he? Just as with sexist language, so too discriminatory scientific labels should be avoided. The solution may thus lie in joint mention, with reference being made, where relevant,

502

Elizabeth Couper-Kuhlen

to “prosodic and phonetic” dimensions of language use in interaction. On this understanding, the present chapter might be more aptly titled: “Prosody and phonetics as social action”. 2.2.

Status of prosody

How central to social interaction is prosody? With recent advances in video technology there has been a concomitant increase in the attention paid to so-called “multi-modality”, i.e. to those embodied and visually based semiotic systems implicated in the conduct of interaction. Prosody is thus no longer an “only child”; it has acquired a number of siblings and must now take its place alongside other nonverbal practices for conducting interaction, including gaze, facial expression, gesture, body posture and kinesics. Ironically, among all the various types of embodiment, prosody – for decades treated as lying “at the edge of language” (Bolinger 1972) – now appears the most strongly linguistic. Despite the fact that many linguists continue to treat it as at best “paralinguistic”, in the larger picture of interactional conduct prosody is much more closely tied to language than is, say, gaze or facial expression. This is in part because prosody is intrinsic to speech and speech-like vocalization, and in part because prosodic resources are primarily auditory and are thus organized linearly in ways in which visible resources are not. So prosody, and research on prosody in interaction, must find a new place in the multimodal world of interaction analysis. And unless language is to be prioritized, that place may not be front stage. Compared to the more dominant, visual semiotic systems deployed in interaction, prosodic and phonetic resources may appear somewhat limited, especially if they are treated as auditory only. Actually there is often a visible component to the prosodic and phonetic formatting of talk: so-called “smile” voice, for instance, is quite naturally embodied in spread lips and is therefore clearly visible. Surges in loudness and pitch are typically accompanied by noticeable changes in jaw and larynx position. The temporal onset and offset of speech is visible, and many articulations, not to mention articulatory settings (e.g. labialization), have clearly visible dimensions. In short, it would behoove prosodists and phoneticians to pay more attention to the visible cues associated with prosodic and phonetic practices. In this respect prosody and phonetics also have a visual contribution to make. But surely their most important contribution to multimodal interaction is time or temporality. In C. Goodwin’s words: “ … time is an intrinsic component of the units used to build events and action within situated human interaction. The location of relevant units is a key task facing both analysts … and participants …” (C. Goodwin 2002: 47). Even in the larger, multimodal picture, temporality remains the backbone of interaction: it is in real time that interactive events emerge and position themselves sequentially and simultaneously. And prosodists and phoneticians can help track this temporal unfolding.

Pragmatics and prosody: prosody as social action

2.3.

503

The way forward

Despite the dominance of multimodality in the world of interaction, prosody and phonetics in social action remain a central concern both for participants and for analysts. Without prosodic and phonetic resources, interactants would lack a crucial means for implementing actions indirectly; they would have no way of determining when a particular action is due or overdue; they would be ill-equipped to mark disjunctive relationships between adjacent pieces of talk or conjunctive relationships between non-adjacent ones; the beginnings of their courses of action would be lacking in projectability and therefore indistinct from endings; their repairs would risk being taken for continuations; their stances towards actions and towards co-participants would remain opaque, etc. It has often been claimed that meaning in interaction is dependent on composition (what sorts of parts) and on position (located where) (e.g. Schegloff 2007: 81). Yet just as crucial is the question of manner (produced how): how high or low, how soft or loud, how fast or slow is a piece of talk? Is it well-timed or delayed, glottalized or not, said with spread lips (“smile” voice) or not? It is answers to such questions that help illuminate the omnipresent issue in interaction: “Why that now?” (Schegloff and Sacks 1973: 299). The way forward in the field of pragmatics and social interaction will thus require continued attention to minute prosodic and phonetic detail. Concomitantly, it will require a flexible and, where needed, prosodically and phonetically explicit transcription system. But it will also demand of its practitioners the willingness and ability to integrate prosodic and phonetic systems with other semiotic modalities in order to fully appreciate the “rich multimodal ecology of sign systems” (Goodwin 2002) which underlies human social interaction. Notes 1. As will become apparent in the course of this chapter, however, the actual delimitation of the field of prosody is currently a subject of debate. 2. By “we” I mean those engaged in the empirical and in vivo investigation of prosody as a pragmatic resource. 3. For reasons of space the following list is necessarily selective (the examples mentioned all involve word repetition, for instance). Here and elsewhere in this chapter, barring indication to the contrary, mentions are restricted to work on English. 4. In Ogden’s most telling case, a second speaker responds to the assessing turn isn’t that good at long last with that’s that’s (.) very good news but in a way which is prosodically and phonetically downgraded. In subsequent talk the latter turn is treated as intimating mild disagreement with the co-participant’s first assessment (2006: 1771). 5. These are interrogative turn-constructional units which propose a new topic of talk, either after the opening section of the call or later in the body of the conversation after a prior topic has been brought to possible closure (Couper-Kuhlen 2004, Schegloff 2007). 6. If only for this reason the negative results of de Ruiter, Mitterer and Enfield’s laboratory experiments on final intonation in Dutch turn-taking (2006) are not surprising.

504

Elizabeth Couper-Kuhlen

7. “Competitive incoming” refers to a would-be speaker trying to take over the floor immediately, although the current speaker is nowhere near completing their turn. 8. It is in part because articulatory and laryngeal settings are often co-implicated in this kind of cohesive linking that some scholars have argued for speaking of phonetic instead of prosodic resources (Local and Walker 2005). See also section 2.1. 9. “Isochronous” refers to a type of rhythmization in which rhythmic intervals have approximately the same temporal duration; “isometric” refers to another type of rhythmization in which rhythmic intervals have the same number of syllables (Auer, CouperKuhlen and Müller 1999). 10. An “increment” is a stretch of same-speaker talk which (a) is produced after a point of possible completion in that speaker’s turn, and (b) is syntactically symbiotic with the prior talk (cf. e.g. Schegloff 1996, Couper-Kuhlen and Ono 2007). 11. “Collaborative completion” is a term used to refer to a participant finishing a turn-constructional unit which their interlocutor has left (provisionally) incomplete. “Collaborative continuation” refers to a participant adding on to a turn-constructional unit which their interlocutor has just (provisionally) completed. 12. This typically happens after the production of uh or uhm (Schegloff 2006). 13. This is especially so when the news being delivered has been designed to elicit surprise (Wilkinson and Kitzinger 2006). Cf. also Local (1996), and Selting (1996b) for German. 14. Szecepek Reed (2006), for example, documents prosodic orientation involving intonation contour, pitch step-up, register, loudness, speech rate and voice quality. 15. One good example of a recent discovery comes from work on crying in interaction (Hepburn 2004, Hepburn and Potter 2007), a follow-up to Jefferson’s (1985) study of laughter. 16. Developed by Paul Boersma and David Weenink, Institute of Phonetic Sciences, University of Amsterdam; available at www.praat.org. 17. Among the options are the autosegmental, laboratory-based ToBI system (Beckman and Ayers 1994), the discourse-based Discourse Transcription system (Du Bois et al 1992) and the CA-based Gesprächsanalytisches Transkriptionssystem: GAT (Selting et al 1998, 2009).

References Anward, Jan 2002 Other voices, other sources. In: Per Linell and Karin Aronsson (eds.), Jagen och rösterna: Goffman, Viveka och samtalet / Selves and Voices: Goffman, Viveka, and Dialogue, 128–147. Linköping: Tema K, Linköpings Universitet. Armstrong, Lilias E. and Ida C. Ward 1926 A Handbook of English Intonation. Cambridge: W. Heffer and Sons. Auer, Peter 1996 On the prosody and syntax of turn-continuations. In: Elisabeth Couper-Kuhlen and Margret Selting (eds.), Prosody in Conversation, 57–100. Cambridge: Cambridge University Press. Auer, Peter, Elizabeth Couper-Kuhlen and Frank Müller 1999 Language in Time: The rhythm and tempo of spoken interaction. New York: Oxford University Press.

Pragmatics and prosody: prosody as social action

505

Beckman, Mary E. and Gayle. M. Ayers 1994 Guidelines for ToBI labeling. Version 2.0. Columbus, OH, Ohio State University, Department of Linguistics. Berger, Peter and Thomas Luckmann 1967 The Social Construction of Reality. Harmondsworth: Penguin. Bolinger, Dwight 1972 Intonation. Selected readings. Penguin: Harmondsworth, Middlesex. Couper-Kuhlen, Elizabeth 1986 An Introduction to English Prosody. Tübingen: Max Niemeyer and London: Edward Arnold. Couper-Kuhlen, Elizabeth 1993 English Speech Rhythm: Form and function in everyday verbal interaction. Amsterdam: Benjamins. Couper-Kuhlen, Elizabeth 1996 The prosody of repetition: On quoting and mimicry. In: Elizabeth CouperKuhlen and Margret Selting (eds.), Prosody in Conversation, 366–405. Cambridge: Cambridge University Press. Couper-Kuhlen, Elizabeth 1999 Coherent voicing: On prosody in conversational reported speech. In: Wolfram Bublitz and Uta Lenk (eds.), Coherence in Spoken and Written Discourse: How to create it and how to describe it, 11–32. Amsterdam: Benjamins. Couper-Kuhlen, Elizabeth 2001 Interactional prosody: High onsets in reason-for-the-call turns. Language in Society 30: 29–53. Couper-Kuhlen, Elizabeth 2004 Prosody and sequence organization: The case of new beginnings. In: Elizabeth Couper-Kuhlen and Cecilia E. Ford (eds.), Sound Patterns in Interaction, 335–376. Amsterdam: Benjamins. Couper-Kuhlen, Elizabeth 2009 A sequential approach to affect: The case of “disappointment”. In: Markku Haakana, Minna Laakso and Jan Lindström (eds.), Talk in Interaction – comparative dimension. Helsinki, Finnish Literature Society. Couper-Kuhlen, Elizabeth Forthc. Some truths and untruths about final intonation in conversational questions. In: Jan P. de Ruiter (ed.), Questions, 94–123. Cambridge: Cambridge University Press. Couper-Kuhlen, Elizabeth and Tsuyoshi Ono 2007 “Incrementing” in conversation. A comparison of practices in English, German and Japanese. Pragmatics 17.4: 513–552. Crystal, David 1969 Prosodic Systems and Intonation in English. Cambridge, Cambridge University Press. Curl, Traci S. 2004 “Repetition” repairs: The relationship of phonetic structure and sequence organization. In: Elizabeth Couper-Kuhlen and Cecilia E. Ford (eds.), Sound Patterns in Interaction, 273–298. Amsterdam, Benjamins.

506

Elizabeth Couper-Kuhlen

Curl, Traci S. 2005 Practices in other-initiated repair resolution: the phonetic differentiation of “repetitions”. Discourse Processes 39.1: 1–44. Cutler, Anne and D. Robert Ladd (eds.) 1983 Prosody: Models and measurements. Berlin: Springer. Du Bois, John W., Stephan Schuetze-Coburn et al. 1992 Discourse Transcription. Santa Barbara Papers in Linguistics 4: 1–225. Ford, Cecilia E. and Sandra A.Thompson 1996 Interactional units in conversation. Syntactic; intonational and pragmatic resources for the management of turns. In: Elinor Ochs, Emanuel A. Schegloff, and Sandra A. Thompson (eds.), Interaction and Grammar, 134–184. Cambridge: Cambridge University Press. Ford, Cecilia E., Barbara A. Fox and Sandra A.Thompson 1996 Practices in the construction of turns: The “TCU” revisited. Pragmatics 6: 427–454. Ford, Cecilia E, Barbara A. Fox and John Hellermann 2004 “Getting past no”: Sequence, action and sound production in the projection of no-initiated turns. In: Elizabeth Couper-Kuhlen and Cecilia E. Ford (eds.), Sound Patterns in Interaction, 233–272. Amsterdam: Benjamins. Freese, Jeremy and Douglas W. Maynard 1998 Prosodic features of bad news and good news in conversation. Language in Society 27:195–219. French, Peter and John K. Local 1983 Turn-competitive incomings. Journal of Pragmatics 7: 17–38. Goldberg, JoAnn 1978 Amplitude shift: A mechanism for the affiliation of utterances in conversational interaction. In: Jim Schenkein (ed.), Studies in the Organization of Conversational Interaction, 199–218. New York, New York Academic Press. Goodwin, Charles 2002 Time in action. Current Anthropology 43: 19–35. Goodwin, Marjorie and Goodwin, Charles 2000 Emotion within situated activity. In: Alessandro Duranti (ed.), Linguistic Anthropology: A Reader, 239–257. Oxford: Blackwell. Goodwin, Marjorie H., Charles Goodwin and Malcah Yaeger-Droh 2002 Multi-modality in girls’ game disputes. Journal of Pragmatics 34: 1621–1649. Gumperz, John 1982 Discourse Strategies. Cambridge: Cambridge University Press. Halliday, Michael A. K. 1967 Intonation and Grammar in British English. The Hague: Mouton. Hepburn, Alexa 2004 Crying: Notes on description, transcription and interaction. Research on Language and Social Interaction 37: 251–290. Hepburn, Alexa and Jonathan Potter 2007 Crying receipts: Time, empathy, and institutional practice. Research on Language and Social Interaction 40.1: 89–116. Holt, Elizabeth and Rebecca Clift (eds.) 2007 Reporting Talk: Reported speech in interaction. Cambridge: Cambridge University Press.

Pragmatics and prosody: prosody as social action

507

Jasperson, Robert 2002 Some linguistic aspects of closure cut-off. In: Cecilia E. Ford, Barbara A. Fox and Sandra A. Thompson (eds.), The Language of Turn and Sequence, 257–286. Oxford: Oxford University Press. Jefferson, Gail 1983 On a failed hypothesis: ‘Conjunctionals’ as overlap-vulnerable. Tilberg Papers in Language and Literature 28: 1–33. Jefferson, Gail 1985 An exercise in the transcription and analysis of laughter. In: Teun van Dijk (ed.), Handbook of Discourse Analysis. Vol. 3, 25–34. London, Academic. Kelly, John and John Local 1989 Doing Phonology. Manchester: Manchester University Press. Klewitz, Gabriele and Elizabeth Couper-Kuhlen 1999 Quote-unquote. The role of prosody in the contextualization of reported speech sequences. Pragmatics. 9.4: 459–485. Linell, Per 2005 The Written Language Bias in Linguistics. Its nature, origins and transformations. London: Routledge. Local, John 1992 Continuing and restarting. In: Peter Auer and Aldo di Luzio (eds.), The Contextualization of Language, 273–296. Amsterdam, Benjamins. Local, John 1996 Conversational phonetics: Some aspects of news receipts in everyday talk. In: Elizabeth Couper-Kuhlen and Margret Selting (eds.), Prosody in Conversation: Interactional studies, 177–230. Cambridge: Cambridge University Press. Local, John 2004 Getting back to prior talk: and-uh(m) as a back-connecting device in British and American English. In: Elizabeth Couper-Kuhlen and Cecilia E. Ford (eds.), Sound Patterns in Interaction, 377–400. Amsterdam: Benjamins. Local, John 2005 On the interactional and phonetic design of collaborative completions. In: W. Hardcastle and J. Beck (eds.), A Figure of Speech: a Festschrift for John Laver, 263–282. Mahwah, NJ: Lawrence Erlbaum. Local, John and John Kelly 1986 Projection and “silences”: Notes on phonetic and conversational structure. Human Studies 9: 185–204. Local, John, John Kelly and William H.G. Wells 1986 Towards a phonology of conversation: Turn-taking in Tyneside English. Journal of Linguistics 22: 411–437. Local, John and Gareth Walker 2004 Abrupt-joins as a resource for the production of multi-unit, multi-action turns. Journal of Pragmatics 36: 1375–1403. Local, John and Gareth Walker 2005 Methodological imperatives for investigating the phonetic organization and phonological structures of spontaneous speech. Phonetica 62: 120–130.

508

Elizabeth Couper-Kuhlen

Mazeland, Harrie 2007 Parenthetical sequences. Journal of Pragmatics 39: 1816–1869. O’Connor, J.D. and Arnold, Gordon F. 1961 Intonation of Colloquial English. London: Longmans. Ogden, Richard 2004 Non-modal voice quality and turn-taking in Finnish. In: Elizabeth CouperKuhlen and Cecilia E. Ford (eds.), Sound Patterns in Interaction, 29–62. Amsterdam: Benjamins. Ogden, Richard 2006 Phonetics and social action in agreements and disagreements. Journal of Pragmatics 38: 1752–1775. Ogden, Richard and Sara Routarinne 2005 The communicative function of final rises in Finnish intonation. Phonetica 62: 160–175. Reber, Elisabeth 2008 Affectivity in Talk-in-interaction: Sound objects in English. PhD diss., University of Potsdam. de Ruiter, Jan Peter, Holger Mitterer and Nick J. Enfield 2006 Predicting the end of a speaker’s turn; a cognitive cornerstone of conversation. Language 82.3: 515–535. Sacks, Harvey, Emanuel Schegloff and Gail Jefferson 1974 A simplest systematics for the organization of turn-taking for conversation. Language 50: 696–735. Schegloff, Emanuel A. 1987 Analyzing Single Episodes of Interaction: An Exercise in Conversation Analysis. Social Psychology Quarterly 50.2: 101–114. Schegloff, Emanuel A. 1990 On the organization of sequences as a source of “coherence” in talk-in-interaction. In: B. Dorval (ed.), Conversational Organization and its Development, 51–77. Norwood, NJ, Ablex. Schegloff, Emanuel A. 1996 Turn organization: One intersection of grammar and interaction. In: Elinor Ochs, Emanuel A. Schegloff and Sandra A. Thompson (eds.), Interaction and Grammar, 52–133. Cambridge: Cambridge University Press. Schegloff, Emanuel A. 1998 Reflections on studying prosody in talk-in-interaction. Language and Speech 41: 235–263. Schegloff, Emanuel A. 2006 On “uh” and “uhm” and some of the things they are used to do. brandial, Potsdam, September 11–13, 2006. Schegloff, Emanuel A. 2007 Sequence Organization in Interaction. A primer in conversation analysis I. Cambridge: Cambridge University Press. Schegloff, Emanuel A. and Harvey Sacks 1973 Opening up closings. Semiotica 8: 289–327. Schubiger, Maria 1958 English Intonation: Its form and function. Tübingen, Niemeyer.

Pragmatics and prosody: prosody as social action

509

Selting, Margret 1996a On the interplay of syntax and prosody in the constitution of turn-constructional units and turns in conversation. Pragmatics 6: 371–388. Selting, Margret 1996b Prosody as an activity-type distinctive cue in conversation: the case of socalled “astonished” questions in repair. In: Elizabeth Couper-Kuhlen and Margret Selting (eds.), Prosody in Conversation, 231–270. Cambridge: Cambridge University Press. Selting, Margret 2000 The construction of units in conversational talk. In: Language in Society 29: 477–517. Selting, Margret 2004 The “upward staircase” intonation contour in the Berlin vernacular: An example of the analysis of regionalized intonation as an interactional resource. In: Elizabeth Couper-Kuhlen and Cecilia E. Ford (eds.), Sound Patterns in Interaction, 201–232. Amsterdam: Benjamins. Selting, Margret 2005 Syntax and prosody as methods for the construction and identification of turnconstructional units in conversation. In: Auli Hakulinen and Margret Selting (eds.), Syntax and Lexis in Conversation, 17–44. Amsterdam: Benjamins. Selting, Margret 2007 Lists as embedded structures and the prosody of list construction as an interactional resource. Journal of Pragmatics 39: 483–526. Selting, Margret, Peter Auer, Birgit Barden, Jörg Bergmann, Elizabeth Couper-Kuhlen, Susanne Günthner, Christoph Meier, Uta Quasthoff, Peter Schlobinski, and Susanne Uhmann 1998 Gesprächsanalytisches Transkriptionssystem (GAT). Linguistische Berichte 173: 91–122. Selting, Margret, Peter Auer, Dagmar Barth-Weingarten, Jörg Bergmann, Pia Bergmann, Karin Birkner, Elizabeth Couper-Kuhlen, Arnulf Deppermann, Peter Gilles, Susanne Günthner, Martin Hartung, Friederike Kern, Christine Mertzlufft, Christian Meyer, Miriam Morek, Frank Oberzaucher, Jörg Peters, Uta Quasthoff, Reinhold Schmitt, Wilfried Schütte, Anja Stukenbrock, and Susanne Uhmann 2009 Gesprächsanalytisches Transkriptionssystem 2 (GAT 2). Gesprächsforschung Online (http://www.gespraechsforschung-ozs.de/). Szczepek Reed, Beatrice 2004 Turn-final intonation in English. In: Elizabeth Couper-Kuhlen and Cecilia E. Ford (eds.), Sound Patterns in Interaction, 97–118. Amsterdam, Benjamins. Szczepek Reed, Beatrice 2006 Prosodic Orientation in English Conversation. Basingstoke: Palgrave Macmillan. Tanaka, Hiroko 2000 Turn-projection in Japanese talk-in-interaction. Research on Language and Social Interaction 33.1: 1–38. Tanaka, Hiroko 2004 Prosody for marking transition-relevance places in Japanese conversation: The case of turns unmarked by utterance-final objects. In: Elizabeth Couper-

510

Elizabeth Couper-Kuhlen

Kuhlen and Cecilia E. Ford (eds.), Sound Patterns in Interaction. 63–96. Amsterdam, Benjamins. Tarplee, Clare 1996 Working on young children’s utterances: Prosodic aspects of repetition during picture labelling. In: Elizabeth Couper-Kuhlen and Margret Selting (eds.), Prosody in Conversation, 406–435. Cambridge: Cambridge University Press. Trager, George. L. and Henry L. J. Smith 1957 An outline of English structure. Studies in Linguistics: Occasional Papers 3: 40–52. Uhmann, Susanne 1992 Contextualizing relevance: On some forms and functions of speech rate changes in everyday conversation. In: Peter Auer and Aldo di Luzio (eds.), The Contextualization of Language, 297–336. Amsterdam: Benjamins. Walker, Gareth 2004 On some interactional and phonetic properties of increments to turns in talk in interaction. In: Elizabeth Couper-Kuhlen and Cecilia E. Ford (eds.), Sound Patterns in Interaction, 147–170. Amsterdam, Benjamins. Wells, Bill and Sue Peppé 1996 Ending up in Ulster. Prosody and turn-taking in English dialects. In: Elizabeth Couper-Kuhlen and Margret Selting (eds.), Prosody in Conversation, 101– 130. Cambridge: Cambridge University Press. Wells, Bill and Sarah Macfarlane 1998 Prosody as an interactional resource: Turn-projection and overlap. Language and Speech 41: 265–94. Wichmann, Anne 2001 Spoken parentheticals. In: Karin Aijmer (ed.), A Wealth of English, 177–193. Göteborg: Gothenburg University Press. Wilkinson, Sue and Celia Kitzinger 2006 Surprise as an interactional achievement: Reaction tokens in conversation. Social Psychology Quarterly 69.2: 150–182. Wright, Melissa 2007 Clicks as markers of new sequences in English conversation. In: Proceedings of the 16 th International Congress of Phonetic Sciences, Saarbrücken, 1069– 1072.

18.

Pragmatics and literature Jacob L. Mey

1.

Introduction

Literary pragmatics is about how language is used in the production and consumption of literary texts. Since much of literature is about telling stories (including the ‘stories’ that we value as lyrical expressions of a poet’s feelings), the notion of narrativity looms large here. To guide the author and reader in their collaborative work of creating and re-creating the fictional universe of the story, the concept of ‘voice’ is of paramount importance. Voices represent characters, as they are created by the author and perceived, ‘re-created’, by the reader; it is the characters that in the end determine the story’s importance rather than the simple plot, or the ‘action’, as it is nowadays called. Important to notice here is that even when the storyteller narrates own or others’ real experiences (as in autobiography, or when expressing a reflection on a factual event), the universe that is created by this narration is still entirely fictional, and so are the characters, real or imagined, that populate the story and the voices that express their respective ‘points of view’. The reader collaborates with the author in identifying (‘re-creating’) the characters, as they are entering on the literary scene and voicing their parts. But the voices do not always sound in perfect unison or dramatic counter-point; when voices become mixed up and ‘clash’, the reader becomes confused and may abandon the common literary enterprise altogether. Thus, voice clashes (along with the related phenomena of transition between, and confusion of, voices) offer an interesting indirect perspective on the roles of authors and readers in the literary enterprise. These roles will be seen as intimately interrelated, indeed dialectically presupposing each other: the literary enterprise is a collaborative effort.

2.

Storytelling and stories

Contrary to what many people believe, storytelling is never a matter of ‘getting the facts, all the facts, and nothing but the facts’. What makes a story interesting is the way we consider those facts as they are presented to us by a narrator; this is what makes a story different from a simple retelling of factual happenings, as in the familiar kind of grade school composition that starts each sentence with “And then …”. Stories are part of our lives: they extend into our lives and branch out from there; but also, stories can only be understood if they are localized, fleshed out in

512

Jacob L. Mey

terms of our daily reality. A story that has no relation to my world is probably not going to interest me very much. Stories of earlier centuries (like the few surviving novels of classical antiquity or the interminable chevaleresque poems of the Middle Ages) appear often boring to a modern reader by their monotonic, endless repetition of the same themes: but these very qualities may have been what endeared this literature to the contemporary readership and contributed to its popularity. Similarly, the 18th century British author Samuel Richardson’s Sir Charles Grandison (an endless epistolary novel in seven ( ! ) volumes, called ‘Series of Letters’) must have received some appreciation among its contemporaries; but already a century later both Jane Austen (in Northanger Abbey) and Aleksander S. Pushkin in Evgenij Onegin saw fit to quote the hapless hero as the paragon of reading-induced boredom; compare Pushkin’s sneering reference to ‘Grandison, the precious bore Whose gallant lines provoke our snore’ (bezpodobnyj Grandison kotoryj nam navodit son …) Evgenij Onegin (ch. III, verse ix; my translation) Stories like Grandison and its predecessors Clarissa and Pamela seem end- and pointless, precisely because we cannot relate to the reality from which they derive and to which they point back. On the other hand, we also find stories in older literature, which today seem as fresh as when they first were written down. I already mentioned Austen and Pushkin: both furnish outstanding examples of stories that are as fascinating and pageturning today as they were when they were written, nearly two centuries ago. For who among us has not chuckled over the pointed ironies of Austen’s characters, or swooned with Tatiana in her romantic, unrequited love for Onegin? Or (to take an even earlier case): Boccaccio’s Decameron is, after three quarters of a millennium, still good for a thrill and a laugh, on a great number of exciting levels. While it is difficult to pinpoint what makes a story succeed, it is perhaps easier to spot at least a few of the causes that make a story fail, or let it fall into oblivion after the story producers and the consumer generation that supported them have passed away. In his Evgenij Onegin, Pushkin satirizes the ennui of the Russian genteel country nobility by pointing to the fabulous array of ‘has-been’ authors who constituted the daily spiritual bread of the country manor, and whose fame faded as their pages yellowed (the incomparable Grandison, whose bore is matched by our snore, is only one among many). In this and similar cases, the interested readership is confined to students of literature or history; the regular consumers will find it difficult to relate to the characters and their voices, as these represent an entirely different social and societal environment. The question thus is what makes for a story, and in particular, for a

Pragmatics and literature

513

good story. Literary scholars and the common readership have pondered about this question for centuries, both theoretically, in studies of narrative, and practically, in university courses of the ‘Creative Writing 101’ type and in popular ‘do-it-yourself’ manuals. In the following, I will first consider what we mean by ‘narrativity’: briefly, the question is about the essential features of story-telling. Next, the ‘voices’ of the narrative will be brought into play, and I will show how they interact, and sometimes conflict. After that, I will identify some, but not all of the linguistic and pragmatic devices that authors use when they tell their stories and that readers rely on for a successful interpretation and consumption of the narrative. And finally, I want to go back to the key figures in the process of literary production and consumption: the authors and the readers, as seen at work in their collaborative efforts, and as perceived through their respective voices.

3.

Narrativity and voices

Narration is about storytelling. But a story has to be told by someone; let us call him or her the ‘narrator’. Already, it is clear that the narrator is different from the author: in every story, the author creates a world of fiction, the narrative, in which the narrator plays an eminent role, even though not always discernible on the surface. Narrativity thus deals with the techniques and devices that a narrator has at his or her command when telling a story; among these, the notion of ‘character’ and the accompanying ‘voice’ are among the most important. To see how character and voice connect and interrelate in sometimes rather oblique and inscrutable ways, consider the case of the ‘homodiegetic’ novel, or stories told in the first person. Clearly, we cannot assume that the speaking ‘I’ is identical to the author, who likewise cannot be held responsible for the narrator’s claim as being his own. When the 19th century Russian Mikhail Yu. Lermontov starts his classic story A hero of our times (1841) with the famous line “I traveled by stagecoach from Tbilisi …”, every reader will understand that this does not mean that Lermontov himself did the traveling: the words are spoken by the narrator (referring to his conversation while sitting on top of the coach with a returning NCO, the ranking shtabskapitan serving in Chechenya). So, the author is not the narrator, not even if the story is told in the first person singular, as in the Lermontov case. The author creates the narrator as the embodiment of the story’s authoritative competence; it is really not important whether this authority assumes the ‘I’-role or pretends to be an outside ‘voice’: ‘Trust the tale, not the teller’, as the old adage has it. What characterizes the story is its ‘narrativity’, its quality of being narrated and received as a story. The latter part is important, but it is often overlooked; Robert Scholes has pointed out that the very idea of narrativity is dependent on the reader actively entering the story world and partici-

514

Jacob L. Mey

pating in the narration: “a process by which a perceiver [author or reader, JM] actively constructs a story from the fictional data provided by any narrative medium” (1982: 60). Thus, on the part of the reader, there corresponds to the act of narration an act of active collaboration, an ‘active narrativity’. However we define narrativity, there is a need to distinguish between author and narrator. A further necessary distinction is that between narrator and actor, and this is where the concept of narrative level comes in. Following Bal (1985), I distinguish between a primary level, where the action unfolds, and a secondary (and possibly further) level(s), where the actors come in and their voices are heard. On the primary level, we have the narrator’s text; on the secondary (“embedded”) level, we are confronted with the actors’ text. The latter text does not necessarily contain only what the actors say or think in the context of the primary level; they may start a story of their own, and this embedded narrative may then again have several levels of its own, where the secondary story is related to the primary one in various ways. A good example is provided by the classic novella by the German author Theodor Storm, Der Schimmelreiter (‘The man on the white steed’, 1888; Engl. transl. ‘The Dyke Master’, 1996), where the narrator on the primary level is the authoritative retired school teacher, who ‘frames’ the story by telling it to his friends in the local pub (in German, this kind of narrative is called a Rahmenerzählung, lit. ‘a story within a (narrative) frame’). But the narrative is not about the teacher and his audience; rather, it concerns the legendary local official in charge of dams and levees (in German: der Deichgraf, ‘the dyke master’), whose expertise and innovative efforts form the secondary narrative, which in fact is the story that we, the readers, co-construct and remember. The teacher’s primary voice, as perceived in his narration and buttressed by his memory, lends credence and local coloring to the embedded story and its secondary voices. Similarly, in one of the most famous collections of stories ever told, the 1001 Nights, the main character, Sheherazade, invents and ‘orchestrates’ the stories, which she then attributes to secondary narrators on the embedded level. These secondary narrators remain implicit, that is, they are heard, but not seen; in contrast, another famous collection of stories, Boccaccio’s Decameron, has the actors on the primary level narrate their own stories as narrators of secondary level-stories. In what I have said so far, I have implicitly assumed that actors ‘speak’ in narratives; in other words, I have attributed them a voice. But how does this happen? More precisely: How do we know whose voices we are hearing in a narrative? How do we decide if what we are ‘hearing’ (and this includes of course ‘reading’) is spoken by the narrator, by an actor (primary or secondary), or by some other (‘third’) party in a story? To better understand, and answer this question, let me introduce another notion: that of ‘focus’, or focalizing, as it is often called (Bal 1985). Focalizing is what makes an uninteresting series of events into a potentially fascinating story. To illustrate my point, I will recount a personal happening in

Pragmatics and literature

515

which some unrelated events somehow suddenly coalesced into something resembling a story.1 At one time during the winter semester of 1994, I was sitting in the sauna at the Crown Aquatic Center of Northwestern University, Evanston, Ill. I was enjoying myself, thinking about nothing in particular, paying not much attention to what was going on either inside or outside of the sauna room. People came in and went out again; some took a shower, some stayed longer, some shorter, some were reading their newspapers, others were talking, some were throwing water on the stove to produce steam and get the temperature up. All this had been going on for quite a while without my paying much attention. But after a while, I became aware of a balding, middle-aged man sitting on a chair outside the sauna and looking in through the big glass door. I remembered him vaguely: he had been inside and gone out, then had come back in again; now, having taken another shower, he was sitting outside, watching. I am directly in his line of vision, and suddenly my perspective is no longer that of one watching a scene with some persons and objects in the background, but of me watching some person watching me. In other words, I start focalizing. My immediate, conscious reaction is, of course, that I am making things up. The man may be there for a number of reasons, and why should he be watching me? Still, something interesting happens the moment I realize I am no longer observing an inconsequential series of events, or looking at some irrelevant piece of humanity, but at somebody who is somebody, even though I don’t know him. When I began to focalize, what happened is that I set in motion a narrative process: in other words, I started making up a story ‘in my head’. Of course, most of this story is just daydreaming, and it happens somewhere below the explicit level of consciousness. For example, depending on my previous experiences (or, if you want, my private hang-ups), I attribute a character to the man, trying to figure out if this is somebody I know, or ought to remember. In a more interesting scenario, I might consider him to be an elderly homosexual, waiting for me to get out of the sauna so he can make a pass at me. Now, even though my story can be said to be pure, self-centered fantasy (and even though I may know it to be exactly that), in terms of narrativity it is still the case that I have given this person a role as a character in my story, an actor in my little play; and the next thing I know is that the character starts speaking, starts having a voice. In my head, I even hear snatches of dialogue: the character is asking me questions (‘Do I come here often; where do I live; what kind of music do I like; do I have a phone number’, and so on), and I listen to myself providing non-committal, evasive answers (‘No, not really; somewhere in South Evanston; no particular preferences; no, I don’t have a phone’, etc.). Thus, my focalization makes an indifferent piece of my environment come alive in my self-constructed narrative; from being a silent prop, the person outside the sauna becomes a ‘talking head’. Focalizing implies vocalizing, that is, providing a voice to a character in a narrative.

516

Jacob L. Mey

No matter that the above fictional dialogue never took place, and never will, and discounting any possible psychological explanations of my sauna daydreaming (earlier experiences, repressed feelings, anxieties, or what have you), what this little anecdote is supposed to show is that we cannot separate focus (and in general, the narrative perspective) and voice. Moreover, focalization presupposes a point of view, inasmuch as it belongs to a character; hence, every focalization happens according to a character’s perspective. Similarly, every character’s voice presupposes the existence of a focalizing perspective in which that character is vocalized. In the next section, I will look at how the characters’ voices interact, and how they sometimes clash.

4.

Characters and voices

In a story, the characters come to life through their voices. The characters are kept apart, and their voices separated, by means of simple typographic means (such as quotation marks or line shifts) or in other, more sophisticated ways; the point of these techniques is to permit the reader or listener to keep track of what is going on, to place and/or reposition the focus of his or her attention. Consider now the following extract: “He [Jack Kemp] joined the Buffalo Bills after an injury. ‘I hit a helmet with my passing hand and dislocated a finger so severely that I had to literally decide what shape I wanted it in. So I put my hand on a football and they put a cast over it.’ In those days football was pretty Darwinian. ‘We’d do anything to survive.’ I glanced at the finger. Yep.” (From an interview with the late former Republican U.S. Congressman Jack Kemp, who had been a professional football player in the major leagues NFL/ AFL; Washington, D.C., October 1994; Parks 1994: 22) In the above story, we have first the voice of the narrator, the journalist who is telling us the story of his interview with Jack Kemp and provides both a historical frame (“He joined the [minor league team] Buffalo Bills after an injury”) and a running comment (“In those days football was pretty Darwinian”). Then there is Kemp’s own, quoted voice, telling us how he “hit a helmet with [his] passing hand and dislocated a finger”. According to the conventions of journalism, we are expected to believe that this was how Kemp spoke while being interviewed; it also provides the background for the continued story (again, in ‘Kemp-speak’) of what happened to the speaker’s hand (“I had to literally decide what shape I wanted it in. So I put my hand on a football and they put a cast over it”). But third, we encounter a voice that has not been heard so far, and a character that unmistakably is neither Kemp nor the journalist-narrator of the interview. The

Pragmatics and literature

517

focus of the narration has changed: what we are hearing in the short utterance “Yep” is the voice of the journalist himself, as an active participant in the narrated scene who suddenly makes his presence in the interview explicit (as opposed to earlier, when he was only the ‘implicit’ narrating character telling us, in a running comment, that “He [Jack Kemp] joined the Buffalo Bills after an injury”). This “Yep” does not belong to the interview, as it is not part of any dialogue; it represents the journalist Parks’s comment to himself on what Kemp has just told him about the consequences of his injury. This mental comment figures as a kind of reality check: Parks looks at the finger in question (and “Kemp’s football-curved hand”, as he calls it later on), and affirms the correctness of the story by inspecting the evidence. But notice that this affirmation is not part of the interview as narrated. It is the author’s internal, unspoken comment, ‘uttered’ to himself: ‘I looked, and there it was all right: a football-shaped finger’. Even though the tone of the commentator’s voice belongs characteristically in spoken language (being recognizable as such by its typically ‘conversational’ nature: ‘yep’, unlike its formal equivalent ‘yes’, belongs in direct, spoken discourse). The comment is not spoken, and the person ‘uttering’ it does not belong to the interview scene. One could perhaps say that Parks, the journalist, intrudes upon the scene by offering his two-bits of commentary ‘aside’, as in an Ibsenian stage direction. His voice is not heard directly, but rather as a kind of ‘ventriloquizing’, as the Russian semiotician and linguist Mikhail M. Bakhtin has called it.2 It is (self-)reported by the character as an unspoken confirmation of what just has been stated by the interviewee in the reported, uttered dialogue. Summing up: In the extract above, we have first the (implicit) narrator, the journalist: his (unheard) voice records the interview for the magazine publishing it, The Smithsonian (the official publication of the Smithsonian Institution in Washington, D.C.). Then, we have the reported voice of the ex-football player-cum-congressman, Jack Kemp: he is quoted by his very words, in ‘direct discourse’. The third voice is that of the journalist, now ‘coming out’ so to speak: he sheds his protective gear and changes his status from ‘implicit’ to ‘explicit’ narrator: “I glanced at the finger”; and he follows up by quoting himself in an internal, likewise unheard but clearly identifiable sotto voce: “Yep”. From this little exercise in ‘voicing’, we see that in narrative, characters tend to maintain their respective voices (and their corresponding stances, or ‘points of view’) throughout the narration. They do not speak out of order or out of turn; their vocalizations help us focalize, such that we do not lose the thread of the story, or are uncertain of who is speaking or what any of the characters is saying. In the next sections, we will see that voices not only collaborate in unison but that they sometimes engage in intra-narrative (and sometimes even extra-narrative) conflicts: the case of the ‘clashing voices’.

518

Jacob L. Mey

5.

When voices clash

5.1.

Free Indirect Discourse (FID)

In the present section, I want to examine the phenomenon of vocalization under the general angle of what I above called ‘focalization’ (often referred to, as I have done above, by the term ‘point of view’). Since the point from which a voice is heard is directly dependent on the place of the ‘voicer’ in the story, it will be useful to examine this place (and its surrounding discourse) in some detail. I will illustrate my findings with appropriate examples culled from literary and other discourse; in particular, many of my examples will be taken from what is often referred to as ‘free indirect discourse’, FID, or ‘represented speech and thought’, as it is also called. The difference between direct and indirect discourse on the one hand, and FID on the other, is that in the former, we have explicit devices indicating the point from which the stream of narration emanates. This point is either localized in the narrative instance (as in most of traditional novel writing), or in one or the other of the actors (as it happens in dialogue); or finally (and exceptionally), in the author him- or herself. Whereas in both direct and indirect discourse, such origins are marked with conventional means (e.g. quotation marks, or some form of a verbum dicendi et sentiendi, or an inquit),3 in the case of FID we have no such formal marks. FID is especially favored among writers who exploit the technique known as ‘stream of consciousness’, a technique often prototypically associated with modern writers (such as Joyce or Hemingway), but by no means a recent invention or innovation. Here is an example from Jane Austen: “ … when [Catherine] thought of Henry, so near, yet so unconscious, her grief and agitation were excessive. The day which she had spent at that place had been one of the happiest of her life. It was there, it was on that day, that the General had made use of such expressions with regard to Henry and herself, had so spoken and so looked as to give her the most positive conviction of his actually wishing their marriage. Yes, only ten days ago had he elated her by his pointed regard – had he even confused her by his too significant reference! And now – what had she done or what had she omitted to do, to merit such a change?” Jane Austen, Northanger Abbey (Chawton edition, pp. 190–191).

In this fragment, even though Catherine is not mentioned as ‘saying’ or ‘thinking’, there is no doubt that the discourse is Catherine’s. The personal pronoun ‘she’ in the second sentence and onwards refers univocally to the novel’s heroine. But since there is no formal marking of this indirect discourse, we call it ‘free’, as opposed to regular indirect discourse, which is preceded by a verbum dicendi or sentiendi (such as ‘Catherine said to herself’, or ‘Catherine thought’).

Pragmatics and literature

5.2.

519

Voices clashing in discourse

The interesting thing about FID is that it seems so natural, so without effort, both on the part of the reader and of the author; yet, this naturalness is an illusion. It does not take much effort to provoke a ‘clash’ of voices, wreaking havoc in the discourse, such that the true origin of a particular voice, the character’s ‘point of view’ is obscured, or even made to disappear. The French semiotician Roland Barthes has suggested using what one could call a literary ‘litmus test’ in order to determine the true provenience of a particular voice in the discourse. The test consists in exchanging the pronominal reference for a narrated third person with a (narrating) first person, and then observing how the various voices get repositioned with regard to their origin in the discourse. The technique is applied by Robert Scholes to one of Ernest Hemingway’s short stories; below, I will repeat Scholes’ experiment and discuss its implications. According to Scholes (1982: 116–117), in Hemingway’s ‘A very short story’ (1965/1925), there are “two third person characters of apparently equal consequence” (1982: 116). Scholes goes on to show how the substitution of the first person for each of the two third persons in the narrative gives an absurd result in the case where the substituted third person in reality was not the original ‘voicer’ of the FID. That is to say, the point of view represented by this character is not that of the ‘I’ character of the story, which actually represents a “covert, first-person narration”, as Scholes remarks; but the same thing could, of course, be said of much of Hemingway’s work, making up its weakness as well as its strength.4 In the original story, the hero, an American soldier who was wounded in the Great War (‘he’), has to undergo an operation. While the hero is being readied for the operation table, he wants to make sure he behaves himself well in the eyes of the Italian nurse, Luz, with whom he has fallen in love; afterwards, during his convalescence, he takes pains to please her by performing some of her chores. “He went under the anæsthetic holding tight on to himself so he would not blab about anything during the silly, talky time … After he got on crutches he used to take the temperatures so Luz would not have to get up from the bed … As he walked back along the halls he thought of Luz in his bed.” (Hemingway 1965: 141–142), Here is Scholes’ first ‘transposition’, with ‘I’ replacing the occurrences of the ‘he’ third person: “I went under the anæsthetic holding tight on to myself so I would not blab about anything during the silly, talky time … After I got on crutches I used to take the temperatures so Luz would not have to get up from the bed … As I walked back along the halls I thought of Luz in my bed.”

520

Jacob L. Mey

And here is the same text with ‘I’ substituted for the ‘Luz’ third person: “He went under the anæsthetic holding tight on to himself so he would not blab about anything during the silly, talky time … After he got on crutches he used to take the temperatures so I would not have to get up from the bed … As he walked back along the halls he thought of me in his bed.” Clearly, the first transposition makes sense, and in fact has the same overall reader effect as does the original story. The occurrences of the third person ‘he’ keep their true ring, even when transposed to the first person; they continue to be attributed to the narrator’s point of view. The omniscient narrator is telling us what happens with the hero; likewise, the hero is able to tell us what is happening directly, the discourse emanating from himself rather than from the narrator. In the same vein, his thoughts are accessible to the omniscient narrator; there is no clash between the voice directly representing ‘he’, ‘his’, and ‘him’, and the voice of the narrator telling us the same things in narrative mode. The second transposition, though, is very odd and voice clashes abound. The overall point of view here is that of the ‘I’ person, the woman called Luz – but she is not the omniscient narrator, so how could she possibly know that the male hero thinks of her in his bed (which is in reality hers)? Similarly for the hero’s reservations about the “blabbing”: unless he has told her this (and unless we are informed that he has done so!), she has no ‘legal’ narrative access to this knowledge. Clashes like these, where the voices of the characters and that of the narrative ‘attributer’ or ‘voice provider’ apparently contradict one another, have their origins in the different points of view that the voices represent: clashing vocalizations are due to incompatible focalizations. While in the first rewriting, the original FID stands uncontradicted as the male protagonist’s discourse, a voice clash happens when we introduce other points of view, and still want to keep alive the pretense of an FID (as it is done in the second rewriting). In the first rewriting, the focalizer and the voice are one and the same, just like in the original; in the second rewriting, the voices are at war. In our example, the clashes between the different voices were made visible by the use of a syntactic device: a pronoun which is being switched around with respect to person reference. Clashes of this kind are, however, not bound to a particular syntactic device such as a referential pronoun; among others, the techniques of changing tenses and moods are frequently used as indicating a change in point of view. A clash between different points of view occurs whenever the changes are not made in accordance with the reader’s normal expectations. In this case, a shifted point of view is usually rendered in a different tense and mood than what had been used in the original discourse. This phenomenon is often called ‘back-shifting’: in FID, a present tense in direct discourse is shifted back into the past, whereas an indicative mood becomes a subjunctive.

Pragmatics and literature

521

Compare the following (constructed) examples (due to Bal 1985: 141): Direct Discourse: [Elizabeth said] ‘I refuse to go on living like this.’ (quotation marks; present tense, indicative) Free Indirect Discourse: Elizabeth would not go on living like this. (no quotation marks; past tense, subjunctive) In the latter case, the focalizing point is not Elizabeth’s (who is quoted in the first example as speaking the words herself, using her own voice); rather, the voice is that of the narrator (the ‘narrative instance’), crawling inside Elizabeth’s head, so to say, and rendering her monologue as one who was privy to the character’s speech and thought. The words, they are the words of Elizabeth; but the voice is that of the implicit narrator: the narrator acts as a Biblical ‘impostor’, taking on the identity of the character. As to the term ‘back-shifting’ itself, this is somewhat misleading. Actually, the point of view that is responsible for the FID may shift not only ‘backwards’ in time and mood, but also in a forward direction. Compare the following stretch of narrative (also due to Jane Austen): “[Catherine] travelled on for about eleven hours without accident or alarm, and between six and seven o’clock in the evening found herself entering Fullerton. A heroine returning, at the close of her career, to her native village, in all the triumph of recovered reputation, and all the dignity of a countess, … is an event on which the pen of the contriver may well delight to dwell; it gives credit to every conclusion, and the author must share in the glory she so liberally bestows. But my affair is widely different: I bring back my heroine to her home in solitude and disgrace; and no sweet elation of spirits can lead me into minuteness. A heroine in a hack post-chaise is such a blow upon sentiment as no attempt at grandeur or pathos can withstand. Swiftly, therefore, shall her postboy drive through the village, amid the gaze of Sunday groups; and speedy shall be her descent from it.” Jane Austen, Northanger Abbey, Chawton edition, pp. 190–191. In this extract, the shifting points of view are clearly marked by the shifts in tense: first, we have a shift from the narrative past tense (“[Catherine] travelled on”) to some gnostic authorial comments offered in present tenses (“A heroine returning … is an event”, and so on). Subsequently, the narrative is extended, in future tense, to embody the author’s own voice, when she is describing to her readers what she is going to do to her heroine: “swiftly shall her post-boy drive … speedy shall be her descent …”.

522

Jacob L. Mey

Admitting that such direct author interventions in the text are the exception in modern literary prose (John Fowles being one, fairly recent exception; earlier authors such as Anthony Trollope literally reveled in it), we have to acknowledge Austen’s skill in avoiding a clash of voices by making her presence on the scene clear, among other things, by using a different tense. Readers usually have no beef with this: if omniscience is the narrator’s privilege, then omnipresence (albeit in differing tense garb) may well be the author’s. But in order to be properly perceived, the textual roles and their voices have to remain clearly separated, not least with the help of the linguistic devices that are at the authors’ and readers’ disposal. The next section will discuss how meaning enters into the picture when it comes to properly distinguishing between voices, and avoiding their clashes. 5.3.

Voicing and reality

Voice clashes often have a semantic background, as when an actor pretends to command a knowledge that is not her or his at the time of speaking: we witness the author sneaking onto the scene, as it were, without any forewarning or contextual restrictions. In cases like these, the rules of discourse organization and voice attribution are not observed: what we get is rampant omniscience and an improbable authorial omnipresence. Often, the author omits to cloak her- or himself in the narrator role when describing events and sentiments that the actors cannot be familiar with; since the readers depend on the narrator for their being properly introduced to the specific parts of the literary scene where the narrative’s events take place, respectively the characters’ thoughts and sentiments are voiced, their entry into what is called the fictional space becomes blocked or hindered. Alternatively, the narrator loses control of the narration, and misconstrues the state of the world-atlarge, as it is reflected inside the narrative; anachronisms and erroneous place attributions (along with other malapropisms) are then apt to occur, all resulting in potential voice clashes. When this happens, the person to blame is ultimately the author who does not respect the limitations that the fictional space imposes on the narrator and the addressees. Here is an example from a modern novel; the narrative is set in the year 1943, and the female protagonist is enrolled in flight training at a Home Guard base somewhere in Texas. “As the summer deepened, the heat thickened. The only time she was ever comfortable was aloft, where the wind and altitude provided natural air-conditioning.” Marge Piercy, Gone to soldiers. (New York: Ballantine Books, 1987, p. 302) The voice heard here is that of the main character, a young airwoman volunteer; however, there is another voice, clashing with hers, that of the author who (without announcing herself or properly introducing her voice) comments on the heat in

Pragmatics and literature

523

terms that belong to a much later period. Air conditioners, as we know them today, were not part of people’s natural reference points in the forties; the most one could imagine in the line of space cooling technology were devices based on evaporation, often in combination with a fan or stuck outside of a car window, blowing humid air into the overheated space and lowering the temperatures by letting the excess heat be absorbed by the evaporating water. Thus, air conditioners were not natural comparison points for people in the early forties, and it is for this reason that the protagonist’s voice in this passage becomes entangled in a reality clash (usually referred to as ‘anachronism’).5 Or, to take another example from the same book (the speaker is an American soldier commenting on various places where he has been assigned): “They say Switzerland is the same scene on a cleaner scale. But you have to rank higher than I do and have more top honchos for your friends to pull that service.” (Piercy 1987: 182) Here, the voice clash is blatant: the soldier, who is speaking in 1943, uses a word that didn’t come into use until well after the war was over: ‘honcho’.6 The character’s voice reveals itself as containing a foreign element; the author (speaking in her own voice-that-knows) intrudes on the narrator’s voice by putting inappropriate words into the mouth of the character (letting him speak with a voice-that-cannot-know). It’s a bit like watching a thirties’ movie in which you suddenly spot a vehicle having flashing turn indicators instead of the traditional, semaphore-type directionals of the earlier period; or spotting a wrong period animal, such as a New World monkey hanging by its tail, displayed in a presumed painting by Hubert van Eyck, when all van Eyck could have possibly known in his days were Old World, tailless monkeys.7 Other instances of this type of reality clash occur characteristically when authors impute knowledge, feelings, or wants to their characters that either are impossible to verify or that could not be imagined in the actual lives of those authors. To take an example: in Susan Sontag’s novel, The Volcano Lover (1992: 182), a scene depicts the deposed Queen of France, Marie-Antoinette, on the scaffold, waiting to be executed, and feeling the guillotine’s yoke chafing around her neck. The narrative cannot be attributed to the queen herself, who is soon to die, and whose voice is only ‘heard’ internally and has left no record. The voice we do hear through the implicit narrator is that of the queen’s sister, Charlotte, Queen of Naples, who (across the waters) conjures up the horrid scene with the cart, the mob, the steps up to the scaffold, even the executioner’s muddy boots and the creaking of the gruesome machinery as the axe starts to fall. Here, the narrator vicariously invests the story with the feelings, observations, emotions, and horrors of the French queen’s sister, as well as providing the reader with a self-styled internal monologue in Marie-Antoinette’s voice.

524

Jacob L. Mey

In this case, too, we are confronted with a shift in point of view that only can be attributed to the twentieth-century author’s invisible appearance, in the guise of her implicit narrator, on the late eighteenth-century scene. Where earlier authors, such as Henry Fielding, would think nothing of entering the narrative scene in person, to openly deliver their narrative homilies to the reader (a tradition that still was alive in the days of Anthony Trollope, and even has been occasionally used by modern authors), in the scene depicted above the author remains unseen, while speaking through a variety of inconsistent mouthpieces. Apparently, however, neither readers nor critics seem object to this kind of ‘vicarious’ FID being ‘smuggled into’ the narrative.8 Contrary to the free indirect discourse that we find in texts like the Jane Austen passages quoted above – a discourse which is duly attributed to one of the characters (usually the hero/ine of the story), in the case of the quotes above, we have a ‘reality clash’ (either an involuntary one, as in the Sontag case, or one exploited for a particular purpose, as it is done by Robertson Davies). In either case, the semantics of the story are not borne out by the reality of the speaking voices. In fact, we are getting close here to the pragmatic aspect of voice clashing; I will deal with this in the next section.

6.

The pragmatics of voicing

6.1.

Positioning the voices

Pragmatic voice clashes are bound up with the notion of ‘narrative level’; by this, I understand the relative spatial positioning of the voices within the text. A voice may be speaking from within the text, as in the case of the narrator’s or an actor’s voice; or it can be heard as speaking from the outside, as it were, on an ‘extra-textual’ metalevel. In the latter case, we may be listening to what the author or another text-external authority has to tell us: their voices may emanate from an ‘authoritative source’, as it is often done in newspaper reporting, or they may be attributed to a scholarly authority (propped up by the proper bibliographic references), and so on. Voice positioning is important for a proper understanding of the text; but also, the proper attribution of a voice, no less than its proper use, is part and parcel of the pragmatics of voicing. Referring to an event is not just a matter of naming place and date; it requires an understanding of why this event is important for the interlocutors, and how a mention of the event should be voiced. As a clear case of the need for proper voicing, consider the way one’s condolences are expressed when we meet with a friend or acquaintance who has gone through the painful personal experience of losing a loved one. Here, correct voicing is so important that most people will not take any chances, but prefer to use some pre-set formula, like

Pragmatics and literature

525

‘Sorry to hear about your loss’, or ‘My prayers and thoughts are with you’ (the latter kind is usually restricted to written use). In the following sections, I will discuss the pragmatics of voicing from a number of angles, including that of ‘territory of information’ and ‘empathy’. I will conclude by offering some thoughts on the linguistic realization of voice positioning, and on some of the clashes that may occur when proper positioning is not observed. 6.2.

Territory of information and empathy

The late Japanese linguist Akio Kamio has, in a number of articles and books (e.g. 1994, 1995, 1997), offered a new perspective on the problem of where to properly position a text’s voice. His concept of ‘territory of information’ implies that we all have access to some information in ways that others normally do not; our territory is so to speak ‘sacrosanct’ in the sense that it cannot be intruded into or operated upon unless the ‘intruder’, the interlocutor, has been properly ratified with regard to his or her rights and duties. In certain cases, the territory is defined so personally that we cannot even refer to phenomena occurring within it without committing a linguistic or social gaffe. In Japanese, for instance, bodily phenomena, such as being hungry or thirsty, belong to the territory defined by the speaking ‘I’. Only the speaker can pronounce “I” about him- or herself, as Karl Bühler (1982: 103) remarked a long time ago. But also, only the speaking ‘I’ has access to his or her own feelings and sensory experiences; thus, saying “You are cold” to a person in Japanese is not proper voicing, since the position of the speaker is outside of the territory where such information is found. The interlocutor can only say something like “You look cold to me”, or “You seem to be cold”; and similarly for other expressions having to do with the ‘ego’ whose territory we are dealing with (like ‘being hungry’, ‘having a problem’, and so on). Here is another example, borrowed from Kamio (1995): In Japanese, it is usually allowable (and hence pragmatically correct) to be asked, and answer a question, about one’s wife’s age.9 But while it is OK for a husband to utter Kanai wa 46 desu ‘My wife is 46’ (a husband is supposed to be familiar with the number of years that have elapsed since his wife’s entry into this world), the same husband cannot properly put himself outside this territory of information by adding a ‘quotative’ maker (in Japanese -tte or soo) to the utterance. Hence both Kanai-wa 46 desu-tte ‘I hear my wife is 46’ and Kanai-wa 46 da-soo desu ‘My wife supposedly is 46’

526

Jacob L. Mey

are pragmatically ill-formed, as they do not reflect the proper positioning of the husband’s voice, viz., inside his and his wife’s shared territory of information. Clearly, the fact that there are limits to one’s territory of information is not particular to Japanese or any other language; but the specific ways in which the territory’s boundaries are expressed linguistically are often strictly bound to the respective territorialities of cultures and languages. Something similar is valid in the cases discussed by Heritage under the general labels of ‘accessibility’ and the ‘right to empathy’ (Heritage 2007). Congratulations, condolences, and other expressions of empathy should not be offered out of place: a faulty positioning of the speaking voice may result in a fatal clash. A student cannot properly congratulate his or her professor on the latter’s ‘brilliant lecturing’ (see Mey 2008 for a case in point). Similarly, as Heritage and Sefi remark, “advice giving carries problematic implications about the knowledge or competence of the intended recipient” (1992: 389); anybody who has witnessed or experienced cases of spontaneously alter-offered, well-intentioned but ill-received advice knows how toe-curlingly irritating such information-territorial transgressions can be. To enter or not to enter another person’s territory is not just a matter of fact; it has eminently to do with the pragmatics of voicing, and in particular with the proper positioning of voices within the proper territory. The linguistic mechanisms that are at our disposal to effect correct voice positioning and avoid voice clashes that result from undue changes of position or unwarranted infraction of territories are numerous. But more often than not they are insufficiently understood and clumsily managed by authors and readers; as a result, a clash of voices may occur. In the following section, I will consider one of the most widely employed mechanisms that are instrumental for a correct pragmatic positioning, to wit: the discourse markers often referred to as ‘sentence adverbials’, and discuss how they function pragmatically. 6.3.

Sentence adverbials

Among the most frequent mechanisms used for positioning the discoursal voices are the so-called ‘sentence adverbs’ (often called ‘sentence adverbials’, since they not strictly have to be what are usually called adverbs, but may be made up of longer stretches, phrases, and even entire sentences functioning ‘adverbially’, that is to say, specifying a verb or verbal expression). Such adverbial expressions have as their scope not just the (immediately) preceding or following verb (as in ‘I admire her greatly’), but may extend over a whole phrase or sentence. Their name, ‘sentence adverbials’, expresses this fact: they act as ‘ad-verbs’, added not just to a single adjective or verb, but potentially to whole phrases or sentences. In this way, they are often equivalent to full-blown adverbial clauses; for instance, in an utterance such as ‘To be honest with you, your article is not good enough for our journal’, I could replace the entire adverbial clause ‘To be honest with you’ by a single

Pragmatics and literature

527

adverb or adverbial phrase: ‘Honestly/In all honesty, your article is not good enough for our journal.’10 Sentence adverbials have the property of ‘coloring’ or characterizing the entire sentence they scope over. Thus, they may be used to express the utterer’s or writer’s viewpoint on what the sentence says; they serve to position the speaker’s or writer’s voice. Consider the adverb ‘happily’ in the following sentence (taken from a letter addressed by a prospective author to a journal editor): “The paper is formatted in MS Word; I will happily send you a diskette if necessary.” In the context, this tells the editor that the sender of the message is willing to provide a diskette with her paper, and moreover that she will be happy to do so. Alternatively, a sentence adverbial may be used to tell the addressee where the message in question is to be placed in the ongoing discourse; again, the sender’s voice is positioned properly thanks to the adverbial. An adverb such as ‘finally’ is often used to tell us that what follows is the last element in a series of arguments, examples, and so on. In this way, ‘finally’ is equivalent to a meta-statement by the author about his or her text: ‘I’m telling you now that the next argument or example is the last one’. And consequently, the sentence adverbial may transmit the additional message that there is no more to talk about, from the speaker’s point of view. Of course, such meta-statements may remain implicit in the text, or be expressed in other ways, as is often the case in highly ritualized text productions, such as sermons, political speeches, or newscasts. A particular inflection of the voice may accompany a summing-up statement (‘For National Public Radio, this is Melissa Brown’); in such cases we do not need a ‘finalizing’ adverbial phrase of the type ‘at the end of the day’, or ‘when all is said and done’. Alternatively, I may make use of the multi-purpose adverb ‘anyway’ to conclude a discussion by indicating my position as the (purportedly) final contributor to the debate, or to voice a need to return to the original argument – the same argument that was expressed in full sentential array by the judge in the medieval farce of Maître Pathelin, who coined the famous expression “Revenons à nos moutons” (literally; ‘let’s go back to our sheep’), meaning: ‘Anyway, enough of these digressions, let’s get back to the subject’. Sentence adverbials are extremely useful devices, not only for positioning the voices of the discourse, but also for assuming or denying speaker’s responsibility for an utterance. Thus, a sentence adverbial like ‘realistically speaking’ may indicate a speaker’s attitude to what is being said; ‘hopefully’ characterizes the speaker’s wishful expectations as to the success of the action announced in an utterance; and so on and so forth. In the language of the press and the courts, the adverbial ‘allegedly’ is universally used whenever there is talk of a person suspected of being involved in a criminal act, but without the speaker wanting to make a formal accusation. The sentence adverb ‘allegedly’ tells us that what we are going to hear does not necessarily reflect the speaker’s beliefs or intentions, but something he or she

528

Jacob L. Mey

has heard somebody else say: the speaking voice is located at a safe distance from the (recording or reflecting) ‘alleging’ voice, which is only properly heard through the mouthpiece of the speaker. Thus, on the level of the informative speech act of ‘assertion’, nothing is said that could incriminate, or create a legal liability for, the speaker person: the catch-all sentence adverbial ‘allegedly’ acts as a disclaimer for the whole utterance. Consider the following constructed example, in which a newspaper reporter quotes what transpired from the courtroom proceedings in the case of an armed robbery: “Allegedly, the defendant entered the 7&11 convenience store on South Congress at around 10 pm, demanded to be handed over the money, and when the store owner declined, proceeded to hit him over the head with an 11 in. pipe wrench”. Here, the speaker implicitly states that he is not responsible for the truth of the allegations: ‘allegedly’ is spoken with a ‘borrowed’ voice; that is to say, somebody (for whom the speaker is not responsible) is making he statement at his or her own risk. The responsibility for the truth of the utterance found within the ‘scope’ of the sentence adverbial is lifted off the shoulders of the speaker; the expression ‘allegedly’ marks the whole reported phrase (“the defendant entered the 7&11 convenience store …” and so on) as something that is merely alleged, and therefore not necessarily true as reported. In other words, the sentence adverbial in question functions as a disclaimer. Whatever is within its scope, is said to be outside the verbal jurisdiction of the speaker, outside his or her territory of information; consequently, the speaker cannot be legally held responsible for the effects of his or her words. Non-adverbial expressions may serve a similar distancing function. As examples, consider the words “Watch my lips”, famously uttered by the first US President Bush (George H. W.), when questioned about his promise not to raise taxes, or the fully-fledged disclaimers of the type ‘I didn’t mean to say that’. This latter type can often result in voice clashes commonly known as ‘paradoxes’: a celebrated case is that of the so-called Moore’s paradox, by which one denies what one just has uttered (classic example: ‘The cat is on the mat but I don’t believe it’, attributed to the philosopher G.E. Moore; Mey 1993: 156). Since sentence adverbials are so unobtrusive and impersonal, as contrasted with explicit, full sentence disclaimers, their voices are positioned in neutral territory, so to speak, while their anonymity prescribes the correct interpretation of the words uttered even in the absence of a concrete utterer: a speaker or writer. We are in a borderland of voices here: on the one hand, we have sentence adverbials that straightforwardly represent the attitude of the speaker like in ‘Strictly speaking, your request is out of order’, where the voice of the speaker is heard as strictly applying certain rules and regulations. On the other, we have cases where the speaker

Pragmatics and literature

529

distances him- or herself from what s/he is going to say, as in “Strictly off the record, I think you’re indulging in a frivolous interpretation of the law”. Here, the speaker wishes to be quoted ‘off the record’, that is, the words spoken are not to be recorded, that is to say, made part of some official account. But the wish to not say something is not equivalent to not saying that very thing; when I say “I don’t mean to insult you, but I think you’re not behaving like a gentleman”, I cannot countermand the insult, as the words have been spoken by what is rightly assumed to be my voice, even though that same voice now tries to distance itself from what it just has said, or is going to say. The speaker dons a speaking mask, so to speak, a persona in the classical sense of the term (like the masks worn on the Greek and Roman dramatic scenes): speaking on the authority of this voice (and possibly employing what is currently on the verge of becoming a standard disclaimer, viz., “speaking on condition of anonymity”), the voice of the speaker assigns the voice of the ‘alleged’ or quoted utterance its correct positioning. We are confronted here with a change of voice that we may experience as normal and legitimate, but also on occasion as representing a clash. The question may be raised whose voices we really are hearing in cases like the above. Most hearers and readers will not usually be in doubt in the normal course of events: it does not take too much reflection on the part of the addressee to correctly position the voices he or she is hearing and attribute them to the proper characters in the narrative or discourse. But there is of course some rhetorical trickery involved in this process. By forcing the addressee to reflect on the positioning of the narrative voices and on the characters’ mutual relationships, the speaker may gain time and improve his or her own position: while the addressee’s defenses are down, the speaker is able to get another thrust in edgewise. The clashing voices of the narrative recall the clanging swords of the classic duel, and it often takes an uncommon skill on the part of the attacked (s)wordsman to position himself correctly in the face of the attack. In less martial contexts, the clashing voices may limit themselves to spreading confusion, as in the following excerpt from a contemporary mystery story, where Viera, the protagonist, suddenly realizes that the character she had trusted, in reality is the main villain. The development of the narrative, as told by the narrator (the story is in the ‘homodiegetic’, that is, the ‘I’ mode, as opposed to the ‘heterodiegetic’, or ‘he/she’ mode) leads us to Viera’s discovery, and thereby opens up a new vision on reality for the story’s protagonist, causing her to reshuffle her ‘voice cards’. But what about Viera’s own voice: where does it position itself? In particular, what does the sentence adverbial ‘subconsciously’ in the following fragment tell us about this? (Viera is conducting an internal monologue): “Subconsciously, I realized …” (Elizabeth Powers, All that glitters. New York: Avon Books, 1981, p. 43).

530

Jacob L. Mey

Viera’s internalizing the changed state of affairs, the new reality, may aptly be described by the verb ‘to realize’, even when uttered by the narrative voice (recall what I said earlier about the omniscient narrator’s position in the story); but how are we going to deal with the sentence adverbial ‘subconsciously’? If the voice we are hearing is the voice of the character Viera, then the utterance ‘subconsciously’ in her monologue is not appropriately attributed to this voice. A person cannot realize something without being conscious of it; hence Viera could not properly utter the sentence fragment in question without provoking a clash of voices and an ensuing confusion in the reader’s mind. This is a pragmatic clash, and we can only solve the dilemma by appealing to a ‘super-voice’ that is not heard directly in the dialogue, but is supposed to be implicitly present: this the hidden voice of the ‘omniscient’ author, inserting itself into the discourse of the character, where it subsumes and suppresses the clashing voices of the monologue. We lend credence to this maneuver because we trust the author; but as readers, we experience a certain uneasiness. This uneasiness may be formulated in the following simple question: How could Viera ‘realize’ what she subconsciously was not even thinking, let alone expressing in words? It is as if the author were telling the character what its subconscious thoughts would sound like, when brought to the surface (where they of course would provoke a clash). By keeping it under wraps, so to speak, the author may avoid the clash, but not eliminate the reader’s uneasiness, caused by the narrator’s attempt to pull the literary wool over our eyes, as she tells us what to believe about the character Viera’s unexpressed state of consciousness – rather a tall order, even for an omniscient narrative instance.

7.

Conclusion: narration and conversation

From a pragmatic point of view, both the literary text and the text generated in conversation are manifestations of human language use. It may be useful to briefly compare the two uses in this concluding section, to see if and how they are different, and if any observed differences can shed light on the use of the literary voices that we have been dealing with so far., If one asks how language use differs from narration to conversation, one notices that the ways we use language in the two cases are rather diverse. Narration may include conversation, and conversely, conversations usually include stretches of narration; yet, their respective ways of ‘doing things with words’ are vastly different. Compare the meticulous wording that is the hallmark of literary prose, with the spontaneous utterances that characterize everyday conversation: in the first case, we have authors judiciously assigning the various voices of the text to their respective characters (where clashing voices should not normally, but do occasionally, occur); in the second, we are confronted with what sounds like a cacophony of

Pragmatics and literature

531

voices, some speaking in unison, some interrupting, some dropping out, some dominating, some subordinated, and so on. The conversation analysts have shown us that (despite apparent irregularity) the business of conversation is conducted according to strict rules of selection, turn-taking, and (non-)overlapping. Beginning with the ground-breaking work of people like Sacks, Schegloff and Jefferson (1974), a new discipline, Conversation Analysis, has seen the light of day and is enjoying great popularity among a crosssection of the linguistics and sociology fields. At first blush, the question ‘Whose voice are we hearing?’ may seem a bit inappropriate in the conversational context. In conversation, we will always be able to identify the current speaker (the one who ‘has the floor’, as it is called). The current speaker is the one who can be singled out by his or her use of the first singular (or sometimes plural) personal pronoun (the ‘speaking I’, as the German psychologist Karl Bühler used to call it; 1982: 103). By the same token, the passages of a novel in which the characters conduct conversation should be the ones that are the least problematic with respect to the positioning of those voices. But on closer inspection, the question of where the voices are coming from turns out to be more complicated. As we have seen in the preceding, voices are not always what they pretend to be, or ‘sound’ like; an implicit voice may represent an extraneous instance, speaking with the authority of a ‘superaddressee’ (as Bakhtin calls this instance), namely: “God, absolute truth, the court of dispassionate human conscience, the people, the court of history, science, and so forth” (1994: 126). Such a voice may not be as easy to detect as, say, the voice of a person using a local or social dialect; but it may manifest itself in numerous indirect ways, e.g. by the authority it invokes, the appeals it makes to the listeners, the value-laden judgments it proffers or reproduces, and so on. When Jane Austen rebels against the “common cant” which “undervalue[s] the labour of the novelist” (1947: 23), we hear the author’s voice deploring the poor appreciation allotted to novelists on the part of those whose voices belong in the specific discourse, ‘common’ to a particular type of readers: those who will pick up an Austen novel, but hide it in shame when caught in the act of reading. Similarly, when Virginia Woolf lets Clarissa Dalloway muse about the important meetings her husband Richard has to attend (“‘Armenians,’ he said; or perhaps it was ‘Albanians’”; Woolf 1990: 119), we hear, through the speaking voice of the protagonist, the implicit background voices of the societal discourse of the time, where the ‘Armenian question’ played such an important role – but not important enough for Mrs. Dalloway to be able to keep ‘Armenians’ apart from ‘Albanians’. Similarly in everyday conversation, voices may appear on the interlocutory scene that are not explicitly heard or observed, but stay hidden in the conversational backdrop. For instance, a person may ask an interlocutor a question that is not appropriate in the context, as it presupposes an authority on the part of the questioner that he or she may not have. The implicit voices we are ‘hearing’ when

532

Jacob L. Mey

such questions are denied an answer are those of society’s powerful instances (such as the police or the judiciary), that pervade the entire fabric of our society. While a policeman or -woman can stop me in the streets of an affluent neighborhood and ask me for my identification if he or she perceives something untoward, a regular neighborhood dweller is not allowed to do the same, except maybe in an act of ‘neighborhood watching’ (or in the US context) of making a ‘citizen’s arrest’. In the latter cases, such a person would be speaking with the (implicit or impersonated) voice of the police official, ‘ventriloquating’ (to use Bakhtin’s colorful expression) the sounds emanating from the official instance, the police ‘superaddressee’. In conversation as in narration, the proper positioning of the speakers and their voices is every bit as important as the produced ‘text’ (narrative or conversational) itself. In the pragmatic analysis of a text (narrative or conversational), correctly placing the voices in the discourse is as crucial for the readers as it is for the interlocutors. Both have to know, at any given moment of the (literary or conversational) interaction, where they stand in relation to the other participants, and where the protagonists of the literary production: viz., the characters (including the author and the narrator) of the narrative, respectively the interlocutors in the conversational situation, are located in the discourse. Pragmatics, when applied to narrative as well as conversational texts, raises a double question: From which point of view are we being addressed as hearers and readers; and: Whose language is being spoken, whose voices are being heard? In the preceding pages, I have tried to make it clear that, in order to answer these (not unrelated) questions, we have to rely on the linguistic and pragmatic concepts that I have referred to in the preceding pages, first of all, the central notion of ‘voice’ and its (harmonious or clashing) manifestations in literary and conversational discourse.

Notes 1. The story has been told earlier in my book When Voices Clash (Mey 2000: 148–150). 2. “The author speaks as it were through language, … a language that has somehow become objectified, that he merely ventriloquates [sic] it” (Bakhtin 1992:299). 3. These terms are taken from traditional Latin (school) grammar. Verba dicendi are, e.g., ‘to say, state, affirm, exclaim, …’; verba sentiendi express the thoughts or feelings of the subject (e.g. ‘to think, mean, believe, …’); inquit is the 3rd person singular of the defective Latin verb inqueo ‘to say, speak’. 4. Of course, as Scholes also remarks, “[t]he third-person narration of the original text is a disguise, a mask of pseudo-objectivity worn by the text for its own rhetorical purposes.” (1982:117) 5. According to the Encyclopaedia Americana, air conditioners for home use began to be installed in the U.S. in the late 40s-early 50s.

Pragmatics and literature

533

6. The word’s first ‘official’ mention (as far as I have been able to ascertain) is in the Oxford American Dictionary (1980). It is said to be derived from the Japanese hancho, ‘leader of a band’; it was first introduced into English, according to Wikipedia, in 1955. 7. And in this way, the ‘clash’ reveals the true date of the forged painting, ‘The Harrowing of Hell’, as the incident is described by the Canadian novelist Robertson Davies in the second volume of his Cornish trilogy, What’s bred in the bone. (Penguin 1985, pp. 350–351) 8. Compare the British critic Natasha Walter’s comments on this episode, in her review of Sontag’s work: “ … there are scenes to remember in the book: Marie-Antoinette on the scaffold, feeling the yoke choking her …” (The Independent, October 2, 1992, p. 27) 9. Or one’s own, for that matter. Other cultures may have other, special socio-cultural conventions of propriety; compare: “Women don’t have ages”, as Talleyrand supposedly remarked to Metternich at the Vienna Congress. 10. Sentence adverbials are among the lesser studied pragmatic discourse operators. The fullest treatment (by now somewhat dated) is by the Canadian-Czech linguist Eva Koktová (1986).

References Austen, Jane 1947 [1815] Northanger Abbey (Chawton edition). London: Allan Wingate. Bühler, Karl 1982 [1934] Sprachtheorie: Die Darstellungsfunktion der Sprache. Stuttgart and New York: Gustav Fischer. Bakhtin, Mikhail M. 1992 Discourse in the novel. In: Michael Holquist (ed.), 259–422. [First Russian publication 1975]. Bakhtin, Mikhail M. 1994 The problem of the text in linguistics, philology and the human sciences. In: Caryl Emerson and Michael Holquist (eds.), 103–131. [First Russian publication 1976]. Bal, Mieke 1985 Narratology: Introduction to the theory of narrative. Toronto and London: University of Toronto Press. Emerson, Caryl and Michael Holquist (eds.) 1994 Mikhail M. Bakhtin: Speech Genres and other late essays. Trans. by Vern W. McGee. Austin, TX: University of Texas Press. [First Russian publication 1986]. Hemingway, Ernest 1965 [1925] A very short story. New York: Scribner. Heritage, John 2007 Territories of knowledge, territories of experience: (Not so) empathic moments in interaction. Keynote speech at the XVth Symposium About Language and Society, Austin (SALSA), Austin, TX, April 14, 2007.

534

Jacob L. Mey

Heritage, John and Sue Sefi 1992 Dilemmas of advice. In: Paul Drew and John Heritage, Talk at work: Interaction in institutional settings, 359–417. Cambridge: Cambridge University Press. Holquist, Michael (ed.) 1992 The dialogic imagination: Four essays by M.M. Bakhtin. Austin, TX: The University of Texas Press. Kamio, Akio 1994 The theory of territory of information: The case of Japanese. Journal of Pragmatics 21(1): 67–100. Kamio, Akio 1995 Territory of information in English and Japanese and psychological utterances. Journal of Pragmatics 24(3): 235–264. Kamio, Akio 1997 Theory of territory of information. Amsterdam: Benjamins. Koktová, Eva 1986 Sentence adverbials in a functional description. Amsterdam: Benjamins. Mey, Jacob L. 1993 Pragmatics: An introduction. Oxford: Blackwell. Mey, Jacob L. 2000 When voices clash: A study in literary pragmatics. Berlin: Mouton de Gruyter. Mey, Jacob L. 2008 ‘Impeach or exorcise?’ Or, what’s in the (common) ground? In: Istvan Kecskes and Jacob Mey (eds.), Intention, common ground and the egocentric speakerhearer, 254–275. Berlin: Mouton de Gruyter. Parks, Edward 1994 Around the Mall and Beyond. The Smithsonian 25(8): 20–25. Piercy, Marge 1987 Gone to soldiers. New York: Ballantine Books. Powers, Elizabeth 1981 All that glitters. New York: Avon Books. Sacks, Harvey, Emanuel A. Schegloff and Gail Jefferson 1974 A simplest systematics for the organization of turn-taking for conversation. Language 50(4): 696–735. Sontag, Susan 1992 The volcano lover: A romance. New York: Farrar Straus Giroux. Scholes, Robert E. 1982 Semiotics and interpretation. New Haven, Conn.: Yale University Press. Storm, Theodor 1996 [1888] The Dyke Master. London, Angel Books. (German original: Der Schimmelreiter. Transl. by Dennis Jackson,). Woolf, Virginia 1990 [1925] Mrs. Dalloway (Harvest edition). San Diego, New York and London: Harcourt Brace and Company.

Part V Methods and tools

19.

Approaching the data of pragmatics Monika Bednarek

1.

Introduction: approaching the data of pragmatics

As the editors of this series, Wolfram Bublitz, Andreas Jucker and Klaus Schneider argue pragmatics is a “transdisciplinary subject” which has a “substantial and wide-ranging significance” “for nearly all areas of language description”. From the broad conceptualization of pragmatics that they envisage, pragmatics includes “approaches that are traditional and contemporary, linguistic and philosophical, social and cultural, text- and context-based, as well as diachronic and synchronic”. In this view, pragmatics shares interests with other related and not always easily differentiable linguistic sub-disciplines, such as discourse studies, semiotics, psychoand sociolinguistics, conversation analysis, or ethnography of speaking (see Wodak 2008: 3–4 for a list of seven dimension these fields have in common), sharing for instance the focus on language use or linguistic (inter)action. There is thus a common commitment of these linguistic sub-disciplines to language in use – the “pragmatics” of language. It seems reasonable to propose that any discipline interested in language use must have some kind of approach to its object of investigation – the data of linguistic (inter)action, or simply, the data of pragmatics. This includes a methodological standpoint from which to approach this data, some kind of research methodology. However, the commitment to particular methods or procedures often fails to deal with a range of important questions: where or from whom should the data be collected and how? (For example, shall we use texts that we obtain in interviews or published self-descriptions? Do we include the news on the notice board and the graffiti in the toilets? Are texts sufficient for our purposes or should we also carry out observations?) Equally undecided is the matter of how the data collection is to be organized. (For example, do we generate our texts in group interviews or individual interviews? In what order shall we do the interviews?) The question of how the material is to be stored also remains open. (For example, will audio recordings be sufficient or do we need video recordings? What rules of transcription shall we apply?) (Titscher et al 2000: 13–14, see e.g. Kreuz and Riordan in this volume on transcription systems).

Thus, while the focus of this chapter is on the use of different data in pragmatics and related disciplines (compare first paragraph above) rather than a general discussion about research methodology, epistemology, or evaluation criteria such as ˇ ermák 2001; Silverman 2004; reliability, validity, replicability (see Taylor 2001; C McNeill and Chapman 2005; Dörnyei 2007), reference will be made to the insights gained from different approaches and methodologies to data and the different per-

538

Monika Bednarek

spectives they provide on the linguistic data. The starting point of this chapter is thus the differentiation between different kinds of data, rather than between different kinds of methodologies (e.g. quantitative vs. qualitative).

2.

Issues of data collection and analysis

Issues that have to be addressed by all researchers using data are whether all data should be systematically collected before the analysis or whether further data can and should be added after some data has already been analyzed. This question relates to the design of data collections or corpora (see e.g. Wynne 2005, Voormann and Gut 2008). Another issue is the difference between inductive (data-driven) and deductive (theory-driven) research. In other words, do we analyze linguistic data according to some pragmatic theory such as politeness (Leech 1983, Brown and Levinson 1987) or do we deduce some pragmatic theory from the analysis of the linguistic data itself? This kind of distinction cuts across the use of different data, and different methodologies, such as quantitative vs. qualitative research. For instance, speaking of large-scale corpus linguistics, Tognini-Bonelli (2001) makes a distinction between corpus-driven and corpus-based linguistics. The term corpusbased is employed “to refer to a methodology that avails itself of the corpus mainly to expound, test or exemplify theories and descriptions that were formulated before large corpora became available to inform language study” (Tognini-Bonelli 2001: 65), and the term corpus-driven refers to an approach where “the commitment of the linguist is to the integrity of the data as a whole, and descriptions aim to be comprehensive with respect to corpus evidence” (2001: 84). Similarly, Bednarek (2006a) speaks of a text-driven methodology, where texts are analyzed trying to make as few a priori theoretical assumptions as possible before analyzing the text data, and a text-based methodology, where texts are analyzed using a previously established theoretical framework. In contrast to corpus-based/corpus-driven research, text-based/text-driven research is based on the manual (and more “qualitative”) analysis of small-scale text corpora rather than on automated quantitative large-scale corpus analyses (Bednarek 2006a). However, what Halliday notes with respect to corpus-based and corpus-driven research is also true for text-based and text-driven analysis: the two approaches represent a continuum rather than a strict dichotomy, and much research is situated somewhere in between (Halliday 2004: 24). This may also be the case more generally with respect to inductive vs. deductive approaches to data analysis in pragmatics. Other issues that need to be addressed by all researchers include legal, ethical and technical considerations, with many research proposals and projects needing the approval of ethics committees at universities, and much data collection and publication requiring sometimes costly copyright consent of respective copyright owners. Most spoken data also require consent of speakers to use and publish these

Approaching the data of pragmatics

539

data. Relevant technical issues concern questions such as: what technology is needed to collect the data (e.g. audio and video recordings, keying in of data, conversion of internet data into other formats), to store the data (e.g. large databases, large electronic storage facilities) and to analyze the data (e.g. computer software)? Such issues can place real constraints on the analysis of linguistic data and should in general be considered before data collection.

3.

Types of data

3.1.

“Non-attested” data

I use the term “non-attested” data to refer to any kind of data that has not been observed or recorded as having been produced by a particular speaker or writer. This covers predominantly the use of introspection – thinking and reasoning about language use based on one’s own knowledge and experience of language. The term “non-attested” does not, however, refer to recourse to anecdotal evidence – data that has been observed by the researcher at a particular time without necessarily having information on when it was observed, by whom it was produced, what the context of use was etc. This kind of ‘attested’ but anecdotal data can sometimes be useful, even if systematic data collection is to be preferred. For example, whereas Bublitz (1998) uses corpus evidence to argue that totally has a negative “semantic prosody” (Louw 1993), Bednarek (2008c) uses informal observations of the use of totally on broadcasts on MTV to argue that, in AE young adult language, totally does not exhibit such a prosody. Much “classic” research in pragmatics seems to rely on introspection, particularly that related to logics and language philosophy (Austin 1975, Searle 1969, Grice 1989, Kompa and Meggle in this volume). Adolphs (2008) goes as far as stating that most research in pragmatics “has not used attested examples of language-in-use” (Adolphs 2008: 7). One example is speech act theory, although more recent approaches in pragmatics study speech acts with respect to elicited (e.g. discourse completion tasks; see below) or corpus data. For instance, Adolphs (2008) focuses on describing and exploring lexico-grammatical patterns that are associated with speech acts. Thus, why don’t you can be related to the act of suggesting and could you just can be related to the act of requesting. She also explores the distinction between direct and indirect speech acts with the help of corpus data. In general, introspection can be useful for forming hypotheses, but these need to be applied to attested language data for confirmation, rejection or modification. Indeed, many corpus linguists would argue that what we think about language use does not always correspond to language use. The focus of introspection also seems to be more on what is possible in language use, rather than on what is actual, typical, salient or frequent.

540 3.2.

Monika Bednarek

Attested data: elicited and non-elicited

Attested data means language data which was actually produced by speakers/ writers on particular occasions. Such data can be elicited (by the researcher) or can have occurred more “naturally”, that is, without being produced in a research context. 3.2.1.

Elicited data

Under the heading of elicited data we can include interviews (involving researchers), verbal reports, elicited narratives, discourse completion tasks, and experimental data. 3.2.1.1. Interviews Interviews are a standard way of collecting data across the social sciences (see e.g. Miller and Glassner 2004), including linguistics. There are various kinds of interview data. Structured interviews use prepared questions, which are often standardized, and thereby pre-define topics, whereas unstructured interviews leave room for topics to be explored, and are more “like an informal conversation” (McNeill and Chatman 2005: 56). Semi-structured interviews feature a combination of the two. While unstructured interviews thus may be more “authentic”, they are also more subjective, less representative, and tend to involve a smaller number of participants. Depending on mode, interviews can be in written or spoken form (with written interviews shading into questionnaires; see below); depending on participants, they can involve only one participant or more than one. Sometimes “focus groups” are set up featuring discussions in which one or more researchers moderate the discussion and interaction between members of a group. There are at least two ways in which interview data are used in research. Either what people say in interviews is seen as a reflection or description of what they “actually” do/think or what people say in interviews is itself analyzed as a particular kind of linguistic construal. Benwell and Stokoe (2006) write: Unstructured interviews and focus groups as data collection methods have been the focus of much debate and discussion across the social sciences, largely due to the way they are often treated as a transparent description of opinion or unmediated report of practice. […] Some commentators argue that interview data must be treated as a researcher-driven, occasioned and situated account of identity and practices. (Benwell and Stokoe 2006: 194)

There is a point in arguing that interviews may tell us what people think but not necessarily what they do or even that interview data are too much influenced by the interview situation, relations between interviewer and interviewee, social values

Approaching the data of pragmatics

541

etc to tell us anything about the topic of the interview. Wetherell et al (2001) ask: “when a researcher interviews someone, how far can their talk be treated as a reflection of something else, such as established memories and ready-formed opinions? Should it instead be analysed as the place and process in which memories and opinions are constituted?” (Wetherell et al 2001: 7). Group discussions in focus groups “can raise new problems of ‘about’-ness: perhaps the interaction is now about group relations rather than the apparent topic of discussion?” (Wetherell et al 2001: 28). It may thus be more useful to use interview data to explore not what participants think about or how they handle the ostensible topic of the interview, but to focus the analysis on other issues of the linguistic data. For example, Galasin´ski (2004) makes use of interviews with 26 male and 29 female participants, where the interview topic was “a variety of issues related to middle age […], the experience of its onset, and its relation to family and working relationships” (Galasin´ski 2004: 24). He then explores the data with respect to emotion and gender, looking in particular at how “emotion narratives” (Galasin´ski 2004: 24) work in the construal of masculinity. In particular, participants can choose to distance themselves from emotions using a variety of “distancing strategies”, such as not directly describing themselves as experiencing a particular emotion. This is the case in this interview extract where a situation is described as being traumatic/dramatic rather than as directly and explicitly referencing the speaker’s emotion (as in the alternative I really hated it or I was really sad): (1) I: and can you describe what it was like when your daughter moved out? IF: well (.) permanently (.) it sort of crept up on us. I mean the night we left her at university (.) she didn’t go in halls, she went into a students’ house. And that night was pretty traumatic, the drive back down the M6 was pretty dramatic. […] (Galasin´ski 2004: 53, my bold) Such distancing strategies can be used by male speakers “to maintain the pretence of meeting the requirements of the dominant model of masculinity” (Galasin´ski 2004: 73). Such use of data thus includes considerations of context and dominant ideologies (e.g. prevailing models of gender), and seems a more interesting way of looking into emotion narratives than interviews where participants are asked about what they think about emotion and gender or where men and women are asked about their emotional experience. In general, interviews have a number of methodological advantages, as they are argued to be a natural and socially acceptable way of collecting information that most people feel comfortable with and which can be used in a variety of situations and focusing on di-

542

Monika Bednarek

verse topics to yield in-depth data. The interviewer’s presence allows for flexible approaches, probing into any emerging new issue, while the interview guide helps to maintain a systematic coverage of the domain. (Dörnyei 2007: 143)

However, interview data are clearly not “naturally occurring”; further, to set up and run interviews and to analyze the resulting data can be challenging and time-consuming; and there are also questions concerning the nature of the data with respect to the interview situation (as discussed above). As already noted, it is problematic to consider interview data as reflecting reality or producing reliable stories of experience; rather, interviews concern “the repetition of familiar cultural tales” (Miller and Glassner 2004: 125) and “the interview is obviously and exclusively an interaction between the interviewer and interview subject in which both participants create and construct narrative versions of the social world” (Miller and Glassner 2004: 125). 3.2.1.2. Verbal reports Verbal reports are self-reflections on participants’ language use, and are primarily used in second language studies, interlanguage studies and cross-cultural research (Woodfield 2007: 2). Verbal reports are employed by researchers to reconstruct cognitive processes in language use, and to find out about speakers’ “pragmatic knowledge and their choice and formulation of speech act strategies” (Woodfield 2007: 2). Verbal reports can be concurrent (on-line) e.g. as “think aloud”, or they can be retrospective (subsequent). However, it is necessary to point out that when interviewing participants after interactions, “participants might not themselves be able to recollect their own interpretations at a later stage” (Adolphs 2008: 41), and that trying to access cognitive processes is problematic because such processes may be unconscious and thus not amenable to self-reporting. As with interviews, it is worth problematizing whether or not verbal reports can be a “reflection” of actual behavior, cognitive processes and knowledge structures. Further, the influence of the researcher and fellow informants on the resulting verbal report cannot be neglected (Woodfield 2007, citing Haastrup 1987). 3.2.1.3. Discourse completion tasks Much cross-linguistic research in pragmatics (e.g. the Cross-Cultural Speech Act Realization Project) or research in Second Language Acquisition into speech acts uses oral or written discourse completion tasks (DCTs). DCTs were first introduced by Levenston (1975) for testing communicative competence and then employed by Blum-Kulka (1982) for speech acts (Rose and Ono 1995: 192). Written DCTs are questionnaires that involve a description of different situations, potentially some dialogue, and empty slots for subjects to produce a particular speech act or to fill in appropriate responses (e.g. to a compliment). For example, to elicit English re-

Approaching the data of pragmatics

543

quests, the following description may be used: “You have been asked by your lecturer to give a seminar paper but you need more time. You go to see her to ask for extra time to prepare your paper. Ask for an extension” (from Woodfield 2007, who also gives an overview of different request situations used in research). Participants in DCT resist such tasks to some extent, commenting on the non-interactive and inauthentic nature of DCTs and their lack of contextual data (Woodfield 2007). Although “only limited efforts have been made to evaluate the validity of this method” (Yuan 2001: 272), it has been suggested that both written and oral DCTs have a number of limitations in that they do not capture dynamic discourse features such as turn-taking, conversational structure, speaker-listener interaction, and pragmatic features (Kasper 2000, cited in Woodfield 2007; see also Yuan 2001). Thus, responses given in DCTs “are shorter in length, simpler in wording, less face-attentive and negotiatory in the negotiation process, and less emotionally involved than naturally-occurring speech” (Yuan 2001: 272). They may capture pragmatic knowledge rather than usage. In as far as utterances in DCT’s are more decotextualized, more decontextualized and less multimodal than natural discourse, they can be argued to neglect the complexity of language usage. Adolphs (2008) notes that “the fact that data generated through DCTs is based around single utterances distorts the overall picture of a speech act which is often negotiated and developed over a number of turns in a dynamic discourse event. It is therefore unclear whether DCT or role play data display the same patterns and attributes as naturally occurring data” (Adolphs 2008: 22).1 A further disadvantage of DCTs is that they are difficult to construct, and there is a “need for rigor in questionnaire construction and for extensive pilot-testing of DCTs to control for various possible effects for prompt, rejoinder, and so forth” (Rose and Ono 1995: 198). As with other questionnaires, general issues that need to be considered concern the sampling of the population of the participants (on sampling see e.g. Dörnyei 2007: 95–100),2 its design (e.g. formulation of situation in terms of appropriate detail, multimodality) and administering the DCT. On the other hand, DCTs, like other questionnaires, allow for control and quantification; they are replicable; they are not time-consuming; and they are versatile. For a more detailed discussion of DCTs in pragmatics see Woodfield (2007), who proposes that DCTs “represent highly constrained instruments of data collection in terms of the degree to which the data is predetermined by the elicitation instrument” (Woodfield 2007: 5). With respect to the distinction between oral and written DCTs, the former share more features of natural speech than the latter (Yuan 2001), and thus may be preferable. Yuan concludes that an oral DCT generates data that are, to a certain extent, representative of natural speech; it allows researcher control; it is efficient in terms of the amount of data it can elicit: and it is ethical. Nevertheless, […] the oral DCT suffers similar drawbacks as do the written DCT [sic] in that it cannot elicit elaborated negotiations and indirect compliment exchanges seen in everyday conversations. (Yuan 2001: 289)

544

Monika Bednarek

3.2.1.4. Experimental data Experimental data in pragmatics and related areas are used in educational research (see Dörnyei 2007), with respect to child language acquisition as well as in cognitive linguistics. For example, experimental research on understanding and recall asks subjects to reproduce sentences with the help of specific cues. Such research has suggested that understanding may proceed with the help of mental models (frames, schemas etc, see Bednarek 2005). With respect to the sentence The fish attacked the swimmer, the word shark worked better as a recall cue than fish, suggesting that readers constructed a mental model for the sentence (Brown and Yule 1983: 251, reporting on research by Anderson et al 1976). Experimental pragmatics bridges pragmatics and psycholinguistics. The 4th Experimental Pragmatics conference in April 2009 included research on topics such as semantic approaches, language development, impaired communication, non-verbal and non-human communication, non literal language, negation, presupposition, reference, conditionals, sentence processing. (http://experimentalpragmatics.org/, last accessed 17 October 2008). With respect to experimental research it is questionable whether it is possible to find out about how people behave “in real life” rather than in the artificial and simplified lab situation. Other problematic issues are related to how many subjects are used and how “representative” they are, and to what degree they are influenced in their language use by the experimental situation and the presence of the experimenter. Some of these factors may be less problematic in field experiments that do not take place in a lab but rather in a “real-world” environment such as a classroom. With the help of a carefully constructed experiment and control groups, it is argued that the advantage of experimental data is that they allow for the testing of hypotheses, the control of variables and the analysis of cause-effect relationships. (For a more detailed discussion of experimentation-based pragmatics see contributions to Noveck and Sperber 2004, Breheny in this volume.) 3.2.1.5. Elicited narratives An important strand of research on language use analyzes elicited narratives, employing various techniques (including the use of non-linguistic data) to elicit narratives from subjects. For example, participants are shown a story represented in a sequence of pictures or a film narrative and are then asked to retell the story. Thus, Bamberg (1991) shows subjects a 24-picture story called “Frog, where are you?” and subsequently asks them to tell the story depicted in this sequence of pictures. The data are then analyzed in terms of narrative structure and narrative activity. Similarly, Labov’s “classic” analysis of the structure of narratives (abstract, orientation, complicating action, evaluation, result/resolution, coda) is based on elicited narratives. Tannen’s (1993) data stem from a project in which small groups of

Approaching the data of pragmatics

545

young Greek and American women watched a film (the “pear film”),3 and afterwards talked about its contents to another woman, who they were told had not seen it (Tannen 1993: 21). Tannen analyzes the women’s accounts of the film and categorizes sixteen types of linguistic evidence such as negatives and repetitions which reveal their expectations or, in cognitive linguistic terms, their schemas (see above). Thus, elicited narratives allow a variety of linguistic enquiries about topics including narrative structure, the construal of narratives, and cognitive phenomena to name but a few. However, they share some disadvantages with interviews, for instance that they are not “naturally occurring”, and that it can be very time-consuming to elicit the data in the first place. With respect to all elicited data discussed in the above sections (interviews, verbal reports, discourse completion tasks, experimental data, and elicited narratives) as well as other kinds of elicited data not discussed here (such as role plays and different types of questionnaire,4 e.g. open or multiple choice questionnaires) the disadvantage is that they are “elicited” rather than “naturally occurring”. In contrast, the following sections look at different types of non-elicited data as used in pragmatics and related research into language use. 3.2.2.

Non-elicited data

Under the heading of non-elicited data we can include observational ethnographic data and different kinds of text/discourse data, including corpora. 3.2.2.1. Ethnographic data Ethnography has its roots in anthropology, in particular research by Bronislaw Malinowski and Franz Boas as well as in the linguistic works of Edward Sapir (Titscher et al 2000: 90). It is self-reflexive, culturally founded and based on participant observation (Titscher et al 2000: 90). Ethnographic methods have emerging research foci, based on what evolves from the researcher’s immersion in a particular context for a period of time (at least 6–12 months). Thus, participant observation is a crucial method in ethnography, which may be complemented by other techniques such as interviews or document analysis (McNeill and Chapman 2005: 89). However, not all participant observation is ethnographic: “A study becomes ethnographic when the fieldworker is careful to connect the facts that s/he observes with the specific features of the backdrop against which these facts occur, which are linked to historical and cultural contingencies” (Baszanger and Dodier 2004: 12). In other words, texts are interpreted in terms of cultural structures or used to reconstrue such structures (Titscher et al 2000: 91). The participant in participant observation can be actively involved in the social group/activity which is the focus of research (covertly or overtly) but need not be. His/her observations can be recorded in observational schemas, reports or field-

546

Monika Bednarek

notes. Data can be recorded and then transcribed or jotted down in notebooks as it occurs. Yuan (2001) points out that the latter type of linguistic data is not very accurate in capturing the details of language use. When participant observation is overt, the observer’s paradox (Labov 1972, cited in Wardhaugh 1986: 18) may play a role. This paradox is that “the aim of linguistic research is to find out how people talk when they are not being systematically observed, but the data are available only through systematic observation” (Wardhaugh 1986: 18–19). Ethnographic data can be argued to show the complexity and dynamics of authentic behavior; ethnographic methods are flexible and open to the observation of new phenomena. However, ethnographic research is time-consuming, not replicable or quantifiable, and not necessarily representative. It may, however, be “ideal for generating initial hypotheses about something totally unknown” (Dörnyei 2007: 133). While different types of ethnographic research exist (Baszanger and Dodier 2004: 19) in linguistics it is usually associated with John Gumperz and Dell Hymes and their ethnography of speaking or communication (e.g. Gumperz and Hymes 1972). For instance, Hymes is interested in speech communities, speech events, speech styles, and ways of speaking as units of analysis (Hymes 1972: 53–58). Within speech acts he proposes to look at various components such as setting, scene, participants, outcomes and goals, key (tone/manner), genres, norms of interaction and interpretation. Key questions in such research are: “Who speaks, with whom, when, where and in what code, about what?” (Titscher et al 2000: 95). 3.2.2.2. Texts/discourse data With respect to the use of spoken and written text/discourse data, we can identify a continuum in terms of size, with analyses ranging from individual case studies of one or just a few texts, to the use of small-scale corpora consisting of a range of texts to the recourse to large-scale corpora of hundreds of thousands or millions of words. Continuum of text/discourse data Individual text(s): case study/ies

small-scale corpus

large-scale corpus

Figure 1

Discourse data also vary in terms of their content and context of production such that analyses can focus on political documents, speeches, media texts, conversations, radio interviews, reality TV talk, books, academic articles, advertisements, songs, poems, television series, university prospectuses, chat room interaction or websites to name but a few. These types of data offer different advantages. For example, written language is usually easier to collect and to study than casual

Approaching the data of pragmatics

547

conversation. Public spoken language (e.g. radio or internet broadcasts) is easier to get hold of than conversations among friends or partners. Printed text is infinitely easier to collect than audiovisual data and more permanent than most Web material. It is already ‘out there’, ready to be gathered, and does not require time-consuming transcription before analysis. Also, the observer’s paradox is not an issue. Unlike, say, informants’ elicited responses or structured interviews, printed material is not in danger of changing just because it is being observed. (Mautner 2008: 32)

In contrast, [t]he variation in the types of contexts recorded as part of different spoken corpora, as well as in the parameters used for collecting data, makes it difficult to accumulate a large enough sample of spoken data that follows the same design principles. As a result, research based on spoken data is often limited in terms of the number of examples of a particular linguistic feature in the corpus, especially once contextual filters have been applied. (Adolphs 2008: 33)

With respect to the recording of spoken data, the observer’s paradox may again be relevant, since people may talk differently when they are being recorded, although there are techniques for minimizing this problem, such as recording people for long periods of time. Spoken data also needs to be transcribed according to specific transcription systems (see Kreuz and Riordan in this volume). Since the emergence of the internet and other technologies, data from the new media (such as emails, chats, social forums, message boards etc) are also increasingly studied in pragmatics (Bublitz 2008; Gruber 2008).5 The majority of the new media incorporate semiotic systems other than language, and multimodal research becomes important. Wodak (2008) even speaks of “the visual turn” (Wodak 2008: 2) in linguistics, necessitating new toolkits and theories. Multimodality is a very vibrant field in linguistics nowadays, with many different approaches and notation/transcription systems (e.g. Hübler 2007; Jewitt 2009). In general, multimodal analysis, whether of “traditional” or of “new” media, can become highly complex and time-consuming (see Pollak 2008 on film analysis). These and other types of spoken and written discourse data may be used in case studies, small-scale corpus studies and discourse analyses, and large-scale corpus linguistics. (Detailed descriptions of different kinds of corpora are given in most introductions to corpus linguistics.) The way in which these data are analyzed varies widely (see below and Andersen as well as Ruehlemann in this volume), but all methodologies have in common the focus on the actual (rather than the possible) and all value authentic language-in-use. There are a variety of approaches that usually analyze individual texts or small-scale collections of texts (including specially designed corpora) such as ethnomethodological conversation analysis and membership categorization analysis (Schegloff 2006; see also Holly and Domke in this volume), interactional sociolinguistics, politeness theory, discursive psychology, critical discursive psychology, content analysis, grounded theory, eth-

548

Monika Bednarek

nographic methods (discussed above), narrative semiotics, SYMLOG (system for the multiple level observation of groups), critical discourse analysis, (German) functional pragmatics, distinction theory text analysis, objective hermeneutics, and systemic functional discourse analysis. See Taylor (2001), Stubbe et al (2003), and Titscher et al (2000) on evaluations of and comparisons between these approaches to discourse data. The advantages of looking at single texts or a limited number of texts or smallscale corpora are that the texts can be analyzed in depth through close-reading and attention to manifold aspects of language use. Such an analysis allows researchers to focus on the localized construal of social phenomena such as identity in particular contexts, often resulting in a complex, rich, interpretive, dynamic and flexible analysis of micro-contexts and capturing the dynamic and negotiatory nature of much language use. If video data are used (Heath 2004), the multimodal context can be taken into account and attention can be paid not only to intonation and prosody but also to gesture, posture, body movement and facial expression. However, while individual case studies allow a focus on significant or important texts and events, they “lose out” on generalizability and representativeness. They raise the question of how far statements made about one particular text are transferable to other contexts. Case studies can also be time-consuming, relatively subjective, and are not easily replicable. Analyses that make use of small-scale corpora or collections of texts are more representative but can be just as subjective depending on the features analyzed and the method of analysis. In addition to the advantages of case studies, they offer observation of recurring regularities, practices or patterns, at least to some extent. Thus, conversation analyses have uncovered regularities of turn-taking and sequencing using collections of spoken data. Small-scale corpus studies have, for instance, been undertaken by Semino and Short (2004) with respect to reported speech and thought, by Bednarek (2006b) focusing on evaluative meaning in a 70,000 word corpus of hard news stories, by Martin and White (2005: 165) on a systemic functional linguistic appraisal analysis in 85 items from journalistic discourse and by Bednarek (2008a) focusing on emotion talk in an 85,000 word corpus comprising conversation, news reportage, fiction and academic discourse. The corpora used in this research are small enough for manual, context-sensitive analysis but large enough to show at least some patterns and allow some generalizability. Coding can be supported through computer assisted manual annotation (CAMA), “where a human annotates the text in terms of patterns that generally computers cannot recognize” (O’Donnell 2007). The latter makes use of smaller corpora because manual coding takes up time but it allows the researcher to explore semantic and pragmatic patterning. O’Donnell has designed computer software that facilitates such analyses: the UAM CorpusTool (http://www.wagsoft.com/CorpusTool/), but other coding/annotation software also exists for use across the social sciences. Finally, a few words on large-scale corpus data. Corpus linguistic applications include discourse analysis, lexicography, stylistics, forensic linguistics, language

Approaching the data of pragmatics

549

variation studies and language teaching (Baker 2006: 2–3). Although corpora have in the past been neglected in pragmatics, corpora are increasingly used in studies of language use, such as e.g. Adolphs (2008) and contributions to Skaffari et al (2005). Compare also the three sub-issues on corpus linguistics in the Journal of Pragmatics (2004: 36/2, 36/6 and 36/9). A crucial advantage of corpus linguistic investigations lies in the representativeness of the data and the use of empirical, systematic, and sophisticated data. The analyses are often reliable and replicable, especially when focusing on formally defined items. Further, corpus linguistics can uncover features of language that are inaccessible to intuition or a close reading type analysis of individual texts. This concerns patterning, typicality of usage, and quantification (e.g. type-token ratios, frequency lists, lexical density) and particular kinds of semantic-pragmatic meanings such as semantic prosody (Louw 1993, Bednarek 2008c). Corpus data also give researchers access to shared connotations (Coffin and O’Halloran 2005) and the experience of language users (Thompson and Hunston 2000: 15). They allow us to study typicality adequately through access to large amounts of actually occurring discourse, and enable easy application of tests of statistical significance through user-friendly software. Thus, hypotheses can be verified, falsified or modified and new language features can be uncovered with the help of large scale corpora. However, there are also some disadvantages6 in that researchers are only able to search the corpus for formally defined items, unless the corpus data have previously been annotated manually or semi-automatically (in which case there is a limitation on the size of the corpus because this process can be quite time-consuming). Social categories such as gender or age are often treated as stable variables rather than as discursively construed (Butler 1999), and many (though not all) corpus analyses fail to take into account reception, the dynamic nature of discourse, its context or textual structure. As Stubbs (2008) notes with respect to keyword analysis, “since the texts have been ripped apart into lists of individual words and/or n-grams, the patterns ignore text segmentation. They are a feature of global textual cohesion, but not textual structure” (Stubbs 2008: 5). It has also not yet been discussed sufficiently to what extent statistical tests are adequate for data of authentic language-in-use, to what extent linguists can evaluate and understand the information provided by statistics correctly, and to what extent they are able to choose the most appropriate statistical tests for their analyses. The majority of corpora are mono-modal, and the incorporation of semiotic systems other than language is not yet wide-spread. With respect to multimodality, Bauldry and Thibault (2006) stress the importance of using multimodal corpora in semiotic analysis while noting that their design is “very much in its infancy” (Bauldry and Thibault 2006: 181). Johnston (2008) has forcefully argued for the need for sign language corpora. Corpus analyses are also usually interested in the repeated, the typical, the frequent, rather than the individual and outstanding (but see e.g. Hoey 2005 on cre-

550

Monika Bednarek

ativity). For example, some texts are “highly valued in the community or […] [have] special significance in some domain such as history or politics. Here we treat texts as artefacts – objects of study in their own right” (Matthiessen 2006: 108). An emphasis on repeated patterns of semiotic behavior neglects the importance of outstanding singular texts that might have an impact on phylogenesis and ontogenesis that goes beyond the impact of “ordinary” singular texts that achieve their impact through repetition. An example from environmental discourse would be the documentary An Inconvenient Truth produced and presented by Al Gore, which has been widely credited with responsibility for a seismic shift in the attention of politicians, governments and the media towards environmental matters and in policy changes around issues of environmental sustainability. In fact, its impact was recognized in the awarding of the Nobel Peace Prize to Gore in 2007 for “efforts to build up and disseminate greater knowledge about man-made climate change, and to lay the foundations for the measures that are needed to counteract such change” (http://nobelprize.org/nobel_prizes/peace/laureates/2007/). The impact of such extraordinary texts, which are more powerful than normal texts produced countless times by speakers over time, presumably depends on diverse factors such as the profile of the author/speaker, the context of reception, and in the case of An Inconvenient Truth, the sophisticated use of semiotic resources (Bednarek, Caple and Hood 2008). Studying such unique texts, then, also provides important insights into the construal of categories, and is an argument for qualitative analyses of individual texts. Further, some publications are more read than others and thus might have greater impact on phylo- and ontogenesis:7 “According to claims, the most likely document that an ordinary English citizen will cast his or her eyes over is The Sun newspaper” (Sinclair 2005: 7). While it is not clear what exactly the impact of such (valued and/or wide-spread) texts is, this has at least two implications for largescale corpus data. On the one hand, it needs to be acknowledged that special texts become “lost” in the corpus where they have only as much weight as any other text: [F]requent patterns of language do not always necessarily imply underlying hegemonic discourses. Or rather, the “power” of individual texts or speakers may not be evenly distributed. A corpus which contains a single (unrepresentative) speech by the leader of a country or religious group, newspaper editor or CEO may carry more weight discursively than hundreds of similar texts which were produced by “ordinary” people. (Baker 2006: 19)

On the other hand, this question clearly needs to be discussed in relation to corpus design. For example, with respect to the quote above, Sinclair asks if we should include more texts from The Sun than from other newspapers in a corpus of British English on account of its importance (Sinclair 2005: 7). In fact, many texts that are included in corpora such as the British National Corpus differ in circulation status (size of readership/circulation level) (Lee 2001: 68). Sinclair notes that issues of

Approaching the data of pragmatics

551

representativeness and balance in corpus design are far from resolved at present (Sinclair 2005: 7–9) and recommends that in the meantime corpus design “should be documented fully with information about the contents and arguments in justification of the decisions taken” (Sinclair 2005: 8). With respect to corpus design (Wynne 2005) other problematic issues include the thorny notion of representativeness (Mahlberg 2004), procedures of sampling (see Mautner 2008, Voormann and Gut 2008), and practicalities of coding and inputting as well as technological exigencies. It may also be difficult to find a corpus that has enough natural occurrences of individual pragmatic phenomena that the researcher is interested in (such as speech acts, see Yuan 2001: 275) and not all pragmatic phenomena can be formally defined – necessitating the use of small-scale corpora at least in a pilot study or as complementary data.

4.

Conclusion: triangulating the data of pragmatics

In view of the above observations, it seems that all data of pragmatics and associated methodologies provide certain insights and perspectives, without necessarily being able to capture the complexity of language in use in its entirety. The use of different data and methodologies contributes variability to the field of pragmatics. With respect to this, Kasper and Dahl (1991: 215) speak of a “double layer of variability”: “variability that reflects the social properties of the speech event, and the strategic, actional, and linguistic choices by which interlocutors attempt to reach their communicative goals” and “the variability induced by different instruments of data collection” (Kasper and Dahl 1991: 215, cited in Woodfield 2007). It thus seems reasonable to argue for a “mixed-methods” approach (Dörnyei 2007) in pragmatics, combining quantitative and qualitative research. Triangulation, or the use of a variety of methods, allows researchers “to cross-check and verify the reliability of a particular research tool and the validity of data collected” (McNeill and Chatman 2005: 23). Researchers arguing for a combination of methods include Holmes (1997), Taylor (2001), Holmes and Schnurr (2005), Baker (2006) and Wodak and Krzy˙zanowski (2008). If there is a team of researchers with various backgrounds, triangulation may be less difficult than if there is one individual researcher only. Large research projects with research teams are thus at an advantage, as they can combine expertise in different research methodologies and distribute the analyses among researchers. An individual researcher may not be able to gain sufficient expertise in various research methods, and there are limitations to the extent to which s/he can undertake different types of analysis for a single research project. One possibility of triangulation for discourse analysis is the “three-pronged” approach suggested by Bednarek (2008a, b, 2009). In essence, such an approach extends previous linguistic research that involves both corpus and discourse analysis, for instance corpus-based

552

Monika Bednarek

discourse analyses (Baker 2006), recent sociolinguistic research (e.g. Holmes and Schnurr 2005), studies in Critical Discourse Analysis (Fairclough 2000; Barker and Galasin´ski 2001: 26; Mautner 2008), and Matthiessen’s (2006) “two-pronged approach” (Matthiessen 2006: 110). The three-pronged approach combines a large-scale quantitative approach with a small-scale corpus analysis to be complemented by qualitative case studies. The large-scale corpus studies use a large (~ more than 500,000 or 1 million words) corpus and computer software to investigate the frequency and distribution of lexico-grammatical items. If the corpus design is reasonable, it can provide a relatively representative sample of language that allows us to formulate hypotheses about general functions of the respective lexico-grammatical items. On account of the large size of the corpus, however, the description is less detailed than would otherwise be possible. In contrast, a smallscale corpus analysis uses a corpus of a size that is amenable to manual or semiautomated analysis, e.g. of semantic or pragmatic meanings, and that still shows us intertextual patterning of such meanings. On account of its smaller size, such a corpus will be less representative, but its analysis will be more detailed and more interpretive. Finally, undertaking one or several case studies allows for qualitative discourse analysis, which can take into account textual development and the sociocultural context (e.g. of production and reception) of the discourses at hand. It will be least representative, in that researchers will be least able to generalize from findings, but its analysis will be most detailed and most interpretive. Thus, applying these three methodologies can provide us with various perspectives on the discourse(s) at hand. However, even with this three-pronged discourse analytical approach, a time span of at least two to three years is necessary for the development of theory, methodology, corpus design, software expertise, and analysis. In terms of its focus on discourse, the analysis is also limited with respect to the various ways in which this discourse can be taken up by its readership or audience. Generally speaking, the development of other mixed-method approaches in discourse analysis and pragmatics is called for, and there is also a need for further evaluation of how to approach the data of pragmatics, as there are only few studies that evaluate data-gathering and analysis in pragmatics (cf. Yuan 2001: 273 on the dearth of such studies within empirical pragmatics). As Rose and Ono (1995) put it: “We are not claiming that one source of data is better than any other; we know too little to make such claims and, frankly, we are sceptical of their value. It would be more productive to ask better for what purpose?” (Rose and Ono 1995: 207).

Notes 1. Research that compares DCT and other data includes Bardovi-Harlig and Hartford (1993), Golato (2003), and Yuan (2001). Rose and Ono (1995) give a useful overview of research on the methodology of DCTs in speech act studies.

Approaching the data of pragmatics

553

2. Sampling procedures are relevant to many approaches to the data of pragmatics: different types include ad hoc sampling, cluster sampling, multi-stage sampling, quota sampling, random sampling, and stratified sampling (Titscher et al 2000). 3. About the history of the “pear film”: In the mid-1970’s Professor Wallace Chafe, a specialist in Native American languages, decided to test how much a simple story will vary from language to language. With his research team, he designed a very simple film to elicit stories from speakers around the world, including uneducated speakers in rural areas. […] “The Pear Stories” film was designed to tap into universal experiences, by showing a man harvesting pears, which are stolen by a boy on a bike. The boy has some other adventures with other children, before the farmer discovers that his pears are missing. The film is six minutes long, in colour, with sound effects but no words. It was filmed in northern California, near the University of California, Berkeley. The man who plays the farmer is a Cuban. The story line is deliberately loose and bland, to avoid imposing a strong U.S. cultural bias. But the film makers deliberately inserted a short scene of a man leading a goat to test descriptions of a background event with no later significance. The scene of falling off the bike and spilling the pears can measure language for cause and effect. And the unusual ping-pong toy tests how people describe an unfamiliar object. The final scene, when the farmer discovers his fruit has been stolen, re-introduces a character who had been off-screen for most of the film, and stimulates speakers to describe emotions and state a moral. (http://www.pearstories.org/docu/ThePearStories.htm, last accessed 14 October 2008) 4. Questionnaires differ with respect to their object of analysis (e.g. factual/behavioral/attitudinal), their use or non-use of scales (e.g. the well-known Likert scale), whether they are open-ended or closed-ended etc (Dörnyei 2007). 5. Computer mediated communication (CMC) can be studied or can itself be made use of as a research tool, with certain implications. On web-based studies compare e.g. Dörnyei (2007: 121–122). 6. Baker (2006: 17–21) discusses some limitations of corpus data and analysis, and Paltridge (2006: 174–176) provides a useful, if brief, summary of and response to criticisms of corpus studies. 7. Additionally, there may be differences in impact depending on who says what to whom how in what context, and mental factors such as “cognitive biases” (Baker 2006: 11) are also involved in discourse production and reception.

References Adolphs, Svenja 2008 Corpus and Context. Investigating Pragmatic Functions in Spoken Discourse. Amsterdam: Benjamins. Anderson, Richard C., James W. Pichert, Ernest T. Goetz, Diane L. Schallert, Kathleen V. Stevens and Stanley R. Trollip 1976 Instantiation of general terms. Journal of Verbal Learning and Verbal Behaviour 15: 667–679.

554

Monika Bednarek

Austin, John L. 1975 How to Do Things with Words. 2nd ed. London: Oxford University Press. Baker, Paul 2006 Using Corpora in Discourse Analysis. London/New York: Continuum. Bamberg, Michael 1991 Narrative activity as perspective taking: the role of emotionals, negations and voice in the construction of the story realm. Journal of Cognitive Psychotherapy 2: 275–290. Bardovi-Harlig, Kathleen and Beverly S. Hartford 1993 Redefining the DCT: Comparing open questionnaires and dialogue completion tasks. Pragmatics and Language Learning 4: 143–165. Barker, Chris and Dariusz Galasin´ksi 2001 Cultural Studies and Discourse Analysis: A Dialogue on Language and Identity. Thousand Oaks: Sage. Baszanger, Isabelle and Nicholas Dodier 2004 Ethnography: relating the part to the whole. In: David Silverman (ed.), Qualitative Research. Theory, Method and Practice, 9–34. London: Sage. Baldry, Anthony and Paul Thibault 2006 Multimodal corpus linguistics. In: Geoff Thompson and Susan Hunston (eds.), System and Corpus. Exploring Connections, 164–183. London: Equinox. Bednarek, Monika 2005 Frames revisited: the coherence-inducing function of frames. Journal of Pragmatics 37: 685–705. Bednarek, Monika 2006a Epistemological positioning and evidentiality in English news discourse – a text-driven approach. Text and Talk 26/6: 635–660. Bednarek, Monika 2006b Evaluation in Media Discourse. London/New York: Continuum. Bednarek, Monika 2008a Emotion Talk across Corpora. Basingstoke/New York: Palgrave Macmillan. Bednarek, Monika 2008b ‘What the hell is wrong with you?’ A corpus perspective on evaluation and emotion in contemporary American pop culture. In: Ahmar Mahboob and Naomi Knight (eds.), Questioning Linguistics. 95–126. Newcastle: Cambridge Scholars Press. Bednarek, Monika 2008c Semantic preference and semantic prosody re-examined. Corpus Linguistics and Linguistic Theory 4/2: 119–139. Bednarek, Monika 2009 Corpora and discourse: a three-pronged approach to analyzing linguistic data. In: Michael Haugh, Kate Burridge, Jean Mulder and Pam Peters (eds.), Selected Proceedings of the 2008 HCSNet Workshop on Designing the Australian National Corpus: Mustering Languages, Sommerville, MA: Cascadilla Proceedings Project. Bednarek, Monika, Helen Caple and Sue Hood 2008 Communicating a sustainable Australia: discourses of persuasion. Unpublished manuscript, Faculty of Arts and Social Sciences, University of Technology, Sydney.

Approaching the data of pragmatics

555

Benwell, Bethan and Elizabeth Stokoe 2006 Discourse and Identity. Edinburgh: Edinburgh University Press. Blum-Kulka, Shoshana 1982 Learning to say what you mean in a second language. Applied Linguistics 3: 29–59. Brown, Penelope and Stephen Levinson 1987 Politeness. Some Universals in Language Usage. Cambridge: Cambridge University Press. Brown, Gillian and George Yule 1983 Discourse Analysis. Cambridge: Cambridge University Press. Bublitz, Wolfram 1998 I entirely dot dot dot: copying semantic features in collocations with up-scaling intensifiers. In: Rainer Schulze (ed.), Making Meaningful Choices in English, 11–32. Tübingen: Narr. Bublitz, Wolfram 2008 ‘Sailing the islands or watching from the dock’: the treacherous simplicity of a metaphor. How we handle ‘new (electronic) hypertext’ versus ‘old (printed) text’. In: Andrea Gerbig and Oliver Mason (eds.), Language, People, Numbers – Corpus Linguistics and Society, 249–273. Amsterdam: Rodopi. Butler, Judith 1999 Gender Trouble: Feminism and the Subversion of Identity. New York: Routledge. Cˇermák, Frantisek 2001 Research Methods in Linguistics. Essential Principles, Based on a General Theory of Science. Prague: Karolinum. Coffin, Caroline and Kieran O’Halloran 2005 Finding the global groove: theorising and analysing dynamic reader positioning using APPRAISAL , corpus, and a concordancer. Critical Discourse Studies 2/2: 143–163. Dörnyei, Zoltan 2007 Research Methods in Applied Linguistics. Oxford: Oxford University Press. Fairclough, Norman 2000 New Labour, New Language? London: Routledge. Galasin´ski, Dariusz 2004 Men and the Language of Emotions. Basingstoke/New York: Palgrave Macmillan. Golato, Andrea 2003 Studying compliment responses: a comparison of DCTs and recordings of naturally occurring talk. Applied Linguistics 24/1: 90–121. Grice, Paul 1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press. Gruber, Helmut 2008 Analyzing communication in the new media. In Ruth Wodak and Michal Krzyz˙ anowski (eds.), Qualitative Discourse Analysis for the Social Sciences, 54–76. Basingstoke/New York: Palgrave Macmillan. Gumperz, John J. and Dell Hymes (eds.) 1972 Directions in Sociolinguistics. The Ethnography of Communication. New York: Holt, Rinehart and Winston.

556

Monika Bednarek

Haastrup, Kirsten 1987 Using thinking aloud and retrospection to uncover learners’ lexical inferencing procedures. In: Claus Faerch and Gabriele Kasper (eds.), Introspection in Second Language Research. 197–212. Clevedon: Multilingual Matters. Halliday, Michael A. K. 2004 The spoken language corpus: A foundation for grammatical theory. In: Karin Aijmer and Bengt Altenberg (eds.), Advances in Corpus Linguistics. Papers from the 23rd International Conference on English Language Research on Computerized Corpora. 11–39. Amsterdam/New York: Rodopi. Heath, Christian 2004 Analysing face-to-face interaction: video, the visual and material. In: David Silverman (ed.), Qualitative Research. Theory, Method and Practice. 266– 282. London: Sage. Hoey, Michael 2005 Lexical Priming. A New Theory of Words and Language. London: Routledge. Holmes, Janet 1997 Women, language and identity. Journal of Sociolinguistics 1/2: 195–223. Holmes, Janet and Stephanie Schnurr 2005 Politeness, humor and gender in the workplace: Negotiating norms and identifying contestation. Journal of Politeness Research 1: 121–149. Hübler, Axel 2007 The Nonverbal Shift in Early Modern English Conversation. Amsterdam: Benjamins. Hymes, Dell 1972 Models of the interaction of language and social life. In John J. Gumperz and Dell Hymes (eds.), Directions in Sociolinguistics. The Ethnography of Communication, 35–71. New York: Holt, Rinehart and Winston. Jewitt, Carey 2009 The Routledge Handbook of Multimodal Analysis. London: Routledge. Johnston, Trevor 2008 Signed languages and corpus linguistics. Paper presented at the 2nd International Free Linguistics Conference, University of Sydney, 11–12 October 2008. Kasper, Gabriele 2000 Data collection in pragmatics research. In: Helen Spencer-Oatey (ed.), Culturally Speaking. Managing Rapport through Talk Across Cultures, 316–341. London/New York: Continuum. Kasper, Gabriele and Merete Dahl 1991 Research methods in interlanguage pragmatics. Studies in Second Language Acquisition 13/2: 215–247. Labov, William 1972 Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Lee, David 2001 Genres, registers, text types, domains, and styles: clarifying the concepts and navigating a path through the BNC jungle. Language Learning & Technology 5: 37–72. Retrieved from http://llt.msu.edu/vol5num3/lee/ on 20 October 2005.

Approaching the data of pragmatics

557

Leech, Geoffrey 1983 Principles of Pragmatics. London: Longman. Levenston, Edward 1975 Aspects of testing the oral proficiency of adult immigrants to Canada. In: Leslie Palmer and Bernard Spolsky (eds.), Papers on Language Testing 1967– 1974, 67–74. Washington, DC: TESOL. Louw, Bill 1993 Irony in the text or insincerity in the writer? In: Mona Baker, Gill Francis and Elena Tognini-Bonelli (eds.), Text and Technology. In Honour of John Sinclair, 157–176. Amsterdam: Benjamins. Mahlberg, Michaela 2004 Corpus design and the words in a dictionary. Lexicographica 20: 114–29. Martin, James R. and Peter R.R. White 2005 The Language of Evaluation: Appraisal in English. Basingstoke/New York: Palgrave Macmillan. Matthiessen, Christian M.I.M. 2006 Frequency profiles of some basic grammatical systems: an interim report. In: Susan Hunston and Geoff Thompson (eds.), System and Corpus: Exploring Connections, 103–142. London: Equinox. Mautner, Gerlinde 2008 Analyzing newspapers, magazines and other print media. In Ruth Wodak and Michal Krzy˙zanowski (eds.), Qualitative Discourse Analysis for the Social Sciences, 30–53. Basingstoke/New York: Palgrave Macmillan. McNeill, Patrick and Steve Chapman 2005 Research Methods. 3rd ed. London: Routledge. Miller, Jody and Glassner, Barry 2004 The “inside” and the “outside”: finding realities in interviews. In: David Silverman (ed.), Qualitative Research. Theory, Method and Practice, 125–139. London: Sage. Noveck, Ira A. and Dan Sperber (eds.) 2004 Experimental Pragmatics. Basingstoke/New York: Palgrave Macmillan. O’Donnell, Mick 2007 The UAM Corpus Tool. Paper presented at ICAME 28, Stratford-upon-Avon, UK, 23–27 May 2007. Paltridge, Brian 2006 Discourse Analysis. London/New York: Continuum. Pollak, Alexander 2008 Analyzing TV documentaries. In: Ruth Wodak and Michal Krzy˙zanowski (eds.), Qualitative Discourse Analysis for the Social Sciences, 77–95. Basingstoke/New York: Palgrave Macmillan. Rose, Kenneth and Reiko Ono 1995 Eliciting speech act data in Japanese: the effect of questionnaire type. Language Learning 45/2: 191–223. Schegloff, Emanuel 2006 A tutorial on membership categorization. Journal of Pragmatics 39: 462–482. Searle, John 1969 Speech Acts. Cambridge: Cambridge University Press.

558

Monika Bednarek

Semino, Elena and Mick Short 2004 Corpus Stylistics: A Corpus-based Study of Speech, Thought and Writing in a Corpus of English Writing. London: Routledge. Skaffari, Janne, Matti Peikola, Ruth Carroll, Risto Hiltunen and Brita Wårvik (eds.) 2005 Opening Windows on Texts and Discourses of the Past. Philadelphia: Benjamins. Silverman, David (ed.) 2004 Qualitative Research. Theory, Method and Practice. 2nd ed. London: Sage. Sinclair, John 2005 Corpus and text – basic principles. In: Martin Wynne (ed.), Developing Linguistic Corpora. A Guide to Good Practice, 1–16. Oxford: Oxbow. Stubbe, Maria, Lane, Chris, Hilder, Jo, Vine, Elaine, Vine, Bernadette, Marra, Meredith, Holmes, Janet and Ann Weatherall 2003 Multiple discourse analyses of a workplace interaction. Discourse Studies 5/3: 351–388. Stubbs, Michael 2008 Three concepts of keywords. Revised version of paper presented to the conference on Keyness in Text, Certosa di Pontignano, University of Siena, June 2007. Retrieved from http://www.uni-trier.de/fileadmin/fb2/ANG/Linguistik/ Stubbs/stubbs-2008-keywords.pdf on 7 August 2008. Tannen, Deborah 1993 What’s in a frame?: surface evidence for underlying expectations. In: Deborah Tannen (ed.), Framing in Discourse, 14–56. Oxford: Oxford University Press. Taylor, Stephanie 2001 Evaluating and applying discourse analytic research. In: Margaret Wetherell, Stephanie Taylor and Simeon Yates (eds.), Discourse as Data: A Guide for Analysis, 311–330. London: Sage in association with The Open University. Thompson, Geoff and Susan Hunston 2000 Evaluation: an introduction. In: Susan Hunston and Geoff Thompson (eds.), Evaluation in Text. Authorial Stance and the Construction of Discourse, 1–27. Oxford: Oxford University Press. Titscher, Stefan, Michael Meyer, Ruth Wodak and Eva Vetter 2000 Methods of Text and Discourse Analysis. Translated by Bryan Jenner. London: Sage. Tognini-Bonelli, Elena 2001 Corpus Linguistics at Work. Amsterdam: Benjamins. Voormann, Holger and Ulrike Gut 2008 Agile corpus creation. Corpus Linguistics and Linguistic Theory 4/2: 235–251. Wardhaugh, Ronald 1986 An Introduction to Sociolinguistics. Oxford: Basil Blackwell. Wetherell, Margaret, Stephanie Taylor and Simeon J. Yates (eds.) 2001 Discourse as Data: A Guide for Analysis. London: Sage in associated with The Open University. Wodak, Ruth 2008 Introduction: Discourse studies – important concepts and terms. In: Ruth Wodak and Michal Krzy˙zanowski (eds.), Qualitative Discourse Analysis for the Social Sciences, 1–29. Basingstoke/New York: Palgrave Macmillan.

Approaching the data of pragmatics

559

Wodak, Ruth and Michal Krzy˙zanowski (eds.) 2008 Qualitative Discourse Analysis for the Social Sciences Basingstoke/New York: Palgrave Macmillan. Woodfield, Helen P. 2007 Problematizing discourse completion tasks: voices from verbal report. Evaluation and Research in Education 20/2. Retrieved from http://hdl.handle.net/1983/943 on 14 October 2008. Wynne, Martin (ed.) 2005 Developing Linguistic Corpora. A Guide to Good Practice. Oxford: Oxbow. Yuan, Yi 2001 An inquiry into empirical pragmatics data-gathering methods: Written DCTs, oral DCTs, field notes, and natural conversations. Journal of Pragmatics 33/2: 271–292.

20.

Experimental pragmatics* Richard Breheny

1.

Introduction

Pragmatics is a mature discipline, focussed largely on the use of language or communication in context. Experimental psychology is a mature discipline, focussed in part on language and communication. It may come as a surprise then to find that Experimental Pragmatics, that is, the development of pragmatics as a theoretical discipline via experimental methods, is more of a fledgling field. While there had been a small but influential tradition in psychological research informed by pragmatic theory (see for instance, Clark 1996), until recently there was comparatively little theoretical pragmatics research founded on the experimental methods of the psychological sciences. Likewise, while there is a strong tradition of psycholinguistic research on the interpretation of discourse in context (see Graesser et al 2003 for a recent handbook on the topic), until recently little of this research had been much interested in the subject matter of pragmatics. Why this state of affairs persisted for so long and how the situation is changing are questions that will be addressed in this entry. In addition, the current state of experimental pragmatics will be reviewed and future prospects surveyed. Recent experimental pragmatics has focussed on Gricean pragmatics, which considers phenomena such as implicature, presupposition, figurative language and, more generally, speaker meaning. Thus the focus of this entry will reflect this bias.

2.

The foundations of pragmatics and experimental methods

Modern experimental psychology aims to develop scientific theories of individual mental, cognitive or psychological capacities; normally of how certain functions, such as vision, planning or memory, are realised. Contemporary pragmatics, particularly as it interfaces with formal linguistics, stems largely from the tradition of analytic philosophy through the work of Austin, Grice, Carnap, Lewis among others. It is not straightforward to apply the methods of experimental psychology to this field of inquiry since thinkers like Grice and Austin were not in the business of proposing accounts of individual pragmatic competencies. For analytic philosophy, questions about how we cognise language in context are separate from critical philosophical questions about meaning in context. If we see pragmatics as a continuation of the analytic tradition, there would seem to be little scope for the application of experimental psychological methods in pragmatics. Things are not that

562

Richard Breheny

different if we consider more socially oriented pragmatics, which takes inspiration from social sciences such as anthropology, sociology and economics. While these disciplines are more empirically grounded, their methods rarely overlap with those of the experimental psychologist. Given the philosophical and social scientific foundations of pragmatics, perhaps we should not be surprised to find that Experimental Pragmatics has been a minority pursuit, at least until relatively recently. However, pragmatics has also flourished not so much as a branch of social sciences like anthropology or of analytic philosophy but as a branch of contemporary linguistics. Given the nature of the more dominant formal end of this discipline, it is more surprising to find the lack of experimental pragmatics. After all, most of contemporary formal linguistics takes itself to be part of the cognitive scientific enterprise, where theory is ultimately aimed at explaining the cognitive psychological underpinnings of linguistic and pragmatic phenomena.

3.

Introspection and its limits in formal pragmatics

Since its inception, the main method for collecting empirical data in formal linguistics has been to introspect on intuitions (and check with a small number of informants if judgments are not sharp). While this methodology has served the discipline well in syntax and semantics, it might be questioned in the case of pragmatics. For pragmatics interfaces with formal linguistics mostly in the domain of generalised pragmatic inference. Probably the best known example of generalised phenomena are the scalar implicatures formalised (to a greater or lesser degree) in the neo-Gricean works of Gazdar (1979), Horn (1989), and Levinson (Atlas and Levinson 1981, Levinson 2000).1 In these works, Grice’s informal account of implicatures is recast so that ‘generalised’ implicatures become default. I.e. the implicatures can be derived solely on the basis of the structure of the sentence used and only become unavailable (‘cancelled’) under specific contextual conditions. While neo-Gricean theory provides a computationally more tractable description of scalar implicature than Grice’s original informal accounts, and hence better meets the desiderata of formal linguistics, the empirical basis for this description is the intuition that these implicatures are default. To take a concrete example, the assumption is that an implicature as in (1b) is available from the use of a sentence like (1a) unless context overrides this (for example by the speaker saying, ‘In fact all of them did’): 1. a. Some of the consultants had a meeting with the director. b. Not all of the consultants had a meeting with the director. Note however that any introspective evidence of the defaultness of (1b) depends on implicit assumptions to the effect that implicatures are available independently of

Experimental pragmatics

563

an utterance situation (real or imagined). To the contrary, one could assume that no implicature becomes available to introspection without a real or imagined utterance situation. In that case, one could observe that it would be dubious to try to ascertain introspectively whether an implicature is available when one’s interest is in whether the implicature in question is available. Although neo-Gricean theory has been around for thirty years or more, no argument has been put forward to defend the implicit assumption behind introspective evidence for the defaultness of certain implicatures. In a recent important set of studies, Geurts and Pouscoulous (2009) have established that introspection on the availability of an implicature introduces a substantial bias in favour of judgements of availability. In their design, Geurts and Pouscoulous had participants perform two tasks. Their inference task mirrors the task undertaken by the theorist when she introspects: participants were asked about whether a putative scalar implicature, like that in (1b), followed from a simple sentence containing the implicature trigger, like (1a). Using separate items in a verification task, participants had to verify the truth of a sentence like (1a) when in fact all consultants had a meeting with the director (contrary to the implicature in (1b)). In a number of different kinds of construction, Geurts and Pouscoulous found that participants were significantly more likely to base their response on an interpretation that included the implicature in the inference task than the verification task. Their conclusion is that, ‘no matter how consistent and robust an introspected inference may be, that doesn’t say anything about the frequency with which it is drawn in practice’ (Geurts and Pouscoulous 2009 p.4.18). These results reinforce the view that it is very important to establish the defaultness of a given pragmatic effect (like scalar implicature) by methods other than introspection. As we will see in the next section, much recent work in experimental pragmatics has focussed on scalar implicatures and is informative on this issue.

4.

Testing for defaultness and preferred readings

One clear statement of a cognitively oriented neo-Gricean theory can be found in Levinson (2000). Levinson argues that there exist utterance-type meanings that include default implicatures, and that these implicatures are actually generated upon encountering the appropriate linguistic stimuli. This thesis has been directly and indirectly tested in experimental work. One set of studies builds on a sentence verification paradigm established in Noveck (2001). In those studies, participants are asked to verify statements like ‘Some elephants have trunks’. Noveck (2001) found that a substantial proportion of participants responded negatively, suggesting that those participants had based their judgement on the meaning of the sentence plus the implicature, ‘not all elephants have trunks’; while a smaller but still substantial number responded yes, apparently basing their response on the plain, ‘logical’ meaning of the sentence. In subsequent studies, Noveck and Poseda

564

Richard Breheny

(2003) and Bott and Noveck (2004), measured response times to these questions and found that it was significantly greater when participants’ responses were based on the enriched meaning – the one that included the implicature. In a follow up study, Bott and Noveck (2004) also found that a group that was put under time pressure to respond as quickly as possible gave significantly more ‘logical’ responses than the group that was not. These results are relevant to the question of whether generalised implicatures are generated by default. According to Levinson (2000), where default implicatures are unavailable this is the result of cancellation – a process akin to revision when a sentence processor has been ‘garden pathed’. As such, responses based on an interpretation of a stimulus without an implicature ought, other things equal, to manifest a greater cost than those based on an interpretation with an implicature. As just discussed, in the speeded verification tasks, Noveck and colleagues found the opposite pattern. In a more direct measure of on-line access of implicatures, Breheny et al (2006) tested Levinson’s theory employing a reading time methodology. In one experiment, participants read short discourses which provided either a context which heavily biased a scalar implicature (as in (2)) or one in which the implicature is generally felt to be unavailable due to lack of relevance (see (3)). The latter kind of contexts are referred to as ‘implicitly cancelling’ contexts: 2. Mary asked John where he would accommodate guests for the weekend of the wedding. John replied that /some of his relatives/ would stay with him. /The rest/ would book into hotels. 3. Mary asked John why he was cleaning his apartment. John replied that /some of his relatives/ would stay with him. /The rest/.would book into hotels. Reading times on the segment ‘the rest’ confirmed that the ‘and not all’ implicature was highly salient in the biasing contexts but not so in the context where the implicature was irrelevant. In line with results in Bott and Noveck (2004), Breheny et al (2006) found that reading times for the segments around the quantifier ‘some’ were significantly longer in the biasing context, like (2). These results again run contrary to the predictions of Levinson’s theory which holds that in the implicit cancellation context, extra work is required to suppress an activated implicature. By contrast, all of these results are in line with a more ‘Gricean’ view whereby implicatures are only available and drawn where warranted by specific features of the context (as per Sperber and Wilson 1986). Breheny et al (2006) present a further study to directly compare the predictions of Levinson’s strong interpretation of neo-Gricean theory and a more fundamentally Gricean view. In this study, there was a simple manipulation of the position in the sentence in which an implicature trigger occurs. For example, participants read one of the discourses in (4a,b):

Experimental pragmatics

565

4. a. Some of the consultants had a meeting with the director. The rest did not manage to attend. b. The director had a meeting with some of the consultants. The rest did not manage to attend. Citation examples presented out of the blue in theoretical discourse or indeed in experimental settings are sometimes referred to as ‘neutral’ contexts (cf. Levinson 2000, Chierchia 2004). There is no agreed definition for ‘neutral’ in this kind of discourse but the suggestion seems to be that the examples are interpreted without any presumptions about the utterance situations in which the items are produced. A more Gricean view would argue that in as far as the examples are given an interpretation by the reader or experimental participant, there ought to be some imagined utterance situation in which these are produced. That is, they are treated as utterances (albeit in written form), produced by a speaker with a purpose. Breheny et al reasoned that if experimental participants treated these items as occurring in imaginary utterance situations, then other things equal, they would accommodate context incrementally. Thus it is more likely with (4a) than (4b) that the utterance is addressing some concern about the consultants, and so in (4a) and not (4b) it is more likely that participants would derive the implicature. So, by the Gricean account, we would expect to see this effect manifested in response times to the anaphor ‘the rest’, which is easier to interpret in contexts where the ‘not all’ inference has been drawn. Stronger versions of neo-Gricean theory, including Levinson’s, predict that in these neutral contexts the implicatures are equally likely to be generated. Thus there should be no difference. In line with the Gricean account, Breheny et al found a difference, suggesting that sentence position has an effect on implicature. In a further test of the strong neo-Gricean view, Grodner and Sedivy (forthcoming) adapted a visual-world design (see Cooper 1974, Tanenhaus et al 1995) from Sedivy et al (1999) – to be discussed below – which involves participants hearing instructions such as in (5): 5. Pick up the tall glass. Modified referential phrases like ‘the tall glass’ often presuppose that there is more than one object that satisfies the nominal description ‘glass’. In a series of papers, Sedivy and colleagues have established that this extra presupposition is effectively a Gricean implicature. Grodner and Sedivy (forthcoming) tested the claim in Levinson (2000) that this inference only requires the appropriate linguistic context to be generated. One group of participants in their study were given information that the speaker producing instructions like (5) was unreliable in certain relevant ways. Grodner and Sedivy reasoned that if the inference is triggered solely by the linguistic form, even though there was reason to think that the speaker may not have intended the implicature, in early eye-movements to the visual array, one should

566

Richard Breheny

see evidence that the implicature was accessed before it could be cancelled. By contrast, if the information about the speaker is allowed to modulate interpretation, then in contrast to the control group, one should see no effect of the implicature in early eye-movements. In line with the off-line and reading time studies reviewed above, Grodner and Sedivy again found that the contextual information about the speaker does modulate the on-line access of implicatures. The findings of these studies point toward the conclusion that even apparently common implicatures are not generated solely on the basis of linguistic context in on-line comprehension. While these findings tend to disconfirm Levinson’s particularly strong version of neo-Gricean theory, other versions of neo-Gricean pragmatics could be adapted as the basis of a cognitive theory of pragmatics, for instance, those found in Horn (1984, 1989), Gazdar (1979) or Sauerland (2004). These accounts presume the default availability of generalised implicatures, rather than actual default generation. The difference is important for predictions in empirical studies that seek to measure participants responses based on the inclusion of an implicature in the interpretation of an utterance. For, just as certain logically derivable entailments of a discourse may not necessarily be generated or accessed in a given circumstance, so it may be that default implicatures may not necessarily be derived or generated in comprehension. The parallels with deductive inferences are important for another reason since it is quite reasonable to view pragmatic description as operating at a normative, rather than cognitive level. The relation between the normative level and the level of individual psychological competence may be very indirect (see Saul 2002). The more indirect relation between neoGricean pragmatics and cognition is suggested in Horn (2006) and simply manifests a more specific version of the position held by Grice (see Warner 2001, Petrus 2010) and others in the analytic tradition.2 Adopting this stance results in a disengagement of pragmatics from empirical, experimental research of the kind undertaken in contemporary psychology. While it remains a very interesting question whether it is viable to take either the normative stance or the more neutral a priori stance towards pragmatics, it is a question which is rarely addressed. For the most part, as mentioned above, those interested in Gricean pragmatics are not philosophers but linguists who work on the assumption that linguistics is in the business of explaining individual behaviour via theories of the cognitive processes that underlie that behaviour. From this perspective, pragmatic theory should be seen as a theory of individual competence, and pragmatic description ought ultimately to be specified in such a way that it becomes open to empirical (dis-) confirmation. Thus an account of pragmatic competence based on neo-Gricean theory according to which implicatures merely become available by default cannot be tested unless we make some additional assumptions about when available pragmatic inferences are accessed. For instance, neo-Gricean theory could be implemented in a constraint-based framework (MacDonald and Seidenberg 2006) so that generalised implicatures would be accessed in the presence of certain cues. However, it

Experimental pragmatics

567

then falls to the proponent of default implicatures to specify in some more detail under what condition implicatures are accessed. It has become clear through the studies in Geurts and Pouscoulous (2009) and the various studies testing Levinson’s strong version of neo-Gricean theory that there is very little evidence that even the most proto-typical scalar implicatures are default. Only when reflecting on the availability of these implicatures does one get a judgement that the implicature is reliably available. In other kinds of task, there is much evidence in support of an alternative, more fundamentally Gricean, view that implicatures are derived if and only if warranted by evidence in the utterance situation. So while it is still perhaps an open empirical question whether neo-Gricean pragmatics offers insights for experimental pragmatics, it is a question which will require a more thorough engagement with experimental research into utterance interpretation to answer. With evidence supporting the idea that implicatures are only derived where warranted by information about the speaker’s likely intentions in context, we can ask if there are any theories of pragmatic competence that take a more fundamentally Gricean stance on utterance interpretation. One theory which is notably ‘Gricean’ in this respect is proposed in Sperber and Wilson (1986), which argues that inferring the speaker’s informative intention is greatly facilitated by the existence of an underlying Relevance mechanism. By contrast, Geurts (2010) is an advocate of a psychologically oriented Gricean position, which adopts Grice’s original ideas about pragmatic inference based on the Co-operative Principle and maxims as the starting point of a theory of pragmatic competence. Similarly Clark (1996) and Tomasello (2008) adopt elements of Grice’s work in their joint-action approach to language use. The viability of more Gricean theories of pragmatic competence has been recently questioned in Sedivy (2003) where it is argued that the rapidity of access to implicatures on line is difficult to square with the apparent complexity of the reasoning that lies behind these implicatures. Breheny et al (under review a) directly test whether very rapid access of implicatures in on-line comprehension could be the result of full-blown pragmatic reasoning on line. The studies involved quantity implicatures that have no linguistic trigger, as where (6a) below implicates (6b) or (6c): 6. a. The student put a book on the dusty shelf b. The student put nothing else on the dusty shelf c. The student put a book nowhere else According to both the standard Gricean account and the Relevance Theory account (see Sperber and Wilson 1995), implicatures such as those in (6b) and (6c) are derived on the basis of assumptions about the purpose (or source of Relevance) of the utterance and on considering whether the speaker could know alternative, more informative propositions. Breheny et al (under review a) report a series of visual-

568

Richard Breheny

world studies based on the ‘look and listen’ paradigm first explored in Altmann and Kamide (1999). In this paradigm, participants are presented with a visual display and listen to spoken discourse. It has been consistently found that eye-movements around the visual scene reflect participants’ understanding of the text. In particular, participants’ looks around the display show that they can anticipate reference to an object on the basis of their understanding of the previous parts of the utterance but before reference is explicitly made. For instance, in Altmann and Kamide (2007) participants hear about an agent who ‘has drunk’ or ‘will drink’ certain things. Where the visual display contains a full glass (of beer) and an empty wine glass among other items, participants look to the empty glass when the verb is in perfect tense and at the full glass when the verb is in the future tense. In both cases, anticipatory looks precede mention of the critical noun, ‘wine’ or ‘beer’. Breheny et al (under review a) adapt this paradigm to explore whether a quantity implicature like (6b) is accessed by participants on-line, given the right biasing context. In a visual display containing a representation of the student putting a book and another item to one location (a table) and a book and nothing else to the shelf, participants do anticipate reference to the shelf before they hear the noun ‘shelf’, given the appropriate context that biases the ‘and nothing else’ quantity implicature. It is also shown that, at an earlier stage in processing, participants are sensitive to what the purpose of the utterance is and that they are aware of more informative alternatives. Thus there is evidence that we can and do access quantity implicatures in incremental interpretation of utterances and that we do so on the basis of sensitivity to the likely purpose of the utterance and reasoning about the speaker’s mental state in the process.

5.

Pragmatic theory meets psycholinguistic theories of the architecture of on-line comprehension

Independently of any question about the correct theoretical account of the availability and/or derivation of implicatures, we can ask questions about the cognitive mechanisms that underpin semantic and pragmatic processing, how these mechanisms interact and how the whole architecture of comprehension develops through infancy and childhood. We will consider developmental research in the next section. In this section, we look at recent research relevant to the architecture of adult comprehension. As we will see, in this domain, higher level pragmatic theory and psycholinguistic theory are able to inform one another to their mutual benefit. As mentioned above, it has been established that we are capable of accessing quantity implicatures in on-line, incremental interpretation of utterances. While Breheny et al (under review a) have shown this for quantity implicatures that have no linguistic trigger, as in (6a-c), Breheny et al (2006) and Breheny et al (under review b) have demonstrated this for the more familiar case where saying ‘some’

Experimental pragmatics

569

implicates ‘not all’. As we will see in this section, there is already a tradition of research in psycholinguistics that has looked at another kind of pragmatic implicature, having to do with establishing the reference of a definite expression. While this research has been relevant to the question of whether any pragmatic implicatures are default, it is equally relevant to architectural issues. Sedivy et al (1999) have established that, in a visual context where there is more than one glass, participants already infer that the referent of ‘the tall glass’ in (5) is the taller of the glasses before they hear ‘glass’, even in the presence of another tall object. Sedivy (2003) reports a study where the scalar adjective ‘tall’ is replaced by a material adjective (‘plastic’, ‘wooden’ etc) and the same effect is found. Grodner and Sedivy (forthcoming) have established that the effect disappears in an ‘unreliable speaker’ condition. All of this evidence points to the conclusion that the inference about the referent based on the modifier alone is a Gricean one: comprehenders assume that speakers will avoid redundancy in their utterances. When they encounter an adjectival modifier in a definite expression in a visual context containing two potential referents of the same type, participants can infer that the likely reason for the presence of the modifier is to distinguish between the two referents. As participants in these eye-tracking studies seem to be able to access this inference very rapidly (between the onset of ‘tall’ and the onset of ‘glass’), we can ask how this is possible. As we have seen above, Sedivy and colleagues have ruled out the theory that rapid access to generalised implicatures only requires formbased cues. As also noted above, one well-motivated alternative hypothesis is that pragmatic inferences are accessed in on-line interpretation via a constraint-based processing mechanism. Such a mechanism is interactive, frequency-based and automatic, meaning that at any point in language processing, the currently favoured hypothesis is based on its likelihood given information from all levels of processing (including contextual information). While a constraint-based approach that simply implements neo-Gricean theory would work for common ‘generalised’ implicatures, it would have to be extended considerably to work for the ad hoc implicatures studied in Breheny et al (under review a) and in the studies reported in Huang and Snedeker (2009) and Grodner et al (forthcoming) to be discussed below. Still, Breheny et al (under review a) argue that a constraint-based architecture could serve to implement more classical Gricean theory, where inferences about the likely purpose of an utterance, and other aspects of the speaker’s epistemic state are made automatically and probabilistically. An alternative to a constraint-based approach to pragmatic processing is one according to which pragmatic implicatures are accessed only once an interpretation of the linguistic expressions is made, based solely on the compositional semantic meaning of those expressions. This alternative is proposed in Huang and Snedeker (2009), who report on similar visual-world studies to those in Sedivy et al (1999), but which have different results. In Huang and Snedeker (2009), partici-

570

Richard Breheny

pants are given instructions like in (7) where the visual display has a girl with socks and a girl with soccer balls: 7. a. Point to the girl that has some of the soccer balls. b. Point to the girl that has all of the socks. While participants can disambiguate the referring expression on hearing ‘all’ in (7b), they cannot disambiguate the referent for (7a) until after the recognition point of ‘soccer’ unless they are able to make a pragmatic inference. Huang and Snedeker found that participants were able to anticipate the correct referent in the ‘some’ condition before the recognition point but that, compared to the ‘all’ condition, they did so at a delay. Huang and Snedeker argue that this delay reflects both that implicatures are accessed only after the compositional meaning of the sentences is activated and that this is a reflection of the architecture of comprehension whereby implicatures belong to a different, pragmatic level of representation than the conventional meaning of a sentence. The findings in Huang and Snedeker (2009) are difficult to square with those of Sedivy and colleagues which suggest the pragmatic inference is accessed very rapidly. So, the question of the time course of accessing implicatures is still open. This is particularly the case as Grodner et al (2010) found very rapid access to the pragmatic implicature in a study almost identical to that of Huang and Snedeker. Breheny et al (under review b) argue that a good Gricean account is available to explain the difference in timing among these studies and present ‘look and listen’ visual-world studies showing that participants can access scalar implicatures of the ‘some’/not all variety as rapidly as information based on the meaning of ‘all’. Although it is certainly too early to tell for sure, there is mounting evidence that the results of Gricean pragmatic reasoning can be accessed as rapidly as the results of deriving information about the interpretation of an utterance from the compositional meaning of the sentence uttered. This evidence suggests that processes that underpin access to pragmatic implicatures can in some cases operate automatically or spontaneously. As mentioned above, this kind of evidence is in line with constraint-based approaches to language processing, although other heuristics-based accounts cannot be ruled out. The evidence does not decide definitively against the idea that there are two levels of representation, semantic and pragmatic, but only against the conjunction of that idea and a modular processing architecture according to which all and only Gricean pragmatic processes are set apart from processes that establish a semantic representation. Having found evidence that, in some cases, genuine Gricean pragmatic inferences are accessed via what appear to be automatic processing mechanisms, we can ask to what extent are such inferences operationalised in this way. Gricean pragmatics explains utterance interpretation as a matter of inferring the speaker’s intentions in the utterance situation and much recent research has explored the question of whether inferences about the speaker’s intentional state in context can

Experimental pragmatics

571

rapidly affect interpretation. Earlier work in this field by Keysar and colleagues tended to support a ‘strategic egocentrism’ model according to which first-pass processes effectively ignored information about the speaker’s mental state (see Keysar and Barr 2005). In a typical study, Keysar et al (2000) have participants play a game where a speaker gives orders to participants to manipulate objects in a grid and where some objects are obscured from the speaker. In critical items, an instruction like, ‘Pick up the tape’ is referentially ambiguous if one cannot ignore items that the speaker is ignorant of, but unambiguous if one only takes the speaker’s perspective into account. Keysar et al (2000) report severe interference from the object the speaker cannot see, as predicted if hearers are initially egocentric in their language processing. Keysar and Barr (2005) argue that the strategic egocentrism strategy is motivated by the apparent difficulty we have in reasoning about the mental states of others, especially manifest when beliefs diverge from one another (see Apperly and Butterfill 2009). More recent research into speaker perspective has explored other factors which may have led to a delay in incorporating speaker’s perspective. For instance, Hanna et al (2003) argues that the use of a recording in Keysar et al (2000) is detrimental to the process of monitoring perspective. In Hanna et al (2003), the instructions were given by a confederate who engages the participant in face to face interaction. The results show much more sensitivity to speaker perspective, virtually without delay. However there is still some small interference. Similar results are reported in Heller et al (2008) which eliminates a global ambiguity from the Hanna et al studies that may have encouraged participants to overcome any ‘strategic egocentrism’ bias. Still better evidence that perspective can be integrated immediately comes from Brown-Schmidt et al (2008) who use a different type of discourse to previous studies. In Keysar et al (2000) and most other perspective research, the speaker gives orders to a participant. In Brown-Schmidt et al’s study, the speaker and hearer are engaged in a joint action and the critical sentences are questions. That contextual factors such as type of discourse can affect results in this way supports a middle position between one according to which information about speakers’ mental states is completely ignored and one according to which such information is always automatically incorporated. One version of the middle position is the Constraint-Based view (Hanna et al 2003, Heller et al 2008, BrownSchmidt et al 2008, among others), according to which cues to mismatched perspective are treated with greater or lesser degrees of certainty and so perspective information is incorporated only where these cues overcome the very strong constraints provided by the form of the speaker’s referring expression. One interpretation of this Constraint-Based view is that processing mechanisms are always attempting to integrate the relevant information about the speaker’s mental state because interpretation systems are geared toward inferring the speaker’s intended meaning. This would be a manifestation of the one-level-of-representation view. A second interpretation of the C-B view would be to see information from the level of

572

Richard Breheny

speaker’s meaning interacting to affect a separate semantic level of representation. Until now, no research has been done to tease these positions apart. A more modular alternative is offered in Barr (2008) where it is proposed that while speaker perspective may always be monitored and may immediately affect anticipation of referents, integration of perspective information with information from the linguistic input is not automatic. Using a similar design to other studies mentioned above but a new method of analysis which discriminates anticipation from integration effects, Barr finds that bottom-up, linguistic information affects early referential hypotheses independently of whether participants manifest awareness of contrary evidence of speaker perspective. As these studies used a recording of a speaker and lacked genuine speaker/hearer interactions, and as they involved a director giving orders, it remains to be seen whether the anticipation-without-integration effect is maintained in more genuinely interactive scenarios across different types of discourse. Supplementary to research on how a speaker’s ignorance may affect reference assignment, recent on-line research has studied the time course of so-called entrainment effects, long documented in off-line interactive research (see Garrod and Anderson 1987, Brennan and Clark 1996). Entrainment is the phenomenon where interlocutors fix on a single description for an object or event throughout a conversation. So the same object may be referred to as a metal pipe or silver tube, but once such a mode of reference is established, subsequent reference among interlocutors tends to maintain this ‘perspective’ on the object. The analysis of entrainment stemming from Brennan and Clark (1996) is that it is the result of a ‘conceptual pact’ among interlocutors who (tacitly) agree on a mode of presentation of a referent for the purposes of a conversation. Off-line research has shown that such entrainment effects can last for days (Markman and Makin 1998) and that speakers do adjust their descriptions according to the knowledge of their audience (Brennan and Clark 1996). The off-line research suggests strongly that entrainment is a partner-specific effect but it is still open whether it results from an intentional act by interlocutors, as Brennan and Clark maintain, or whether some other factors are at play. In an early on-line study, Metzing and Brennan (2003) had participants play a game with an experimenter moving objects around a grid. The experimenter (speaker) would fix on a description for a certain object (‘silver tube’) for three moves and then the experimental conditions manipulate both whether the pact was maintained and who the speaker was. In measures of visual responses around the grid, it was found that when the description was changed, significantly more interference was created with the same speaker than with a new one. However, when descriptions were maintained, there was no effect of speaker. That is, participants fixed on the correct referent as rapidly with a new speaker as with the old. While the results of the different description condition support the conceptual pact account, it would be expected that there should be an effect of speaker both in the case where the pact is

Experimental pragmatics

573

broken and when it is maintained. Kronmuller and Barr (2007) offer an account of Metzing and Brennan’s results according to which the partner-specific effect is the combined result of an automatic pre-emption of other descriptions due to a link made in memory between the referent and the description, and a somewhat delayed application of inference based on pragmatic principles. This account explains the lack of difference in same-description condition of Metzing and Brennan’s study and predicts that the interference effect for the same-speaker in the different-description condition occurs at a delay. This is the result Kronmuller and Barr found in a visual world study similar to that of Metzing and Brennan. Their conclusion is that there are two mechanisms at play in the maintenance of descriptions in conversation. However, it is somewhat mysterious why an automatic pre-emption mechanism would develop, except as a result of something like Gricean pragmatic principles. So even if Kronmuller and Barr are right, it may still be the case that entrainment is the result of a heuristic that serves joint interaction in discourse. On the other hand, Brown-Schmidt (2009) argues that the lack of interactivity in Kronmuller and Barr’s studies could explain the weakness of their entrainment effects. Brown-Schmidt employed a slightly different design to the Metzing and Brennan (2003) study, employing critical items that would encourage entrainment effects and using more participants. She found both an effect of speaker in the same- and different-speaker condition. In a second study, she employed the same items but with a recording rather than a live confederate and found that the effects of speaker were diminished greatly. The conclusion seems to be that entrainment effects that are the product of genuine real-time interactions among speakers can occur spontaneously or automatically. Whether, in addition, there exist pre-emption effects, and if so why they arise, are still open questions. To sum up this section, we have considered a sample of psycholinguistic research which sheds light on the mechanisms that underpin access to pragmatic inferences. We have seen that there is no strong support for the idea that genuinely Gricean pragmatic inference processes are somehow set apart from processes that establish a separate semantic representation. The idea that Gricean inferences could not automatically or spontaneously be derived because of the complexity of the mental state reasoning involved also needs to be reconsidered, for two reasons. Firstly, across a number of different domains, it has been demonstrated that genuine intentional pragmatic inferences are accessed as rapidly as any other information relevant to an interpretation. Secondly, the idea that inferences about mental states are per se especially difficult compared to any other kind of non-demonstrative inference ought to be re-examined. Indeed, as we will see in the next section, there is good reason t to think that in the domain of language use and communication, humans develop a keen ability to infer the contents of goal states from the earliest possible stage. Given the rapidity of Gricean inference in some cases and the relative open-endedness of their confirmation, the challenge for both psycholinguistics and prag-

574

Richard Breheny

matic theory is to explore the mechanisms that deliver them. We have ruled out the idea that we only get rapid pragmatic inferences when these can be associated with particular forms (like ‘some’ or adjectival modification). In fact we have seen evidence that any ad hoc quantity implicature can be accessed very rapidly. Thus, if we are to adopt a Constraint-Based model or any other frequency-based model for pragmatic inference, we need to specify which elements of an utterance situation are being automatically monitored by hearers. Breheny et al (under review a) suggest that one important place to start is with expectations of the conversational purpose or, more generally, the source of relevance for an utterance. In the case of quantity implicatures, this could be coupled with a mechanism that monitors alternatives. One way to explore the latter mechanism is via the more general mechanisms that anticipate the completion of an utterance. These are evident in the visualworld research mentioned above (see Altmann and Kamide 1999, Barr 2008) but much more so in ERP (Event Related Potential) research focussed on the N400 effect.3 Generally speaking, it seems clear that future psycholinguistics research into the mechanisms that underpin utterance interpretation would benefit from insights of theoretical pragmatics. At the same time, pragmatic theory (at least the cognitively oriented variety) can greatly benefit from research into on-line processing. In particular, one important question for recent philosophical and linguistic research is whether there is justification for positing a separate semantic level of representation of the truth conditions of an utterance in context which is independent of the level at which the speaker’s intentions are represented. If we treat this theoretical question as also a cognitive question (as do Borg 2004, Carston 2002), then the type of research reviewed above could be informative to the question. However, while much evidence supports a one-level view, the two-level view cannot be ruled out given the current state of research.

6.

The development of Gricean pragmatics in language use

Any good theory of pragmatic competence ought not just inform the modelling of on-line processing in adults but should also be able to explain how the relevant mechanisms involved develop. This section will review research into the application of intentional pragmatic principles in early language use. We will see that, while our understanding of child pragmatics has been transformed in the recent past, the focus has been on getting certain facts right about what, if any, limitations children have with regards to common pragmatic inferences. The emerging picture is of children being capable of deriving any pragmatic inferences that adults derive as early as anyone can ascertain. Where there has been apparent divergence between child responses in experimental studies and adults, this has been partly due to inadequate methodology. The interesting theoretical question regarding devel-

Experimental pragmatics

575

opmental research is how children so rapidly achieve full pragmatic competence and what lies behind any genuine differences between children and adults regarding the implementation of pragmatic abilities in conversation. Perhaps in contrast to research into adult comprehension, child language research has, for a long time, focussed its attention on children’s intentional pragmatic abilities. The watchword for most work on children’s syntactic and semantic abilities has been pragmatic plausibility. For instance, when asking a child whether she finds a statement about an acted-out situation true or false, a good test for whether a child is capable of ruling out ungrammatical construals of a string involves scenarios where grammatical construals are false but an ungrammatical construal is true. For example, in Crain and McKee (1985) experimental scenarios made ‘He ate pizza and Ninja Turtle danced’ true only on the ungrammatical construal according to which the turtle ate pizza and danced, while the grammatical construal is false. The methodological principle of pragmatic plausibility in child studies is to make the grammatical construal relevant. In this case, by having another character consider eating pizza. The theory is that children in these studies operate with a ‘principle of charity’ whereby they are liable to ignore divergences from proper usage if not doing so does not lead to a plausible construal. Whether children should operate with such a principle more than adults is something of an open question. Indeed, as we will see below, the real question now for child pragmatic research is why children’s responses often diverge from adults’. Any explanation based on a ‘principle of charity’ would have to be motivated more fully than hitherto, while there may be alternative explanations for such divergences. While early work in syntax and semantics using Truth Value Judgement Tasks (TVJT) tended to assume some pragmatic competence on the part of toddlers, early research directly testing children’s access to common pragmatic inferences suggested that pragmatic competence might not be fully developed in some respects. Here we will focus on children’s understanding of scalar quantity implicatures and the pragmatic dimension of Principle B (cf. below) of binding theory. In a landmark paper, Noveck (2001) reports studies using a sentence verification task involving quantificational expressions like ‘All elephants have trunks’ or ‘Some elephants have trunks’. As mentioned in section 4, adult participants in these studies tend to favour a response to the ‘some’ items that is based on an interpretation of the utterance that includes the quantity implicature ‘Not all elephants have trunks’. Although numbers vary from study to study, the proportion of responses that incorporate the implicature rarely falls below two thirds. Noveck (2001) reports that children’s response rates do not approach that of adults until 7 years in some cases and later with more complex items. Rather, children tend to favour the ‘logical’ response, saying that ‘Some elephants have trunks’ is true. In a follow-up study, Papafragou and Musolino (2003) used act-out scenarios and a truth-value judgement task to see if the task demands of sentence verification (requiring access to long-tem memory) might have been a factor. Indeed, Papafra-

576

Richard Breheny

gou and Musolino (2003) found younger children were performing like adults in rejecting statements like ‘Some of the horses jumped the fence’ but still five year old children showed little sign of having incorporated the implicature into their understanding of such statements. In an attempt to determine whether it is the inherent difficulty of the Gricean inference that is at issue, Papafragou and Musolino ran a felicity judgement task which required the children only to evaluate whether the puppet had described the scenario well. They found significantly more children were able to evaluate an underinformative utterance appropriately. The conclusion that could be drawn from this is that, while children do have expectations of informativity appropriate to drawing quantity implicatures, something about the TVJT causes the children to diverge from the adult response pattern. This conclusion is bolstered by the results of another kind of felicity judgement task (Guasti et al 2005), where children are given a forced choice between two images, one that depicts a scenario where all of the characters perform the action and one where some but not all do. In this kind of study, five year-old children robustly choose the scenario consistent with expectations of informativity. Katsos and Bishop (under review) find that five year old children succeed at the latter felicity judgement task where the implicature is not associated with any lexical item, but ad hoc. Kastsos and Bishop also find that children’s own descriptions of scenarios always respect Gricean expectations of informativeness. Pouscoulous et al (2007) explore the idea that children do not readily incorporate the implicature into their understanding of the experimental items due to the task demands resulting from the combination of the metalinguistic nature of the TVJT and the extra effort required to derive quantity implicatures. In a series of experiments that were aimed at successively lightening task demands, Pouscoulous et al did find a marked difference in performance. In particular, by changing the task from truth evaluation to an act-out task, they found that even four-year-old children’s rate of with-implicature response was above chance. They also found that children’s with-implicature response rates increased when the items involved the apparently less complex quantifier ‘quelques’ rather than the more complex ‘certains’. One important conclusion from these child studies looking at quantity implicature seems to be that there really is no principled limitation on young children’s ability to derive quantity implicatures. They seem clearly aware of expectations of informativeness, they produce appropriately informative utterances and they can interpret utterances with quantity implicatures. So we could say that the basics of even four-year-old children’s pragmatic competence is in place. However, even in the act-out task of Pouscoulous et al (2007) there are still differences across age groups which look like a trend toward typical adult responses. Pouscoulous et al (2007), following Noveck (2001), propose that this is due to the cost of the extra inference involved in the enriched interpretation. This view is supported by the adult

Experimental pragmatics

577

studies reported in Bott and Noveck (2004) where it is found that adults put under time pressure tend to respond more often without the implicature. When considering the difference between performance in act out tasks and TVJTs, Katsos and Bishop (under review) argue that processing cost may not be the whole story but that children may be more inclined to assent to underinformative utterances in TVJTs due to the operation of a principle of charity. In light of the fact that even four-year-old children can derive quantity implicatures, it is clear that they must operate with such a principle, as do adults when faced with a statement that clearly could be construed as true or false. However, the puzzle with the developmental trend found in these tasks lies more with adult responses. Note that it is not as though the adults tend to give a ‘correct’ response while children deviate from this. In some respects, the child’s ‘logical’ response is more correct than that of the adults. If it is true that the principle of charity operates both with adults and children, then the question really is why adults tend not to see the alternative interpretation. We may be able to find the answer to this question from the adult studies reviewed above. There we found very rapid access to quantity implicatures in adults and speculated that adult access to this type of implicature rests on a combination of automatic, frequency based heuristics. So it may be that what develop between childhood and adulthood are the heuristic mechanisms that both lighten processing load and make judgements more certain when it comes to valuation tasks. The tale of children’s abilities with quantity implicature has played itself out with only minor modifications when it comes to the development of the so-called Principle B effects studied by syntacticians. According to the favoured theory of these effects, due to Reinhart (see Reinhart 1983, 2006), informants judge that the noun phrases in (8a) cannot be co-referential due to a pragmatic parsimony principle similar to the one we might invoke to account for the contrastive implicatures studied by Sedivy and colleagues: 8. a. Mama Bear washed her. b. Mama Bear washed herself. The idea is that if the speaker wished to express the proposition that Mama Bear washed Mama Bear, then she would have used the reflexive form as in (8b), as it is unambiguous. Many child studies of the Principle B phenomena have shown that children abide by the syntactic principle restricting genuine binding but seem to fail to appreciate that (8a) should not normally be used to mean (8b) for pragmatic reasons (see Chien and Wexler 1990, Grodzinsky and Reinhart 1993, Thornton and Wexler 1999). However, Conroy et al (2009) have challenged certain methodological aspects of these studies and show that 4 year old children are capable of rejecting the pragmatically inappropriate interpretation of (8a) given the right kind of contextual support. Conroy et al conclude, contra Chien and Wexler (1990) that children do have the requisite pragmatic abilities to rule out co-reference. But as with the implicature studies, they do recognise that there is a kind of developmen-

578

Richard Breheny

tal trend. Conroy et al (2009) follow Reinhart (2006) in proposing that the divergence in child responses from the adult norm may have to do with the processing resources involved. But, again, it may be that the full developmental story turns on the development of the heuristics which make adults confidently reject the infelicitous utterance. To sum up this section, we have seen two examples of research on pragmatic development where, after initial indications to the contrary, it seems children have full Gricean pragmatic competence as early as anyone can tell. We have also see that the methods used in child studies have thrown up some intriguing questions as to why children’s behaviour in some studies deviate from that of adults. Finding out what develops and how, may shed light on the steady-state adult competence, while research on the mechanisms that underlie adult processing may in turn shed light on child pragmatic development.

7.

The ontogenesis of pragmatics – Studies in pre-linguistic communication

One conclusion we may draw from studies of toddlers’ pragmatic abilities is that as soon as they have access to the syntactic binding principles, they can rule out pronominal co-reference on pragmatic grounds, and as soon as they grasp the meaning of indefinite quantifiers like ‘some’, they can derive quantity implicatures. We may then wonder how these pragmatic abilities are ever-present through the development of linguistic abilities. The answer coming from pre-linguistic research seems to be that it is because infants are fully competent Gricean communicators before they acquire language. In a series of interactive pointing production and comprehension tasks, Tomasello and colleagues (see Tomasello 2008, Tomasello et al 2007) have shown that children treat communicative gestures as involving the specifically reflexive intention first brought to light in Grice’s work on speaker meaning (see Grice 1957). That is, not only do infants see communication as involving an intention that another person obtain information but as doing so on the basis of recognising the intention to inform. This overtness condition has been analysed in a number of ways by philosophers and cognitive scientists but in practice it requires certain cues on the part of the communicator, such as eye-contact or interactive, contingent behaviour, and certain cues on the part of audience to signal uptake. In elicited production tasks, Liszkowski et al. (2004) elicited infant points to an interesting event and, in one condition, the adult experimenter simply attended to the event in response, rather than also acknowledge interest to the child. In a baseline condition, the experimenter both attended to the event and acknowledged uptake with the child. In the no-uptake condition, as compared to the baseline, children repeated the communicative act until the experimenter responded appropriately.

Experimental pragmatics

579

In a comprehension experiment, Behne et al (2005) engaged infants in a game where a prize is hidden in one of two containers. In one experimental condition, the experimenter engages the infant with ostensive cues (particularly, eye contact) and points to a container. Behne et al found that at 14 months, infants are above chance in choosing the indicated container, while at 18 months they are virtually at ceiling. In another condition, the experimenter forms the same hand gesture, fashioned in the shape of a pointing, but does not engage the child in eye contact or use other ostensive cues. In this condition neither age group favoured the ‘indicated’ container. In more recent research, evidence has accumulated that infants understand communicative acts not simply in terms of what is relevant to them but in terms of what the communicator would think is relevant (see Liebal et al 2009, and particularly, Southgate et al 2009). The overall picture is that by the time children approach the one- or two-word phase, they are already employing sophisticated pragmatic abilities. It now seems clear that these abilities appear to develop throughout infancy, from around 4 months of age. For instance, Farroni et al (2003) show that 5 month-old infants tend to follow shifts in eye direction only if the shift is preceded by a period of eye contact with the infant. Similar research, summarised in Csibra and Gergely (2009) and Csibra (2010) points to the conclusion that the specifically human mode of communication involving the recognition of complex intentions develops along pre-determined pathways prior to the development of fuller social cognitive abilities at the end of the first year. Thus human pragmatic competence may have an innate basis and involve specific neural structures (see Grossman et al 2008). How this competence relates to more general human social cognitive abilities and to general linguistic abilities remain intriguing open questions (see Tomasello 2008, Breheny 2006, Csibra 2010).

8.

Summary

In the past decade, there has been a significant upsurge in experimental research in the domain of intentional pragmatics, that is, research into our understanding of speaker’s meaning based on inferences about intentions in context. This upsurge has come from different sources, more or less independently. On the one hand, ‘Experimental Pragmatics’ has attempted to apply experimental methodology to illuminate theoretical debates among linguists.4 On the other hand, psycholinguists have become increasingly interested in how information about speakers’ intentions inform rapid on-line interpretation processes. Meanwhile, there has been a revolution in attitudes towards toddlers’ linguistic pragmatic abilities; and finally research originating mostly in the infant social cognitive domain has raised the exciting possibility that intentional pragmatic abilities are part of the basic human makeup.

580

Richard Breheny

Taking this research together, we are by far better informed about intentional pragmatic abilities now than ten years ago. But there are still very many empirical questions that remain open. This paper has offered a small sample of recent work and highlighted some of those open questions.

Notes * This work was carried out with the support of a grant from the UK Arts and Humanities Research Council (Ref: AH/E002358/1) and from the Centre for the Study of Mind in Nature (Oslo), ‘Linguistic Agency’ project. 1. For a general overview of (types of) implicature cf. Huang this volume. 2. Although it is argued in Saul (2002) that Grice’s account of conversational implicature should be understood to be a normative account, there is very little textual evidence for this and far more evidence that Grice would have wished to remain neutral on this issue (see Petrus 2010). 3. See van Berkum in press for an overview; N400 refers to a negativity that appears in an ERP reading 400 ms after the offset of a stimulus. It is widely thought of as an indicator of the processing of meaning. Larger N400 suggests extra semantic processing work (as when there is semantic anomaly). 4. Here we have reviewed work focussed on quantity implicatures, but there is other work on metaphor (see Glucksberg 2003, Rubio-Fernandez 2007), presupposition (see Schwarz 2007, Chemla and Bott 2010) and other areas of pragmatics (see Noveck and Sperber 2004 for an early collection).

References Altmann, Gerry T.M. and Yuki Kamide 1999 Incremental interpretation at verbs: restricting the domain of subsequent reference. Cognition 73: 247–264. Altmann, Gerry T.M. and Yuki Kamide 2007 The real-time mediation of visual attention by language and world knowledge: Linking anticipatory (and other) eye movements to linguistic processing. Journal of Memory and Language 57: 502–518. Apperly, Ian A. and Stephen A. Butterfill 2009 Do humans have two systems to track beliefs and belief-like states? Psychological Review 116(4): 953–970. Atlas, Jay and Stephen C. Levinson 1981 It-clefts, informativeness and logical form: radical pragmatics (revised standard version). In: Peter Cole (ed.), Radical pragmatics, 1–61. New York: Academic Press. Barr, Dale 2008 Pragmatic expectations and linguistic evidence: Listeners anticipate but do not integrate common ground. Cognition 109: 18–40.

Experimental pragmatics

581

Behne, Tanya, Malinda Carpenter and Michael Tomasello 2005 One-year-olds comprehend the communicative intentions behind gestures in a hiding game. Developmental Science 8: 492–499. Berkum, Jos J.A. van in press The neuropragmatics of ‘simple’ utterance comprehension: An ERP review. To appear in: Ulrich Sauerland and Kazuko Yatsushiro (eds.), Semantic and pragmatics: From experiment to theory. New York: Palgrave. Borg, Emma 2004 Minimal Semantics. Oxford: Oxford University Press. Bott, Lewis and Ira A. Noveck 2004 Some Utterances are Underinformative: The Onset and Time Course of Scalar Inferences. Journal of Memory and Language 51(3): 437–457. Breheny, Richard 2006 Communication and folk psychology. Mind and Language 21(1): 74–107. Breheny, Richard, Napoleon Katsos and John Williams 2006 Are Scalar Implicatures Generated by Default? Cognition 100: 434–463. Breheny, Richard, Heather J. Ferguson and Napoleon Katsos under review, a How quantity implicatures are accessed in incremental utterance interpretation: Evidence for full-blown Gricean inferencing. Breheny, Richard, Heather J. Ferguson and Napoleon Katsos under review, b On the timecourse of access of conversational implicatures in incremental interpretation. Brennan, Susan E. and Herbert H. Clark 1996 Conceptual pacts and lexical choice in conversation. Journal of Experimental Psychology: Learning, Memory, and Cognition 22: 1482–1493. Brown-Schmidt, Susan 2009 Partner-specific interpretation of maintained referential precedents during interactive dialog. Journal of Memory and Language 61: 171–190. Brown-Schmidt, Susan, Christine Gunlogson and Michael K. Tanenhaus 2008 Addressees distinguish shared from private information when interpreting questions during interactive conversation. Cognition 107: 1122–1134. Carston, Robyn 2002 Thoughts and utterances. Oxford: Blackwells. Chemla, Emmanuel and Lewis Bott 2010 Processing presuppositions: dynamic semantics vs pragmatic enrichment. Ms Institut Jean Nicod (http://www.emmanuel.chemla.free.fr/Material/ChemlaBott-ProcPresProj.pdf) Chien, Yu-Chin, and Kenneth Wexler 1990 Children’s knowledge of locality conditions in binding as evidence for the modularity of syntax and pragmatics. Language Acquisition 1: 225–295. Chierchia, Gennaro 2004 Scalar implicatures, polarity phenomena and the syntax/pragmatics interface. In: Adriana Belletti (ed.), Structures and beyond, 39–103. Oxford: Oxford University Press. Clark, Herbert H. 1996 Using Language. Cambridge: Cambridge University Press.

582

Richard Breheny

Conroy, Anastasia, Eri Takahashi, Jeff Lidz and Colin Phillips 2009 Equal Treatment for All Antecedents: How Children Succeed with Principle B. Linguistic Inquiry 45: 446–486. Cooper, Roger M. 1974 The control of eye fixation by the meaning of spoken language: A new methodology for the real-time investigation of speech perception, memory, and language processing. Cognitive Psychology 6: 84–107. Crain, Stephen and Cecile McKee 1985 The acquisition of structural restrictions on anaphora. In: Proceedings of NELS 16. Ed. by Stephen Berman, Jaewoong Choe and Joyce McDonough, 94–110. Amherst: University of Massachusetts, GLSA. Csibra, Gergely 2010 Recognizing communicative intentions in infancy. Mind and Language 25: 141–168. Csibra, Gergely and Gyorgy Gergely 2009 Natural Pedagogy. Trends in Cognitive Science 13: 148–153. Farroni, Teresa, Eileen M. Mansfield, Carlo Lai and Mark H. Johnson 2003 Infants perceiving and acting on the eyes: Tests of an evolutionary hypothesis. Journal of Experimental Child Psychology 85: 199–212. Garrod, Simon and Anthony Anderson 1987 Saying what you mean in dialog: A study in conceptual and semantic co-ordination. Cognition 27: 181–218. Gazdar, Gerald 1979 Pragmatics: Implicature, Presupposition, and Logical Form. New York: Academic Press. Geurts, Bart 2010 Quantity Implicature. Cambridge: Cambridge University Press. Geurts, Bart and Nausicaa Pouscoulous 2009 No scalar inferences under embedding. Semantics and Pragmatics 2(4): 1–34. Glucksberg, Sam 2003 The psycholinguistics of metaphor. Trends in Cognitive Sciences 7: 92–6. Graesser, Arthur C., Morton Ann Gernsbacher Susan R. Goldman 2003 Handbook of discourse processes. Mahwah, NJ: Lawrence Erlbaum. Grice, H. Paul 1957 Meaning. Philosophical review 66: 377–388. Reprinted in Grice (1989, 213–223). Grice, H. Paul 1989 Studies in the way of words. Cambridge, MA: Harvard University Press. Grice, H. Paul 2001 Aspects of Reason. Oxford: Oxford University Press. Grodner, Daniel and Julie Sedivy in press The effects of speaker-specific information on pragmatic inferences. In: Neal J. Pearlmutter and Edward Gibson (eds.), The Processing and Acquisition of Reference, Cambridge, MA: MIT Press. Grodner, Daniel J., Natalie M. Klein, Kathleen M. Carbary and Michael K. Tanenhaus 2010 “Some,” and possibly all, scalar inferences are not delayed: Evidence for immediate pragmatic enrichment. Cognition.

Experimental pragmatics

583

Grodzinsky, Yosef and Tanya Reinhart 1993 The innateness of binding and coreference. Linguistic Inquiry 24: 69–101. Grossmann, Tobias, Mark H. Johnson, Sarah Lloyd-Fox, Anna Blasi, Fanni Deligianni, Clare Elwell and Gergely Csibra 2008 Early cortical specialization for face-to-face communication in human infants. Proceedings of the Royal Society, London B, 275: 2803–2811. Guasti, Maria Teresa, Gennaro Chierchia, Stephen Crain, Francesca Foppolo, Andrea Gualmini and Luisa Meroni 2005 Why children and adults sometimes (but not always) compute implicatures. Language and Cognitive Processes 20(5): 667–696. Hanna, Joy, Michael K. Tanenhaus and John C. Trueswell 2003 The effects of common ground and perspective on domains of referential interpretation. Journal of Memory and Language 49: 43–61. Heller Daphna, Daniel Grodner and Michael K. Tanenhaus 2008 The role of perspective in identifying domains of reference. Cognition 108: 831–836. Horn, Laurence 1984 A new taxonomy for pragmatic inference: Q-based and R-based implicature. In: Deborah Schiffrin (ed.), Meaning, Form and Use in Context, 11–42. (GURT ’84). Washington: Georgetown University Press. Horn, Laurence 1989 A Natural History of Negation. Chicago, IL: University of Chicago Press. Horn, Laurence 2006 The border wars: a neo-Gricean perspective. In: Klaus von Heusinger and Ken Turner (eds.), Where Semantics Meets Pragmatics, 21–48. Amsterdam: Elsevier. Huang, Yi Ting, and Jesse Snedeker 2009 On-line Interpretation of Scalar Quantifiers: Insight into the Semantic-Pragmatics Interface. Cognitive Psychology 58: 376–415. Katsos, Napoleon and Dorothy Bishop under review The development of informativeness from the speaker’s and the comprehender’s perspective. Keysar, Boaz, Dale J. Barr, Jennifer A. Balin and Jason S. Brauner 2000 Taking perspective in conversation: The role of mutual knowledge in comprehension. Psychological Sciences 11: 32–38. Keysar, Boaz, and Dale J. Barr 2005 Coordination of action and belief in communication. In: John C. Trueswell and Michael K. Tanenhaus (eds.), Approaches to Studying World Situated Language Use: Bridging the Language-as-Product and Language-as-Action, 71–94. Cambridge, MA: MIT Press. Kronmuller, Emundo and Dale J. Barr 2007 Perspective-free pragmatics: Broken precedents and the recovery-from-preemption hypothesis. Journal of Memory and Language 56: 436–455. Levinson, Stephen 2000 Presumptive Meanings. Cambridge, MA: MIT Press. Liebal, Kristin, Tanya Behne, Malinda Carpenter and Michael Tomasello 2009 Infants use shared experience to interpret pointing gestures. Developmental Science 12(2): 264–271.

584

Richard Breheny

Liszkowski, Ulf, Malinda Carpenter, Anne Henning, Tricia Striano and Michael Tomasello 2004 12-month-olds point to share attention and interest. Developmental Science 7: 297–307. MacDonald, Maryellen C. and Mark S. Seidenberg 2006 Constraint satisfaction accounts of lexical and sentence comprehension. In: Matthew J. Traxler. and Morton Ann Gernsbacher (eds.), Handbook of Psycholinguistics. 2nd ed., 581–611. London: Elsevier. McKoon, Gail and Roger Ratcliff 1992 Inference during reading. Psychological Review 99: 440–466. Markman, Arthur B. and Valerie S. Makin 1998 Referential communication and category acquisition. Journal of Experimental Psychology: General 127: 331–354. Metzing, Charles, and Susan E. Brennan 2003 When conceptual pacts are broken: partner-specific effects on the comprehension of referring expressions. Journal of Memory and Language 49: 201– 213. Noveck, Ira A. 2001 When children are more logical than adults: Investigations of scalar implicature. Cognition 78/2: 165–188. Noveck, Ira, and Andres Posada 2003 Characterising the time course of an implicature. Brain and Language 85: 203–210. Noveck, Ira A. and Dan Sperber (eds.) 2004 Experimental Pragmatics. New York: Palgrave. Papafragou, Anna and Julien Musolino 2003 Scalar implicatures: Experiments at the Semantics-Pragmatics interface. Cognition 86: 253–282. Petrus, Klaus 2010 Introduction: Paul Grice, Philosopher of Language, but more than that. In: Klaus Petrus (ed.), Meaning and Analysis: New Essays on H. Paul Grice, 1–30. New York: Palgrave. Pouscoulous, Nausicaa, Ira A. Noveck, Guy Politzer, and Anne Bastide 2007 A developmental investigation of processing costs in implicature production. Language acquisition 14: 347–376. Reinhart, Tanya 1983 Anaphora and semantic interpretation. London: Croom Helm. Reinhart, Tanya 2006 Interface strategies. Cambridge, MA: MIT Press. Rubio-Fernandez, Paula 2007 Suppression in Metaphor Interpretation: Differences between Meaning Selection and Meaning Construction. Journal of Semantics 24: 345–371. Saul, Jennifer 2002 Speaker meaning, what is said and what is implicated. Nous 36: 228–248. Sauerland, Ulrich 2004 Scalar implicatures in complex sentences. Linguistics and Philosophy 27: 367–391.

Experimental pragmatics

585

Sedivy, Julie C. 2003 Pragmatic versus form-based accounts of referential contrast: Evidence for effects of informativity expectations. Journal of Psycholinguistic Research 32: 3–23. Sedivy, Julie C., Michael K. Tanenhaus, Craig G. Chambers and Gregory N. Carlson 1999 Achieving incremental semantic interpretation through contextual representation. Cognition 71: 109–147. Schwarz, Florian 2007 Processing Presupposed Content. Journal of Semantics 24(4): 373–416. Southgate, Victoria, Coralie Chevallier and Gergely Csibra 2009 Sensitivity to communicative relevance tells young children what to imitate. Developmental Science 12(6): 1013–1019. Sperber, Dan and Deirdre Wilson 1986 Relevance: Communication and Cognition. Oxford: Blackwell. Sperber, Dan and Deirdre Wilson 1995 Postface to 2nd Edition. In: Dan Sperber and Deidre Wilson, Relevance: Communication and cognition, pl. add+ rede. 2nd ed. Oxford: Blackwell. Tanenhaus, Michael K., Michael J. Spivey-Knowlton, Kathleen M. Eberhard and Julie C. Sedivy 1995 Integration of visual and linguistic information during spoken language comprehension. Science 268: 1632–1634. Thornton, Rosalind and Kenneth Wexler 1999 Principle B, VP-ellipsis, and interpretation in child grammar. Cambridge, MA: MIT Press. Tomasello, Michael 2008 Origins of Human Communication. Cambridge, MA: MIT Press. Tomasello, Michael, Malinda Carpenter and Ulf Lizskowski 2007 A new look at infant pointing. Child Development 78: 705–22. Warner, Richard 2001 Introduction: Grice on reasons and rationality. In: H. Paul Grice (2001), vii–xxxviii.

21.

Corpus-based pragmatics I: qualitative studies1 Gisle Andersen

1.

Contextualising corpus-based pragmatics

This paper concerns the use of computational methods for collection and analysis of language data as a basis for research in pragmatics. The aim is to describe advantages and problems associated with the use of corpora, and to provide both a bird’s-eye view of the field as well as an in-depth account of a limited set of topics of particular salience. A main claim to be argued here, with reference to several papers, is that, as a method for empirical research, corpus-based pragmatic analysis is “essentially inductive and interpretative” (Mosegaard Hansen 1998: 214) and can never be performed without the reliance on the analyst’s intuition (cf. especially section 3). Corpus-based studies are conceptually related to other ways of doing empirical pragmatics, such as anthropological or ethnological field work, sociolinguistic interviews and the use of experimental techniques and elicitation tests (Chafe 1992). The insights offered by the studies reviewed in the current survey have emerged via meticulous analysis of corpus data, as opposed to other empirical methods. The studies discussed here are further distinguished from those discussed in Christoph Rühlemann’s contribution to the current volume, which focuses on quantitative studies. However, corpus-based studies are rarely exclusively qualitative or quantitative but commonly describe pragmatic functions as well as quantifiable distributional aspects, involving for instance comparisons between users and user groups within or across corpora. The current study, then, describes that body of pragmatics literature that has its main focus on functional aspects but focuses less on measurable distributional differences. The term ‘discourse linguistics’ has been used as “an umbrella term for discourse analysis, discourse studies, text linguistics, pragmatics, conversation analysis and other related approaches to the study of discourse” (Virtanen 2009: 49), a list of topics which is broadly consistent with the topics to be reviewed in this chapter. The application of corpus-linguistic method in this knowledge field has been relatively limited, especially compared to its application in the study of more formal topics such as syntax, morphology and lexis (cf. McEnery and Wilson 2001; Kennedy 1998). Explorations of corpora have nevertheless been fruitful in pragmatics, as evidenced by a series of studies which investigate how interaction is sequentially and interactionally organised, how the attitudes and perspectives of speakers and hearers are manifested in spoken discourse, how texts are coherently structured beyond the sentence level, etc. Among the topics studied are turn-taking phenomena, pauses and repairs, discourse markers, hedging and vague language,

588

Gisle Andersen

small talk, speaker attitudes and so on. (cf. section 2). Despite the variety of topics, there are many aspects of pragmatics that have not been subjected to corpus-based investigations to the extent they could have been, and I shall argue in favour of an extended application of corpora in pragmatics, to capture a wider range of phenomena within “the cognitive, social and cultural science of language and communication” (Verschueren 2009: 1). Corpus linguistics and pragmatics can be seen as two related, but historically distinct research traditions. I endorse the generally accepted view expressed by corpus linguists such as Leech (1992), Kennedy (1998), McEnery and Wilson (2001) and Virtanen (2009), that pragmatics and corpus linguistics are ontologically different academic pursuits, in that the former is a branch of linguistics – albeit many-faceted and complex, while the latter is not. Rather, corpus linguistics is regarded as a certain way of performing empirical studies in linguistics, a “research approach” (Biber 1988) and “a methodological basis for doing linguistic research” (Leech 1992: 105; cf. also Kennedy 1998: 7). Empirical pragmatics crucially depends on authentic language data, and corpus linguistics is one methodological approach which may be used for pragmatics research. Like other kinds of empirical language data, corpora are documentations of the choices made by language users and surface manifestations of the underlying communicative competence of the speakers whose language the corpus represents (de Beaugrande 1999; cf. also section 3). In the words of Chafe (1992:88), “corpora provide a particularly useful means of access to the language-related capabilities of the mind”. A notable recent development in corpus linguistics has been the increased focus on corpus-driven research (as opposed to corpus-based research; TogniniBonelli 2001: 84–87; Sinclair 2004), which “differs from the standard practice of linguistics in that it makes minimal a priori assumptions regarding the linguistic constructs that should be employed for the analysis” (Biber 2009: 276). In corpusbased research, the researcher approaches the data equipped with a set assumptions as to its linguistic structure and expected findings. Corpus-driven research, on the other hand, explores linguistic structures that emerge from the corpus, thus inductively “exploiting the potential of a corpus to identify linguistic categories and units that have not been previously recognised” (Biber 2009: 278). One of the main claims that could be made about pragmatics research using corpora is that it is – by and large – corpus-based and not corpus-driven (Sandford 2008). Most studies have as their starting point a previously recognised structure, where the main value of the empirical work results from the investigator’s meticulous analysis of the range of different functions associated with particular linguistic forms, as evidenced by the naturally occurring discourse contained in the corpus. Examples would be a structure like the it-cleft in English (e.g. Collins 2005) or an individual word form like really; see e.g. Stenström (1986b), who provides a detailed account of this form and its various functions in the London-Lund Corpus. The corpus method principally involves inspecting data via a search term that yields relevant

Corpus-based pragmatics I: qualitative studies

589

results, i.e. the method known as “one-to-one searching” (Ädel and Reppen 2008: 2). Often, however, the inspection of extracted corpus data may well result in interesting observations of functional categories not previously accounted for or even recognised, such as the observation made by Andersen (2001: 139ff) that the invariant British English form innit (from isn’t it or ain’t it) has an extended function as a marker of mutual manifestness (common ground) directed towards the previous speaker’s utterance, in addition to its more generally recognised use as a tag question which modifies a proposition of the current speaker. Most corpusbased pragmatics studies are thus form-based. The computerised corpus lends itself well to this kind of approach, in which relevant instances of a type can be identified as tokens in actual use. The methodological counterpart is the function-based approach, which takes as its basis a particular pragmatic function and describes its possible realisations in actual discourse. Functions that have been investigated include conversational openings and closing, topic transition, making a request, expressing disagreement or the like. An example of a study where this method is successfully applied is Mauranen’s (2004) study of hedges, items which make propositions more or less fuzzy and add “indeterminacy of one kind or another to an utterance” (Mauranen 2004: 194). She shows that in a contextually constrained corpus, the Michigan Corpus of Academic Spoken English (MICASE), this function is typically performed by a category of items kind of, sort of, or something, or so, somewhat, just, and a little bit, items which share some functional characteristics as hedges but which perform specialised functions in different contexts of use. In both the form and function-based approaches, the role of the corpus is essentially to provide a repository of linguistic examples to be accounted for. Until recently, it has been less customary to carry out corpus-driven studies in pragmatics, despite the fact that this approach has been applied widely and successfully in related fields, particularly phraseology and lexicography (cf. Sinclair 1991; Partington 1998; McEnery et al. 2006; Biber 2009). Notably, the corpus-based and corpus-driven approaches should be seen as complementary rather than conflicting. As Biber argues, “corpus-driven research is not in any way superior to corpusbased research. The two require radically different methods, and thus they offer the possibility of uncovering radically different perspectives on language structure and use” (Biber 2009: 279). Studies in pragmatics are likely to gain from using corpora in manners which go beyond the dominating pattern of one-to-one search used in corpus-based studies. That this combination may be fruitful can be demonstrated by a recent paper by Walsh et al. (2008). Their study of vagueness markers in academic discourse is essentially a synthesis of “two methodologies not always seen as complementary, corpus linguistics and conversation analysis. These have much to offer each other as they provide both quantitative and qualitative insights respectively” (Walsh et al. 2008: 16). Using this combinatory approach, the authors show that vague language varies considerably according to mode of discourse

590

Gisle Andersen

(managerial, classroom, conversational) and that “these qualitative differences are a valuable means of understanding the complex relationship between language and learning” (Walsh et al. 2008:9). Furthermore, discourse functions are often associated with processes of language change, involving the emergence of new discourse items, realised either as new forms, such as the English phrase yeah right used as an irony marker in Norwegian (Andersen 2006), or as new uses of old forms, such as the new quotative marker BE like (Tagliamonte and Hudson 1999; Andersen 2001; Rühlemann, this volume). These cases exemplify processes of change (borrowing vs. structural reanalysis) that are observable using corpus-driven methods. Studies in pragmatics should investigate changes in vocabulary and in collocational and colligational patterns, with a view to revealing new ways of organising discourse, structural change, categorical reanalysis and emergent grammaticalisation (Traugott and Heine 1991; Lindquist and Mair 2004). In other words, there are good reasons for complementing the well-tested and approved method of corpus-based pragmatics with a corpus-driven approach. The current chapter focuses on studies where corpora have been applied successfully in pragmatics research. The chapter is organised as follows. Section 2 provides an historical overview of the field of corpus-based pragmatics, giving a brief description of a wide range of studies in order to show the breadth of this field. Section 3 concerns method and has the two-fold purpose of arguing in favour of a corpus-based approach as well as problematising the hermeneutical aspects of this method. Section 4 provides an in-depth account some particularly salient topics and contributors in corpus-based pragmatics. These include discourse markers, prosody, listener activities (backchannels) and corpus-based models of discourse structure. The final section reiterates some of the main claims and observations made in the previous sections.

2.

An overview of the field

Corpora are compilations of naturally occurring spoken or written language that can be accessed on a computer. Such compilations may be monolingual or multilingual and may represent general language or specific domains (professional/academic corpora). This study will show that corpora form a well-prepared basis for systematic, descriptive studies of instances of actual speech, of language variation and of how social context constrains communicative practices. Spoken corpora have a favoured position in pragmatics. A spoken corpus is a collection of speech data made accessible via a computer, containing at least transcriptions of speech but often also audio and/or video files (for problems and types of transcription, cf. Roger J. Kreuz and Monica A. Riordan, this volume). Corpora have been available for linguistic research from the 1960s. Since the early days of corpus linguistics, there has been an interest in using corpora as a

Corpus-based pragmatics I: qualitative studies

591

means of exploring functional and contextual aspects of language use. However, following the compilation of the first-generation written corpora, the Brown University Corpus (Francis 1967) and the Lancaster-Oslo/Bergen corpus (Johansson and Hofland 1994), the bulk of studies were within other fields of linguistics, particularly syntax, semantics, morphology and lexicology (cf. Altenberg 1986 for an overview). Though few and far between, early pragmatics studies include linguistic stylistics (Enkvist 1973) and aspects of text linguistics, addressing such issues as information structure, fronting and topicalisation (Gustafsson 1982, 1983, 1985; Davidson 1977). The parallel structure and content of the Brown and LOB corpora enabled reliable comparative studies of American and British English that focused on lexical and grammatical differences (e.g. Hofland and Johansson 1986; Algeo 1988). The ground-breaking compilation of the first spoken corpus, the London Lund Corpus of Spoken English (LLC, Svartvik and Quirk 1980), triggered a range of studies in pragmatics (Altenberg 1986). The corpus was based on recordings of spoken data from the Survey of English usage (SEU), which in its pre-computerised version had resulted in Crystal’s detailed empirical work on intonation (Crystal 1972, 1975; Crystal and Quirk 1964), Haegeman’s (1984) empirically based account of pragmatic conditionals like the story, if so it may be termed, is weak and loose and Kussmaul’s (1978) early study of Indikatoren von Sprechakten (speech act markers). LLC inspired research into phenomena relating to speech production and performance, such as ‘slips of the tongue’ (Garnham et al. 1981), repairs (Hedström 1984) and hesitation (Oreström 1977). Several corpus-based studies were collected in what must be seen as a seminal volume on conversational discourse, namely Enkvist (1982). This volume addresses an array of phenomena relating to the study of speech, including turn-taking, feedback and repair (Faerch and Kasper 1982; Oreström 1982; Stenström 1982), intonation and prominence (Brazil 1982), pragmatic particles (Östman 1982), referential expressions (Thavenius 1982) and functional aspects of negation (Tottie and Paradis 1982). The book contains pioneering work on theoretical and methodological topics such as transcription, segmentation and annotation of conversational data (Loman 1982; Svartvik 1982) and the validity of contrastive methods (Sajavaara 1982). Also notable is the variety of languages represented by the corpus data described in this volume (English, French, German, Hebrew, Russian and Swedish). A topic which received early attention is discourse markers (dealt with more extensively in section 4). The early contributions include Erman’s work on pragmatic expressions like you know, you see and I mean (1986, 1987), Stenström’s work on really (1986b) and Svartvik’s investigation on the various functions of well (1980). Moreover, Stenström’s (1982, 1984) extensive work on question/response sequences (cf. section 4) dates from this period, as do some accounts that deal specifically with tag questions (Nässlin 1984; Bublitz 1979). The availability

592

Gisle Andersen

of both spoken and written corpora also inspired comparative work, particularly Biber’s (1986, 1987, 1988) work focusing on the lexical, grammatical and pragmatic differences between these two modes of communication. In addition, LLC formed the basis of an at least partly corpus-based account of conversational phenomena, Wardhaugh (1985), which has been criticised for its superficiality and lack of an analytical apparatus and technical terminology (Garnham 1986). A second wave of corpus-based studies came with the development of additional English language corpora, such as the Kolhapur Corpus of Indian English (Shastri 1988), the Birmingham Collection of English Texts (Renouf 1986; Sinclair 1987), the Tosca Corpus (van den Heuvel 1988), The Nijmegen Corpus (Aarts and Meijs 1986) and The Helsinki Corpus of English Texts, representing historical English (Rissanen 1986; Kytö and Rissanen 1988). Moreover, the introduction of formal annotation schemes for intonation in speech corpora was a timely and relevant innovation that became highly influential to pragmatics research (Svartvik 1987, 1988; Knowles 1986; Knowles and Lawrence 1987; Taylor and Knowles 1988; Knowles et al. 1992). This sparked research on the pragmatic aspects of prosody (Altenberg 1987; Nevalainen 1987a) and studies exploring the correlation between intonation and type of speech act (Geluykens 1986, 1987a, 1988) and between intonation and information structure (Nevalainen 1987b). The annotated corpora enabled studies of the functional characteristics of pauses and other types of prosodic segmentation (Stenström 1986a, 1988). The range of pragmatic phenomena accounted for increased in the 1980s and came to include investigations of dislocation (Geluykens 1987b; Aijmer 1989), evidentiality and epistemic modality (Westney 1986; Holmes 1988; Coates 1987; Biber and Finegan 1989), turn-taking (Brodda 1988) and direct vs. indirect speech (Aijmer 1986). There has been a gradual increase in the range of domains of language use that are represented by corpora, and especially in the period from the 1990s, the growth in size and coverage has been remarkable. The second generation of corpora includes large, general-purpose corpora like the 100 million word BNC corpus (Crowdy 1995; Aston and Burnard 1998) and the 20 million word Birmingham Corpus. A notable coordinated effort is the International Corpus of English (ICE, Greenbaum 1991, 1992, 1996; Banjo 1996), which has been used for cross-varietal studies of the English spoken in many Anglophone societies. The Santa Barbara Corpus of Spoken American English (CSAE, Chafe et al. 1991) constitutes the American spoken component of ICE. The family of corpora also came to include corpora for specific purposes, such as foreign learner corpora like the International Corpus of Learner English (ICLE, Altenberg 1997; Granger 1997) specialised LSP corpora (Faber and Lauridsen 1991; Fang 1993) such as corpora of doctor-patient interactions (Thomas and Wilson 1996), and parallel corpora such as the EnglishNorwegian Parallel Corpus (ENPC, Johansson and Hofland 1994) and the EnglishSwedish Parallel Corpus (ESPC, Altenberg and Aijmer 2000). Finally, corpora that focus on regionally and socially restricted varieties of English include the Bergen

Corpus-based pragmatics I: qualitative studies

593

Corpus of London Teenage language (COLT, Haslerud and Stenström 1995; Stenström et al. 1988). The growth in availability of corpora of various kinds has led to a growth in corpus-based studies in pragmatics, although the relative growth has been larger in other fields of linguistics. Several important contributions can be mentioned. With respect to contrastive studies, Granger and Tyson (1996) and Granger (1997) compare discourse aspects of how native and non-native writers use connectives and participle clauses in English academic writing. Work on phonology and intonation advanced greatly through efforts in the SEC/MARSEC corpora (Bald 1990; Bauer and Holmes 1996; Knowles et al. 1996; Wichmann 1996; Nevalainen 1992). Finally, studies in pragmatics came to cover an ever wider range of topics, including amplifiers (Altenberg 1991), thanking, apologies and requests (Aijmer 1996), register (Biber 1993, 1994, 1995), conversational humour (Fill 1992), expletives and taboos (Stenström 1991, 1995, 2006) and backchannels and response signals (Bublitz 1988; Tottie 1991). In its initial phase, corpus linguistics was practically synonymous with English corpus linguistics. The establishment of the ICAME (International Computer Archive of Modern and Medieval English) organisation in 1979, and its coordinated development and distribution of many English language corpora, have provided a particularly fruitful environment for corpus-based research. Recently, however, corpora for a number of other languages have been compiled, including large, national corpus-building initiatives such as the Croatian National Corpus, the Czech National Corpus, the Spoken Dutch Corpus, the Gothenburg Spoken Language Corpus (Swedish), and a number of other, more targeted corpora developed in connection with specific research projects, especially in sociolinguistics (Andersen 2010). For example, de Haan (2002) uses the Uppsala corpus of contemporary Russian texts to explore different means of expressing modality and negation in Russian and the ambiguity of different markers of modality. Similarly, Schelfhout et al. (2005) explore the Dutch National Corpus, observing that speakers’ use of interjections varies significantly with textual genre as well as with grammatical, functional and prosodic factors of the utterances in which they o ccur. Among the international corpora available, the C-ORAL-ROM project has launched comparable spoken corpora in four Romance languages, French, Italian, Portuguese and Spanish (Cresti and Moneglia 2005). A study like Schneider (2007) draws on no less than 22 corpora of spoken French, Italian and Spanish in a comprehensive comparative account of parenthetical clauses like je crois, je dirais, je pense, je sais pas and je trouve, and its equivalents in the other Romance languages. The Spanish Corpus Oral de Lenguaje Adolescente (COLA) enables comparative studies of the varieties of Spanish used by adolescents in major Hispanophone cities, and has been explored extensively in pragmatics studies (Stenström and Jørgensen 2009). In contrast, the Corpus de Referencia del Espanol Actual (CREA) represents Spanish spoken by adults. Downing (2006) explores CREA in

594

Gisle Andersen

a contrastive study of the stance marker surely and its Spanish cognates. Furthermore, the Italian spoken corpus Lessico dell’italiano parlato (LIP, De Mauro et al. 1993) forms the basis of Waltereit’s (2006) study on the role of speaker motivations in the sense of “abstractions of communicative intentions” (Waltereit 2006: 62) in the grammaticalisation from lexical items to discourse markers of Italian forms such as insomma and diciamo. The Lancaster Corpus of Mandarin Chinese (McEnery and Xiao 2004) is explored by Ming and Chen (2010) in a recent study of how the discourse function of noun phrases affects the word order in noun phrases with a relative clause. Factors such as the information status, animacy, and high versus low discourse salience of head NPs have a bearing on whether a Mandarin Chinese relative clause immediately precedes the head NP or whether it is separated from its head by a numeral-classifier sequence consisting of yi (‘one’) and a classifier. Noun phrases representing abstract entities with a low discourse salience are contrasted with concrete (animate or inanimate) entities with a high discourse salience, and the relative order of the components are explained with reference to the principle of relevance (Sperber and Wilson 1995), as it “seems to indicate the direction in which the relevance is to be sought” (Ming and Chen 2010: 188). Used with a fully realised and highlighted discourse function of introducing a salient participant, the numeral classifier tends to precede the relative clause; otherwise it tends to follow the relative clause. The authors argue that this shows “that clause level structural analysis does not suffice to explain the underlying factors that govern the choice of alternative means of expressions in a language. By contrast, corpus data can be tremendously useful in helping us arriving at a realistic account of the structure of human language” (Ming and Chen 2010: 188). Despite notable studies like the ones mentioned, there is still a considerable bias towards English-based studies in pragmatics in the literature (McCarthy and Carter 2004b).2 Importantly, however, it should be pointed out that the availability of corpora representing different languages, including parallel corpora consisting of original and translated texts, has triggered significant comparative work. A recent volume that reports a number of multilingual studies is Aijmer and SimonVandenbergen (2006). The authors argue that comparative work is interesting for its potential to lead to new knowledge about “what is universal and what is language-specific” (Aijmer and Simon-Vandenbergen 2006: 3), for its potential in investigating etymological or semantic cognates, and for revealing the effects of translation on pragmatic aspects. These general claims are substantiated by the individual papers in the volume, such as Johansson (2006) and Hasselgård (2006), who jointly demonstrate the applicability of parallel corpora in determining the meaning and distribution of the Norwegian English cognate words vel/well and nå/ now, respectively. Moreover, Altenberg’s (2006) account of adverbial connectors shows how cross-linguistic similarities are largely discourse-driven. In the languages compared, Swedish and English, “contrastive and exemplifying connectors often serve to highlight the initial element, syntactically and prosodically, thereby

Corpus-based pragmatics I: qualitative studies

595

indicating a significant break or shift in the discourse. Hence, despite the syntactic differences between English and Swedish, there are interesting parallels between the two languages in the way that connectors adjust to, and signal, the progression of discourse” (Altenberg 2006: 35f). Similarly, Foolen (2006: 70) compares Dutch toch and German doch, showing that the multilingual approach makes it clear “more than a monolingual study could have done, that the language specific uses of ‘the same’ item are indeed rather specific, but at the same time that in all these applications a single schematic meaning was preserved”. The distributional differences of the two items are well accounted for; e.g., judging by translators’ choices when handling doch/toch as an adversative connective, ‘[a]pparently, German does not ‘like’ a bare doch in first sentence position’ (Foolen 2006: 65). This might be explained by German doch’s status as a conjunction. As modal particles, only toch can occur in sentence-final tag position, where it functions as a checking device, inviting an affirmative response from the hearer. As a whole, the collective volume shows that the contrastive corpus-based approach provides a new perspective on fundamental properties of discourse markers, such as their multifunctionality and core meanings. The most recent advances in corpus linguistics are connected with increased data processing and storage capacity, the emergence of web-based corpora, and the (semi-) automatic compilation of large monitor corpora which are constantly updated with new texts representing an “unbroken, chronological data flow” (Renouf 2007: 36). These recent advances have changed the “face of corpus linguistics” (Renouf and Kehoe 2006: 3) to a “much richer cycle of events, consisting not just of small, medium and large primary corpus building, but of specialised and multidimensional secondary corpus building; not just of corpus analysis but also of corpus evaluation; not just of an initial application of theory but of self-reflection and a new concern with theory in the light of experience” (Renouf and Kehoe 2006: 3). The tangible outputs of these processes are ‘super-corpora’ containing hundreds of millions of words, like the Bank of English at the University of Birmingham, The Corpus of Contemporary American English (COCA; cf. Davies 2009) and the Norwegian Newspaper Corpus (Andersen 2005). These continually growing corpora are particularly apt for studying change in progress and for revealing innovations in language use, including the development of new and borrowed lexical items (Renouf 2007; Furiasso and Hofland 2007). Web-based corpora, aka. ‘cyber-corpora’ (Renouf 2007) use web crawler technology to harvest data which are subsequently filtered, annotated and transformed into a corpus. This has resulted in a growing body of corpus-based studies using the web as its prime source of data (Kilgarriff and Grefenstette 2003; Hundt et al. 2007). Although the exploitation of these large corpora in pragmatics has so far been limited (but see section 3), the recent developments forecast a promising future for corpus-based pragmatics.

596 3.

Gisle Andersen

Methodological issues

An important event in the history of corpus linguistics was the 1991 Nobel Symposium on Corpus Linguistics, which is comprehensively reported in Svartvik (1992). This volume does not focus on issues in pragmatics in particular, but it contains an impressive combination of papers that justifies its mention in this context. It presents a theoretical and methodological basis with great scientific validity for any type of corpus linguistic work, despite subsequent major technical developments that go well beyond what the pioneers of corpus linguistics could have foreseen (cf. above). The influential contributors, among them Chafe, Nelson Francis, Fillmore, Halliday, Johansson, Greenbaum, Rissanen and Quirk, discuss a number of theoretical issues and issues in observational methodology, corpus design and corpus use. For instance, Chafe provides a strong argument for corpus linguistic method in cognitive linguistics: “the acceptance of the value of corpora is closely tied to the acceptance of the view that language cannot be separated from what the mind is and what the mind does and how the mind interacts with the universe in which it is located. The modular view of language can get along without corpora; the integrated view requires them” (Chafe 1992: 81). Chafe further contextualises corpus linguistics as an observational technique, placing “the use of corpora within a larger repertoire of observational alternatives” (Chafe 1992: 84). Unlike introspective techniques that involve making judgements regarding invented language, corpus linguistics involves the observation of overt behaviour. Further, corpus linguistics involves natural language, which distinguishes this method from artificial methods like experimental techniques and elicitation tests. Corpora thus reflect what people actually say and write. Ontologically, corpus linguistics shares its characteristics with ethnography, in that it is both a natural and behavioural observation technique. However, the introspection/observation dichotomy is problematic, especially from the point of view of pragmatics. As Milroy (1987) points out, “[i]n practice, generalizations based on analyses of corpuses (sic!) are usually derived using an additional method – either introspection or some kind of experimental procedure. … entirely corpus-based linguistic description of the kind attacked by Chomsky (1957: 15; 1965: 2) is extremely rare and probably represents a concept whose value is largely polemic” (Milroy 1987: 4f). The corpus-based approach to pragmatic phenomena relies on supplementing observations of overt linguistic behaviour with the intuitive knowledge of the analyst. Pragmatics concerns meaning arrived at on the basis of inferential processes. The empirical researcher uses spoken data as his/her primary stimuli on the basis of which he/she makes assumptions as regards the communicative intention of the observed interlocutors. Observable forms in a corpus, such as tokens of the word please, represent instantiations of pragmatic acts performed by speakers, such as the speech act function of requesting somebody to do something, the attitudinal function of signalling tenta-

Corpus-based pragmatics I: qualitative studies

597

tiveness or the politeness function of face-threat mitigation. Naturally, it can be hard to judge regarding speaker intentions. This methodological challenge is well accounted for in Mosegaard Hansen’s (1998) comprehensive account of discourse particles in French, where she argues that “[a]n interpretative approach of course has an inescapably subjective dimension, which means that one cannot in any sense prove the correctness of one’s analysis, but one can only hope to demonstrate it as convincingly as possible” (Mosegaard Hansen 1998: 214). The complexity of spoken language data poses a challenge to the compilers and users of corpora. The study of linguistic phenomena can be enhanced by annotations of various types of morpho-syntactic and discourse-structural information, and contributions like Edwards and Lampert (1993) and Leech et al. (1995) offer systematic accounts on the art of transcription and annotation of spoken discourse. The availability of prosodically annotated corpora has provided uniform methods for studies of intonation and its pragmatic effects (cf. e.g. Stenström 1984; Aijmer 1996). One objection that has been raised against the first generation of corpusbased studies is the fact that they regard the transcriptions themselves as the primary data, while the sound files are not consulted due to their unavailability to the analyst. Although the unavailability of audio files can be explained with reference to historical, practical and technological details (Wichmann 2007), this remains a justifiable scientific objection for at least two reasons. Firstly, like phonetics and phonology, many studies in conversation analysis and discourse analysis need audio data because the interpretation depends on audible phenomena that could never be fully represented in the transcripts: “there is far more in the speech signal to be accounted for than could ever be captured in an orthographic transcription” (Wichmann 2007: 76), and “[n]ot even the most detailed transcription can faithfully represent the ‘raw’ data of spoken interaction” (Mosegaard Hansen 1998: 207; cf. also section 4). Prosodic and segmental phenomena play a more significant role than earlier anticipated in turn-taking, discourse marking, etc. (cf. Couper-Kuhlen 1996; Venditti and Hirschberg 2003; Wichmann 2007). For example, Walker (2004) uses a variety of (mostly telephone-based) corpora to explore how features such as pitch movements, glottal aspects, lingual and labial closures are systematically associated with the signalling of transition relevance. Secondly, it may be problematic for corpus users to rely fully on the judgements of the transcriber, since transcribers often differ with respect to their interpretations of spoken data. Ensuring inter-annotator agreement is a major challenge in corpus compilation, and a study such as Raymond et al. (2002) has shown that even experienced transcribers have different interpretations, particularly when annotating data at phonetic or prosodic levels. This underlines, of course, the analyst’s need for access to the raw spoken data when using corpora (Andersen 2001: 211ff). Fortunately, new technology has made it much easier to access audible spoken data in corpus linguistics. The state-of-the-art spoken corpora contain hypertext

598

Gisle Andersen

links between the transcripts and the audio and/or video data, cf. corpora such as COLA, COLT, ICE-GB and NoTa (Johannessen and Hagen 2008; Andersen 2010). There is thus a tendency in corpus building to move away from ambitious paralinguistic annotation schemes aimed at “transcribing infinity” (Cook 1990: 1) to implementing a more a rudimentary, but nevertheless searchable, prosodic transcription coupled with the actual sound files (Wichmann 2007). Importantly, video data facilitates the analysis of gestures and facial expressions as part of conversational behaviour, but the utilisation of video material in corpus-based pragmatics is still in its infancy. The importance of using video data can be illustrated with reference to minimal response signals (backchannels). Since active listenership may be realised verbally or non-verbally, typically by a gesture such as nodding, the articulated realisations cannot and should not be seen as separate from non-verbal realisations performing the same discourse function (Bjørge 2009). In other words, a comprehensive study requires a holistic approach to language data, in which all aspects of utterances are investigated (cf. also Selting and Couper-Kuhlen 2001; Couper-Kuhlen and Ford 2004). The fruitfulness of such an approach can be shown with reference to a recent dissertation, Loehr (2004), that explores the functional parallelism between intonation and gestures. Specifically the study demonstrates the ability of gestures – like intonation – to signal completeness/incompleteness, information status, focus, emphasis, contrast, and even discourse structure; in fact that hand movements reflect the on-going status of discourse, while “[i]ntonation signalled the discourse status much less subtly” (Loehr 2004: 235). The exploration of the interaction between gesture and intonation is a relatively recent academic pursuit, and studies have looked at temporal alignments of the two modalities and the mapping between their respective hierarchical structures. Interestingly, “there is strong evidence, provided initially by Kendon (1972, 1980) and McNeill (1985, 1992), that gesture and speech in general are two surface facets of a single underlying thought being expressed” (Loehr 2004: 1). Despite the potential richness of corpus annotation, such a connection could not have been explored using merely transcription-based methods.3

4.

Corpora as basis for the study of topics in pragmatics

In the following, I address topics in pragmatics that have been devoted particular attention from a corpus-based perspective, namely discourse markers (with a special focus on Aijmer 2002), prosody (focusing on Wichmann’s (1994, 2000, 2004, 2005) work), listener activities/backchannels (Bublitz 1988; Tottie 1991), and corpus-based models of discourse structure (Sinclair and Coulthard 1975; Stenström 1984, 1994; Sinclair 1996, 2004).

Corpus-based pragmatics I: qualitative studies

4.1.

599

Discourse markers

It is perhaps not surprising that ‘discourse markers/particles’ or ‘pragmatic markers/particles’ are among the most focused topics in the literature, as this class lends itself easily to form-based retrieval in corpora.4 Studies include items like well, like, kind of, sort of, I mean etc. in English (Brinton 1996: 32), eh bien, donc and allors in French (Mosegaard Hansen 1998), aber, doch, and ja in German (Abraham 1991), etc. With respect to terms and inventory, Fischer (2006) provides a good overview of conceptual and terminological differences. The meanings of discourse markers are acknowledged to be “excruciatingly hard to state” (Wierzbicka 1991: 341). They make only a limited contribution to propositional content, but contribute considerably at the pragmatic level, to the organisation of discourse (for instance as markers of speaker transition or by marking boundaries in the discourse), to marking a speaker’s attitude towards the propositional content of the utterance (e.g. the Norwegian particle vel signalling reduced speaker commitment; Fretheim 1981) and to various functions pertaining to the speaker-hearer relation, such as you know indicating common ground between speaker and hearer (Jucker and Smith 1998). Among the substantial contributions are Lenk’s (1998) study of connectives and Andersen’s (2001) sociolinguistic study based on COLT and a comparable subset of the BNC. The former focuses on anyway/anyhow, however, still, incidentally, actually and what else in British and American English represented by LLC, CSAE and ICE-GB, whereas Andersen (2001) describes two discourse markers, like and innit, showing their pragmatic function, grammaticalisation and distribution (cf. Rühlemann, this volume). These two studies are theoretically based in Sperber and Wilson’s (1986/1995) relevance theory and analyse the pragmatic function of discourse markers as procedural constraints on relevance, as does Jucker (1993) in his account of well. Comparative and cross-linguistic aspects of discourse markers are covered by Aijmer and Simon-Vandenbergen (2006) and Fischer (2000), analysing German ja, also, ne, oh and ach and English yes, yeah, oh and well. Several of the studies in Jucker and Ziv (1998) and Andersen and Fretheim (2000) are also based on corpora. Furthermore, Mosegaard-Hansen (1998) presents a detailed analysis of the six French markers bon, ben, eh bien, puis, donc and alors. She bases her study on “a collection of recorded and transcribed corpora obtained from the archives of the UFR de linguistique francaise of the Université de Paris III” (Mosegaard-Hansen 1998: 201). Among the markers investigated are donc and alors, forms which both originate in temporal anaphoric expressions (deriving from the Latin forms dunc/illa hora, ‘at that time’/‘at that hour’, respectively). While donc is a monosemous marker of mutual manifestness, in the sense of Sperber and Wilson (1995), the much more common marker alors is multiply polysemous and its core use involves a shift in perspective, e.g. a new topic, marking a result or a conclusion, functions which are derived metonymically from their diachronic sources.

600

Gisle Andersen

Aijmer’s (2002) contribution is illustrative of both the advantages of the corpus-based approach, as well as some of its problems. She describes “a class of items with fuzzy boundaries” as they are used in LLC with occasional comparisons with LOB and COLT. The inventory of forms includes now signalling emphasis or transition, oh and ah and their functions as discourse particles signalling topic shift, clarification or elaboration, just as a marker of speaker involvement, affect or rhetorical use, the evidential particles (hedges) sort of and kind of, vagueness markers like and all that, and that sort of thing, etc., and actually used with evidential functions. The value of Aijmer’s contribution lies first of all in its extensive qualitative analysis of individual tokens in relevant contexts of use, as well as her firm reliance on data, in what she considers a “bottom-up approach”. She develops very specific criteria needed to distinguish discourse particles from other (often adverbial) uses of the items investigated, and focuses on core meanings and their derivative relation to the discourse functions of individual items, including their grammaticalisation. A fundamental notion in Aijmer’s account is that of “indexicality”, that is to say, the potential of discourse particles to “point to the speaker’s epistemic attitude to the utterance and affective attitude to the hearer as well as to the preceding and following discourse” (Aijmer 2002: 55). This capacity can be likened to the indexicality of deictic expressions like anaphoric pronouns, in the sense that they “point to the presence of some entity in the immediate situation-athand” (Ochs 1996: 414), including the speaker and hearer, spatial and temporal dimensions of the speech situation, epistemic stances like certainty/uncertainty, attitudes towards hearer relationship, social rank, group identity and attitudes towards the speech act or some of its characteristics. Aijmer argues convincingly that discourse particles must be construed as polysemous items in what can be seen as a ‘modified meaning minimalist’ programme, having a single core meaning and various related senses, as opposed to the proposed homonymy analysis in a “meaning maximalist” programme, in which particles are seen as having multiple meanings determined by the context of use, explicitly refuted by Östman (1995) and Mosegaard Hansen (1998). It is precisely their polysemy that can explain the grammaticalisation and the diachronic and synchronic relation between a discourse particle and its predecessor. For example, the temporal and topic transitional meanings of the discourse particle now are seen to co-occur as a result of a development from adverb to discourse particle. Aijmer performs a meticulous qualitative analysis of each individual item, and the treatment of each particle includes descriptions of a variety of features, including the pragmatic-functional range, collocational analysis, prosody, grammaticalisation and relation to the word’s core meaning, and, where relevant, comparisons with other particles, such as the comparison of now with well with their respective adverbial functions. A significant aspect of this contribution is the collocational analysis of each particle. Specific collocational characteristics are viewed as a functional clue “to assist the hearer when a particle is ambiguous between different senses” (Aijmer

Corpus-based pragmatics I: qualitative studies

601

2002: 28). To exemplify, the pragmatic value of oh is dependent on its co-occurrence with well, as the collocation oh well specifically indicates surprise and acceptance, at the same time belittling or trivialising a previous utterance (Aijmer 2002: 104). This can be contrasted with the collocation ah well, which has an upgrading effect, signalling the importance of the upcoming message. Collocations also have a potential for distinguishing between marker and non-marker (i.e. usually adverbial) use of an item, in that now in the contexts of well now, now then and now look are invariably analysed as discourse markers. However, Aijmer does not fully explore the possibility that collocations may also be markerhood indicators. It is for example notoriously difficult to draw the line between marker and non-marker use of just (Tottie 1986), in other words the restrictive adverb with a propositional meaning of ‘exactly’ or ‘only’. Aijmer devotes relatively limited attention to the collocational patterns of just (but see Erman 1997) and it seems fair to assume that the collocational analysis could have been fruitfully extended to include descriptions of how specific phrases are (perhaps exclusively) associated with marker use, while others are not. An objection that could be made towards empirical studies of discourse markers in general, is that collocational features are largely underexplored. There are good grounds for claiming, as does Aijmer herself that “[a]mong the characteristic properties of discourse particles which have not received enough attention in the literature is their ‘clustering’ (collocations with other particles)” (Aijmer 2002: 31; see also Vicher and Sankoff 1989). In studies of discourse markers, the mention of collocational features is often coincidental rather than systematic. Aijmer’s work is illustrative in this respect, as her account of just amounts to a mere listing of the top most frequent collocational patters (as in Aijmer 2002: 161), and the collocational analysis for actually is ignored altogether. Moreover, the presumably common collocation oh really is ignored in the section on collocations, albeit acknowledged as a unit with a separate status as a ‘topicaliser’ in the discussion of pragmatic functions. Aijmer gives no reason for her selection of the seven items studied. Her only reference to criteria for inclusion is frequency in LLC, but we are not told whether this frequency as discourse particles or overall frequency of the forms in question. One might well ask why a particle like well is not accounted for in her description; presumably this is because she finds this particle sufficiently covered in the abundant previous literature. Besides, the common use of like as a discourse marker is only devoted sporadic attention, illustrating a difference between LLC and subsequent English corpora. 4.2.

Prosody

The availability of spoken corpora has prompted many studies on prosody, from the early LLC-based studies using prosodic transcriptions as primary data, to studies using instrumental analysis of speech data as a basis for research. In prag-

602

Gisle Andersen

matics, many corpus-based studies fail to explore prosody systematically as part of the analysis. With respect to the literature on discourse markers outlined above, this general shortcoming is surprising, since prosody may be crucial for the determination of markerhood status, and since grammaticalisation processes tend to be accompanied by reduction in form. Aijmer’s (2002) study is exceptional in that it utilises the prosodic annotations of LLC extensively, which adds to the credibility of the analysis. An important effort in the study of intonation is the Intonational Variation in English (IViE) Corpus, containing speech data and intonation transcriptions from nine urban dialects of British English. The data represent five different speaking styles, including conversation, map task dialogues, retold stories, reading passages and sentences. Fletcher et al. (2004) use this corpus in a comparative study of the high-rising terminal (HRT) in four varieties of English. HRT refers to the use of the high-rising intonation contour – generally associated with yes/no questions in English – in declarative utterances. With respect to the function of this feature, Guy and Vonwiller suggest that “HRT correlates with the semantic complexity of the text and therefore the need for checking to see if the audience is understanding what is being said” (1989: 25). Previous studies of Australian and New Zealand English have shown that HRT is associated with low prestige varieties, is socially stigmatised, occurs most often in the telling of narratives and is used predominantly by young adolescent females as part of narratives (Allan 1990). Contrary to this, Fletcher et al. (2004) show that it is the male speakers in IViE who use HRTs the most. According to the authors, the advantage of using the IViE corpus for this study is that it “provides sufficient coverage of intonational phenomena across a very large corpus comprising several, very different, varieties of British English” (Fletcher et al. 2004: 407). Several contributions by Wichmann (1994, 2000, 2004, 2005) exemplify how a corpus-based method provides a valuable complement to previous conversation analytic work based on experimentally controlled perception studies (e.g. CouperKuhlen and Selting 1996). Much of her work is based on SEC, while more recent studies explore the British component of the ICE corpus (ICE-GB). Although corpora which “capture speech only in its natural habitat” (Wichmann 2004: 1527) often do not maintain the sound quality needed for instrumental analysis, Wichmann’s work firmly demonstrates how an auditory analysis of the data can lead to enlightening scientific results. Wichmann (2000) analyses how intonation is used to signal the structure of spoken texts and its role in turn-taking and in signalling cooperation and clarifying the relationship between utterances in conversational discourse. For example, major shifts in topic are distinguishable from units with close cohesive ties on prosodic grounds. Furthermore, Wichmann (2004) shows the advantages of using a speech corpus in a study of please-requests. She describes the pitch contour and accentual status, focusing on how the realisation varies according to the speech situation and the speech act function of utterances,

Corpus-based pragmatics I: qualitative studies

603

as well as the utterance position of please. Her study explores formulaic expressions in which please occur, notably yes please and please do, the ‘isolate’ please itself, which is used to “reinforce a more explicit request” (Wichmann 2004: 1539), as well as the “exaggerated segmental realisation [phj li::z]” (Wichmann 2004: 1540), where the epenthetic schwa leads to what is in effect a disyllabic realisation, usually expressing scorn or disapproval. The fact that her study is based on a corpus representing speakers of varying backgrounds and different contexts of use enables the author to relate her findings of prosodic patterns to the social distance and symmetrical/asymmetrical power relations between interlocutors. This provides an interpretational perspective which supplements the existing accounts of the discourse effects of terminal contours, e.g. of a falling contour as contributing ‘a feeling of closure’ (Croft 1995) and a final rise as signalling ‘openness’ (Cruttenden 1997). At the same time she redefines the role of please itself used as a face-threat mitigator in requests as “a gesture of courtesy that contextualises the accompanying request as occurring within a known set of rights and obligations” (Wichmann 2004: 1544). The importance of intonation for utterance interpretation is further corroborated in Wichmann’s (2005) study of emotional requests, which also demonstrates the value of investigating low-frequency items. The intonation of please help yourself with a high onset on please and a falling tone on self functions as a neutral invitation, while please help yourself, with a fall on please, is much more pleading, functioning as an emphatic, emotional exhortation. Collectively, Wichmann’s corpus-based work has contributed consideraby to our understanding of the relation between intonation and speaker attitude, showing that the attitudinal information carried by intonation is not an inherent feature of specific intonation contours but a result of context-based pragmatic inference. Other corpus-based discourse studies focus on the segmental aspects of prosody. Two early contributions by Stenström (1990a, 1990b) explore prosodic segmentation and pauses in monologue and dialogue. Stenström’s (1990b) category of “pauses” is taken to include so-called “verbal fillers” like hedges (sort of), softeners (you know), stallers (well) and initiators (anyway). Stenström points out that pauses play important roles in reformulations and as turntakers, turnyielders and turnholders. Based on the same corpus, Romero (1994) explores the role of pauses as content-bearing elements, showing that filled and unfilled pauses contribute to the realisation of Halliday’s (1985) theme/rheme structure of utterances. The interpersonal function of the filled pause is to indicate a new move in the conversation, which may or may not involve a speaker shift. In their textual function, filled pauses are used as the starting point of the following message. Finally, Romero (1994) argues that filled pauses also have an ideational function of signalling an element referred to elsewhere in the discourse, typically by anaphora. Though he may be right in claiming that filled pauses carry “meaning” of a structural and interactional nature, this ideational function seems dubious. Though the filled pause admittedly co-occurs with anaphoric referential expressions, it seems unjustified to

604

Gisle Andersen

claim that the pause “has to be considered a meaning-carrying particle which conveys semantic information in terms of structure and content” (Romero 1994: 502), as this seems to suggest that pauses are capable of taking on referential meanings and contributing to the propositional content of utterances. 4.3.

Listener activities and backchannels

Studies in corpus-based pragmatics have become increasingly concerned with the role of listeners in conversation. Corpus data can be effectively used to demonstrate how interactants actively signal listenership, while allowing the other speaker to continue to talk. Response signals convey comprehension and agreement or disagreement and are often also furnished with attitudinal meanings, signalling involvement of various types (positive/negative evaluation) and of various strengths, ranging from weak doubt or scepticism to enthusiastic agreement. Listenership can be expressed by auditory or visual signals. By far, the existing literature has mostly focused on the auditory side, which includes backchannels such as mm, mhm or uh-huh, attitudinally salient interjections such as wow or damn, discourse markers such as oh, yes/yeah and okay, lexical repetitions, etc. To a much lesser degree, non-verbal signals such as nods, furrowed brows and gestures have been studied (Norrick and Fischer 2009). The term backchannel first appears in the pioneering work of Yngve (1970), referring to the “existence of what I call the back channel, over which the person who has the turn receives short messages such as yes and uh-huh without relinquishing the turn” (Yngve 1970: 568). The LLC has been instrumental in providing an empirical basis for the study of listenership in discourse. Oreström (1983) provides an early LLC-based account which describes backchannels in the wider context of turn-taking which includes a rather narrow set of items (m, yes, yeah, mhm, no, quite, aha, good) but excludes tokens preceded by the markers you know and you see. Similarly, Bublitz (1988) uses a subset of the LLC (excluding meetings and interviews) in a study of “recipient action” which includes hearer activities and “certain reactive or ‘recipient’ speaker contributions” (Bublitz 1988: 141). Acknowledging a “generally prevailing tendency to agree with, assent to and support the fellow-speaker’s attitudes and statements” (Bublitz 1988: 153), Bublitz makes a distinction between primary speakers, secondary speakers and hearers. Unlike a secondary speaker, who makes a minor contribution to the topic at hand, the hearer “signals that he is taking note of what has been said and meant” and “uses particular hearer signals taken from a closed set” (Bublitz 1988: 153). Bublitz’s study is a bottom-up approach where a careful qualitative analysis of backchannels precedes generalisations about their basic functions, their high flexibility with respect to discourse placement, and their formal features (usually short and monosyllabic). The closed paradigm of forms includes yes, yeah, yep, uhuh, ɘhɘ, m, mhm, I see and really (with a falling tone). Bublitz excludes no from the category on the grounds

Corpus-based pragmatics I: qualitative studies

605

that it provides an “expression of stating the speaker’s position as to the current topic” (Bublitz 1988:181), thus its function is more than simply taking note of the previous speaker’s utterance. Tottie (1991) provides a comparison between the British and American English represented by LLC and CSAE. This contribution is important as it addresses theoretical and methodological problems in connection with corpus-based listenership research. It focuses on differences between supportive and regulative functions of backchannels, that is, the dual function of signalling understanding/agreement and encouraging the speaker to continue. Tottie points out that “[t]he term backchannel has … been extended to designate ‘backchannel message’” (Tottie 1991: 256). In fact, in its current use, the term is multiply polysemous, as it is also used to refer to forms that can realise the supportive and regulative functions of listenership signals, as well as actual instantiations (tokens) in real data (Andersen 2009). Tottie also problematises the fact that the two datasets may not necessarily be comparable in all respects, for example since “the LLC transcription records tone … in detail but does not account for lengthening, which is shown in the CSAE transcription” (Tottie 1991: 260). This illustrates that the information available to a researcher does not always make corpora fully equivalent. A more substantial difference concerns the data themselves, not the transcription conventions. Tottie suggests that, in order to draw justifiable conclusions in a comparative corpus-based study, it is necessary to control for a variety of non-linguistic factors in order to avoid that emerging differences are artefacts of biased data. She explicitly discusses factors such as the speakers’ age, gender and affinity with each other, subject matter and degree of consensus, as well as whether or not the conversations are surreptitiously recorded. Ideally, in comparative studies, the data should match on all of these points, certainly if one is to make any sense of quantitative comparisons. Tottie acknowledges that her corpus data match in some of these respects, but not all. Nevertheless, she is able to demonstrate convincingly the qualitative differences in intonation, syllabicity, articulation and length of the British and American variety of mm. Tottie raises the issue of what distinguishes a backchannel from a turn, stating that “the most important principle is that backchannel status can be determined only on the basis of the following utterance” (Tottie 1991: 260). It is problematic that Tottie links backchannelling so closely to observable speech production, without taking speaker/listener intentions into account. A backchannel such as mm may be intended as mere supportive device without the listener intending to take the floor. But it may well be that the primary speaker chooses not to continue his talk, in which case the “backchanneller” cedes the opportunity and there is a change of speaker. It is problematic to argue that this changes the functional characteristics of the backchannel item to any great extent. Therefore, a sharp division between an mm that does not involve change of speaker and one that does so does not seem justified. Tottie’s study concludes with a valuable proposal for a future research pro-

606

Gisle Andersen

gramme on backchannelling, proposing the need to investigate “whether subject matter, sex, and number of participants determine the use of backchannelling” and “how their prosodic realization is linked to discourse function” (1991: 270). Although many of her questions still remain unanswered, subsequent corpus-based research has provided important insights into this field of study. For example, McCarthy (2002; cf. also 2003) studies response tokens used in everyday conversation, focusing on differences between the British CANCODE corpus and the Cambridge-Cornell Corpus of Spoken North American English. Moreover, the conversation analytical work of Gardner (1998) argues that research-based findings on receipt tokens should have bearings on how language is taught, concluding that it would be “worth considering exposing learners to these interactional exponents, and providing opportunities for practice and feedback” (Gardner 1998: 220–221). Heintz (2003) compares the backchannel behaviour of German and American English speakers, demonstrating differences between Germans who are proficient in English and those that are not. Concerned with the use of English as a lingua franca (ELF), Bjørge’s (2009) recent investigation is valuable as it sees verbal and non-verbal aspects of backchannelling in context. She investigates video recordings of ELF interaction involving students partaking in simulated business negotiations, showing that backchannel behaviour varies according to conflict level. Some of these results are corroborated by Andersen (2009, based on a similar dataset), who considers backchannels in polylogic discourse, with a particular focus on the interactional placement of backchannels, that is, the direction of backchannelling between two teams representing a seller and buyer. Finally, Kjellmer (2009) considers the insertion point of backchannels, observing that, although most often inserted at the end of a completed utterance, backchannel insertions often occur when the utterance has not been completed. What this signifies is that “the information-content of an utterance can be conveyed to the listener before it has been expressed in toto” (Kjellmer 2009: 107). 4.4.

Models of discourse structure

Given the short history of corpus-based pragmatics, special attention should be paid to scholarly contributions which are less concerned with micro-phenomena but which focus on the theoretical basis of this research domain, addressing basic issues such as the units of discourse and their sequential organisation, the mapping of lexical, morphological and syntactical form with illocutionary function, participant relations, textual co-production, etc. Inspired by theoretical precursors like Firth (1935) and Sinclair (1966), the seminal contribution by Sinclair and Coulthard (1975) was original in its exploration of naturally occurring speech in an academic atmosphere which massively discarded the notion that studying real data might have any significance. Their choice of the classroom as the venue for collecting samples of speech was motivated by their wish to “control as many of

Corpus-based pragmatics I: qualitative studies

607

the potential variables as possible” (Sinclair and Coulthard 1975: 6) and overcome the many problems inherent in natural conversation and opt for “a more simple type of spoken discourse, one which has a much more overt structure, where one participant has acknowledged responsibility for the direction of the discourse, for deciding who shall speak when, and for introducing and ending topics” (Sinclair and Coulthard 1975: 6). Based on a restricted corpus of transcribed speech data, they develop an analytical apparatus which, despite subsequent revisions (e.g. Sinclair and Brazil 1982; Sinclair and Coulthard 1992) and criticism (Stubbs 1984), still remains influential and has been inspirational for a wealth of research. They propose a discourse model of exchange and transaction structure which shows that spoken discourse can be systematically described according to its structure, just as syntax can be. The model, which is ‘closely modelled on Halliday’s Categories of a Theory of Grammar’ (sic.; Sinclair and Coulthard 1975: 24; cf. Halliday 1961), encompasses a hierarchy of discourse components, specified as a ranking scale containing the four components transaction, exchange, move and act. In its original version, the model also included a top-level of ‘lesson’ which has in later revisions of the theory been dismissed as a stylistic type specific to the usage domain of classroom discourse. At each level, discourse units are divided into classes which realise an element of a structure, and the internal structure of the unit is described. Exchanges have a three-part structure containing the elements initiation, response, feedback (IRF-structure), with the corresponding classes of move: opening, answering and follow-up (Sinclair and Coulthard 1975: 26). Despite later criticism for example concerning the unnatural nature of classroom discourse (in which teachers ask pseudo-questions) upon which the model is based, this contribution effectively involves no less than the inception of a research field, discourse analysis. A decade later, Stenström (1984) develops this model further in a comprehensive analysis of question/response strategies. Her contribution is valuable in that it uses LLC as a test bed for the discourse model in an analysis of authentic conversational data. She adds empirical support to the theory by accounting for “how the different levels defined in the model of analysis were manifested in the corpus” (Stenström 1984: 94). As its title suggests, the primary focus is on exchanges. Although there is a bias towards the analysis of responses, as opposed to the initiation and feedback moves, Stenström provides a detailed account which represents a much wider user domain than the original empirical basis for the model, ‘selected to represent the greatest possible variety of speech situations’ (Stenström 1984: 262), ranging from informal conversation to ‘procedural dialogue’ (planned, public discourse; e.g. courtroom interaction). She extends the model in various ways by introducing notions such as topic-bound transactions (anaphorically related to the previous context) versus proposition-bound transactions (forward-oriented) and by specifying optional and obligatory moves in exchange sequences. Her study incorporates categories like hesitation, pauses, hedges, reformulations and the like,

608

Gisle Andersen

and shows how they are linguistically realised in a corpus which, for its time, was both extensive and varied. The “Stenströmian model” (Kirk 2002) is further developed into what the author herself refers to as “a synthesis of much of the work that has been done so far in the analysis of conversation” (Stenström 1994: xii). Kirk (2002) provides a good illustration of how the model can be applied in the analysis of real corpus data, arguing that “although there are many competing analytic models available (e.g. Edwards and Lampert 1993; Eggins 1996; Langford 1994; etc.), none is as rigorous or as syntactically systematic as that of Anna-Brita Stenström” (Kirk 2002: 144). Despite Kirk’s praise, the model has raised criticism as regards the conceptual nature of discourse units and their relation to each other, and to the hierarchy of units, “theoretical difficulties [which] are of course problems of the subject in general, and are not specific to the book” (Knowles 1995: 122). The relation of the turn (not part of Sinclair and Coulthard’s model) to the rest of the hierarchy is particularly problematic, as an exchange boundary may come in the middle of a turn, and since a turn boundary may come in the middle of a move, for instance if a speaker finishes off someone else’s word in joint utterance production. In his more recent work, Sinclair has introduced the notion of “extended units of meaning”, first in the seminal paper Sinclair (1996), and more comprehensively in Sinclair (2004). He presents the hypothesis that “the notion of a linguistic item can be extended … so that units of meaning are expected to be largely phrasal” (Sinclair 1996: 82). Although the fixedness of word strings is most obvious in idiomatic/proverbial usage and phrasal verbs, Sinclair’s model extends the concept of phrasehood to words in general. Corpus evidence shows that each word in the language has its own unique lexico-grammatical profile and that different senses of for instance a noun such as bank have their own distinctive sets of collocates. Sinclair argues that these observations force the linguist to reassess the notion of meaning. His model includes structural categories with an increasing level of abstraction: collocation (a frequent co-occurrence of words, such as his example naked eye), colligation (the co-occurrence of grammatical choices, e.g. choice of preposition with/to in front of the naked eye), semantic preference (“the collocation of a lexical item with items from a specific (more or less general) semantic subset” Bednarek 2008: 119) and semantic prosody (persistent collocation with items with specific connotations; Louw 1993; cf. also Hunston 2009). Despite the term – and importantly for the current study – Sinclair views semantic prosodies as attitudinally significant and places them “on the pragmatic side of the semantics/ pragmatics continuum” (Sinclair 1996: 87). Indeed, speakers’ choice of a semantic prosody is what links meaning to communicative purpose: “Without it, the string of words just means – it is not put to use in a viable communication … all subsequent choices within the lexical item relate back to the prosody” (Sinclair 1996: 88). This can explain the pragmatic difference between semantically equivalent sentences like I will make no promises and I will not make any promises. Hence, Sinclair’s analysis has an explanatory power to account for the creation of new

Corpus-based pragmatics I: qualitative studies

609

pragmatic expressions like make no mistake, which turns out strikingly more common than any other collocation of make no in a large corpus such as COCA. Sinclair (2004) develops the analytical model further by specifying the core (the invariable component of a lexical item) and the semantic prosody as the two obligatory components of a lexical item, and the remaining three (collocation, colligation and semantic preference) as optional components, by relating the categories of co-selection to the paradigmatic and syntagmatic dimensions of language. Furthermore, Stubbs (1996, 2001) is another linguist who argues that meaning is a largely a phraseological phenomenon and whose corpus-based accounts also contribute to the rigour and maturity of the field of corpus-based discourse analysis. Like the previous models, he draws on the British neo-Firthian tradition by Halliday and Sinclair, but this account is more sociologically oriented than the ones referred to above. It takes into account the status of speakers and listeners and the importance of social institutions for the interpretation of the meaning of texts. Stubbs’s linguistic analysis, largely based on collocation patterns, focuses on how users’ lexis and grammatical choices (e.g. active vs. passive voice) contribute to different meanings that can be interpreted in a social context. The rapidly growing interest in corpus-based pragmatics is underlined by several collective volumes which for reasons of space have to be mentioned in passing. Adolphs (2008), Ädel and Reppen (2008) and Biber et al. (2007) are significant contributions that are exclusively or predominantly corpus-based and which address the main theoretical and methodological tenets of this field, while Baker (2006) provides an enlightening evaluation of a variety of methodologies for corpus-based discourse analysis.

5.

Concluding remarks

The overview presented above has demonstrated the breadth and complexity of the field of corpus-based pragmatics. Most studies consider overt and easily retrievable features such as discourse markers, connectives, hedges etc., and are formbased (cf. section 1). There are fewer studies where corpora are used to explore a particular discourse function and its possible realisations, although such studies appear to be on the increase. Relevant examples would be Deutschmann’s (2006) study which explores particular speech act function – apologies – and its various realisations, and Stenström et al. (2002) who demonstrate how quotations are marked by different structures, including BE + like, GO and SAY as reporting verbs, voice modulation and zero-quotations. There appears to be an ongoing increase in the different topics that are explored by corpora. Schmid (2001) studies various patterns for expressing presuppositions. Corpora are also used in the study of tropes such as metaphor (Sandford 2008) and

610

Gisle Andersen

hyperbole in recent contributions such as Norrick (2004) and McCarthy and Carter (2004b), who argue that this approach “can add substance to the already healthy existing literature on rhetorical strategy, conventional and creative metaphor, irony and over- and understatement.” (2004b: 158). These contributions illustrate the problem associated with one-to-one searching, as it assumes an a priori understanding of the forms that may realise a particular discourse function. This is especially the case in studies of creative language processes such as the use of tropes. I have argued in favour of supplementing corpus-based pragmatics studies with a corpus-driven approach. Corpus-based pragmatics studies require a hermeneutical and holistic approach to language data, because face-to-face interaction involves multiple layers of information. Video-based research methods are therefore valuable. Modern multimodal corpora developed with state-of-the-art technology provide a promising outlook for such studies in the future.

Notes 1. I am grateful to the editors and the anonymous reviewer for valuable comments on the first draft of this paper. 2. This can for example be seen from the studies included in three recent special issues of Journal of Pragmatics focusing on corpus linguistics (e.g. McCarthy and Carter 2004a), of which a vast majority is based on English. 3. Another shortcoming of some corpora is that the data are only available as short snippets of sound and not as longer stretches of spoken dialogue. For many research purposes, access to longer text passages is a prerequisite, e.g. studies of the structure and development of narratives in conversation. 4. In the current account, I opt for the term “discourse markers”, which seems the most commonly used, especially in the most recent literature. For discussions on terminology, cf. Brinton (1996); Andersen (2001), Fischer (2006) and Mosegaard Hansen (2006), who argues that “discourse particle” is misleading on formal grounds.

References Aarts, Jan and Willem Meijs (eds.) 1986 Corpus Linguistics II. New Studies in the Analysis and Exploitation of Computer Corpora. Amsterdam: Rodopi. Abraham, Werner (ed.) 1991 Discourse Particles. Amsterdam: Benjamins. Ädel, Annelie and Randi Reppen (eds.) 2008 Corpora and Discourse: The Challenges of Different Settings. Amsterdam: Benjamins. Adolphs, Svenja 2008 Corpus and Context: Investigating Pragmatic Functions in Spoken Discourse. Amsterdam: Benjamins.

Corpus-based pragmatics I: qualitative studies

611

Aijmer, Karin and Anne-Marie Simon-Vandenbergen (eds.) 2006 Pragmatic Markers in Contrast. Amsterdam: Elsevier. Aijmer, Karin 1986 Why is actually so frequent in spoken English? In: Gunnel Tottie and Ingegerd Bäcklund (eds.), English in Speech and Writing: A Symposium, 119–129. Stockholm: Almqvist and Wiksell. Aijmer, Karin 1989 Themes and tails: The discourse functions of dislocated elements. Nordic Journal of Linguistics 12: 137–154. Aijmer, Karin 1996 Conversational Routines in English. London: Longman. Aijmer, Karin 2002 English Discourse Particles: Evidence from a Corpus. Amsterdam: Benjamins. Algeo, John 1988 British and American grammatical differences. International Journal of Lexicography 1: 1–31. Allan, W. Scott 1990 The rise of New Zealand intonation. In: Allan Bell and Janet Holmes (eds.), New Zealand Ways of Speaking English, 115–128. Clevendon: Multilingual Matters. Altenberg, Bengt and Karin Aijmer 2000 The English-Swedish Parallel Corpus: a resource for contrastive research and translation studies. In: Christian Mair and Marianne Hundt (eds.), Corpus Linguistics and Linguistic Theory, 15–34. Amsterdam: Rodopi. Altenberg, Bengt 1986 ICAME bibliography. ICAME News 10: 62–79. Altenberg, Bengt 1987 Prosodic patterns in spoken English. Studies in the Correlation between Prosody and Grammar for Text-To-Speech Conversion. Lund: Lund University Press. Altenberg, Bengt 1991 Amplifier collocations in spoken English. In: Stig Johansson and Anna-Brita Stenström (eds.), English Computer Corpora, 127–147. Berlin: Mouton de Gruyter. Altenberg, Bengt 1997 Exploring the Swedish component of the International Corpus of Learner English. In: Barbara Lewandowska-Tomaszcyk and Patrick James Melia (eds.), PALC’97: Practical Applications in Language Corpora, 119–132. Lódz: Lódz University Press. Altenberg, Bengt 2006 The function of adverbial connectors in second initial position in English and Swedish. In: Karin Aijmer and Anne-Marie Simon-Vandenbergen (eds.), Pragmatic Markers in Contrast, 11–38. Amsterdam: Elsevier. Andersen, Gisle and Thorstein Fretheim 2000 Pragmatic Markers and Propositional Attitude. Amsterdam: Benjamins.

612

Gisle Andersen

Andersen, Gisle 2001 Pragmatic Markers and Sociolinguistic Variation: a Relevance-Theoretic Approach to the Language of Adolescents. Amsterdam: Benjamins. Andersen, Gisle 2005 Assessing algorithms for automatic extraction of anglicisms in Norwegian texts. In: Proceedings from The Corpus Linguistics Conference Series, Vol. 1, no. 1, Corpus Linguistics 2005. Andersen, Gisle 2006 Engelske importord i norsk. Språknytt 3/2006: 27–32. Andersen, Gisle 2009 Responses in academic and professional discourse. Paper presented at the 11th International Pragmatics Conference, Melbourne, 12–17 July 2009. Andersen, Gisle 2010 How to use corpus linguistics in sociolinguistics. To be published in Anne O’Keeffe and Michael McCarthy (eds.), Routledge Handbook of Corpus Linguistics, 547–562. Londin: Routledge. Aston, Guy and Lou Burnard 1998 The BNC Handbook. Exploring the British National Corpus with SARA. Edinburgh: Edinburgh University Press. Baker, Paul 2006 Using Corpora in Discourse Analysis. London: Continuum. Bald, Wolf-Dietrich 1990 An example of phonological reduction in English. In: S. Ramsaran (ed.), Studies in the Pronunciation of English. A Commemorative Volume in Honour of A.C. Gimson, 317–322. London: Routledge. Banjo, Ayo 1996 The sociolinguistics of English in Nigeria and the ICE project. In: Sidney Greenbaum (ed.), Comparing English Worldwide: The International Corpus of English, 239–248. Oxford: Clarendon Press. Bauer, Laurie and Janet Holmes 1996 Getting into a flap! /t/ in New Zealand English. World Englishes 15: 115–124. Beaugrande, Robert de 1999 Linguistics, sociolinguistics, and corpus lingistics: Ideal language versus real language. Journal of Sociolinguistics, 3: 128–139. Bednarek, Monika 2008 Semantic preference and semantic prosody re-examined. Corpus Linguistics and Lingustic Theory 4: 119–139. Biber, Douglas (ed.) 1994 Sociolinguistic Perspectives on Register. Oxford/New York: Oxford University Press. Biber, Douglas and Edward Finegan 1989 Styles of stance in English. Lexical and grammatical marking of evidentiality and affect. Text 9: 93–124. Biber, Douglas, Ulla Connor and Thomas A. Upton 2007 Discourse on the Move: Using Corpus Analysis to Describe Discourse Structure. Amsterdam: Benjamins.

Corpus-based pragmatics I: qualitative studies

613

Biber, Douglas 1986 Spoken and written textual dimensions in English: Resolving the contradictory findings. Language 62: 384–414. Biber, Douglas 1987 A textual comparison of British and American writing. American Speech 62: 99–119. Biber, Douglas 1988 Variation across Speech and Writing. Cambridge: Cambridge University Press. Biber, Douglas 1993 An analytical framework for register studies. In: Douglas Biber and Edward Finegan (eds.), Sociolinguistic Perspectives on Register, 31–56. New York: Oxford University Press. Biber, Douglas 1995 Dimensions of Register Variation. A Cross-Linguistic Comparison. Cambridge: Cambridge University Press. Biber, Douglas 2009 A corpus-driven approach to formulaic language in English: Multi-word patterns in speech and writing. International Journal of Corpus Linguistics 14.3: 275–311. Bjørge, Anne-Kari 2009 Conflict or cooperation: the use of backchannelling in ELF negotiations. English for Specific Purposes 29: 191–203. Brazil, David 1982 Impromptuness and intonation. In: Nils Erik Enkvist (ed.), Impromptu Speech: a Symposium, 277–290. Åbo: Åbo Akademi. Brinton, Laurel J. 1996 Pragmatic Markers in English: Grammaticalization and Discourse Functions. Berlin: Mouton de Gruyter. Brodda, Benny 1988 Tracing turns in the London-Lund Corpus with BetaText. Literary and Linguistic Computing 3: 94–104. Bublitz, Wolfram 1979 Tag questions, transformational grammar, and pragmatics. Papers and Studies in Contrastive Linguistics 9: 5–21. Bublitz, Wolfram 1988 Supportive Fellow-Speakers and Cooperative Conversations. Amsterdam: Benjamins. Chafe, Wallace L 1992 The importance of corpus linguistics to understanding the nature of language. In: Jan Svartvik (ed.), Directions in Corpus Linguistics. Proceedings of Nobel Symposium 82, 79–97. Berlin: Mouton de Gruyter. Chafe, Wallace L., John W. Du Bois and Sandra A. Thompson 1991 Towards a new corpus of Spoken American English. In: Karin Aijmer and Bengt Altenberg (eds.), English Corpus Linguistics. Studies in Honour of Jan Svartvik, 64–82. London: Longman.

614

Gisle Andersen

Chafe, Wallace 1992 The importance of corpus linguistics to understanding the nature of language. In: Jan Svartvik (ed.), Directions in Corpus Linguistics. Proceedings of Nobel Symposium 82 Stockholm, 4–8 August 1991, 79–97. Berlin: Mouton de Gruyter. Chomsky, Noam 1957 Syntactic Structures. The Hague: Mouton. Coates, Jennifer 1987 Epistemic modality and spoken discourse. Transactions of the Philological Society 85(1): 110–131. Collins, Peter 2005 It-clefts and wh-clefts: Prosody and pragmatics. Journal of Pragmatics: 38: 1706–1720. Cook, Guy 1990 ‘Transcribing Infinity: Problems of Context Presentation’. Journal of Pragmatics 14: 1–24. Couper-Kuhlen, Elizabeth and Cecilia E. Ford (eds.) 2004 Sound Patterns in Interaction: Cross-linguistic Studies from Conversation. Amsterdam: Benjamins. Couper-Kuhlen, Elizabeth and Margret Selting (eds.) 1996 Prosody in Conversation. Cambridge: Cambridge University Press. Couper-Kuhlen, Elizabeth 1996 The prosody of repetition: on quoting and mimickry. In: Elizabeth CouperKuhlen and Margret Selting (eds.), Prosody in Conversation: Interactional Studies, 366–405. Cambridge: Cambridge University Press. Cresti, Emanuela and Massimo Moneglia (eds.) 2005 C-ORAL-ROM: Integrated Reference Corpora for Spoken Romance Languages. Amsterdam: Benjamins. Croft, William 1995 Intonation units and grammatical structure. Linguistics 33: 839–882. Crowdy, Steve 1995 The BNC spoken corpus. In: Geoffrey Leech, Greg Myers and Jenny Thomas (eds.), Spoken English on Computer. Transcription, Mark-up and Application. 224–234. London: Longman. Cruttenden, Alan 1997 Intonation. Cambridge: Cambridge University Press. Crystal, David and Randolph Quirk 1964 Systems of Prosodic and Paralinguistic Features in English. The Hague: Mouton. Crystal, David 1972 The intonation system of English. In: Dwight Bolinger (ed.), Intonation, 110–136. Harmondsworth: Penguin. Crystal, David 1975 The English Tone of Voice. Essays on Intonation, Prosody and Paralanguage. London: Arnold. Davidson, Benjamin D. 1977 Aspects of information structuring in modern spoken English. In: Wolf-Diedrich Bald and Robert Ilson (eds.), Studies in English Usage: The Resources of

Corpus-based pragmatics I: qualitative studies

615

Present-day English Corpus for Linguistic Analysis, 157–170. Frankfurt am Main: Peter Lang. Davies, Mark 2009 The 385+ Million Word Corpus of Contemporary American English (1990– 2008+): Design, Architecture, and Linguistic Insights. International Journal of Corpus Linguistics 14: 159–90. Deutschmann, Mats 2006 Social variation in the use of apology formulae in the British National Corpus. In: Antoinette Renouf and Andrew Kehoe (eds.), The Changing Face of Corpus Linguistics, 205–222. Amsterdam/New York: Rodopi. Downing, Angela 2006 The English pragmatic marker surely and its functional counterparts in Spanish. In: Karin Aijmer and Anne-Marie Simon-Vandenbergen (eds.), Pragmatic Markers in Contrast, 39–58. Amsterdam: Elsevier. Edwards, Jane A. and Martin D. Lampert 1993 Talking Data: Transcription and Coding in Discourse Research. Hillsdale: Lawrence Erlbaum. Eggins, Suzanne 1996 Analysing Casual Conversation. London: Continuum. Enkvist, Nils Erik (ed.) 1982 Impromptu Speech: a Symposium. Åbo: Åbo Akademi. Enkvist, Nils Erik 1973 Linguistic stylistics. Janua Linguanun Series Critica 5. The Hague: Mouton. Erman, Britt 1986 Some pragmatic expressions in English conversation. In: Gunnel Tottie and Ingegerd Bäcklund (eds.), English in Speech and Writing: a Symposium, 131– 147. Stockholm: Almqvist and Wiksell. Erman, Britt 1987 Pragmatic Expressions in English: a Study of You Know, You See, and I Mean in Face-To-Face Conversation. Stockholm: Almqvist and Wiksell. Erman, Britt 1997 ‘Guy’s just such a dickhead’; the context and function of just in teenage talk. In: Ulla-Brite Kostsinas, Anna-Brita Stenström and Anna-Malin Karlsson (eds.), Ungdomsspråk i Norden, 82–95. Stockholm: Institutionen för nordiska språk. Faber, Dorrit and Karen M. Lauridsen 1991 The compilation of a Danish-English-French corpus in contract law. In: Stig Johansson and Anna-Brita Stenström (eds.), English Computer Corpora. Selected Papers and Research Guide, 235–243. Berlin: Mouton de Gruyter. Faerch, Claus and Gabriele Kasper 1982 Phatic, metalingual and metacommunicative functions in discourse. Gambits and repairs. In: Nils Erik Enkvist (ed.), Impromptu Speech: a Symposium, 71–104. Åbo: Åbo Akademi. Fang, Alex C. 1993 Building a corpus of computer science. In: Jan Aarts, Pieter de Haan and Nelleke Oostdijk (eds.), English Language Corpora: Design, Analysis and Exploitation. 73–78. Amsterdam: Rodopi.

616

Gisle Andersen

Fill, Alwin 1992 Joking in English and German: A contrastive study. In: Christian Mair and Manfred Marcus (eds.), New Departures in Contrastive Linguistics. Vol. 2, 21–31. Innsbruck: Institut für Anglistik, Universität Innsbruck. Firth, John R 1935 The technique of semantics. Reprinted in 1957 in Papers in Linguistics, 1934–51,7–23. Oxford: Oxford University Press. Fischer, Kerstin 2000 From Cognitive Semantics to Lexical Pragmatics: The Functional Polysemy of Discourse Particles. Berlin: Mouton de Gruyter. Fischer, Kerstin 2006 Approaches to Discourse Particles. Amsterdam: Elsevier. Fletcher, Janet, Esther Grabe and Paul Warren 2004 Intonational variation in four dialects of English: the high rising tune. In:, Sun-Ah Jun (ed.) Prosodic Typology. The Phonology of Intonation and Phrasing, 390–410. Oxford: Oxford University Press. Foolen, Ad 2006 Polysemy Patterns in Contrast: The Case of Dutch toch and German doch. In: Karin Aijmer and Anne-Marie Simon-Vandenbergen (eds.), Pragmatic Markers in Contrast, 59–72. Amsterdam: Elsevier. Francis, W. Nelson 1967 The Brown University Standard Corpus of English: Some implications for TESOL. In: Betty Wallace Robinett (ed.), On Teaching English to Speakers of Other Languages, 131–35. Washington, D.C.: TESOL. Fretheim, Thorstein 1981 Ego-dempere og alter-dempere. Maal og minne 1–2: 86–100. Furiasso, Cristiano and Knut Hofland 2007 The retrieval of false anglicisms in newspaper texts. In: Roberta Facchinetti (ed.), Corpus Linguistics 25 Years on, 347–364. Amsterdam/New York: Rodopi. Gardner, Rod 1998 Between speaking and listening: the vocalisation of understandings. Applied Linguistics 19.2: 204–224. Garnham, Alan 1986 Review of R. Wardhaugh, ‘How Conversation Works’. Journal of Literary Semantics 15: 153–154. Garnham, Alan, Richard Shillcock, Gordon Brown, Andrew Mill and Anne Cutler 1981 Slips of the tongue in the London-Lund Corpus of spontaneous conversation. Linguistics 19: 805–817. Geluykens, Ronald 1986 Questioning intonation: An empirical study into the prosodic feature ‘rising intonation’ and its relevance for the production and recognition of questions. Antwerp Papers in Linguistics 48. Geluykens, Ronald 1987a Intonation and speech act type: An experimental approach to rising intonation in declaratives. Journal of Pragmatics 11: 483–494.

Corpus-based pragmatics I: qualitative studies

617

Geluykens, Ronald 1987b Tails (right-dislocations) as a repair mechanism in English conversation. In: Jan Nuyts and Georges de Schutter (eds.), Getting One’s Words into Line: On Word Order and Functional Grammar, 119–129. Dordrecht: Fortis. Geluykens, Ronald 1988 On the myth of rising intonation in polar questions. Journal of Pragmatics 12: 467–485. Granger, Sylviane 1997 On identifying the syntactic and discourse features of participle clauses in academic English: native and non-native writers compared. In: Jan Aarts, Inge de Mönnink and Herman C. Wekker (eds.), Studies in English Language and Teaching in Honour of Flor Aarts, 185–198. Amsterdam/Atlanta: Rodopi. Granger, Sylviane and Stephanie Tyson 1996 Connector usage in the English essay writing of native and non-native EFL speakers of English. World Englishes 15: 17–27. Greenbaum, Sidney (ed.) 1996 Comparing English Worldwide: The International Corpus of English. Oxford: Clarendon Press. Greenbaum, Sidney 1991 ICE: the International Corpus of English. English Today 28: 3–7. Greenbaum, Sidney 1992 A new corpus of English: ICE. In: Jan Svartvik (ed.), Directions in Corpus Linguistics. Proceedings of Nobel Symposium 82, 171–179. Berlin: Mouton de Gruyter. Gustafsson, Marita 1982 Textual aspects of topicalization in a corpus of English. ICAME News 6: 46–76. Gustafsson, Marita 1983 Fronting of adverbials in four genres of English. In: Sven Jacobson (eds.), Papers from the Second Scandinavian Symposium on Syntactic Variation, 7–17. Stockholm: Almqvist and Wiksell. Gustafsson, Marita 1985 Topicalizations revisited. In: Jorma Tommola and T. Vinanen (eds.), Working Papers in English Studies 6, 43–50. Publications of the Department of English, Turku: University of Turku. Guy, Gregory and Julia Vonwiller 1989 The meaning of an intonation in Australian English. Australian Journal of Linguistics 4: 1–17. Haan, Ferdinand de 2002 Strong modality and negation in Russian. In: Randi Reppen, Susan M. Fitzmaurice and Douglas Biber (eds.), Using Corpora to Explore Linguistic Variation, 91–110. Amsterdam: Benjamins. Haegeman, Lillian 1984 Pragmatic conditionals in English. Folio Linguistica 13: 485–502. Halliday, Michael A. K. 1961 Categories of the theory of grammar. Word 17: 241–92.

618

Gisle Andersen

Halliday, Michael A.K. 1985 An Introduction to Functional Grammar. London: Edward Arnold. Hardy, Donald E. and Amy Leuchtmann 1996 Topic versus cohesion in the prediction of causal ordering in English conversation. Discourse Processes 21: 237–254. Haslerud, Vibeke and Anna-Brita Stenström 1995 The Bergen Corpus of London Teenager Language (COLT). In: Geoffrey Leech, Greg Myers and Jenny Thomas (eds.), Spoken English on Computer. Transcription, Mark-up and Application, 235–242. London: Longman. Hasselgård, Hilde 2006 “Not now” – on Non-Correspondence between the Cognate Adverbs now and nå. In: Karin Aijmer and Anne-Marie Simon-Vandenbergen (eds.), Pragmatic Markers in Contrast, 93–114. Amsterdam: Elsevier. Hedström, Karl Erik 1984 A study of repairs in speech. In: Stockholm Papers in English Language and Literature 4, 69–101. Stockholm: Department of English, Stockholm University. Heinz, Bettina 2003 Backchannel responses as strategic responses in bilingual speakers’ conversations. Journal of Pragmatics 35: 1113–1142. Hofland, Knut and Stig Johansson 1986 The Tagged LOB Corpus: KWIC Concordance. Microfiche. Bergen: Norwegian Computing Centre for the Humanities. Holmes, Janet 1988 Doubt and certainty in ESL textbooks. Applied Linguistics 9: 21–44. Hundt, Marianne, Carolin Biewer and Nadja Nesselhauf 2007 Corpus Linguistics and the Web. Amsterdam: Rodopi. Hunston, Susan 2009 Semantic prosody revisited. In: Rosamund Moon (ed.), Words, Grammar, Text – Revisiting the Work of John Sinclair, 85–103. Amsterdam: Benjamins. Johannessen, Janne B. and Kristin Hagen (eds.) 2008 Språk i Oslo: Ny forskning omkring talespråk. Oslo: Novus. Johansson, Stig and Knut Hofland 1994 Towards an English-Norwegian parallel corpus. In: Udo Fries, Gunnel Tottie and Peter Schneider (eds.), Creating and Using English Language Corpora, 25–37. Amsterdam: Rodopi. Johansson, Stig 2006 How well can well be translated? On the English discourse particle well and its correspondences in Norwegian and German. In: Karin Aijmer and Anne-Marie Simon-Vandenbergen (eds.), Pragmatic Markers in Contrast, 115–138. Amsterdam: Elsevier. Jucker, Andreas and Sara Smith 1998 And people just you know like “Wow”: discourse markers as negotiating strategies. In: Andreas Jucker and Yael Ziv (eds.), Discourse Markers: Descriptions and Theory, 171–201. Amsterdam: Benjamins. Jucker, Andreas and Yael Ziv (eds.) 1998 Discourse Markers: Descriptions and Theory. Amsterdam: Benjamins.

Corpus-based pragmatics I: qualitative studies

619

Jucker, Andreas 1993 The discourse marker well: a relevance-theoretical account. Journal of Pragmatics 19.5: 435–52. Kendon, Adam 1972 Some relationships between body motion and speech: An analysis of an example. In: Aron Siegman and Benjamin Pope (eds.), Studies in Dyadic Communication, 177–216. New York: Pergamon Press. Kendon, Adam 1980 Gesticulation and speech: Two aspects of the process of utterance. In: Mary Ritchie Key (ed.), The Relationship of Verbal and Nonverbal Communication, 207–227. The Hague: Mouton. Kennedy, Graeme 1998 An Introduction to Corpus Linguistics. London: Addison Wesley Longman. Kilgarriff, Adam and Grefenstette, Gregory 2003 Introduction to the Special Issue on Web as Corpus. Computational Linguistics 29: 1–15. Kirk, John 2002 Applications of the Stenström model of discourse structure. In: Angela Hasselgren and Leiv Egil Breivik (eds.), From the COLT’s Mouth … and Others’, 141–166. Amsterdam: Rodopi. Kjellmer, Göran 2009 Where do we backchannel? On the use of mm, mhm, uh huh and such like. International Journal of Corpus Linguistics 14.1: 81–112. Knowles, Gerry and Lita Lawrence 1987 Automatic intonation assignment. In: Roger Garside, Geoffrey Leech and Geoffrey Sampson (eds.), The Computational Analysis of English. A Corpusbased Approach, 139–148. London: Longman. Knowles, Gerry, Briony Williams and Lita Taylor (eds.) 1996 A Corpus of Formal British English Speech. The Lancaster/IBM Spoken-English Corpus. London: Longman. Knowles, Gerry, Lita Taylor and Briony Williams 1992 A Corpus of Formal British English Speech. London: Longman. Knowles, Gerry 1995 Review of Stenström 1994. ICAME Journal 19: 120–123. Knowles, Gerry 1986 Tone grouping by numbers. In: Lancaster Papers in Linguistics 44. Lancaster: Department of Linguistics and Modern English Language, University of Lancaster. Kussmaul, Paul 1978 In fact, actually, anyway …: Indikatoren von Sprechakten im informellen gesprochenen Englisch. Die Neueren Sprachen 27: 357–369. Kytö, Merja and Matti Rissanen 1988 The Helsinki Corpus of English Texts: Classifying and coding the diachronic part. In: Merja Kytö, Ossi Ihalainen and Matti Rissanen (eds.), Corpus Linguistics Hard and Soft, 169–179. Amsterdam: Rodopi. Langford, David 1994 Analysing Talk. London: Macmillan.

620

Gisle Andersen

Leech, Geoffre 1992 Corpora and theories of linguistic performance. In: Jan Svartvik (ed.), Directions in Corpus Linguistics. Proceedings of Nobel Symposium 82, 105–122. Berlin: Mouton de Gruyter. Leech, Geoffrey, Greg Myers and Jenny Thomas (eds.) 1995 Spoken English on Computer. London: Longman. Lenk, Uta 1998 Marking Discourse Coherence: Functions of Discourse Markers in Spoken English. Tübingen: Narr. Lindquist, Hans and Christian Mair (eds.) 2004 Corpus Approaches to Grammaticalization in English. Amsterdam: John Benjamins. Loehr, Daniel P. 2004 Gesture and Intonation. PhD. Dissertation, Georgetown University, Washington, DC. Loman, Bengt 1982 The segmentation problem in the study of impromptu speech. In: Nils Erik Enkvist (ed.), Impromptu Speech: a Symposium,105–130. Åbo: Åbo Akademi. Louw, Bill 1993 Irony in the text or insincerity in the writer? The diagnostic potential of semantic prosodies. In: Mona Baker, Gill Francis and Elena Tognini-Bonelli (eds.), Text and Technology. In Honour of John Sinclair, 157–176. Amsterdam: Benjamins. Mauranen, Anna 2004 They’re a Little Bit Different … Observations on Hedges in Academic Talk. In: Karin Aijmer and Anna-Brita Stenström (eds.), Discourse Patterns in Spoken and Written Corpora, 173–197. Amsterdam: Benjamins. Mauro, Tullio de, Federico Mancini, Massimo Vedovelli and Miriam Voghera 1993 Lessico di Frequenza dell’Italiano Parlato. Milan: Etas. McCarthy, Michael and Ronald Carte 2004a Introduction. Journal of Pragmatics 36: 147–148. McCarthy, Michael and Ronald Carter 2004b “There’s millions of them”: hyperbole in everyday conversation. Journal of Pragmatics 36: 149–184. McCarthy, Michael 2002 Good listenership made plain: British and American non-minimal response tokens in everyday conversation. In: Randi Reppen, Susan M. Fitzmaurice and Douglas Biber (eds.), Using Corpora to Explore Linguistic Variation, 73–90. Amsterdam: Benjamins. McCarthy, Michael 2003 Talking Back: “Small” Interactional Response Tokens in Everyday Conversation. Research on Language and Social Interaction 36.1: 33–63. McEnery, Tony and Andrew Wilson 2001 Corpus Linguistics. 2nd ed. Edinburgh: Edinburgh University Press. McEnery, Tony and Richard Xiao 2004 The Lancaster Corpus of Mandarin Chinese: A Corpus for Monolingual and Contrastive Language Study. Paper presented at LREC 2004. May 2004, Lisbon.

Corpus-based pragmatics I: qualitative studies

621

McEnery, Tony, Richard Xiao and Yukio Tonio 2006 Corpus-based Language Studies: An Advanced Resource Book. New York: Routledge. McNeill, David 1985 So you think gestures are nonverbal? Psychological Review 92.3: 350–371. McNeill, David 1992 Hand and Mind: What Gestures Reveal about Thought. Chicago: University of Chicago Press. Mey, Jacob L. 2004 FoI Corpus Linguistics Part III – Introduction. Journal of Pragmatics 36: Editorial. 1519. Milroy, Leslie 1987 Observing and Analysing Natural Language. Oxford: Blackwell. Ming, Tao and Liang Chen 2010 A discourse-pragmatic study of the word order variation in Chinese relative clauses. Journal of Pragmatics 42: 168–189. Mosegaard Hansen, Maj-Britt 1998 The Function of Discourse Particles. Amsterdam: Benjamins. Mosegaard Hansen, Maj-Britt 2006 A dynamic polysemy approach to the lexical semantics of discourse markers. In: Kerstin Fischer (ed.), Approaches to Discourse Particles, 21–41. Oxford: Elsevier. Nässlin, Siv 1984 The English Tag Question: A Study of Sentences Containing Tags of the Type ‘Isn’t It?’, ‘Is It?’. Stockholm: University of Stockholm. Nevalainen, Terttu 1987a A quantitative survey of nuclear tone variation in English. In: Proceedings of the Eleventh International Congress of Phonetic Sciences. Vol 3, 180–183. Tallin: Institute of Language and Literature, Academy of Sciences of the Estonian S.S.R. Nevalainen, Terttu 1987b Adverbial focusing and intonation. Lingua 73: 141–165. Nevalainen, Terttu 1992 Intonation and discourse type. Text 12: 397–427. Norrick, Neal and Kerstin Fischer 2009 Listener activities and responses: Panel Introduction at the 11th International Pragmatics Conference, Melbourne, 12–17 July 2009. Norrick, Neal 2004 Hyperbole, extreme case formulation. Journal of Pragmatics 36.9: 1727– 1739. Ochs, Elinor 1996 Linguistic resources for socializing humanity. In: John J. Gumperz and Stephen Levinson (eds.), Rethinking Linguistic Relativity, 407–438. Cambridge: Cambridge University Press. Oreström, Bengt 1977 Why /dhi/ Book? SSE Report, Survey of Spoken English, Lund University.

622

Gisle Andersen

Oreström, Bengt 1982 ‘When is it my turn to speak?’. In: Nils Erik Enkvist (ed.), Impromptu Speech: a Symposium, 131–145. Åbo: Åbo Akademi. Oreström, Bengt 1983 Turn-taking in English Conversation. Lund: Liber/Gleerup. Östman, Jan-Ola 1982 The symbiotic relationship between pragmatic particles and impromptu speech. In: Nils Erik Enkvist (ed.), Impromptu Speech: a Symposium, 147– 178. Åbo: Åbo Akademi. Östman, Jan-Ola 1995 Pragmatic particles twenty years after. In: Brita Wårvik, Sanna-Kaisa, Hiltunen, and Risto Hiltunen (eds.), Organization in Discourse: Proceedings from the Turku Conference, 95–108. Turku: University of Turku. Partington, Alan 1998 Patterns and Meanings: Using Corpora for English Language Research and Teaching. Amsterdam: Benjamins. Raymond, William D., Mark Pitt, Keith Johnson, Elizabeth Hume, Matthew Makashay, Robin Dautricourt and Craig Hilts 2002 An Analysis of Transcription Consistency in Spontaneous Speech from the Buckeye Corpus. ICSLP-02, Denver. Renouf, Antoinette 1986 The exploitation of a computerised corpus of English text. In: Michele Rivas (ed.), Actes du VIIIeme Colloque G.E.R.A.S, 123–136. Paris: Universite de Paris-Dauphine. Renouf, Antoinette 2007 Corpus development 25 years on: from super-corpus to cyber-corpus. In: Roberta Facchinetti (ed.), Corpus Linguistics 25 Years on, 27–50. Amsterdam/ New York: Rodopi. Renouf, Antoinette and Andrew Kehoe (eds.) 2006 The Changing Face of Corpus Linguistics. Amsterdam: Rodopi. Rissanen, Matti 1986 Variation and the study of English historical syntax. In: David Sankoff (ed.), Diversity and Diachrony. 97–109. Amsterdam: Benjamins. Romero, Jesus 1994 Ahm, ehm … you call it theme? A thematic approach to spoken English. Journal of Pragmatics 22: 495–509. Sajavaara, Kari 1982 Contrastive analysis and impromptu speech. In: Nils Erik Enkvist (ed.), Impromptu Speech: a Symposium, 195–206. Åbo: Åbo Akademi. Sandford, Daniel 2008 Discourse and metaphor: A corpus-driven inquiry. Corpus Linguistics and Lingustic Theory 4: 209–234. Schelfhout, Carla, Petro-Arno Coppen, Nelleke Ostdijk and Frans Van Der Slik 2005 Interjections in Dutch: a corpus-based approach, available at http://www.neerlandistiek.nl/publish/articles/000100/article.pdf. Schmid, Hans-Jörg 2001 ‘Presupposition can be a bluff’: How abstract nouns can be used as presupposition triggers. Journal of Pragmatics. 33.10: 1529–1552.

Corpus-based pragmatics I: qualitative studies

623

Schneider, Stefan 2007 Reduced Parenthetical Clauses as Mitigators. Amsterdam: Benjamins. Selting, Margret and Elizabeth Couper-Kuhlen (eds.) 2001 Studies in Interactional Linguistics. Amsterdam: John Benjamins. Shastri, S.V. 1988 The Kolhapur Corpus of Indian English and work done on its basis so far. ICAME Journal 12: 15–26. Sinclair, John M. (ed.) 1987 Looking up. An Account of the COBUILD Project in Lexical Computing. London: Collins ELT. Sinclair, John M. 1966 ‘Indescribable English’. Inaugural lecture, University of Birmingham, 1966 (unpublished). Sinclair, John M. 1991 Corpus, Concordance, Collocation. Oxford: Oxford University Press. Sinclair, John M. 1996 The search for units of meaning. Textus 9: 75–106. Sinclair, John M. 2004 Trust the Text: Language, Corpus and Discourse. London: Routledge. Sinclair, John M. and Malcolm Coulthard 1975 Towards an Analysis of Discourse: the English used by Teachers and Pupils. London: Oxford University Press. Sinclair, John M. and Malcolm Coulthard 1992 Towards an analysis of discourse. In: Malcolm Coulthard (ed.), Advances in Spoken Discourse Analysis, 1–34. London: Routledge. Sinclair, John M. and David Brazil 1982 Teacher Talk. Oxford: Oxford University Press. Sperber, Dan and Deirdre Wilson 1986 Relevance: Communication and Cognition. Oxford: Blackwell. Sperber, Dan and Deirdre Wilson 1995 Relevance: Communication and Cognition. 2nd ed. Oxford: Blackwell. Stenström, Anna-Brita and Annette Myre Jørgensen 2009 Youngspeak in a Multilingual Perspective. Amsterdam: Benjamins. Stenström, Anna-Brita, Gisle Andersen and Ingrid Kristine Hasund 2002 Trends in Teenage Talk: Corpus Compilation, Analysis and Findings. Amsterdam: Benjamins. Stenström, Anna-Brita, Gisle Andersen, Kristine Hasund, Kristine Monstad and Hanne Aas 1998 User’s Manual to Accompany The Bergen Corpus of London Teenage Language (COLT). Bergen: Department of English at the University of Bergen. Stenström, Anna-Brita 1982 Feedback. In: Nils Erik Enkvist (ed.), Impromptu Speech: a Symposium, 319–340. Åbo: Åbo Akademi. Stenström, Anna-Brita 1984 Questions and Responses in English Conversation. Malmo: Gleerup/Liber. Stenström, Anna-Brita 1986a A study of pauses as demarcators in discourse and syntax. In: Jan Aarts and Willem Meijs (eds.), Corpus Linguistics II. New Studies in the Analysis and Exploitation of Computer Corpora, 203–218. Amsterdam: Rodopi.

624

Gisle Andersen

Stenström, Anna-Brita 1986b What does really really do? Strategies in speech and writing. In: Gunnel Tottie and Ingegerd Bäcklund (eds.), English in Speech and Writing: A Symposium, 149–163. Stockholm: Almqvist and Wiksell. Stenström, Anna-Brita 1988 Adverbial commas and prosodic segmentation. In: Merj Kytö, Ossi Ihalainen and Matti Rissanen (eds.), Corpus Linguistics Hard and Soft, 221–236. Amsterdam: Rodopi. Stenström, Anna-Brita 1990a Adverbial commas and prosodic segmentation. In: Jan Svartvik (ed.), The London-Lund Corpus of Spoken English: Description and Research, 253–266. (Lund Studies in English 82.) Lund: Lund University Press. Stenström, Anna-Brita 1990b Pauses in monologue and dialogue. In: Jan Svartvik (ed.), The London-Lund Corpus of Spoken English: Description and Research. 211–252. (Lund Studies in English 82.) Lund: Lund University Press. Stenström, Anna-Brita 1991 Expletives in the London-Lund Corpus. In: Karin Aijmer and Bengt Altenberg (eds.), English Corpus Linguistics. Studies in Honour of Jan Svartvik, 239–253. London: Longman. Stenström, Anna-Brita 1994 An Introduction to Spoken Interaction. London: Longman. Stenström, Anna-Brita 1995 Taboos in teenage talk. In: Gunnell Melchers and Beatrice Warren (eds.), Studies in Anglistics, 71–79. Stockholm: Almqvist and Wiksell. Stenström, Anna-Brita 2006 Taboo words in teenage talk: London and Madrid girls’ conversations compared. In: Clare Mar-Molinero and Miranda Stewart (eds.), Language Variation and Change, 115–138. Amsterdam: Benjamins. Stubbs, Michael 1984 Review of Sinclair and Brazil 1982. Applied Linguistics 5: 71–74. Stubbs, Michael 1996 Text and Corpus Analysis. London: Blackwell. Stubbs, Michael 2001 Words and Phrases: Corpus Studies of Lexical Semantics. Oxford: Blackwell. Svartvik, Jan (ed.) 1992 Directions in Corpus Linguistics. Proceedings of Nobel Symposium 82 Stockholm, 4–8 August 1991. Berlin: Mouton de Gruyter. Svartvik, Jan and Randolph Quirk (eds.) 1980 A Corpus of English Conversation. Lund: Gleerup/Liber. Svartvik, Jan 1980 ‘Well’ in conversation. In: Sydney Greenbaum, Geoff Leech and Jan Svartvik (eds.), Studies in English Linguistics, 167–177. London: Longman. Svartvik, Jan 1982 The segmentation of impromptu speech. In: Nils Erik Enkvist (ed.), Impromptu Speech: a Symposium, 131–145. Åbo: Åbo Akademi.

Corpus-based pragmatics I: qualitative studies

625

Svartvik, Jan 1987 Graphic English prosody. In: Ishrat Lindblad and Magnus Ljung (eds.), Proceedings from the Third Nordic Conference for English Studies, 335–346. Stockholm: Almqvist and Wiksell. Svartvik, Jan 1988 A computerized corpus of spoken English. In: Vinca Liedloff (ed.), Standardization in Computerized Lexicography, 33–52. Saarbrücken: Institut der Gesellschaft zur Förderung der angewandten Informationsforschung e.V. an der Universität des Saarlandes. Tagliamonte, Sali and Rachel Hudson 1999 Be like et al. beyond America: The quotative system in British and Canadian youth. Journal of Sociolinguistics 3: 147–172. Taylor, Lita and Gerry Knowles 1988 Progress report on the Spoken English Corpus. In: Merja Kytö, Ossi Ihalainen and Matti Rissanen (eds.), Corpus Linguistics Hard and Soft, 237–244. Amsterdam: Rodopi. Thavenius, Cecilia 1982 Exophora in English conversation: a study of third person pronominal reference. In: Nils Erik Enkvist (ed.), Impromptu Speech: a Symposium, 291–305. Åbo: Åbo Akademi. Thomas, Jenny and Andrew Wilson 1996 Methodologies for studying a corpus of doctor-patient interaction. In: Jenny Thomas and Mick Short (eds.), Using Corpora for Language Research, 92– 109. London Longman. Tognini-Bonelli, Elena 2001 Corpus Linguistics at Work. Amsterdam: Benjamins. Tottie, Gunnel 1986 The importance of being adverbial. Adverbials of focusing and contingency in spoken and written English. In: Gunnel Tottie and Ingegerd Bäcklund (eds.), English in Speech and Writing: A Symposium, 93–118. Stockholm: Almqvist and Wiksell. Tottie, Gunnel and Carita Paradis 1982 From function to structure: Some pragmatic determinants of syntactic frequencies in impromptu speech. In: Nils Erik Enkvist (ed.), Impromptu Speech: a Symposium, 307–317. Åbo: Åbo Akademi. Tottie, Gunnel 1991 Conversational style in British and American English: The case of backchannels. In: Karin Aijmer and Bengt Altenberg (eds.), English Corpus Linguistics. Studies in Honour of Jan Svartvik, 254–271. London: Longman. Traugott, Elizabeth Closs and Bernd Heine (eds.) 1991 Approaches to Grammaticalization. Amsterdam: Benjamins. Van den Heuvel, Theo 1988 TOSCA: An aid for building syntactic databases. Literary and Linguistic Computing 3: 147–151. Venditti, Jennifer J. and Julia Hirschberg 2003 Intonation and discourse processing. In: Maria Josep Solé, Daniel Recasens and Joaquín Romero (eds.), Proceedings of the 15th International Congress of Phonetic Sciences, 107–114. Barcelona, Spain.

626

Gisle Andersen

Verschueren, Jef 2009 Introduction: The pragmatic perspective. In: Jef Verschueren and Jan-Ola Östman (eds.), Key Notions for Pragmatics, 1–27. Amsterdam: Benjamins. Vicher, Anne and David Sankoff 1989 The emergent syntax of pre-sentential turn openings. Journal of Pragmatics 13: 81–97. Virtanen, Tuija 2009 Discourse linguistics meets corpus linguistics: theoretical and methodological issues in the troubled relationship. In: Antoinette Renouf and Andrew Kehoe (eds.), Corpus Linguistics, Refinements and Reassessments, 49–66. Amsterdam: Rodopi. Walker, Gareth 2004 The phonetic design of turn endings, beginnings, and continuations in conversation. PhD thesis, Department of Language and Linguistic Science, University of York. Walsh, Steve, Anne O’Keeffe and Michael McCarthy 2008 ‘… post-colonialism, multi-culturalism, structuralism, feminism, post-modernism and so on and so forth’: A comparative analysis of vague category markers in academic discourse. In: Annelie Ädel and Randi Reppen (eds.), Corpora and Discourse: The Challenges of Different Settings, 9–30. Amsterdam: Benjamins. Waltereit, Richard 2006 The rise of discourse markers in Italian: a specific type of language change. In: Kerstin Fischer (ed.), Approaches to Discourse Particles, 61–76. Amsterdam: Elsevier. Wardhaugh, Ronald 1985 How Conversation Works. Oxford: Basil Blackwell, London: Andre Deutsch. Westney, Paul 1986 How to be more or less certain in English: Scalarity in epistemic modality. International Review of Applied Linguistics 24: 311–320. Wichmann, Anne 1994 F0 troughs and prosodic phrasing: an investigation into the linguistic information contained in a speaker’s base-line when reading. Hermes 13: 205–212. Aarhus School of Business. Wichmann, Anne 1996 Prosodic style: a corpus-based approach. In: Gerry Knowles, Anne Wichmann and Peter Alderson (eds.), Working with Speech. Perspectives on Research into the Lancaster/IBM Spoken English Corpus, 168–188. London/New York: Longman. Wichmann, Anne 2000 Intonation in Text and Discourse. London: Longman. Wichmann, Anne 2004 The intonation of Please-requests: a corpus-based study. Journal of Pragmatics 36: 1521–1549. Wichmann, Anne 2005 Please: from courtesy to appeal: the role of intonation in the expression of attitudinal meaning. English Language and Linguistics 9.2: 229–253.

Corpus-based pragmatics I: qualitative studies

627

Wichmann, Anne 2007 Corpora and spoken discourse. In: Roberta Facchinetti (ed.), Corpus Linguistics 25 Years on, 73–88. Amsterdam/New York: Rodopi. Wierzbicka, Anna 1991 Cross-cultural Pragmatics: The Semantics of Human Interaction. Berlin: Mouton de Gruyter. Yngve, Victor 1970 On getting a word in edgewise. In: Papers from the 6th Regional Meeting, Chicago Linguistic Society, 567–577. Chicago: Chicago Linguistic Society.

22.

Corpus-based pragmatics II: quantitative studies Christoph Rühlemann

Introduction Due to the massive dependence of pragmatic phenomena on context, corpora, as a relatively decontextualized medium, have long been seen by some researchers as unfit for use in pragmatic research. Nonetheless, corpus linguistic analyzes, both qualitative and quantitative in orientation, have produced a wealth of new insights into key pragmatic phenomena. The aim of this paper is to illustrate key quantitative corpus studies into phenomena of pragmatic interest. The paper is divided into six sections. The opening section addresses the question of how context-sensitive corpora are. Section 2 presents a case study into semantic prosody, an attitudinal phenomenon at the semantics/pragmatics interface. Section 3 presents a relevance-theoretic study of the pragmatic marker like. Section 4 is concerned with quantitative studies on reference. Section 5 introduces corpus research into speech acts. The concluding section looks to the future, outlining recent attempts at building multimodal corpora.

1.

Pragmatics and quantitative corpus linguistics: a troubled relationship?

Pragmatics is concerned with meaning in context. Because speakers can mean more than they say, pragmatics is “the art of the analysis of the unsaid” (Mey 1991: 245). The relationship between pragmatics, thus understood, and corpus linguistics is seen by some as a troubled one. The reason is simple: corpora record text, not meaning, and they record context only crudely, particularly in spoken corpora. Spoken corpora are based on transcriptions made from audio recordings. Since tape recorders cannot filter out non-speech noises, thus selectively ‘listening’ to speech – as humans can – many, and often large, stretches of corpus transcripts may be inaudible or unintelligible. Further, the information provided in spoken corpora about the contexts in which the spoken texts were produced is fairly minimal; we get to know who (sex, age, class, etc.) was talking to whom when. Moreover, other types of context are recorded only abstractly: we learn in what kinds of setting the talk occurred, or what type of interaction it was. Yet another type of ‘abstract’ context, which is considered so crucial that in most large corpora the data are categorized according to it, is ‘type of situation’, or register, such as academic writing, fiction, and conversation. The types of context largely missing from

630

Christoph Rühlemann

spoken corpora include the almost infinite wealth of concrete situational, nonverbal, and social context that conversationalists in their specific contexts of situation are connected to (cf. Cook 1990). In sum, we may have impoverished textual evidence, and only rudimentary or abstract contextual evidence. Therefore, corpora have long been seen by some researchers as unfit for use in pragmatic research. Indeed, some pragmatic features inevitably escape corpus linguistic analysis. This is because in part-of-speech (POS)-tagged corpora, only those phenomena can be studied fully whose lexical form(s) and pragmatic function(s) display a straightforward one-to-one relationship. This relationship is found, for example, in the words sorry and pardon, which are regularly included in apologies (cf. Jucker 2009). The form-function match is already weaker in compliments which need not necessarily be realized using typical conventionalized patterns (cf. Jucker et al. 2008). Where there is a complete form-function mismatch, as in cases of conversational implicature, a quantitative corpus study will be useless: what listeners take to be implicated in an utterance cannot be retrieved exhaustively from a corpus but can only be inferred (post hoc) with varying degrees of confidence. What a corpus can do even in those cases where the form-function mismatch of a phenomenon prevents exhaustive searches, is to provide the analyst with illustrative examples that are not only attested and, in this sense, authentic but also embedded in their co-texts, thus giving some evidence of the context in which they were used. Such corpus illustrations can usefully complement, or even replace, the invented and often completely decontextualized examples that have formed the basis of much pragmatic enquiry. Another approach to studying pragmatics corpus-linguistically is to use pragmatically-annotated corpora. For example, a small subcorpus of the Michigan Corpus of Academic Spoken English (MICASE) is tagged for some speech acts (cf. Maynard and Leicher 2007) and in the Corpus of Verbal Response Mode (VRM) Annotated Utterances, all utterances are coded twice: once for their literal meaning and once for their pragmatic meaning, using a principled taxonomy of speech acts (cf. Stiles 1992). However, the number of such pragmatically-annotated corpora is still small and they are coded for selected aspects of pragmatic interest only. Given these limitations, it may be surprising that quantitative corpus analyses of pragmatic phenomena have grown into a large body of literature and produced a wealth of new insights. In the following sections I will present some corpus linguistic studies into pragmatic units. The approach will be selective: only a few key studies can be presented. References to related studies, however, will be provided. Unless otherwise indicated, the illustrative examples will be taken from the British National Corpus (BNC), a 100-million-word corpus of contemporary British English (cf. Hoffmann et al. 2008).

Corpus-based pragmatics II: quantitative studies

2.

631

The semantics/pragmatics interface: semantic prosody

One of the clearest strengths of corpus linguistics is the analysis of what Sinclair (1991) refers to as the ‘idiom principle’, that is, broadly, how language patterns at the level of phraseology. A very large body of corpus linguistic literature points to ‘idiom’ phenomena such as collocation, collostruction, colligation and so forth as important building blocks of phraseological patterning. Another, crucial, phraseological phenomenon is ‘semantic prosody’, which I will focus on in this section. Semantic prosodies probably best illustrate Channell’s (2000) claim that some pragmatic phenomena can only be revealed in studies of large corpora. This is because semantic prosodies are normally hidden not only from introspection but also from observation of small numbers of examples. I will first briefly outline essential characteristics of semantic prosody and then illustrate the phenomenon in a case study of BREAK out. Semantic prosody is closely related to ‘semantic preference’, which is defined by Stubbs as “the relation, not between individual words, but between a lemma or word-form and a set of semantically related word forms” (2002: 65). Stubbs cites the example of the adjective large, which often collocates with nouns denoting ‘quantity’ and ‘size’ such as number, scale, part, amounts, etc. Semantic prosody, on the other hand, can, as a rule of thumb (Partington 2004: 149), be seen as a subcategory of semantic preference. The distinguishing feature of semantic prosody is that the type of semantic preference is related to polarity: the collocates of a word are either typically positive or, more often, negative. Because of the inherently evaluative nature of semantic prosody, alternative terminologies have been proposed, such as ‘pragmatic meaning’ (Channell 2000), ‘emotive prosody’ (Bublitz 2003) or ‘evaluative prosody’ (Morley and Partington 2009). Further, because of the evaluation they convey, semantic prosodies are said to be “on the pragmatic side of the pragmatics/semantics continuum” (Sinclair 2004: 34).1 Bublitz (1996) and Morley and Partington (2009) also note the cohesive role of semantic prosody, which the latter describe as “the mechanism which shows how one elemental type of meaning – evaluative meaning – is frequently shared across units of discourse and, by ensuring consistency of evaluation or evaluative harmony, plays a vital role in keeping the discourse together, in its cohesion” (2009: 139). An illustrative example from the Corpus of Contemporary American English (COCA) (cf. Davies 2009) is (1): (1) Hello, everyone. I am Dr. Sanjay Gupta. Your health, there is nothing more important. # And now that winter weather has set in, there are certain health and safety issues tat [sic] you need to be more concerned about, including how to avoid succumbing to sickness. (COCA: CNN_YourHealth 2002) This example involves the phrasal verb SET in, of which Sinclair (1991) observed that its subjects “refer to unpleasant states of affairs” (Sinclair 1991: 74). Sinclair cites as main vocabulary for SET in words like rot, decay, malaise, despair, ill-will,

632

Christoph Rühlemann

etc. In (1), the subject is winter weather, a rather neutral phrase. However, winter weather is negatively evaluated. This becomes clear when we consider the context: health and safety issues, need to be concerned about, and succumbing to sickness are clear indices of the presence of negative evaluation. The evaluative harmony, to use Partington’s term, in the excerpt is thus ensured by the accumulation of negative items. See also Louw’s (1993: 173) observation that “in many cases semantic prosodies ‘hunt in packs’ and potentiate and bolster one another”, an observation Bublitz (2003: 387) refers to as “the need to establish a common emotive ground by accumulating equi-polar means.” Excerpt (1) is illustrative of yet another crucial aspect of semantic prosody: it can serve as an effort-saving device “lighten[ing] the burden on the listener and free[ing] the speaker of the tedium of labouring a point” (Morley and Partington 2009: 144). This is, in Louw’s (1993: 157) words, because semantic prosody is defined as “an aura of meaning with which a form is imbued by its collocates.” In other words, because SET in typically co-selects negative subjects, the negativity of its collocates ‘colours’ the meaning of SET in in such a way that it is perceived as negative even if it co-occurs with a subject that is not clearly negative, such as winter weather. Louw’s notion of ‘imbuing’ has met with criticism. Whitsitt (2005), for example, argues that “there is no evidence for assuming that we can see the results of a diachronic process of imbuing” (2005: 296; emphasis in original; see also Bednarek 2008; Morley and Partington 2009). Whitsitt cites as evidence words such as alleviate and heal, which habitually co-occur with clearly unpleasant words and yet “certainly [do] not come to have an unpleasant meaning because of that company” (2005: 297). On the other hand, the fact that speakers/writers can achieve certain rhetorical effects such as irony by deliberately deviating from typical collocational patternings, as in outbreak of sanity, shows that the “underlying semantic prosody clearly persists, even and especially, in collocational clashes” (Morley and Partington 2009: 150). How can this evaluative persistence be explained? An answer that has recently been given is the theory of ‘priming’ (Hoey 2005). As a word, such as SET in, is learnt “through encounters with it in speech and writing, it is loaded with the cumulative effects of those encounters such that it is part of our knowledge of the word that it co-occurs with other words” (Hoey 2003), such as, in the case of SET in, heavily negative items like rot, decay and so on. That is, in (1), the speaker need not spell out that winter weather is taken as bad, for example by adding a negative adjective such as cold, severe, or harsh. S/he can trust the addressee to know the primings of SET in, that is, a set of ‘instructions’ on how to use it, how it normally interacts with other items. Among these primings is a word’s semantic prosody, “an instruction, which tells the reader ‘when you find me in a text, read the surrounding discourse in a favourable/infavourable light, unless there’s something else around which tells you not to’” (Partington forthcoming). The evaluation conveyed by semantic prosodies may be obvious or ‘hidden to the naked eye’. Morley and Partington (2009: 151) suggest that it is “best con-

Corpus-based pragmatics II: quantitative studies

633

sidered as a cline”, with items such as murder and venerable on the ‘overt-evaluation’ end expressing fairly clearly unfavourable and favourable evaluation, items such as peddle and fraught with somewhere in the middle, and items such as SET in and not + BUDGE (see below) occupying the ‘covert-evaluation’ end where the evaluation can only be identified using corpus linguistic methods. The linguistic analysis and description of semantic prosody faces several challenges. One problem is that while it may be easy to achieve a consensus that words such as rot, decay etc. describe negative states of affairs, the detection of negative or positive semantic prosody may be much less straightforward in other cases. Whether a word is used and interpreted positively or negatively depends in part on the wider context in which it is used and in part on the speaker who uses it. Stubbs (1996) cites the example of intellectuals, whose collocates – e.g., activist, student, leftwing, liberal – may have a negative ‘ring’ in some circles but be positively evaluated in others (cf. Bednarek 2008: 122). Semantic prosody therefore “cannot be ‘objectively’ derived from corpus data and requires a lot of inference on the part of the analyst” (Bednarek 2008: 132).2 A second problem is the evaluative inconsistency of semantic prosodies. Prosodies can be ‘switched off’ or even reversed. Morley and Partington (2009: 149) discuss the example of not + BUDGE, which usually carries a negative prosody. However, under certain circumstances, for example when the subject is first person, it can be used favourably. An example of positive evaluation in the use of not + BUDGE is given in Hunston (2007): (2) The Prime Minister rejected resounding calls for the resignation of the government, “I will not budge”, he said Further, the distribution of semantic prosody and word form may be ‘asymmetrical’ (Bublitz 1996). That is, it may be misleading to claim that ‘verb X’ or ‘noun Y’ has positive or negative prosody because in many cases different forms of one and the same word have different evaluative tendencies. Consider the phrasal verb BREAK out: when used transitively, as in Just as linotype operators at the Sun were breaking out the champagne to celebrate the arrival of £1,000 pay packets (CHU 308), its prosody is clearly positive (cf. Louw 2000); when used intransitively, as in Fires keep breaking out (A18 1206), its prosody is decidedly negative (cf. the discussion below). Another problem is that semantic prosodies can be subject to variation across registers. O’Halloran (2007: 15) demonstrates that, for example, the past tense form erupted “has largely positive associations in the sports report register, but largely negative ones in the hard news register.” That is, the non-register specific concept of semantic prosody may, in some cases, have to be replaced by the concept of ‘register prosody’, which indicates that “some prosodies have probabilistic relationships to register” (O’Halloran 2007: 4; see also Bublitz 1996). Finally, semantic prosody cuts across the traditional lexis/grammar dichotomy. The GET-passive has been shown to have a negative prosody in that it typically at-

634

Christoph Rühlemann

tracts past participles sharing an ‘adversative’ core meaning (Carter and McCarthy 1999; Stubbs 2002; Rühlemann 2007b). Typical examples are GET stuck, GET caught, and GET killed. Further, semantic prosody may also be observed at a level below the word. Rühlemann (2006) demonstrates that the prefix dys-, as in dyslexia, dysfunction, dysplasia, etc., as well as the prefix dis-, when used as a productive morpheme, as in disease, disappear, disabled, etc. form words that have negative prosody. In the remainder of this section a case study into the pattern ‘(inanimate) N + (intransitive) BREAK out’ will bring into relief some of the above said. A search for the verbal lemma BREAK immediately followed by out finds 1,126 occurrences in the BNC. (3) shows a random excerpt from the concordances. As is suggested by the inanimate nouns (underlined in (3)) preceding the string, negative things such as hostilities, war, mayhem etc. break out. (3) if hostilities ever broke out. We’ve When war broke out he returned to complete mayhem broke out. Hats were flung a fight broke out at the Zuwaya resembling panic broke out among the ladies Is this association with negative items systematic? Table 1 presents the 15 most frequent nouns preceding BREAK out in the BNC; right hand column gives the Z-scores, which measure the strength of attraction between collocates. Table 1.

Top 15 nouns collocating with BREAK out

No.

Collocate

Freq.

Z-score

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

war fire fighting fight row world violence blaze riot argument hostility scuffle rioting way sweat

259 89 46 39 36 32 20 17 17 17 16 14 13 12 11

167.8 82.8 113.8 75.3 48.5 12.6 29.4 72.4 44.7 16.3 43.5 132.3 82.1 1.1 37.2

Corpus-based pragmatics II: quantitative studies

635

As can be seen from Table 1, BREAK out most commonly collocates with inanimate nouns that share physical violence and, to a lesser degree, fire meanings. War, the most extreme manifestation of violence, is not only by far the most frequent collocate but also has the highest Z-score; note also that the 259 occurrences of war + BREAK out account for more than a fifth of all occurrences of BREAK out. War is thus a very strong collocate of BREAK out. Other nouns in the violence group include fighting (rank 3), fight (4), row (5) etc.; world, ranked 6th, occurs in combination with war forming the compound world war. The fire group includes the nouns fire (2) and blaze (8). Inspection of concordance lines suggests that it is invariably destructive, not warming or cosy, fire that breaks out. Since destructive fire is an inevitable concomitant of war, we might subsume the fire meanings under the heading of the violence meanings, defining the negative prosody of BREAK out thus: BREAK out co-selects inanimate nouns that typically express violence (including destructive fire) meanings. BREAK out illustrates that the concepts of semantic prosody and semantic preference are not mutually exclusive in that the phrasal verb is a semantic prosody with a clear semantic preference. Interestingly, there are six occurrences of peace + BREAK out in the BNC. Inspection of the concordance lines suggests that covert evaluation is the driving force behind these uses. Consider (4) (from the Economist, 1991): (4) Where will the dollar head when peace breaks out? (ABH 2961) In (4), the prospect of peace (at the time of the Gulf war) is taken as a threat to the dollar, which has risen before the war but, it is suggested, might fall when the war is over. That peace may have bad consequences for the dollar is not made explicit. Rather, the negative ‘aura of meaning’ surrounding BREAK out is tacitly projected onto peace. In sum, prosodies are a useful resource for the diagnosis of overt and covert speaker evaluation. Further, as Louw argues, there is the same well-calculated collocational deviance in instances of irony (Louw 1993). The prospect, then, is that semantic prosodies may help to computationally uncover irony, a pragmatic phenomenon that has so far escaped quantitative empirical study. Moreover, semantic prosodies may assist in the study of persuasion: “Propaganda, advertising and promotional copy will now be gradable against the semantic prosodies of the whole language” (Louw 1993: 173). Exploiting the diagnostic potential of semantic prosodies for these kinds of speaker meaning has only just begun and more insights may be expected from this promising avenue of research.

3.

Pragmatic markers: like in a relevance-theoretic perspective

Pragmatic markers are words or phrases that do not add so much to the propositional content of utterances as they metalingually flag how discourse relates to

636

Christoph Rühlemann

other discourse. Pragmatic markers thus play a crucial role in facilitating processes of pragmatic inference. Because they are lexically relatively fixed and thus relatively easily retrievable from a corpus, corpus research into pragmatic markers has been extremely productive.3 This section presents Andersen (2001), a corpus study on a marker that is particular in that it attends to several discourse functions and, not surprisingly, has attracted a wealth of recent research: like. Andersen’s study, which is based on the Bergen Corpus of London Teenage Language (COLT), is notable because the approach to like is informed by relevance theory (Sperber and Wilson 1995). This theory holds that human communication is geared towards the maximisation of relevance such that from the wealth of information and stimuli in our cognitive environments we tend to select that information and those stimuli that are most relevant to us in any given communicative situation. Andersen identifies five broad functions of like used as a pragmatic marker. They are illustrated in (5)-(9). In (5), like carries out an approximation function indicating that the numerical information is approximate rather than precise; in (6), like is in an exemplification function. The common denominator of uses of like classified as approximation and exemplification is “that they involve non-identical resemblance between the encoded and the communicated concepts” (Andersen 2001: 237). The third function Andersen identifies is ‘metalinguistic use’, which concerns the speaker’s relation to the proposition “in terms of its formal linguistic characteristics” (Andersen 2001: 243); in example (7), like marks a particular word choice (reminder) as potentially inexact or inappropriate. Another function like can carry out is as ‘interpretive use’, a term Andersen prefers over the more commonly used term ‘quotative’ on the grounds that like may be used to preface not only presentations of speech, as in (8), but also “gestures and facial expressions that can be seen as metarepresentations of speaker attitude” (Andersen 2001: 254; see also Buchstaller 2008). Finally, Andersen stresses the role of like as a hesitational/linking device, that is, its function in accompanying false starts, self repairs and cut-off utterances and in providing a discourse link between syntactically distinct units of discourse (Andersen 2001: 259). In (9), for example, like acts as a pause filler granting the speaker planning time (note the co-occurrence with the hesitation form er): (5) I mean I’ve been in two shops now there’s fifty pound difference like, you know (KB2 2401) (6) think it’s the way he looks, like, if you know what I mean, you know like […] pull his face and like, look over glasses (KB1 768) (7) PS007 >: Oh yes, I was assuming that … erm … a I do, I didn’t really set it out as a formal agenda just as a PS002 >: Just as a note PS007 >:a reminder, like. (KB0 1337)

Corpus-based pragmatics II: quantitative studies

637

(8) Yeah that’s what I, why, that’s what I said to Susanna and she was like don’t be ridiculous! (KP5 2163) (9) PS03T >: well you should wipe the outside of the frame didn’t you? PS03S >: er, yes, like, you like just flick round it (KBB 6472) The functional distribution of like in the COLT is shown in Table 2: Table 2.

Functional distribution of like (cf. Andersen 2001: 266)

Function

%

Approximation Exemplification Metalinguistic focus Interpretive use (quotative) Hesitational/discourse link

21 19 19 7 35

As is shown in Table 2, the hesitational/linking functions in the corpus are most salient, followed by approximation, exemplification, and metalinguistic use. The quotative function, by contrast, is relatively infrequent. This may be contrary to expectations, given the higher rates for quotative like in Tagliamonte and Hudson (1999). The discrepancy between Andersen and Tagliamonte and Hudson, however, may be due to the fact that quotative like is a recent development that has been spreading to many regional varieties of English (cf. Buchstaller 2008) and that Tagliamonte and Hudson’s data were collected in 1996 whereas the COLT data stem from the early 1990s. Do all of the different functions of discourse marker like have a common root? Andersen (2001) interprets discourse marker like as a ‘looseness marker’, that is, as “a signal that the relation between an utterance and its underlying thought is not a one-to-one relationship, but a relation of non-identical resemblance” (2001: 230). As such, it is a highly co-constructive item inviting recipients to collaborate in the negotiation of meaning.

4.

Reference

Reference undoubtedly touches on one of the most fundamental questions concerning language: how can we speak and, in speaking, communicate something meaningful about things, people, and states of affairs? The answers linguists and, more importantly, philosophers have suggested are by far too varied and complex to go into in sufficient detail in this short section (see Schwarz-Friesel and Consten this volume). A few remarks must suffice.

638

Christoph Rühlemann

It is often claimed that referring expressions refer to their referents, such that, for example, the expression ‘the present U.S. American president’ ‘refers’ to Barack Obama. On this view, the expression denotes an object in the real world; see Russell’s (1905) influential paper, entitled “On denoting”. Denotation, in Russell, is taken to be “a relation between an expression, considered in abstraction, and the thing that is the expression’s referent or denotatum” (Lycan 2008: 19). Russell distinguishes between meaning and denotation. An expression such as ‘the present U.S. American prime minister’ would have to be assigned meaning because it could be perfectly understood, but, since the US political system does not provide for a prime minister, it would fail to denote anything. In Russell’s view, the sentence “The present U.S. American prime minister is wise” would be considered to be false. A number of objections have been brought against Russell’s theory of definite descriptions, chiefly by Strawson’s (1950) article aptly entitled “On referring”. Strawson thought of referring not as an abstract relation between an expression and an object but argued that referring “is not something an expression does; it is something that someone can use an expression to do” (1950: 326). The variable that he brings into the equation is “the context of an utterance […] and by ‘context’ I mean, at least, the time, the place, the situation, the identity of the speaker, the subjects which form the immediate focus of interest, and the personal histories of both the speaker and those he is addressing” (1950: 336). Whether ‘the present U.S. American president’ is used to refer to Mr Obama or his predecessor or his successor or any other U.S. American president crucially depends on the circumstances of its use. Also, the sentence ‘The present U.S. American prime minister is wise’ would not be seen as lacking truth-value; since it builds on a false presupposition, the question whether the statement is true or false simply does not arise. On this view, which links reference intimately to context and, hence, inference, reference is deeply pragmatic. The following two subsections section present corpus-based research into two types of referring expressions: definite noun phrases and deixis. 4.1.

Reference through definite noun phrases

One type of material that corpus linguists standardly work with are frequency lists, that is, lists in which the words in a corpus are ranked in order of their frequency in that corpus. Table 3 displays what are, according to Kilgarriff (1998), the 20 most frequent items in the conversational subcorpus and the written subcorpus of the BNC.

Corpus-based pragmatics II: quantitative studies

Table 3.

639

20 most frequent items in the conversational subcorpus (C) and the written subcorpus (W) of the BNC C

W

Rank

Frequency

Item

Tag

Frequency

Item

Tag

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

167,640 135,217 128,165 115,247 92,239 90,886 77,611 68,846 62,382 58,810 48,322 47,391 43,977 42,241 41,654 38,515 35,156 34,901 34,837 34,477

i you it the ’s and n’t a that yeah he to they do oh what is of was in

pnp pnp pnp at0 vbz cjc xx0 at0 dt0 itj pnp to0 pnp vdb itj dtq vbz prf vbd prp

5,776,384 2,789,403 2,421,302 1,939,617 1,695,860 1,468,146 1,892,937 1,845,350 1,839,964 1,834,957 1,768,898 1,606,027 1,605,749 1,603,178 1,590,305 1,580,267 1,561,041 1,490,673 1,435,574 1,426,207

the of and a in to is to was it for with he be on i that by at you

at0 prf cjc at0 prp to0 vbz prp vbd pnp prp prp pnp vbi prp pnp cjt prp prp pnp

Table 3 reveals clear differences in the ways that language users refer in conversation (C) and writing (W) respectively. Among the 20 most frequent items in C, there are (i) more personal pronouns (the pronoun they is not included in the top 20 in W) and (ii) far less NP-related items (prepositions and the definite and the indefinite article) than in W: there are ten NP-related items in W but only four in C. Table 3 further shows that the definite article the is by far the most common word in writing (roughly twice as common as the next one, the preposition of). Indeed, the is not only the most frequent word in writing but also the most frequent word in the whole of the BNC (cf. Kilgarriff 1998) and in most other general corpora, such as the Cambridge International Corpus (CIC) (cf. McCarthy 1998) or the Bank of English (BoE) (cf. Sinclair 1999). With its roughly 6 million occurrences in the 90-million-word written subcorpus of the BNC, the accounts for more than six percent of all word tokens in that subcorpus. It is ironic that, a good 100 years on,

640

Christoph Rühlemann

Russell’s verdict that “to the philosophical mathematician [the] is a word of great importance” (1919: 167) should be empirically confirmed by corpus frequency counts. This evidence from the BNC suggests two major interpretations: (i) the most common referring technique in writing is the use of both the definite NP and, less importantly, the indefinite NP, while (ii) in conversation, reference is expressed most commonly by means of pronouns. While this latter finding will be discussed in the next subsection (5.2), the remainder of this subsection takes a closer look at the use of the definite article both in conversation and writing. The overriding function of the definite article is to specify “that the referent of the noun phrase is assumed to be known to the speaker and the addressee” (Biber et al. 1999: 263). That is, broadly speaking, the functions as a marker of given information. It does so in a variety of ways. In a study carried out on the Longman Spoken and Written English (LSWE) Corpus, Biber et al. (1999) identify the following reference patterns of definite NPs marked by the.4 They are illustrated by relevant examples: (10) Anaphoric: A MAN died and a girl was badly injured when fire ripped through a house yesterday. The girl, who had been clinging to a third-floor window ledge, fell just as firemen were about to grab her. (CBF: 2921) (‘the girl’ is understood as the girl that ‘was badly injured when fire …’; reference is backward-looking) (11) Indirect anaphoric: A woman died yesterday after being knocked down by a shoplifter fleeing with a £2.58 descaler. Mrs Lillian Amelia Smith, 81, sustained a fractured skull at the store at Newham, east London, on Friday. Police are treating the incident as murder. (AKH: 260) (‘the incident’ can be identified via inference as the events described in the preceding report) (12) Cataphoric: The next main point is about the complexity of the system. (HHW: 3638) (‘the complexity’ in question is identifiable via the of-phrase following it; reference is forward-looking) (13) Situational: I think there might be parcel for you at the door is it? (FPU: 358) (based on ‘frame knowledge’ the reference of ‘door’ is understood as ‘the front door’ where parcels are normally delivered) (14) Generic:

Just for once, a Frenchie has conceded that the Brits do something better. (A0C 1386)

Corpus-based pragmatics II: quantitative studies

641

(‘the’ denotes ‘the whole class of Brits’ without specific reference to particular Brits) (15) Idiom: But, quite frankly, what’s the point? (A5Y: 230) (‘the’ is an integral part of the idiom ‘what’s the point?’; it has no referring or denoting function; see also Searle 1969: 72) Intriguing results emerge from the analysis of the extent of use of these patterns in different registers by Biber et al. (1999). The registers considered include three written registers (fiction, news, and academic writing) and conversation. The results are summarized in Table 4: Table 4.

Percentage use of reference patterns in four registers (C: conversation; F: fiction; N: news reportage; A: academic writing) (cf. Biber et al. 1999: 266)

Situational Anaphoric Indirect anaphoric Cataphoric Generic Idiom Uncertain

C

F

N

A

55 25 5 5