Raymond Chapman Linguistics and Literature

man

An introduction to literary stylistics

T o Norman Denison

__________.

Linguistics and Literature An introduction to literary stylistics

Raymond Chapman Senior Lecturer in English, London School of Economics and Political Science

Edward Arnold

© Raymond Chapman, 1973 First published 1973 by Edward Arnold (Publishers) Ltd., Я5 Hill Street, London W iX 8LL Reprinted with corrections and glossary, 1974

Cloth edition ISBN: о 7131 5685 6 Paper edition ISBN: о 7131 j686 4

All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise without the prior permission of Edward Arnold (Publishers) Ltd.

Printed in Great Britain by Whitstable Litho Ltd., Whitstable, Kent

( Contents

Preface

vi

Note on Reading

vii

i

Allies or Opponents?

I

2 Literature and Stylistics

9

3

Language, Literature and History

21

4

Speech and Writing

32

5

Syntax

44

6

Words and Meanings

58

7

The Language of Rhetoric

72

8

Rhythm and Metre

85

9

Beyond the Sentence

100

Glossary of Linguistic Terms

11 3

Index

118

Preface

T h is book is addressed to those who are interested, as students, teachers or general readers, in literature or linguistics. Its aim is to add them to the increasing num ber o f people who are interested in both and have found that the two disciplines can illum inate one another in m any ways. It is an introductory study, which does not pretend to give all the answers; indeed, stylistic criticism does not encourage anyone to claim to know all the answers. W h at is attem pted here is a presentation o f some possible methods o f approach to the basic problems. Learning to ask the right sorts o f question is perhaps the most im portant part o f an academ ic training. I am grateful to m y colleagues N orm an Denison, Jean Aitchison and D avid Durkin for advice on some points o f linguistics. Elizabeth Johnson and Betty Sm ale have bravely tackled the vagaries o f m y handw riting and typing. T h e staff o f the London L ib rary are often thanked in prefaces for their helpfulness; I should like to add to the list.

Note on Reading

Although literary stylistics is still a com paratively new study, a great deal o f w ork in this field has been published and any sug­ gestions for further reading must necessarily be selective. E ach chapter is followed by a list o f books and essays w hich are particu­ larly relevant to the topics which have just been discussed. M uch important w ork is to be found in periodicals; but as m any o f these are specialist journals not likely to be accessible to all readers, reference has been m ade only to articles w hich have later been included in books. T h e serious student w ill need to consult a full bibliography such as: R . W . B ailey and D . M . Burton, English Stylistics (Cam bridge, Mass., 1968, Massachusetts Institute o f Technology Press). L. T . M ilic, Style and Stylistics (London, 1967, C ollier-M acm illan). T h e following books, w hich are frequently mentioned in the text and reading lists, are identified by one or two words in italics: Chatman: E ssays-S. Chatm an and S. R . Levin, eds., Essays on the Language o f Literature (Boston, 1967, H oughton M ifflin). Chatman: S ty le-S . Chatm an, ed., Literary Style: a Symposium (London, 19 71, O xford University Press). Fowler- R . Fowler, ed., Essays on Style and Language (London, 1966, R outledge and K egan Paul). Leech- G . N . Leech, A Linguistic Guide to English Poetry (London, ідбд, Longm ans). Minnis - N . M innis, ed., Linguistics at Large (London, 1971, Gollancz). Nowottny- W . Nowottny, The Language Poets Use (London, 1962, Athlone Press).

In ad dition to the books listed at the end o f each chap ter, the follo w in g are recom m en d ed :

H. S. Babb, ed., Essays in Stylistic Analysis (New Y ork, 1972, H arcourt Brace). M . W . Croll, Style, Rhetoric and Rhythm (Princeton, N. J ., 1966, Princeton U niversity Press). D . C . Freem an, ed., Linguistics and Literary Style (New Y ork, 1970, H olt, R inehart and W inston). R . F o w ler et al., The Languages o f Literature (L on d on , 19 7 1, R o u tled ge

and K egan Paul). G . H ough, Style and Stylistics (London, 1969, R outledge and K egan Paul). H . M . H ulm e, Explorations in Shakespeare's Language (London, 1962, Longm ans). M . Joseph, Shakespeare’s Use o f the Arts o f Language (New Y ork, 1969, C olum bia U niversity Press). S. R . Levin, Linguistic Structures in Poetry (The H ague, 1962, M outon). D . Lodge, The Language o f Fiction (London, 1966, Routledge and K eg an Paul). J. M iles, Style and Proportion (Boston, 1967, Brown). A . C . Partridge, The Language o f Renaissance Poetry (London, 1971, Andrd Deutsch). T . Sebeok, ed., Style in Language (New York, i960, John W iley). L . Spitzer, Linguistics and Literary History (Princeton, 1948, Princeton U niversity Press). S. U llm ann, Language and Style (Oxford, 1964, Blackwell).

I

Allies or Opponents ?

O ne o f the problems raised b y the linguistic theory o f Noam Chom sky is the status o f sentences like this: Colourless green ideas sleep furiously. Here we have a sequence o f words w h ich m u st be accounted ‘accept­ able’ English on gram m atical criteria, since it responds to analysis by any reasonable m ethod o f classification, but w hich can hardly be seen as an ‘acceptable’ part o f m eaningful discourse in known varieties o f English as a com m unicating language. It is difficult to find a context in which that particular utterance could be used. Students o f literature m ay respond differently to a sentence that deviates from the expectations o f everyday usage. T h e sentence quoted above is not too remote from the type o f language in which we learn to accept, and even to admire, indicative statements such as: H er fist o f a face died clenched on a round pain (D ylan Thom as, ‘In M em ory o f A nn Jones’) or: N o, I ’ll not, carrion comfort, Despair, not feast on thee (Hopkins, ‘ Carrion Com fort’) in which the gram m ar follows a ‘ normal’ pattern but in which scarcely any o f the form-words are associated in fam iliar ways. A gain , Chom sky and his followers have been m uch concerned with the problem o f am biguity in linguistic statements. It is apparent that the earlier m ethod o f I С analysis (see p. 7) does not help to give an unequivocal m eaning for the sentence: T h e police were ordered to stop drinking after midnight.

2

Linguistics and Literature

which has four possible interpretations that can be sorted out only b y a deep analysis leading perhaps to rephrasing. In our everyday use o f language we rightly regard such uncertainty o f m eaning as undesirable and do our best to avoid it. Y e t critics continue to argue about the interpretation o f the last lines o f K eats’s ‘ O de on a G recian U rn ’ : W hen old age shall this generation waste, T h o u shalt rem ain, in midst o f other woe T h a n ours, a friend to man, to whom thou say’st, Beauty is truth, truth beauty,— that is all Y e know on earth, and all ye need to know. W hat exactly is the U rn supposed to be ‘saying’— the phrase ‘Beauty is truth, truth beau ty’, followed by the poet’s own comment, or the whole o f the last two lines? In the absence o f quotation marks from the original edition, we m ay be in doubt. T h e argum ent must certainly take into account the question whether K eats elsewhere uses ‘y e ’ as a singular pronoun. T h e point is that the uncertainty is a source o f stim ulating discussion, adding to the interest o f the whole poem rather than detracting from it. T h ere exists, then, a type o f discourse in w hich we can apply evaluative words like ‘clever’, ‘interesting’, and even ‘ brilliant’ to usage w hich m ight evoke very different adjectives if it occurred in a different situation. W e m ay justify such discrimination by assigning Thom as, Hopkins and K eats to the realm o f ‘literature’ when we are discussing certain written texts which they have left to posterity. T h e same tolerance m ight not extend to other aspects o f their writing, such as their letters or their critical essays, and would certainly not be extended to an encounter w ith them in polite conversation if such were possible— though in practice w e m ight carry over some o f the respect attached to their literary reputation. W hat is literature? T h e question m ay find an answer, or rather a num ber o f answers, as we exam ine some specimens o f language in the course o f this book. A t this stage it is easier to say w hat literature is not. First, it is not sim ply that w hich is w ritten as opposed to that w hich is spoken. It is true that w e speak loosely about the ‘literature’ which a m anufacturer sends out to promote his product. Y et none o f us w ould include such areas when considering the possible scope o f a syllabus in English Literature. N or w ould w e include recipe books, telephone directories, Acts o f Parliam ent or guidebooks to

Allies or Opponents?

з

ancient buildings, though all o f these are written and m ay be of interest for the study o f language in general. T h e distinction is not always quite so clear. I f Pilgrim's Progress is counted as literature, w hat about H obbes’s Leviathan? Is M atthew A rnold’s Culture and Anarchy literature as unquestionably as is his poetry? T here is not likely to be a perfect test o f admissibility to determine all cases; rather a spectrum o f linguistic utterance, at one end o f which are specimens o f undisputed literature, at the other a much larger corpus that cannot be so labelled. T h ere will always be an area o f doubt and it need not greatly trouble us. T w o features perhaps w ill be noticeable in those works w hich form a society’s literature in the more specialized sense o f the word. O ne is the interest attaching to the w riter’s choice o f framework for the dis­ courses w hich together m ake up his extant work. H e w ill have used a method o f organizing and connecting w hat he wants to say on a particular topic, generally in a unit which other writers have used and w hich critics have labelled as the species and sub-species known as genres. W e approach a writer as a novelist or a dramatist or a poet, with narrower categories in mind— lyric, epic, elegy and so on. There are dangers that these labels w ill becom e artificial and restrictive, imposing a rigidity which is not inherent in the work itself; and o f course m any writers have practised different genres. Nevertheless, there is a special interest attaching to Bunyan’s choice o f prose allegory for Pilgrim's Progress, and M ilton’s choice of epic verse for Paradise Lost, w hich does not attach to the fact that textbooks are written in referential prose. T h e other distinguishing feature o f literature brings in a word which has been given m an y interpretations: ‘im agination’ . For the m om ent it is sufficient to say that the m eaning is not confined to that o f fantasy or even to the creation o f characters and episodes which never had a ‘real’ existence. It means that the linguistic utterance which involves im agination has a quality beyond the use o f words to convey referential meaning. A work o f literature m ay indeed offer information; it m ay, and probably w ill, have a m eaning­ ful content which can be paraphrased in referential prose. But such a paraphrase will certainly seem ‘less’ than the original; it will have ‘lost’ something, it w ill be ‘poorer’ . T h e search for the positive quality im plied b y those negative words is an im portant part o f literary criticism. Literature, then, seems to offer language w hich is different from w hat m ay be loosely termed the ‘norm al’ or ‘everyday’ usage o f a

4

Linguistics and Literature

speech-comm unity, yet w hich is intelligible to the members o f that com m unity i f they are w illing to apply a special standard o f accept­ ability. Literary language has been chosen and m anipulated by its user w ith greater care and com plexity than the average languageuser either can or wishes to exercise. I f this distinctive use is recog­ nized, it m ay be possible to discuss intelligently a w riter’s individual ‘style’ . These are questions to w hich we must return in more depth. Even in these general terms, they seem to lim it the study o f language to a particular area, and thus to deny the omnivorous appetite for data shown b y modern linguistics. T h e question ‘w h at is linguistics?’ m ay be answered more precisely than the equivalent question about literature, but it cannot be fu lly answered in a few words. Readers o f this book w ill probably have some basic knowledge o f linguistics. For those who have not, the books listed at the end o f this chapter w ill prove helpful. Briefly, the study o f linguistics is concerned with language as an observable phenomenon o f hum an activity, both in its general principles and in the particular realizations w hich we call ‘languages’— English, French, M ala y, A rab ic and so on. C learly, literature is created from the basic m aterial o f linguistic study and is allied to it in a w a y that the other arts like music and painting are not. Y e t it w ould be a sad error to regard linguistics as valuable only in connection w ith the study o f literature. Linguists are interested in every form o f language use, and also in the under­ lying ‘rules’ which govern potential as well as actual use. Literature occupies only a very small area o f the total language m ap, and we have already found reason to suppose that it is a rather unusual area. Should the linguist then eschew the literary creations o f a language? This attitude tended to prevail during the founding period o f modern linguistics. Ferdinand de Saussure, with his insistence on the prim acy o f everyday speech, was little interested in the written language and even less in the specifically literary: in his view, they were special uses which were com paratively unim portant in the study o f language as a whole. His pupil Charles Bally, who began the systematic study o f w hat we now call ‘stylistics’, again gave scant attention to literature. Leonard Bloomfield, while paying a scholar’s respects to the cultural value o f literature, did not value it highly as a field o f linguistic investigation; it deviated too m uch from the common denom inator and was tainted with the association o f classical

Allies or Opponents?

5

philology which the new linguistics was trying to leave behind. Bloomfield’s words are worth quoting since some linguists would still subscribe to them: T h e linguist . . . studies the language o f all persons alike; the individual features in w hich the language o f a great writer differs from the ordinary speech o f his time and place interest the linguist no more than the individual features o f any other person’s speech, and m uch less than the features that are common to a ll speakers. {Language, pp. 21-2) W hen those words w ere written, the study o f linguistics was still fighting for its autonom y and needed to emphasize in w hat respects it differed from traditional language studies. T o d a y most o f the early anxieties have been outgrown and linguists are ready to re-open the frontiers that were closed in defence. T h e tools o f linguistics can be used in related disciplines w ithout reducing linguistics itself to a mere technology or service-station. T h e literature o f a language offers a corpus o f m aterial for linguistic study. It is, as we shall see, deviant in some respects from the more orthodox field o f the linguist’s concern. It is mostly written; it is m ostly o f the past; and it presents features peculiar to itself which are not found in other areas o f expression. T h e more im portant consideration is that literature is the work o f men who were specially sensitive to the language o f their time and who used the skill o f language to m ake perm anent their vision o f life. T h e y m anipulated language to m ake it contain a unique series o f ex­ periences and interpretations. T h a t, surely, is enough reason for bringing every available scholarly skill to bear on its elucidation. Co-operation without suspicion is needed from the literary scholar as well. Despite the outstanding critical work done in recent years b y writers with both literary and linguistic training, there is still a general disapproval o f linguistics when it impinges on literary subjects. It is regarded as ‘ too scientific’ ; its m athem atical diagrams and term inology, its developm ent o f theory from em pirical observa­ tion, its refusal to be prescriptive about ‘good’ or ‘ bad ’ usage, all serve to alienate the m ore traditional literary scholar. Y e t some o f these approaches are precisely w hat criticism needs for its con­ tinuing vitality. This does not m ean that all other critical approaches must be

6

Linguistics and Literature

cast aside in the euphoria o f w hat is new. T h e kind o f study on w hich we are going to em bark does not yield the whole truth about literature. Nevertheless, it is literary criticism and not some strange, im proper use o f literary m aterial. It is a proper concern o f literary study, but not the total concern. Frank Palm er has put the point in words w hich seem to me to adm it no refutation: No linguist should ever hope to explain the aesthetic values of literature b y linguistic investigation any more than the values o f great music can be explained simply by a careful exam ination o f the score. But literature no less than everyday speech is language and as such is a proper subject for linguistic investigation, even if there are some who w ould regard the linguistic analysis o f a poem as a kind o f blasphemy. {Minnis, p. 252) W e shall go on therefore in the belief that students o f literature and o f linguistics have each something to gain from that which was once considered to be the other’s province. This book does not set out to be a survey o f either linguistics or literary-critical methods, but rather to suggest the kind o f investigation w hich m ay be helpful to both disciplines. It must be remembered that, although linguistics is now an autonomous discipline, it is not a homogeneous one. T here are m any schools, theories and methodologies; probably no linguist is fully acquainted with all o f them and certainly no attem pt can be m ade to represent each o f them here. I shall work m ainly on the assumptions w hich are com m on throughout the world o f linguistic study, and i f some ways o f approach are particularly useful for literary texts they can be used without prejudice to their status in that world. Sam uel Johnson offered two reasons for not issuing a list of subscribers to his edition o f Shakespeare: one was that he had lost all the names, the other that he had spent all the m oney. I can offer two different but equally cogent reasons for the fact that most o f the examples used in this book are drawn from English literature. O ne is that they w ill be fam iliar, or at least comprehensible, to everyone who can read the book. T h e second is, very sim ply, that English is the only literature with which I can claim a m ore than superficial acquaintance. But I believe that the approaches will be valid, mutatis mutandis, for other literatures as well. O ne more caveat before em barking on the real business. T h e end-

Allies or Opponents?

7

product o f literature, the text, is always capable o f linguistic in­ vestigation. By definition, literature is the art that uses language. T h e starting-point, however, m ay be quite different. It m ay be a historical occasion, an em otional experience, a desire to rebuke and reform society. It m ay be a pattern o f m etre, or a sequence o f sounds, or a collocation w hich cannot be regarded as an analysable linguistic utterance until it has been developed, or an im age first presented to the visual sense. T e d Hughes has thus described a poem: A n assembly o f living parts m oved b y a single spirit. T h e living parts are the words, the images, the rhythms. T h e spirit is the life w hich inhabits them w hen they all work together. It is impossible to say w hich comes first, parts or spirit.1 Spirits cannot be confined: but they can be investigated.

F U R T H E R R E A D IN G

A lthough this book is intended to give no difficulty to those who are not linguistic specialists, it makes no pretence to be a course in linguistics. Readers to whom the subject is com pletely new w ould do well to read D avid Crystal, What is Linguistics? (2nd edn., London, 1969, Edw ard Arnold) and to follow it with C rystal’s longer book Linguistics (Harmondsworth, 1971, Penguin Books) or F. Palm er, Grammar (Harmondsworth, 1971, Penguin Books); both o f the latter deal w ith I C (Im m ediate Constituent) Analysis which was referred to on p. I. Those who w ant to go beyond these introductory works w ill probably be taking a course o f instruction and receiving advice on reading; for any who are relying on private study, the appropriate book is J. Aitchison, Teach Yourself General Linguistics (London, 1972, English Universities Press). T h e outstanding figures in m odern linguistics can be approached through their own work: F . de Saussure, Cours de Linguistique generate (1916); English trans­ lation, Course in General Linguistics (London, i960, Peter O w en). E. Sapir, Language (N ew Y ork, 1921, H arcourt Brace). L .Bloom field, Language (London, 1935 and 1950, A llen and U nw in). 1 Poetry in the Making (London, 1967, Faber and Faber), p. 17.

8

Linguistics and Literature

N . Chom sky, Syntactic Structures (The H ague, 1957, M outon) and Aspects o f the Theory o f Syntax (Cam bridge, M ass., 1965, M assa­ chusetts Institute o f Techn ology Press); Chom sky’s books, especially the second, are not easy and should not be attem pted before reading one or m ore o f the introductory works mentioned above. A good simplified explanation o f his theories is J. Lyons, Chomsky (London, 1970, Collins). T h e issues raised in this chapter are discussed b y G; Steiner, ‘Linguistics and Literature’ (Minnis, pp. 113-36); М . A . K . H alliday, ‘T h e Linguistic Study o f L iterary T exts’ (Chatman: Essays, pp, 21723); R . Jakobson, ‘Linguistics and Poetics’ (Chatman: Essays, pp. 296-322); R . Fowler, ‘ Linguistic T h eo ry and the Study o f Literature’ (Fowler, pp. 1-29). Readers with special interest in English studies who w ant a linguistic approach without going too deeply into general linguistics should read R . Q uirk, The Use o f English (London, 1962, Longm an), followed b y H . A . Gleason, Linguistics and English Grammar (New York, 1965, H olt, Rinehart and W inston).

2

Literature and Stylistics

T h e w ord ‘language’ is not easily defined without reference to the context in w hich it is used. Consider these four sentences, each o f which could be constructed and understood w ithout difficulty b y a native speaker o f English: A ll hum an beings possess the power o f language. L atin is a synthetic language. W e w ere delighted b y the old m an’s hom ely language. Contracts should be draw n up in proper legal language. It is clear that in each o f these sentences ‘language’ is m ade to cover a different area o f reference. This kind o f indeterm inacy causes little trouble in everyd ay conversation but is unacceptable in serious linguistic study where a more precise division o f the area is necessary. Linguists have generally adopted the term inology o f de Saussure, who used langage to describe the faculty o f hum an speech in general, langue for the totality o f a particular language-system and parole for an act o f speech b y an individual user o f that system. W e need not consider all the axioms and definitions on w hich he based his theory. T h ere is no problem in accepting langue as the total resources available to anyone who is a m em ber o f a speech-comm unity and shares w ith others a system w hich can be given a nam e such as English, or French, or K o rean , or T agalog. Inform ation about langue is collected in dictionaries, grammar-books and studies o f pronunciation. T h a t w hich is observed for the com pilation o f these books is parole— or rather, enough examples o f parole to allow the form ation o f general rules. A similar distinction is that m ade b y Chom sky between ‘com petence’ and ‘perform ance’, and the need for these distinctions becomes apparent as soon as w e think seriously about the subject. Saussure’s descriptions could be respectively substituted for the

io

Linguistics and Literature

w ord ‘language’ in each o f the first three sentences above. T h r fourth sentence uses ‘language’ with reference not to any specific utterance but to an abstraction within the wider system. T h e notion o f ‘legal language’, ‘m edical language’, ‘religious language’ is fam iliar enough, but w hat exactly does it mean? T h e abstraction is composed from a large num ber o f paroles in w hich certain linguistic features recur with high enough frequency to be significant. Each of these ‘languages’ is unquestionably part o f a certain langue, showing enough com mon features to be intelligible in general pattern, if not in every detail, to most users o f that langue. N o special linguistic skill is required to pick out some o f these recurrent features. In English legal documents, for instance, w e m ay be struck b y the high incidence o f conjunctional phrases— ‘without let or hindrance’, ‘ the messuage or dwelling-house’, ‘the last will and testament’ ; and a little work w ith the dictionary w ill show that these often contain a word o f O ld English origin linked with one of R om ance origin. W hen w e speak o f ‘religious language’ w e are probably thinking o f the Book o f Com m on Prayer and the A uthorized Version o f the Bible. H ere w e find similar linked phrases— ‘when w e assemble and m eet together’, ‘sore let and hindered’, ‘confirm and strengthen’ . In the gram m ar o f this ‘language’ we find the second person singular pronoun ‘ thou’ , w ith its oblique forms and the associated verbal inflection -est which is obsolete in most types o f present-day English. Sim ilar observations can be m ade about m ore widespread com­ munications. T h e average academ ic lecture w ill contain few or no direct imperatives, but a higher proportion o f com plex sentences than w e should expect to hear in a conversation on the bus. News commentators on the broadcasting m edia have adopted the use of the simple present for a proxim ate future— ‘at the end o f this bulletin we talk to a correspondent’, ‘we hear the views o f people in the streets’ . T h e basis o f such observations is the choice o f certain linguistic features in place o f others. From recognizing a greater frequency o f these features than is found in other types o f utterance, we can go on and analyse enough specimens to allow the form ulation o f a de­ scription. T h a t which is described will be something that the non­ linguist recognizes without analysis; something that can be parodied and im itated. T h e set o f features w hich is accepted as fully appropri­ ate in one situation m ay seem com ic or distasteful if it occurs in another.

Literature and Stylistics

n

Mislead o f talking about ‘legal language’ and so on it is better to і їїII these distinctive usages styles. T h e notion o f ‘legal style’ or 'litigious style’ is, like all other attempts to categorize language, m.tele possible b y the perform ance o f users. W hen a user directs his I» rl'ormance towards a particular style, he is adopting a register. I'lie adoption o f register m ay be deliberate and w ith awareness o f a і (-cognized style, as when a barrister speaks in court. W hen the »,ime barrister speaks to his small children at home he w ill use a illllerent register, one w hich is less form alized and more instinctive. (Most parents use a special register for children, despite frequent IMolestations to the contrary.) Almost every individual has com mand o f a num ber o f registers which he uses in different situations o f his life— a t work, at home, willi friends o f his own age and sex, speaking at a public m eeting >ind so on. Choice o f register is constrained by the circumstances o f • ummunication rather than by the content. M an y native users o f Knglish w ould give the same message in different forms according in their relationship w ith the recipient, for exam ple: W e hope to arrive at approxim ately four o’clock. W e’ ll be there about four. W e’ll turn up fourish. The com mon adoption o f a register by a num ber o f people in a certain recurring situation creates a style. A n established style m ay make the use o f appropriate register obligatory. It should be m ade clear that this use o f the w ord ‘register’ is quite different from its use in music and phonetics to denote the scale o f pitch covered by an instrument or voice. T h e linguistic study o f different styles is called stylistics. T h e foregoing description has shown that styles are the product o f social situation: o f a common relationship between language-users. Stylistics is thus a part o f sociolinguistics— language studied in relation (o society. Sociolinguists are interested in the effect upon language o f speakers’ groups according to ethnic, social, class or other divisions. Stylistic features m ay derive from more tem porary associations as well, those w hich concern a speaker’s working or leisure time only. Hut every style is used for com m unication w ithin a group, large or small, close-knit or scattered, w ith features w hich are accepted as com m unicative by members o f the group. Now this is clearly something different from the use o f ‘style’

12

Linguistics and Literature

in more traditional approaches to language. Literary critics and commentators on the quality o f written language have m ade us so fam iliar w ith a certain conception o f ‘style’ that it m ay be necessary to m ake a deliberate change in our thinking. ‘ Good style’, or some­ times sim ply ‘style’, has been used as a description o f w riting that was in some w ay praiseworthy, skilful or elegant. Q uiller-C ouch rem arked, ‘Style in w riting is m uch the same as good manners in other hum an intercourse.’1 This kind o f evaluation need not be dismissed as unintelligent, but it is not the concern o f linguistics to m ake these judgem ents. Its use to the student o f literature is to provide techniques on which evaluative judgem ent can be based. In linguistic analysis, however, ‘style’ is not an ornam ent or a virtue; it is not something to be characterized as ‘good’ or ‘b ad ’ in any absolute sense. N or is it confined to written language, or to literature, or to any single aspect o f language. T h ere is no use o f language that is not open to stylistic investigation. Some areas, however, are richer in m aterial for such investigation because they show a high incidence o f special features. Instead o f a dogm atic evaluation o f any linguistic specimen as ‘good’ or ‘ bad ’, it is m ore reasonable to consider to w hat style it belongs and then to ask whether its features are appropriate to that style as com m only observed. W e return to the point that the same referential content m ay be expressed in different ways. I f someone says, ‘T h e sun is rising’, w e accept his statement as one appropriate to most conversational situations. W e accept the message and are unlikely to give conscious attention to the m ode o f com munication. Y e t there are other ways o f expressing the same point w hich w ill m ake us more aware o f m anner as well as m atter. T h e diffused daylight w hich precedes the passage o f the sun above the horizon is due to refraction, reflection, and scattering o f the light o f the sun b y the atmosphere. But look, the m orn, in russet m antle clad, W alks o’er the dew o f yon high eastern hill. In m eeting either o f these statements, we recognize the presence o f something more than the ‘com m on core’ o f the language. Either o f 1 A. T . Quiller-Couch, The Art of Writing (Cambridge, 1916, Cambridge University Press).

Literature and Stylistics

13

1linn would probably cause m irth or embarrassment if it came out in colloquial conversation. T h e first, however, w ould be acceptablc in a style generally described as ‘scientific’ ; the second is recognized, with detached tolerance or deep involvement according to personal predispositions, as ‘literary’ . Recognition o f style is prim ary; other questions m ay follow. Stylistics, then, is not confined to literature, although its applica­ tion to literature is the concern o f this book. In some respects, literature is the most difficult type o f language to approach stylistic­ ally, because o f the diversity and com plexity which will appear in 1lie course o f investigation. T h e difficulty is more than compensated l>y the special value o f the m aterial being investigated. A t this stage we m ay indeed adm it a concept o f quality and excellence, which can he scrutinized by the m ore objective criteria o f linguistics. Y et can we really think o f literature as a style? Is there a discernible Iі terary style, as there is a legal or liturgical style? O bviously literature is not confined to any aspect o f hum an experience, nor does it exclude any. W ithin a given langue, any parole could be incorporated into the literature using that langue. Literary style is not something to be described by a few salient characteristics; but careful study o f literary texts will show that literary stylistics is a viable study. Like all m eaningful use o f langue, literature contains a great deal of ‘common core’ w hich w ould cause no surprise in any situation. It also contains a higher incidence o f special or deviant features than non-literary styles. Between these extremes, it is possible to observe that literary style shows more careful and consistent use o f the regular patterns o f the language: the ‘ rules’ o f traditional gram m ar, which drew examples from literature because o f its regularity as m uch as its prestige. As w e shall see, m uch o f the most striking literary language appears deviant when it is really using, w ith singular economy and compression, the resources available to all native speakers. T o say that literary language is more careful is another w ay o f saying that it is m ore conscious in formation. Literature uses language as an artistic m edium , not sim ply for com m unication or even expression. It is not spontaneous, w hatever theories o f spon­ taneous inspiration m ay sometimes have been canvassed. I t is considered and developed in a w a y that is impossible for everyday conversation, or even for the more deliberate registers adopted for certain styles. W hile other styles show recurrent features, literature is dis­ tinguished b y w hat can be described overall as pattern. T h e text will

14

Linguistics and Literature

show selection and arrangem ent o f items that contribute to the total cffect; elements that w ould be absent or incidental in other styles are im portant for the fulfilm ent o f purpose. Poetry shows such patterning devices as metre, rhym e, assonance, alliteration; prose m ay contain similar devices, less regularly arranged. Both types o f literary discourse w ill have careful and often unexpected selection o f words and syntactic constructions. Figures o f rhetoric w ill give unusual prom inence to certain items. W e m ay therefore add a third to the two distinguishing marks o f literature suggested in the previous chapter: the use o f special devices w hich heighten the effect o f linguistic acts through patterning. Literature m ay be m uch m ore than w ould norm ally be understood as a ‘style’, but there is value in attem pting to treat it as one. A n y profitable approach through linguistics must deal with literature as an exam inable part o f the available realization o f langue. Special, heightened and prestigious as it m ay be, it cannot deviate too far from the expectations o f the speech-comm unity i f it is to find any readers at all. Such deviations as occur can be discerned and de­ scribed by methods applicable to more fam iliar and hum bler paroles. L ike other styles, it has features not shared by all users at all levels; but, as in all styles, these features can be utilized only in association with ‘com mon core’ features. Extremes are generally dangerous and distorting. In some periods and cultures literary language has gained such prestige that other styles have been jud ged good or bad according to their resemblance to it. Non-literary users have tried to incorporate literary features into personal com m unication. T h e developm ent o f national languages has been affected b y the prestige o f a dialect used for literature; individual writers have left their m ark on common speech. This kind o f influence, by no means undesirable in itself, has had the unfortunate effect o f isolating literature from regular methods o f investigation. T h e other extreme, w hich has already been mentioned, was the dismissal o f literature by some m odern linguists as too deviant for their attention. It is now generally accepted that any description o f a langue must take account o f all its different realizations. T h e present investigation sees literary style as deriving its strength from the ‘com m on core’ , even in respect o f features w hich are usually thought to be distinctively ‘literary’ . Its deviations do not break down com ­ m unication with ‘com mon core’ users. C ertainly it is sometimes necessary to m ake the kind o f adjustment or allowance which in the

Literature and Stylistics

/5

p.ist was vaguely named ‘poetic licence’ . A t the same time, the very extension o f literature into all aspects o f hum an experience means 11 iiit its style is less exclusive than some others w hich are the preserves of smaller groups. T h e literary writers have themselves been divided about the par­ ticularity o f their style. T h e course o f English poetry, for instance, shows a succession o f swings between belief in the special nature of poetic diction and insistence that the criterion o f poetic greatness was its closeness to everyday speech. A t the end o f the sixteenth century, Spenser tended to the first view and Donne to the second. Restoration poets in turn reacted against the school o f Donne, which seemed to have becom e artificial and remote from real life. M ost eighteenth-century poets shared the opinion o f G ray that ‘ the language o f the age is never the language o f poetry’, until W ords­ worth cam e with his standard o f poetry written from ‘a selection o f language really used by m en’ . O ne task o f literary stylistics is, w ithout taking sides in this dispute, to determine how far and in w hat respects a poet’s language in fact shows deviant features. Another is to note how a writer uses generally accepted features to special effect. It is necessary to p ay close attention to particular writers, since literature shows far more diversity o f individual usage than do other styles. This fact creates a link between modern stylistics and the m ore traditional w ay o f discussing a writer’s style in the sense o f the too-often quoted dictum o f Buffon— ‘L e style est l ’homm e meme’ . Although we have tried to postulate a literary style, as parallel to a legal, m edical or religious style, it is apparent that we are led (o something that is far from being homogeneous. In practice, the examination o f a single writer, or o f the com m on features o f a school or literary period showing com m on aims and influences, w ill yield the most satisfactory results. W hatever can realistically be said about literary style as a whole is w orth saying as a contribution to critical theory and to understanding o f how language works. In the present study, examples from particular writers w ill be used as evidence for general principles, but the aim o f the whole is to prepare readers for close stylistic study o f m ore lim ited areas. U nlike other styles, literature does not and cannot exclude any aspects o f langue. H ere the notion o f register is im portant to stylistics, as w e approach the w riter as an individual user o f resources poten­ tially available to other members o f the speech-community. Like other users, he is in a relationship o f com m unication, though one o f

іб

Linguistics and Literature

m onologue rather than dialogue. T h e reciprocal attitudes o f writers and readers towards one another is an im portant part o f the sociology o f literature and n aturally enters into stylistics. E ach writer w ill in fact choose registers according to varyin g factors in the situation. T h e lim iting factors m ay be mentioned first, although they generally produce less interesting results. Like other styles, literature m ay constrain an individual to adopt a particular register. Pre­ scriptive forces tend to operate in literary culture, not always o f the same kind or o f the same intensity. Even periods o f revolt, like the Rom antic movement, bring their own inhibitions and pressures to replace the old. T h e totally permissive society is no more a reality in literature than in hum an organization as a whole. Sometimes the pressures are more overt and more clearly codified, as in the negative attitude to poetic diction in the eighteenth century, which objected to certain words and phrases as foreign, technical, or too com m on­ place in their associations. T h ere was a positive pressure as well, in the attem pt to prescribe words specially or solely suitable for poetic expression. Johnson tended towards the first attitude and Addison towards the second, while Pope fought in both camps. T o some extent, then, the critics and writers o f literature m ay create a style w hich demands that a certain register be adopted, others avoided, by those w ho seek acceptability within the group o f style-conformists. T h e rules seldom endure for a long period o f time, since those who break them can equally claim to have incorporated their register into a new literary style. W hether his attitude to prevailing fashion be one o f acquiescence or o f rebellion, however, a w riter is unlikely to stick to any single register for the whole o f his literary output. It is the business o f literary stylistics to recognize and exam ine the different registers encountered. A w riter w ill perhaps change in this as he changes his attitude to the currently prestigious literary style. T h ere w ill be other reasons too for the adoption o f different registers. H e m ay be aware o f addressing different groups o f readers, as a speaker changes register in m oving from one set o f acquaintances to another. Examples can be seen in books written specifically for children by authors who have generally worked at the adult level. Com pare the narrative o f K ingsley’s The Heroes or o f D ickens’s Child's History o f England w ith that o f their novels, and note such features as m ore direct secondperson address, shorter sentences, a smaller range o f lexical items. Sometimes the difference o f register w ill be linked w ith difference o f genre or literary kind, as when D onne’s sermons are com pared w ith

Literature and Stylistics

iy

his poetry. T h e sermon, although ‘literary’ in some features, w ill draw from the religious style and from the oratorical. Its direct address is different from the assumption o f an interlocutory personam the lyric. Y e t this latter assumption, widely adopted by poets, brings register-change. T h e im aginary critic or questioner m ay be m ade to speak in a w ay that heightens b y contrast the poet’s own address to the reader. Such are the interpolations o f ‘A rbuthnot’ in Pope’s Prologue to the Satires and the hearty indictm ent that begins Housm an’s apologia in no. L X I I o f A Shropshire Lad: ‘Terence, this is stupid stuff: Y o u eat your victuals fast enough; T h ere can ’t be m uch amiss, ’tis clear, T o see the rate you drink your beer. But oh, good Lord, the verse you make, It gives a chap the belly-ache . . .’ and so on, with colloquial contractions o f auxiliaries, near-cliches such as ‘fast enough’ , lexical items like ‘belly-ache’ seldom found in the literary style, and reference to physical activity. A t last the poet starts to reply in the same register and then turns to a more formal narrative conclusion without second-person reference w hich lifts his poetry as a whole into a different world from the commonsense standard o f ‘victuals’ and ‘beer’ : T here was a king reigned in the East: There, w hen kings w ill sit to feast, T h ey get their fill before they think W ith poisoned m eat and poisoned drink. H e gathered all that springs to birth From the m any-venom ed earth . . . I tell the tale that I heard told. M ithridates, he died old. A sim ilar register-change m ay be observed in the novel written in the first person. Consider the ‘adult’, sophisticated and ironical register o f the narrator in David Copperfield, contrasted with the child’s speech o f the rem embered boy-D avid; or the m ature Butler’s persona O verton reporting the conversation o f Ernest Pontifex as child and undergraduate. T h e novel and the dram a must use a num ber o f registers since they still ‘contain’ characters conceived as beings o f separate existence

i8

Linguistics and Literature

com m unicating in an im aginary but recognizable society. It is no novelty in literary criticism to study how the most skilful writers clearly differentiate the speech o f their characters, while the incom petent or unpractised m ake no significant variation; stylistic study considers how these effects are made. T h e question o f written dialogue introduces the idea o f dialect— the distinctive system o f a group o f users o f a langue, with common regional or class identity. Here, w riting so clearly interprets speech that m ore consideration must be given to it in a later chapter. Finally, the speeches attributed to a character in literature w ill yield samples o f an im aginary idiolect— the choice from langue m ade by an individual at a given stage in his life. T h e stylistic investigator m ay com pare a fictional or dram atic idiolect with examples drawn from real speakers o f sim ilar age, class, education or region, and m ay thus offer some serious evidence for the frequent discussion o f whether a character is or is not ‘ true to life’ . A fruitful subject for stylistic study is the deliberate m ixing of registers w ithout clear identification o f speakers; the shift o f utterance is m arked b y shift o f register and not by any extraneous pointers. Such m ixing, generally frowned upon in the past as an offence against decorum, is w idely used in recent and current literature. I f we met the ‘scientific’ description o f dawn, previously quoted, in a work w hich we had accepted as ‘literary’, w e m ight suppose a deliberate contrast for ironic or other effect. For the modern reader is habituated to the raiding o f other registers, w hich Joyce carried out in Ulysses and Eliot in The Waste Land. T h e com plexity o f society, the uncertainty o f personal identity, the realization o f coinherence in a com m on hum an predicam ent— all this can be suggested through the changing viewpoints shown b y selection o f different register-features. A n exam ple o f m ixed registers, which has been noticed by other critics but is too good not to use, is H enry R eed ’s poem , ‘Nam ing o f Parts’ . H ere no typographical devices show the transition to and fro between the words o f the arm y instructor and the thought-observations o f the poet-recruit. A ll is done by linguistic placing and selection, with use o f the cliches from the teaching m anual fam iliar to thousands o f soldiers, sw itching to the kind o f language expected from a literate and im aginative user. T h e sexual im agery drawn from the am bivalence in some o f the instructional items points a contrast between the sterile destructiveness o f w ar and the natural life-force. T h e first and the last two stanzas must serve to illustrate:

Literature and Stylistics

ig

T o d a y we have nam ing o f parts. Yesterday, W e had daily cleaning. A n d tomorrow morning W e shall have w hat to do after firing. But today, T o d a y we have nam ing o f parts. Japonica Glistens like coral in all o f the neighbouring gardens, A nd today we have nam ing o f parts. . . . A nd this you can see is the bolt. T h e purpose o f this Is to open the breech, as you see. W e can slide it R ap id ly backwards and forwards: we call this Easing the spring. A n d rapidly backwards and forwards T h e early bees are assaulting and fum bling the flowers: T h e y call it easing the Spring. T h ey call it easing the Spring: it is perfectly easy I f you have any strength in your thum b: like the bolt, A nd the breech, and the cocking-piece, and the point of balance, W hich in our case w e have not got; and the almond-blossom Silent in all the gardens and the bees going backwards and forwards, For today we have nam ing o f parts. T h e full potential o f the langue can be indicated in a short extra­ polation from different registers. T h e critic who uses stylistic method needs to be aw are o f the potentiality and to be able to recognize the styles and registers available to the literary writer. A part from the kind o f proscription imposed b y convention, the writer is free to select from langue where he w ill, aided by his peculiar sensitivity to the use o f language. T h e critic should try to respond with equal freedom,- with as m uch sensitivity as he can claim , and with knowledge o f the basic m aterial from which literature is made.

FURTH ER

R E A D IN G

A n exam ination o f w hat is m eant by ‘style’ in modern linguistics could well begin with N. E. Enkvist, J. Spencer and M . J. Gregory, Linguistics and Style (London, 1964, O xford U niversity Press). T h e principal styles o f present-day English are surveyed in D. D avy and D. Crystal, Investigating English Style (London, 1969, Longm an). O n register, dialect and idiolect see also C . Barber, Linguistic Change in Present-day English (Edinburgh, 1964, O liver and Boyd),

20

Linguistics and Literature

pp. 16-33; A- E. D arbyshire, A Description o f English (London, 1967, Edward A rnold ), pp. 1-29 and 159-77; R . Q uirk, ‘Linguistic, U sage and the U ser’ (Minnis, pp. 297-313). P. G uiraud, La Stylistique (5th edn., Paris, 1967, Presses U niversitaires de France) is an interesting introduction b y a French linguist. Readers w ho w an t to go m ore deeply into sociolinguistics should start with P. P. G iglioli, ed., Language and Social Context (Harmondsworth, 1972, Penguin Books). Discussions o f the specific nature o f literary language are m any; the present-day debate owes m uch to I. A . R ichards, Principles o f Literary Criticism (London, 1924, R outledge and K e g a n Paul). A n im portant though less easily accessible book is I. H ungerland, Poetic Discourse (Berkeley and L os Angeles, 1958, California U niversity Press). V a lu a b le shorter discussions are Nowottny, pp. 1-25; Leech, pp. i - 19; B. Lee, ‘T h e N ew Criticism and the Lan gu age o f Poetry’ (Fowler, pp. 29-52); A . L . Binns, ‘ Linguistic R eading: two suggestions o f the qu ality o f literature’ (Fowler, pp. 118-34); J- G. Ransom , ‘W anted: an O ntological C ritic’ (Chatman: Essays, pp. 269-82). T raditional views o f poetic diction are surveyed b y F. W . Bateson, English Poetry and the English Language (Oxford, 1934, O xford U n i­ versity Press) and S. A . L eonard, The Doctrine o f Correctness in English Usage (M adison, 1929, W isconsin U niversity Press). O nce the new understanding o f ‘style’ has been reached readers w ill lose nothing and m ay w ell profit from reading some o f the older, m ore prescriptive works such as H. W . and F. G . Fowler, The King's English (Oxford, 1906, O xford U niversity Press) and A . T . Q uiller-Couch, The Art o f Writing (Cam bridge, 1916, C am ­ bridge U niversity Press). A m odern linguistic approach to correct usage is J. W arburg, Verbal Values (London, 1966, Edw ard Arnold). A good recent intro­ ductory book is G. W . Turner, Stylistics (Harmondsworth, 1973, Penguin Books).

з Language, Literature and History

W hen applicants for a university course in linguistics are asked about their motives they very often say that they are interested in the developm ent o f language. T h ey w ant to learn more about, ‘how words change their m eanings’ , ‘the history o f English gram m ar’, or ‘ the influence o f A nglo-Saxon’ . T h ey have to be gently told that these matters are not am ong the prim ary concerns o f m odern linguistics. In nineteenth-century philology, historical study played a large part; the reaction against it was vigorous and perhaps m ore sweeping than we should wish to perpetuate. O n ce again we can turn to Saussure for a fundam ental explana­ tion. H e separated the diachronic study o f language, which traces developm ent through the past, from the synchronic view o f the total state o f a langue at a given point o f time. T h e chosen point is usually that o f present-day observation, and some linguists claim to have practically no concern w ith w hat happened in the past. Although historical linguistics is a respectable branch o f the subject, it is a com paratively minor one. Y e t literature comes to us m ainly from the past. T h ere are good academ ic grounds for not attem pting to study the literature o f our own tim e without some knowledge o f that w hich has gone before. A n exam ination o f literary style as it appears at the present d ay w ould certainly be both interesting and revealing, but if literary stylistics is to be o f deeper value educationally or in personal appreciation, it must be effective for the literature o f all periods. T h e stylistic study o f a w riter or school w ill be synchronic in its concern for total performance using the linguistic code available at the time. D iachronic considerations must enter as w e look back from our present position: the code by which w e formulate our reactions to literature and verbalize our judgem ents is not identical with the code understood by Chaucer, Shakespeare or Dryden, or even Browning.

nil

Linguistics and Literature

T h e apparent conflict is not irreconcilable. Even Saussure recog­ nized the intersection o f synchronic and diachronic in every linguistic act. H is concern was not to abolish diachronic study, but to avoid confusion o f the two approaches. W hen w e examine a literary text w e are m aking a double extrapolation— from the timeaxis o f historical developm ent and from the performance-axis o f all accessible linguistic acts m ade at that point o f time. Questions based on our know ledge o f both axes must be asked; description o f performance requires knowledge o f code. T h e critic needs to think clearly about this intersection. H e w ill use the tools o f m odern linguistic scholarship w ithout supposing that they were som ehow present in the mind o f a past writer— a sup­ position w hich is clearly absurd as soon as it is formulated but which has a w a y o f being insidiously troublesome beneath the surface. T h ere is a possible analogy in the history o f prices and wages, related to currency changes. O ld records o f paym ent can be con­ verted to m odern decim al currency w hich w ill give the student an idea o f w hat things used to cost. Effective conversion needs two quite different skills. First, there is the ability to work in the old currency and see whether the accounts are accurate in their own terms— to discover whether the m an who left the record had got his sums right. Secondly, the methods o f econom ic history can help to relate form er prices and wages in real terms and to show their significance for the wider considerations o f society at that time in the past. W orking with the old currency is like using diachronic knowledge. A lthough linguistic performance does not yield to notions o f right and wrong like a set o f accounts, the accepted usage o f society in any period produces its own rules. W e need to know, as it were, the ‘value’ o f words at a given time; the semantic equivalent in present-day speech m ay be discovered, as groats and shillings can be converted into decim al currency. T h e popular interest in ‘how words change their m eaning’ in fact represents a real concern o f stylistics. Sem antic change can cause serious misunderstanding o f w hat a writer was in fact trying to say. Evaluation o f performance will not be helpful if it proceeds from the wrong starting-place. W ords in isolation are dangerous traps for aspiring linguists. Semantics is a difficult and still experim ental branch o f study, which finds little profit in single lexical items. Y e t there are certain conversions that have to be m ade before closer analysis can begin.

Language, Literature and History

23

When Johnson wrote o i Lycidas that it was in ‘ the form o f a pastoral: easy, vulgar and therefore disgusting’, his opinion was literary and not moralistic. ‘ Easy’ must be converted to ‘over-simple’, ‘vulgar’ to ‘com m onplace’ and ‘disgusting’ to ‘distasteful to refined literary sensibility’, before any further com m ent can be attem pted. O r when Sam W eller enquires about his ‘m other-in-law’ w e m ay waste time looking for evidence that he was m arried unless we can make the im mediate conversion to ‘stepm other’ and thus identify the character to whom he is referring. D iachronic information is not confined to a glossary o f changes in meaning. A glossary will be prim arily concerned w ith the denotations of words. F ull comprehension, in all styles and especially in litera­ ture, depends on grasping the connotations— the em otive ambience of words, their associations and the emotions which they m ay arouse. It is simple enough to explain thou as the second person singular pronoun, subjective case, and to add that it is now archaic except in a few special registers. But this does not take us very far in a study o f Twelfth Night when the text gives the advice o f Sir T o b y to Sir A ndrew about how to convey a challenge to V io la , disguised as the young m an Cesario: T a u n t him with the licence o f ink: if thou thou’st him some thrice, it shall not be amiss. (III. ii) This seems very curious, so long as it is read only at the level o f traditional gram m ar. Its effect depends on the social connotation o f thou in that period. Sir T o b y rightly uses it in speaking to his friend and equal Sir Andrew , but it w ould be a sharp insult for the latter to use it in w riting to a mere acquaintance who seemed to be m oving on the same social level. But the second thou is placed and inflected as a verb: is this some bold and unique device o f Shakespeare’s, w hich m ight seem too im aginative for the character o f Sir T oby? Recourse to other contem porary evidence reveals both connotation and usage. W hen Edw ard Coke was conducting the prosecution o f R alegh for treason, he m ade the same insult in the same w ay: A ll that he did was at thy instigation, for I thou thee, thou traitor. A nd Stubbes related o f his wife that:

if,/

Linguistics and Literature She was never heard to give the He nor so m uch as to thou any ill anger.

I ii lliis instance we have looked along the horizontal axis o f l .li/.ibclhan and Jacobean English to establish a basis for criticism o f a specific text. It is possible also to trace a particular item along the vertical axis o f historical development. For instance, the form o f the negative im perative norm ally used at the end o f the sixteenth century is frequently exemplified in Shakespeare, as: N o, faith, m y coz, wish not a m an from England {Henry V, IV . iii) I love thee not, therefore pursue m e not (A Midsummer Night’s Dream, II. ii) By 1700 the auxiliary do was in general use for this construction, having grown from rarity (it appears a few lines after the first o f the above quotations) to dom inance. This change gives a standpoint for the criticism o f K eats’s Ode to Melancholy: No, no, go not to Lethe and o f Housman:

k

T ell me not here, it needs not saying. As time passes, the older construction becomes more remote so that its use in poetry is a conscious choice by the poet w hich stands apart from everyday discourse. Its function is no longer straightforward but suggests literary artifice b y the exploitation o f archaic usage w ith its traditional associations. Exam ples from one m ore field o f linguistics m ay serve to establish the need for a historical approach. K now ledge o f phonology as a diachronic study will often correct false conclusions about rhymes w hich are no longer ‘good’ rhymes according to present-day Received Pronunciation: I ’ll tune thy elegies to trumpet-sounds A n d write thy epitaph in blood and wounds. (Montrose) H ere thou, great Anna! whom three realms obey, Dost sometimes counsel take— and sometimes tea. (Pope)

Language, Literature and History

25

In these and a great m an y other instances the poet produced a perfect rhym e. Change in the pronunciation o f one o f the rhym ing words makes it appear to the m odern reader as i f there had been an incom petent or deliberate lapse from the pattern o f the poem. A lthough they are o f little im portance in the mainstream o f present-day linguistics, these considerations are o f the first im port­ ance to the student o f literary stylistics. T h e existence o f new and exciting developments in the subject as a whole m ay endanger the status o f slightly less up-to-date approaches. D iachronic study does not divert research into unprofitable channels provided its limits are defined and observed. In literary stylistics, the linguist calls on the aid o f historical research, ju st as in other branches o f linguistics he calls on sociology or psychology or anthropology. U sing the resources o f history in stylistic criticism confronts us with another and m ore closely linguistic question. W hat status do we afford to literary texts in relation to the language as a whole? C oncentration on one style has the danger that it w ill come to be seen as the style with particular prestige by which others are to be judged. U n til recently the literature o f a culture was seen as the highest linguistic usage, as the level to w hich all utterances should aspire and b y which they could be somehow graded. A ppeal to the ‘best authors’ was the justification o f lexicographers, grammarians and those who sought to teach ‘style’ in the old sense. T h e appeal was especially strong in Britain, where the absence o f any kind o f A cadem y m eant that a w eight o f literary usage could be taken to justify a prescriptive rule. G reat writers were, am ong their other excellences, models for im itation. T h e notion o f a model in this sense— as one m ight speak o f a ‘model answer’ to an exam ination question— is not congenial to the spirit o f m odern linguistics. A m ore useful approach is the attem pt to create a m odel in the m ore philosophical sense o f a construct w hich adequately represents reality and serves for the explanation and evaluation o f specific instances. It is possible to describe a m odel o f language in this w ay, as the norm o f usage acceptable to the native speakers in a given com m unity at a given time. D u e allowance is then m ade for the distinctive usages o f different styles which depart to a greater or lesser extent from the com mon stock. In this perspective, the language o f literature is often notably deviant. T h e concept o f deviation is an im portant one in stylistics and we shall return to it. It arises as soon as we set particular linguistic acts against the apparent norm and it appears in two ways. First, there

26

Linguistics and Literature

is the statistical deviation which w ould m ake H ousm an’s ‘T ell me not here’ a m inority usage in the corpus o f late nineteenth-century standard English. A deviant feature, lexical, syntactic or phono­ logical, can sim ply be noted as an infrequent item in the total. Stylistic study w ill seek to account for it and to ju d ge its effect w ithin the whole text. T h e significance o f deviation m ay depend on the precision or delicacy w ith w hich the parameters are drawn. T h e same archaic form o f the negative im perative would not be deviant in liturgical usage w hich perpetuates sixteenth-century E ng­ lish and yields such examples as ‘lead us not into tem ptation’ , ‘enter not into judgem ent’, ‘or else come not to that H oly T a b le’ . It is possible to recognize a second and m ore interesting kind o f deviance in w hat appears to be the novel and distinctive usage o f a particular writer. Shakespeare’s use o f the verb ‘spanieled’ could be statistically recorded as a single exam ple not found elsewhere in the surviving evidence o f Elizabethan English. W h at is more interesting is to regard it as an item devised for a particular context, using lexical, m orphological and syntactic methods perm itted by the norm but never before com bined in that w ay: the substantive spaniel, the past tense verbal inflection -ed and the relative sentenceposition generally given to a transitive verb, producing T h e hearts T h a t spanieled m e at heels, to w hom I gave T h eir wishes. (Antony and Cleopatra, IV .xii) T h ere is no great difficulty here, nor in m an y other examples o f deviation such as: T h e achieve of, the mastery o f the thing (Hopkins) O n ce below a tim e I lordly had the trees and leaves (Thomas) Y e t as w e go farther back in time, there is less and less evidence on which to base the judgem ent o f norm and deviation. T h e concept is less useful for older than for m ore recent literature because it becomes m ore difficult to construct a model. So m uch o f w hat has been preserved from earlier states o f the language is literary that we cannot cast a survey wide enough to say w h at was normative.

Language, Literature and History

27

I here is also a danger that the attitude to literary language which sees it as m ainly deviant can harden into a new kind of prncriptivism , w ith literature as a pathological condition o f the language and common speech as the healthy norm. A t one time it was possible to set exam ination questions asking for comments on such 'errors o f gram m ar’ by the masters as Shakespeare’s ‘Y oung Ferdinand, whom they suppose is drowned’, or Byron’s ‘T here let him lay’ . Nowadays the crim e is more likely to be that o f deviation from the lowest com mon denom inator. W ith these caveats in m ind, the idea o f deviation can be an extremely useful approach to at least some kinds o f literary language. I ii'viation is not a pejorative term if we m aintain an objective view o f the m any possible styles. A fter due observation it is indeed valid to adjudge the deviation o f the feeble-minded or the aphasic to be undesirable because it hinders com munication, and the deviation of the poet com m endable because it heightens awareness and under­ standing. A n y such judgem ents must depend on a synchronic picture o f language gained from adequate sam pling o f a given period. Therefore the student o f literary stylistics needs to go beyond the limits o f ‘literature’ and ‘non-literature’, w hich w e have already seen to be scarcely m eaningful. T h e evidence already quoted for the connotation o f Elizabethan thou is an example. Sim ilarly, features o f language which m ay appear to be specifically and solely literary may in fact be shared w ith other registers: sharing o f features is not to be confused with the kind o f deliberate borrowing from other registers which has been noticed. It is no doubt easy to find something stilted and artificial in Byron’s: T h e angel o f death spread his wings on the blast A n d breathed in the face o f the foe as he passed . . . (‘T h e Destruction o f Sennacherib’) or in Evelyn G reen’s: l’he angel o f death had not come alone— there was Another with him. {Only a Child) Surely this is remote from any other use o f language in the nineteenth century— totally deviant from w hat anyone could ever have spoken.

28

Linguistics and Literature

Y e t here is John Bright, speaking in the House o f Com mons in

«855: T h e angel o f death has been abroad through the land; you m ay almost hear the beating o f his wings. T h e interchange between literary and non-literary language is not confined to discrete features. W e are accustomed, certainly in European cultures, to think o f a ‘standard’ form o f a national language, local or class variants on it being regarded as dialects. In m any instances, however, the standard form is traceable back to a dialect which gained prestige through being the m edium o f literature. A period o f great literary production, with associated confidence in execution, can elevate a dialect to superior status and cause it to be adopted b y educated members o f the com m unity for written and— usually later— for spoken com m unication, w ithout area restriction. This status was gained by W est Saxon in preConquest England and by East M idland at the beginning o f the fifteenth century. T h e same kind o f thing happened to Francien in France, Castilian in Spain and Tuscan in Italy. In all these instances, econom ic and social factors also played their part; it can be argued that the literature o f these dialects owed its strength to favourable external conditions. L u th er’s choice o f Saxon can certainly be seen to have a direct influence on the developm ent o f Germ an, partly through the deliberate aid o f later writers and grammarians. W ithout over-simplifying a com plex question, it is im portant to realize that literature helps to affect the diachronic developm ent o f the langue which is its medium. Ernst Cassirer’s words are w orth remembering: N o poet can create an entirely new language. H e has to adopt the words and he has to respect the fundam ental rules o f his language. T o all this, however, the poet gives not only a new turn but also a new life . . . T h e Italian language, the English language, the G erm an language, were not the same at the death o f D ante, o f Shakespeare, o f Goethe, as they had been at the day o f their birth.1 Prestige, status and high quality are notions which are scarcely congenial to the m ore descriptive and analytical approach o f m odern linguistics, but w hich cannot be altogether excluded from 1 An Essay on Man (New Haven, 1944, Y ale University Press).

Language, Literature and History

2g

literary stylistics. Probably no one can m ake a useful study o f literature without some sense o f its excellence as a product o f the human mind. T h e value o f new methods depends on starting from the right place. W e do not say, ‘This is a piece o f work by a great writer which must be respected as a specimen o f the best use of l.itiguage.’ T h e consideration is rather, ‘K n ow in g w hat we do about language in general, and about the state o f this particular speechi (immunity at the time when this work was written, we are in a position to m ake an informed judgem ent.5 There is o f course no reason for excluding literary texts from any linguistic investigation or for refusing to treat them exactly the same was as other paroles, although some linguists w ould regard literary language as too deviant to yield much useful information. Linguistic­ ally, the literary style can be described w ithout offering any view .1bout its merit, just as legal style can be described without raising questions o f justice or liturgical style without raising questions of belief. H owever, a special interest in any style usually implies some commitment to the intrinsic im portance o f that which is spoken or written in that style. O ne other m atter is worth remembering while we still have historical considerations in mind. Linguistic theory, and interest in the nature o f language, is not an invention o f the twentieth century. Close attention was given to Classical Greek, L atin and Sanskrit by grammarians to whom they were living languages, and nearly every age has produced some views about how language works or should be made to work. T h e revolution in linguistics that has taken place in our own time is com parable to the Einsteinian revolution in physics. It has had the sim ilar result o f discrediting a great deal of what was formerly believed and m aking it impossible for intelligent use to be m ade o f the old approach as a basis for investigation. T here are, nevertheless, reasons for giving some attention to the history o f linguistic theory and for not closing the mind to all that has gone before. Literature was written b y m en whose view o f language was that o f their own age. This is not to say that all im agina­ tive writers were deeply concerned with linguistic theory or spent much tim e 'in earnest colloquy with contem porary grammarians. Y et no-one who uses an artistic medium can rem ain totally unaware o f the w ay in which that m edium operates in the com m unity as a whole. O n e specific point is that the m ajority o f English writers before the latter part o f the nineteenth century were educated m ainly through the study o f Latin, by schoolmasters to whom it was

jjo

Linguistics and Literature

axiom atic that English could and should be described in Latinbased gram m atical terms. Also, and this is a w orthy discipline for all whose study requires historical perspective, the fact that any theory can no longer be accepted totally does not m ean either that its originators were fools or that it contained nothing fit to be remembered. Literature can still be illum inated b y the judgem ents o f those to whom it was fresh, the latest expression o f im agination through language. O n e o f the greatest mistakes that can be m ade is to regard the past as homogeneous. It is easy to laugh at Lord M onboddo in the eighteenth century calling G othic ‘ the parent o f all the different dialects o f the T eu ton ic’ . Y e t everyone today must honour the wisdom o f John W allis in 1653 com plaining that gram marians had previously ‘forced our tongue too m uch into the pattern o f L atin ’ . Everyone— well, almost everyone. T h ere are still teachers o f English whose consciences should be touched by the words o f W allis. I f they venture the P layer’s excuse, ‘I hope we have reformed that indifferently with us’, they can be met w ith H am let’s admonition, ‘ Reform it altogether’ . N ot all the wisdom o f the past has been universally recognized. N ot all the attitudes o f the past need be eschewed by those who have the wisdom o f the present. T h ere are ideas w hich can be lifted out o f their shaky fram ework and fitted into the new. Abuses should never make us disregard right uses.

F U R T H E R R E A D IN G

Historical linguistics can be studied in W . P. Lehm ann, Historical Linguistics: an Introduction (N ew Y ork, 1962, H olt, Rinehart and W inston), or, in the fram ework o f Chom skyan linguistics in, R . D . K in g, Historical Linguistics and Generative Grammar (Englewood Cliffs, N. J ., 1969, Prentice-H all). T h e developm ent o f English is surveyed in G . L . Brook, A History o f the English Language (London, 1958, A nd re Deutsch), w hich includes an extensive bibliography; aspects particularly relevant to the present study are exam ined in W . F. Bolton, A Short History o f Literary English (2nd edn., London, 1967, E dw ard A rnold). H. C . W yld, Historical Study o f the Mother Tongue (London, 1906, John M urray) is an old but not obsolete application o f philology to the diachronic study o f English. G ood shorter studies are: K . D . U itti, ‘ Philology: Factualism and

Language, Literature and History

31

History’ (Chatman: Style, pp. 111-32); A . R odw ay, ‘B y A lgebra to Augustanism’ (Fowler, pp. 53-8); J. Norton-Sm ith, ‘ Chaucer’s Kpiatolary Style’ (Fowler, pp. 157-65); R . F. Law rence, ‘T h e I’ortnulaic T h eory and its A pplication to English A lliterative I’or try’ (Fowler, pp. 166-83); M - W . Croll, ‘T h e Baroque Style in Prose’ (Chatman: Essays, pp. 341-61); J. M iles, ‘ Eras in English Poetry’ (Chatman: Essays, pp. 175-96). T h e history o f ideas about language— which is not the same thing us historical linguistics— is covered b y R . H. Robins, A Short History o f Linguistics (London, 1967, Longm an).

Speech and Writing

Tennyson has left us a b rief word picture o f how his Morte d'Arthur was read aloud: T h e poet little urged, But with some prelude o f disparagement, R ead , m outhing out his hollow oes and aes, Deep-chested music . . . Tennyson was not the first m ajor poet to be aw are o f the distinctive features in his own or other voices, but he was the first whose voice was recorded. It is still possible to hear the old gram ophone record, too faint and technically poor to give m uch idea o f the living reality, b u t yet a link between two eras in the study o f language. T h e developm ent o f efficient means o f recording the hum an voice was contem porary with the recognition o f speech as the prim ary mode o f language, both as a m atter o f chronological developm ent and in its im portance as a corpus o f performance. T h e portable taperecorder makes it possible to assemble synchronic evidence before change invalidates it. W ith the application o f phonetic methods to the spoken utterance, w e are no longer satisfied with such verbal descriptions o f speech as ‘ hollow oes and aes’ . It would not be true to say that the m ajority o f linguists regard all w ritten language as unim portant or deviant— though the feeling is not unknown— but certainly language no longer means chiefly written language as it did for the classical philologists. W riting cannot be esteemed as it was when no other record o f usage, past or present, was available. T h e student o f literature, still dependent on written texts, m ay once more seem to be working against the mainstream o f linguists. I f certain readjustments can be m ade, however, the result is beneficial to both disciplines. In the first place, it is unwise to exaggerate the distance between

Speech and Writing

33

iprerli and writing in a society like our own where so m any people 1 |>. 213-24 for a survey o f the history o f English gram m ar and its rurrent developments. Applications o f gram m ar to literary criticism are m ade by I). D avie, Articulate Energy (London, 1955, R outledge and K egan Paul) and F. Berry, Poet's Grammar (London, 1958, R outledge and K.cgan Paul). Particular studies o f single gram m atical items o f the type m entioned on p. 51 w ill be found in G . R . H am ilton, The Tell-tale Article (London, 1949, Heinem ann); L. Spitzer, Linguistics and Literary History (Princeton, 1948, Princeton U niversity Press)— nee pp. 10-14 f° r discussion o f the use o f the phrase d cause de by Charles-Louis Phillippe; H . W einreich, ‘T h e T extu al Function o f the French A rticle’ (Chatman: Style, pp. 221-40). T h e syntax o f literature is considered by Leech, pp. 44-6; Nowottny, pp. 187-222 (a detailed exam ination o f a poem b y D ylan Thom as); W. N . Francis, ‘ Syntax and L iterary Interpretation’ (Chatman: Essays, pp. 209-16); S. R . Levin, ‘Poetry and Gram m aticalness’ (Chatman: Essays, pp. 224-30); D. D avie, ‘ Syntax and M usic in Paradise Lost’ in F. Kerm ode, ed., The Living Milton (London, i960, R outledge and K egan Paul). A diachronic survey is m ade by W . E. Baker, Syntax in English Poetry 1870-1930 (Berkeley and Los Angeles 1967, California U niver­ sity Press). O n the Prague School and aktualisce see P. L. G arvin, A Prague School Reader on Esthetics, Literary Structure and Style, (W ashington, D .C ., 1964, Georgetown U niversity Press).

6 Words and Meanings

‘ W h at do you read, m y lord?’— ‘W ords, words, words.’ T h e exchange between H am let and Polonius m ight be echoed by the literary critic who is asked to state the basic m aterial o f his study. It is not, however, an answer that entirely commends itself to the presentd ay linguist, whose attention is directed m ore towards syntax and phonology than towards the words w hich had traditionally seemed the irreducible atom ic components o f language. It is not that words are no longer held to be im portant, rather that attempts to think o f them as things in themselves, apart from other features o f language, raise difficulties. Even the title o f this chapter could be criticized as imprecise, for the definition o f word is not straightforward. W hen language is seen prim arily as speech, it becomes apparent that words are not neatly segmented as they are b y spaces in graphological realization. T h e pauses in speech do not consistently correspond with wordendings; m any languages, including English, do not m ake it clear to a foreign listener where the utterance is divided into words. Even the w ritten page is full o f com plications in this respect. Bloomfield m ade an advance when he defined a word as ‘a m inim al free form ’— the smallest unit o f m eaning that can exist in isolation, but this does not help us unreservedly. Is newspaper-seller a word, or petrol-station, or computer-programmer? T h e y certainly convey ‘bits’ o f m eaning which w e do not autom atically break into smaller units w hen we m eet them in com m on use. So too we can m ake total response to the epithets in Jo yce’s phrase ‘the bullockbefriending bard ’ or Shakespeare’s ‘world without end hour’, although they do not follow the regular adjective pattern. A t the other extreme, we m ay regard an affix as less than a word. Y e t people w ill speak confidently about ‘different isms and ologies’, or respond to a sentence like, ‘Some w ere in favour o f the idea, but most were very anti’, without filing a com plaint o f deviance. A gain , in an attem pt to m ake a count o f all the words in present-

Words and Meanings

59

ilny English, how do we assess the set teach, teaching, teacher, teachable, in >1.1у nothing o f the change to taught? I f a foreigner learns the form Ituch and has some knowledge o f methods o f word-formation, how 111.my words has he learned? Even more im portant, how m any words 11 и he learned in recognizing as units the sequence o f sounds which me written dow n as pipe, match, box, balance? For each o f these, and for 111.my other ‘words’, the dictionary offers a num ber o f apparently ililTcrent meanings. These are some o f the simpler and more obvious problems— ilirrc are m any others— w hich confront those who are trying to deal linguistically w ith ‘words’ . It is im portant to recognize that they exist and not to suppose that words can be treated as isolated linguistic phenomena. T h e traditional m ethod o f language-teaching was concerned w ith accidence, syntax and vocabulary; and indeed it цспегаїїу worked well enough in the hands o f a good teacher for the practical acquisition o f a language. T h e structuralists thought instead o f phonology, gram m ar and semantics, breaking some of the rigid divisions w hich prevented deeper understanding o f language as a hum an phenomenon. Chom sky and his followers prefer to discuss gram m ar as possessing phonological, syntactic and Ncmantic aspects. A ll this is o f the greatest im portance in linguistics and m ay help our present study— if only because there is still no definitive theory o f semantics and it is exciting to follow w hat is being done in this field. It will not do too m uch harm, however, i f we continue to use the term ‘w ord’ and to pursue words in their relationships to one another. Literary writers in all ages have experienced w hat T . S. Eliot called, ‘ the intolerable wrestle w ith w ords’. A lthough they m ay have formulated no linguistic theories, they knew well enough that m eaning is not to be sought only a t the level o f the single word. It is contained in the smaller units as well: in the affixes, and in the inflections which are few in modern English but were once numerous. R ecognition o f m eaning w ithin a smaller unit than the w ord makes it possible to compose new units w hich w ill themselves be more readily recognized in their own right. M eaningful neologisms depend on competence w hich splits the seemingly atom ic word and takes from it something that still communicates. H ow ever m uch we m ay dislike neologisms like motorcade or washeteria, however m uch we deplore the etym ological inaccuracy o f paratroop, w e cannot deny their semantic function. T h e y take their places in the paradigms

бо

Linguistics and Literature

o f sim ilarly classifiable words, with Shakespeare’s enskyed and C arro ll’s chortle, and with the less overtly derived coinages like Spenser’s blatant. It is, however, m eaning that spreads beyond word-boundaries w hich is o f the greatest interest. I f we look a t the lexicon o f any langue— the store o f words available to its users at a given tim e— we are presented w ith countless possibilities o f com bination. T h e lexicon is neither infinite nor static in itself. T h ere will alw ays be the hypothesis o f phonem ic sequences w hich are derived from the phonology but are inadmissible as words because they convey no agreed m eaning, and the lexicon is constantly losing items which becom e archaic, as well as receiving neologisms. Y et even a lexicon m uch smaller than that o f present-day English offers a seemingly infinite series o f syntagm atic and paradigm atic choices. A syntagm atic sequence is correctly realized, appropriate choices from the lexicon are inserted in their places— and we once again m arvel at the power o f hum an beings to generate new and unique sentences that are im m ediately comprehensible. N o single user w ill possess the whole lexicon, and performance does not draw on the whole range even o f w h at is theoretically possessed. Y e t a skilful writer has a large potential choice and exercises it widely. His choices are am ong the matters to be exam ined through stylistics. W e have perhaps seen enough to be w ary o f some o f the words used in the preceding paragraph. It has becom e clear that literary writers have a habit o f going beyond the conventions o f com m on speech in questions o f w h at is ‘correct’, w hich choices are ‘ap­ propriate’, even w hat is to be regarded as ‘com prehensible’, and in other matters. O n e thing they share with the rest o f us, though with different intensity— the tension between freedom and constraint w hich lies beneath all linguistic performance. For most o f us the tension is slightly and rarely felt, as when we are ‘feeling for a w ord ’, ‘at a loss for words’, ‘trying to put it better’ . T h e degree o f tension in literary creation was expressed by T . S. E liot in Burnt Morton: Words strain, C rack and sometimes break, under the burden, U nder the ten­ sion, slip, slide, perish,D ecay with imprecision, will not stay in place, W ill not stay still. T h e freedom o f choice becomes anarchic w ithout restriction. W e have seen how the rules o f syntax operate to reduce the num ber o f possible choices as a sequence progresses; and also that the literary

Words and Meanings

61

Hylr sometimes defies the prohibition. Syntagm atic deviation is I timparatively simple to detect and to judge. Paradigm atic deviation U ii different m atter, since the choice from the paradigm must be |udgcd with regard to m eaning and is therefore less readily referable In Ilie rules. Y e t in this relationship too each choice is to some rul n it restricted by w h at precedes and restrictive o f w hat follows. The restrictions m ay be imposed by external forces, and these need to be recognized in any stylistic approach. T h e influence of II itics— and o f creative writers themselves— in the imposition of 'lii^h style’ and ‘poetic diction’ has already been mentioned. It is mi influence which belongs to literary history but which cannot be discounted in criticism; and one exam ple must stand for m any. In the Impartial Critick (1718 ), John Dennis makes a prescriptive comment on W aller’s couplet: So Jove from Id a did both hosts survey, A n d when he pleas’d to thunder, part the fray. 'Is not that a noble similitude?’ Dennis asks, and answers thus: Yes; but the word F ray is altogether unw orthy o f the greatness o f the thought and the dignity o f heroic verse. F ray is fitter to express a quarrel between drunken bullies than between the G recian and T rojan heroes. This is one kind o f basis for a writer's choice, and it is o f a kind with the advice given in m anuals o f good writing, headed by the work o f the Fowler brothers, The King's English, w hich is adm irable in its own terms. Form al considerations, too, m ay condition the choice o f words: phonological requirements o f rhym e and alliteration, as well as metrical ones. Fashion, form, m eaning— and the im ponderable personal factor w hich is most interesting o f all— m ay seem a heavy concentration o f arm am ent on one little word. Y e t such concentra­ tion m ay be one of the factors which distinguish literature from other linguistic styles. T o put it in basic terms, it is because a writer takes such care with his language that w e m ay believe it worthwhile to apply some special technical methods to the result. Criticism which pays regard only to discrete words w ill not greatly heighten perception or increase response, but any and every critical approach to a work o f literature is m ade through the words w hich constitute

6г

Linguistics and Literature

it. T h e famous Prague M anifesto o f 1926, a signpost to the develop­ m ent o f literary stylistics, recognized that there is no getting away from the words. George Steiner interprets part o f its statement thus: T h e study o f a poem is an attem pt to register exhaustively the semantic elements or signal structure o f w hich that poem is made and through which, alone, it reaches our consciousness. {Minnis, p. 123) T h e use o f the phrase ‘signal structure’ here emphasizes a basic truth about language: words are signs, not things. W e all know, of course, that the sounds or letters w hich m ake up the word tree are not identical w ith any tangible vegetable growth. T h e word points our attention, to a particular tree or to a concept formulated from a num ber o f observed trees, without itself partaking o f a single characteristic that could be called ‘ tree-like’ . This is clear, except when w e react em otively or superstitiously to words as if they some­ how are the things that they denote, or if we are stupid enough to find something uniquely correct in tree and are incredulous that any sensible person w ould call the same object arbre or baum or albero. T h e identification o f words w ith things is o f some psychological and anthropological interest; it has im plications for our present purpose too. T h e word nightingale is not a small brow n bird that sings by night; neither is rossignol, luscinia, Philomel, or light-winged Dryad o f the trees. Y e t all these point to the same creature— the first in w hat we should call a foreign language, the second in technical zoological description, and the others . . .? W e are back with the question o f appropriate register, for the last two are clearly ‘literary’ and acceptable only in a certain kind o f context. E ach o f the four has a place where it seems to fit, isolated from others where it w ould be awkward or deviant. W e adjust our expectations and m eet it without surprise, once w e have accepted that a particular register is being used. N ow it m ay be felt that this close consideration o f single words is rem oved from daily speech and listening and is somewhat artificial. T h e point m ay readily be conceded; the linguist J. R . Firth held that words operate in social situations where we p ay little attention to single items such as would receive separate entries in a glossary. O u r response is holophrastic, m ade to a total m eaning and not to the sequence o f separate meanings. Single words are noted only when they are brought into prom inence by being particularly striking,

Words and Meanings

63

illm(dieting or shocking. In other words, when they are foregrounded шиї possibly appear deviant. The response to a work o f literature is properly to the whole text, 11ml us an intelligent response to any linguistic realization must be In Ilie whole. Y e t single words im pinge on the m ind more often in lllrrature than in other styles, and it is part o f the w riter’s art that lliey should do so. T h e foregrounding m ay be done b y formal devices 11I prosody, b y syntagm atic deviance and b y choice that is un­ expected in the register. Now we have seen that literature can and does avail itself o f all registers in a langue; no register can be excluded— even though we may learn to recognize a distinctively literary register, or several. W e do not know w h at to expect as we do in non-literary situations; we ill) not know where we are, and that is one reason w h y literature is rxciting and im portant. M ost com m unication in life is carried on with an unconscious prediction o f probabilities and rejection o f im ­ probabilities. In buying a railw ay ticket, it is extrem ely likely that single, please, change w ill be heard, even m ore extrem ely unlikely that dragon, tribal or syntax w ill be. In a sim ilar situation at a French station we could with some assurance exclude tu, though it m ight be very frequent in another situation. Nor is the determ inant only semantic or social: it w ould be startling to hear prognostication, disenchantment or glacial in a w eather forecast, though on some grounds thpy m ight seem quite appropriate. No such inhibitions constrain the literary writer, and the response to his work must be open and receptive. Y e t the balance is not all one w ay and although literature m ay seem open-ended in its possibilities, it does in fact act as something o f a controlling influence. This is not, or not solely, b y reason o f prescriptivism am ong its practitioners and critics, but b y the very fact o f its existence as part o f a com m unity’s culture, as a set o f perm anent and prestigious linguistic realizations. Sooner or later in every age, and despite the intentions o f successive reformers, literature creates its own stylistic variations from the spoken norm. L iterary critics have perhaps been wiser than linguists in under­ standing w h at literature does for the words that it uses. W ords which are lifted from the lexicon for a particular use m ay be returned to it with signs o f their honour still upon them. A single use m ay dignify a word and give it life after m any o f its contemporaries have faded into archaism : this is true, for understanding if not for active use, o f the Authorized Version kine and the Shakespearean bourne.

64

Linguistics and Literature

M isplaced revivals o f the ‘ye olde’ type, however, leap from the holophrase like an obscenity. M ore often, the power o f the word comes from repeated use. T h in k w h at T h ackeray m ade o f the obscure word snob, o f how W ordsworth and the other Rom antic poets conditioned our response to nature. W ords are not things, but they can acquire associations w hich affect our w ay o f under­ standing things. A gain the tension: for the word culled from the lexicon does not com e untested by the speech-community. L ike a human being, its distinctiveness is partly owed to the influence o f birth and environ­ ment. Its user has a certain responsibility to honour its accepted m eaning and its proper placing in the syntax. But these m ay not be beyond all doubt; usage can blur and blunt m eaning as well as sharpening it, and the result m ay be the am biguity against w hich m anuals o f good w riting w arn us and in w hich poets rejoice. Since the publication o f W illiam Empson’s Seven Types o f Ambiguity in 1930, the word ambiguity has been used som ewhat loosely in literary criticism to describe any feature in a text w hich could be interpreted in m ore than one w ay. T h ere is no need to be over-fussy about nom enclature at this point, bu t it is as well to note that in linguistics ambiguity is usually taken with reference to the problem o f sentences w hich seem identical in surface structure bu t have different deep meanings. T h e ambiguities o f d aily speech are, generally, unintentional and call for clarification as soon as they are detected. T h ey m ay be phonic— ‘I m eant I ’d have a pear, the fruit, not a pair, two’ ; or semantic— ‘D o you m ean funny, peculiar, or funny, ha, ha?’ or syntactic, as when we question whether running water means w ater w hich runs, or the process o f causing water to run. A n y o f these m ay occur in literature, but in this style they are m uch more likely to be studied and intentional. T h e words o f literary language m ay be in conflict, but it is conflict to which they are deliberately set on, in contrast to the random brawls o f words in colloquial use. T h e type o f phonic am biguity known as the pun is fam iliar to all. T h e phonic identity or close similarity o f two or m ore words is exploited in a m anner w hich brings their different meanings into juxtaposition. Its deterioration in the hum our o f the pantom im e and the V ictorian com ic periodical should not m ake the modern reader despise its use in foregrounding w ith m ore serious intent. It can be explicit, when the words in question are realized as separate units:

Words and Meanings

65

I ’ll gild the faces o f the grooms withal, For it must seem their guilt. (Macbeth, Il.i) or im plicit when we are left to deduce two meanings from one unit: T his counsellor Is now most still, most secret, and most grave, W ho was in life a foolish prating knave. (Hamlet, I ll.iv ) George Eliot has a fine use o f the second type in Daniel Deronda (Chapter 23) when Gw endolen is distressed after a hum iliating interview w ith Klesm er and her unsuspecting m other enters saying, 'I see by the wheel-marks that Klesm er has been here.’ T h e surface meaning relates to ‘ the sound o f his departing wheels getting more distant on the gravel’ which Gwendolen has ju st heard, but the phonic identity with weal relates to the suffering w hich is not physical but has already been given a physical m etaphor b y the author— 'E very w ord that Klesm er had said seemed to have been branded into her m em ory’. L iterary am biguity can draw on phonic, semantic and syntactic features. Shakespeare’s punning on his own nam e w ith the com plex Elizabethan associations o f will with sexual as well as mental desire in Sonnet 143 is well known: So w ill I pray that thou mayst have thy W ill. Donne has an even m ore adm irable piece o f w ordplay in ‘A H ym n to G od the Father’ with its repeated refrain: W hen T h o u has done, Thou hast not done; F or I have more and the conclusion: A nd having done that, T h o u hast done; I fear no more. Rem em bering the pronunciation o f Donne as /d A n/ we have here a use o f language w hich heightens the uncertainty and spiritual anguish that is finally resolved by faith. It exploits the pun through

66

Linguistics and Literature

the syntactic am biguity o f ‘T h o u hast’, w hich could lead to the series: Thou hast done, finished, accomplished . . . or to: T h o u hast Donne, H erbert, M arvell . . . T h e two series are equally possible in positive and negative verbs so that Donne offers us a com plex dance o f interpretations: W hen you have finished, you still have not got m e . . . W hen you have got me, that is not the end . . . W hen you seem to have finished, you have not really ended . . . and these are not all the possibilities. T h e point is that we do not seek the one ‘correct’ interpretation, for any m eaning which the language can bear is correct within the poem. A unit w hich most people would think o f as ‘one w ord’ m ay carry a num ber o f meanings, b y association with certain contexts. Thus pipe can be any tubular object, a musical instrument or a piece o f apparatus for smoking; a hand can be on a clock or watch as well as at the end o f the arm. M ultiple m eaning or polysemy is o f consider­ able linguistic im portance, and the process o f extension is a concern o f historical linguistics. M ost o f the time, we are able to distinguish the intended m eaning by the usual process o f m ental adjustment to context and register: we don’t expect to find tobacco pipes in the school recorder band. T h e literary language, however, again refuses to give us com fortable divisions o f m eaning beyond which im agination need not stray. It often forces us to accept polysemy not as a feature from w hich w e select but as one in which we meet the w riter’s intention without restriction. T hus W hitm an in ‘T h e Im prisoned Soul’ : A t the last, tenderly, From the walls o f the powerful, fortressed house, From the clasp o f the knitted locks— from the keep o f the well-closed doors, L et me be wafted. W e are not allowed to interpret locks solely as ‘door locks’ and exclude ‘locks o f hair’ w ith its suggestion o f binding hum an re­

Words and Meanings

67

lationship. A lthough fortressed points to the architectural m eaning o f keep, the wider com mon sense o f ‘retention’ is equally present. T h e writer m ay indeed call in the aid o f context to distinguish the meanings o f polysemic words; but his intention is not necessarily to elucidate a single m eaning but rather to em phasize the uncertain­ ties o f d aily usage and to point from this to an ironical comment on the hum an predicam ent. W hat has already passed in the action allows us to follow the jealousy o f Leontes through the meanings o f play as ‘reaction’, ‘sexual intercourse’ and ‘ theatrical acting’ : Go, play, boy, play: thy m other plays and I P lay too; but so disgraced a part, whose issue W ill hiss m e to m y grave. ( The Winter's Tale, I.ii) A sim ilar extension is m ade in H ousm an’s poem, ‘L an cer’, with its repeated italicized line: Oh who would not sleep with the brave? w here the successive juxtapositions w ith other lines bring out the literal sense o f the youth desiring comradeship o f com m unal life, the sexual thoughts o f the girls who w atch the soldiers pass, and the sleep o f death w hich is to be the recruit’s fate. Polysem y m ay allow a writer to work on two levels concurrently, apparently relating one set o f events while really indicating some­ thing different. W e m ove here towards m etaphor, which must be a separate concern, but it is interesting to see how a chosen im age can be m aintained by word-choice appropriate to the register in w hich we should norm ally expect to find it, while the m etaphorical relation to hidden m eaning is deferred. For exam ple, George H erbert sustains the im age o f G od as the landlord in the poem ‘Redem ption’ by use o f legal terms w hich are in perfect registeragreem ent with the opening statement: H avin g been tenant long to a rich Lord N ot thriving, I resolved to be bold, A n d m ake a suit unto him, to afford A new sm all-rented lease, and cancel th ’old. In heaven at his m anor I him sought: T h ey told me there that he was lately gone

68

Linguistics and Literature A bo u t some land, w hich he had dearly bought L on g since on earth, to take possession.

T h e overt legal narrative is acceptable in its own right but com ­ placence is jarred b y the intrusion o f heaven and on earth and the reader is alerted for the conclusion: T h ere I him espied W ho straight, Tour suit is granted said, and died H ere than is another exam ple o f register-mixing: we have already seen one in H enry R eed’s ‘N am ing o f Parts’ . Selection is m ade o f words current in more than one register o f the speech com munity. H owever, the writer m ay not confine him self to any normal register bu t rather create his own b y choices that w ould seem odd or questionable in that context in everyday use. It is useful, though w ithout attem pting to draw any impassable line, to distinguish between two ways in w hich a w riter’s selection o f a single word m ay seem adm irable. W e w ill assume that there is no syntagm atic deviation and that the choice is paradigm atic w ithin a context that is free from apparent am biguity. O f course, the associations and figurative applications o f words m ay still operate even when there is no obvious polysemy. In the first w ay, there is no deviation; the achievem ent is in tackling the problem o f synonymous words. It m ay well be argued that there are no perfect synonyms, since choice must be conditioned by register, dialect and em otive association. H ow ever, the problem o f word-selection is difficult and is not m uch aided b y the brief definitions o f a dictionary or the listings o f a thesaurus. O n e o f the most effective ways o f finding out w hat a word means in current usage is b y asking people whether they w ould readily use it in a given sentence. Consider the words in italics in each o f the following quotations; space does not allow extended quotation, but it is advisable to look up the whole passage if possible. W h at other words could the writer have chosen w hich are seem ingly synonymous? W ould they have been equally effective, or more or less effective? This is the right sort o f critical question to ask, although there is no single answer that is ‘right’ as a sum is right or wrong. A m an so various, that he seemed to be N ot one, but all m ankind’s, epitome. (Dryden, Absalom and Achitophel)

Words and Meanings

6g

T en thousand saw I at a glance, Tossing their heads in sprightly dance. (W ordsworth, ‘Daffodils’) W as she guilty or not? She said not; but who could tell w hat was truth w hich cam e from those lips; or if that corrupt heart was in this case pure? (Thackeray, Vanity Fair, Ch. 18) In that enormous silence, tiny and unafraid, Com es up along a w inding road the noise o f the Crusade. (Chesterton, ‘Lepanto’) Last night at the Jackson’s Agnes had displayed a brisk pity that m ade him wish to w ring her neck. (Forster, The Longest Journey, Ch. 26) A gainst these we m ay set the choices that seem deviant because we should not norm ally regard them as available at the point o f developm ent w hich the text has reached. It is not that they come from a different register— that, as we have seen, is an easily detect­ able device— but rather that the paradigm atic list that we should expect to construct in order to choose a filler for this particular space w ould not contain it. T h e blood-dim med tide is loosed, and everywhere T h e ceremony o f innocence is drowned. (Yeats, ‘T h e Second C om ing’) Here, w e m ay feel, the m etaphor o f tide is well sustained in the participle drowned, and we are happy to accept a poetic personifica­ tion o f innocence, but ceremony pulls us up sharply. A rather rhetorical public speaker m ight use the same statement w ith body o f innocence, or figure, or person, or martyr, or victim . . . W e cast through the paradigm to find a word that can com bine the abstraction o f inno­ cence w ith the physical nature o f drowning in a tide. But Yeats was not prepared to close the list on these terms and he m ade a choice w hich was syntagm atically acceptable but w hich extended instead o f further containing the m etaphor o f his vision. T h e reader m ay, if he wishes, say that the choice was a bad one; but the oddity o f the choice foregrounds the w ord and demands that response to it is made. H ere are some more examples o f paradigm atic deviance. T h e

jo

Linguistics and Literature

reader w ill make his own response to them, based on the w ay in w hich he reacts to the defeat o f regular linguistic expectation. It is necessary to consider the force o f the chosen word in relation to other possibilities o f the same class which m ight be considered more likely; also whether m eaning is heightened or blurred b y the deviation: But most through m idnight streets I hear H ow the youthful harlot’s curse Blasts the new-born infant’s tear, A nd blights w ith plagues the m arriage hearse. (Blake, ‘London’ ) A rth u r followed him up the staircase . . . into a dim bed-cham ber, the floor o f w hich had gradually so sunk and settled, that the fireplace was in a dell. (Dickens, Little Dorrit, Bk. I, Ch. 2) I am the m an who looked for peace and found M y own eyes barbed. (Sidney K eyes, ‘W ar Poet’) N ow as I was young and easy under the apple boughs A b o u t the lilting house and happy as the grass was green. (D ylan Thom as, ‘Fern H ill’) T h e last quotation, o f course, shows a second and more startling deviation b y defeating expectation o f the stereotyped simile ‘happy as the d ay was long’ . It brings us to consider the treatment which writers give another type o f usage fam iliar in daily speech but often regarded as distinctively literary.

F U R T H E R R E A D IN G

T h e im portant though difficult study o f semantics can begin with G . N . Leech, Towards a Semantic Description o f English (London, 1969, Longm an), followed b y S. U llm ann, Semantics (Oxford, 1962, Blackwell). R . A . W aldron, Sense and Sense Development (London, 1967, Andr6 Deutsch) is also useful. A n older but influential work is С . K . O gden and I. A . Richards, The Meaning o f Meaning (10th ed., London, i960, R outledge and K eg an Paul). W . Empson, Seven Types o f Ambiguity (London, 1930, C hatto and Windus) referred to on p. 64 is conveniently available in a cheaper

Words and Meanings

ji

edition (Harmondsworth, 1961, Penguin Books); see also Em pson’s later book, The Structure o f Complex Words (London, 1951, C hatto and W indus). T h e work o f J. R . Firth is best approached through his Papers in Linguistics 1934-51 (London, 1957, O xford U niversity Press); some o f his ideas are developed b y М . A . K . H allid ay, A . M cIntosh and P. Strevens, The Linguistic Sciences and Language Teaching (London, 1964, Longm an). Also useful are: C . S. Lewis, Studies in Words (Cam bridge, i960, Cam bridge U niversity Press); B. Groom , The Diction o f Poetry from Spenser to Bridges (London, 1956, O xford U niversity Press); Nowottny, pp. 26-48, 146-73; Leech, pp, 42-4, 131-6, 105-21.

7 The Language of Rhetoric

T h e question o f lexical deviation has em phasized something w hich has becom e increasingly apparent in the course o f our study. L iterary language does not function prim arily for the purpose o f conveying inform ation verifiable by reference to experience which is not linguistic, or at all events not verbalized. Certainly this inform ative function is not excluded, and literature can increase our knowledge and understanding o f the external w orld as discerned b y sensory perceptions. Some literary fashions have m ade m uch o f the need to be objectively ‘true to life’ and have earned such labels as ‘realism ’ and ‘naturalism ’, It does not dem and m uch critical judgem ent, however, to distinguish this kind o f creation from the factual reporting which makes no claim to be in the literary style. A novel like Z o la ’s Germinal or U pton Sinclair’s The Jungle is m uch m ore— and also m uch less— than a report based on official interrogation and w eighing o f evidence. M uch more, because it incorporates the author’s desire to touch the emotions, to cause shock and to persuade into action; m uch less, because the ‘facts’ are selected and arranged in a w ay that does not totally reproduce a verifiable situation. T h e qu ality o f im agination, w hich we have noted as one o f the distinguishing marks o f literature, comes into service. W hen it is not disciplined, the result is crude sensationalism; w hen it is inadequate, the result is a piece o f non-literature in w hich the overt message is too m uch for the m edium , as in m any V ictorian novels dedicated to a specific social reform. T h e use o f language to persuade or influence, even to prom ote action, is b y no means inim ical to literature. T h e fallacy o f sup­ posing that even non-literary language is prim arily to inform has been assailed from Coleridge to Chomsky. T h e use o f specific linguistic devices to m ake desired effects does not isolate literature from the com m on core o f the langue. Close attention given in recent

The Language o f Rhetoric

73

years to literary language has brought renewed respect for the attitude to ‘style’ w hich was dom inant for a very long time in the past— the idea that literature produced a set o f ‘models’ for the generation o f desired linguistic effects. T h e study o f rhetoric rested on a special kind o f attitude to language as a faculty through which the recipient— reader or auditor— could be influenced in the m anner desired by the writer or orator. L anguage could be m anipulated into recognized ‘figures’, the categorizing and exem plifying o f which was a proper concern o f the critic from classical G reece to the beginning o f the R om an tic era. T h e relegation o f these ‘figures of speech’ to textbooks, to be learned b y rote and reproduced in examinations, gave them a bad nam e which they are now losing. T h e attention o f both critics and creative writers has been directed towards figurative language w ith a concern com parable to that shown in the sixteenth and seventeenth centuries. For the Renaissance writer, skill in rhetoric was a necessary part o f the prevailing attitude to language in which excitem ent about new ly discovered flexibility was in tension w ith anxiety about the status o f the vernacular tongue. It was desirable that English should be proved capable o f accom m odating the figures traditional to Greek and L atin. M anuals like P eacham ’s Garden o f Eloquence (1577) and Puttenham ’s Art o f English Poesy (1589) did not seek to impose artificial constraints on creation but rather to codify contem porary practice and thereby guide both the poet and the daily user o f language. T h e very fact o f parody, as in Shakespeare’s Sonnet w hich begins ‘ M y mistress’ eyes are nothing like the sun’ , proves the ready acceptance o f rhetorical conventions. T h e strength o f literature was seen to lie in its controlled use o f features w hich did not destroy regular com m unication but w ere developed from fam iliar usage and arranged for the best effect. A ll this is not far from the idea o f foregrounding through devia­ tion, w ith the related recognition o f a langue as the totality o f available resources from w hich different styles and dialects are drawn. So influential a critic as Northrop F rye has called for ‘a wholesale revival o f the lexicon o f Renaissance rhetoric’. T h is lexicon m ay indeed help us to discuss the special kind o f ‘reality’ which literature often presents: the creation o f an experience pointing to no such perceivable objectivity as we m ight expect in a conversation or a news com m entary, but unquestionably evoking response from the reader and becom ing part o f his individual situation.

у4

Linguistics and Literature

This response requires no withdraw al from the instinctive response to specific performance by any m em ber o f the shared speech-community. It is not a switch o f attention, but a further step in the distancing process which gives civilized language com ­ m unicative value not possessed by the anim al noise which auto­ m atically accompanies a given stimulus. Thus lily is not a white flower, but a sign w hich evokes our precedent knowledge o f a white flower. This is the denotative use; a recipient w ith developed awareness o f the language m ay let his thoughts dwell on the word lily and recall Pre-Raphaelite paintings and aesthetic notions o f purity w hich are am ong the connotative associations. So he will be prepared to respond to K eats: I see a lily on thy brow, W ith anguish moist, and fever dew . . . (‘L a Belle D am e sans M erci’ ) w ith acceptance o f the figurative use o f the word, in which both denotative and connotative senses help to reach an understanding still farther aw ay from the basic referent. Sim ilarly, branch can aid com m unication about trees, can evoke ideas o f dynam ic growth in biological or political or religious registers, and can be accepted as purely figurative when used by M arlow e: C u t is the branch that m ight have grown full straight A n d burned is A po llo ’s laurel bough T h a t sometime grew w ithin this learned man. (Doctor Faustus) In all this, as w e have seen in other features o f language, it is wiser to regard m eaning as a spectrum rather than as a set o f enclosed cells: figurative use does not em erge all o f a sudden but shades off, both diachronically and synchronically, from connotation. So once again the language o f literature is seen to be not far from the conventions o f daily speech and to be am enable to the same methods o f investigation. Before looking at more literary examples, let us think further about the relationship o f rhetoric to non-literary styles. T h e label ‘figures o f speech’ was not such a misnomer as it appeared w hen these usages were tabulated in textbooks o f pre­ scriptive gram m ar. T h e apparent detachm ent was a result o f the

The Language o f Rhetoric

75

insistence that written language was the only type w orthy o f serious study, and o f the raiding o f past literature for examples to support names o f figures w hich w ere treated as having reality in their own right. W ithout departing from traditional nom enclature, or even traditional practice in adducing literary examples, w e can find the figurative expressions alive in d aily usage. T h ere is no difficulty about agreeing with Bloomfield that ‘poetic m etaphor is largely an outgrowth o f the transferred uses o f ordinary language’ . I f we contend that the language o f literature constitutes a style o f the langue, any viable gram m ar m ust be able to accom m odate the usages o f literary writers. Conversely, the writers must be open to the judgem ent o f the gram mar. W e w ill look at some— not all— o f the established figures in the dual relationship o f literature and common usage. O n e distinction o f types w hich has sometimes been overlooked is o f considerable linguistic im portance— the difference between tropes and schemes that was generally made b y the great rhetoricians. Tropes depend essentially on paradigm atic relationships, schemes on syntagmatic. Most o f the more fam iliar textbook ‘figures o f speech’ are tropes. T h ey take us on from the lexical deviations w hich were discussed at the end o f the previous chapter and for w hich the nam e metaphor there becam e necessary. T h e y are the result o f unusual choices from the items which the gram m ar makes available in a given pattern. S im ile is the root-notion o f tropes: the com parison derived from likeness perceived between two referents. T h ere is clearly a very wide range o f choice here, and the successful literary simile will point a likeness not usually discerned yet not so far-reached as to be purely subjective and therefore uncom m unicative. A t least one item generally refers to something perceptible b y the senses, which foregrounds the other item b y its actuality. T h e com parison m ay be directly between noun and noun: T h y soul was like a star, and dwelt apart: T h o u hadst a voice whose sound was like the sea. (W ordsworth, ‘ London, 1802’) She smiled as she saw how big his mouth was, and his chin so small, and his nose curved like a switchback, with a knob at the end. (V irginia W oolf, The Voyage Out)

уб

Linguistics and Literature

or between a quality shared by the two items: His legions, angel forms, who lay entranced T h ick as autum nal leaves that strow the brooks In Vallom brosa. (M ilton, Paradise Lost, Book i ) O ld as a coat on a chair; and his crushed hand as unexpressive as a bird’s face. (Terence T iller, ‘Egyptian Beggar’) or between action w hich makes a verb act as the link: H e trod the ling Like a buck in spring (K ipling, ‘T h e Ballad o f East and W est’) Words flower like crocuses in the hanging woods (Sidney K eyes, ‘W illiam W ordsworth’) A ll varieties are fam iliar in speech, repeated until worn into cliches: ‘a face like thunder’ ; ‘as cool as a cucum ber’ ; ‘ M arch comes in like a lion and goes out like a lam b’ . M e t a p h o r is a term sometimes used to include the m ore particular types o f figure, such as those discussed below. W hile it m ay be convenient to consider them more specifically, they certainly have the nature o f m etaphor which makes analogy b y compression o f the simile so that the overt ground o f likeness is not verbalized. T h e im plicit comparison contained in a m etaphor is the essence o f figurative language and must be exam ined m ore closely later. For the moment a few examples will establish the relation o f literary m etaphor to com mon usage: I feed a flame within, which so torments me T h a t it both pains m y heart, and yet contents me. (Dryden, ‘H idden Flam e’) T h o u still unravished bride o f quietness, T h o u foster-child o f silence and slow Tim e (Keats, ‘ O de on a G recian U rn ’) But somewhere some word presses O n the high door o f a skull (Stephen Spender, ‘F all o f a C ity ’)

The Language o f Rhetoric

77

Л metaphor can be opened into simile and compressed again: (Mrs. Skewton) had a sharp eye, verily, a t picquet. It glistened like a bird ’s, and did not fix itself upon the gam e, but pierced the room from end to end, and gleam ed on harp, performer, listener, everything. (Dickens, Dombey and Son, C h. 21) I lere the stereotyped ‘sharp eye’ m etaphor develops into the animate кішіїе o f the bird, and the m etaphoric possibilities o f both ideas are exploited w ith ‘ pierced’ and ‘gleam ed’ . Little com m ent is needed on the wide range o f com m on metaphor, which falls into at least four degrees o f being figurative in the awareness o f users and recipients: (i) T h e obvious and blatant m etaphor w hich is always in danger o f becom ing ludicrous b y associating w ith others in ‘mixed metaphor’ o f the type, ‘ I smell a rat, I see it floating in the air, but I hope to nip it in the bud’ . (ii) T h e m etaphor w hich is accepted as figurative because it puts an idea more vividly and forcefully than abstraction could do but does not seem seriously deviant in any register: ‘in the light of experience’ , ‘ the hub o f activity’ . (iii) T h e m etaphor w hich is not regarded as figurative at all except when attention is draw n to it b y gross ‘m ixing’ or by the difficulty o f finding a non-m etaphorical word to fill the same space: ‘ the foot o f the hill’, ‘a bottleneck in production’, ‘blanket legisla­ tion.’ (iv) T h e m etaphor w hich is totally ‘dead’ because its literal meaning is lost or obsolescent and known only to the student o f language: ‘ ponder’, ‘depend’, ‘preposterous’ . T h is type is meta­ phorical only in a historical view. S y n e c d o c h e is the m etaphorical use o f part o f the referent to stand for the whole: F air stood the wind for France W hen we our sails advance (M ichael D rayton, ‘A gincourt’) In cam e M rs. Fezziw ig, one vast substantial smile. (Dickens, A Christmas Carol) L a y your sleeping head, m y love, H um an on m y faithless arm. (W . H . A uden, ‘ L ay your sleeping head’)

y8

Linguistics and Literature

T h e figure is fam iliar in everyday use such as ‘hand’ for ‘workman* — an exam ple satirized b y Dickens in Hard Times. M e t o n y m y is the use o f some feature contiguous or closely associated with the referent: Sceptre and Crow n M ust tum ble down, A n d in the dust be equal m ade W ith the poor crooked scythe and spade. (James Shirley, ‘T h e glories o f our blood’) Cedant arm a togae, concedant laurea laudi (Cicero, De Officiis) W e accept daily such m etonym y as ‘ Crow n property’ ; ‘coppers’ for small coins; and ‘ the pen is m ightier than the sword’, a quotation from Bulw er Lytton that has become а cliche. M e i o s i s is conscious understatement, with its special type L i t o t e s which uses a negative construction to foreground an intended positive emphasis: H e was a m an, take him for all in all (Hamlet I.ii) A thing o f beauty is a jo y for ever: Its loveliness increases; it w ill never Pass into nothingness. (Keats, Endymion) A very British turn o f phrase: ‘quite a few people’ , ‘could be worse’, ‘not b ad ’, ‘by no means unlikely’ . . . H y b e r b o l e is conscious overstatement w hich foregrounds the theme by paradigm atic choices that would norm ally seem excessive in the context: I w ill love thee still, m y dear, T ill a ’ the seas gang dry. (Burns, ‘M y love is like a red, red rose’) Cover her face; mine eyes dazzle: she died young (Webster, The Duchess o f Malfi, IV .ii) A very un-British feature o f ordinary speech? ‘T errib ly sorry’ ; ‘it’s aw fully good o f you’ ; ‘nobody could have been kinder’ . . .

The Language o f Rhetoric

уд

Tropes, then, are richly varied and unpredictable in the items which they include. Schemes make the foregrounding effect through development o f normal syntactic patterns by repetition and ju x ta ­ position; if they are deviant at all, it is b y unusual frequency, not by unexpected choice. Rhetoricians have nam ed m any types o f ihem; a very few must serve for illustration. A n ap h o ra, sometimes used o f verbal repetition in general, is specifically the repetition o f a w ord or phrase a t the beginning o f successive stages o f the chosen pattern: After the torchlight red on sw eaty faces After the frosty silence in the gardens After the agony in stony places (T. S. Eliot, The Waste Land) N ow N ow N ow N ow

to the banquet we press; for the eggs and the ham; for the mustard and cress, for the strawberry jam ! (W . S. G ilbert, The Sorcerer)

D ead, your M ajesty. D ead, m y lords and gentlemen. D ead, R igh t Reverends and W ron g Reverends o f every order. D ead, men and wom en born with H eavenly compassion in your hearts. (Dickens, Bleak House, Ch. 47) E p is t r o p h e

uses repetition at the end o f successive stages:

I f you did know to whom I gave the ring, I f you did know for whom I gave the ring, A n d would conceive for w hat I gave the ring, A nd how unw illingly I left the ring, W hen naught w ould be accepted but the ring, Y o u w ould abate the strength o f your displeasure. (Shakespeare, The Merchant o f Venice, V .i) T h e device is very fam iliar in the refrains o f songs and the repeated last line o f stanzas in forms like the ballade. S y m p l o c e repeats at the beginning and the end: W e are the hollow men W e are the stuffed men (T. S. Eliot, ‘T h e H ollow M en ’)

8o

Linguistics and Literature

and see also W yatt’s poem ‘A n d w ilt thou leave m e thus?’ which repeats this line at the beginning o f each stanza and the line, ‘ Say nay! say n a y!’ at the end o f each. A n a d ip lo s is links the end o f one stage to the beginning o f the next b y repetition, as D onne’s sonnets La Corona each take the last line o f a sonnet to be the opening line o f the next in the sequence. Ernest Dowson’s ‘Flos L un ae’ opens and closes each stanza with the line ‘ I would not alter thy cold eyes’ . E p iz e u x is repeats a w ord or phrase without any break at all: A n d when he falls, he falls like Lucifer, N ever to hope again. (Shakespeare, Henry VIII, IH .ii) Sun is torn in coloured petals on the water, T h e w ater shivering in the heat and the north wind (R ex W arner, ‘N ile Fishermen’ ) W e are certainly fam iliar w ith repetition in the syntax o f daily speech; but w e do not dignify it with technical names, in the w ay that w e recognize the appearance o f certain tropes. Repetition can be used consciously for emphasis— It’s cold outside, bitterly cold’ ; or to establish a phatic sense o f sharing— ‘I t ’s a shame,isn’t it?’ ‘Yes, it’sreally a sham e.’ M ore often, it is used unconsciously and is associated with users w ho have not a highly developed linguistic skill. A n y overheard conversation in a public place is likely to yield the kind o f repetition that Eliot reproduced in The Waste Land: W hen L il’s husband got dem obbed, I said— I didn’t m ince m y words, I said to her m y s e lf. . . Nevertheless, patterned repetition is constantly found in literary language, and also in the religious register: for exam ple, in litanies. Its appearance in the ritual and incantatory language o f diverse cultures cannot be overlooked b y the linguist. It should now be clear that even the seemingly extreme usages o f the literary style can be approached with the understanding that they grow from the com mon core o f language. T h e y do not call for modifications in the explanatory model o f langue. T h e y do, however, dem and that we recognize the special dimension o f literature in which these figures do not appear by autom atic response or by rapid register-choice. T h ey are planned and given performance as

The Language o f Rhetoric

81

contributing to a situation in w hich creator and recipient share a linguistic act without sharing the outward event w hich is linguistic­ ally signified. T h e literary use o f figurative language is evaluated in a whole text. In everyday speech, any figurative expressions can be paraphrased with no loss except possibly that o f emphasis, or vivid effect, or suitability to register: the paraphrase o f literary language loses w hat is essential to its kind o f performance. Equally, the response o f the recipient is affected, but not limited, by his recognition that figures can occur in non-literary styles. T h e rules o f interpretation w hich are applied almost autom atically to the conversational use o f ‘I smell a rat’ or ‘ as bold as brass’ must be applied m ore carefully and thoroughly to gain the full appreciation of literary rhetoric. In speech we seldom consciously hold the literal \ and figurative meanings o f an expression in balance at the same time, except when they are forced upon us b y a deliberate pun or by accidental m ixed m etaphor. In the exploration o f literature w e learn to be on the alert for likenesses not seen before; and if the w riter has done his work well, w e have only the text itself to guide our awareness o f the levels involved. A ll that we have seen in respect o f tropes has shown that some kind o f likeness is the basis o f every metaphor. It w ill be convenient now to consider metaphor as the basis o f figurative language, without continually digressing to consider the more delicate sub-divisions of rhetorical theory. T h ere must be some likeness, i f m etaphor is to com m unicate at all and not to lose all contact with reality. T h e point o f contact between figurative and norm ative use m ay be very slight; and it is an indication o f its success if it is one not realized before. M etaphor w ill focus attention on some aspect o f the referent w hich makes analogy possible. M etaphor often makes a bridge between levels o f experience w hich are not norm ally considered to be expressible in the same terms. T h e bridging can be o f m any types: here b y w ay o f exam ple are three o f the most frequent. (i) O n e type o f sensory perception is expressed in terms o f another: A nnihilating all that’s made T o a green thought in a green shade (M arvell, ‘Thoughts in a G arden’) I f music be the food o f love, play on (Shakespeare, Twelfth Night, I.i)

8г

Linguistics and Literature H er fist o f a face died clenched on a round pain (D ylan Thom as, ‘ In M em ory o f A nn Jones’)

and we m ay recall Chesterton’s parody o f m odern verse: So sorry i f you have a green pain G naw ing your brain aw ay . . . W hen I have a pain, I never notice the colour. (ii) A non-human referent is given hum an attributes: So I unto m yself alone will sing; T h e woods shall to me answer, and m y echo ring (Spenser, ‘Epithalam ion’ ) Flakes o f soot . . . as big as full-grown snowflakes— gone into m ourning, one m ight im agine, for the death o f the sun. (Dickens, Bleak House, C h . i) (iii) A n abstraction is treated as if it were anim ate; A terrible beauty is born (W. B. Yeats, ‘ Easter 1916’) Bon chevalier m asque qui chevauche en silence, L e M alheur a perce m on vieux coeur de sa lance (Paul V erlain e, Sagesse, I) In the analysis o f m etaphor w e are in effect reconverting the thought back into the fuller statement o f simile. A simile is tripar­ tite: one thing is likened to another, and the ground o f likeness is specified. T h e terms tenor, vehicle and ground have been applied in the elements o f simile and the application will be shown in the following examples, where the superscribed T stands for ‘ tenor’, V for ‘vehicle’ and G for ‘ground’ : t

o

v

I have seen old ships sail like swans asleep (J. E. Flecker, ‘T h e O ld Ships’) T

V

G

A n eye like M ars to threaten and com mand (Shakespeare, Hamlet, III, iv)

His legions, angel forms, who lay entranced G

v

T h ick as autum nal leaves (M ilton, Paradise Lost, Bk. I) M etaphor omits the ground, w hich has to be sought and supplied; it is in the kernel sentence bu t not found in the sentence as performed. Thus when w e meet the m etaphor that closes A u den’s ‘ In M em ory o f W . B. Y eats’ : In the prison o f his days T each the free m an how to praise we supply the ground o f constraint w hich links ‘prison’ with the inescapable progression o f time verbalized as ‘days’ . Sometimes both tenor and ground must be supplied by the reader: T h eir path lay upw ard, over a great bald skull, h a lf grass, h a lf stubble (E. M . Forster, The Longest Journey, Ch. 12) H ere the vehicle ‘skull’ leads to the unexpressed ‘hill’, with the ground o f shape and bareness. Frequently more than one step is needed to find the kernel: Tiger, tiger, burning bright (Blake, ‘T h e T ig e r’) ‘Bright tiger’ is acceptable without recourse to m etaphor; but we have to link the ground ‘brightness’ with the unexpressed vehicle ‘fire’, w hich leads to ‘burning’ as the actual item to be the vehicle: (a) tiger (is as) bright (as a fire) (a fire) bum s bright tiger burning bright I hope that enough has been quoted to show that the freedom o f choice in formation o f m etaphor is immense. Freedom must, however, be in tension w ith the linguistic restraints o f syntax and lexicon discussed in previous chapters. A lthough, as w e have seen, tropes are based on paradigm atic relationship and schemes on

84

Linguistics and Literature

syntagm atic, yet every linguistic act is answerable to both systems. R andolph Q uirk puts the point precisely: A m etaphor involves sim ultaneously a paradigmatic relation between the literal elem ent it replaces and the figurative one it introduces, and a syntagmatic relation between the literal and m etaphorical elements in the linguistic environment. {Minnis, p. 308) T h e purpose o f stating a likeness creates a gap which has to be filled from the lexicon: the structure o f the m etaphor requires that the chosen item shall fit into the syntagm atic pattern. Thus, m etaphor is not different in kind from other utterances for which, as we have seen, certain possibilities are open and others closed.

FURTHER

R E A D IN G

A n excellent general introduction to the study o f figurative language in literature is P. Dixon, Rhetoric (London, 1971, M ethuen), which includes a useful bibliography. I. A . Richards, The Philosophy o f Rhetoric (London, 1936, Routledge and K eg an Paul), has been influential; it is the source o f the terms ‘ tenor’ and ‘vehicle’ used in this chapter. C . Brooke-Rose, A Grammar o f Metaphor (London, 1958, Seeker and W arburg), is particularly good on analysis o f the types o f m etaphorical analogy. R . W ellek and A . W arren, Theory o f Literature (London, 1949, Jonathan C ape), has an im portant study in C hapter 15, ‘ Im age, M etaphor, Sym bol, M yth ’ . Some o f the most valuable recent work on rhetoric w ith a stronger linguistic emphasis has been done b y G . N . Leech: see Leech, pp. 73-86, and his ‘ Linguistics and the Figures o f R hetoric’ {Fowler, pp. 135-56). Nowottny, pp. 49-98, should also be read. For the Renaissance approach to the subject see L . A . Sonnino, A Handbook to Sixteenth-century Rhetoric (London, 1968, Routledge and K eg an Paul). Descriptions o f the traditional figures as used in literary and form al writing w ill be found in H . W . Fowler, A Dictionary o f Modern English Usage (Oxford, 1926, O xford University Press). W . C. Booth A Rhetoric o f Irony (Chicago, 1974, University of Chicago Press) studies a particular literary device in detail.

8 Rhythm and Metre

T h e relationship between speech and w riting has already been discussed as a concern o f literary stylistics. A lthough the two modes o f realization are distinct and not to be confused, the switch from one to another is not difficult in a literate society. T h e units o f inform ation do not precisely correspond in the two, but an alpha­ betical system o f writing shows more or less an attem pt to reproduce speech-sounds visually, although historical change and lack o f phonetic understanding m ay result in something far short o f per­ fection. It is possible to represent speech in w riting by using alpha­ betical letters with the aid o f diacritical marks, or extra symbols, or both: this is the basis o f phonetic and phonem ic transcription. Y e t phonemes, necessary as a concept in phonetic analysis, are not the only constituents o f speech. O r to pu t it in less technical terms: our conversations do not consist only o f sounds. There can be foreign learners o f English— or any other tongue— who achieve faultless reproduction o f the separate sounds but are almost un­ intelligible when they open their mouths. I f each syllable in an utterance is given equal stress and an equal following pause, the recipient loses nearly all the features w hich enable m eaningful response. A n exam ple w ould be the following sentence, here transcribed graphologically, in w hich the oblique strokes represent breath-pauses and the superscribed vertical marks represent breath-force or stress; no account is taken o f the segmentation into ‘words’ : I I I I I I I I I I I I I John/must/give/you/a/def/in/ite/an/swer/by/Fri/day L eavin g aside the difficulties o f syllabic division in traditional spelling, it is clear that the pronunciation thus represented would not be heard in any native variety o f English. T o give any indication o f actual utterance we have to show fewer pauses and greater

86

Linguistics and Literature

variation in stress. By using a horizontal line for a light or secondary stress, and a cross for absence o f stress, we can overcome at least some o f the inadequacy o f representation: I

X

—

X

X

—

X

X

I

x

x

|

x

Johnm ustgiveyou/adefiniteanswer/byFriday or, w ith different emphasis but no change o f basic ‘ m eaning’ : —

|

—

x

x

|

x

x

x

x

x

|

x

Johnm ustgiveyou/adefiniteanswer/byFriday A good m any other arrangements o f pause and stress are con­ ceivable for this simple utterance within the bounds o f normal con­ versation: it is clear that m uch more than a succession o f sounds is involved. Elements o f stress, pitch and duration make up the intonation w hich gives distinctive pattern to the dialects and idiolects o f an ethnic language. It is chiefly through intonation that w e are enabled to ‘place’ a person without necessarily understanding all that he says, or to parody the features o f a foreign language w hich we speak im perfecdy. In graphological realization there is practically no indication o f intonation; it is silently and subconsciously supplied b y the reader, or inserted b y the actor or reciter, helped to some extent by punctua­ tion. A phonem ic transcription cannot rely on ordinary punctuation which, although originally an aid to oratory, is incomplete and uncertain in the complexities o f speech. A m ore delicate system o f m arking is used for these elements, w hich are the supra-segmental or prosodic division o f phonetics. T h e word ‘prosodic’ at once reminds us that literary critics have for centuries been concerned with the regular patterns o f verse and have tried to represent them visually b y various systems o f scansion. For convenience we can use the word rhythm for the distinctive but variable pattern in the spoken utterances o f a langue; the deliberate use o f a regular and recurrent pattern in a literary composition w ill be called metre. General linguistic study is not concerned w ith m etre as such, but m etre can be exam ined linguistic­ ally in relation to the rhythm o f spoken language. T h e latter is an extensive study and conclusions about it are b y no means definitive; a few principles are im portant for the present. W e can recognize in speech the principle o f ‘equal tim ing’ or isochronism, w hich breaks utterances into segments o f approxim ately

Rhythm and Metre

87

rqual duration in sym pathy w ith the pulses o f breathing on which the sounds are produced. T h e segmentation is not the same in all languages; in French, for exam ple, and more m arkedly in Japanese, it is syllabic. In English the unit is generally larger than the syllable, containing one stressed syllable and a variable num ber o f un­ stressed. T h ere are roughly equal time intervals from each stress to the next, though obviously not w ith metronom ic precision. T h e different markings o f the sentence examined above show how the system works, and also show that any m ajor deviation from it breaks aw ay from recognizable speech. T h e m ain stresses tend to fall on form-words (nouns, verbs, adjectives, adverbs) rather than on the other parts o f speech whose function is m ainly gram m atical or structural, and stress is thus closely related to the gram m ar and lexicon o f the langue. This is not an infallible guide, however, particularly in English where stress can fall alm ost on any item that is to be foregrounded. For instance, i f we give the instruction: P u t those books and papers on the table the utterance is most likely to m ove fairly rapid ly to find its m ain stress on table. It is, however, quite reasonable to say: Put those books and papers on the table (not the others) P u t those books and papers on the table (not just the books) Put those books and papers on the table (not underneath) W hen w e dwell too long on single examples out o f context the result soon appears unreal or slightly com ic, but unusual stressing is continually used in most conversations. T h e duration which depends on stress must not be confused with the question o f syllable-quantity; some syllables are longer than others because o f their phonetic m ake-up, apart from w hatever length is imposed on them in connected utterance b y force and pause o f breathing. Thus sit /sit/ is shorter than seat /si:t/ w hich in turn is shorter than seed /si:d /; two consonants give greater length to the preceding vowel than does a single consonant— Stan is shorter than stand. This factor enters into the total prosodic effect o f an utterance. A nother feature o f speech is the apparently negative but actually

88

Linguistics and Literature

very im portant one o f silence. It appears in the b rief pauses between segments and the longer ones between sentences, the latter units not alw ays corresponding to the syntactic structure that would qualify as a sentence in traditional gram m ar. A tape-recording of conversation w ill reveal pauses o f considerable duration o f which the original participants were not aw are at the time. W ithin an utterance, the placing o f a b rief pause can distinguish the different m eanings o f identical sequences o f phonemes w hich m ay or m ay not be fully distinguishable b y context: some addresses : summer dresses take Greater London : take G rey to London L iterary critics and theorists were aware o f such prosodic features long before any scientific study o f speech was m ade. T here have been a num ber o f theories about exactly how English poetry is constructed and consequently how it can best be scanned. M an y o f the theories, like those developed in form al gram m ar, have leaned too heavily on the classical languages and obscured the close relationship between literary and spoken English. T h ere was for a long time concentration on the feature ju st noted o f syllable quantity, w hich was the basis o f L atin verse prosody. T h e Rom antics, w ith their principle— at least in the first generation— o f bringing poetry back to daily speech-rhythm, looked to w hat Coleridge in the preface to ‘ Christabel’ called ‘a new principle: nam ely that o f counting in each line the accents, not the syllables’ . T h e principle was perhaps new in theory rather than in practice, but it becam e im portant in nineteenth-century thinking about prosody. T here was a revival o f interest in the accentual metre o f O ld English poetry, w hich had depended on a fixed num ber o f stresses in each line, w ith considerable freedom in the num ber o f unstressed syllables. W e have already seen that this is in fact the basis o f English speech-rhythm. Students o f English literature w ill be fam iliar with the elaboration o f the basic theory in O ld English, its debased form in the alliterative metre o f the later m edieval period as used by L angland, and its transformation into the ‘sprung rhythm ’ o f G erard M anley Hopkins. W h at Hopkins wrote about his theory o f prosody, as well as his actual poetry, is interesting source-material for literary stylistic study. Som e examples have been quoted in earlier chapters as illustrative o f other features and m ay be re­ exam ined, bearing in m ind that he worked on the basis o f stresses

Rhythm and Metre

8g

wliich could follow one another without intermission or be separated by an indeterminate num ber o f unstressed syllables. U nlike the ( )ld English poets, he sometimes used rhym e and did not always incorporate alliteration o f the stressed syllables. I f his poetry is read aloud w ith the confidence o f consecutive but planned speech, 11 presents fewer difficulties than its appearance on the printed page m ay suggest. T h e same is true o f m any other late Victorian and more recent poets who did not expatiate so fully on their prosodic theories. There is o f course no lack o f theories at present; perhaps there liave never been so m any serious attempts to find out w hat lies beneath the writing o f verse. T h ere is no place here for the explora­ tion o f different approaches. I f we stick to the simplest idea o f stress-metre we shall have to exclude m any im portant ideas for the sake o f simplicity. O nce we have looked carefully at the relationship of metre to language, however, the w ay is open for further work. There is a great deal o f argum ent about metrics, but general agreement that the subject can be explored linguistically and that it does relate to fam iliar speech. T h a t pure stress metre is ‘n atural’ to English is indicated b y its adoption in nursery rhymes and children’s com m unal play-rhymes: I I I I T h ere was an old w om an that lived in a shoe I I I I I W ee W illy W inkie runs through the town I I I I A pril F ool’s gone and past, I I I I Y o u ’re the biggest fool at last These clearly follow spoken stresses and not any regular pattern o f syllabic stress. Y e t their users are aw are that they are somehow ‘different’ from ordinary talking and that the satisfaction which they give demands greater attention to stress placem ent. In fact the child knows, o f course w ithout precise form ulation, that poetic composition rests on two sorts o f pattern. T here is the rhythm which gives the intonation b y w hich speech is accepted as ‘norm al’ in a national or local com munity, and there is the m etre w hich follows more precise patterns and can be given the codification o f ‘rules’ .

go

Linguistics and Literature

R hythm can, and most frequendy does, exist without metre: but m etre draws its being from the existence o f rhythm . In the divisions o f literature metre is considered an adjunct o f verse— a term which w ill suffice to m ake the technical distinction from prose without creating any difficulties about the boundaries o f poetry. Prose shares the quality o f rhythm , in the sense in w hich w e have been associating it w ith spoken intonation. A nything w hich is consistendy written in metre, then, ceases b y definition to be prose; but the appearance o f occasional metre in prose can be the result either o f chance or o f design: more often, perhaps, o f a spilling over from exceptionally sensitive awareness o f norm al rhythm . W e are most likely to become aware o f m etre in prose w hen reading aloud, but the pattern m ay be strong enough to present itself even in silent reading and to invite scansion: X

-

|

X

X

—

I

X

—

I

X

X

-

I

X

О poor mortals,/ how ye make/ this E arth bitter/ for each other (Carlyle, The French Revolution, V , 5) X

—

I

X

X

—

X

|

X

X

X

|

—

I

X —

N o m ore firing/ was heard at Brussels/ the pursuit/rolled miles aw ay (Thackeray, Vanity Fair, C h . 33) T h ere is need for m ore w ork yet on the questions o f metre in prose; so far no effective notation has been developed, so that analysis is com pelled to depend on the scansion used for verse, which soon proves too rigid for more than a b rief extract. G eorge Saintsbury in his time pursued the problem w ith vigour but his findings were over-elaborate and sometimes clearly idiosyncratic. It is pretty certain that closer linguistic study w ill cause some revision o f the dem arcation between verse and prose, as we learn m ore about the patterns o f speech rhythm . It is salutary to remember that understanding o f H ebrew poetry, with its characteristic patterning through types o f parallelism , was lost in translation until Louth rediscovered it some two hundred years ago. Even in the days o f most rigid classical prescriptivism, however, the poets themselves were seldom so far rem oved in their practice from ordinary speech rhythm as the Rom antics and their disciples m ay have believed. O u r leading poets have constantly, in their own terms, echoed the judgem ent o f R obert Graves: ‘O n e o f the most difficult problems is how to use natural speech rhythm s as variations on a m etrical norm ’ . ( The Crowning Privilege). W e can look back

Rhythm and Metre

gi

long before W ordsworth and Coleridge, to the best later m edieval poetry— to Chaucer who ‘knew that a poet could avoid dullness by using the rhythms o f com m on speech’.1 T h e com paratively barren period in literature which followed Chaucer, accom panied b y changes in the English language, led to the m etrical uncertainty o f W yatt and Surrey— though W yatt m ay have been m ore consciously concerned to bring the spoken rhythm into his work— and Skelton’s monotonous clutching at frequent stresses. E arly dram atic blank verse over-compensated by a series o f regular lines, ‘end-stopped’ so that sense and m etre generally pause together: A n d thus experience bids the wise to deal. I lay the plot, he prosecutes the point, I set the trap, he breaks the worthless twigs A n d sees not that wherewith the bird is limed. (Thom as K y d , The Spanish Tragedy, IH .iv) Here we are only too aware o f five stresses in each line, each alter­ nating with an unstressed syllable: it is a different kind o f verse from that o f the later Shakespeare and his contemporaries: W h y, ’ tis impossible thou canst be so wicked, O r shelter such a cunning cruelty, T o make his death the m urderer o f m y honour. T h y language is so bold and vicious, I cannot see w hich w ay I can forgive it W ith any modesty. (Thom as M iddleton, The Changeling, IH .iv) N or was the change felt only in dram atic verse; the Elizabethan lyric was often drawn towards the steady beat o f music rather than the looser rhythm o f speech; for example: U pon m y lap m y sovereign sits A nd sucks upon m y breast; M eantim e his love maintains m y life A nd gives m y sense her rest. Sing lullaby, m y little boy, Sing lullaby, mine only jo y . (R ichard Rowlands, ‘L u llab y’) 1 A. C. Partridge, The Language of Renaissance Poetry (London, 1971, Andre Deutsch), p. 29.

g2

Linguistics and Literature

:

T h e stresses follow the controlled movements o f the hand on the lute; it needed Donne to bring back to the lyric the pulses o f breath w ith minor syllables clustered around each emphasis: I wonder, by m y troth, w hat thou and I D id till we loved? were we not weaned till then? But sucked on country pleasures, childishly? O r snorted w e in the seven sleepers’ den? O th er periods o f literature can yield sim ilar examples o f poetry draw ing closer to speech rhythm or diverging from it. I f the divergence gives the kind o f poetry that can be called ‘artificial’, this need not correspond w ith a critical response o f good versus bad. W hether poetry should always be close to current speech is a question not to be answered by a simple axiom ; w hat is certain, however, is that m etre assumes an im portant function w hen it makes us most aw are o f underlying speech rhythm . It restores to poetry some of the im m ediacy that is lost b y form al graphological presentation. It works, w ith the syntactic and lexical features that we have discussed, to m eet expectation w ith surprise. For m etre can disturb the normal run o f emphasis, just as the breath-force in speech can be directed onto structural words which are norm ally unstressed. T h e stresses o f m etre can give unexpected prom inence to a syllable, and corresponding lightness to another. It can thus foreground items w hich have no apparent lexical or gram m atical support. In the following examples, the italicized syllables have m etrical prom inence which gives, respectively, strong negation; exclam atory appeal; and the contrast between a verb used successively in positive and negative form w hich is com m only found in speech: I I I I I No, no, go not to Lethe, neither twist I

-

I

I

-

X

X

I

X

X

I

Wolfsbane, tight-rooted, for its poisonous wine (Keats, ‘O d e to M elancholy’ ) X

I

X

I

—

X

X

I

X

I

M y m other bore m e in the southern wild, X

I

X

I

X

I

X

I

X

I

A nd I am black, but 0 , m y soul is white (Blake, ‘T h e Little Black B oy’)

Rhythm and Metre x

I

x

I

X

—

| —

x

93

|xx

T o be or not to be, that is the question (Hamlet, I ll.i)

In Shakespeare’s time, question was trisyllabic.) I'urther, this foregrounding can link items or anticipate con­ nections w hich m ight otherwise be missed. T here is no need to work nut the whole scansion to understand w hat is done b y m etrical nlrcss to aid progressive tension here: T h e first word that Sir Patrick read So loud, loud, laughed he; T h e next word that Sir Patrick read T h e tear blinded his e’e. (Anon., ‘Sir Patrick Spens’) or indeed in this more sophisticated example from C lough’s Amours C atholic U niversity o f A m erica Press).

9 Beyond the Sentence

Sentences have been described b y H. W einreich as ‘ the Hercules columns o f linguistics’ (Chatman: Style, p. 221), in recognition o f the fact that recent linguistic theory has tended to deal with no unit larger than the sentence. A gram m ar is regarded as satisfactory if it can generate all, and only, the acceptable sentences o f a language. There has been m uch attention to the phonemes and morphemes from which sentences are formed, com paratively little to sequence o f sentences. It is plain to everyone, linguists and non-linguists alike, that very little hum an com m unication through language is confined to isolated sentences. T h e concern o f literary criticism in particular is w ith the total text which constitutes a ‘w ork’ o f the author, whether it be as short as a Japanese haiku or as long as War and Peace. T h e exercise o f reducing a work, or a section o f it, to its basic components is a valuable means o f finding informed response to the whole. T h e ‘pieces’ o f a literary work are as interesting as the separate parts o f a com plex m achine and as essential for under­ standing. T h e student whose concern is m ainly literary m ay think that they are about as useless too, once they are parted from the total structure. In fact stylistics, w hatever style is being investigated, cannot proceed very far without recognition o f units above the sentence. Even the b rief examples used in previous chapters have sailed a little w ay beyond the Hercules pillars without, it is hoped, falling over the edge o f the discoverable world. A unit o f linguistic per­ formance w hich stands com plete in itself is com m only called a discourse. T h e nam e gives no inform ation about size, style or quality. A t the low er end o f the scale it can be a single im perative— ‘Stop!’— and the upper end is com pletely open as far as analysis is concerned, depending on factors o f planning and endurance w hich are not linguistic phenom ena. A discourse is the effective or, in H allid ay’s

Beyond the Sentence

101

description, ‘operational’ unit o f language, as the sentence is the syntactic unit. T o be o f more than im m ediate and limited value in com m unication, a sentence must stand in relation with other sentences. Y e t the sentence is not m erely a theoretical unit: we reach the fullness o f a discourse, however long, through a linear progression o f sentences encountered in the order w hich the performer gives them. T h e m ain reason w h y com paratively little work has been done on discourse is the difficulty o f creating linguistic ‘models’ from which a kind o f gram m ar o f discourse could emerge. W e simply do not know enough about how sentences build up into larger units. So at this point, o f vital im portance for stylistic study, we are left to use a good deal o f com mon intellegence about com munication, some o f the traditional approaches to whole texts, and the sense o f exciting research in progress. T h e reader m ay care to look at some o f the work done by Z . S. Harris, to whom we owe the term ‘discourse analysis’, and the following definition: Discourse analysis is a m ethod o f seeking in any connected discrete linear m aterial, whether language or language-like, which contains more than one elem entary sentence, some global structure characterising the whole discourse (the linear m aterial), or large sections o f it. Particulars o f the short but im portant work from which this quotation comes are given at the end o f the chapter. T h e reader should be warned that H arris works m ainly on technical rather than literary texts and that m uch o f his analysis has a forbiddingly diagram m atic appearance. It is, however, a pointer to the kind o f work w hich linguists are likely to attem pt m ore and more in the future. In dealing with the approach to whole works, the lack o f space for extended quotation is obviously a handicap. T h e next step is to m ake a few general observations, with only m inim al reference, in the hope that the reader w ill go to the texts to verify— or perhaps dispute— the assertions that are made. First— and this is commonsense but needs to be kept in m ind— a discourse m ay reveal m eaning and significance w hich is not appar­ ent in the isolated sentence. A sequence o f words which has the appearance o f an unacceptable sentence m ay prove acceptable in

юг

Linguistics and Literature

a com pleted structure. T h e foreign learner who says ‘I ’ll m ake m y possible’ for ‘I ’ll do m y best’ is told that he is being com pletely unidiom atic. But an interviewer o f candidates for admission to a course m ight well say at the end o f the proceedings, ‘I ’ll m ake m y possible acceptances into an alphabetical list.’ Literature gives sentences w hich yield only partial comprehension on first en­ counter but w hich show richer significance w hen the work is fully known. ‘ I thought the K in g had m ore affected the D uke o f A lb a n y than C orn w all.’ K en t’s words that open King Lear convey very little to the reader or auditor com ing to the play for the first time: they compose a gram adcally well-form ed sentence w ith obscure semantic reference. It requires five acts to reveal the full sense o f those proper nouns. N ext, any discussion o f recurrent linguistic features must take account o f the whole discourse i f it is to be at all useful. Leo Spitzer counselled the student to read over a text m any times until he recognized one or m ore recurrent stylistic idiosyncrasies, which he w ould then try to explain from the writer’s psychology.The final assault on the work was to return with this knowledge and look for its appearance in yet other features. This is a m ethod w hich perhaps leaves too m uch to the reader’s own psychological presuppositions; but it demands close and repeated attention to the whole discourse, without w hich no critical conclusions are w orth m uch. T h e pursuit o f a single word, phrase or im age through the discourse can be valuable in two ways. First it can, as Spitzer suggested, lead to better understanding o f the author’s whole achievem ent. Additional inform ation from outside the text, not necessarily only psychological, can illum inate the discourse and the discourse can in turn confirm w h at is known externally. Yeats uses the rare w ord ‘gyre’ in m ore than one poem and his usage links w ith non-poetic statements o f his personal m ythology in w hich cyclic notions play a considerable part. O r the occurrence o f the w ord ‘sweet’ more than thirty times in Shakespeare’s Sonnets leads to consideration not only o f his own affective response to life but also to the description o f him as ‘Sweet M r. Shakespeare’ in the contem porary Return from Parnassus and to the praise o f Francis Meres— ‘the sweet w itty soul o f O vid lives in mellifluous and honey-tongued Shakespeare . . . his sugared sonnets’. T h e significance o f the recurrent feature m ay be purely internal to the discourse, as when a character is given a keyword or catch-

Beyond the Sentence

103

phrase which identifies him and perhaps adds to his convincing existence. Dickens frequently works the trick— and indeed over­ works it— as with Captain C uttle’s ‘W hen found, make a note on’ and Miss M ow cher’s ‘volatile’ . In Iris M urdoch’s novel The Bell, the sixth-form schoolboy T o b y applies the adjective ‘rebarbative’ to a variety o f people, things and situations. His first use o f it seems unlikely and out o f character; only through the whole novel do we come to understand that he has recently added it to his vocabulary and that its repetition and pejorative sense precisely express his immature but observant and critical character. In this kind o f analysis statistical methods can be useful and have been employed particularly in attempts to check questions o f disputed authorship. It w ill be apparent that the pleremes (‘full words’ like nouns, verbs, adjectives and adverbs) w ill be o f m ain importance and we must not expect too m uch from recurrence o f the words in other classes— the kenemes— whose function is m ainly syntactic. Y e t these last cannot be neglected and have been found significant in some analyses. For instance, the definite article occurring with great frequency m ay indicate a desire to particularize and be specific rather than general: G eorge Rostrevor H am ilton found it so in his study o f m odern poetry, The Telltale Article. As well as the recurrence o f words and phrases, discourse analysis must consider syntactic recurrence and sentence-pattern. O ne exam ple o f the former is G eorge H erbert’s frequent use o f the firstperson past definite to open a poem; a glance at the index o f first lines in The Temple w ill confirm this and, I hope, im pel the reader to further investigation. H erbert had com e to this devotional serenity from years o f debate about whether to take holy orders; he continually sees him self in past action o f rebellion, defiance or lack o f faith. Another exam ple, w hich subsumes a good deal o f the R om antic response to life, is the frequency o f direct apostrophe to anim al or inanim ate themes in the two generations o f the English R om antic poets. For recurrent patterns o f sentence structure, the reader is referred to studies cited at the end o f the chapter. Since we are again dealing w ith the sentence, it is im portant to rem em ber that the linguistic content o f a sentence m ay not con­ form to the graphological realization o f a sequence o f words starting w ith a capital letter and ending with a full stop. M an y o f these w ritten ‘sentences’ are really two or more sentences in their deep structure. H arris gives the exam ple from his analysis o f T h u rber’s story ‘T h e V e r y Proper G an d er’ in which the sentence:

Ю4

Linguistics and Literature

H e was strutting in his front yard, singing to his children and his wife is a transform of: H e was strutting in his front yard, he was singing to his children and his wife and the second part is a further transform of: H e was singing to his children and he was singing to his wife. This kind o f understanding helps us to determ ine w hat structures and patterns really make up a discourse under consideration. Readers who are fam iliar with Chom sky’s methods o f transformation m ay wish to use them; but simpler methods o f arriving at the kernels o f m eaning can be stylistically perfectly valid. N ow it is clear that, whether or not a given sentence can be divided, a sequence o f sentences must be connected to create a discourse. Everyone is accustom ed to recognize and nam e units beyond the sentence: the paragraph in prose, the stanza in verse. T h e existence o f verse-paragraphs as stylistic units is also recognized; such divisions are often m arked graphologically in m odern poetry and can be discerned in the long poems w ithout stanza-divisions o f M ilton, D ryden, Pope, W ordsworth, Browning and a great m any others. Beyond these w e have the divisions proper to the different literary ‘kinds’ : chapters, scenes and acts, cantos and books. E very unit depends on two factors: connection and silence. T o take connection first, we have the opinion o f Coleridge that ‘a close reasoner and a good w riter in general m ay be known b y his connectives’ . W ithout going so far, it is apparent that connectives are am ong the essential features o f discourse: a random order o f sentences w hich were themselves well-formed w ould be meaningless. E ach sentence in a discourse is a step forward in the linear m aterial w hich H arris describes: it is also a glance back at w h at has just been form ulated. U nderstanding depends on overt or concealed reference to the precedent. T here is no established list o f connectives; how m an y are observed depends largely on the precision with w hich m inor differences are categorized. H ere are some o f the most frequent, w ith examples taken from G eorge O rw ell’s Nineteen EightyFour. T h e y are not m eant to prove anything about O rw ell’s w riting,

Beyond the Sentence

105

for which a m uch fuller survey would be required; it is, however, reasonable to suggest that a good range and variety o f connectives is a m ark o f quality. 1. Conjunctions and conjunctive adjectives such as however, furthermore, nevertheless. T h e appearance o f a simple co-ordinating conjunction at the start o f a sentence is frowned on b y some critics, but in defiance o f m any acceptable examples: I f you m ade unexpected movements they yelled at you from the telescreen. But the craving for food was grow ing upon him. 2. Pronom inal linkage w ith a preceding noun: W inston did not bu y the picture. It w ould have been an even m ore incongruous possession than the glass paperweight. 3. R epetition o f a keyword or proper nam e, either identically or in a different gram m atical form: W inston was gelatinous w ith fatigue. Gelatinous was the right word. O n e d ay a chocolate-ration was issued. T h ere had been no such issue for weeks or months past. 4. U se o f a synonymous or related word or phrase: H e knew that sooner or later he w ould obey O ’Brien’s summons. Perhaps tomorrow, perhaps only after a long delay— he was not certain. 5. D eictic words— ‘pointers’ like the, this, that— either governing a noun or referring back to the whole statement: W hen he cam e back his m other had disappeared. This was already becom ing norm al at that time. 6. R epetition o f opening structure (there was an exam ple o f this from Dickens in C hapter 5): He confessed to the assassination o f em inent P arty members . . . He confessed that he had been a spy . . . He confessed that he was a religious believer . . . He confessed that he had murdered his

іоб

Linguistics and Literature

wife . . . He confessed that for years he had been in personal touch w ith Goldstein. 7. Class-member relationships, or relationship o f the parts o f a referent to the whole: W inston picked his w a y up the lane through dappled light and shade, stepping out into pools o f gold w herever the boughs parted. U n d er the trees to the left o f him the ground was misty with bluebells. 8. Looser semantic connection without repetition o f items: E very thing faded into mist. T h e past was erased, the erasure was forgotten, the lie becam e truth. 9. C lear sequence o f events in successive time: A hand fell lightly on his shoulder. H e looked up. T h e other element in division o f units is silence: the negative pole o f the current, as connection is the positive. T h e gap on the page, small and unnoticed between sentences, more apparent between paragraphs and unm istakable between chapters and similar large units, corresponds to the breaks in experience o f linguistic com munication. In all our encounters w e find similar short or long vacuities: the pause for breath, the interval for refresh­ ments, the retirem ent to sleep. N o discourse goes on for ever, though the reasons against its doing so are not linguistic. There w ill be a beginning, interm ittent pauses and an end. A n apparent exception is found in a w ork like J o yce’s Finnegans Wake, starting graphologically half-way through a sentence and ending with the first h a lf o f the same sentence, so that reading could begin anywhere and continue indefinitely. B ut this is a planned deviation, m otivated b y the Joycean obsession w ith cyclic return. T h e reader w ill w ant to attem pt his own analysis o f texts, en­ couraged b y the fact that there is no rigid system which must be followed. A n y exam ination w hich increases understanding and response w ill be w orth while. H ere, by w ay o f exam ple and not o f prescription, is a possible approach to the first two paragraphs o f D . H . Law ren ce’s Aaron's Rod (1922). It is not suggested, in this

Beyond the Sentence

107

or in other attempts at discourse analysis, th a t the writer worked out and deliberately incorporated all the features which can be extrapolated from his w ork. T h e skilful h an d lin g o f any medium will develop the best possibilities that the m edium contains. R ichard Aldington considered this novel to be one o f those works o f Law rence which w ere ‘w holly im provisations, begun a t random, w ith no more coherence and structure than the v e ry im portant ones o f L aw rence’s com pelling personality and brillian t writing’ . H elping us to understand w hat m akes for ‘brilliant w ritin g’ is a function o f stylistics. T h ere was a large brilliant evening star in the early twilight and underfoot the earth was half-frozen. I t was Christmas Eve. Also, the w ar was over, and there was a sense o f relief that was almost a new m enace. A m an felt the violence o f the nightmare released now into the general air. Also there had been another w rangle am ong the m en on the pit-bank th at evening. A aro n Sisson was the last m an on the little black railway-line clim bing the hill home from work. H e was late because he had attended a m eeting o f the m en on the bank. H e was secretary to the M iners’ U nion for his colliery, and had heard a good deal o f silly w rangling that left him nettled. W e are not out o f order i f we take into account information about Law ren ce’s philosophy and our reading o f the whole novel— for intelligent discourse analysis is not likely to follow from the first sentence-by-sentence reading o f the text. W e can bear in mind the theme o f contrast between the individual and society, o f a m an finding his true self against its pressures. i. W e look at the connectives, identifying them by the numbers already used with the examples from O rw ell. T h ere was . . . (opening impersonal phrase o f situation) It was Christmas E ve (semantic and form al link with ‘evening’ - 4)

Also . . . (conjunction— 1) . . . the violence o f the nightm are . . . (looser semantic link w ith ‘a new m enace’— 8) A lso . . . (conjunction— 1) A aro n Sisson . . . (opens new paragraph, with semantic and form al connection ‘m en’/‘m an’— 4) H e was late . . . (pronominal linkage— 2) H e was secretary . . . (pronominal linkage— 2)

ro8

Linguistics and Literature

Thus, looking at the full graphological sentences without seeking their com ponent sentences, w e find an interesting distribution o f connectives. 2. Parallels o f certain features can be seen to link the two p ara­ graphs: . . . large, brilliant evening star . . . . . . little, black railw ay line . . . (syntagm atic parallels with lexical items in contrast) . . . u n d erfo o t. . . . . . clim bing . . . (loose semantic connection and contrast o f items im m ediately following the parallel structures, setting the individual in the overall scene.) . . . w rangle . . . . . . w rangling . . . (formal and semantic parallels link the conclusion o f each paragraph). 3. Items from a com mon semantic area give unity and location: . . . pit-bank . . . . . . bank . . . . . . colliery . . . 4. T h e single item man is twice juxtaposed with its plural form in successive sentences, the first time with impersonal reference and the second time with personal. 5. W ords and phrases giving a sense o f time are used w ith semantic contrast: 1st. paragraph— evening/early twilight/Christmas Eve/that evening = expectation 2nd. paragraph— last man/late = disappointment 6. Syntactic parallels in two successive sentences help to fore­ ground the antecedent item ‘A aron Sisson’ : H e was late . . . he had attended . . . H e was secretary . . . he had heard . . . 7. These features all contribute to the pattern o f the two p ara­ graphs w hich go from the general to the particular, from the

Beyond the Sentence

iog

impersonal to the personal, setting the individual ‘A aron Sisson against the cold, disturbed background. T h e first paragraph is bu ilt up o f sentences whose m ain verb, stated or concealed, is the past definite was: (i) (ii) (iii) (iv) (v) (Vi) (vn)

there (it) it the W ar there that (there)

was was was was was was (was)

a star early twilight Christmas Eve over a sense o f relief almost a new menace a feeling o f violence

Item (vii) is the underlying structure o f the last sentence, with ‘a man’ used in an accepted, if now old-fashioned, impersonal sense. The appearance o f ‘m an’, however, leads into the changed tense o f the next sentence, which is the ‘hinge’ between the two paragraphs: there

had been

another wrangle

T h e second paragraph is dom inated b y A aron Sisson as the subject o f each sentence; again there is first a sequence o f was: A aron Sisson

was (was) was was (was)

the last m an clim bing the hill late secretary nettled

but twice the progression is broken with the pluperfect form which switches attention back to the ‘w rangle’ o f the first paragraph: A aro n Sisson

had had

attended a m eeting heard . . . wrangling

This is a simple analysis o f a passage from w hat would by com mon consent be called a novel. W hile leaving plenty o f room for dispute about nom enclature and assignment o f particular texts, critics ever since Aristotle have recognized the notion o f ‘kinds’ or ‘genres’ in literature. These raise their own expectations that must affect the response: they have their own conventions or norms w hich m ake further categories within the literary style. T h e sense

no

Linguistics and Literature

o f w h at is proper to a particular genre is by no means confined to classical, m edieval or Renaissance criticism, or to the products o f a period w hen notions o f ‘poetic diction’ were dominant. H ere is a vast subject which soon spreads beyond stylistic con­ siderations. But, for exam ple, we accept the m e o f first-person or third-person narrative in the novel, or the alternation which occurs in a book like Bleak House. Continuous second-person address w ould be considered deviant, though its occasional appearance with direct address to the reader was readily accepted in the fiction o f the eighteenth and nineteenth centuries, and regular in the epistol­ a ry type like Humphry Clinker. It has been extensively used in this century b y Beckett and Genet. W e accept that the ‘I ’ o f poetry is not alw ays a predication o f the poet’s ‘real’ experience or opinion, but m ay be: (i) the im aginary statement o f a created character: ‘ I am m onarch o f all I survey’ (Cowper) (ii) dram atic statement o f the poet’s inner experience: ‘ I struck the board and cried, “ N o m ore!” ’ (Herbert) (iii) a generalized state o f m ind given particularity: ‘ I wake and feel the fell o f dark, not d a y ’ (Hopkins) (iv) actual experience narrated, perhaps verifiable from other texts: ‘ I wandered lonely as a cloud’ (W ordsworth) and m an y other things beside. A gain , dram a w ill dispense w ith the ‘he said’ type o f interpolation required in narrative fiction, while the appearance o f dialogue form w ith characters’ names and ‘stage directions’ in a novel, as in the N ighttow n episode in Ulysses, is regarded as deviant. In both fiction and dram a w e not only accept but positively require the shifting o f registers in order to differentiate characters and their situations. In poetry the abrupt switch o f register causes surprise though (as w e have seen) not necessarily condem nation. C an we go beyond the genre-expectations o f a given literature? O u r examples have been taken mostly from literature written in English over the last four hundred years— a noble but tiny fragm ent of the w orld’s literature. T o attem pt further exploration stretches the resources o f stylistics beyond present development. Y e t the possibility o f quest into regions now obscure is there. Linguists seek

Beyond the Sentence

in

continually for the universal principles underlying language as a human activity: schools o f com parative literature proliferate. O n e o f the foundation stones o f m odern linguistics was the discovery o f languages different in m any essentials from those previously studied b y scholars and the realization that valid statements about language cannot exclude any languages w hich m en use. T h ere are already pioneering efforts, like the work o f V . Propp on the folk-tale, classifying plots found in m any speech-communities through structure and syntax. Better known, though adm ittedly controversial, is the comprehensive vision o f Claude Levi-Strauss, an anthropologist who has been m uch influenced b y Saussurean linguistics. H e seeks a com m on principle in all hum an thinking, as expressed in m yth, custom and ritual— manifestations with which literary critics too have becom e concerned. R olan d Barthes leads the ‘semiologists’ who seek, and sometimes claim to possess, a key to all hum an ‘signs’ , not solely in linguistic codes. Little is yet established; but w e m ay reasonably wonder i f the existence o f similar ‘kinds’ in literature w idely separated in time and space m ay not yield some clue through discourse analysis. Som e o f the kinds named b y Aristotle are still viable to-day, and the em ergence o f epic poetry and dram atic action w ith dialogue can be traced in cultures o f w hich he knew nothing. It m ay be that there is an approach to be found through stylistics to fresh understanding o f the w ay in which hum an beings verbalize their experience. Conjecture points to the future. For the present, a last w ord o f rem inder that the study o f larger units must incorporate techniques used on smaller segments, while at the same time introducing new considerations. O n w hatever scale, and in whatever langue, a literary text is always a piece o f linguistic performance. A nd this, like the last page o f Finnegans Wake, is where w e cam e in. T h e reader will, I hope, continue his study o f literary texts knowing that close reading and dissection do not destroy, but rather enhance, the precious gift o f delight.

F U R T H E R RE AD IN G

V arious kinds o f discourse analysis are carried out in m any o f the books and articles already recommended. T h e w ork o f Z. S. Harris mentioned on p. 101 is to be found in Discourse Analysis Reprints (The H ague, 1963, M outon) and is im portant but not easy.

112

Linguistics and Literature

Good shorter studies are: R . O hm ann, ‘Literature as Sentences’

0Chatman: Essays, pp. 231-8); M . Riffaterre, ‘ Stylistic context’ {Chatman: Essays, pp. 4 3 1 -4 1 );} . M cH . Sinclair, ‘T a k in g a Poem to Pieces’ {Fowler, pp. 68-81); М . A . K . H alliday, ‘Linguistic Function and L iterary Style: an In quiry into the L anguage o f W illiam G olding’s The Inheritors' {Chatman: Style, pp. 330-68); F. S. Scott and others, English Grammar (London, 1968, Heinem ann), pp. 203-11. O n the application o f statistics to stylistics see: G . U . Y ule, The Statistical Interpretation o f Literary Vocabulary (London, 1944, C am ­ bridge U niversity Press); and L . Dolezel and R . W . Bailey, Statistics and Style (New Y ork, 1969, Elsevier). C . Levi-Strauss is best approached through his Structural Anthro­ pology (N ew Y ork, 1963, Basic Books and London, 1968, Penguin Press), especially C hapter 2, ‘Structural Analysis in Linguistics and in Anthropology’, and C hapter 11, ‘T h e Structural Study o f M yth ’ . A general introduction to his work is E. Leach, Uvi-Strauss (London, 1970, Collins); see also G . Steiner, ‘ Orpheus with his M yths: Claude Levi-Strauss’, pp. 248-60 o f Language and Silence (Harmonds­ worth, 1969, Penguin Books). T h e work o f V . Propp m entioned on p. 111 is not easily accessible in English; a translation o f his ‘ M orphology o f the Folk T a le ’ appeared in the International Journal o f American Linguistics (The Research Center in Anthropology, Folklore and Linguistics, Bloom ington, Indiana, 1958); there is a brief critique o f Propp in R . Jakobson, ‘ Linguistics and Poetics’ {Chatman: Essays, pp. 323-36). T h e more difficult fields o f general linguistics which are briefly referred to in the foregoing chapter soon pass beyond the scope o f this book. Those who wish to venture farther m ay start with: J. H. Greenberg, ed., Universals o f Language (Cam bridge, Mass., 1963, Massachusetts Institute o f Technology Press); and B. L . W horf, Language, Thought and Reality (Cam bridge, Mass., 1956, Massachusetts Institute o f Technology Press).

Glossary of Linguistic Terms

Accent: prom in ence o f a spoken sylla b le; also a distinctive p ro­ n un ciatio n w ith social or region al characteristics. Affix: elem ent ad d ed to a base to form a w ord , in itially a prefix as ияаЫе, fin a lly a suffix as k ind n m . Alliteration: repetition o f in itial consonantal sounds as yast and yU rious’ , ‘w h en to the sessions o f jw e e t silent th o u gh t’ . Ambiguity: doubtful m ea n in g o f sentences sharin g the same surface structure b u t w ith d ifferen t deep structures. Anacoluthon:

ch an ged or incom p lete g ra m m atica l sequence.

Analytic: (lan guage) d ep en d in g m ain ly on w ord -ord er to show sy n tactic relationships. Aphasia: loss of, or defect in, the fa cu lty o f speech and com ­ prehension. Archaism:

w ord or gra m m a tica l construction no longer in use.

Assonance: half-rhym e b y rep eatin g a vo w el w ith different final consonants: late, make. Case: v a ria b le form o f noun, pronoun or ad jective, sh ow in g its syn tactic relationship to oth er w ords in the sentence. Clause: structure co n tain in g a finite verb , form in g eith er a sentence or a constituent p a rt o f a sentence. Code:

set o f sym bols w ith agreed com m u n icative valu e.

Colloquial: inform al, fam iliar u ttera n ce; ge n e ra lly spoken as con ­ versation but som etim es w ritten , as in personal letters. Competence: speaker’ s total know ledge o f a a cq u a in ta n ce w ith its system (C hom sky).

lan gu a g e

th rough

Complex sentence: sentence consisting o f one p rin cip le clause and one or m ore subordin ate clauses.

114

Linguistics and Literature

Connotation: m ean in g in con text, as related to associated ideas and em otions. Consonance: half-rhym e b y rep eatin g differen t p reced in g v o w e ls : fit, cut.

a

final

consonant

w ith

Consonant: speech sound u ttered b y p a rtial or com plete obstruction o f the ou tgo in g a irstream ; letter representing such a sound. Corpus: b o d y o f lingu istic perform ance from w h ich ab o u t u sage are d raw n .

deductions

Deep structure: u n d e rly in g syn tax o f a sentence, y ield in g m ean in g w h ich m a y be con cealed b y its surface structure. Deictic:

(w ord) w ith function o f po in tin g or dem on strating.

Denotation: m ean in g th rough d irect reference w ith o u t extraneous association. Deviation: linguistic u sage considered exp ectation s o f users o f the lan gu ag e. Diachronic:

to

d ep art from

nbrm al

relatin g to stu d y o f lan gu a g e in historical perspective.

Diacritical: (m ark) ad d ed to a letter to m ake it stan d for a differen t sound, or to in trod u ce inform ation not otherw ise shown in the text. Dialect: distin ctive speech o f a grou p w ith in a lan gu a g e-co m ­ m u n ity, determ in ed b y region, class or occu pation . Discourse: linguistic p erform ance stan din g as a u n it com plete in itself, irrespective o f len gth . Ellipsis: om ission o f p a rt o f an u tteran ce w h ich c a n be rea d ily understood and supplied b y the recipien t, e.g. ‘ these apples are b etter th an those’ (sc. apples). Etymology:

stu d y o f d eriva tion o f form and m ea n in g in w ords.

Foregrounding: utteran ce.

stylistically g iv in g special p rom in en ce to p a rt o f an

Generative grammar: system o f g ra m m ar offerin g ex p licit rules from w h ich all, an d only, the po ten tial sentences o f a lan gu a g e m a y be form ed. Graphology: system o f w ritin g , in clu d in g spelling, p u n ctu a tio n and p a rag ra p h in g.

Glossary o f Linguistic Terms Holophrase:

115

group o f w ords understood as a single un it o f m eanin g.

Idiolect: d istin ctive usage o f a single m em b er o f a lan gu agecom m u n ity. Idiophone:

distin ctive sound in an idiolect.

Immediate constituent analysis: progressive division o f sentences into sim pler segm ents un til the basic elem ents are isolated. Inflection: va riab le elem ent in a w ord , m a rk in g its gra m m atica l fun ction in a sentence. Intonation:

pattern o f the rise and fall o f pitch in speech.

Isochronism: prin ciple o f d ivid in g spoken utteran ces into segm ents o f a p p ro x im ately eq u a l d u ration . Keneme: ‘em p ty w o rd ’ , existin g on ly as p a rt o f the gra m m ar o f a lan g u a g e and m eaningless in isolation. Kernel sentence: one o f the g ro u p o f basic sentences in a lan gu ag e from w h ich all oth er scntenccs o f that lan gu a g e m a y be d erived b y the rules o f tran sform ation al-gen erative gram m ar. ‘ Langage':

the hum an fa cu lty o f speech in gen eral (de Saussure).

‘ Langue': the total con ten t o f a lan gu a g e shared b y a com m u n ity o f speakers (de Saussure). Lexicon: V o c a b u la r y : the total resource o f w ords a va ila b le to users o f a lan guage. Liquid:

C on sonan t articu lated w ith ou t ob stru ction or friction.

Metre: regu lar pa ttern in g in verse, based on syllab ic stress or q u an tity . Morpheme:

M in im u m form al un it o f m eanin g.

Morphology: lan gu ag e. Neologism:

arran gem en t o f m orphem es in the gram m ar o f a

N ew w ord in trod u ced into the lexico n o f a lan gu ag e.

Onomatopoiea: Paradigm:

im itation o f n atu ra l sounds, e.g. bang, quack.

set o f form s representing form al chan ges in a w ord-class.

Paradigmatic: (relationship) b etw een an item in an u tteran ce and oth er item s o f the sam e w ord-class p o ten tia lly a v a ila b le in that position.

п6

Linguistics and Literature

Paralinguistic: (signs) co m m u n ica tin g w ith o u t a rticu latin g norm al arran gem en ts o f phonem es, e.g. sighing, lau gh ter. ‘Parole':

specific u tteran ce, a ct o f speaking or w ritin g (de Saussure).

Performance: a ctu al as opposed to poten tial u tteran ce b y a lan gu ageuser (C h o m sk y ); sim ilar to parole. Phatic: (lan gu age) used to establish relationship rath er than for com m u n ication o f m essage. Philology: trad ition al d iach ro n ic w ritten texts.

study o f lan gu a g e,

based

on

Phoneme: m inim u m significan t sound-unit, c a p a b le o f con veyin g a ch an ge o f m ean in g w h en one is replaced b y another. Phonemics: id en tification and description o f the phonem es o f a lan gu a g e. Phonetics: analysis, description and classification o f speech-sounds, n ot confin ed to a single lan gu ag e. Phonology: stu d y o f the phonem es o f a lan g u a g e as form ing a system o f speech. Plereme: ‘full w o rd ’, referrin g to an id en tifiab le ob ject, action or q u ality . Plosive: C on son an t prod u ced b y obstruction o f airstream follow ed b y sudden release. Polysemy: M u ltip le m eanin g, e.g. light as ‘p a le ’, ‘b rig h t’, ‘not h e a v y ’ , ‘triv ia l’ . Prosody:

stu d y o f versification.

Prosodic: referrin g to p roso d y; also, the elem ents in speech other than phonem es, e.g. stress, pitch , volu m e. Received pronunciation: type o f British speech cu rren tly reckoned as stan dard, show in g no specific d ialectal features. Referential: ‘d ictio n a ry ’ m ea n in g o f a w o rd : its referen ce to som e­ thin g in the w orld o f experience. Register: lingu istic perform ance show ing distin ctive features chosen w ith rega rd to extern al circum stances. Rhythm: distin ctive p attern o f stresses and in ton ation in a lan gu age. Semantics:

stud y o f m eanin g.

Semiology:

stu d y o f sym bolic systems, in clu d in g lan gu age.

Glossary o f Linguistic Terms Sentence:

ny

m inim um com plete utterance.

Sibilant: consonant w ith hissing sound caused b y pressure o f the ton gu e on the hard palate. Sociolinguistics: structures. Stress:

study o f la n g u a g e in relation to social groups an d

relative force o f b reath in u tteran ce o f syllab le or w ord.

Structuralism: system .

study o f la n g u a g e as a coh eren t and interdependen t

Surface structure: a p p a ren t syn tactic structure o f a sentence, not a lw ays yield in g the tru e m ea n in g o f the sentence. Syllable:

grou p o f phonem es uttered w ith one b reath pulse.

Synchronic: Synonym: Syntax:

relatin g to stu d y o f lan gu a g e a t a single po in t o f tim e. w ord precisely or n ea rly the sam e in m ean in g as another.

relationship o f w ords to one an oth er in form ing sentence.

Synthetic: (lan guage) d ep en d in g m ain ly on inflections to show syn tactic relationships. Transformational grammar: system w hich allow s the conversion o f deep structures into surface structures b y ordered stages. Typographical: Voiced: Voiceless:

R e la tin g to printin g.

(sound) prod u ced w ith vib ration o f the vo ca l cords. (sound) p rod u ced w ith o u t vib ra tio n o f the vo ca l cords.

Vowel: speech sound u ttered w ith o u t ob stru ction or friction in the ou tgo in g airstream ; letter representing such a sound.

Index

Addison, Joseph, 16 Aldington, Richard, 107 alliteration, 14, 61, 88-9 ambiguity, i, 2, 64-6, 68 anadiplosis, 80 anaphora, 79 Arnold, Matthew, 3, 53 assonance, 14, 38 Auden, W . H ., 77, 83 Austen, Jane, 42 Austin, Alfred, 54 Bally, Charles, 4, 42 Barthes, Roland, 111 Beckett, Samuel, n o Betjeman, John, 54 Bible, Authorized Version of, 63 Blake, W illiam, 70, 83, 92 blank verse, 91, 96-8 Bloomfield, Leonard, 4, c;8 Bright, John, 28 Browning, Robert, 21, 51, 94, 104 Buffon, Georges, 15 Bunyan, John, 3 Bums, Robert, 78 Butler, Samuel, 17 Byron, Lord, 27, 95 Carlyle, Thomas, 90 Carroll, Lewis, 37, 60 Cassirer, Ernst, 28 Chaucer, Geoffrey, 21, 49, 91 Chesterton, G . K ., 69, 82 Chomsky, Noam, 1, 9, 46, 55, 59, 72, 104 Cicero, 78 Clough, A . H ., 93, 95 Coke, Edward, 23

Coleridge, Samuel Taylor, 72, 88» 91, 104 connectives, 103-4, 107-8 connotation, 23, 74 Cowper, W illiam, n o Crabbe, George, 95 Cummings, E. E., 40 deep structure, 56, 64, 103-4 de la Mare, W alter, 38 Dennis, John, 61 denotation, 23, 74 deviation, stylistic, 1, 4-5, 13-15, 25-7> Зб, 45-56, 58, 62-3, 68-70, 75. 77» 79. 93. 95. 106, n o dialect, 14, 18, 28, 35, 46, 48, 86 Dickens, Charles, 16-17, 23, 38, 40-41. 52, 70, 77-9, 82, 103, 105,

110 discourse analysis, 10 0 -in Donne, John, 15-16, 80, 92, 94 Dowson, Ernest, 80 drama, 17, 34, n o - 11 Drayton, M ichael, 77 Dryden, John, 21, 68, 76, 104 Eliot, George, 65 Eliot, T . S., 18, 59-60, 79-80, 96 Empson, W illiam, 64 epistrophe, 79 epizeuxis, 80 figurative language, see rhetoric, figures of Firth, J. R ., 62 Flecker, James Elroy, 82 foregrounding, 48, 51, 63, 78-9, 87,

93. 108

Index Forster, E. М ., 6g, 83 Fowler, H. W . and F. G ., 61 free verse, 36-7, 96-8 Frye, Northrop, 73 Garvin, F. L ., 48 Gascoyne, David, 98 Genet, Jean, n o genre, literary, 3, 104, 109-1 1 Gilbert, W . S., 79 graphology, 33, 35-7, 39-41, 44, 55, 85-6, 92-3, 103-4, 106-7 Graves, Robert, 90 G ray, Thomas, 15 Green, Evelyn, 27 Halliday, М . A . K ., 48, 100 Hamilton, G . R ., 103 Harris, Z . S., 101, 104 Herbert, George, 37, 67, 103, n o historical linguistics, 21-30, 37, 50,

66 Hobbes, Thomas, 3 Hopkins, Gerard Manley, 1, 2, 26, 50-51, 88-9, 110 Housman, A . E., 17, 24, 26, 67 Hughes, Ted, 7, 97 hyperbole, 78 idiolect, 18, 46, 86 intonation, 34, 36, 86, 98 isochronism, 86, 95, 97 Johnson, Samuel, 6, 16, 23 Joyce, James, 18, 42, 51, 55-6, 58, 106, I I O- I I Keats, John, 2, 24, 48, 74, 76, 78, 92, 94 Keyes, Sidney, 70, 76 Kingsley, Charles, 16 Kipling, Rudyard, 76 K yd , Thomas, 91, 93 Langland, W illiam, 88 Lawrence, D. H ., 106-9 legal style, 9 -11, 13, 15, 29, 67-8 L6vi-Strauss, Claude, 111 lexis, 16-17, 22-3, 44-5, 47, 53, 56, 58-70, 75, 83-4, 92, 96, 108

I IQ

literature, definitions of, 1-6, 13-15, 27> 61, 63, 72 litotes, 78 Lowth, Robert, 90 Luther, M artin, 28 Lytton, Bulwer, 78 M acaulay, Lord, 48 M acGonagall, William, 54 Marlowe, Christopher, 74 Marvell, Andrew, 81 meaning, see semantics meiosis, 78 Meredith, George, 51 Meres, Francis, 102 metaphor, 67, 69, 75-7, 81-4 metonymy, 78 metre, 14, 36, 61, 86, 88-98 Middleton, Thomas, 91, 93 Milton, John, 3, 35, 76, 83, 95, 97, 104 Monboddo, Lord, 30 Montrose, Marquis of, 24 morphology, 48, 50-51 Murdoch, Iris, 103 myth, 111 Nash, Ogden, 50 novel, 17, 39, 109-10 nursery rhymes, 89 onomatopoeia, 37-8 Orwell, George, 47, 104-7 Owen, Wilfred, 37 Palmer, Frank, 6 paradigmatic relation, 45-7, 59-61, 68-9, 75, 78, 83-4 paragraph, 35, 104, 106-9 Partridge, A . C ., 91 Pater, W alter, 42 Peacham, Henry, 73 phonology, 24, 33, 36-9, 55, 58-61, 64-5, 85-7, 93 polysemy, 66-8 Pope, Alexander, 16-17, 24, 39, 51,

95, 104

Prague School, 48, 62 prescriptivism, 16, 19, 25, 45-6, 61, 63, 74, 90

1 20

Linguistics and Literature

Propp, V ., h i prosody, 63, 85-98 punctuation, 35, 41, 86 Puttenham, George, 73 Quiller-Couch, Arthur, 12 Quirk, Randolph, 84 • Racine, Jean, 95 Ralegh, W alter, 23 Reed, Henry, 18, 68 register, i i , 15-19, 23, 27, 34, 37» 46, 48-9, 62-3, 66-9, 74, 77, 80-81, 96, n o religious style, ІО-ІІ, 13, 15, 26, 29, 49, 74, 80 rhetoric, figures of, 14, 51, 72-84 rhyme, 14, 24-5, 37, 46, 61, 93 rhythm, 36, 86, 88-92, 94-7 Romantic movement, 16, 64, 73, 88, 103 Rowlands, Richard, 91, 93 Saintsbury, George, 90 Saussure, Ferdinand de, 4, 9, 21, i n scansion, 36, 86, 88, 90, 93, 95-7 schemes, 75, 79, 83 scientific style, 12-13 Scott, F. S., 44 semantics, 22-3, 47, 52, 59, 62-7, 106, 108 sentence, 33, 44-6, 88, 100-07 Shakespeare, William, 21, 23-4, 26-7, 39. 46, 49-50, 52, 58, 60, 63, 65» 67, 73, 78-9, 81-2, 91, 93, 96-7, 102 Shaw, George Bernard, 40 Shelley, P, B., 46 Shirley, James, 78 Sidney, Philip, 52 silence, 35, 104, 106 simile, 51, 70, 75-6, 82 Sinclair, Upton, 72 Skelton, John, 91 Smollett, Tobias, n o sociolinguistics, I I speech, 18, 32-43, 46, 48, 52, 58, 74, 80, 85-92, 94-7

Spender, Stephen, 76 Spenser, Edmund, 15, 60, 82 Spitzer, L., 102 Steiner, George, 62 stress, 34, 36, 85-9, 92-3 Stubbes, Philip, 23 style, definitions of, n -1 3 , 25, 73 surface structure, 55-6, 64-5 Surrey, Earl of, 91 Swinburne, A . C ., 46 symploce, 79 synecdoche, 77-8 syntax, 44-56, 58-61, 63-5, 68-9, 75, 79-8o, 83-4, 88, 92-3, 95-6, 103, 108, i n Tennyson, Lord, 32, 38, 95 Thackeray, W . М ., 64, 69, 90 Thomas, Dylan, 1, 2, 26, 50, 56, 70, 82 Thurber, James, 103 Tiller, Terence, 56, 76 transformational-generative gram­ mar, 55-6, 104 Trollope, Anthony, 41 tropes, 75, 79-81, 83 Verlaine, Paul, 82 Waller, John, 61 Wallis, John, 30 W arner, Rex, 80 Webster, John, 78 Weinreich, H., 100 Wells, H. G ., 36 Whitman, W alt, 66 Wilde, Oscar, 48 Williams, Charles, 98 Williams, W illiam Carlos, 54 Woolf, Virginia, 42, 75 words, see lexis Wordsworth, W illiam , 15, 54, 64, 69, 75. 9 1» I04> 110 W yatt, Thomas, 91 Yeats, W. B., 37, 53, 69, 82, 102 Zola, Emile, 72

This lively introductory study of stylistics - the lin ­ guistic study of different styles, in a sociological or literary sense - is written for those who are interested in any aspect of literary criticism or linguistics. Consider­ able antagonism has been generated by recent applica­ tions of linguistic techniques to the field of literature: much of this work has failed to find a wide acceptance in departments of literature, and moreover many linguists feel that literature is not a proper area for linguistic study. In this book Raymond Chapman shows that the tw o disciplines can illuminate each other in many ways: that linguistic analysis can make a precise and stimulating contribution to literary criticism, and that literature provides a rich and varied field for linguistic study that does not in any way reduce linguistics itself to a mere technology or 'service station'. He views literary lan­ guage as a distinctive 'style' - or indeed a number of styles - which nevertheless springs always from the common core of language and can be investigated by the same techniques as are applied to other languagestyles. His examination of various areas of stylistic analysis is supported by specific reference to literary works over a wide range of periods and authors: 'Any useful stylistics must be valid in all literary traditions. Its exponents have a similar duty to that of the grammarian : to offer a system which can accommodate all acceptable realizations, w ithout taking refuge behind the shelter of ''exceptions''.'

95p net

Edward Arnold

ISBN . 0 7131 5686 4