peter fonagy the development of psychopathology from infancy to adulthood

A

R

T

I

C

L

E

THE DEV DEVELO ELOPM PMENT ENT OF PSY PSYCH CHOPA OPATHO THOLOG LOGY Y FRO FROM M INFAN INF ANCY CY TO AD ADULT ULTHOO HOOD: D: THE MYS MYSTER TERIOU IOUS S UNFO UN FOLD LDIN ING G OF DI DIST STUR URBA BANC NCE E IN TI TIME ME

PETER FONAG FONAGY Y

University College London ABSTRACT: A model for the development of this mechanism is offered as well as evidence for it from

five areas: (1) the nature of the association of early attachment and later cognitive functioning, (2) accumulating evidence for the association between secure attachment and the facility with which internal states are understood and represented, (3) the limited predictive value of early attachment classification, (4) the stu studie diess of the bio biolog logica icall func functio tions ns of att attach achmen mentt in oth other er mam mammal malian ian sp speci ecies, es, and (5) fac factor tor ana analyt lytic ic studies of adult attachment scales that suggest the independence of attachment type and attachment quality. The author tentatively proposes that attachment in infancy has the primary evolutionary function of generating a mind capable of inferring and attributing causal motivational and epistemic mind states, and through these arriving at a representation of the self in terms of a set of stable and generalized intentional attributes thus ensuring social collaboration, whereas attachment in adulthood serves the evolutionary function of protecting the self representation from the impingements that social encounters inevitably create. Severe personality pathology arises when the psychological mechanism of attachment is distorted or dysfunctional and cannot fulfill its biological function of preserving the intactness of self  representations. A partir de las siguientes cinco a´reas, ofrecemos un modelo para el desarrollo de este mecanismo, ası´ como prue prueba ba del mism mismo: o: (i) la natu naturalez ralezaa de la asoci asociacio acio ´ n entre la tempr temprana ana afectividad afectividad y el funcionamiento cognitivo posterior, (ii) la acumulacio´ n de pruebas para determinar la asociacio´ n entre una afectividad segura y la facilidad con la cual los estados internos son comprendidos y representados, (iii) el limitado valor de prediccio´ n de la clasificacio´ n de la temprana afectividad, (iv) los estudios de las funciones biolo bi olo´gicas de la afectividad afe ctividad en e n otras especies especi es mamı´feras, ´feras, y (v) los estudios analı´tico ´ticoss de fac factor tores es de las escalas de la afectividad adulta que sugieren la independencia del tipo y calidad de la afectividad. El autor autor propone propone tenta tentativam tivamente ente que la afect afectivida ividad d en la infan infancia cia tiene como func funcio io´n prim primaria aria de evol evolucio ucio´n ´n la de generar una mente capaz de deducir y atribuir estados mentales causales, de motivos y de conociRESUMEN:

Paper submitted to the Infant Mental Health Journal February 2001, based on a plenary address given at the World Association of Infant Mental Health Congress, Montreal, July 28, 2000. The author would like to acknowledge the contribution of his colleagues at the Child and Family Center, Menninger Clinic, Topeka, Kansas. This article is in large part based on data gathered by them. In particular, the author is indebted to Drs. Helen Stein, Jon Allen, Martin Maldonado, and Jim Fultz. The author would also like to acknowledge Dr. Anna Higgitt’s guidance in the writing of  the paper. This article is part of an ongoing collaboration with Drs. Mary Target, George Gergely, and Elliot Jurist. A coauthored book covering these themes has recently been published by Other Press of New York (Fonagy, Gergely, Jurist, & Target, 2002). The final preparation of this manuscript was superbly aided by Dr. Elizabeth Allison. Direct correspondence to: Peter Fonagy, Psychoanalysis Unit, Sub-Department of Clinical Health Psychology, University College London, 1-19 Torrington Place, London WC1E 7HB, UK; e-mail: [email protected].

 From Infancy to Adulthood 

●

213

mientos, y por medio de esto llegar a una representacio mientos, representacio ´ n del prop propio io ser en te´rminos ´rminos de un grup grupo o de atributos intencionales, estables y generalizados, para asegurar ası´ la colaboraci co laboracio o´n social, socia l, en tanto que la afectividad afectiv idad en la e´poca ´p oca adulta sirve a la l a funcio f uncio´n de evolucio´ n de proteger la representa r epresentacio cio´n propia propi a de las impresiones que los encuentros sociales crean inevitablemente. Una grave patologı´a ´a de la personalidad aparece cuando el mecanismo sicolo´ gico de la afectividad esta´ distorsionado o funciona mal y no puede llevar a cabo sus funciones biolo´gicas de preservar intactas las representaciones de uno mismo. RE´SUME´ : Un mode`le `le pour le de´veloppement ´veloppement de ce me´canisme ´canisme est offert ainsi que la preuve qu’il fonc-

tionne, dans cinq domaines: (i) la nature de l’association de l’attachement pre´coce et du fonctionnement cognitiff subse´q cogniti ´quent, uent, (ii) ( ii) preuve pr euve de l’associa l’association tion entre e ntre l’attache l ’attachement ment se´ cure et la facilite faci lite´ avec a vec laquelle laquel le les e´tats ´tats internes sont compris et repre´sente ´sente´s, ´s, (iii) la valeur de pre´diction ´diction limite´e ´e de la classification de l’attachement, (iv) les e´tudes ´tudes des fonctions biologiques de l’attachement chez d’autres espe`ces mammife`res `re s et (v) des e´tudes ´t udes de facteur analyti analytique que d’echell d’echelles es d’attachem d’ attachement ent adulte qui sugge sugg e`rent `re nt l’inde l’i nde´pendance ´pe ndance entre le type d’att d’attachem achement ent et la quali qualite te´ d’att d’attachem achement. ent. L’au L’auteur teur prop propose ose prov provisoir isoirement ement que l’att l’attachem achement ent durant la petite enfance a la l a fonction f onction e´volutionnai ´vo lutionnaire re principa principale le de ge´ne´r ´rer er un esprit capable d’infe´re ´rerr et d’attribuer d’attri buer des e´tats ´t ats de motivati motivation on et e t des e´tats ´t ats e´piste ´pi ste´miques ´m iques causals, et au travers tr avers sont capables ca pables d’arriver  d’ arriver  a` une repre´sentation ´se ntation du self en termes d’un set d’attri d’attributs buts intentio intentionnels nnels stables et ge´ne ´n e´ralise´s assurant par  la` meˆme ˆme une collabora collaboration tion sociale, alors que l’attac l’attachement hement a` l’aˆge ˆge adulte sert de fonction e´volutionnai ´vo lutionnaire re pour prote´ger ´ger la repre´sentation ´sentation du moi des empie`tements `tements que les rencontres sociales cre´ent ´ent ine´vitablement. ´vitablement. Une pathologi pathologiee se´ve ´ve`r `ree de la personnali personnalite te´ de´coule ´co ule lorsque le me´canisme ´c anisme d’attache d’attachement ment psychosoci psychosocial al est de´forme ´forme´ et dysfonctionn dysfonctionnel el et ne peut pas remplir sa fonction biologique de pre´servation ´servation de la nature intacte de la repre´sentation ´sentation du moi. ZUSAMMENFASSUNG: Ein Modell fu¨ r die Entwicklung dieses Mechanismus wird angeboten und Beweise

aus fu¨ nf Bereichen: (i) Das Wesen der Verbindung der fru¨ hen Bindung und der spa¨teren, ¨teren, kognitiven Leistung, (ii) zunehmende Beweise fu¨ r die Verbindung zwischen sicherer Bindung und der Einfachheit mit der innere Zusta Z usta¨nde ¨n de verstanden verstan den und abgebildet abg ebildet werden, (iii) (i ii) die Begrenztheit Begre nztheit der Voraussagefa Vora ussagefa¨h ¨higkeit igkeit der Klassifikationen der fru¨ hen Bindung, (iv) die Studien der biologischen Natur der Bindung bei anderen Sa¨ugern ¨u gern und (v) ( v) faktorenanalyti faktor enanalytische sche Studien Studi en zur Bindung Bind ung bei Erwachsenen, Er wachsenen, die eine Unabha Unab ha¨ngigkeit ¨ng igkeit des Bindungst Bindu ngstyps yps von der – qual qualita ita¨t ¨t wahr wahrsche scheinlich inlich machen. Der Auto Autorr neigt dazu vorz vorzusch uschlagen lagen,, dass Bindung im Kleinkindesalter die prima¨re ¨re evolutiona¨re ¨re Funktion hat einen Geist zu entwickeln, der kausale motivationale und epistemiologische Geisteszusta¨nde ¨nde herstellen und zuordnen kann und durch diese zu einer Repra¨sentation ¨sentation des Selbsts kommt – im Besi Besitz tz einer Menge von stab stabilen ilen und generalisierba generalisierbaren ren intentionalen Attributen –, wodurch soziales Zusammenleben gesichert wird, wohingegen Bindung im Erwachsenenalter der evolutiona¨ren ¨ren Funktion dient die Repra¨sentation ¨sentation des Selbsts vor den Nadelstichen zu schu¨ tzen, die durch soziale Begegnungen unausweichlich erzeugt werden. Schwere Perso¨ nlichkeitspathologie tritt dann auf, wenn die psychologisch psychologischen en Mechanismen der Bindung gesto¨ rt, oder dysfunktional sind und ihre biologische Funktion, na¨mlich ¨mlich den Erhalt der Intaktheit der Selbstrepra¨sentation, ¨sentation, nicht mehr erfu¨ llen kann.

214

●

 P. Fonagy

*

*

*

PREAMBLE

Where Has All the Mystery Gone?

Recently, very unusually for a clinician-academic like myself, I saw three new patients for  assessment on the same day. The three new male cases were very different (a depressed journalist with sexual problems, a young man soon to be married but worried about his history of  bipolar illness, and an adolescent with violent behavior problems), but they shared a single common theme. My friend, the late George Moran, taught me to try during initial assessments to elicit the patient’s theory of their problems. Whether this is a good idea or not, I always ask something like: “Why do you think this has happened to you?” or “Why do you think people like you get depressed?” On this particular day I was surprised because all three men came came up with identical answers: “I think it is well established to be a chemical imbalance caused by my genes,” said the groom with the bipolar disorder; “I think from my mother I inherited a tendency to look for the negative” answered the journalist, and “I’ve been told I have bad genes that make me hit people” repli replied ed the adolescent. adolescent. In each case, as they answered, I had the impression of time collapsing. There was no space between the moment their father’s sperm penetrated their mother’s ovum and the present moment. Of course, in each case I was able to call upon the natural human desire to create a meaningful life narrative and to explore how their experiences had assisted or hindered their  capaci cap acity ty to cope wit with h the dif difficu ficulti lties es the they y brought. brought. But But while while the they y exp expoun ounded ded the their ir naı¨ve ¨ve nativ nativist ist views, there was no room for dialogue. There was just one simple message: don’t ask what causes my problems, don’t probe my memories or thoughts or feelings; there is nothing to know, the answer lies in my genes. There was no room for human mystery!

THE DEMI DEMISE SE OF SOCI SOCIALIZA ALIZATION TION:: PAREN PARENTING TING VERSUS GENETICS

There are three primary agents of socialization of children in western society: families, peer  group, and day-care centers or schools (Maccoby, 2000). The emphasis, both professional and cultural, has been on the family as an agent of socialization. For the best part of the past centur cen tury, y, bot both h psy psycho cholog logica icall the theori ories es and com common mon-se -sense nse psy psycho cholog logica icall vie views ws wer weree in agr agreem eement ent in identifying experience with parents as pivotal in shaping an individual’s values, beliefs, character and, naturally, dysfunctions in adaptation. It is interesting to note that, of the two psychological approaches which dominated the last century, learning theory and psychoanalysis, it was the latter that retained some emphasis on the constitutional delimiters to socialization (e.g., Freud, 1920). The last quarter of the 20th century saw a dramatic realignment of developmental theories. The emergence of a cognitive mental science prompted the translation of some learning and many psychodynamic principles into the language of information processing, with presumed mentall opera menta operations tions on past experience creating creating predi predictabl ctablee biases and disto distortion rtionss in mental representations (e.g., Bandura, 1977). Cognitive behavioral approaches to development and

 From Infancy to Adulthood 

●

215

The views of socialization that emerged from cognitive social learning theory have underscored the important role of the child in determining their their socialization experience. Clearly, mothering an infant high in emotionality must elicit quite a different set of maternal behaviors than the mothering of a sociable, unemotional infant. This realization was critical in radically moderating the parent-blaming tendency of early early psychopathologists. These transactional models of child-to-parent effects, however, while both highly desirable and totally appropriate in the way that they were originally proposed and stated, unfortunately became a major plank in the argument of those proposing a nativist revival—but more of this later. The ontogenetic frame of reference of cognitive social learning theory for the most part maintained the environmentalist tradition of psychoanalytic and learning theories. Developmental psychopathology, permeated with the dialectic of social learning theory, came to dominate child psychiatric epidemiology, under the leadership of Norman Garmezy with other giants such as Michael Rutter, Alan Sroufe, Robert Emde, Dante Cicchetti, and others.. The key resea others research rch question came to be the myste mysterious rious unfolding unfolding of distu disturbance rbance through through time, the integration and interaction of person and environmental characteristics in the generation of psychological disturbance through ontogenetic development. Notwithstanding its explicit pli cit commi commitme tment nt to a dia dialec lectic tic tra transa nsacti ctiona onall mod model el (e. (e.g., g., Gar Garmez mezy, y, Mas Masten ten,, & Tel Telleg legen, en, 198 1984), 4), developmental psychopathology always retained its emphasis on socialization, particularly intrafamilial socialization (e.g., Cicchetti, Cicchetti, 1987; Sameroff, 1995). Attachment theory became one of the guiding frameworks of the approach, and John Bowlby was, to some degree posthumously, recognized by many as one of its pioneers (Sroufe, 1986). Thus, notwithstanding the domina dom inance nce of cog cognit nitive ive psy psycho cholog logy y and soc social ial lea learni rning ng the theory ory,, dev develo elopme pmenta ntall psy psycho chopat pathol hology ogy remained a broad church, and many psychodynamic concerns were retained. The principal concerns of developmental psychopathologists in the last quarter of the 20th century were mostly around risk factors, with risk factors associated with the family occupying a most important role. Developmental psychopathology of the early years of development was particularly concerned with social and cultural facets of risk, parent–infant relationships, epistemic and motivational mental states that influence parenting (e.g., Belsky, 1984), the interaction of economic and social disadvantage with parenting (e.g., McLoyd, 1990), the distorting influences of past experience on emotional and cognitive structures of the child (e.g., Fox, Platz, & Bentley, 1995), and parental behaviors as mediators of the gross social inequalities that became an increasing source of concern for social scientists of the Thatcher and Reagan years (e.g., Pettit, Bates, & Dodge, 1997). Permeating all these ideas was the notion that the unfolding of psychopathology occurred in the context of the child’s primary socialization environment: the family. The family, in particular the parents, provided the backdrop against which this unfolding occurred: their characteristics were crucial to the developmental choices the child would make, their actions and collaboration critical to both treatment and prevention. But alas, all this was too good to continue. The Findings from Behavi Behavior or Genetics

Over the last decade of the 20th century, perhaps partly triggered by the excitement of the human genome project, but also by research designs of increasing statistical sophistication, quantitative behavior genetics was unleashed on early development research. For some time, it seemed as if research in genetics had all but eliminated the place for classical socialization theories with an emphasis on parenting, such as attachment theory, and had refuted all theories

216

●

 P. Fonagy

(p. 7). He went on to say that he doubted whether any undesirable trait displayed by a child could be significantly modified by anything a parent does. The biological biological (genet (genetic) ic) movement of the 1990s highlighted highlighted a numbe numberr of issue issuess of parti particcular relevance to early developmentalists: 1.

The overall connection between early parenting and socialization outcomes turns out to be quite weak, and in longitudinal studies, parenting accounts for negligible proportions of the variance. There is very limited evidence that might link early relationship experiences to the development of psychopathology. Most observed associations between socialization socialization and disor disorder der have been reint reinterpre erpreted ted in terms of rever reverse se causality: causality: the child’s disorder causes family dysfunction rather than the other way around. For  example, critical parental attitudes are more commonly observed in children who have suffered from a psychological disorder for longer (Hooley & Richters, 1995), suggesting that the parents’ exposure to psychopathology increases the likelihood of parental criticism, rather than the other way around.

2.

Correlations between characteristics of early parenting and later child behavior, even in prospective studies, can be reinterpreted in a model in which the child’s genetic characteristics are seen as determining the parent’s response, rather than assuming that parenting influences the child. For example, the observed associations between parenting sensitivity and attachment classification may be driven by the behavior of the child and accounted for by the child’s genetically determined predispositions (the so-called child to paren parentt effec effects). ts). It is also inter interestin esting g that aspects aspects of a fami family’s ly’s experience experience of its own interactions are genetically determined. Thus, according to the findings of the Colorado Adoption Project, the parents’ report of warmth and negativity in the family and the child’s report of achievement orientation appear to be genetically determined, suggesting that aspects of the family environment are susceptible to the influence of  the child’s child’s genetic genetically ally rooted rooted charact characterist eristics ics (Deat (Deater-De er-Deckard, ckard, Fulker Fulker,, & Plomi Plomin, n, 1999). Parental Parent al warm warmth th is influen influenced ced by the parents’ genetic endowment, endowment, thus the association association between warmth and lack of pathology may well be spurious (Losoya, Callor, Rowe, & Goldsmith, 1997).

3.

The relative contributions of genes and environment are estimated by examining the observed correlation between two siblings with the correlation that would be expected on the basis of the degree of genetic material the two have in common. Thus, MZ twins who share all genetic material should resemble each other (correlate on a trait) about twice as much as DZ twins. Behavior genetic models of twin and adoption studies partition variability into genetic and environmental components by subtracting the proportion of variability on a specific trait accounted for by shared genes (h ( h 2) from 100 ( E 100 h 2). In most domains h 2 is 50–60%, with less than half left to E. Two large-scal large -scale, e, highhigh-qualit quality, y, commu communitynity-based based studi studies es — the Virg Virginia inia Twin Study (Eaves et al., 1997) and the Non-shared Environment and Adolescent Development (NEAD) project (Reiss et al., 1995a)—have confirmed that most types of childhood psychopathology have quite substantial genetic components. For example, heritability estimates for ADHD range from 54 to 82%. More or less the only psychological disorder  of childhood with a negligible genetic component is separation anxiety. Even for this disorder, heritability estimates for girls are substantial (31–74%), the low figures come ϭ

Ϫ

 From Infancy to Adulthood 

●

217

environmentally mediated family influences, such as children who are read to learning to read sooner than those who are not read to, are, in fact, mostly mediated by the shared share d genetic predispositi predisposition on of careg caregiver iver and offspring, offspring, and are there therefore fore in thems themselves elves unimportant (Harris, 1998; Rowe, 1994). Other studies have shown a number of peer  influences to be actually genetically determined (e.g., Jacobson & Rowe, 1999). Most recently, an analysis of the Colorado adoption project showed that many of the milder  adversee effec advers effects ts on social adjustment adjustment associated associated with parental parental divorce are in fact geneti genetic: c: the “divorce gene” causes adjustment problems in children even if they are adopted into nondivorcing families (O’Connor, Caspi, DeFries, & Plomin, 2000). 5.

In so far as behavior genetic studies showed family environment to matter, it was environment specific to each child within the same family (nonshared environment) that mattered. Environment may be partitioned into a shared and a nonshared component. If the trait under scrutiny has a shared environmental component, then both MZ and DZ twins would be significantly correlated on the trait, while if nonshared environmental factors are involved, then the siblings should not be correlated. Shared environmental influences may be estimated in adoption studies by comparing the correlation of adopted children and their adopted siblings with children in other households. If shared aspects of the environment such as parenting were indeed formative, then adopted siblings living in the same home should be significantly more alike than unrelated children across households. After the genetic and shared environmental components are estimated, what remains is the nonshared environment ( E ( Eus 100 h 2  E ). The nonshared environment appears to be the bulk of the environmental componentt — sha nen shared red envi environ ronmen ment, t, an ins instan tance ce of whi which ch wou would ld be par parent ental al sen sensit sitivi ivity, ty, acc accoun ounts ts for almost no variance (Plomin, 1994). Adopted children, it seems, are no more like their the ir ado adopte pted d sib siblin lings gs tha than n unr unrela elated ted chi childr ldren en gro growin wing g up in a dif differ ferent ent hou househ sehold old (Plomin (Plom in & Berge Bergeman, man, 1991). Such finding findingss are parti particular cularly ly striking in the case of twin studies where there are in-built controls for age, gender, temperament, and birth order. This is important because the relatively weak observed effects of the shared environment have been used to suggest that environments generally assumed to be toxic by developmental psychopathology (such as high level of parental conflict, divorce, inconsistent discipline, parental psychiatric disturbance, multiple moves, death of the parent or even relative social disadvantage and neighborhood effects) are either of less importance impor tance than previ previously ously thought, or, more likely, are actual actually ly genet geneticall ically y mediated (Plomin, Chipuer, & Neiderhiser, 1994). Plomin (1994) put this quite elegantly: “So often we have assumed that the key influences on children’s development are shared: their parents’ personality personality and child childhood hood exper experiences iences,, the quali quality ty of their paren parents’ ts’ marriage relationship, childrens’ educational background, the neighbourhood in which they grow up, and their parents’ attitude to school or to discipline. Yet to the extent that these influences are shared, they cannot account for the differences we observe in children’s child ren’s outcomes” outcomes” (p. 23). More or less the only disorders disorders with a substa substantial ntial shared shared environmental component are oppositional defiant disorder and conduct disorder (Goldsmith, Buss, & Lemery, 1997). ϭ

Ϫ

Ϫ

s

6.

It has been argued that even nonshared environmental effects may be better understood as genetic in origin. Genetically influenced aspects of children’s behavior may be responsible for provoking specific observed responses in parents and other people. This

218

●

 P. Fonagy

rental behavior rather than to his/her genes. As much as 20% of the variability in how parents treat adolescents may be due to the genetic characteristics of the adolescent (O’Connor, Hetherington, Reiss, & Plomin, 1995). There is evidence, for example, from studies of adopted children that authoritarian parenting may be elicited by the child’s resistive or distractible behavior (Ge, Conger, Cadoret, Neiderhiser, & Yates, 1996). The celebrated results of the NEAD study, showing that adolescents who are observed to receive preferentially negative treatment from the parents relative to their  twin siblings are at greater risk of developing depression and antisocial symptoms, while the more positively treated siblings are actually protected from those disorders (Reiss et al., 1995b), may also be an instance of a pseudoenvironmental effect. The correlation between parental conflict-negativity and adolescent disorder could also be accounted for by genetic factors in the adolescent (Neiderhiser, Reiss, Hetherington, & Plomin, 1999). These are just the highlights. It seems that over the past 10 years we, both developmental professionals and the lay public, have unconsciously switched from a primarily psychosocial model of child and adult development to a genetic-biological frame of reference that often a priorii exclu prior excludes des consideration consideration of child child– – paren parentt relat relationshi ionships. ps. In an informal study, we asked 20 consecutive parents referred to an outpatient child community mental health clinic about the likely cause of their child’s problems. It surprised no one that they all put brain chemistry at the top of the list. It was more surprising that “bad genes” came second, peers third, and early life experiences a poor fifth, just ahead of food additives. Why is this happening? The excitement of novelty, of scientific discovery, must have something to do with it. But there is more to it than that. The reduction of the mind to chemicals was appealing even to Freud. Although our consciousness, our free will, our mind is undoubtedly our most treasured possession, it is also the source of all our sadness, conflict, pain, suffering, and misery. The reduction of models of pathology to a principally genetic mode of causation is undoubtedly a relatively comfortable solution for all of us. But like all comforts, it comes at a price. I now intend to argue that the case for reducing the emphasis on parenting, particularly the emphasis on the early attachment relationships, is based on false evaluations of behavior  genetics data. I will also suggest that perhaps in the past our emphasis on the role of parenting was somewhat naı¨ve ¨ve in trying to see the parents’ influence simply in terms of relationship quality, or in terms of mechanisms of internalization, introjection, identification and so forth. I will try to show that (1) early attachment experiences may well be key moderators of the expression of individual genotype, and (2) that the primary evolutionary function of attachment may be its contribution to the ontogenetic creation of a mental mechanism that could serve to moderate psychosocial experiences relevant for gene expression. SCRUTI SCR UTINIZ NIZING ING TH THE E CA CASE SE FOR GEN GENETI ETICS CS

As we have seen, the case for genetic determinants rests on two pillars: the first is the weakness of the socialization evidence, and the second is the findings of quantitative behavior genetics. The Current Case for Socialization Socialization

Classical reviews extensively cited by behavior geneticists tended to reveal weak correlations between betwe en parenting parenting and socia socializat lization ion outcomes outcomes open to alternative, alternative, genet genetic ic inter interpreta pretations tions.. How-

 From Infancy to Adulthood 

●

219

Family Center carried out operationalized clinical ratings of almost 150 infants. Over 70% of  the children have so far been followed up, two to four years later. Infant behavior problems strongly predicted later behavioral difficulties in a number of areas. Parenting observed in infancy predicted preschool emotional difficulties. In particular, parents who were both neglectful glect ful and hostil hostilee had child children ren with more behavioral problems. problems. This remained remained true even when behavioral behavi oral problems in infan infancy cy were controlled controlled for. Observed negle neglect ct and hosti hostility lity correlated correlated .36 with the child’s behavioral problems noted at four years, even when infant behavior was already controlled for. This implies that parenting had a predictable effect beyond that which could be explained explained in term termss of the parent responding responding specifical specifically ly to the precursors precursors of the child child’s ’s behavioral problems. There is increasing support for the hypothesis that many of the effects of socioeconomic disadvantage, social adversities, family transitions, and other risk factors for the development of psychological disturbance are mediated by distortions in intrafamiliar social relation relationships, particularly in the relationships between the child and the parent. For example, a large-scale study of 9,000 children in the UK, living in various family configurations (step-, single-pare singl e-parent, nt, and nonste nonstep p famil families) ies) found that the quali quality ty (degr (degree ee of negativity) negativity) of child children’s ren’s relationship with their caregiver was related to adjustment outcome, regardless of the degree of the biolo biological gical relationship relationship betwe between en child and careg caregiver iver (Dunn, Deate Deater-De r-Deckard, ckard, Pickering, Pickering, O’Connor, & Golding, 1998). An accumulation of findings links the course of childhood disorder to characteristics of parent–child relations, specifically expressed emotion (EE). The extent of family criticism, hostility, and emotional overinvolvement predicted recovery rates one year after hospitalization in a sample of children hospitalized for depression irrespective of treatment method, ilness duration or comorbidity (Asarnow, (Asarnow, Tompson, Hamilton, Hamilton, Goldstein, & Guthrie, 1994). Studies of children with physical disorders also imply that the course of  these disorders is likely to be affected by parenting characteristics. For example, Mrazek and colleagues (Mrazek et al., 1999) studied the association between early parental behavior and the later onset of asthma in 150 children at genetic risk of developing the disorder. Poor  maternal coping and parenting problems were associated with the child child developing the problem before the age of three years. Parenting characteristics appeared to accelerate the onset of the disease but did not increase the overall risk for the disorder. It appears that the over twofold increase of the risk of psychopathology in children with chronic physical illness is mediated by the quality of family functioning under the stress of the child’s disorder (Cohen, 1999). Not surprisingly, then, family-based interventions tend to be effective effective in reducing the psychological complications of chronic physical illness in children (Walker, Johnson, Manion, & Cloutier, 1996). Family conflict is associated with poor adherence in both asthma and diabetes (Dumont et al., 1995; Weinstein & Faust, 1997) and decreasing family conflict appears to consistently improve impro ve illne illness ss outcome (Weinstein, (Weinstein, Faust, McKee McKee,, & Padma Padman, n, 1992). Better-controlled studies are accumulating, and current reviews of the socialization literature tend to yield more encouraging conclusions (Maccoby, 2000). In the meantime, the question has shifted from attempts to establish the nature or extent of parenting influence upon child development to questions concerning the range of factors that influence parent–child relationshi relat ionships. ps. Chara Characteri cteristics stics that have been relia reliably bly demons demonstrat trated ed to be signi significant ficant in this context include: (1) parents’ childhood relationships or experiences with their own parents (Simons, (Simo ns, Beama Beaman, n, Conger Conger,, & Chao, 1993) with advers adversee early experiences experiences increasing increasing the likelihood of family conflict and divorce (O’Connor, Thorpe, Dunn, & Golding, 1999); (2) economic and social disadvantage disadvantage that is linked to poorer parenting parenting practices practices (e.g. (e.g.,, Hughe Hughes, s, Deate Deaterr-

220

●

 P. Fonagy

parents have spent in their current family setting is likely to be related to the quality of their  relationships with their children (Hetherington, Bridges, & Insabella, 1998); (5) biological relatedness (the experience of “ownness” in relation to a child) enhances emotional relatedness between parent and child (Dunn, Davies, O’Connor, & Sturgess, 2000); (6) parent–child relationships are strongly influenced by the relationship history of the other parent (Dunn et al., 2000) as well as by aspects of the sibling relationships (Dunn et al., 1998), which suggests that a systemic perspective is essential to arriving at an adequate account. These are simply some illustrative recent findings, but they stress the complex character of the interaction between the social environment provided by the family and aspects of parent–child relationships with a potential to influence childhood outcome. The recognition recognition that the parenting parenting environment environment varies in significance significance acros acrosss devel development opmental al stages and even within each developmental phase across across time introduces a considerable degree of furth further er compl complexity exity into the interpretation interpretation of socia sociall envir environment onment to behavi behavioral oral outcome outcome correlations. Therapeutic experience teaches us of the exceptional influence of certain “key moments”” of inter ments interactio action n betwe between en thera therapist pist and patie patient, nt, when experi experience ence brought into clear relief  relief  by a confluence of circumstance and intrapsychic factors factors suddenly enables therapeutic change. This idea is full fully y elabo elaborated rated in the work of Danie Daniell Stern Stern,, Ed Tronick, Karlen Karlen Lyons-Ruth, Lyons-Ruth, and the Boston group on therapeutic process (Stern et al., 1998). They argue that change in therapy might be a function of special moments of attunement between therapist and patient. More generally, the same could apply to special moments of influence of parents on children—  naturally both with a positive and negative valence—moments of parenting influence, which, however, key and formative, might be the needles in a socialization haystack. It is hard to imagine how observational research, focusing on aggregate or time-sampled behaviors, can hope to captur capturee signi significant ficant numbers of such key momen moments. ts. The corre correlatio lations ns between observed observed parenting and child outcome may never reflect the true influence of parenting. In any case, however strong the associations between parenting and socialization, the possibility that the more parsimonious explanation is genetic cannot be ruled out. It is the apparent strength of genetic findings, the massive proportion of variability accounted for, no matter how specific the trait, which casts such a dark shadow over developmental psychopathology. The room they leave for socialization, for the mysteries and uncertainty of development, is limited. But are the genetic findings as unequivocal as they seem? The evidence from behavior genetics should be interpreted with caution. The reasons are: (1) methodological, (2) conceptual, and (3) empirical. METHODO METH ODOLOGI LOGICAL CAL PROB PROBLEMS LEMS WITH THE EVID EVIDENCE ENCE FROM BEHA BEHAVIOR VIOR GENE GENETICS TICS

Methodologically, the contrast of identical and fraternal twins confounds genetic similarity and environmental influence. It has been claimed that identical twins have more similar environments than fraternal ones (e.g., they have more friends in common, they are treated more similarly by the parents) (Reiss, Neiderhiser, Hetherington, & Plomin, 2000). The status of  inferring environmental effects by a process of subtraction using the additive model has also been questioned (Turkheimer, 1998). In particular, E is often estimated by subtraction, with no attempt at a direct measure of environmental factors. If the estimate for heritability is high, simple additive models,  E must be low. In reality, G and E combine to generate a phenotype. In simple however, this interaction would be pooled with genetic effects. Further, measurement error is

 From Infancy to Adulthood 

●

221

The source of data about pathology significantly influences heritability estimates. The agreement between raters and methods in the Virginia Twin study was only moderate, with parent child agreement rarely above .3 for most clinical conditions (Hewitt et al., 1997). The use of parents as a source of data about themselves as well as their child creates an inbuilt geneticc bias. Heritabilit geneti Heritability y estim estimates ates are inflat inflated ed by the use of parents’ reports reports of child behavior  behavior  rather than behavioral observation or self reports. It is not surprising that when parents rate a child’ss aggression, child’ aggression, the corre correlatio lation n with the paren parent’s t’s aggression aggression should increase increase (Mil (Miles es & Carey Carey,, 1997). It is more challenging to explain why the heritability of internalizing disorders would be 60–72% when the informants were the parents and 11% when the informant was the child (Eaves et al., 1997). CONCEPTU CONC EPTUAL AL PRO PROBLEM BLEMS S WITH THE EVIDE EVIDENCE NCE FROM BEHAVIOR BEHA VIOR GENE GENETICS TICS

At a conceptual level, we may certainly question question the notion of nonshared environment, because it merely refers to intersibling differences, not to their environment. In fact, shared environments could as easily serve to make children in the family different from one another as to increase intrafamilial similarity, because shared environments may be experienced very differently by two children. A further conceptual problem concerns heritability estimates based solely solel y on indivi individual dual diff difference erences. s. Such estimates remove share shared d enviro environment nmental al effects such as secularr trends secula trends,, and are strictly restricted restricted to the environment environment studied. Height, IQ, and the prevalence of a number of psychological disorders have increased markedly over the last century, undoubtedly as a consequence of environmental changes, yet current behavioral genetic methods of estimating environmental effects preclude consideration of these. EMPIR EM PIRICA ICAL L PRO PROBLE BLEMS MS WI WITH TH TH THE E EVI EVIDE DENCE NCE FRO FROM M BEHAVIOR BEHA VIOR GENE GENETICS TICS

Empirically, we could point to studies where environmental determinants revealed substantial effects after  after geneti geneticc influe influences nces had been excluded (Johnson, Cohen, Brown, Smailes, Smailes, & Bernstein, 1999). We could raise questions about the actual (rather than assumed) differential responsivenes sponsi venesss of caregi caregivers vers to sibli siblings. ngs. Evidence on just how diffe differentl rently y sibli siblings ngs are treat treated ed is actually quite mixed. In one of the only behavioral genetic studies to actually look at the child’s environment rather than simply infer it, Reiss, Plomin, Hetherington, and colleagues found direct evidence for the notion of the nonshared environment (Reiss et al., 2000). The difference between the degree of coerciveness of parenting between two twins was more predictive than the absolute level of negativity (Reiss et al., 1995b). However, Judy Dunn’s naturalistic observational studies of siblings actually suggest that whilst cross-sectionally parents may appear  to be treating siblings differently, looked at longitudinally children at various ages receive comparable treatment (Dunn & McGuire, 1994). Regardless of the ultimate conclusion concerning the differential treatment of siblings, the fact that studies of social development tended to look at single children implied that they have on the whole underestimated the impact of  parenting and other shared environmental influences. As we shall see later on, there may be specific pressures in family systems for differential responses of siblings as part of the need for each each pers person on withi within n the sys system tem to have have a uni unique que rol role. e. Int Intere eresti stingl ngly, y, the pre pressu ssure re for dif differ ferenc encee may be greater when genetic differentiation is least. Further, experimental manipulations of 

222

●

 P. Fonagy

size of parent training for children with ODD is around 1 (Serketich & Dumas, 1996). More relevant relev ant in this context, accumu accumulati lating ng evide evidence nce suppor supports ts the usefu usefulness lness of exper experiment imental al interventions with parents such as home visitation visitation (e.g., Olds et al., 1998), with long-term beneficial effects in reducing risk of criminality and delinquency. Of course, the impact of environmental manipulation is often not as large as one would hope; moreover, long term follow-ups in treatment treat ment studie studiess are relati relatively vely rare rare and even even quite quite impre impressive ssive change changess initi initiated ated by exper experiment imental al interventions dissipate (Fonagy, Target, Cottrell, Phillips, & Kurtz, 2000). As clinicians, our main objection to behavior genetic data would not be methodological, conceptual, or empirical but rather pragmatic. Genetic effects may well be indirect as well as direct. Even a high genetic loading for a certain environmental hazard does not mean that the consequence conseq uencess assoc associated iated with that risk facto factorr would necessarily necessarily be geneti genetically cally rather than environmentally mediated. If child abuse, for example, were found to have a large genetic component, its toxic effects would still be via the destruction of trust in the world for the abused child, rather than via a purely genetic process. The implications of behavior genetic data for  clinical intervention are thus quite limited. THE TH E RO ROLE LE OF EX EXPE PERI RIEN ENCE CE IN TH THE E EX EXPR PRES ESSI SION ON OF THE GENO GENOTYPE TYPE

It is universally agreed that developmental psychopathology involves a gene–environment interaction. Empirically, this interaction term has proved to be quite hard to find. Plomin’s systematic review of the literature, now admittedly somewhat dated, only found evidence for  relatively isolated examples (Plomin, DeFries, McLearn, & Rutter, 1997). Some quantitative behavior genetic studies strongly suggest interactive processes in which a genetic vulnerability is triggered by environmental exposure. For example, the the classic Finnish adoptive family study of schizophrenia suggests that children with a schizophrenic biological parent were more likely to develop a range of psychiatric problems if, and only if, they were adopted into dysfunctional families (Tienari, Wynne, Moring, Lahti, & Naarala, 1994). Bohman (1996) reported that criminality only appeared to be associated with a genetic risk if children whose biological parents were criminals were adopted into dysfunctional homes. Genetic risk may or may not become manifest, depending on the quality of the family environment to which a child is exposed. But if this is such a pervasive process, why is the quantitative behavior genetic evidence evide nce so spars sparse? e? The obvious answer is that the environment that triggers the expression of a gene is not objective. The child’s experience of the environment is what counts, and this is a function of  appraisal. The manner in which environment is experienced will act as a filter in the expression of genotype into phenotype. Here we touch on the pivotal role of parenting parenting for genetic research, particular parti cularly ly attachment attachment theory. theory. The primary concern concern of attachment attachment theor theory y is with the inter interactio action n of multiple layers layers of representations representations in generating generating developmental outcomes. Data Data from genetics calll for cal for exact exactly ly such such sop sophis histic ticati ation on in in und unders erstan tandin ding g the way gen genes es may or may not be exp expres ressed sed in particular individuals. This idea was beautifully elaborated by a founding father of WAIMH, Robert Emde (1988), over a decade ago. The pathway between genes and phenotypes is a tortuous one, along which genetics and the envir environment onment constantly constantly inter interact. act. Internal and exter external nal stimuli, steps in the development development of  the brain, hormones, stress, learning, and social interaction, alter the binding of transcription regulators regul ators (Kandel, 1998). For example, although risk facto factors rs operate in combi combination nation,, there is

 From Infancy to Adulthood 

●

223

depend not only on the nature of those factors, but also on the way the child experiences them, which in turn, in many instances will be a function of attachment and other intrapsychic experiences. perie nces. The exper experienti iential al filter that attachment attachment provides may, in turn, be a funct function ion of eithe either  r  genetic or environmental influences, or of their interaction (Kandel, 1998). Thus, intrapsychic representational processes are not just consequences of environmental and genetic effects—  they may be the critical moderators of these effects. This point has substantial clinical significance, because a child’s understanding of his/her  environment is more readily modifiable than the environment itself, or the genes with which the environment interacts (Emde, 1988). In attachment theory, intrapsychic perspective may be helpf helpful ul in consi considering dering,, not just what precipitates precipitates a disor disorder, der, but also which processes processes influinfluence the course of the disorder for better or worse. Up until the last five years this idea was theory—now the collaboration of molecular geneticists and attachment theory is making it a reality. I shall give just one example of this powerful paradigm. An important result has recently been published from the Budapest Infant–Parent Study (Lakatos, Toth, Nemoda, Ney, Sasvari-Szekely, & Gervai, 2000). These workers found an association between the DRD4 receptor III exon polymorphism and disorganized attachment classification in 12-month-old infants. Considerable evidence over the years has linked behavioral problems in both children and adults with the seven-repeat allele of the DRD4 gene. In particular, ADHD has been implicated (Faraone et al., 1999) although not all studies concur  (Castellanos et al., 1998). The review by Swanson and colleagues (Swanson et al., 2000) confirmed the likely role of the seven-repeat allele of this gene in making the postsynaptic receptor recept or subse subsensiti nsitive. ve. Comin Comings gs et al. (1999) report findings relat related ed to impul impulsive, sive, compulsive, compulsive, addictive behaviors that indicate a greater complexity then a sole focus on the seven versus non-seven alleles of the DRD4 gene. In view of recent findings that have linked disorganized attachment in infancy to clinical conditions in middle childhood, it may be particularly important that in this study 71% of the infants classified as disorganized were found to have at least one seven-repeat allele, in contrast with only 29% of the nondisorganized group. Thus, infants classified as disorganized were more than four times more likely to be carrying this allele. This finding is consistent consistent with observations observations that neurological neurological (Pipp-Sieg (Pipp-Siegel, el, Siegel Siegel,, & Dean, 1999) and neonatal behavioral organization (Spangler, Fremmer-Bombik, & Grossman, 1996) may anticipate a disorganized attachment classification. It might at first sight seem at odds, however, with the classical observation that disorganized infant attachment was linked to unresolved loss or trauma in the mother (Main & Hesse, 1990). A recent prospective study led by Pat Hughes, in which I had the privilege of participating, confirmed that mothers with a histor his tory y of perinata perinatall ber bereav eaveme ement nt were far mor moree lik likely ely to hav havee dis disorg organi anized zed inf infant antss tha than n con contro trols. ls. Whereass almost 45% of the mothe Wherea mothers rs who had lost their last baby during pregnancy pregnancy had infan infants ts classified as disorganized at one year, only 20% of the control mothers matched for age, SES, and education did so. Adult attachment interviews collected before the birth of the child picked up the risk for disorganization. Lack of resolution of mourning mediates the association of  still-birth experience and disorganization of the “replacement infant” in the Strange Situation. But only 62% of the mothers with unresolved AAI classification had infants classified as disorganized, although specificity was relatively high (over 80% of disorganized infants had unresolved mothers). It seems that lack of resolution of mourning may be a necessary but not sufficient condition for disorganization. Because only a third of the children in the Budapest study with the seven-repeat allele showed disorganized attachment, checking for the presence

224

●

 P. Fonagy

tioning less efficiently, for which the seven-repeat allele of the D 4 receptor may be a marker. It has been suggested that the mesolimbic dopamine system controls behavior motivated by reward (Robbins & Everitt, 1999), and less sensitive D 4 dopamine receptors (Van Tol, Vu, Guan, Chara, Bunzow, & Civeili, 1992) could further distort the signal value of the mother’s response. respo nse. The revie review w by Swanso Swanson n et al. (2000) suggests that dopamine underactivity underactivity comprocompromises attentional attentional systems that might exagerate exagerate the impact of subtl subtlee anoma anomalies lies of the mother’s behavior in relation to her infant (e.g., momentary dissociation, frightened or frightening behavior, etc.; see Solomon & George, 1999). This is clearly speculative, but is a readily testable hypothesis that is in line with the general interactional model that we propose. To summarize, a posttraumatic state in the mother might possibly interact with another  dopamine receptor, causing dysfunctional attachment organization and subsequent psychological disturbance. All these findings are rather tenuous at the moment, but all three are consistent with the notion of early experience with the caregiver triggering gene expression. Taken together, they suggest quite a fruitful line of investigation, which, given the relative facility with which samples can be collected and analyzed, might become an important adjunct to most of  our work in developmental psychopathology. THE TH E GEN GENESI ESIS S OF AN APP APPRAI RAISAL SAL ME MECHA CHANIS NISM M

We have seen that the importance of family environment may have been underestimated in behavioral genetics research, for methodological, conceptual, as well as empirical reasons. I have also tried to construct a prima facie case for a representational system that forms an active filter between the genotype and the phenotype. In other words, the mental processing of experience is critical for the expression of genetic material, and therein lie substantial interactions between gene and environment. I would like to conclude by suggesting that the genesis of the appraisal mechanism is intrinsically linked to human attachment. Perhaps more than shaping the quality of subsequent relationships (for which evidence is lacking), the early relationship environment serves to equip the individual with a processing system. The creation of this representational system is arguably the most important evolutionary function for attachment to a caregiver. Adopting this perspective helps redress the prevailing bias against the centrality of the family as the major force in socialization, but it also shifts the emphasis from content of experience to psychological structure or mental mechanism, and involves expanding on current curre nt ideas of the evolut evolutionar ionary y funct function ion of attac attachment. hment. John Bowlby, a major Darwin scholar, was impressed by the obvious selection advantages of infant protest at separation, i.e., protection from predation (Bowlby, 1969). Given that phylogenetica logene tically lly and ontoge ontogenetic netically, ally, infancy is a perio period d of extre extreme me risk, natural selection selection would inarguably favor individuals with a capacity for attachment. There has been a revolution in evolutionary theory since Bowlby’s time. We now realize that “survival of the fittest” cannot guarantee the natural selection of a behavior. Only the reproduction of genetic material can achieve this (Hamilton, 1964). This is the theory of inclusive fitness. One does not need to survive and reproduce oneself for one’s genes to be replicated. For example, some organisms will forgo reproduction to ensure the reproductive potential of their  genetically close relatives. The concept of “incl “inclusive usive fitness” places attachment attachment theor theory y at the center stage of evolutionary evolutionary sociobiology as a key behavioral mechanism mediating the establishment of genetic proximity. Attachment Attac hment is the process that ensure ensuress that we know whose survival will advantage advantage the reproduction of our genes. {Of course it may have additional evolutionary functions. It is possible

 From Infancy to Adulthood 

●

225

of helping nonkin if, and only if, they have done something for one, might also be underpinned by the mec mechan hanism ism of att attach achmen ment. t. Att Attach achmen mentt is lik likely ely to min minimi imize ze the adv advers ersee eff effect ectss of  “cheaters,” individuals who do not reciprocate equitably in groups over time and to whom we are unlikely to become attached. This would be a good example of a further interesting facet of evolution, how a mechanism that evolved for one purpose (the protection of the vulnerable infant) may be put to good biological use in the context of the adaptive problems of subsequent developmental phases. But all these potential biological functions would apply as readily to animal models of attachment as the human infant. If the biological function of attachment is to be a pillar in our argument for the importance of parenting, we need to restrict ourselves to uniquely human capacities. The generally recognized components of attachment behaviors that serve to establish and maintain proximity are: (1) signals that draw the caregivers to their children (e.g., smiling); (2) aversive behaviors (such as crying), which perform the same function; and (3) skeletal muscle activity (primarily locomotion) that brings the child to the caregiver. There is a fourth component that provides a better evolutionary rationale for the entire enterprise of human attachment, going beyond the issue of physical protection. According to Bowlby, at about the age of three, behaviors signifying a goal-corrected partnership begin to emerge. The central psychological processes that mediate goal-corrected partnerships are the internal working models. Bowlby’s original concept has been thoughtfully elaborated by some of the greatest minds in the attachment field, and no attempt to duplicate this will be undertaken here. However, it might be helpful to summarize the four representational systems that are implied in these reformulations: (1) expectations of interactive attributes of early caregivers created in the first year of life life and subsequentl subsequently y elaborated; elaborated; (2) (2) event repres representat entations ions by which genera generall and specifi specificc memories of attachment-related experiences are encoded and retrieved; (3) autobiographical memories by which specific events are conceptually connected because of their relation to a continuing personal narrative and developing self-understanding; (4) understanding of the psychological characteristics of other people (inferring and attributing causal motivational mind states such as desires and emotions and epistemic mind states such as intentions and beliefs) and differentiating these from those of the self. Thus, a key developmental attainment of the IWM is the creation of a processing system for the self (and significant others) in terms of a set of stable and generalized intentional attributes, such as desires, emotions, intentions, and beliefs inferred from recurring invariant patterns in the history of previous interactions. The child comes to be able to use this representational system to predict the other’s or the self’s behavior in conjunction with local, more transient intentional states inferred from a given situation. Classically, in attachment theory this phase change from behavior to representation is generally regarded as a modification of the attachment system propelled by cognitive development (Marvin & Britner, 1999). Our contention here is the reverse: rather than seeing the biological role of attachment shifting ontogenetically as a consequence of other, biologically driven, maturational changes, we propose that a major selective advantage conferred by attachment on humans was the opportunity it afforded for the development of social intelligence and mea meanin ning g mak making ing.. The cap capaci acity ty for “in “inter terpre pretat tation ion,” ,” whi which ch Bog Bogdan dan (19 (1997) 97) defi defined ned as “organisms making sense of each other in contexts where this matters biologically” (p. 10), becomes uniquely human when others are engaged “psychologically in sharing experiences, information and affects” (p. 94). The capacity to interpret human behavior—to make sense of  each other—requires the intentional stance: “treating the object whose behavior you want to

226

●

 P. Fonagy

it is also a product of the complex psychological processes engendered by close proximity in infancy to another human being, the attachment figure. It is not the IWM. It does not contain representa repre sentations tions of of experienc experiences, es, and is not a repos repository itory of personal personal encoun encounters ters with the caregi caregiver. ver. Rather, it is a mechanism for processing new experiences. We could  could call call it an internal working model, if the meaning of this term was clarified: Bowlby applied IWM interchangeably to a processor of experience—a computational module—and as a repository of experiences. The IIM overlaps the notion of a “Theory of mind,” the ability to attribute independent mental states to self and others to explain and predict behavior. It has been suggested that this ability arises from a dedicated domain-specific and possibly modular cognitive mechanism (Leslie & Thaiss, 1992). Studies that have identified selective impairments in theory of mind in patients with Asperger’s and autism support this notion (Happe & Frith, 1996). Recent attempts to link theory of mind to specific areas of the brain (Frith & Frith, 1999) have been consistent with viewing this unique human capacity as a hard-wired constitutionally firmly established maturational function, such as language or vision. We have taken an alternative view, linking theory of mind more closely to social than to cognitive development. Theory of  mind is in many ways an unsatisfactory term, and tends to be operationalized in terms of  relatively simple experimental tasks that call for the child’s appreciation of a puppet’s putative perspective (e.g., Wimmer & Perner, 1983). Others refer to the capacity capacity as mentalization mentalization (Fonagy,, Ste agy Steele ele,, Mor Moran, an, Ste Steele ele,, & Hig Higgit gitt, t, 199 1991; 1; Mor Morton ton & Fri Frith, th, 1995), implying implying a pro proces cesss whereby mental representations of mental states are created. In previous work we have referred to mentalization that is concerned with emotionally charged (attachment) relationships as reflective function (Fonagy & Target, 1997). This term, however, conflates the measurement of  this capacity on the basis of attachment related narratives (Fonagy, Steele, Steele, & Target, 1997) with an assume assumed d psycho psychologic logical al capac capacity ity that may be invol involved ved to greate greaterr or lesse lesserr extent in generating narratives variably imbued with mental state contents. In any case, mentalization is just one aspect of the IIM. We conceive of the IIM as that collection of neurocognitive mechanisms that are naturally selected to evolve under the influence of early social interactions (predominantly) with the attachment figure, and that serve to mediate the impact of the quality of early relationships relationships into adult adult personality functioning. These mechanisms include the secondorder representation of affect and through this its regulation, the regulation of attention, particularly ticul arly effort effortful ful control control,, alongside alongside aspect aspectss of mental mentalizati ization, on, both impli implicit cit and expli explicit cit (Fonag (Fonagy y & Target, in press). However, it is likely to have many more components than these, all of  which have the function of taking forward early relationship experiences into adult personality functioning. THE TH E ON ONTOG TOGENE ENESIS SIS OF THE IIM

How is the IIM creat created ed out of the secur securee base? Clearly, Clearly, there must be biological biological preparedness, preparedness, but in our view this is not separa separable ble from the infant’s infant’s experienc experiencee of the careg caregiving iving envir environment onment.. Jon Allen (personal communication) drew attention to the possiblility of integrating the innate and constructivist perspectives in the development of representational redescription suggested by Karmiloff-Smith (1992) in her book, Beyond Modul Modularity arity.. Representational redescription involves making what is in the mind available to the mind via some kind of endogenous feedback mechanism. She proposes that the innate equipment drives attention in the service of  modulee const modul constructi ruction. on. She elabo elaborates rates the princ principle iple in the context of langu language age acquisition acquisition but it applies with equal force to the development development of an interpersonal representational representational system: “There

 From Infancy to Adulthood 

●

227

build up linguistic representations that are domain-specific. Since we process language very rapidly, the system might with time close itself off from other influences—i.e., become relatively modularized” (p. 36). In answering this question we are drawing upon George Gergely’s and John Watson’s model (Gergely & Watson, 1996, 1999), which we have elaborated in a coauthored monograph (Fonagy (Fonag y et al., 2002). The Gerge Gergely-Wa ly-Watson tson model assumes that the human infant has an atten atten-tional bias or filter for detecting contingencies between its own behavior and the responses in the environment environment that are conti contingent ngent with it. This mechanism mechanism works in the service of representation building for the mentalization capacity. Our core idea is that the attachment context provides the setting in which the infant can develop a sensitivity to self-states, through what Gergely has termed “psychofeedback” or  social biofeedback. This is basically the development of a second-order symbolic representational system for motivational and epistemic mind states. Secondary representations of selfstates are established through this social biofeedback loop, even though the meta-representational capacity itself may be an innate potential waiting to be “fed” by (tailored) environmental input. The internalization internalization of the mother’s mirroring mirroring response of the infant’s distress (caregiving behavior) comes to represent an internal state. Empathic emotion comprises social biofeedback insofar as the mother’s expressiveness is yoked to the infant’s emotional state. The infant internalizes the mother’s empathic expression by developing a secondary representation of his emotional state with the mother’s empathic face as the signifier and his own emotional arousal as the signified. The mother’s expression tempers emotion to the extent that it is separate and different from the primary experience, although crucially it is not recognized as the mother’s experience, but as an organizer of a self-state. It is this this “intersubjectivity,” which is the bedrock of the intimate connection between attachment and self-regulation. It should be noted that intersubjectivity in this context may be a misnomer. The infant at this stage is unaware that he is seeing the other’s subjective state. It is likely that the infant does not yet know that others have internal feelings. At this level of human proximity the other’ oth er’ss sub subjec jectiv tivee sta state te is aut automa omatic ticall ally y ref referr erred ed to the sel self. f. In inf infanc ancy y the con contin tingen gentt res respon pondin ding g of the attachment figure is thus far more than the provision of reassurance about a protective presence. It is the principal means by which we acquire understanding of our own internal states, which is an intermediate step in the acquisition of an understanding of others as psychological cholog ical entities— the intentional intentional stance. In the first year, the infant only has primary awareness of being in a particular, internal, emotional state. Such awareness is noncausal or epiphenomenal in that it is not put to any functional use by the system. It is in the process of psychofeedback that these internal experiences are more closely attended to and evolve a functional role (a signal value) and a function in modul modulating ating or inhib inhibiting iting action. Thus, it is atta attachment chment processes processes that ensure the move from primary awareness of internal states to functional awareness. In functional awareness a feeling of anger may be used to simul simulate ate and so to infer the other’s correspondin corresponding g mental state or may be used used to to serve serve a signa signall value value to dire direct ct acti action. on. A fina finall lev level el of awa awaren reness ess is refl reflect ective ive awa awaren reness ess,, where the individual can make a causal mind state become the object of attention before, and without it causing action. Whereas functional awareness is intrinsically coupled with action, reflective reflect ive awareness is separ separate ate from it. It has the capacity to move away from physical reality reality and may be felt to be not for real. Many studies provide evidence consistent with this model. For example, a study carried out by us showed that the rapid soothing of distressed six-month-olds could be predicted on

228

●

 P. Fonagy

sponders were far more likely to manifest multiple affect states (complex affects). We interpreted these results as supporting Gergely and Watson’s notion of the mother’s face being a secondary representation of the infant’s experience—the same and yet not the same. This is functional funct ional awareness awareness with the capac capacity ity to modula modulate te affec affectt state states. s. A further set of studies, performed by Gergely and his colleagues in Budapest as well as our lab in London with an ongoin ongoing g repli replicatio cation n in Topeka, KS, explor explored ed one-year-olds’ one-year-olds’ underunderstanding of conflicting affect. In one study recently presented at ICIS, 12-month attachment classification, particularly secure and disorganized, was found to be predicted by infant behavior at 6.5 month monthss in a modifie modified d still paradigm (Koo´s, Gerge Gergely, ly, Gerva Gervai, i, & To´ th, 2000). The paradigm involves the mother being instructed according to the still face protocol, but facing a mirror where the infant has a choice between looking at the mother’s face or looking at a perfectly contingent image (themselves). Infants classified as securely attached six months later  engaged in significant amounts of active testing of their mirror self-image only when their  mother became temporarily inaccessible (the still-face period). By contrast, babies who went on to manifest disorganized attachment six months later were drawn to the image of their fully contingent conti ngent self-mov self-movement ementss all through through the labor laboratory atory testing. testing. Interes Interestingl tingly, y, the Koo´ s et al. study also demonstrated that following the still-face period the infants who engaged in more contingency testing looking at their self-image showed more positive affect following the procedure. This led to more successful affect regulation in disorganized infants than in secure ones. Yet seeking for perfect contingency in attempts to detect internal states in the context of human interaction will be of limited effectiveness in the long run. It characterizes the dissociative style of attention organization that is typical of disorganized attachment. EVIDENCE EVID ENCE FOR THE INTE INTERPER RPERSONA SONAL L INTE INTERPRET RPRETIVE IVE MECHANISM

Is there any evidence for an IIM that evolves out of the attachment relationship, with its efficiency conditioned by attachment security? First, there is unequivocal evidence from two decades of longitudinal research that secure attachment in infancy is strongly associated with the precocious development of a range of capacities that depend on interpretive or symbolic skills, such as exploration and play, intelligence and language ability, ego resilience and ego control, frustration tolerance, curiosity, self-recognition, social cognitive capacities, and so on. Attachment security foreshadows cognitive competence, exploratory skill, emotion regulation, communication style and other outcomes. In our view, this is not because of the general impact of attachment security on the child’s self-confidence, initiative or ego functioning, or other  broader personality processes, but rather because attachment processes provide the key evolutionarily prepared paths for an interpersonal interpretive capacity to develop. Thus, it is not the first attachments that are formative, it is not attachment security per se that predicts good outcome on this dazzling array of measures; rather, the features of the interpersonal environment that generate attachment security during the first year of life also prepare the ground for the rapid and competent ontogenetic evolution of interpersonal interpretation. One problem with attempting to trace some of the long-term outcomes of secure attachment in infancy has been the appropriately conservative strategy of controlling for numerous aspects of this interpretive capacity. Controlling for verbal fluency or even IQ removes a part of the variability in which the attachment relationship arguably plays a causal role. But this is not an issue which this article has scope to explore.

 From Infancy to Adulthood 

●

229

(Cassidy, Kirsh, Scolton, & Parke, 1996) found that securely attached kindergarteners were less likely to infer hostile intent in stories with ambiguous content and this bias appeared to mediate their superiority in sociometric status. In the London Parent–Child Project, Miriam and Howard Steele and Juliet Holder and I reported precocious performance on theory of mind tasks amongst five-year-olds with a history of secure attachment in infancy. Since then, other  investigato invest igators rs have also reported this finding (Meins, Fernyhough, Fernyhough, Russe Russel, l, & Clar Clark-Car k-Carter, ter, 1998). Third,, in a relat Third relatively ively full exploration exploration of findin findings gs linki linking ng early attachment attachment and later develdevelopment, Ross Thompson (1999) concludes that “the strength of the relationship between infant securi sec urity ty and lat later er soc socioio-per person sonali ality ty fun functi ctioni oning ng is mod modest est”” (p. 280 280). ). The ass associ ociati ations ons are stronge str ongerr con contem tempor porane aneous ously ly tha than n the they y are pre predic dictiv tively ely.. Wit Within hin the con contex textt of the pre presen sentt the theory ory,, it is not how the content of internal working models is determined by early experience that is of interest; rather, the extent to which early experience can jeopardize the very existence of a model, the processing skills required to deal with interpersonal interaction, the robustness of  the model, the extent to which this interpretative mechanism can function under stress and process emotionally charged information, is the focus of the present enquiry. Thus, attachment classification might or might not be stable from infancy through middle childhood to adolescence. As prediction comes from the IIM, not from attachment security per se, this is of no great concern. The focus of study should not be attachment security, which achieved significance as a correlate of the IIM, but has little stability, and possibly little predictive value. Rather, the interpersonal interpretive mechanism, which is a genetically defined capacity, probably localized in the medial prefrontal cortex, is the mechanism of predictive significance. Studies with patients with orbital-frontal and medial-frontal lesions have repeatedly suggested specific deficits in tasks that call for thinking about mental states in others (Channon & Crawford, 2000; Stuss, Gallup, Gallup, & Alexa Alexander, nder, 2001). Both PET and FMRI studies where subjects subjects were asked to make inferences about the mental states of others found activity associated with mentalizing in the medial prefrontal cortex (Frith & Frith, 1999; Gallagher, Happe, Brunswick, Fletcher, & Frith, 2000). In addition, activity was elicited in the tempero-parietal junction (Gallagher et al., 2000). There is independent evidence for the developmental vulnerability of this structure from PET scan studies of Romanian adoptees who were deprived of the interpersonal experiences that we think might generate the IIM (Perry, 1997). Independently, we, of course, know that the attachm attachment ent classific classification ation of these these adoptee adopteess remains remains disor disorganize ganized d at age three and their socia sociall behavior is abnormal at age eight (e.g., Chisolm, 1998). We also have evidence that the mentalizing capacity of individuals maltreated in early childhood continues to have significant limitations limit ations (Beeghly (Beeghly & Cicch Cicchetti, etti, 1994). The front frontal al lobes have a centr central al role to play in social behavior, personality, personal memories, and self awareness (Adolphs, Tranel, Damasio, & Damasio, 1995; Channon & Crawford, 2000; Rogers et al., 1999). Damage to these frontal areas has been consistently associated with social and personality deficits that are consistent with the notion of the loss of interpersonal interpretive capacity, viz. impaired social judgements, impaired pragmatics, deficient self-regulation, and impoverished association of social situations with personal affective markers (e.g., Craik, Moroz, Moscovitch, Stass, Winocur, & Tulving, 1999). Fourth, Myron Hofer’s work with rodent pups identified regulatory interactions within the mother–infant relationship that have clear analogies to what is proposed here (Polan & Hofer,

230

●

 P. Fonagy

ioral system. Hofer’s view is analogous to ours in that he proposes that the attachment “relationship provides an opportunity for the mother to shape both the developing physiology and the behavior behavior of her offspr offspring ing through through her her patterned patterned inter interactio actions ns with her infan infant” t” (Polan & Hofer Hofer,, 1999, p. 177). Attachment is not an end in itself—it is a system adapted by evolution to fulfill key ontogenetic physiological and psychological tasks. Hofer’s Hofer ’s refor reformulat mulation ion of attac attachment hment in term termss of regul regulatory atory processes, processes, hidden but observable within the parent – infant interaction, provides a very different way way of explaining the range of phenomena usually discussed under the heading of attachment. The traditional attachment model is clearly circular. The response to separation is attributed to the disruption of a social bond, the existence of which is inferred from the presence of the separation response. What is lost in “loss” is not the bond but the opportunity to generate a higher order regulatory mechanism: the mechanism for appraisal and reorganization of mental contents. We conceptualize attachment as a process that brings complex mental life into being from a complex and adaptable behavioral system. Some, but by no means all, of such mental function is unique to humans. The mechanisms that generate these (the attachment relationships) have evolutionary continuity across nonhuman species. Just as in rat pups the ontogenetic development of biological regulators crucially depends on the mother–infant unit, so in human development, psychological interpretive capacity evolves in the context of the repetitive interactions with the mother. Fifth, in a series of studies at the Menninger Clinic we explored the factor structure of a number of self-report measures of adult attachment. These studies cannot be described in detail here, but on both community and clinical samples we found very similar results across three investigations. In the first study (Allen, Huntoon, Fultz, Stein, Foagy, & Evans, 2000) two measures of adult attachment style, the Relationship Questionnaire Questionnaire (Bartholomew & Horowitz, 1991) and the Adult Attachment Scale (Collins & Read, 1990), were administered to 253 individuals (99 female trauma patients and 154 community controls). In a principal component analysis a two-component solution accounted for all eigen values greater than 1 and 67.2% of  the total variance. The factor space provided a reasonable two dimensional solution with a secure–fearful axis and a dismissive–preoccupied axis. We found the same two factors: a secure–fearful axis and a dismissive–preoccupied axis in a replication study by Stein et al. (2002), which used five adult attachment questionnaire measures, again on a mixed population. When we plotted the subjects in the sample, both patients and community controls, on the same two principal principal components components it was clear that although although the secure secure– – fearf fearful ul axis was excel excellent lent at distinguishing the community sample from the patient group, the preoccupied–dismissive axis did not distinguish the groups well. What was also clear was a somewhat unexpected relationsh relat ionship ip betwe between en compon component ent score scores. s. Altho Although ugh the overa overall ll corre correlatio lation n betwe between en the two scales was negligible as one might expect, the discrimination between preoccupied and dismissive was somewhat greater towards the middle point of the secure–fearful dimension. One way of interpreting these data is to assume that security represents an experience of  safety in closeness, while fearfulness relates to a disorganization of attachment. Fearfulness appears to be specific to attachment relationships, as nonattachment relationships rarely score highly on this dimension. The dismissing attachment style appears to offer protection to the self by isolation, whereas in enmeshed preoccupation self-protection is perhaps afforded by an amplification of the other, by a denial or subjugation of the self. We would argue that the safety to fearfulness dimension corresponds to the quality of  functioning of the Interpersonal Interpretative Mechanism. At the high end, individuals are

 From Infancy to Adulthood 

●

231

chological mechanism crucially underpinning attachment is somewhat weaker (as a function of attachment history or biology), the capacity for sustaining a clear distinction between self  and other also becomes weaker. In such a situation situation the individual will will require speci specific fic strat strategies egies to accommodate to interpersonal encounters. The two prototypical strategies are the avoidant and resistant strategies. But why are such strategies necessary? Both serve to protect the self in the context of  intensee inter intens interperso personal nal relat relations. ions. We assum assumee that these strat strategies egies may be neces necessary sary because the self, which is as we have seen the product of the other, forever remains vulnerable to social influence. To avoid such instability, against a background of a relatively insecure internal working model, the individual can either deliberately withdraw and enhance the self-representation relative to the other representations (dismissing), or protectively overamplify and exaggerate the other representation (preoccupied). In either case, the strategies in representational terms are about deliberately separating the other from the self-representation. Neither of these strategies is inherently pathological, although both signal a certain degree of weakness. At the extreme end of the safety to fearfulness dimension, there can be no strategy because the attachment system is not there to sustain a consistent set of defences. In these cases, the lack of the interpretive interpretive mechanism that susta sustains ins social relations functions functions so poorly that the capacity to arrive at representations of the motivational or epistemic mind states of the other independent of those of the self are profoundly compromised. This is attachment disorganization, or rather, the absence of the mental function that sustains attachment. Thus, we conceive of attachment disorganization as at the opposite end of attachment security and as an indicator of the regular failure of the interpersonal interpretive mechanism. The concept of empathy has been variously defined but in this context the availability of  a mechanism that allows one individual to take another’s perspective and infer and to some degree experience their emotional state of mind is central. We believe that a psychological mechanism, which we have labeled the IIM-a, is responsible for this. Other workers have independently come to very similar theoretical conclusions (e.g., Corcoran, 2000). Darwin considered sympathy to be the core moral emotion as it involved the automatic experiencing of the other’s distress, which gives rise to altruistic attempts to offer comfort or relief. Some severely severe ly antis antisocial ocial psychopathic psychopathic individuals individuals lack an adult moral sense, hence the term “moral imbecility.” These individuals cannot feel a sense of right and wrong, guilt or remorse, because the neural mechanisms (IIM-a) underlying these experiences are impaired and these individuals cannot feel empathy or sympathy (Blair, Jones, Clark, & Smith, 1997). We have for a number  of years maintained that the ability to predict and experience the emotions of others is mediated by the fun functi ctiona onall connectio connection n of cognitiv cognitivee rep repres resent entati ations ons of oth other er min minds ds in the med medial ial pre prefro fronta ntall cortex (IIM-c) with emotional control centers (IIM-a) in the OFC and temporal lobe. We argue that these functional connections are impaired in individuals with antisocial personality disorder, some of whom have no IIM-a function (cannot feel empathy at all) but most of whom may simply have poor connections between the IIM-a and IIM-c systems. The purely cognitive task of identifying belief states in the other has been associated with activation of medial prefrontal foci around BA8 using SPECT (Baron-Cohen, Ring, Moriarty, Schmitz, Costa, & Ell, 1994), PET (Blair et al., 1997), and fMRI (Gallagher et al., 2000). Innate emotional responses to facially expressed emotion, particularly fear or sad expressions, underpin the acquisition of empathy and morality, or the moral emotions (Blair et al., 1997). Extreme ASPD (high psychopathy) individuals do not respond autonomically to distressed faces. They appear  to be insensitive to the distinction between moral transgressions that result in harm to someone

232

●

 P. Fonagy

(AM) and the impairment of this structure accounts for the limited morality of the ASPD individuals (Blair, Morris, Frith, Perrett, & Dolan, 1999). Others have suggested that early damage to the OFC is associated with severely impaired and immature moral judgment and evidence of OFC lesions are associated with loss of autonomic responsiveness to social stimuli (Anderson, Bechara, Damasio, Tranel, & Damasio, 1999). The infant’s understanding of their own emotional responses (a precursor of empathy by definition) itself arises arises out of a complex process of early mirroring between caregiver and child, which leads to the creation of a second-order representation of the child’s emotional state (Gergely & Watson, 1996, 1999). I suggest that some individuals with limited capacity for  empathy will show limited AM responses to depiction of distress in children’s faces, while others will show no AM deficit but demonstrate limited OFC activation. Those with AM impairment are likely to have failed to acquire a proper understanding of their own emotional responses that directly led to a failure of empathy. Those with insensitive caregivers (perhaps disorganized early attachments) would not have acquired the second order representation of  emotional states and this level of incapacity manifests in OFC dysfunction. A third level of  failure may occur as a consequence of a disconnection between the medial frontal and the orbital frontal areas (IIM-a and IIM-c), whereby the individual might experience affect in relation to an other’s distress but this is inappropriately or inadequately linked to a representation of the belief and intentional state of that person. This latter pattern of functional disconnection might be most likely to arise as a function of severe social adversity at later stages of  development. To summarize, at least five strands of converging evidence suggest that a key selective advantage of attachment might be the development of an understanding of internal states: (1) secure attachment is associated with favorable outcomes across a wide range of relevant tasks; (2) secure attachme attachment nt predicts predicts precocious precocious performance performance in tasks tasks specifi specifically cally calli calling ng for symbol symbolic ic capacity; (3) the class of early attachment classification has less predictive weight than whether  attachment experiences occurred; (4) attachment has been demonstrated to have other ontogenetic biological functions in mammalian species that have analogies or that may parallel the evolutionary function for attachment proposed here; (5) the factor structure of adult attachment scales sca les separ separate atess out a fac factor tor to do with type type of attach attachmen mentt (pe (perha rhaps ps the Int Intern ernal al Wor Workin king g Mod Model) el) and the quality of attachment (perhaps the Interpersonal Interpretive Mechanism). CONCLUSIONS

Although what Although what is propose proposed d here may sound sound radical, radical, when closel closely y scrut scrutinize inized d it actua actually lly contai contains ns sadly very little that is new. I suggest that infant attachment functions, in part at least, to facilitate the development of an interpersonal interpretative capacity. The quality of the early relationship plays a major role in determining the robustness of that capacity but attachment security per se is less relevant to later development. The interpretive capacity, in turn, has a key role in the processing of social experience. The level of functioning of the IIM will be reflected reflect ed in an indivi individual’ dual’ss abili ability ty to funct function ion in close interperson interpersonal al relationship relationshipss with without out needing to have recourse to strategies for amplifying the distinction between self and other representations senta tions.. The unfol unfolding ding of disturbance disturbance over time is condit conditioned ioned by the inter interpreti pretive ve capac capacity ity —  we speculate that the expression of pathogenic genotypes is made more likely by the poor  functioning of a mechanism designed to differentiate the psychological state of self and other. This is a function of immense importance, as the laborious move from genotype to phe-

 From Infancy to Adulthood 

●

233

cites Francois Jacob, who wrote in Of Flies, Mice and Men (1998): “The century that is ending has been preoccupied with nucleic acids and proteins. The next one will concentrate concentrate on memory and desir desire. e. Will it be able to answer the questions they pose?” REFERENCES

Adolphs, R., Tranel, D., Damasio, H., & Damasio, A.R. (1995). Fear and the human amygdala. Journal of Neuroscience, 15(9), 5879–5891. Allen, J.G., Huntoon, J., Fultz, J., Stein, H.B., Foagy, P., & Evans, R.B. (2000). Adult attachment styles and current attachment figures: Assessment of women in inpatient treatment from trauma related psychiatric disorders. Topeka, KS: Menninger Clinic. Anderson, S., Bechara, A., Damasio, H., Tranel, D., & Damasio, A. (1999). Impairment of social and moral behavior related to early damage in human prefrontal cortex. Natural Neuroscience, 2(11), 1032–1037. Asarnow, J.R., Tompson, M., Hamilton, E.B., Goldstein, M.J., & Guthrie, D. (1994). Family-expressed emotion, childhood-onset depression, and childhood-onset schizophrenia spectrum disorders: Is expressed emotion a nonspecific correlate of child psychopathology or a specific risk factor for depression? Journal of Abnormal Child Psychology, 22(2), 129–146. Axelrod, R. (1984). The evolution of cooperation. New York: Basic Books. Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall. Baron-Cohen, S., Ring, H., Moriarty, J., Schmitz, B., Costa, D., & Ell, P. (1994). Recognition of mental state terms. Clinical findings in children with autism and a functional neuroimaging study of normal adults. British Journal of Psychiatry, 165(5), 640–649. Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: A test of a fourcategory model. Journal of Personality and Social Psychology, 61, 226–244. Beeghly, M., & Cicchetti, D. (1994). Child maltreatment, attachment, and the self system: Emergence of  an internal state lexicon in toddlers at high social risk. Development and Psychopathology, 6, 5–  30. Belsky, J. (1984). The determinants of parenting: A process model. Child Development, 55, 83–96. Blair, R., Morris, J., Frith, C., Perrett, D., & Dolan, R. (1999). Dissociable neural responses to facial expressions of sadness and anger. Brain, 122(5), 883–893. Blair, R.J., Jones, L., Clark, F., & Smith, M. (1997). The psychopathic individual: A lack of responsiveness to distress cues? Psychophysiology, 34(2), 192–198. Bogdan, R.J. (1997). Interpreting minds. Cambridge, MA: MIT Press. Bohman, M. (1996). Predisposition to criminality. Swedish adoption studies in retrospect. In M. Rutter  (Ed.), Genetics of criminal and antisocial behaviour. Chichester: Wiley. Bowlby, J. (1969). Attachment and loss, vol. 1: Attachment. London: Hogarth Press and the Institute of  Psycho-Analysis. Cassidy, J., Kirsh, S.J., Scolton, K.L., & Parke, R.D. (1996). Attachment and representations of peer  relationships. Developmental Psychology, 32, 892–904. Castellanos, F.X., Lau, E., Tayebi, N., Lee, P., Long, R.E., Giedd, J.N., Sharp, W., Marsh, W.L., Walter, J.M., Hamburger, S.D., Ginns, E.I., Rappoport, J.R., & Sidransky, E. (1998). Lack of an association between a dopamine-4 receptor polymorphism and attention defict hyperactivity disorder: Genetic and brain morphometric analyses. Molecular Psychiatry, 3, 431–434.

234

●

 P. Fonagy

Chisolm, K. (1998). A three year follow-up of attachment and indiscriminate friendliness in children adopted from Russian orphanages. Child Development, 69, 1092–1106. Cicchetti, D. (1987). Developmental Developmental psychopathology in infancy: infancy: Illustration from from the study of maltreated youngsters. Journal of Consulting and Clincal Psychology, 55, 837–845. Cohen, M.S. (1999). Families coping with childhood chronic illness: A research review. Family System Health, 17, 149–164. Collins, N.L., & Read, S.J. (1990). Adult attachment, working models and relationship quality in dating couples. Journal of Personality and Social Psychology, 58, 644–663. Comings, D.E., Gonzalez, N., Wu, S., Gade, R., Muhleman, D., Saucier, G., Johnson, P., Verde, R., Rosenthal, R.J., Lesieur, H.R., Rugle, L.J., Miller, W.B., & MacMurry, J.P. (1999). Studies of the 48 bp repeat polymorphism of the DRD4 gene in impulsive, compulsive, addictive behaviors: Tourette syndrome, ADHD, pathological gambling, and substance abuse. American Journal of Medical Genetics, 88, 358–368. Corcoran, R. (2000). In S. Baron-Cohen & D. Cohen & H. Tager-Flusberg (Eds.), Understanding other  minds. Perspectives from developmental cognitive neuroscience (pp. 358–391). Oxford: Oxford University Press. Craik, F.I.M., Moroz, T.M., Moscovitch, M., Stuss, D.T., Winocur, G., & Tulving, E. (1999). In search of the self: a positron emission topography study. Psychological Science, 10, 26–34. Deater-Deckard, K., Fulker, D. W., & Plomin, R. (1999). A genetic study of the family environment in the transition to early adolescence. Journal of Child Psychology and Psychiatry, 40, 769–795. Dennett, D. (1987). The intentional stance. Cambridge, MA: MIT Press. Dumont, R.H., Jacobson, A.M., Cole, C., Hauser, S.T., Wolfsdorf, J.I., Willett, J.B., Milley, J.E., & Wertlieb, D. (1995). Psychosocial predictors of acute complications of diabetes in youth. Diabetic Medicine (Chichester), 12(7), 612–618. Dunn, J., Davies, L.C., O’Connor, T.G., & Sturgess, W. (2000). Parents’ and partners’ life course and family experiences: links with parent–child relationships in different family settings [In Process Citation]. Journal of Child Psychology and Psychiatry, 41(8), 955–968. Dunn, J., Deater-Deckard, Deater-Deckard, K., Pickering, Pickering, K., O’Connor, O’Connor, T.G., & Golding, J. (1998). Children’s adjustment and prosocial behaviour in step-, single-parent, and non-stepfamily settings: Findings from a community study. study. ALSPAC ALSPAC Study Study Team. Avon Longitudinal Longitudinal Study Study of of Pregnancy and Childhood. Journal of Child Psychology and Psychiatry, 39(8), 1083–1095. Dunn, J., & McGuire, S. (1994). Young children’s non-shared experiences: A summary of studies in Cambridge and Colorado. In E.M. Hetherington, D. Reiss, & R. Plomin (Eds.), Separate social worlds of siblings. Hillsdale, N.J.: Erlbaum. Eaves, L.J., Silberg, J.L., Meyer, J.M., Maes, H.H., Simonoff, E., Pickles, A., Rutter, M., Neale, M.C., Reynolds, C.A., Erikson, M.T., Heath, A.C., Loeber, R., Truett, K.R., & Hewitt, J.K. (1997). Genetics and developmental psychopathology: 2. The main effects of genes and environment on behavioral problems in the Virginia Twin Study of Adolescent Behavioral Development. Journal of  Child Psychology and Psychiatry, 38(8), 965–980. Emde, R.N. (1988). Development terminable and interminable. I. Innate and motivational factors from infancy. International Journal of Psycho-Analysis, 69, 23–42. Faraone, S.V., Biederman, J., Weiffenbach, B., Keith, T., Chu, M.P., Weaver, A., Spencer, T.J., Wilens, T.E., Frazier, J., Cleves, M., & Sakai, J. (1999). Dopamine D4 gene 7-repeat allele and attention deficit hyperactivity disorder. American Journal of Psychiatry, 156, 768–770. Fonagy, P., Gergely, G., Jurist, E., & Target, M. (2002). Affect regulation, mentalization and the development of the self. New York: Other Press.

 From Infancy to Adulthood 

●

235

Fonagy, P., Steele, M., Steele, H., & Target, M. (1997). Reflective-functioning manual, version 4.1, for  application to adult attachment interviews. London: University College London. Fonagy, P., & Target, M. (1997). Attachment and reflective function: Their role in self-organization. Development and Psychopathology, 9, 679–700. Fonagy, P., & Target, M. (in press). Early intervention and the development of self-regulation. Psychoanalytic Inquiry. Fonagy, P., Target, M., Cottrell, D., Phillips, J., & Kurtz, Z. (2000). A review of the outcomes of all treatments of psychiatric disorder in childhood (MCH 17-33). London: National Health Service Executive. Fox, R.A., Platz, D.L., & Bentley, K.S. (1995). Maternal factors related to parenting practices, developmental expectations, and perceptions of child behavior problems. Journal of Genetic Psychology, 156, 431–441. Freud, S. (1920). Beyond the pleasure principle. In J. Strachey (Ed.), The standard standard edition of the complete psycholo psyc hological gical works of Sigm Sigmund und Freud (Vol. 18, pp. 1– 64). London: London: Hoga Hogarth rth Press. Frith, C.D., & Frith, U. (1999). Interacting minds—A biological basis. Science, 286(5445), 1692–1695. Gallagher, H.L., Happe, F., Brunswick, N., Fletcher, P.C., Frith, U., & Frith, C.D. (2000). Reading the mind in cartoons and stories: an fMRI study of “theory of mind” in verbal and nonverbal tasks. Neuropsychologia, 38(1), 11–21. Garmezy, N., Masten, A.S., & Tellegen, A. (1984). The study of stress and competence in children: A building block for developmental psychopathology. Child Development, 55, 97–111. Ge, X., Conger, R.D., Cadoret, R., Neiderhiser, J., & Yates, W. (1996). The developmental interface between nature and nurture: A mutual influence model of child antisocial behavior and parent behavior. Developmental Psychology, 32, 574–589. Gergely, G., & Wats Gergely, Watson, on, J. (199 (1996). 6). The socia sociall biofeedback biofeedback model of paren parental tal affec affect-mir t-mirrorin roring. g. Inte Internati rnational onal Journal of Psycho-Analysis, 77, 1181–1212. Gergely, G., & Watson, J. (1999). Early social-emotional development: Contingency perception and the social biofeedback model. In P. Rochat (Ed.), Early social cognition: Understanding others in the first months of life (pp. 101–137). Hillsdale, NJ: Erlbaum. Goldsmith, H.H., Buss, K.A., & Lemery, K.S. (1997). Toddler and childhood temperament: Expanded content, stronger genetic evidence, new evidence for the importance of environment. Developmental Psychopathology, 33, 891–905. Hamilton, W.D. (1964). The genetic evolution of social behaviour. Journal of Theoretical Biology, 7, 1–  52. Happe, F., & Frith, U. (1996). Theory of mind in autism. In E. Schopler & G.B. Mesibov (Eds.), Learning and cognition in autism. New York: Plenum. Harris, J.R. (1998). The nurture assumption: Why children turn out the way they do. Parents matter less than you think and peers matter more. New York: Free Press. Hetherington, E.M., Bridges, M., & Insabella, G.M. (1998). What matters? What does not? Five perspectives on the association between marital transitions and children’s adjustment. American Psychology, cholo gy, 53(2), 167– 184. Hewitt, J.K., Silberg, J.L., Rutter, M., Simonoff, E., Meyer, J.M., Maes, H., Pickles, A., Neale, M.C., Loeber, R., Erickson, M.T., Kendler, K.S., Heath, A.C., Truett, K.R., Reynolds, C.A., & Eaves, L.J. (1997). Genetics and developmental psychopathology: 1. Phenotypic assessment in the Virginia Twin Study of Adolescent Behavioral Development. Journal of Child Psychology and Psychiatry, 38(8), 943–963.

236

●

 P. Fonagy

Hughes, C., C., Deater-Deckard, K., & Cutting, Cutting, A. (1999). “Speak “Speak roughly roughly to your little boy?” Sex differences in the relations between parenting and preschoolers’ understanding of mind. Social Development, 8, 143– 160 160.. Jacobson, K.C., & Rowe, D.C. (1999). (1999). Genetic and environmental influences influences on the relationships between family connectedness, school connectedness, and adolescent depressed mood: Sex differences. Developmental Psychology (Washington, DC), 35(4), 926–939. Johnson, J.G., Cohen, P., Brown, J., Smailes, E.M., & Bernstein, D.P. (1999). Childhood maltreatment increases risk for personality disorders during early adulthood. Archives of General Psychiatry, 56, 600—605. Kagan, J. (1989). Unstable ideas: Temperament, cognition and self. Cambridge, MA: Harvard University Press. Kandel, E.R. (1998). A new intellectual framework for psychiatry. American Journal of Psychiatry, 155, 457–469. Kandel, E.R. (1999). Biology and the future of psychoanalysis: A new intellectual framework for psychiatry revisited. American Journal of Psychiatry, 156, 505–524. Karmiloff-Smit Karmiloff -Smith, h, A. (199 (1992). 2). Beyo Beyond nd modu modularit larity: y: A deve developm lopmental ental persp perspectiv ectivee on cogn cognitive itive science. Cambridge, MA: MIT Press. Kendler, K.S., Neale, M.C., Prescott, C.A., Kessler, R.C., Heath, A.C., Corey, Corey, L.A., & Eaves, L.J. (1996). Childhood parental loss and alcoholism in women: A causal analysis using a twin-family design. Psycholog Psyc hological ical Medic Medicine, ine, 26, 79– 95. Koo´ s, O., Gergely, G., Gervai, J., & To´ th, I. (2000). The role of infant-generated stimulus contingencies in affect regulation and the development of attachment security. Paper presented at the 12th Biennial International Conference on Infant Studies (ICIS), Brighton. Krueger, R.F., Moffitt, T.E., Caspi, A., Bleske, A., & Silva, P.A. (1998). Assortative mating for antisocial behavior: Developmental and methodological implications. Behavior Genetics (New York, NY), 28(3), 173–186. Laible, D.J., & Thompson, R.A. (1998). Attachment and emotional understanding in pre-school children. Developmental Psychology, 34, 1038–1045. Lakatos, K., Toth, I., Nemoda, Z., Ney, K., Sasvari-Szekely, M., & Gervai, J. (2000). Dopamine D4 receptor (DRD4) gene polymorphism is associated with attachment disorganization in infants. Molecular Psychiatry (Houndmills), 5(6), 633–637. Leslie, A.M., & Thaiss, L. (1992). Domain specificity in conceptual development: Neuropsychological evidence from autism. Cognition, 43(3), 225–251. Losoya, S.H., Callor, S., Rowe, D.C., & Goldsmith, H.H. (1997). Origins of familial similarity in parenting: A study of twins and adoptive siblings. Developmental Psychopathology, 33, 1012–1023. Maccoby, E.E. (2000). Parenting and its effects on children: On reading and misreading behaviour genetics. Annual Review of Psychology, 51, 1–27. Main, M., & Hesse, E. (1990). (1990). Parents’ Parents’ unreso unresolved lved traumatic traumatic expe experienc riences es are relate related d to infan infantt diso disorgan rganized ized attachment status: Is frightened and/or frightening parental behavior the linking mechanism? In M. Greenberg, D. Cicchetti, & E.M. Cummings (Eds.), Attachment in the preschool years: Theory, research and intervention (pp. 161–182). Chicago: University of Chicago Press. Marvin, R.S., & Britner, P.A. (1999). Normative development: The ontogeny of attachment. In J. Cassidy & P.R. Shaver (Eds.), Handbook of attachment: Theory, research and clinical applications (pp. 44–  67). New York: Guilford. McLoyd, V.C. (1990). The impact of economic hardship on black families and children: Psychological

 From Infancy to Adulthood 

●

237

Miles, D., & Carey, G. (1997). Genetic and environmental architecture of human aggression. Journal of  Personality and Social Psychology, 72, 207–217. Morton, J., & Frith, U. (1995). Causal modeling: A structural approach to developmental psychology. In D. Cicchetti & D. J. Cohen (Eds.), Developmental psychopathology. Vol. 1: Theory and methods (pp. 357–390). New York: John Wiley. Mrazek, D.A., Klinnert, M., Mrazek, P.J., Brower, A., McCormick, D., Rubin, B., Ikle, D., Kastner, W., Larsen, G., Harbeck, R., & Jones, J. (1999). Prediction of early-onset asthma in genetically at-risk children. Pediatric Pulmonology (Philadelphia, PA), 27(2), 85–94. Neiderhiser, J.M., Reiss, D., Hetherington, E.M., & Plomin, R. (1999). Relationships between parenting and adolescent adjustment over time: Genetic and environmental contributions. Development and Psychopathology, 35, 680–692. O’Connor, T.G., Caspi, Caspi, A., DeFries, J.C., & Plomin, Plomin, R. (2000). Are associations associations between between parental divorce and child children’ ren’ss adjus adjustment tment gene geneticall tically y media mediated? ted? An adop adoption tion stud study. y. Deve Developme lopmental ntal Psy Psycholo chology gy (New York, NY), 36, 419–428. O’Connor, T.G., Hetherington, E.M., Reiss, D., & Plomin, R. (1995). A twin-sibling study of observed parent-adolescent interactions. Child Development (Chicago, IL), 66(3), 812–829. O’Connor, T.G., Thorpe, K., Dunn, J., & Golding, J. (1999). Parental divorce and adjustment in adulthood: Findings from a community sample. The ALSPAC Study Team. Avon Longitudinal Study of  Pregnancy and Childhood. Journal of Child Psychology and Psychiatry, 40(5), 777–789. Olds, D., Henderson, C.R., Jr., Cole, R., Eckenrode, J., Kitzman, H., Luckey, D., Pettitt, L., Sidora, K., Morris, P., & Powers, J. (1998). Long-term effects of nurse home visitation on children’s criminal and antisocial behaviour: behaviour: 15 year follow-up of a randomized controlled trial. Journal of the American Medical Association, 280, 1238–1244. Perry, B. (1997). (1997). Incubated in terror: Neurodevelopmental Neurodevelopmental factors in the “cycle of violence.” violence.” In J. Osofsky (Ed.), Children in a violent society (pp. 124–149). New York, NY: Guilford Press. Pettit, G.S., Bates, J.E., & Dodge, K.A. (1997). Supportive parenting, ecological context, and children’s adjustment: A seven year longitudinal study. Child Development, 68, 908–923. Pipp-Siegel, S., Siegel, C.H., & Dean, J. (1999). Neurological aspects of the disorganized/disoriented attachment classification system: Differentiating quality of the attachment relationship from neurological impairment. In J.I. Vondra & D. Barnett (Eds.), Atypical attachment in infancy and early childhood among children at developmental risk (vol. 64, pp. 25–44): Monograph of the Society for Research in Child Development. Plomin, R. (1994). Genetics and experience: The interplay between nature and nurture. Thousand Oaks, CA: Sage Publications Inc. Plomin, R., & Bergeman, C.S. (1991). The nature of nurture: Genetic influences on “environmental” measures. Behavior and Brain Sciences, 14, 373–386. Plomin, R., Chipuer, H.M., & Neiderhiser, J.M. (1994). Behavioral genetic evidence for the importance of non-shared environment. In E.M. Hetherington, D. Reiss, & R. Plomin (Eds.), Separate social worlds of siblings (pp. 1–31). Hillsdale, NJ: Erlbaum. Plomin, R., DeFries, J.C., McLearn, G.E., & Rutter, R. (1997). Behavioral genetics (3rd ed.). New York: W.H. Freeman. Polan, H.J., & Hofer, M. (1999). Psychobiological origins of infant attachment and separation responses. In J. Cassidy & P.R. Shaver (Eds.), Handbook of attachment: Theory, research and clinical applications (pp. 162–180). New York: Guilford. Reiss, D., Hetherington, E.M., Plomin, R., Howe, G.W., Simmens, S.J., Henderson, S.H., O’Connor,

238

●

 P. Fonagy

Reiss, D., Hetherington, E.M., Plomin, R., Howe, G.W., Simmens, S.J., Henderson, S.H., O’Connor, T.J., Bussell, D.A., Anderson, Anderson, E.R., & Law, T. (1995b). Genetic questions questions for environmental studies. Differential parenting and psychopathology in adolescence. Archives of General Psychiatry, 52, 925–936. Reiss, D., Neiderhiser, J., Hetherington, E.M., & Plomin, R. (2000). The relationship code: Deciphering genetic and social patterns in adolescent development. Cambridge, MA: Harvard University Press. Robbins, W.T., & Everitt, B.J. (1999). Motivation and reward. reward. In M.J. Zigmond, F.E. Bloom, Bloom, S.C. Landis, J.L. Roberts, & L.R. Squire (Eds.), Fundamental neuroscience (pp. 1246–1260). San Diego: Academic Press. Rogers, R.D., Everitt, B.J., Baldacchino, A., Blackshaw, A.J., Swainson, R., Wynne, K., Baker, N.B., Hunter, J., Carthy, T., Booker, E., London, M., Deakin, J. F., Sahakian, B.J., & Robbins, T.W. (1999). (199 9). Dissociable Dissociable deficits deficits in the decis decisionion-makin making g cogn cognition ition of chron chronic ic amph amphetami etamine ne abus abusers, ers, opia opiate te abusers, patients with focal damage to prefrontal cortex, and tryptophan-depleted tryptophan-depleted normal volunteers: evidence for monoaminergic mechanisms. Neuropsychopharmacology, 20(4), 322–339. Rowe, D. (1994). The limits of family influence: Genes, experience and behaviour. New York: Guilford Press. Rutter, M. (1999). Psychosocial adversity and child psychopathology. British Journal of Psychiatry, 174, 480–493. Sameroff, A.J. (1995). General systems theories and developmental psychopathology. In J. Cicchetti & D.J. Cohen (Eds.), Developmental psychopathology: Vol 1. Theory and methods (pp. 659–695). New York: Wiley. Scarr, S. (1992). Developmental theories for the 1990s: Development and individual differences. Child Development, 63, 1–19. Serketich, W.J., & Dumas, J.E. (1996). The effectiveness of behavioural parent training to modify antisocial behaviour in children: A meta-analysis. Behaviour Therapy, 27, 171–186. Simons, R.L., Beaman, J., Conger, R.D., & Chao, W. (1993). Childhood experience, conceptions of  parenting and attitudes of spouse as determinants of parental behaviour. Journal of Marriage and the Family, 55, 91–106. Solomon, J., & George, C. (1999). Attachment disorganization. New York: Guilford. Spangler, G., Fremmer-Bombik, E., & Grossman, K.E. (1996). Social and individual determinants of  infant attachment security and disorganization. Infant Mental Health Journal, 17, 127–139. Sroufe, L.A. (1986). Bowlby’s contribution to psychoanalytic theory and developmental psychopathology. Journal Journal of Child Psyc Psycholo hology gy and Psy Psychiatr chiatry, y, 27, 841– 849. Stein, H., Koontz, A.D., Fonagy, P., Allen, J.G., Fultz, J., Brethour, J.R., Jr., Allen, D., & Evans, R.B. (2002). Adult attachment: What are the underlying dimensions? Psychol Psychother, 75(Pt 1), 77–  91. Stern, D., Sander, L., Nahum, J., Harrison, A., Lyons-Ruth, K., Morgan, A., Bruschweilerstern, N., & Tronick, E. (1998). Non-interpretive mechanisms in psychoanalytic therapy: The “something more” than interpretation. International Journal of Psycho-Analysis, 79(5), 903–921. Stuss, D.T., Gallup, G.G., & Alexander, M.P. (2001). The frontal lobes are necessary for “theory of  mind.” Brain, 124(Pt 2), 279–286. Swanson, J.M., Flodman, P., Kennedy, J., Spence, M.A., Moyzis, R., Schuck, S., Murias, M., Moriarity, J., Barr, C., Smith, M., & Posner, M. (2000). Dopamine genes and ADHD. Neuroscience and Biobehavioral ehav ioral Reviews, Reviews, 24(1 24(1), ), 21– 25. Thompson, R.A. (1999). Early attachment and later development. In J. Cassidy & P.R. Shaver (Eds.),

 From Infancy to Adulthood 

●

239

of schizophrenia: Implications for family research. British Journal of Psychiatry, 23(Suppl 164), 20–  26. Turkheimer, E. (1998). Heritability and biological explanantion. Psychological Review, 105, 1–10. Van Tol, H.H.M., Wu, C.M., Guan, H.C., Ohara, K., Bunzow, J.R., & Civelli, O. (1992). Multiple dopamine dopa mine D4 recep receptor tor variants variants in the human population. population. Nature, Nature, 358, 149– 152. Walker, J.G., Johnson, S., Manion, I., & Cloutier, P. (1996). Emotionally focused marital intervention for coup couples les with chron chronically ically ill child children. ren. Jour Journal nal of Cons Consultin ulting g and Clini Clinical cal Psy Psycholo chology gy (Was (Washingt hington, on, DC), 64(5), 1029–1036. Weinstein, A.G., & Faust, D. (1997). Maintaining theophylline compliance/adherence in severely asthmatic children: The role of psychologic functioning of the child and family. Annals of Allergy, Asthma, and Immunology (McLean, VA), 79(4), 311–318. Weinstein, A.G., Faust, D.S., McKee, L., & Padman, R. (1992). Outcome of short-term hospitalization for children with severe asthma. Journal of Allergy and Clinical Immunology, 90(1), 66–75. Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of  wrong beliefs in young children’s understanding of deception. Cognition, 13, 103–128.