The Role of Attachment in Personality Development and ...

6

The Role of Attachment in PersonalityDevelopment and Psychopathology

JOHN nowtnv

During the tirst third of this century the two great pro­ponents of developmental psychiatry were Adolf Meyer andSigmund Freud. Both believed that the seeds of mental healthand ill health were sown in childhood and that to understandthe present-day functioning ofa person it is necessary to knowhow he or she has become the man or woman we meet withtoday.

In their approaches to the field Meyer and Freud took verydifferent routes. Initially Freud focused on traumatic familyrelationships, including incest, but soon, for reasons that remainobscure, he claimed that the real-life events he had originallyinvoked as pathogenic had in fact never occurred and that thepatient had only imagined them. Thenceforward the emphasiswas on fantasy. Consequently Freud`s interests became focusedon a person's internal world of mental processes, especially onthe' powerful influence that unconscious processes have on theway a person feels, -thinks, and behaves and, above all, on thedefensive processes that actively keep them unconscious.

Meyer, by contrast, continued to emphasize the part playedby real-life events in shaping personality, but he was never veryspecific about the nature ofthe ones that matter, nor did he

Material in this chapter is reprinted from "Developmental PsychiatryComes of Age," American journal of Psychiatr)-, Vol. 1411, l987; and from"Attachment and Loss: Retrospect and Prospect," American joumal of Ortho­Pljdllhffj, Vol. 52, l982. Permission from the Editors to reprint is gratefullyacknowledged.

229

230 jonx B0\\'Ll3Yadvance any theory of how being exposed to some event orsituation affects a person's mental state. Nonetheless, Meyer'sapproach played a major part in promoting the mental hygienemovement and child psychiatry. ln both these revolutionizingmovements the notion that the environment in which a childgrows up plays a critical part in determining his future mentalhealth has always been a stubbornly held if ill-defined assump­tion. Moreover, these are the fields in which Freud's originalideas about the role of childhood trauma have not only persistedbut have borne valuable fruit.

We can see now that the tremendous strides being madetoday in developmental psychiatry owe a great deal to boththese pioneers. To Adolf Meyer is due the credit for havingcontinued to emphasize the influential role of the events _andsituations a person meets with during his development. To Sig­mund Freud is due the credit for having emphasized the in­fluence on how a person thinks, feels, and behaves that isexerted by his internal world, namely by the way he perceives,construes and structures the events and situations he encoun­ters. Today we know that the central task of developmentalpsychiatry is to study the endless interaction of internal andextemal and how the one is constantly influencing the other,not only during childhood but during adolescence and adultlife as well.

Evidence that happenings within the family during child­hood and adolescence play a major role in determining whethera person grows up to be mentally healthy is now formidable,and a review of the important epidemiological findings hasrecently appeared (Rutter, 1985). For that reason, in this chap­ter l am presenting only a limited amount of data, some epi­demiological and retrospective and some ethological andprospective, and am giving most attention to the conceptualframework within which I believe the different sets of lindingscan best be comprehended. _

As an example of retrospective data l draw on the Endingsof the group led by George Brown and Tirril Harris which hasbeen undertaking sophisticated epidemiological studies to ex­plore the roles of family experience and other social variablesas antecedents oftlepressive and anxiety disorders in atlult lil`c.

AT`l`A(lHMEN`l' AND |‘£ttsoN.-\t-|'t‘\‘ l)E\'l'll.0l’MEN'l` 231

As an example of prospective data I draw on the work of thedevelopmental psychologists who have been inspired by andare busy expanding the brilliant pioneering studies of` MaryAinsworth into the development ofa young child`s capacity tomake`intimate, emotionally mediated relationships with par­ents. Since the findings from these two quite different ap­proaches are highly compatible, the resulting scientific structurecan be likened to a trilithon made up of two stout pillars ofevidence and a crosspiece of theory. A major conclusion is that,whatever influence variations in genetic endowment may exerton personality de\'elopment and psychopathology, an immenseinfluence is unquestionably exerted by environmental variablesof the kinds now being systematically explored.

For many years sensitive clinicians with a .psychoanalyticorientation have been aware that a person`s mental slate isdeeply influenced by whether his intimate personal relation­ships are warm and harmonious or tense, angry, and anxious,or else emotionally remote, possibly nonexistent. Among cli­nicians so oriented a variety of terms are in use: significantother, dependency, symbiosis, object relations. Nevertheless,although a field of manifest importance to psychiatry, there hasbeen no agreement how best to conceptualize it. A number ol'theoretical systems have been proposed, almost all of them de­rived in some sort of way from psychoanalysis, and a mountainofjargon has accumulated. Yet none has generated sttstainedproductive research. ln scientific quarters the whole field hasstruggled for recognition.

ln recent years a new conceptual framework, known asattachment theory, has been proposed (Bowlby, 1969, l973,1980) which is providing explanations for existing data andpromoting rapidly expanding research programs in develop­mental psychology and developmental psychopathology. WhereasFreud in his scientific theorizing felt confined to a conceptualmodel that explained all phenomena, whether physical or bi­ological, in terms ofthe disposition of energy, today we haveavailable conceptual models of much greater variety. Manydraw on such interrelated concepts as organization. pattern.and information, while the purposeful activities of biologicalorganisms can be conceived in terms of control systems struc­

232 Jonn tsowtuvtured in certain ways. The world of science in which we live isradically different from the world Freud lived in at the turn ofthe century, and the concepts available to us immeasurably bet­ter suited to our problems than were the very restricted onesavailable in his day. Before describing this new conceptualframework, some account of why new concepts have been in­troduced and how the field has evolved may be of interest.

The Evolution of a New Conceptual Framework

Deprivation of Malemal Care and IL9 I ll Effrcls

The origins of the new framework lie in observations madeduring the late 1930s and early 1940s. At that time a numberof clinicians on both sides of the Atlantic, mostly working in­dependently of each other, were making observations of the ill,effects on personality development of prolonged institutionalcare and/or frequent changes ofmother-ligurc during the earlyyears of life. Influential publications followed, including thoseof Lauretta Bender (Bender and Yarnell, 1941; Bender, 1947),Dorothy Burlingham and Anna Freud (1942, 1944), WilliatnC-oldfarb (1943a, b, and c; and six papers summarized in Cold­farb, 1955), David Levy (1937), René Spitz (1945, 1946), andmyself (Bowlby, 1940, 1944). As each ofthe authors was aqualified analyst (except Goldfarb, who trained later), it is nosttrprise that the findings created little stir outside analytic cir­cles.

Then, in late 1949, an imaginative young British psychia­trist, analytically oriented and recently appointed to be Chiefofthe Mental Health Section ofthe World Health Organization,stepped in. Requested to contribute to a United Nations studyof the needs of homeless children, Ronald Hargreaves' decidedto appoint a short-term consultant to report on the mentalhealth aspects of the problem and, knowing of my interest inthe field, invited me to undertake the task. For me this was ztgolden opportunity. After five years as an army psychiatrist, I

‘ Ronald H2l’gI'C2\'CS'§ premature death it\ |962 when professor of psy­chiatry at Leeds was a grievous loss to |)l'C\'Cllll\'C psycltiatry. _

.-\'l`|`.'\(Z1lMEN'l` .-\3l) 1’l'lR5()N:\l_l`|`\' l)l-L\'£l..Ol'3lLN'l`

had returned to child psychiatry determined to explore furtherthe problems 1 had begun working on before the war at theLondon Child Guidance Clinic; I had already appointed as myfirst research assistantjames Robertson, a newly qualified psy­chiatric social worker who had worked with Anna Freud in theHampstead Nurseries during the war. The six months l spentwith the World Health Organization in 1950 gave me the chancenot only to read the literature and discuss it with the authorsbut also to meet many others in Europe and the United Stateswith experience of the field. Soon after the end of my contractI submitted my report, published early in 1951 as a WHOmonograph entitled Marr-rnal Care am! Men/ul Health. ln it 1reviewed the far from negligible evidence then available re­garding thc adverse influences on personality development ofinadequate maternal care during early childhood. called atten­tion tothe acute distress of young children who find themselvesseparated from those they know and love, and made recom­mendations of how best to avoid, or at least mitigate, the short­and long-term ill effects. During the next few years this reportwas translated into a dozen languages and appeared also in acheap abridged edition in English (Bowlby, 1965).

Influential though the written word may often be, it hasnothing like the emotional impact ofa movie. Throughout the19505 Rene Spitz's early film Grizj? A Peril in hjancy (1947),followed soon after by _]a|nes Robertson‘s A Two-Year-Old Goeslo Hospital (1952), had an enormous influence. Not only did thetwo films draw the attention of professional workers to theimmediate distress and anxiety of young children in an insti­tutional setting; they also proved powerful instruments for pro­moting changes in practice. In this field Robertson was to playa lasting part (e.g., Robertson, 1958, 1970).

Although by the end ofthe 1950s a great many of thoseworking in child psychiatry and psychology and in social work,and some also of those in pediatrics and sick children's nursing,had accepted the research findings and were implementingchange, the sharp controversy aroused by the early publicationsand films continued. Psychiatrists trained in traditional psy­chiatry and psychologists who adopted a learning theory ap­proach never ceased to point to the deficiencies of the evidence

234 _IOHN uowt.urand to the lack of an adequate explanation of how the types ofexperience implicated could have the effects on personalitydevelopment that were claimed. In addition, many psychoan­alysts, especially those whose theory focuses on the role of fan­tasy in psychopathology to the relative exclusion of the influenceof real-life events, remained unconvinced and sometimes verycritical. Meanwhile research continued. For example, at YaleSally Provence and Rose Lipton (1962) were making a system­atic study of institutionalized infants in which they comparedtheir development with that of infants living in a family. At theTavistock Clinic members of my small research group wereactive collecting further data on the short-term effects on ayoung child of being in the care of strange people in a strangeplace for weeks and sometimes _months at a time-sec especiallythe studies by Christoph Heinicke (1956) and Heinicke andWestheimer (1966)-while 1 addressed myself to the theoreticalproblems posed by our data.

Meanwhile the field was changing. One important influ­ence was the publication in 1962 by the World Health Orga­nization ofa collection of articles in which the manifold effectsof the various types of experience covered by the term depri­vation qf maternal crm: were reassessed. Of the six articles, by farthe most comprehensive was by my colleague Mary Ainsworth(1962). In it she not only reviewed the extensive and diverseevidence and considered the many issues that had given rise tocontroversy but also identilied a large number of problemsrequiring further research.

A second important influence was the publication, begin­ning during the late iiftics, of Harry Harlow`s studies of theeffects of maternal deprivation on rhesus monkeys; and onceagain film played a big part. Harlow's work in the United Stateshad been stimulated by Spitz`s reports. In the United Kingdomcomplementary studies by Robert Hinde had been stimulatedby our work at the 'l`a\'istock. For the next decade a stream olexperimental results from those two scientists (see summariesin Harlow and Harlow, 1965; Hinde and Spencer-Booth, 1971),courting on top of the .-\insworth review, undermined the op­position. Thereafter, nothing more was heard ofthe inherent

/t't'r.tci~|:\t|ar~t1‘ AND i>t;Rso:<.aL1Tv t>t;\'t»;tor~MExT 235

implausibility of our hypotheses. and criticism became moreconstructive.

Much of course remained uncertain. Even if the reality ofshort-term distress and behavioral disturbance is granted, whatevidence is there, it was asked. that the ill effects can persist?What features of the experience, or combination of features,are responsible for the distress? And, should it prove true thatin some cases ill effects do persist, how is that to be accountedfor? How comes it that some children seem to come throughvery unfavorable experiences relatively unharmed? How im­portant is it that a child should be cared for most of the timeby one principal caregiver? In less developed societies, it wasclaimed (wrongly as it turns out), multiple mothering is notuncommon. In addition to all these legitimate questions, more­over, there were misunderstandings. Some supposed that ad­vocates ofthe view that a child should be cared for most of thetime by a principal mother-figure held that that had to be thechild's natural mother-the so-called blood-tie theory. Otherssupposed that, in advocating that a child experience a warm,intimate, and continuous relationship with his mother (or per­manent mother-substitute), proponents were prescribing a re­gime in which a mother had to care for her child 24 hours aday, day in and day out, with no respite. In a lield in whichstrong feelings are aroused and almost everyone has some sortof vested interest, clear unbiased thinking is not always easy.

The monograph Malemal Care and Mental Health has twoparts. The first reviews the evidence regarding the adverseeffects of maternal deprivation; the second discusses means forpreventing it. What was missing, as several reviewers pointedout, was any explanation of how experiences subsumed underthe broad heading of maternal deprivation could have the ef­fects on personality development of the kinds claimed. Thereason for this omission was simple: the data were not accom­modated by any theory then current, and in the brief time ofmy employment by WHO there was no possibility of developinga new one.

The Child'.t Lic to H is :Wal/ter: Allaclimcnt

At that time it was widely held that the reason a childdevelops a close tie to his mother is that she feeds him. Two

236 joim sowtnvkinds of drive are postulated, primary and secondary. Food isthought of as primary; the personal-relationship, referred toas "dependency," as secondary. This theory did not seem to meto fit the facts. For example, were it true, an infant of a yearor two should take readily to whomever feeds him and thisclearly was not the case. An alternative theory, stemming fromthe Hungarian school of psychoanalysis, postulated a primitiveobject relation from the beginning. In its best-known version,however, the one advocated by Melanie Klein, the mother‘sbreast is postulated as the lirst object, and the greatest emphasisis placed on food and orality and on the infantile nature ofdependency. None of these features matched my experienceof children.

But if the current dependency theories were inadequate,what was the alternative?

During the summer of 1951 a friend mentioned to me thework of Konrad Lorenz on the following responses ofducklingsand goslings. Reading about this and related work on instinctivebehavior revealed a new world, one in which scientists of highcaliber were investigating in nonhuman species many of theproblems with which we were grappling in the human, in par­ticular the relatively enduring relationships that develop inmany species, lirst between young and parents and later be­tween mated pairs, and some of the ways in which these de­velopments can go awry. Could this work, I asked myself, castlight on a problem central to psychoanalysis, that of “instinct”in humans?

Next followed a long phase during which I set about tryingto master basic principles and to apply them to our problems,starting with the nature of the child's tie to his mother. HereL,orenz's work on the following response ofducklings and gos­lings (1935) was ol` special interest. lt showed that in someanimal species a strong bond to an individual mother-ligurecould develop without the intermediary of food: for theseyoung birds are 1mtf1'zll{\'/1arm1Lrl)ut feed themselves by catchinginsects. Here then was an alternative model to the traditionalone, and one that had a number of features that seemed possiblyto fit the human case. Thereafter, as my grasp of ethologiealprinciples increased and l applied them to one clinical problem

.-\'l'1`.-\CHl\lEN'l` AND PERSON.-\Ll'l'\' Dl-l\'l-lLOl'§1liN`l` 237

after another, 1 became increasingly confident that this was apromising approach. Thus, having adopted this novel point ofview, I decided to "follow it up through the material as long asthe application of it seems to yield results" (to borrow a phrasefrom Freud, 1915).

From 1957. when "The Nature of the (2liild`s 'lie to HisMother" was first presented (Bowlby. 1958) through 1969,when Attachment appeared until 1980 with the publication ofLats, l concentrated on this task. The resulting conceptualframework’ is designed to accommodate all those phenomenato which Frettd called attention-for example, love relations,separation anxiety, mourning, defense, anger, guilt, depres­sion, trauma, emotional detachment. sensitive periods in earlylife-and so to off`er an alternative to the traditional metapsy­cliology of psychoanalysis and to add yet another to the manyvariants ofthe clinical theory now extant. 1-low successful theseideas will prove only titnc will tell.

As Kuhn has emphasized. any novel conceptual l`rameworkis difficult to grasp, expecially so for those long familiar witha previous one. Of the many difficulties tnet with in under­standing the framework advocated, I describe only a few. Oneis that, instead of starting with a clinical syndrome oflater yearsand trying to trace its origins retrospectively, 1 have started witha class of childhood traumata and tried to trace their sequelaeprospectively. A second is that, instead of starting with the pri­vate thoughts and feelings of a patient, as expressed in freeassociations or play, and trying to build a theory of personalitydevelopment f`rom those data, I have started with observationsofthe behavior of children in certain sorts of defined situation,including records of the feelings and thoughts they express,and have tried to build a theory of personality developmentfrom there. Ot11er difficulties arise from my use of conceptssuch as control system (instead of psychic energy) and devel­opmental pathway (instead of libidinal phase) which, althoughnow firmly established as key concepts in all the biological sci­

' This is the term Thomas Kuhn (1974) now uses to replace "paradigm,"the term he used in his earlier work (1962).

238 JOHN uowtuvences, are still foreign to the thinking of a great many psy­chologists and clinicians.

Having discarded the secondary drive (dependency) theoryof the child`s tie to his mother, and also the Kleinian alternative,a first task was to formulate a replacement. This led to theconcept of attachment behavior as a special class of behaviorwith its own dynamics distinct from the behavior and dynamicsof either feeding or sex, the two sources of human motivationfor long widely regarded as the most fundamental. Strong sup­port for this step soon came from Harlo\v's finding that inanother primate species-rhesus macaques-infants show amarked preference for a soft dummy "mother," despite its pro­viding no food, to a hard one that does provide it (Harlow andZimmermann, 1959).

Attachment behavior is any form of behavior that resultsin a person attaining or maintaining proximity to some otherclearly identified individual who is conceived as better able tocope with the world. It is most obvious whenever the personis frightened, fatigued, or sick, and is assuaged by comfortingand caregiving. At other times the behavior is less in evidence.Nevertheless, for a person to know that an attachment figureis available and responsive gives him a strong and pervasivefeeling of _security, and so encourages him to value and continuethe relationship. While attachment behavior is at its most ob­vious in early childhood, it can be observed throughout the lifecycle, especially in emergencies. Since it is seen in virtually allhuman beings (though in varying patterns), it is regarded as anintegral part of human nature and one we share (to a varyingextent) with members of other species. The biological functionattributed to it is that of protection. To remain within easyaccess of a familiar individual known to be ready and willingto come to our aid in an emergency is clearly a good insurancepolicy-whatever our age.

By conceptualizing attachment in this way, as a fundamen­tal form of behavior with its own internal motivation distinctfrom feeding and sex, and of no less importance for survival,the behavior and motivation are accorded a theoretical statusnever before given them-though parents and clinicians alikehave long been intuitively aware of their importance. Hitherto

.=t'|'i',\cmn;N'i‘ AND v£itsoN.u_|'rr l)l-1\' 1-;i.om1£N'i‘ 239

the terms dvpcitderiq and d1p¢'ndcnq'11crdl1a\'e been used to refc rto them, but these terms have serious disadvantages. In the firstplace, dependency has a pejorative liavor; in the second, it doesnot imply an emotionally charged relationship to one or a veryfew clearly preferred individtials; and, in the third, no valuablebiological function has ever been attributed to it.

lt is now thirty years since the notion ol` attachment waslirst advanced as a useful way of conceptualizing a form ofbehavior of central importance not only to clinicians and todevelopmental psychologists but to every parent as well. Duringthat time attachment theory has been greatly clarified and am­plified. The most notable contributors have been Robert Hinde,who in addition to his own publications (e.g., 1974) has con­stantly guided my own thinking, and Mary Ainsworth, whofrom the late fifties on has pioneered empirical studies of at­tachment behavior both in Africa (1963, 1967) and in the U.S.(Ainsworth and Wittig, 1969; Ainsworth, Blehar, Waters, andWall, 1978) and has also helped greatly to develop theory (e.g.,1969, 1982). Her work, together with that of her students andothers influenced by her (e.g., Sroufe and Waters, 1977; Sroufe,1985; Main, 1985; Bretherton, 1987), has led attachment theoryto be widely regarded as probably the best supported theoryof socioemotional development yet available (Rajecki, Lamb,and Obmascher, 1978; Rutter, 1980; Parkes and Stevenson­Hinde, 1982).

Because my starting point in developing theory was obser­vations of behavior, some clinicians have assumed that the the­ory amounts to no more than a version of behaviorism. Thismistake is due in large part to the unfamiliarity of the concep­tual framework proposed and in part to my own failure in earlyformulations to make clear the distinction to be drawn betweenan attachment and attachment behavior. To say ofa child orolder person that he is attached to, or has an attachment to,someone means that he is strongly disposed to seek proximityto and contact with that individual and to do so especially incertain specilied conditions. The disposition to behave in thisway is an attribute of the attached person, a persisting attributewhich changes only slowly over time and which is unaffectedby the situation of the moment. Attachment behavior, by con­

240 JOHN uowtavtrast, refers to any of the various forms of behavior the personengages in from time to time to obtain or maintain a desiredproximity.

There is abundant evidence that almost every child habit­ually prefers one person, usually his mother-figure, to go towhen distressed but that in her absence he will make do withsomeone else, preferably someone he knows well. On theseoccasions most children show a clear hierarchy of preferencesso that, in extremity and with no one else available, even akindly stranger may be approached. Thus, while attachmentbehavior may in differing circumstances be shown to a varietyof individuals, an enduring attachment, or attachment bond,is confined to very few. A child who fails to show such cleardiscrimination is likely to be severely disturbed.

The theory of attachment is an attempt to explain bothattachment behavior, with its episodic appearance and disap­pearance, and also the enduring attachments that children andother individuals make to particular others. In this theory thekey concept is that of behavioral system. As this concept maybe unfamiliar to many readers, its full exposition is postponeduntil after a description is given of how some of the many otherphenomena of central concern to clinicians are explained withinthe new framework.

Old Problems in a New Light

Srpn ration A nxirly

First, new light is thrown on the problem of separationanxiety, namely, anxiety about losing, or becoming separatedfrom, someone loved. Why "mere separation" should cause anx­iety has been a mystery. Freud wrestled with the problem andadvanced a number of hypotheses (Freud, l92(i; Strachcy,1959). Every other leading analyst has done the same. With nomeans of evaluating these hypotheses, many divergent schoolsof thought have proliferated.

The problem lies, l believe, in an unexamined assumption,made not only by psychoanalysts but by more traditional psy­chiatrists as well, that fear is aroused in a mentally healthy

:\'l'l`.-\Cl'lM l-INT AND l'l-QR5ON.\l.l'l`\' l)l:.\'l-ll.Ol‘Ml-QN`l' 2-l l

person only in situations that everyone would perceive as in­trinsically painful or dangerous, or that are perceived so by aperson only because of his having become conditioned to them.Since fear of separation and loss does not lit this formula. an­alysts have concluded that what is feared is really some othersituation; and a great variety of hypotheses have been advanced.

The difficulties disappear, however, when an ethologicalapproach is adopted. For it then becomes evident that man,like other animals, responds with fear to certain situations notbecause they carry a high risk of pain or danger but becausethey signal an increase of risk. Thus, _just as animals of manyspecies, including man, are disposed to respond with fear tosudden movement or a marked change in level of sound orlight because to do so has survival value, so are many species.including man, disposed to respond to separation l`rom a po­tentially caregiving ligure and for the same reasons.

When separation anxiety is seen in this light, as a basichuman disposition, it is only a small step to understand why itis that threats to abandon a child, often used as a means ofcontrol, are so very terrifying. Such threats, and also threats ofsuicide by a parent, are, we now know, common causes of in­tensilied separation anxiety. Their extraordinary neglect in tra­ditional clinical theory is due, I suspect, not only to aninadequate theory of separation anxiety but to a failure to giveproper weight to the powerful effects, at all ages, of real-lifeevents.

Not only do threats of abandonment create intense anxiety;they also arouse anger, often also of intense degree, especiallyin older children and adolescents. This anger, the function ofwhich is to dissuade the attachment ligure from carrying outthe threat, can easily become dysfunctional. It is in this light,l believe, that we can understand such absurdly paradoxicalbehavior as the adolescent reported by Burnham (1965) who.having murdered his mother, exclaimed, “l couldn't stand tohave her leave me."

Other pathogenic family situations are readily understoodin terms of attachtnent theory. One fairly common example isseen when a child has such a close relationship with his motherthat he has difficulty in developing a social life outside the

242 _|oHN tsowtiii'family, a relationship sometimes described as symbiotic. ln amajority of such cases the cause of the trouble can be traced tothe mother, who, having grown up anxiously attached as a resultof a difficult childhood, is now seeking to make her own childher attachment figure. Far from the child being overindulged,as is sometimes asserted, he is being burdened with having tocare for his own mother. Thus, in these cases, the normal re­lationship of attached child to caregiving parent is found to beinverted.

Mounting

While separation anxiety is the usual response to a threator some other risk of loss, mourning is the usual response toa loss after it has occurred. During the early years of psycho­analysis a number of analysts identified losses, occurring duringchildhood or in later life, as playing a causal role in emotionaldisturbance, especially in depressive disorders. By 1950 a num­ber of theories about the nature of mourning, and other re­sponses to loss, had been advanced. Moreover, much sharpcontroversy had already been engendered. This controversy,which began during the thirties, arose from the divergent the­ories about infant development that had been elaborated inVienna and London. Representative examples of the differentpoints of view about mourning are those expressed in HeleneDeutsch’s "Absence of Grief" (1937) and Melanie Klein's“Mourning and lts Relation to Manic-Depressive States" (1940).Whereas Deutsch held that inadequate psychic developmentrenders children unable to mourn, Klein held that they notonly can mourn but do. ln keeping with her strong emphasison feeding, howe\-er, she held that the object mourned was thelost breast; in addition, she attributed a complex fantasy life tothe infant. Opposite though these theoretical positions are, bothwere constructed using the same methodology, namely, infer­ence about earlier phases of psychological development basedon observations made during the analysis ofolder, emotionallydisturbed subjects. Neither theory had been checked by directobservation of how ordinary children of different ages respondto a loss.

Approaching the problem prospectively as I did lcd me to

/\`l`\`ACll!\l1~1N'l`.-\ND1‘ER$0N;\l.1'1`\’ 1lL\`1Q1.01'5lEN'l` 2-13

different conclusions. During the early 19505 Robertson and1 had generalized the sequence of responses seen in youngchildren during temporary separation from the mother as thoseof protest, despair. and detachment (Robertson and Bowlby.1952). A few years later. when reading a study by Marris (1958)ofhow widows respond to the loss oftheir husband, 1 was struckby the similarity ofthe responses he describes to those observedin young children who have suffered a loss or separation. Thisled me to a systematic study of the literature on mourning,especially the mourning of healthy adults. The sequence ofresponses that commonly occur, it became clear, was very dif­ferent from what clinical theorists had been assuming. Not onlydoes mourning in mentally healthy adults last far longer thanthe six months often suggested in those days, but several com­ponent responses widely regarded as pathological were foundto be common in healthy mourning. These include anger di­rected at third parties, the self, and sometimes at the personlost; disbelief that the loss has occurred (misleadingly termeddenial); and a tendency. often though not always unconscious,to search for the lost person in the hope of reunion. The clearerthe picture of mourning responses in adults became, the clearerbecame their similarities to those observed in children. Thisconclusion when first advanced (Bowlby, 1960, 1961) was muchcriticized, but it has now been amply supported by a numberof subsequent studies (e.g., Kliman, 1965; Parkes, 1972; Fur­man, l974; Raphael, 1982).

Once an accurate picture of healthy mourning has beenobtained it becomes possible to identify features that are trulyindicat.ive of pathology. It becomes possible also to discern manyof the conditions that promote healthy mourning and thosethat lead in a pathological direction. The belief that childrenare unable to mourn can then be seen to derive from gener­alizations that had been made from the analyses of childrenwhose mourning had followed an atypical course. ln many casesthis had been due either to the child’s never having been givenadequate information about what had happened or to therehaving been no one to sympathize with him and help him grad­ually come to terms with his loss, his yearning for his lost parent,his anger, and his sorrow.

244 JOHN nowuavThe next step in this reformulation of theory was to con­

sider how defensive processes could best be conceptualized, acrucial step since defensive processes have always been at theheart of psychoanalytic theory. Before addressing this problem,however, it is necessary to consider some of the other basicfeatures of the new conceptual framework.

Control Systems, Defensive Processes and InternalWorking Models

Exploration from a Secure Base

In addition to attachment behavior, and antithetical to it,is the urge to explore the environment, to play, and to takepart in varied activities with peers. When an individual of anyage is feeling secure he is likely to explore away l`rom his at­tachment figure. But when alarmed, anxious, tired, or unwellhe feels an urge toward proximity. This is the typical patternof interaction between child and parent known as explorationfrom a secure base. Provided the parent is known to be acces­sible and will be responsive when called upon, a healthy childfeels secure enough to explore. At lirst these explorations arelimited both in time and space. Around the middle of the thirdyear, however, a secure child begins to becotne conlidentenough to increase time and distance away-iirst to half daysand later to whole days. As he grows into adolescence his ex­cursions are extended to weeks or months, but a secure homebase remains indispensable nonetheless for optimal functioningand mental health. No concept within the attachment frame­work is more central to the development of personality thanthat of the secure base; as shown below, it is when an individualdoes not have a secure base that development deviates towardpathology.

Control Syslenu

To explain the behavior described above, attachment the­ory draws on two concepts which, though hitherto unemployedin the field of personality development, are now in regular usein physiology and cognitive psychology. First, it is postulated

,\'t'rAct-|M£N'r AND PERSON/\Ll`l`Y DI-;\`ELOl’MEN'l` 245

that within the central nervous system there exists a specialcontrol system, analogous to the physiological control systemsthat maintain physiological measures, such as blood pressureand body temperature, within set limits. ln a way analogous tophysiological homeostasis, this attachment control system main­tains a person’s relation to his attachment figure between certainlimits of distance and accessibility. As such it can be regardedas an example of what can usefully be termed environmentallzomeostasit (Bowlby, 1973). Among situations that activate care­seeking are anything that frightens a child or signals that he istired or unwell. Among situations that terminate careseekingand release him for other activities are comfort and reassurance.Since control systems are themselves sources of activity, tradi­tional theories of motivation which invoke a build-up of psychicenergy or drive are rendered obsolete.

It is in the nature of a control system that it can operateeffectively only within a specified environment. For example,the system regulating body temperature cannot maintain anappropriate temperature when environmental conditions be­come either too hot or too cold, which means that wheneverconditions exceed certain limits the whole organism becomesstressed or dies. Viewed in an evolutionary perspective, it isevident that variation and natural selection have resulted ineach organism's physiological systems being so constructed thatthey operate effectively in the environment to which the specieshas become adapted and that they will become stressed or failin others.

An evolutionary perspective is necessary also if we are tounderstand psychological stress and the environmental condi­tions that cause it. Like other control systems, the system gov­erning human attachment behavior is constructed so as topromote survival in the environment in which the human spe­cies evolved. Since in such an environment it is essential forsurvival that an infant or older child has an attentive and re­sponsive caregiver to go to in an emergency, it follows that theattachment system will be constructed so as to operate mostefficiently in interaction with a person the chilcl believes willrespond promptly and effectively when called upon. It is hardlysurprising, therefore, that a failure to respond by the familiar

246 JOHN tsowiin'caregiver, even if due to physical absence, or a failure to re­spond appropriately, will always cause stress and sometimes betraumatic.

This leads to consideration of the third major componentof human nature relevant to this exposition-namely, caregiv­ing, the prime role of parents and the complement of attach­ment behavior. When looked at in terms of evolution theory,the occurrence of altruistic care ofthe young is readily under­stood, as it serves to promote the survival of offspring (andoften of other relatives as well) and thereby the caring individ­ual's own genes. Nevertheless, this form of explanation consti­tutes a radical shift from most psychological theorizing, indudingFreud's, which has mistakenly assumed that individuals are bynature essentially selfish and th_at they consider the interests ofothers only when constrained to do so by social pressures andsanctions. Nothing, l believe, that stems from an ethologicalperspective has more far-reaching implications for understand­ing human natu_re than this reappraisal of altruism.

Drferuivc Proccsm

Although as a clinician 1 have inevitably been concernedwith the whole range of defenses, as a research worker I_ havedirected my attention especially to the way a young child be­haves toward his mother after a spell in a hospital or residentialnursery unvisited. In such circumstances it is common for achild to begin by treating his mother almost as though she werea stranger; then, after an interval, usually of hours or days, hebecomes intensely clinging, anxious lest he lose her again, andangry with her should he think he may. ln some way all hisfeeling for his mother and all the behavior toward her we takefor granted, keeping within range of her and most notablyturning to her when frightened or hurt, have suddenly van­ished-only to reappear again after an interval. That was thecondition james Robertson and l termed delac/nnrnt and whichwe believed was a result of some defensive process operatingwithin the child.

What is so very peculiar about this strange detached be­havior is of course the absence of attachment behavior in cir­rumstances in which we would confidently expect to see it. Even

/\'|"1',\t;i|s1|;x'r AND |°t;RsoNAL1'i'\' i>ia\=i;Loi»xit;x'|' 247

when he has hurt himself severely such a child shows no signof seeking comfort. Thus, signals that would ordinarily activateattachment behavior are failing to do so. This suggests that insome way and for some reason these signals are failing to reachthe behavioral system responsible for attachment behavior, thatthey are being blocked off and the behavioral system itselfthereby immobilized. What this means is that a system con­trolling such crucial behavior as attachment can in certain cir­cumstances_ be rendered either temporarily or permanentlyincapable of being activated, and with it the whole range offeeling and desire that normally accompanies it is renderedincapable of being aroused.

In considering how this deactivation might be effected, lturned to the work of the cognitive psychologists_(e.g., Dixon,l97l, 1981; Norman, 1976), who during the past twenty yearshave revolutionized our knowledge of how we perceive theworld and how we construe the situations we are in. This rev­olution in cognitive theory not only gives unconscious mentalprocesses the central place in mental life that analysts havealways claimed for them, but presents a picture of the mentalapparatus as well able to shut off information of certain spec­ilied types and of doing so selectively, without the person beingaware of what is happening.

In the emotionally detached children described earlier andalso, I believe, in adults who have developed the kind of per­sonality that Winnicott (1960) describes as “false self' and Kohut(1977) as “narcissistic,°' the infortnalion being blocked off is ofa very special type. Far from being an instance of the routineexclusion of irrelevant and potentially distracting informationthat we engage in all the time and that is readily reversible,what is excluded in these pathological conditions are the signals,arising from both inside and outside the person, that wouldactivate their attachment behavior and that would enable themboth to love and to experience being loved. In other words, themental structures responsible for routine selective exclusion arebeing employed-one might say exploited-for a special andpotentially pathological purpose. This form of exclusion l referto-for obvious reasons-as defensive exclusion, which is, ofcourse, only another way of describing repression. And,just as

248 _|o||N uowt_tsvFreud regarded repression as the key process in every form ofdefense, so I see the role of defensive exclusion.” A fuller ac­count of this information processing approach to the problemof defense, in which defenses are classified into defensive pro­cesses, defensive beliefs. and defensive activities, is given in anearly chapter ofLo.ss (Bowlby. l980).

I tilmtal Working /lforlrls

Returning now tothe attachment control system within thechild, it is evident that for it to operate efficiently it tnust haveat its disposal as much information as possible about self andattachment figure, not only in regard to their respective loca­tions and capabilities but in regard also to ho\v each is likely torespond to the other as environmental and other conditions

>-change. Qbservations lead us to conclude that toward the endof the first year oflife a child is acquiring a considerable knowl­'edge of his immediate world and that during subsequent yearsthis knowledge is best regarded as becoming organized in theform of internal working models, including models of self andmother in interaction with each other. The function of thesemodels is to simulate happenings in the real world, therebyenabling the individual to plan his behavior with all the advan­tages of insight and foresight. The more adequate and accuratethe simulation. of course, the better adapted is the behaviorbased on it likely to be (]ohnson-Laird, 1982). Although ourknowledge of the rate at which these models develop duringthe earliest years is still scanty, there is good evidence that bythe fiftI1 birthday most children are using a sophisticated work­ing tnodel of mother or mother-substitute which includesknowledge of her interests. moods. and intentions. all ol' whichhe can then take into account (Light, 1979). With a cotnple­mentary model of himself, he is already engaging in a complexintersubjective relationship with his mother, who of course hasher own working models both of her child and ol' herself; Be­cause these models are in constant use, day in and day out. their

` As Spiegel (lflfll) points Otll. my term :lr/rmiw mrlmiuu carries at mean­ing very similar lo sl|||i\'llI\‘§ trlrrlivr irmllrnliun.

A'l`l'ACHMENT AND l’ERSONALl'1`\' DEVELOPMENT 249

infiuence on thought, feeling, and behavior becomes routineand operates largely outside of awareness.

Since as clinicians we know that long before a child reachesthe age of 5 the patterns of interaction between him and hismother range vastly in diversity, from smooth-running andhappy to being filled with friction and distress of every kindand degree, and are also apt to persist, the more we know abouthow they originate the better. It is here that recent research bydevelopmental psychologists has made huge strides.

Patterns of Attachment and Their Determinants

Three principal patterns of attachment present during theearly years are now reliably identified, together with the familyconditions that promote them. Ong_g` these patterns is con­sistent with the child's developing heztlthily and two are pre­dictive of disturbed_U'e@op|nent.T'V‘lTit'lT‘patte'§any`U1cindividual develops during these years is found to be pro­foundly influenced bythe way his parents (or other parent-typeligures) treat him. This conclusion, as important as it isig­troversial, is unpopular i| " circles and is constantly chal­lenged. Yet the evidence for it is now weighty and derives froma number of prospective research studies of socioemotionaldevelopment during the lirst 5 years. This research tradition,pioneered by Ainsworth during the 1960s (Ainsworth et al.,1978; Ainsworth, 1985), has since been exploited and ex­panded, notably in the U.S. by Main (Main and Stadtman, 1981;Main and Weston, 1981; Main, Kaplan, and Cassidy, 1985),Sroufe (1983, 1985), and Waters (Waters, Vaughn, and Ege­land, 1980; Waters and Deane, 1985) and in Germany by Gross­mann (Grossmann, Grossmann, and Schwan, 1986).

The pattem of attachment consistent with healthy devel­opment is that of secure attachment, in which the individual isconfident that his parent (or parent ligure) will be available,responsive, and helpful should he encounter adverse or fright­ening situations. With this assurance, he feels bold in his ex­plorations of the world and also competent in dealing with it.This pattern is found to be promoted by the ready availability

250 Joitx iiowriivof a parent, in the early years especially the mother, sensitiveto t_he child's signals and lovingly responsive when he seeksprotection, comfort. or assistance.

A second pattern is that of anxious resistant attachment inwhich the individual is uncertai nt will be avail­able, responsive, or helpful when called upon. Because of thisuncertainty the individual is always prone to separation anxiety,tends to be clinging, and is anxious about exploring the world.This pattern is promoted by a parent who is available and help­ful on some occasions but not on others, b iorisiiid,s in¢§@0ntr<>i_mHUT§5cl"E`nt inwhich the individual has no con carehe will be responded to helpfully; on the contrary, he expectsto be rebuffed. Such an individual attempts to live his life with­out the love and support of others. This pattern is the resultof the individu:tl's mother constantly rebufling him when heapproaches her for comfort or protection. The most extremecases result from repeated rejection and ill treatment, or pro­longed institutionalization. Clinical evidence suggests that if itpersists this pattern leads to a varietLoLpersonality disordersfrom cgmpulsively self-sufficient individuals to persistentlyrle­ling uent oges, w

There is much evidence that, at least in families wherecaregiving arrangements continue stable, the pattern ofattacb­ment between child and nkther, once established, tends to mer­§. For examplfin two cfi`lTerent samHes (Cilifcnnian andCerman) the patterns of attachment to the mother at 12 monthswere found, with but few exceptions, still to be present at 6years (Main, Kaplan, and Cassidy, 1985; Wartner, 1986). Fur­thermore, prospective studies in Minneapolis (Sroufe, l983)have shown that the pattern of attachment characteristic of thepair, as assessed when the child is l2 months old, is highlypredictive also of behavior outside the home in a nursery group3'/2 years later. Thus children who showed a secure patternwith the mothergat 12 months are likely to be descri§d 5&e|rnursery teachers as cheerful and cootgmtive, popular withother cliilifn, re§lient and resourceful. Those whdsliowedan anxious avoidant pattern are likely to be described later asq

f\`I`l`ACl<lMl-.N`l` AND t't_Rsox,tt_|'r\' l)l»L\'l-_l.()l’\lEN`l` 251

emotionally insulated, ho§le_oLa@>o_c_ial, and as unduly seek­ir1g o showed an anxi@s Fe§ t¥ert{are also likely to be described as unduly seeking of attention

and either as tensefimpulsive, 3 \’Cand heTplc§.' " *'“Kmple confirmation of the teacher`s descriptions comes

from independent observers and laboratory assessments ofthesame children (Sroufe, l983; Lalfrenier and Sroufe, 1985). Sim­ilarly, an experimental study done in Germany (Liitltenhaus,Grossmann, and Grossmann, 1985) shows that at 3 years of agechildren earlier assessed as securely attached respond to po­tential failure with increased effort, whereas the insecurely at­tached do the opposite. In other words, the securely attachedchildren respond with confidence and hope that they can suc­ceed, while the insecure are already showing signs of helpless­ness and defeatism.

In a number of these studies detailed observations havebeen made of the way the children’s mothers treated them.Great variability is seen, with high correlations between amother's style of interaction and the child's pattern of attach­ment to her. For example, in one such study (Matas, Arend,and Sroufe, 1978), made when the children were 2'/2 years old,mothers were observed while their children were attemptinga task they could not manage without a little help. Mothers ofsecure toddlers enabled their children to focus on the task,respected their attempts to complete it on their own, and re­sponded with the required help when called upon; communi­cation between mother and child was harmonious. Mothers ofinsecure infants were less sensitive to their toddlers' state olmind, either not giving support and help when appealed to orelse intruding when a child was striving to solve the problemhimself.

In discussing these and similar findings, Bretherton (1987)etnphasizes the easy flow of communication between motherand child in the secure partnerships and concludes that thiseasy communication is possible only when a mother is intuitivelyalive to the crucial part she plays in providing her child witha secure base, variously encouraging autonomy, providing nec­

252 jonn nowuwessary help, or giving comfort according to her child's state ofmind.

Mothers of insecure infants deviate from this sensitive pat­tern of mothering in a great variety of ways. One, commonamong mothers of avoidant infants, is to scoffut a child's bidsfor comfort and support (Main, Kaplan, and Cassidy, 1985).Another, well-known to clinicians, the effects of which are nowbeing observed by developmentalists (Sroufe, _|acobvitz, Man­gelsdorf, DeAngelo, and Ward, 1985; Main and Solomon,1986), is the failure to _rgpect a child's desire for autonomy,and the discouragemint of ex lcTx_tE. This is seen usuzmfnmothers w o, not avtng had a secure home base during theirown childhood, consciously or unconsciously seek to invert therelationship by making an attachment ligure of the child. Inthe past this has too often been labeled “overindulgence" or"spoiling,” which has led to appalling confusion about what isbest for a child.

It is not difficult to understand why patterns of attach ment,once developed, tend to persist. One reason is that the way aparent treats a child, for better or worse, tends to continueunchanged; another is that each pattern tends to be self-per­petuating. A secure child is a happier and more rewarding childto care for, and also is less demanding than an anxious one. Bycontrast, an anxious, ambivalent child is apt to be whiny andclinging, while an anxious, avoidant child keeps his distanceand is ill-tempered and prone to bully other children. In eachof the latter two cases the child's behavior is likely to elicit anunfavorable response from the parent, encouraging the devel­opment of vicious circles.

Although for the reasons given patterns, once formed, arelikely to persist, this is by no means necessarily so. Evidenceshows that during the lirst two or three years pattern of :tt­tachment is a property of the relationship-for example, pat­tern of child to mother may differ from pattern to father-andalso that if the parent treats the child differently the patternwill change accordingly. These changes are among the evidencereviewed by Sroufe (1985) indicating that stability of pattern,when it occurs, cannot he attributed to the child's inborn tem­perament, as has often been claimed. On the contrary, the evi­

ATl`.-\(Zl~lMIiN`I` AND l’ERSONALl`l`\' DEV!-l|.Ol’M£N`l` 253

dence points unmistakably to the conclusion that a host ofpersonal characteristics, traditionally termed temperamentaland often ascribed to heredity, are in fact environmentally in­duced. True, neonates differ from each other in many manyways. Yet the evidence is crystal clear from repeated studiesthat infants described as difficult during their early days areenabled by sensitive mothering to become happy, easy toddlers.Contrariwise, placid newborns can be turned into anxious,moody, demanding, or awkward toddlers by insensitive or re­jecting mothering. Not only did Ainsworth demonstrate this inher original study, but it has been found again and again insubsequent ones. Those who attribute so much to inborn tem­perament will have to think again.

Thus, during the earliest years features of personality cru­cial to psychiatry remain relatively open to change because theyare still responsive to the environment. As a child grows olclcr,however, clinical evidence shows that both the pattern of at­tachment and the personality features that go with it becomeincreasingly a property ofthe child himselfand also increasinglyresistant to change. This means that he tends to impose it, orsome derivative of it, upon new relationships, as that with theteacher in the Minneapolis study. Similarly, experience showsthat he tends also to impose it, or some derivative of it, on afoster mother or a therapist.> These tendencies to impose earlielpatterns on new nga­tionships, Siwin some measure to persist in doing so despite': :§>TFUare,' 'tlf_[&enomena_th_at gaye birthto'p'syEl'i>analWsEy are als5the phenomena that duringave led an increasing number of analysts toembrace an interpersonal or object relations version of psy­choanalytic theory and, in my own case, to advance the attach­ment version with its postulate of internal working models ofself and attachment figure in interaction. Thus far, therefore,the picture presented can be looked upon as a much modifiedand updated variant of traditional psychoanalytic thinking inwhich great emphasis is placed on the particular pattern intowhich each personality comes to be organized during the earlyyears, with its own distinctive working models or, to use thetraditional term, internal world, and on the strong influence

254 jour# nowtnvthereafter that each individuals models have in shaping his orher life.

An Altemative Framework

During the time it has taken to develop the conceptualframework described here, Margaret Mahler has been con­cerned with many of the same clinical problems and some ofthe same features of children's behavior; she has also beendeveloping a revised conceptual framework to account forthem, set out fully in The Psychological Birth of the Human Infant(Mahler, Pine, and Bergman, 1975). To compare alternativeframeworks is never easy, as Kuhn (1962) emphasizes, and noattempt is made to do so here. Elsewhere (Bowlby, 1981) 1describe what 1 believe to be some ofthe strengths ofthe frame­work I favor, including its close relatedness to empirical data,both clinical and developmental, and its compatibility with cur­rent ideas in evolutionary biology and neurophysiology; whatI see as the shortcomings of Mahler’s framework are trenchantlycriticized by Peterfreund (1978) and Klein (1981).

In brief, Mahler's theories of normal development, includ­ing her postulated normal phases of autism and symbiosis. areshown to rest not on observation but on preconceptions basedon_ traditional psychoanalytic theory, which holds that each non­organic psychiatric syndrome represents a regression to a phaseof development in which features of the syndrome are part ofnormal development. Adopting that model has the effect ofignoring almost entirely the remarkable body of new infor­mation about early infancy that has been built up from carefulempirical studies over the past two decades. Although the keyconcepts with which Mahler's theory is built are very different,some of the clinical implications are not dissimilar; there is infact one central idea that is shared (Mahler et al., 1975). Thisis Mahler's concept ofa return to base for "emotional refueling,"which, as we have seen, is described here as the use of anattachment figure :is a secure base from which to explore.

,\'i'r/\t:iim;N'1' AND l’ERS()N;\l.l'l`\' l)l§\'EL()l‘MEN`I` 255

Continuity and Discontinuity: Vulnerability andResilience

Drurlopmnilal /’aIl|wa_)'.\

One of the long-running debates between analysts and oth­ers in our field turns on the extent to which it is believed thatpersonality, once developed during the early years, is open tochange. Analysts have emphasized a strong tendency towardstability and continuity and have explained it as due to thepowerful influence of the individual's existing internal worldon how he construes and responds to every new situation. Criticshave emphasized the extent to which an individual's perform­ance can change given changed conditions of life. Today I hopewe can agree that, if we are to do _justice to the plethora of dat'now available, we must abandon simple dichotomics. First,sharp distinction must be drawn between current functioning,C5Cl1CC or absence of psychiatric dis­order, and personality structure, measured in terms of greateror lesser vulnerability to adverse life events and situations.Linkedclosely to degrees of vulnerability, moreover, are thevery different ways that people feel about their lives, either asmostly enjoyable and to be lived to the full or as a burden tobe endured; as an emotionally rich and varied experience oran emotional desert.

Second, abandoning the traditional psychoanalytic modelof ighagdf development, lixations, and regressions, we arewise at all times to think in terms of the interactions and trans­actions that are constantly occurring between an ever devel­oping personality and the environment, especially the peoplein it. This means that it is necessary to think of each personalityas moving through life along some developmental pathway,with the particular pathway followed always being determinedby the interaction of the personality as it has so far developedand the environment in which it presently finds itself. If asdevelopmental psychiatrists we adopt these ways of thinking,we need to picture each personality as moving through lifealong its own unique pathway. So long as family conditions arefavorable, the pathway will start and continue within the boundsof healthy and resilient development; but should conditions

256 joHN uowuavbecome sufficiently unfavorable at any time, it may deviate toa lesser or greater extent toward some form of disturbed andvulnerable development. Conversely, should an infant be borninto unfavorable conditions, the pathway along which it devel­ops may become deviant very early; yet once a`gain, should therebe change, in this case for the better, there is a chance of thedeviation diminishing to a greater or lesser extent. Examplesof two such deviant pathways are illustrated in Figures l and2.

This leads me back to patterns of attachment, since myhypothesis is that the pathway followed by each developingindividual and the extent to which he or she becomes resilientto stressful life events is determined to a very sign eby the pattern of attachment developed during the early years.Furthermore, this implies that, in identifying the family ex­periences that result in different children developing differentpatterns of attachment, my colleagues are identifying also someof the major determinants of each person`s future resilience orvulnerability to life's hazards, as well as the extent to which heor she will be able to enjoy life.

So far, unfortunately, the detailed prospective studies de­scribed earlier of how some ofthe different patterns of attach­ment come into being, and of their relative persistence, havenot been carried beyond the sixth year. This means that thchypothesis, however plausible, will be without rigorous testingfor some years to come. Meanwhile, however, those who ad­vance the hypothesis point to much supporting evidence com­ing from the other major school of research referred to earlier,namely, epidemiological studies of adults who are sufferingfrom one or another ol' a wide range of psychiatric disabilitiesincluding anxiety and depressive states, suicidal behavior, bor­derline conditions, and sociopathic personality.

From among a large array ofepidemiological studies l havechosen the work of George Brown and Tirril Harris to illustratemy thesis because it is not only ofthe highest quality but directsattention to variables of the kind that the prospective researchalready described is showing to be so very influential. ln ad­dition, they use a sophisticated model of developmental path­ways for interpreting their lindings.

A'l'I`ACHMEN'l` AND I’ERSONALl'l`Y I)I€\’EL()l’MEN`l` 257

Range of potentiallyhealthy pathways

D l l +20' Psychotherapy x : 1begins ` 0 l I I <215' Mother dies I I I

I l~l§~ w__ tilGuiltinducing gilltechniques ol \|[

5 _ discipline W0

|~'|t-t'kE I. Dcvcloptllcntal pzltltway dcviuting to\\~:||'d anxious uttnclnncnt anddepression.

Range ol potentiallyhealthy pathways

20-I + II I A1 O \ I <75- = : Quarrel with mother° p I l

3 10" I I QHelpful teacherllif ‘F th d t_ a er esers5 ll I

Unstable home

0 .

F ttavtuz 2. Developmental patlnt-ay dcviating toward hostility and delinquency

258 JOHN uowtnvEpidemiological Studies

Brown and Harris have studied four large and represent­ative samples of the female population] two in inner Londonboroughs, one in an outer borough, and one in a remote ruralarea of the Scottish Highlands. Their aim has been to identifywomen suffering from depression ofa clinically significant de­gree and to compare them to other women in the same com­munity who are not depressed in order to discover whetherthere are any family or other social variables that distinguishthe two groups. The findings of the first such study, conductedin the inner London borough of Camberwell, were publishedten years ago and are becoming well known (Brown and Harris,1978). In this study they identified four classes of variable thatwere found significantly more often among the depressedmembers of the population. Three of these concern currentevents and conditions: (a) a severe adverse event, usually in­volving an important personal loss or disappointment, that hadoccurred within the year prior to onset; (b) the absence of acompanion in whom to confide; and (c) chronically difficultliving conditions, including extremely bad housing and re­sponsibility for caring for a number of children under the ageof 14. The fourth variable associated with current incapacitatingdepression is an historical one, namely, (d) a wc=man's loss of hertnother due to death or prolonged separation before her elev­enth birthday. In subsequent studies of two large and mainlyworking class samples, in Walthamstow and Islington, findingsof a similar sort have been the rule (Bifulco. 1985; Brown,Harris, and Bifulco. 1986; Harris, Brown, and Bifulco, 1987;Bifulco, Brown, and Harris, 1987). The annual prevalence ofa current affective disorder in women who had lost their motherbefore their eleventh birthday compared to those who had notwere. in the three samples, 43% vs. 14%, 25% vs. 7%. and 34%vs. 17%. Moreover, in the second and third studies it was found

‘ Among samples of depressed women there is often also a higher inci­dence. compared to controls. uf loss of father during childhood: in two ol'these studies the effects ofthis have been explored. Since the detailed analysessliuw that loss of father has much less influence on vulnerability than loss ofmother. further rcl`erent°e lu it is Ul|\f\|C(|.

A'|“t'.=tt:iisn;N'i' AND 1'£itsoN,\t.|'i'\' l)lf.\'l-ll.Ol'MEN'l` 259

that women who had lost their mother during adolescence(between their eleventh and seventeenth birthdays) were alsomore prone to develop depression than the controls, thoughless so than those who had lost their mother when they wereyounger. The consistency of these findings, together with find­ings from a long-term prospective study reported by Wads­worth (1984), lend strong stipport to the clinically derivedhypothesis that childhood loss of mother is likely to render aperson excessively prone to develop psychiatric symptoms, andto do so especially when current personal relationships gowrong.

In their Walthamstow study, members ofthe Brown-Harrisgroup undertook a prolonged and extensive interview with eachwoman in order to gain as much information as possible aboutthe family circumstances during her childhood and adolescenceand also about her subsequent life course, including, for ex­ample, the jobs she had had, her boyfriends, any premaritalpregnancy, and her husband and children (if any). With thesample deliberately structured to include a high proportion ofwomen who had suffered a childhood loss or prolonged sep­aration, it became possible to analyze out the extent to whichfamily variables other than loss may have contributed to awoman`s current vulnerability.

ln keeping with expectations, it was found that both thefamily circumstances that had led to the childhood loss and theadequacy of care a girl had received afterward were of greatconsequence. The worse the family circumstances before theloss and the more inadequate the care after it, the more vul­nerable to depression the girl had become.

The very detailed analyses of these data carried out by theresearchers have led them to explain their findings in terms ofdevelopmental pathways that, through the continuing inter­action of social and personality factors, e\'entuate in the plightof`a working class London girl who, having lost her mother forany reason, is at appallingly high risk of becoming depressed.In the first place, they found, such a loss carried with it a highlikelihood that she would thereafter receive inadequate care;in the Walthamstow study the chances were found to be nobetter than fifty-fifty. Subsequently, her troubles were likely to

260 JOHN aowtuvsnowball. Should the care she received have been poor, riskswere high that she would become pregnant before marriage orthat she would venture into an early and ill-advised marriage.These two occurrences were found to be strongly associatedwith later depression, in part because they usually resulted inher living in very unfavorable conditions at high risk of suf­fering a severe adverse event and with no one in whom toconfide. For these reasons a working class girl who had lost hermother was all too likely to find herself on a slippery slope.

Figure 3, adapted from one designed by Brown and Harris,attempts to illustrate these interacting processes. The thick up­per line represents the girl's mental state. The middle line rep­resents what may for convenience be described as her familyenvironment. The thin lower line represents her socioeconomicenvironment, in this case urban working class. The broken linesindicate the endless ways in which mental state and environmentinteract, with environmental happenings affecting mental stateand mental state in turn affecting the way the girl or youngwoman deals with aspects of her environment.

Two further findings, both from the Islington study, il­lustrate this interaction (Bifulco, Brown, and Harris, 1987). Agirl who received inadequate care, either in or out of her family,is found twice as likely to have developed a negative self-imagewhen compared to a girl who received adequate care. The lig­ures are 54% compared to 27%. Similar differences are re­ported from Edinburgh (by Ingham, Kreitman, Miller,Sashidharan, and Surtees, 1986). Furthermore, the marriageshe enters into is more likely to fail than is that of the morefortunate girl. Here the figures are 36% and 23%. The highrisk of this happening is evidently the result of a chain of :td­verse happenings. For example, when a girl has no caring homebase she may become desperate to find a boyfriend who willcare for her; that, combined with her negative sell'-image, makesher all too likely to settle precipitately for some totally unsuitableyoung man. Premature pregnancy and childbirth are then likelyto follow, with all the economic and emotional dilliculties thatentails. Moreover, in times of trouble, the effects of her previousadverse experiences are apt to lead her to make unduly intensedemands on her husband and, should he fail to meet them, to

depressed

3u

I

current helplessnesshelplessness and low self-esteemChildIOUS

SCVCIC' €VCn

BX@»CUII

/ A

P t >\\ /‘ \ 1' ,"\ / \\ / //< /\ // \ / \ // \ / ~ /// \ // weUnsu port, \ HQ/ /r mar \rn amy \\ \\\a

QU

3fuUU5

iO3U1S:zS§u

S

auachmem "/¢ /4 t// // )// / \_/ _/Loss of Lack of ,mo care ’ ` /\' smuraon gf ° "aa

Working _clas ' 7

Ex

A.2C5.___

56:>sc:I,_:o

'H

1UE.C-:1Q1:

.QQU\

?E.U-JQ...lh2?E

L'3Q.U7:Ei'U6llI1);:u.

262 JOHN nowuivtreat him badly. No wonder one in three of these marriagesbreak up.

Gloomy though these conclusions are, we must rememberalways that a disastrous outcome is not inevitable. The moresecure an attachment a girl has experienced during her earlyyears, we can confidently predict, the greater will her chancesbe of escaping the slippery slope.

Scientific Problems and Practical Implications

In concluding this chapter I refer, all too briefly, first tosome of the many scientific problems that clamor for attentionand, second, to some of the practical implications of our presentknowledge.

A first scientific task is to explore and test the hypothesis,stated earlier, that each person‘s resilience or vulnerability tostressful life events is determined to a very significant degreeby the pattern of attachment he or she develops during theearly years and, more especially, to clarify to what degrees andin what ways these early developed patterns influence subse­quent development. Since, however, there are many subpat­terns of each of the three main patterns, to elucidate theinfluence of each will be a long and time-consuming task. Anintegral part of such research is the further examination ofprecisely what environmental conditions, impinging earlier orlater, enable a person to retain or attain a greater degree ofresilience and. conversely, what push him or her toward greaterdegrees of vulnerability.

Yet another substantial group of variables to be taken intoaccount are the heritable differences which must be assumedto exist between different individuals in their capacity to dealwith the various environmental hazards, especially inadequateparenting, that influence vulnerability.

In undertaking such a series of research programs an earlyrequirement is the development of psychological methods lorassessing patterns of attachment and their derivatives at eachphase of the life cycle. Main and other developmental psy­chologists are already making some promising moves in this

A'I`l`ACHMENT AND PIQRSON.-\l.l l`\` l)E\'rLl.Ol'l\1liNl 263

direction (see the review by Ainsworth, l9S5). while l-lansburg`sSeparation Anxiety Test is proving its worth (Hansburg, l9S6).To cast light on problems of continuity and discontinuity. bothof patterns of attachment and also of different degrees of re­silience and vulnerability, prospective studies following person­ality development through different phases of the life cycle andin different environments are plainly indispensable, despitetheir being very costly.

This brief sketch represents what will be an enormous pro­gram of research in developmental psychopathology (Sroufeand Rutter, l984) and one that will clearly require generationsof research workers. Both in magnitude and in -the biologicalprinciples informing it, it can best be compared to the vastprogram of research which has been undertaken in immunol­ogy. Here, as in our own field, research workers are concernedwith the extraordinarily complex interactions and transactionsthat occur between an organism as it develops over the yearsand the array of hazards present and potentially present in itsenvironment. Concepts analogous to those in our lield includethe degree to which an organism is immune to a wide varietyof hazards and the extent to which an existing state of immunitywill persist or change over time.

What then are the practical implications of our presentknowledge? For the clinician concerned with the assessmentand treatment ofa wide range of psychiatric patients of all agesit provides a developmental psychopathology that is biologicallybased, coherent, and already empirically well-supported. Assuch it can provide guidance for understanding how a patient`sproblems and symptoms have evolved in his transactions withthe particular environments he has encountered from infancyto the present day. Furthermore, it can provide guidance forplanning therapeutic intervention. Though much detailed workis still required in these areas. promising starts have been made(e.g., Heard, 1981, 1982; Belsky and Nezworski, 1988; Lieber­man and Pawl, in press). Elsewhere (Bowlby, 1977, 1988) l haveoutlined the approach to analytic psychotherapy that l believeshould be adopted in the light of current knowledge.9 What, however, is even more important, is the liriu guid­ance that present knowledge gives for prevention. As I have

264 _|o|-in nowtnvemphasized throughout, l believe there is already sufficientevidence, coming from diverse and independent sources, thatpoints to the very substantial influence on personality devel­opment and mental health of the way an individual's parents(or in some cases parent substitutes) treat him or her. Givenaffectionate, responsive, and encouraging parents whothroughout infancy, childhood, and adolescence provide a boyor girl a secure base from which to explore the world and towhich to return when in difficulty, it is more than likely thathe or she will grow up to be a cheerful, socially cooperative,and effective citizen, and to be unlikely to break down in ad­versity. Furthermore, such persons are far more likely thanthose who come from less stable and supportive homes to makestable maniages and to provide their children the same favor­able conditions for healthy development that they enjoyedthemselves (Dowdney. Skuse, Rutter, Quinton, and Mrazek,1985; Quinton and Rutter, I985). These, ofcourse, are age-oldtruths, but they are now underpinned by far more solid evi­dence than ever bclbre.

Unfortunately, as we also know, there is another and lesshappy side to the picture. Children and adolescents who growup without their home base providing the necessary supportand encouragement are likely to be less cheerful, to [ind life,especially intimate relationships, diflicult, and to be vulnerablein conditions of adversity. In addition, they are likely to havedifliculties when they come to marry and have children of theirown. It is fortunate, of course, that despite these handicapssome manage to struggle through, though often at a muchgreater cost to their emotional life than meets the undiscerningeye. Nor must the lbrtunatc exceptions blind us to the rule.Thus, to take an analogy from physiological medicine. the factthat some heavy smokers survive is no argument for tobacco.

lf, by exploiting this new knowledge, the burden of mentalill health is to be reduced. Western societies will have to givefar more attention to measures that encourage and supportstable family life than any do at present. What the social, eco­nomic, and political implications of such measures might be, ifthey are to be effective, will require very serious study and maywell be far-reaching. Experts from many fields will be required.

A'l'l`.~\(1IlMliN'l` ,mn l‘liRS()N.-\l.l'l`\' ni:\'t;|_t>\~st£><1' 265

Yet substantial progress in this important area cannot be ex­pected until members of the mental health professions speakwith a more united voice than they do at present. Let us hopethat we are now within sight of that happening.

References

Ainsworth, M.D. (l£l62). `l`he effects ol` maternal deprivation: A review olEndings and controversy in the context of research strategy. ln: Drpri­vatian ttf Mntrnutl (Inrr: A Rra.ut°.umrn! nf [Lt Eff:-rl_\. l’nhlic Health PapersI4. Geneva: World Health Organization.

i(l963), The development ol' inl`ant-mother interaction among theCanda. ln: DcIcrminant.r Of/llf-l1P|lBt’ll(|'l|i0llfi Vol. 2. ed. B.l\l. Foss. London:Methuen.

1 (1967). lnfancy in Uganda: lnfant care and the growth ofattachment.Baltimore: _johns Hopkins Press.

--(l969). Object relations. dependency and attachment: A theoreticalreview ofthe ll]l2\|‘\l~ll1()ll'ICI' relationship. Child Dnvlu/nnrnl, 40:969- I 025.

l (l982). Attachment: Retrospect and prospect. ln: The Place of Atlnrlt­ment in I-Immun Brhrnfiour. ed. C.M. Parkes S: _]. Stevenson-Hincle. NewYork: Basic Books.

1(l985), l. Pattems of inl`ant-mother attaclnnent: Antecedents and el`­fects on development. ll. Attachments across the life span. Bull. .\’.l’.Acad. Mrd., 6:77l-812.

-Blehar, l\l.C.. Waters. 1-I.. Sc Wall. S. (l';l78). Paltrrm nf Attnclnnrnt.~Assmrd in the Strrmgr Situation nur! al I-Iumr. Hillsdale. N._].: Erlbaum.

-l 8: Wittig, B.A. (l969). Attachment and exploratory behaviour of one­year-olds in a strange sitttation. ln: Drtrrminuntt if infant bcluiviaur: Vol.4. ecl. B.M. Foss. London: Methuen.

llelsky, _]., 8: Nezworski, T.. ecls. (IQSS). Clinical Implimtimu of Aunrltmmt.Hillsdale, N._].: Erlbaum.

Bender, L. (l947), Psychopathic behaviour disorders in children. ln: Handbookof Correctional Psycltalogjr, ed. R.M. Lindner 8: R.V. Seliger. New York:Philosophical Library.

-8: Yamell. H. (l94l). An ohservation nursery. Amer. _/. Pqcliint.,97:|l58-l I7-1.

llifulco, A.'l`.M. (l985), Death ol' mother in childhood and clinical depressionin adult life: A biographical approach to actiology. Doctoral thesis.University of London.

-- Brown, G.W., 8: Harris. `l`.O. (l987), (Lhildhood loss ol' parent andadult psychiatric disorder: The Islington study. journal of Affective Du­ordcrs, l2:ll5-l28.

Buwlby, _|. (l940), The inlluence ol' early environment in the developmentol' neurosis and neurotic character. lnlcnu1t._/_ P.\)‘(,l0'Alllll., 2l:l5-1-l78.

1-(l944), Forty~four juvenile thieves: Their characters and home lil'e.ln!cmat._]. Psycho-Anal.. 25:19-52. l07-l27.

i (l95 I), Maternal Care and Mmtnl Hcalth. Geneva: WHO.

266 tom: uowun'1 (1958), The nature of the chi|d's tie to his mother. lnlrnutl.]. Psycho­

Anal., 39:350-373.1 (l960), Grief and mourning in infancy and early childhood. Thr Pry­

choanalylir Study 1] Ihr Child, |519-52. New York: International Univers­ities Press.

1 (l96l), Processes ol' mourning. lnLrntal._]. Psycho-Anal., 42:3l7-340.1 (l965). Child Carr and lltf Cru-wth ofLot/r. Harmondsworth: Penguin.1 (l969). .llllnrlnnrnr and Lott: Vol. I. Alla(/:mf-nl, 2nd ed. New York: Basie

Books.1 (1973). Allurlnurul und Luis: Vol. 2. Separation: AlLl’|]'l:I and Anger. New

York: Basic Books. _1 (1977). The making antl breaking ofalleetional bonds. Bril._/_ Psycltial..

l30:20l-2l0,42l-l3l.1 (1980). Allaclnnrnl and Lau: Vol. 3. Loss: Sadnus and Drprruiovt. New

York: Basic Books.1 (1981), Psychoanalysis as a natural science. Ink-mat. Rev. Psycho-Anal.,

81243-256.1 (l983). Attachment. conttnunicttion and the therapeutic process. ln:

Th: Secure Bu."-_ New York: Basie Books.Bl'Cll'\Cl'l0|'|, I- (1937). Nw' |>ers|»ecti\'es on attachment relations: security,

communication and internal working models. ln: Handbook afltrfmu Dr­unlopmznl, e<l._|. Osolsky. ‘Znd ed. New York: Wiley.

llrown, G.W. & Harris. `l`. (li-178). The Sorial Origins rf Drprnuiatt: A Study ryP.1)'dll0{f|? Diwnlrr in \‘l'amm. Lonclon: `l'a\'isloclt.

-1 1 8: Bifulco. A. (1986). Long-term effects ol' early loss ol' parent.ln: Depression in Childhood: Detrelopmrnlnl Pmfrcrliws, ed. M. Rutter, C.lzard. 8: l’. Read. New York: Guilford.

Bttrlingham. D., 8: Freutl, A. (I9-I2). Young children in war-time London.London: Allcn 8: Unwin.

-11 ( I9-I-1), In/ants Wil/taut Familia. London: Allen 8: Unwin.Burnham, D.l... (l965). Separation anxiety. Arch. Gen. Psjrclthtt., l3:3-I6-358.Deutsch, H. (l937), Absence of grief. Psjthormul. Quarl.. 6:l2-22.Dixon, N.F. (l 97 l ). Subliminal Prrrrplivm: Thr Nature ju Conlruvrrsy. London:

McGraw-Hill.1 (l98l), l’rt:eonst'iotts Processing. New York: Wiley.Dnwdney. L., Skuse. D.. Rutter. M., Quinton. D.. & Mrazelt. D. (l935). 'l`Ite

nature and quality ol' patenting provicled by women misecl in institutions._/_ Child Pg-cltol. Ed l'\j;fhiul.. 26:599-625.

Ftettd. S. (l9l5). Repression. Sl(|1ldl1NlEd|h0l|, I-l: I4 l-l58.1 (l92G), lnltihitions, symptoms and anxiety. Slumlnrd Edition. 20:87-I74.

London: Hogarth Press. l€|59.I-`ttrtn:|n, E. (l97-I). :l Chtltlit l’rtrrul Dirt. New Haven: Yale University l’ress.(:<»Itll`arb. W. (l943:t). The elleet ul' early institutional care on adolescent

pCI’S0l1alily. Chi/rl l)f1'l'I||/llllfllf, I-1:2l3-223.1 (l943b), `l`l\e ellects ol' early institutional care on adolescent person­

ality. Exfu-rimrulul lialttmlimt, l2: l06.1 (l943t‘). lnfatnt rearing and problem behavior. Autrr. Orlho/rs)-rhiul.,

l3:2"l9-265.1 (l955). Ettmtiunztl and intellectttztl tmtseqttenres nl` psyrhologic (lep­

ATl`.»\Cl<lMl-QNT AND l'liRS()NAl.l'l`Y Dl;\'F.L()|'Ml£N l'

rivation in infancy: A re\':tltt:ttion. ln: l’tyrl|n/mlltultigy nf Childhood, cd.P.H. Hoch &_|. Zubin. New York: (lrune 8: Stratton. pp. |05-1 19.

(lrosstnann, l(.E., Grnsstnann. K.. 8: Scltwan. A. (1986). (Iaptttring the widerview of attachment: A reanalysis of Ains\\~nrtlt`s strange situation. ln:Mrasurirtg Emotions in lryintls um/ Chilrlrnt: Vol. 2. ed. C.E. lzard & P.B.Read. New York: (Iantbriclge University Press. pp. 124-171.

Hansbttrg, HA. (19S6).Rrsmrr/:rs in Sr/mrrtlimt Anxirhv. Malabar. Fla.: Krieger.llarluw, H.F., 8: Harlow. M.l(. (1965). 'l`he aflectional systems. ln: Bt-Itat-iam'

|fNotthutnan Primulrs: Vol. 2. etl. A.M. Schrier. l-l.F. Harlow, & l". Stollnitz.New York: Academic |’ress.

-i S: Zimmermann. R.R. (1959). Affectinnal responses in the infant mon­key. Scirnrz, 130:421-432.

Harris, T.A.. Brown, C.W.. 8: Bifuleu. A. (1987). Loss of parent in childhoodand adult psychiatric disorder: The role of social class position and pre­marital pregnancy. Psychol. Med., 17:163-184. _

Heard, D.H. (1981). The relevance of attachment theory to child psychiatricpractice- Child Psychol. U Pxjrchial., 22:89-96.

l(l982). Family systems and the attachment dynamic._/. Fam. Thrr..4:99-I 16.

Ht-inicke. C. (1956). Some eflects of separating two-year-old children fromtheir parents: A comparative study. Human Rrlutim|.t. 9: 105-176.

-1- Sr Westheimer. l. (1966). Hrirf Sr/narnliuns. New York: InternationalUttirersities l’ress.

Hintle, R.A. (1974). lliological Bases ol' Human Social lleltariour. New York:McGraw-Hill.

-1 8: Spencer-Booth, Y. (1971). Effects of briel' separation front motheron rhesus monkeys. Scimcr, l73:I l 1-I 18.

Ingham, _|.G., Kreitman, N.B.. Miller, l’.M., Sashidharan, S.P.. 8: Surtees.l’.C. (1986), Self-esteem, vulnerability and psychiatric disorder in thecommunity. Bril._]. Psychiul.. 148:375-385.

Irthnson-Laird. P.N. (l983),»\-lcttlullllndrlr. Cambridge: Cambridge UniversityPress.

Klein, Melanie (1940), Mourning and its relation to tnanie-depressive states.ln: Love, Guilt and Rtpamlinn and Olhrr Paprrs. l92I-I9-I6. London: Ho­garth, 1947.

Klein. Milton (1981). On Mahler's autistic and symbiotic phases: an expositionand evaluation. Pxyrlmanal. (sf Conlrmp. Thought, 4:69-105.

Klitnan, G. (1965), Psyfltolngkal Emrrgmcirs If Childhood. New York: Grune8: Stratton.

Koltut. H. (1977), Thr Reslumlion :JI/tr St# New York: International Uni­versities Press.

Kuhtt. T.S. (1962). The Sfrurlurr 14' Srimlyic Rn-ululiuns. 2nd ed. Chicago:University of Chicago Press. 1970.

-1 (1974), Second thoughts on paradigms. ln: `l`he Structure of ScientificTheory, ed. l-`. Suppe. Urbana: University of Illinois l’rcss, pp. 459-499.

l.:tFrcnier. P., 8: Sroufe, L.A. (1985), Proliles of peer cutnpetence in the pre­school: Interrelations between measures, influence of social ecology andrelation to the attachment history. D:-wlupmml Psychol., 2 I :56-69.

Levy. D. (1937), Pritnary affect hunger. Amt.-r._/. P.\)'(,If|lf., 94:643-652.1.iel>erntan, A.F., lk l’aw|, _|.H. (in press). Disorders of attztchmcnt and secure

268 Jonw ttowt_ttvbase behaviour in the second year: Conceptual issues and clinical inter­vention. In: Alktrltmrnl in the Preschool Years: Theory, Research and lnlcr­vention. ed. M. Greenberg, D. Cicchetti, 8: M. Cummings. Chicago:University of Chicago Press.

Light, P. (1979). Development of a child's sensitivity to people. London:Cambridge University Press. _Lorenz, K.Z. (l935). Der Rumpan in tler Umvelt des Vogels._]. Om. licrl., 83.

English trans. in: Imlinrlive Brlu1viour_ ed. C.H. Schiller. New \’ork: ln­ternational Universities Press. l957, pp. 83-l28.

Lotkenhaus, P.. C-rossmann, K.E., 8: Grossmann, K. (l985), lnl"ant-motherattachment at twelve months and style ol' interaction with a stranger atthe age of three years. Child Druclopnmu. 56: l 538-1542.

Mahler, M.S., Pine. F.. 3: Bergman. A. (1975), Thr Psychological Birth of IhrHuman Irjanl. New York: Basic Books.

Main, M., Kaplan, N., & Cassidy._]. (l985), Security in infancy, childhood andadulthood: A move to the level ol' representation. ln: Growing Points inAllachmnu: Theory and Rcsrurclt, ed. l. Bretherton Sc E. Waters. Chicago:University ol' Chicago Press. pp. 66-104.

-T 8: So|omon,_]. (l986), Discovery ol' an insecure-disorganized/dtsorientedattachment pattern. ln: Aff(Cll1'(DfI'fl0I)lHfl|l in limutq, ed. 'l`.B. Brazelton8: M. Yogman. Norwood. N._].: Ablex.

-- 8: Stadtman,_|. (li-ISI), lnl'ant response to rejection of physical contactby the mother: Aggression, avoidance and conllict. j. Amor. Acad. ChildPsjchial., 20:292-307.

--8: Weston. D.R. (|9Sl). The quality ol' tht: t0<ldler`s relationship tomother and to father: Related to conllict behaviour and the readinessto establish new relationships. Child Dnrrlopmrnr, 52:932-940.

Marris, P. (1958), Widows and Their Families. London: Routledge 8: KeganPaul. ­Matas. L., Arend, R.A., & Srnufe. L.A. (1978). Continuity ol' adaptation in

the second year: The relationship between quality ol' attachment andlater competence. Child Dnvrlopmntl. -19:5-I7-556.

Norman, D.A. (l976), .\lrmury and all¢~ittiun: I ulmrlurlion lo lnmmn iiqormnliouprocessing. 2nd etl. New York: Wiley.

Parkes, C.M. (l972), Brrrot-rmenl: Slurlics of Grief in Arlnll Lifr. New York:International Universities Press.

l- 8: Stevenson-Hincle. _|., eds. (|982), The Place of Almcltntml in HmmmBehaviour. New York: Basic Books.

l’eterl`reund. E. (l978), Some critical comments on psychoatnalytic concep­tualizations of inllmcy. lnlrmut._/. I-’.g~rho-Amtl., 59:-127-14 l.

Provence, S., 8: Lipton, R.(J. (l962), IIJIIYILS in Imlilttlioiu. New York: Inter­national Univetsities l’ress.

Quinton. D.. 8: Rutter. M. (l985), Parenting hehaviottr of mothers raisetl incare. ln: Loitgiltttliital Studies in Child I’.tyrI|olu_q)~ und Psyrlriulry: I’mrIicolLesson.: from /lm-arrlt Iisprricitre, ed. /‘\.R. Nicol. New York: Wiley, pp.|57-261.

Rajecki, D.W., lamb, M.l£., 8: Obtnascher, l‘. (l9`78), 'l`owart.ls a general theoryol' infantile attachment: A comparative review of aspects ofthe socialbond. Behavioural if llmin Srintcm, 3:-I l 7-IG4.

Raphael, B. (l982), 'l`he young child :n\d the deattli ofa parent. ln: The Plocr

/\T|`f\C| tm;N'r ANU t't;ttsoN.\t_t't'\' l)E\`l‘1L()l'MEN'l` 269

nf/lilac/tmcvu in Human Behnvior, ed. C.M. Parkes andj. Stevenson-Hinde.New York: Basic Books, pp. I3 I-l50.

Robertson, _]. (l952), Film: A Two-Yrnr-Old Gan lo Hospital. New York: NewYork University Film Library.

l (1958), Film: Going to I-Iospiml with Molhrr. London: Tavistock (IhildDevelopment Research Unit: New York: New York University Film Li­hrary.

-__ (l970), Young C/iilrlrrn in Hm/iilnl. 2nd ed. London: 'l`a\'istock.L- 8: l3owlby,_]. (l952), Responses of young children to separation frotn

their mothers. Courrirr Crnlrr In/rrnnlianulr Eijanrr, ‘2:l3l-H2.Rutter, M., ed. (1980). Srirnlyic Fnnnduliuns rf lhtfclnpmrnlul P.\_\(/|lt|ll)°, Lon­

don: Heinetnann Medical Books.-(l985), Resilience in the face of ad\'et'sity: Protective factors and rc­

sistance to psychiatric disorder. Iiril. j. Piyrlii/it., 147598-6| l.Spiegel, R. (l93l), Re\'ic\v of Lass: Srulnrss and Drprmion by john Bowlby.

Amer. _]. Psycholhrr., 35:598-600.Spitz. R.A. (l945), Hospitrtlistnz An enquiry into the genesis of psychiatric

conditions in early childhood. Thr P.\']‘F/l0l|l|(|hTfl[ Study ofllu Child, I :53-7-l.New York: International Universities Pres.

l (l946). Anatclilic depression. The P.\)'clnmnnl~rlic Sludy nf lhr Child,2:3l3-342. New York: lntenmtional Universities Press.

l (1947), Film: Grirf' A Prril in In/nnrjv. New York: New York UniversityFilm Library.

Sroufc. L.A. (1983), Inlhnt-citregirer :tttachment and patterns of adaptationin preschool: The roots of maladnptation and competence. ln: Minnl-tomSymposium in Child Ps)'rhoIogy.‘ Vol. I6. ed. M. Perlmtttter. Minneapolis:University of Minnesota Press, pp. -l l-Bl.

-1 (1985), Attachment classilication from the perspective of infant-c:tre­giver t'elationships and infant temperament. Child Drvrloprnrnt. 56: I-l4.

_-_]acobvitz. D., Mangelsdorf, S.. Deftngelo, E.. & Ward. MJ. (l985).Generational boundary dissolution between mothers and their preschoolchildren: A relationship systems approach. Child Dct-clopinnu, 56:317-325.

-- Sc Rutter. M. (1984), The domain of developmental psychopathology.Child Development, 55: l7-29.

l- 8: Waters, E. (IQ77), Attachment as an organizational construct. ChildDevelopment, 48: l 184-I I 99.

Str:tchey,_|. (1959), Editor`s introduction to "Inhihitions, symptoms and anx­iety." Slandnrd Edition, 20:77-86. London: Hogarth Press, l959.

Wadsworth, M.EL]. (198-I), Early stress and associations with adult health.behaviour and parenting. ln: Strm and Disability in Childhood, ed. N.R.Butler 8: B. Corner. Bristol:_|ohn Wright.

Wirtner, U.G. (l986), Attachment in infancy and at age six, and children`sself-concept: A follow-up of a German longitudinal study. Doctoral dis­sertation, University of Virginia.

Waters, E., 8: Deane, RE. (l£|B5), Defining and assessing individual differ­ences in attaclnnent relationships: Q-methodology and the organizationof behaviour in infancy and early childhood. ln: Growing Poinu in At­tachment Theory and Rcmirch, ed. I. Brelherton Sc E. Waters. Chicago:Chicago University Press. pp. -I l-65.

-1- Vaughn, B.l'l., 8: Egeland, B.R. (l980), Individual differences in in­

270 _|OHN sowuwfantimother attachment relationships at age one: Anlecedents in neonatalbehaviour in an urban. economically disadvantaged sample. Child Devel­oprnzru, 51:208-216.

Winnicott, D.W. (l960), Ego distortion in terms oflrue and false self. ln: TluMaluralional Process and ilu Facilitating Environment. New York: Interna­tional Universities Press, pp. |40-152.