He was
completely
lost and desperate after her departure, and refused to look at her when she visited him a fortnight later.
Bowlby - Separation
Retrospectively, we might imagine that his work on the powerful emotional bonds between children and their mothers would have been welcomed within the psychoanalytic world.
It was not.
It was certainly not ignored, however; the persuasiveness of Bowlby's presentation made that impossible.
Bowlby's initial work on attachment was given an airing in the most influential psychoanalytic publications of his day and was then attacked by the most prestigious analytic authorities. That he was virtually exiled by the psychoanalytic community was, ultimately, fortuitous. It led to the establishment around Bowlby's work of a research tradition into the nature of emotional attachment, separation and loss that has generated some of the most remarkable, reliable and provocative empirical data of the past fifty years.
What was it that got Bowlby into so much trouble? He introduced the project as a working out of the implications of observations of the responses to loss young children. It was Bowlby's commitment to the importance of his observations that was the problem. The psychoanalysis of Bowlby's day was an interpretive system that was almost hermetically sealed. Freud had provided the Rosetta Stone for deciphering the underlying structures of mind, and Melanie Klein had extended his psychic excavations to the deepest recesses. Armed with their decoding devices, psychoanalysts felt they could see beneath the superficial, behavioral levels of human interaction to the underlying instinctual conflicts and fantasies that generated their deeper meanings.
Bowlby had been asked by the World Health Organization in 1950 to study and offer advice on the plight of homeless children, real children in the real world. The ravaging impact of maternal deprivations and separations made an enormous impression on
-vii-
him -- to his surprise. It is a tribute to how much Bowlby's work has changed our consciousness that we must strain to imagine why he would have experienced the magnitude of such an impact as a surprise.
In the understanding of mid-twentieth century experts, small children had basic physical needs that required tending; the complex emotional relationships involving a unique sense of interpersonal connection evolved only later. In the conditioning paradigms that dominated American psychology in those days, the caregiver was a "secondary-reinforcer," who became important to the child only by virtue of being associatively linked with physical ministrations. And in the language of the prevailing psychoanalytic theory the mother was, similarly, a "need-gratifying object," whose significance developed gradually through her role in
6
satisfying drive pressures. Separations from mothers shouldn't matter terribly much, as long as the child's needs were taken care of.
But separations did matter, Bowlby discovered. Observations of actual children suggested that maternal deprivation was extremely traumatic. Bowlby s position was a little like that of the wife in the old story who discovers her philandering husband in bed with another woman. "Darling" pleads the husband, "who are you going to believe, me or your lying eyes? " Bowlby decided to believe his eyes.
Any good psychological theory applied cleverly can account for virtually any data. Surely the instinct theory of Bowlby's day could be stretched to accommodate his observations. But for Bowlby, the center of gravity of the data on trauma in real children in the real world did not match up with the center of gravity of Freud's drive theory, privileging impersonal, body- based needs and fantasies. So, as Bowlby announces in this work, a new instinct theory was required, in which the powerful emotional attachment between child and mother is understood not as derivative of more basic processes but as fundamental in itself.
This was no minor addition; it challenged psychoanalytic metapsychology at its core. Bowlby liked to recount experiences formative for him as a student listening to case presentations that emphasized unconscious, instinct-based fantasies at the British Psychoanalytic Society; he remembered one conference at which he felt moved to rise and state emphatically, "But there is such a thing as a bad mother! "
Bowlby was not working completely alone; he had conceptual fellow travelers, especially in the Interpersonal Psychiatry of the American Harry Stack Sullivan and the psychoanalytic contributions of other innovators like W. R. D. Fairbairn, Donald Winni-
-viii-
cott and Hans Loewald. Part of what drew so much fire in Bowlby's direction, however, was his clarity. Sullivan was a tortured, blocked writer. Fairbairn was often tedious and difficult. Winnicott was poetic and elusive. Loewald was extremely subtle and often obscure. Bowlby wrote with lucidity and power.
It was amply apparent from the very beginning and throughout that Bowlby regarded his contributions as a direct challenge to certain basic tenets of Freudian theory. And he had data on children in the real world to back them up. And he identified himself very much as a scientist, offering testable hypotheses. And, his links with other scientists, especially the ethologists of his day, made his position extremely persuasive. For the psychoanalytic establishment of the time, this was simply too much to bear. Bowlby became much more interested than the average psychoanalyst in what actually goes on between people in the real world, and the neighboring discipline of ethology provided powerful explanatory concepts for understanding what he had been observing in children's reactions to separation and loss.
Both Freud and Bowlby were extremely involved with Darwin's contribution (one of Bowlby's last works was a biography of Darwin), but their Darwins were very different. Freud's Darwin was part of the first wave of reaction to the extraordinary implications of the theory of evolution; one of Freud's projects was to work out the implications for human psychology of Darwin's demonstration of the continuity between, so-called lower and so- called higher forms of animal life. Freud's fascination with primitivism and his reliance on
7
bestial metaphors are thematic throughout his writings. And Freud's structural model of the psyche is a re-creation, on a microcosmic level, of Darwin's sweeping account of the evolution of species: lower level, primitive energy of the id is transformed by the reality- oriented ego into higher level, aim-inhibited resources for activities consistent with the cultural values of the superego. Ontogeny recapitulates phylogeny.
Bowlby draws on a different Darwin. Like the Ego psychologist Heinz Hartmann, Bowlby was most interested in what Darwin taught about animal adaptation to environmental conditions and niches. In this second volume of his Attachment and Loss trilogy, Bowlby actually refers to Freud as pre-Darwinian because Freud did not grasp the importance of the principle of "natural selection" in Darwin's theory of the evolution of species. Bowlby, like Darwin, was interested in what animals do to maximize their chances for survival. Whereas Freud's Darwin lent himself to the study of internality and unconscious, primitive states, Bowlby's Darwin lent himself
-ix-
to a behavioral analysis of what small children and mothers actually do with each other. Thus, among the most vivid of Bowlby's contributions is his account of the five component instincts that insure the baby's proximity to the mother--the underpinnings for both healthy attachment and traumatic separations and loss.
One advantage of this behavioral emphasis has been that Bowlby's ideas have been applied, with extraordinary effectiveness, to the empirical research tradition that Mary Ainsworth and Mary Main have done so much to develop. Another advantage is the ease with which Bowlby's observations have been adopted by popular pediatric practice and social policy planners to prioritize the personal, emotional dimensions of childcare. Despite our difficulties as a society in investing the resources necessary for psychologically healthy child- evelopment, we have come to take as axiomatic the principle of "bonding" between parents and children and the importance of emotional warmth and security in caregiving. Bowlby's work was central to this enriched understanding.
Until recently, the disadvantage of Bowlby's behavioral emphasis has been the relative underdevelopment of the psychodynamic dimension within attachment theory, which has made the bridge to other psychoanalytic theorizing more difficult. Bowlby's concept of "working models" had an overly schematic, mechanistic feel to it, which lacked the richness of psychoanalytic investigations of the inner world. But the more recent attachment literature (Fonagy) has taken a more inward turn in exploring the concomitants of secure and insecure attachments in the textures of conscious and unconscious subjectivity. And finally psychoanalysis itself has begun to catch up with bowlby. The recent relational turn in psychoanalysis ( Mitchell and Aron) has made bridges between Bowlby's work and contemporary psychoanalytic thought much more compelling.
Part of what makes Bowlby's early work so inspiring thirty years later is his blend of openness and persistence. His intellectual curiosity seems to have known no bounds, and he continued to draw upon many diverse sources for ideas and conceptual tools, including child- observation, ethology, systems theory, and information-processing. Bowlby also knew from early on that he was onto something very important, and his integrity and perseverance in pursuing what mattered, despite criticism, has benefitted us all.
8
Attachment and Loss has been one of the most influential works of this century. Stephen A. Mitchell New York City, October 1999
-x-
Preface
IN the preface to the first volume of this work I describe the circumstances in which it was begun. Clinical experience of disturbed children, research into their family backgrounds, and an opportunity, in 1950, to read the literature and to discuss problems of mental health with colleagues in several countries led me, in a report commissioned by the World Health Organization, to formulate a principle: 'What is believed to be essential for mental health is that the infant and young child should experience a warm, intimate and continuous relationship with his mother (or permanent mother-substitute) in which both find satisfaction and enjoyment' (Bowlby 1951). To support this conclusion evidence was presented for believing that many forms of psychoneurosis and character disorder are to be attributed either to deprivation of maternal care or to discontinuities in a child's relationship with his mother figure.
Though the contents of the report proved controversial at the time, most of the conclusions are now accepted. What has plainly been missing, however, is an account of the processes through which the many and varied ill effects attributed to maternal deprivation or to discontinuities in the mother-child bond are brought into being. It is this gap that my colleagues and I have since striven to fill. In doing so we have adopted a research strategy that we believe is still too little exploited in the field of psychopathology.
In their day-to-day work, whether with disturbed children, disturbed adults, or disturbed families, clinicians have of necessity to view causal processes backwards, from the disturbance of today back to the events and conditions of yesterday. Though this method has yielded many valuable insights into possible pathogenic events and into the kinds of pathological process to which they appear to give rise, as a research method it has grave limitations. To complement it, a method regularly adopted in other branches of medical research is, having identified a possible pathogen, to study its effects prospectively. If the pathogen has been correctly identified and the studies of its effects in the short and long term are skilfully executed, it then becomes possible to describe the processes set in train by the pathogenic agent and also the ways by which they lead to
-xi-
the various consequent conditions. In such studies attention must be paid not only to the processes set in train by the pathogen but also to the very many conditions, internal and external to the organism, that affect their course. Only then can some grasp be had of the particular processes, conditions, and sequences that lead from a potentially pathogenic occurrence to the particular types of disturbance with which the clinician was in the first place concerned.
In adopting a prospective research strategy my colleagues and I early became deeply impressed by the observations of our colleague, James Robertson, who had recorded, both on
9
paper and on film, how young children in their second and third years of life respond while away from home and cared for instead in a strange place by a succession of unfamiliar people, and also how they respond during and after return home to mother ( Robertson 1952; 1953; Robertson & Bowlby 1952). During the period away, perhaps in residential nursery or hospital ward, a young child is usually acutely distressed for a time and is not easily comforted. After his return home he is likely to be either emotionally detached from his mother or else intensely clinging; as a rule a period of detachment, either brief or long depending mainly on length of separation, precedes a period during which he becomes strongly demanding of his mother's presence. Should a child then come to believe, for any reason, that there is risk of a further separation he is likely to become acutely anxious.
Reflecting on these observations we concluded that 'loss of mother figure, either by itself or in combination with other variables yet to be clearly identified, is capable of generating responses and processes that are of the greatest interest to psychopathology'. Our reason for this belief was that the responses and processes observed seemed to be the same as those found to be active in older individuals who are still disturbed by separations they have suffered in early life. These comprise, on the one hand, a tendency to make intensely strong demands on others and to be anxious and angry when they are not met, a condition common in individuals labelled neurotic; and, on the other, a blockage in the capacity to make deep relationships, such as is present in affectionless and psychopathic personalities.
From the start an important and controversial issue has been the part played in the responses of children to separation from mother by variables other than that of separation per se;
-xii-
these include illness, the strange surroundings in which a child finds himself, the kind of substitute care he receives while away, the kind of relations he has both before and after the event. It is plain that these factors can greatly intensify, or in some cases mitigate, a child's responses. Yet evidence is convincing that presence or absence of mother figure is itself a condition of the greatest significance in determining a child's emotional state. The issue is already discussed in Chapter 2 of the first volume, where a description is given of some of the relevant findings, and is taken up again in the first chapter of this one, where attention is given to the results of a foster-care project undertaken in recent years by James and Joyce Robertson in which they 'sought to create a separation situation from which many of the factors that complicate institutional studies were eliminated; and in which the emotional needs of the children would be met as far as possible by a fully available substitute mother' ( Robertson & Robertson 1971). 1 Study of the Robertsons' findings has led to some modification of views expressed in earlier publications, in which insufficient weight was given to the influence of skilled care from a familiar substitute. In parallel with the empirical studies of my colleagues, I have myself been engaged in studying the theoretical and clinical implications of the data. In particular, I have been trying to sketch a schema able to comprehend data derived from a number of distinct sources:
observations of how young children behave during periods when they are away from mother and after they return home to her;
observations of how older subjects, children and adults, behave during and after a separation from a loved figure, or after a permanent loss;
observations of difficulties found during clinical work with children and adults who, during childhood or adolescence, have either experienced a long separation or a loss or had grounds to fear one; these include various forms of acute or chronic anxiety and
10
depression, and difficulties of every degree in making and maintaining close affectional bonds, whether with parent figures, with members of the opposite sex, or with own children.
____________________
1 In addition to their written report the Robertsons have published a series of films on the
children fostered, particulars of which are given in the list of references at the end of this volume.
-xiii-
First steps towards formulating a theoretical schema were taken in a series of papers published between 1958 and 1963. The present three-volume work1 is a further attempt at a formulation.
Volume 1, Attachment, is devoted to problems originally tackled in the first paper of the series, 'The Nature of the Child's Tie to his Mother' ( 1958b). In order effectively to discuss the empirical data regarding the development of that tie and to formulate a theory to account for it, it proved necessary to discuss first the whole problem of instinctive behaviour and how best to conceptualize it. In doing so I drew heavily on findings and ideas contributed by ethologists and also on ideas derived from control theory.
This, the second volume, deals mainly with problems of separation anxiety and covers ground originally tackled in two further papers of the original series, 'Separation Anxiety' ( 1960a) and 'Separation Anxiety: A Critical Review of the Literature' ( 1961a). Once again, in order to comprehend better the problems before us -- the distress occurring during a separation and the anxiety often evident after it -- it has proved desirable first to discuss a broad range of related phenomena and theory, notably the various forms of behaviour taken to be indicative of fear and the nature of the situations that commonly elicit fear. This discussion occupies Part II of the volume; it provides a background against which are considered, in Part III, the great differences in susceptibility to fear and anxiety that are found when one individual is compared with another. Since many of the data required for the completion of this task are missing, much extrapolation is necessary and the resulting picture is patchy. In some places it can be painted in detail, in others only impressionistically. The aim is to provide clinicians and others with principles on which they can base their actions, and research workers with problems to explore and hypotheses to test.
The third volume, Loss, will deal with problems of grief and mourning and with the defensive processes to which anxiety and loss can give rise. It will comprise a revision and amplification of material first published in the remaining papers of the earlier series -- 'Grief and Mourning in Infancy and Early Childhood' ( 1960b), 'Processes of Mourning' ( 1961b), and 'Pathological Mourning and Childhood Mourning' ( 1963).
____________________
1 In the preface to the first volume I refer only to a second volume. During further work,
however, it has become apparent that a third volume will be required. -xiv-
11
Meanwhile two colleagues, Colin Murray Parkes and Peter Marris, have written books in which they approach problems of loss in a way close to my own. The books are Bereavement by Parkes ( 1972) and Loss and Change by Marris (in press).
In the preface to the first volume it was explained that the frame of reference from which I start is that of psychoanalysis. The reasons are several. The first is that my early thinking on the subject was inspired by psychoanalytic work -- my own and others'. A second is that, despite all its limitations, psychoanalysis and its derivatives remain by far the most used of any present-day approach to psychopathology and psychotherapy. A third and most important is that, whereas many of the central concepts of my schema -- object relations (better termed affectional bonds), separation anxiety, mourning, defence, trauma, sensitive periods in early life -- are the stockin-trade of psychoanalytic thinking, until the last decade or two they have been given scant attention by other behavioural disciplines.
Nevertheless, although the initial frame of reference is that of psychoanalysis, there are many ways in which the theory advanced here differs from the classical theories advanced by Freud and elaborated by his followers. A number of these differences are described already in the first chapter of the earlier volume. Others are referred to throughout the present volume, notably in Chapters 2, 5, and 16.
Note to the Paperback Edition
A paperback edition of this volume has provided the opportunity to include a number of additional notes referring to some later published work. These notes appear on pp. 409 14; the points in the text to which they relate are indicated by a dagger (? ). Details of the additional references are given on pp. 436 -7, following the original references. The index has been considerably expanded.
-xv-
Acknowledgements
IN the first volume of this work I listed the many colleagues and friends without whose help over the years these volumes could not have been written; and it is a great pleasure to express to them all once again my very warmest thanks. My debt to them is deep and lasting.
In the preparation of this volume I am indebted especially for help given by Robert Hinde, Mary Salter Ainsworth, and David Hamburg, each of whom read drafts of all or most of the material and offered a great many valuable criticisms and suggestions. James Robertson scrutinized the first chapter and proposed a number of improvements. Others who have contributed in different ways are Christoph Heinicke, Colin Murray Parkes, and Philip Crockatt. To all of them I am deeply grateful for the time and trouble they have given.
To the preparation of the script my secretary, Dorothy Southern, has again brought her customary care and enthusiasm. Library services have again been provided with unfailing efficiency by Ann Sutherland, and editorial assistance, similarly, by Rosamund Robson. The index has been prepared with great care by Lilian Rubin. To each of them my warmest thanks are due.
12
The many bodies that have supported the research for which I have been responsible at the Tavistock Institute of Human Relations since 1948 are listed in the first volume. Throughout the time that this volume has been in preparation I have been a part-time member of the External Scientific Staff of the Medical Research Council.
For permission to quote from published material, thanks are due to the publishers, authors, and others listed below. Bibliographical details of all the works cited in the text are given in the list of references at the end of the volume.
George Allen & Unwin Ltd, London, and Aldine Publishing Co. , Chicago, in respect of Four Years Old in an Urban Community by J. and E. Newson; Dr I. C. Kaufman, Dr L. A. Rosenblum, and Science in respect of 'Depression in Infant Monkeys Separated from their Mothers' (copyright 1967 by the American Association for the Advancement of Science); Methuen & Co.
-xvii-
Ltd, London, in respect of "'Attachment and Exploratory Behaviour of One-year-olds in a Strange Situation'" by M. D. S. Ainsworth and B. A. Wittig, in Determinants of Infant Behaviour, Vol. 4, edited by B. M. Foss; Dr R. F. Peck and Dr R. J. Havighurst in respect of The Psychology of Character Development, published by John Wiley & Sons, Inc. , New York; University of Chicago Press, Chicago, in respect of The Structure of Scientific Revolutions by T. S. Kuhn; University of London Press Ltd, London, in respect of Truancy by M. J. Tyerman.
Acknowledgement is due also to Tavistock Publications Ltd, London, for permission to include, in Chapter 21 of this volume, material that appears in Support, Innovation, and Autonomy edited by R. Gosling; and to the Journal of Child Psychology and Psychiatry for permission to reproduce, as the basis of Appendix I, a paper that was first published in that journal in 1961.
-xviii-
13
Part I
SECURITY, ANXIETY, AND DISTRESS
-1-
Chapter 1
Prototypes of Human Sorrow
Unhappiness in a child accumulates because he sees no end to the dark tunnel. The thirteen weeks of a term might just as well be thirteen years.
GRAHAM GREENE, A Sort of Life
Responses of young children to separation from mother 1
A generation has now passed since Dorothy Burlingham and Anna Freud recorded their experiences of caring for infants and young children in the setting of a residential nursery. In two modest booklets published during the second world war ( Burlingham & Freud 1942; 1944) they describe the immense problem of providing for young children who are out of mother's care. In particular they emphasize how impossible it is in a nursery setting to provide a child with a substitute figure who can mother him as well as his own mother can. When the Hampstead Nurseries were reorganized so that each nurse could care for her own little group of children they tell how the children became strongly possessive of their nurse and acutely jealous whenever she gave attention to another child: 'Tony (3 1/2) . . . would not allow Sister Mary to use "his" hand for handling other children. Jim (2-3) would burst into tears whenever his "own" nurse left the room. Shirley (4) would become intensely depressed and disturbed when "her" Marion was absent for some reason. '
Why, it may be asked, should these children have become so strongly possessive of their nurse and so deeply distressed whenever she was missing? Was it, as some traditionalists might suppose, that they had been spoiled by having been given too much attention and allowed too much their own way? Or
____________________
1 Although throughout this book the text refers usually to 'mother' and not to 'mother figure',
it is to be understood that in every case reference is to the person who mothers a child and to whom he becomes attached. For most children, of course, this person is also his natural mother.
-3-
was it, by contrast, that since leaving home they had been subjected to too many changes of mother figure and/or had too limited access to whoever in the nursery was acting temporarily as their mother figure? On how we answer these questions turn all our practices of child- rearing.
14
Not only did children in these nurseries become intensely possessive and jealous of their 'own' nurse but they were also unusually prone to become hostile towards her or to reject her, or else to retreat into a state of emotional detachment, as the following records illustrate:
Jim was separated from a very nice and affectionate mother at 17 months and developed well in our nursery. During his stay he formed two strong attachments to two young nurses who successively took care of him. Though he was otherwise a well adjusted, active and companionable child, his behaviour became impossible where these attachments were concerned. He was clinging, over-possessive, unwilling to be left for a minute, and continually demanded something without being able to define in any way what it was he wanted. It was no unusual sight to see Jim lie on the floor sobbing and despairing. These reactions ceased when his favourite nurse was absent even for short periods. He was then quiet and impersonal.
Reggie, who had come to our house as a baby of 5 months, went home to his mother when he was 1 year 8 months, and has been with us ever since his return to the nursery 2 months later. While with us, he formed two passionate relationships to two young nurses who took care of him at different periods. The second attachment was suddenly broken at 2 years 8 months when his 'own' nurse married.
He was completely lost and desperate after her departure, and refused to look at her when she visited him a fortnight later. He turned his head to the other side when she spoke to him, but stared at the door, which had closed behind her, after she had left the room. In the evening in bed he sat up and said: 'My very own Mary-Ann! But I don't like her. '
These observations, made in the pressure of wartime and recorded anecdotally with all too little detail, none the less cast a shaft of light on the nature of many forms of psychiatric disturbance. States of anxiety and depression that occur during
-4-
adult years, and also psychopathic conditions, can, it is held, be linked in a systematic way to the states of anxiety, despair, and detachment described by Burlingham & Freud, and subsequently by others, that are so readily engendered whenever a young child is separated for long from his mother figure, whenever he expects such a separation, and when, as sometimes happens, he loses her altogether. Whereas during later life it is often extremely difficult to trace how a person's disturbed emotional state is related to his experiences, whether they be those of his current life or those of his past, during the early years of childhood the relationship between emotional state and current or recent experience is often crystal clear. In these troubled states of early childhood, it is held, can be discerned the prototype of many a pathological condition of later years.
It is, of course, a commonplace that most children who have had experiences of these kinds recover and resume normal development, or at least they appear to do so. Not infrequently, therefore, doubts are expressed whether the psychological processes described are in reality related so intimately to personality disturbances of later life. Pending much further evidence, these are legitimate doubts. Nevertheless, reasons for holding to the thesis are strong. One is that data from many sources can be arranged and organized into a pattern that is internally consistent and consistent also with current biological theory. Another is that many clinicians and social workers find the resulting schema enables them to understand better the problems with which they are grappling and so to help their patients or clients more effectively.
15
Why some individuals should recover, largely or completely, from experiences of separation and loss while others seem not to is a central question, but one not easily answered. In living creatures variation of response is the rule and its explanation is often hard to fathom. Of all those who contract poliomyelitis less than 1 per cent develop paralysis, and only a fraction of 1 per cent remain crippled. Why one person should respond one way and another another remains obscure. To argue that, because 99 per cent recover, polio is a harmless infection would obviously be absurd. Similarly, in the field under consideration, to argue that because most individuals recover from the effects of a separation or loss these experiences are of no account would be equally absurd.
Nevertheless the problem of differential response remains
-5-
important. Conditions likely to be playing a part can be considered under two main heads: those intrinsic to or closely associated with the separation itself, notably the conditions in which a child is cared for while away from mother;
those present in the child's life over a longer period, notably his relations with parents during the months or years before and after the event.
Here we consider variables in category (a). Those in category (b) are discussed in the later chapters of Part III.
We start by reviewing observations of how children behave when cared for in one of two very different settings. The first is an ordinary residential nursery, in which a child finds himself in a strange place with strange people none of whom is sufficiently available to give him more than very limited mothering. The second is a foster home in which a child receives the fulltime and skilled care of a foster mother with whom he has become in some degree familiar beforehand.
Conditions leading to intense responses
In our early studies children were observed during stays in institutional settings and it was on the basis of these observations that the sequence of responses which we term protest, despair, and detachment came first to be delineated ( Robertson & Bowlby 1952). Since then two further studies have been conducted by colleagues in the Tavistock Child Development Research Unit, the first by Christoph Heinicke ( 1956) and the second by Heinicke & Ilse Westheimer ( 1966). Although in each of these investigations only a handful of children were observed (six in the first and ten in the second), the studies are unique for the care of their design and the amount of systematic observation. Moreover, for each sample of separated children a contrast group was selected and observed: in the first study it was a fairly well- matched group of children observed during their first weeks of attendance at a day nursery; in the second it was a similarly matched group of children observed while living in their own homes. Heinicke & Westheimer treat their data statistically and also describe in some detail the behaviour of individual children.
In the larger investigation ( 1966), work was conducted in three residential nurseries. Arrangements and facilities were fairly similar. In each, a child belonged to a defined group of
16
-6-
children and was cared for mainly by one or two nurses. Ample opportunities were available for free play either in large rooms or outdoors in a garden. Before a child entered the nursery, contact was made with the family by a psychiatric social worker ( Ilse Westheimer), who was also responsible either then or later for collecting full information about the family and the child. Arrival at the nursery was observed; and in the course of his stay a child was observed during free play on six occasions each week. Each of the two observers (one male, Christoph Heinicke, and one female, Elizabeth Wolpert) observed for a period of at least half an hour during each of the three sample periods into which the week was broken (Monday and Tuesday; Wednesday and Thursday; Friday, Saturday, and Sunday). The method used, of categorizing behavioural units in terms of agent, object, relation, mode, and intensity, had been used in the earlier study and had been shown to be reliable.
In addition to the categorized observations of free behaviour, similarly categorized observations were made of every child's behaviour in standardized doll-play sessions; and a number of other records of each child's stay in the nursery were kept.
It was originally intended to select the separated children in accordance with the five criteria used in the first study, namely: (i) that the child had had no previous separations of more than three days, (ii) that he fell within the age-limits of fifteen to thirty months, (iii) that he did not enter the nursery with a sibling, (iv) that he was living with both mother and father at the time the separation occurred, and (v) that there was no evidence that being placed in a nursery indicated a rejection by his parents. Because of the difficulty of obtaining cases, however, these criteria had to be modified to allow greater latitude.
Although most of the children had had either no separations or only very brief ones prior to the one being studied, in one case the length of previous separation was four weeks and in two it was three weeks. The age-range was slightly extended and ran from thirteen to thirty-two months, instead of from fifteen to thirty months. But the most marked departure from the previous criteria was that four of the children entered the residential nursery in the company of a sibling: in three cases this was a four-year-old sibling and in one case the sibling was younger. The remaining two criteria remained unmodified: each of the children was living with both mother and father at the time of separation, and there were no indications that he
-7-
was being rejected by the parents by being placed in the nursery.
The reason that the ten children studied were cared for in a residential nursery was that, in a family emergency, neither relatives nor friends were available to take temporary care of them. In the case of seven families, mother was to be away in hospital having a new baby. In two others mother was to be in hospital for some other form of medical attention. In the tenth case, the family became homeless.
Among much else in their book, Brief Separations ( 1966), Heinicke & Westheimer describe behaviour typical of the ten children during their time in the nursery, and, similarly, behaviour typical of the children after they had returned home. In the paragraphs that follow some of their principal findings are presented. Every one of the patterns reported had also been
17
observed and recorded by Robertson during his earlier, less systematic though more extensive, studies.
Behaviour during Separation
The children arrived at the nursery in the care of one or both parents. Four of them, brought by father, stayed close to him and seemed already subdued and anxious. Some of the others, who had come with mother or both parents, seemed more confident and were ready to explore the new environment. They ventured forth on short or long excursions and then returned.
When the moment came for the parent (s) to depart, crying or screaming was the rule. One child tried to follow her parents, demanding urgently where they were going, and finally had to be pushed back into the room by her mother. Another threw himself on the floor and refused to be comforted. Altogether eight of the children were crying loudly soon after their parents' departure. Bedtime was also an occasion for tears. The two who had not cried earlier screamed when put in a cot and could not be consoled. Some of the others whose initial crying had ceased broke into renewed sobs at bedtime. One little girl, who arrived in the evening and was put straight to bed, insisted on keeping her coat on, clung desperately to her doll, and cried 'at a frightening pitch'. Again and again, having nodded off from sheer fatigue, she awoke screaming for Mummy.
Crying for parents, mainly for mother, was a dominant response especially during the first three days away. Although it decreased thereafter, it was recorded sporadically for each of
-8-
the children for at least the first nine days. It was particularly common at bedtime and during the night. In the early hours of her second day of separation one child, Katie, aged eighteen months, awoke screaming and shouting for Mummy. She remained awake and continued to cry for mother until noon. During the early days away a visit from father led to renewed crying. Another little girl whose father visited her on the third day cried frantically and continuously for twenty minutes after he left.
Searching for mother occurred also and was particularly evident in Katie. After the first week, Katie stopped crying for mother and, instead, seemed content to sit on the nurse's lap watching television. From time to time, however, she demanded that they go upstairs. When asked what she hoped to find there her unhesitating reply was 'Mummy'.
Oriented as they were to their missing parents, these small children were in no mood either to cooperate with the nurses or to accept comfort from them. Initially the children refused to be dressed or undressed, refused to eat, refused the pot. During the first day all but one child, the youngest, refused to be approached, picked up, or comforted. After a day or two resistance abated, but even at the end of two weeks over onethird of the nurses' requests and demands were still being resisted.
Nevertheless, although resistance to the nurses continued to be frequent, the children also began occasionally to seek some sort of reassuring or affectionate response from them. At first these bids for affection were not discriminating but before the end of the second week a few children were beginning to exhibit preferences. For example, one little girl, Gillian, who had refused any dealings with the nurses during the early days, had by the sixth day singled
18
out one nurse and seemed happy sitting on her lap. When the nurse left the room, moreover, Gillian looked longingly at the door. Even so Gillian's feelings for her nurse were not unmixed: when the nurse returned Gillian walked away.
The children's relations with the two research observers were also not unmixed. During the first day most of the children seemed friendly to at least one of the observers. Subsequently they made a point of avoiding the observers by moving away, turning their back, leaving the room, shutting their eyes, or burying their head in a pillow. Especially dramatic were certain occasions when a child broke into a panic the moment
-9-
one of the observers entered the room. On seeing him, or her, a child might scream and run to cling to a nurse. Sometimes a child would show marked relief as soon as an observer had left.
Needless to say, the observers were as unobtrusive as possible. In general their role was not to initiate interaction but to respond in a friendly way whenever a child approached them. Nevertheless, part of the plan was that, fairly late during each of the observation periods, the observer 'actively though cautiously approached the child to see how he would react'. In later chapters of this volume (Chapters 7 and 8) it will be seen that, unwittingly, this plan resulted in conditions that, in combination, are likely to be especially frightening. In some degree, at least, the children's fear of the observers must be attributed to these circumstances.
All but one of the ten children brought with them to the nursery a favourite object from home. For the first three days or so they clung closely to their object and became extremely upset if a nurse, trying to be helpful, happened to take hold of it. Subsequently, however, the children's treatment of their favourite object changed: at one moment they would cling to it, at another throw it away. For instance, one little girl alternated between carrying her rag doll about in her mouth, like a mother cat with a kitten, and flinging it away shouting 'All gone'.
Hostile behaviour, though infrequent, tended to increase during the two weeks of observation. It often took the form of biting another child or ill-treating the favourite object brought from home.
A breakdown in sphincter control was usual. Of the eight children who had attained some degree of control before arriving in the nursery, all but one lost it. The exception, Elizabeth, aged two years eight months, was the oldest of the children.
Although certain kinds of behaviour were common to all or almost all the children, in other ways the children differed. For example, four were constantly active whereas two others preferred to stay in a single spot. A few rocked; others, who seemed constantly on the verge of tears, continually rubbed their eyes.
It will be recalled that four of the children entered the nursery with a sibling, in three cases a four-year-old child and in one case a younger one. As had been expected, the frequency and intensity of the responses typical of children staying in a
-10-
19
residential nursery were much diminished in these children. They cried less and showed fewer outbursts of marked hostility. During the early days especially, siblings constantly sought each other's company, talked and played together. To outsiders they presented a united front, with exclamations such as 'She's not your sister, she's my sister'.
Behaviour at and after Reunion
Inevitably in a situation of this kind the length of time children remain away from home varies. In this study six children were away from twelve to seventeen days; the other four were away for weeks, the periods being seven, ten, twelve, and twenty-one weeks respectively. The ways in which the individual children responded when they returned home differed in many respects; and part of the difference was related to the length of time they had been away, a finding confidently forecast from the results of Robertson's earlier observations.
In this phase of the study two principal lessons learnt from earlier Tavistock studies were applied. The first was to make continuous first-hand observations of how a child responded on first meeting his mother again and during the next few hours. The second was to pay especial attention to a child's response to the visit to his home of an observer he had seen regularly in the nursery. Accordingly, with three research workers available, the following dispositions were made.
One worker, Ilse Westheimer, who had made contact with each of the families before a child went to the nursery, continued contact while the child was away, e. g. by visiting mother in hospital, and was on the spot to make observations when the child was reunited with his parents. Except for one brief visit, she avoided the children while they were in the nursery. In a complementary role, the two observers, Christoph Heinicke and Elizabeth Wolpert, who were responsible for all the observations made on the children in the nursery, took no part in liaising with the families; and they avoided visiting a child at home after he had returned there until a planned visit was made exactly sixteen weeks after his return. 1 (The only exception to this arrangement occurred because Ilse Westheimer was unavailable on one occasion when a child was returning
____________________
1 During the intervening sixteen weeks contact with the home was maintained by Ilse
Westheimer, who also administered doll-play procedures during the sixth and the sixteenth weeks after reunion, and during the equivalent weeks for the control children.
-11-
home, and observation of the return was in this case made by Elizabeth Wolpert. )
In seven cases Ilse Westheimer met mother at the nursery, witnessed the meeting of child and mother, and then drove them home. In three others she met father at the nursery, witnessed the meeting of child and father, and drove the pair home to mother. (In one case they picked mother up on the way, at the hospital in which she had been a patient. )
On meeting mother for the first time after the days or weeks away every one of the ten children showed some degree of detachment. Two seemed not to recognize mother. The other eight turned away or even walked away from her. Most of them either cried or came close to tears; a number alternated between a tearful and an expressionless face.
20
In contrast to these blank, tearful retreats from mother, all but one of the children responded affectionately when they first met father again. Furthermore, five were friendly to Ilse Westheimer as well.
As regards detachment, two findings of earlier studies were clearly confirmed in this one. The first is that detachment is especially characteristic of the way in which a separated child behaves when he meets his mother again, and is much less evident with father; the second is that the duration of a child's detachment from mother correlates highly and significantly with the length of his time away.
In nine cases detachment from mother persisted in some degree throughout almost the first three days of reunion. In five children it was so marked that the mother of each complained, characteristically, that her child treated her as though she were a stranger; none of these children showed any tendency to cling to mother. In the other four, detachment was less pronounced: phases during which they turned away from mother alternated with phases during which they clung to her. Only one child, Elizabeth, who was the oldest and whose separation was among the shortest, was affectionate towards her mother by the end of the first day home. Both Elizabeth and the four who alternated soon showed that they were afraid to be left alone and became far more clinging than they had been before they had gone away.
There is reason to believe that after a very prolonged or repeated separation during the first three years of life detachment can persist indefinitely; the problems to which this gives rise will be discussed in Volume III. After briefer separations
-12-
detachment gives way after a period lasting usually hours or days. It is commonly succeeded by a phase during which a child is markedly ambivalent towards his parents. On the one hand, he is demanding of their presence and cries bitterly if left; on the other, he may become rejecting, hostile, or defiant towards them. Of the ten children studied eight showed a noticeable degree of ambivalence, and in five of them it persisted for not less than twelve weeks. Among determinants of the length of time ambivalence lasts one of the most influential is likely to be the way a mother responds to it.
It will be evident from the descriptions given that when a child returns home after a period away his behaviour presents his parents, especially his mother, with great problems. How a mother responds depends on many factors: for example, the kind of relationship she had with the child before separation and whether she believes that a disturbed and demanding child is better treated with comfort and reassurance or with discipline. A variable to which Westheimer ( 1970) has drawn attention is the way in which a mother's feelings for her child may change in the course of a long separation from him, lasting many weeks or months, during which she does not see him. Warm feelings are apt to cool and family life to become organized on lines that leave no place into which the returning child can fit.
There is abundant evidence that after a child has been away from home in a strange place and in the care of strangers he is liable to be very frightened lest he be taken away again. This was borne in on Robertson during some of his early studies. Children who had been in hospital tended to panic, he found, at the sight of anyone in a white coat or nurse's uniform, and they indicated clearly their fear of being returned to hospital. Several children were apprehensive
21
when Robertson himself visited them at home. They took care to avoid him and, provided they were not in a detached condition, clung close to mother.
In the Heinicke-- Westheimer study one of the two observers who had been present in the nursery visited each child sixteen weeks after his return home. All the children seemed clearly to remember the observer and reacted with strong feeling; all but one made 'a desperate attempt' to avoid the observer. The mothers were much surprised that their child should be so afraid, and affirmed that other strangers who visited the house did not arouse such reactions.
-13-
The case of Josephine, who had been aged two years when she was in the nursery for thirteen days and was now aged two years and four months, illustrates the anxious hostile behaviour seen typically on the occasion of these visits. 1
When CH approached the door of the suburban home, he could hear Josephine making all kinds of excited, joyous noises. As her mother opened the door, however, Josephine at once exclaimed 'No', ran over to the staircase, sat down, ejaculated another 'No', and then picked up the gollywog she had had in the nursery and threw it at the visitor. Mother, observer, and child then went to sit in the garden. Josephine could not sit still, however, and remained excited throughout. She pulled clothes off the line and began to throw them on the grass. Though this seemed deliberately provocative, her mother at first did nothing.
Josephine became more excited, ran about vigorously and repeatedly threw herself in the air and landed on her bottom, but she seemed to ignore any pain she may have inflicted on herself. Later she became aggressive towards her mother, threw herself at her and began to bite, first her mother's arm and then her necklace. Mother was surprised by this behaviour since nothing like it had occurred for some time, and she now restricted it.
Throughout this time Josephine had been afraid of the observer and had assiduously avoided him. When he walked towards her a very worried look came over her face, she cried 'Mummy' and went over to her. Although Josephine continued to react to any approach of the observer by running away, she would try to sneak up to him and to hit him on his back as long as he remained still. Sometimes she ran away and then turned towards him and hit him suddenly. Finally, while the observer sat quietly in one place, Josephine crept close enough to cover him with a small blanket, whereupon she exclaimed 'All gone'. She then uncovered him again.
Mother remarked that the way Josephine had treated the observer was quite different from the way she had treated other strangers, and she expressed surprise that Josephine should so anxiously avoid someone she had not seen for sixteen weeks.
Support for the view that the children were responding in a specially fearful way to the visitor they had known in the nursery comes from a comparison of their reactions with those
____________________
1 The case report is adapted and abridged from that given in Heinicke & Westheimer (
1966). -14-
22
of the matched group of children who had not been away. For the control group, a period of weeks was designated that was treated as the equivalent of the separated children's period of separation, and a succeeding period of weeks was treated as the equivalent of the separated children's period of reunion. During the period equivalent to separation, the control children were visited in their homes by either Christoph Heinicke or Elizabeth Wolpert who administered the same doll-play procedures as they had administered to the children in the nursery. In both samples each child was given either two or three such sessions. In the case of the separated children these were on the third and eleventh days of separation, and, when separation lasted more than three weeks, again a few days before return home. (In the case of all the control children sessions were given on the third, eleventh, and twenty-first days. ) At the end of the sixteenth week of the period equivalent to reunion each of the control children was visited by one of the two workers who had administered the doll-play procedures, just as the separated children were. The response of the control children was quite different from that of the separated children. In every case the control children seemed to recognize the visitor and then approached. 1
At one time it was supposed by critics of our thesis concerning separation that distress seen in a child during a period away from mother, and increased ambivalence and anxiety seen after it, must betoken an unfavourable relationship between child and mother before the event, or reflect perhaps a child's anxiety about his mother's pregnancy or illness. Yet observations of healthy children from thoroughly satisfactory homes,
____________________
1 It was not only at the sixteenth-week visit that the control children behaved differently
towards the observer when compared with the separated children. When the control children were visited in their homes during the period equivalent to separation they treated CH and EW in a far more friendly way than the separated children had done, and nothing resembling panic was seen. It must be remembered, however, that, compared with the separated children, the control children had a very different experience of the observers during the period equivalent to separation. Whereas the separated children were observed six times each week during play in the nursery as well as given two (or three) doll-play sessions, the control children had only the three dollplay sessions. It is possible, therefore, that some part of the difference in the ways in which the two groups of children treated the observers was due to their different experiences of them.
-15-
who are separated from mother for one of many different reasons, show that, whatever contribution other variables may make, when a young child is in a strange place with strange people and with mother absent, protest, despair, and detachment still occur. The only children so far observed in such conditions who appear undisturbed have been those who have never had any figure to whom they can become attached, or who have experienced repeated and prolonged separations and have already become more or less permanently detached. There can be no doubt that a number of variables when combined with absence of mother figure increase the degree of disturbance seen. For example, the more strange the surroundings and the people, or the more painful any medical procedures, the more frightened a child is likely to be and the greater will be his disturbance, both during and after the separation. Again, however, observations of how very differently young children respond to any or all of these conditions when mother is with them make it clear that these conditions of themselves are not sufficient
23
to cause more than transient distress and that, in determining the sequence of protest, despair, and detachment, a key variable is mother's presence or absence.
Bowlby's initial work on attachment was given an airing in the most influential psychoanalytic publications of his day and was then attacked by the most prestigious analytic authorities. That he was virtually exiled by the psychoanalytic community was, ultimately, fortuitous. It led to the establishment around Bowlby's work of a research tradition into the nature of emotional attachment, separation and loss that has generated some of the most remarkable, reliable and provocative empirical data of the past fifty years.
What was it that got Bowlby into so much trouble? He introduced the project as a working out of the implications of observations of the responses to loss young children. It was Bowlby's commitment to the importance of his observations that was the problem. The psychoanalysis of Bowlby's day was an interpretive system that was almost hermetically sealed. Freud had provided the Rosetta Stone for deciphering the underlying structures of mind, and Melanie Klein had extended his psychic excavations to the deepest recesses. Armed with their decoding devices, psychoanalysts felt they could see beneath the superficial, behavioral levels of human interaction to the underlying instinctual conflicts and fantasies that generated their deeper meanings.
Bowlby had been asked by the World Health Organization in 1950 to study and offer advice on the plight of homeless children, real children in the real world. The ravaging impact of maternal deprivations and separations made an enormous impression on
-vii-
him -- to his surprise. It is a tribute to how much Bowlby's work has changed our consciousness that we must strain to imagine why he would have experienced the magnitude of such an impact as a surprise.
In the understanding of mid-twentieth century experts, small children had basic physical needs that required tending; the complex emotional relationships involving a unique sense of interpersonal connection evolved only later. In the conditioning paradigms that dominated American psychology in those days, the caregiver was a "secondary-reinforcer," who became important to the child only by virtue of being associatively linked with physical ministrations. And in the language of the prevailing psychoanalytic theory the mother was, similarly, a "need-gratifying object," whose significance developed gradually through her role in
6
satisfying drive pressures. Separations from mothers shouldn't matter terribly much, as long as the child's needs were taken care of.
But separations did matter, Bowlby discovered. Observations of actual children suggested that maternal deprivation was extremely traumatic. Bowlby s position was a little like that of the wife in the old story who discovers her philandering husband in bed with another woman. "Darling" pleads the husband, "who are you going to believe, me or your lying eyes? " Bowlby decided to believe his eyes.
Any good psychological theory applied cleverly can account for virtually any data. Surely the instinct theory of Bowlby's day could be stretched to accommodate his observations. But for Bowlby, the center of gravity of the data on trauma in real children in the real world did not match up with the center of gravity of Freud's drive theory, privileging impersonal, body- based needs and fantasies. So, as Bowlby announces in this work, a new instinct theory was required, in which the powerful emotional attachment between child and mother is understood not as derivative of more basic processes but as fundamental in itself.
This was no minor addition; it challenged psychoanalytic metapsychology at its core. Bowlby liked to recount experiences formative for him as a student listening to case presentations that emphasized unconscious, instinct-based fantasies at the British Psychoanalytic Society; he remembered one conference at which he felt moved to rise and state emphatically, "But there is such a thing as a bad mother! "
Bowlby was not working completely alone; he had conceptual fellow travelers, especially in the Interpersonal Psychiatry of the American Harry Stack Sullivan and the psychoanalytic contributions of other innovators like W. R. D. Fairbairn, Donald Winni-
-viii-
cott and Hans Loewald. Part of what drew so much fire in Bowlby's direction, however, was his clarity. Sullivan was a tortured, blocked writer. Fairbairn was often tedious and difficult. Winnicott was poetic and elusive. Loewald was extremely subtle and often obscure. Bowlby wrote with lucidity and power.
It was amply apparent from the very beginning and throughout that Bowlby regarded his contributions as a direct challenge to certain basic tenets of Freudian theory. And he had data on children in the real world to back them up. And he identified himself very much as a scientist, offering testable hypotheses. And, his links with other scientists, especially the ethologists of his day, made his position extremely persuasive. For the psychoanalytic establishment of the time, this was simply too much to bear. Bowlby became much more interested than the average psychoanalyst in what actually goes on between people in the real world, and the neighboring discipline of ethology provided powerful explanatory concepts for understanding what he had been observing in children's reactions to separation and loss.
Both Freud and Bowlby were extremely involved with Darwin's contribution (one of Bowlby's last works was a biography of Darwin), but their Darwins were very different. Freud's Darwin was part of the first wave of reaction to the extraordinary implications of the theory of evolution; one of Freud's projects was to work out the implications for human psychology of Darwin's demonstration of the continuity between, so-called lower and so- called higher forms of animal life. Freud's fascination with primitivism and his reliance on
7
bestial metaphors are thematic throughout his writings. And Freud's structural model of the psyche is a re-creation, on a microcosmic level, of Darwin's sweeping account of the evolution of species: lower level, primitive energy of the id is transformed by the reality- oriented ego into higher level, aim-inhibited resources for activities consistent with the cultural values of the superego. Ontogeny recapitulates phylogeny.
Bowlby draws on a different Darwin. Like the Ego psychologist Heinz Hartmann, Bowlby was most interested in what Darwin taught about animal adaptation to environmental conditions and niches. In this second volume of his Attachment and Loss trilogy, Bowlby actually refers to Freud as pre-Darwinian because Freud did not grasp the importance of the principle of "natural selection" in Darwin's theory of the evolution of species. Bowlby, like Darwin, was interested in what animals do to maximize their chances for survival. Whereas Freud's Darwin lent himself to the study of internality and unconscious, primitive states, Bowlby's Darwin lent himself
-ix-
to a behavioral analysis of what small children and mothers actually do with each other. Thus, among the most vivid of Bowlby's contributions is his account of the five component instincts that insure the baby's proximity to the mother--the underpinnings for both healthy attachment and traumatic separations and loss.
One advantage of this behavioral emphasis has been that Bowlby's ideas have been applied, with extraordinary effectiveness, to the empirical research tradition that Mary Ainsworth and Mary Main have done so much to develop. Another advantage is the ease with which Bowlby's observations have been adopted by popular pediatric practice and social policy planners to prioritize the personal, emotional dimensions of childcare. Despite our difficulties as a society in investing the resources necessary for psychologically healthy child- evelopment, we have come to take as axiomatic the principle of "bonding" between parents and children and the importance of emotional warmth and security in caregiving. Bowlby's work was central to this enriched understanding.
Until recently, the disadvantage of Bowlby's behavioral emphasis has been the relative underdevelopment of the psychodynamic dimension within attachment theory, which has made the bridge to other psychoanalytic theorizing more difficult. Bowlby's concept of "working models" had an overly schematic, mechanistic feel to it, which lacked the richness of psychoanalytic investigations of the inner world. But the more recent attachment literature (Fonagy) has taken a more inward turn in exploring the concomitants of secure and insecure attachments in the textures of conscious and unconscious subjectivity. And finally psychoanalysis itself has begun to catch up with bowlby. The recent relational turn in psychoanalysis ( Mitchell and Aron) has made bridges between Bowlby's work and contemporary psychoanalytic thought much more compelling.
Part of what makes Bowlby's early work so inspiring thirty years later is his blend of openness and persistence. His intellectual curiosity seems to have known no bounds, and he continued to draw upon many diverse sources for ideas and conceptual tools, including child- observation, ethology, systems theory, and information-processing. Bowlby also knew from early on that he was onto something very important, and his integrity and perseverance in pursuing what mattered, despite criticism, has benefitted us all.
8
Attachment and Loss has been one of the most influential works of this century. Stephen A. Mitchell New York City, October 1999
-x-
Preface
IN the preface to the first volume of this work I describe the circumstances in which it was begun. Clinical experience of disturbed children, research into their family backgrounds, and an opportunity, in 1950, to read the literature and to discuss problems of mental health with colleagues in several countries led me, in a report commissioned by the World Health Organization, to formulate a principle: 'What is believed to be essential for mental health is that the infant and young child should experience a warm, intimate and continuous relationship with his mother (or permanent mother-substitute) in which both find satisfaction and enjoyment' (Bowlby 1951). To support this conclusion evidence was presented for believing that many forms of psychoneurosis and character disorder are to be attributed either to deprivation of maternal care or to discontinuities in a child's relationship with his mother figure.
Though the contents of the report proved controversial at the time, most of the conclusions are now accepted. What has plainly been missing, however, is an account of the processes through which the many and varied ill effects attributed to maternal deprivation or to discontinuities in the mother-child bond are brought into being. It is this gap that my colleagues and I have since striven to fill. In doing so we have adopted a research strategy that we believe is still too little exploited in the field of psychopathology.
In their day-to-day work, whether with disturbed children, disturbed adults, or disturbed families, clinicians have of necessity to view causal processes backwards, from the disturbance of today back to the events and conditions of yesterday. Though this method has yielded many valuable insights into possible pathogenic events and into the kinds of pathological process to which they appear to give rise, as a research method it has grave limitations. To complement it, a method regularly adopted in other branches of medical research is, having identified a possible pathogen, to study its effects prospectively. If the pathogen has been correctly identified and the studies of its effects in the short and long term are skilfully executed, it then becomes possible to describe the processes set in train by the pathogenic agent and also the ways by which they lead to
-xi-
the various consequent conditions. In such studies attention must be paid not only to the processes set in train by the pathogen but also to the very many conditions, internal and external to the organism, that affect their course. Only then can some grasp be had of the particular processes, conditions, and sequences that lead from a potentially pathogenic occurrence to the particular types of disturbance with which the clinician was in the first place concerned.
In adopting a prospective research strategy my colleagues and I early became deeply impressed by the observations of our colleague, James Robertson, who had recorded, both on
9
paper and on film, how young children in their second and third years of life respond while away from home and cared for instead in a strange place by a succession of unfamiliar people, and also how they respond during and after return home to mother ( Robertson 1952; 1953; Robertson & Bowlby 1952). During the period away, perhaps in residential nursery or hospital ward, a young child is usually acutely distressed for a time and is not easily comforted. After his return home he is likely to be either emotionally detached from his mother or else intensely clinging; as a rule a period of detachment, either brief or long depending mainly on length of separation, precedes a period during which he becomes strongly demanding of his mother's presence. Should a child then come to believe, for any reason, that there is risk of a further separation he is likely to become acutely anxious.
Reflecting on these observations we concluded that 'loss of mother figure, either by itself or in combination with other variables yet to be clearly identified, is capable of generating responses and processes that are of the greatest interest to psychopathology'. Our reason for this belief was that the responses and processes observed seemed to be the same as those found to be active in older individuals who are still disturbed by separations they have suffered in early life. These comprise, on the one hand, a tendency to make intensely strong demands on others and to be anxious and angry when they are not met, a condition common in individuals labelled neurotic; and, on the other, a blockage in the capacity to make deep relationships, such as is present in affectionless and psychopathic personalities.
From the start an important and controversial issue has been the part played in the responses of children to separation from mother by variables other than that of separation per se;
-xii-
these include illness, the strange surroundings in which a child finds himself, the kind of substitute care he receives while away, the kind of relations he has both before and after the event. It is plain that these factors can greatly intensify, or in some cases mitigate, a child's responses. Yet evidence is convincing that presence or absence of mother figure is itself a condition of the greatest significance in determining a child's emotional state. The issue is already discussed in Chapter 2 of the first volume, where a description is given of some of the relevant findings, and is taken up again in the first chapter of this one, where attention is given to the results of a foster-care project undertaken in recent years by James and Joyce Robertson in which they 'sought to create a separation situation from which many of the factors that complicate institutional studies were eliminated; and in which the emotional needs of the children would be met as far as possible by a fully available substitute mother' ( Robertson & Robertson 1971). 1 Study of the Robertsons' findings has led to some modification of views expressed in earlier publications, in which insufficient weight was given to the influence of skilled care from a familiar substitute. In parallel with the empirical studies of my colleagues, I have myself been engaged in studying the theoretical and clinical implications of the data. In particular, I have been trying to sketch a schema able to comprehend data derived from a number of distinct sources:
observations of how young children behave during periods when they are away from mother and after they return home to her;
observations of how older subjects, children and adults, behave during and after a separation from a loved figure, or after a permanent loss;
observations of difficulties found during clinical work with children and adults who, during childhood or adolescence, have either experienced a long separation or a loss or had grounds to fear one; these include various forms of acute or chronic anxiety and
10
depression, and difficulties of every degree in making and maintaining close affectional bonds, whether with parent figures, with members of the opposite sex, or with own children.
____________________
1 In addition to their written report the Robertsons have published a series of films on the
children fostered, particulars of which are given in the list of references at the end of this volume.
-xiii-
First steps towards formulating a theoretical schema were taken in a series of papers published between 1958 and 1963. The present three-volume work1 is a further attempt at a formulation.
Volume 1, Attachment, is devoted to problems originally tackled in the first paper of the series, 'The Nature of the Child's Tie to his Mother' ( 1958b). In order effectively to discuss the empirical data regarding the development of that tie and to formulate a theory to account for it, it proved necessary to discuss first the whole problem of instinctive behaviour and how best to conceptualize it. In doing so I drew heavily on findings and ideas contributed by ethologists and also on ideas derived from control theory.
This, the second volume, deals mainly with problems of separation anxiety and covers ground originally tackled in two further papers of the original series, 'Separation Anxiety' ( 1960a) and 'Separation Anxiety: A Critical Review of the Literature' ( 1961a). Once again, in order to comprehend better the problems before us -- the distress occurring during a separation and the anxiety often evident after it -- it has proved desirable first to discuss a broad range of related phenomena and theory, notably the various forms of behaviour taken to be indicative of fear and the nature of the situations that commonly elicit fear. This discussion occupies Part II of the volume; it provides a background against which are considered, in Part III, the great differences in susceptibility to fear and anxiety that are found when one individual is compared with another. Since many of the data required for the completion of this task are missing, much extrapolation is necessary and the resulting picture is patchy. In some places it can be painted in detail, in others only impressionistically. The aim is to provide clinicians and others with principles on which they can base their actions, and research workers with problems to explore and hypotheses to test.
The third volume, Loss, will deal with problems of grief and mourning and with the defensive processes to which anxiety and loss can give rise. It will comprise a revision and amplification of material first published in the remaining papers of the earlier series -- 'Grief and Mourning in Infancy and Early Childhood' ( 1960b), 'Processes of Mourning' ( 1961b), and 'Pathological Mourning and Childhood Mourning' ( 1963).
____________________
1 In the preface to the first volume I refer only to a second volume. During further work,
however, it has become apparent that a third volume will be required. -xiv-
11
Meanwhile two colleagues, Colin Murray Parkes and Peter Marris, have written books in which they approach problems of loss in a way close to my own. The books are Bereavement by Parkes ( 1972) and Loss and Change by Marris (in press).
In the preface to the first volume it was explained that the frame of reference from which I start is that of psychoanalysis. The reasons are several. The first is that my early thinking on the subject was inspired by psychoanalytic work -- my own and others'. A second is that, despite all its limitations, psychoanalysis and its derivatives remain by far the most used of any present-day approach to psychopathology and psychotherapy. A third and most important is that, whereas many of the central concepts of my schema -- object relations (better termed affectional bonds), separation anxiety, mourning, defence, trauma, sensitive periods in early life -- are the stockin-trade of psychoanalytic thinking, until the last decade or two they have been given scant attention by other behavioural disciplines.
Nevertheless, although the initial frame of reference is that of psychoanalysis, there are many ways in which the theory advanced here differs from the classical theories advanced by Freud and elaborated by his followers. A number of these differences are described already in the first chapter of the earlier volume. Others are referred to throughout the present volume, notably in Chapters 2, 5, and 16.
Note to the Paperback Edition
A paperback edition of this volume has provided the opportunity to include a number of additional notes referring to some later published work. These notes appear on pp. 409 14; the points in the text to which they relate are indicated by a dagger (? ). Details of the additional references are given on pp. 436 -7, following the original references. The index has been considerably expanded.
-xv-
Acknowledgements
IN the first volume of this work I listed the many colleagues and friends without whose help over the years these volumes could not have been written; and it is a great pleasure to express to them all once again my very warmest thanks. My debt to them is deep and lasting.
In the preparation of this volume I am indebted especially for help given by Robert Hinde, Mary Salter Ainsworth, and David Hamburg, each of whom read drafts of all or most of the material and offered a great many valuable criticisms and suggestions. James Robertson scrutinized the first chapter and proposed a number of improvements. Others who have contributed in different ways are Christoph Heinicke, Colin Murray Parkes, and Philip Crockatt. To all of them I am deeply grateful for the time and trouble they have given.
To the preparation of the script my secretary, Dorothy Southern, has again brought her customary care and enthusiasm. Library services have again been provided with unfailing efficiency by Ann Sutherland, and editorial assistance, similarly, by Rosamund Robson. The index has been prepared with great care by Lilian Rubin. To each of them my warmest thanks are due.
12
The many bodies that have supported the research for which I have been responsible at the Tavistock Institute of Human Relations since 1948 are listed in the first volume. Throughout the time that this volume has been in preparation I have been a part-time member of the External Scientific Staff of the Medical Research Council.
For permission to quote from published material, thanks are due to the publishers, authors, and others listed below. Bibliographical details of all the works cited in the text are given in the list of references at the end of the volume.
George Allen & Unwin Ltd, London, and Aldine Publishing Co. , Chicago, in respect of Four Years Old in an Urban Community by J. and E. Newson; Dr I. C. Kaufman, Dr L. A. Rosenblum, and Science in respect of 'Depression in Infant Monkeys Separated from their Mothers' (copyright 1967 by the American Association for the Advancement of Science); Methuen & Co.
-xvii-
Ltd, London, in respect of "'Attachment and Exploratory Behaviour of One-year-olds in a Strange Situation'" by M. D. S. Ainsworth and B. A. Wittig, in Determinants of Infant Behaviour, Vol. 4, edited by B. M. Foss; Dr R. F. Peck and Dr R. J. Havighurst in respect of The Psychology of Character Development, published by John Wiley & Sons, Inc. , New York; University of Chicago Press, Chicago, in respect of The Structure of Scientific Revolutions by T. S. Kuhn; University of London Press Ltd, London, in respect of Truancy by M. J. Tyerman.
Acknowledgement is due also to Tavistock Publications Ltd, London, for permission to include, in Chapter 21 of this volume, material that appears in Support, Innovation, and Autonomy edited by R. Gosling; and to the Journal of Child Psychology and Psychiatry for permission to reproduce, as the basis of Appendix I, a paper that was first published in that journal in 1961.
-xviii-
13
Part I
SECURITY, ANXIETY, AND DISTRESS
-1-
Chapter 1
Prototypes of Human Sorrow
Unhappiness in a child accumulates because he sees no end to the dark tunnel. The thirteen weeks of a term might just as well be thirteen years.
GRAHAM GREENE, A Sort of Life
Responses of young children to separation from mother 1
A generation has now passed since Dorothy Burlingham and Anna Freud recorded their experiences of caring for infants and young children in the setting of a residential nursery. In two modest booklets published during the second world war ( Burlingham & Freud 1942; 1944) they describe the immense problem of providing for young children who are out of mother's care. In particular they emphasize how impossible it is in a nursery setting to provide a child with a substitute figure who can mother him as well as his own mother can. When the Hampstead Nurseries were reorganized so that each nurse could care for her own little group of children they tell how the children became strongly possessive of their nurse and acutely jealous whenever she gave attention to another child: 'Tony (3 1/2) . . . would not allow Sister Mary to use "his" hand for handling other children. Jim (2-3) would burst into tears whenever his "own" nurse left the room. Shirley (4) would become intensely depressed and disturbed when "her" Marion was absent for some reason. '
Why, it may be asked, should these children have become so strongly possessive of their nurse and so deeply distressed whenever she was missing? Was it, as some traditionalists might suppose, that they had been spoiled by having been given too much attention and allowed too much their own way? Or
____________________
1 Although throughout this book the text refers usually to 'mother' and not to 'mother figure',
it is to be understood that in every case reference is to the person who mothers a child and to whom he becomes attached. For most children, of course, this person is also his natural mother.
-3-
was it, by contrast, that since leaving home they had been subjected to too many changes of mother figure and/or had too limited access to whoever in the nursery was acting temporarily as their mother figure? On how we answer these questions turn all our practices of child- rearing.
14
Not only did children in these nurseries become intensely possessive and jealous of their 'own' nurse but they were also unusually prone to become hostile towards her or to reject her, or else to retreat into a state of emotional detachment, as the following records illustrate:
Jim was separated from a very nice and affectionate mother at 17 months and developed well in our nursery. During his stay he formed two strong attachments to two young nurses who successively took care of him. Though he was otherwise a well adjusted, active and companionable child, his behaviour became impossible where these attachments were concerned. He was clinging, over-possessive, unwilling to be left for a minute, and continually demanded something without being able to define in any way what it was he wanted. It was no unusual sight to see Jim lie on the floor sobbing and despairing. These reactions ceased when his favourite nurse was absent even for short periods. He was then quiet and impersonal.
Reggie, who had come to our house as a baby of 5 months, went home to his mother when he was 1 year 8 months, and has been with us ever since his return to the nursery 2 months later. While with us, he formed two passionate relationships to two young nurses who took care of him at different periods. The second attachment was suddenly broken at 2 years 8 months when his 'own' nurse married.
He was completely lost and desperate after her departure, and refused to look at her when she visited him a fortnight later. He turned his head to the other side when she spoke to him, but stared at the door, which had closed behind her, after she had left the room. In the evening in bed he sat up and said: 'My very own Mary-Ann! But I don't like her. '
These observations, made in the pressure of wartime and recorded anecdotally with all too little detail, none the less cast a shaft of light on the nature of many forms of psychiatric disturbance. States of anxiety and depression that occur during
-4-
adult years, and also psychopathic conditions, can, it is held, be linked in a systematic way to the states of anxiety, despair, and detachment described by Burlingham & Freud, and subsequently by others, that are so readily engendered whenever a young child is separated for long from his mother figure, whenever he expects such a separation, and when, as sometimes happens, he loses her altogether. Whereas during later life it is often extremely difficult to trace how a person's disturbed emotional state is related to his experiences, whether they be those of his current life or those of his past, during the early years of childhood the relationship between emotional state and current or recent experience is often crystal clear. In these troubled states of early childhood, it is held, can be discerned the prototype of many a pathological condition of later years.
It is, of course, a commonplace that most children who have had experiences of these kinds recover and resume normal development, or at least they appear to do so. Not infrequently, therefore, doubts are expressed whether the psychological processes described are in reality related so intimately to personality disturbances of later life. Pending much further evidence, these are legitimate doubts. Nevertheless, reasons for holding to the thesis are strong. One is that data from many sources can be arranged and organized into a pattern that is internally consistent and consistent also with current biological theory. Another is that many clinicians and social workers find the resulting schema enables them to understand better the problems with which they are grappling and so to help their patients or clients more effectively.
15
Why some individuals should recover, largely or completely, from experiences of separation and loss while others seem not to is a central question, but one not easily answered. In living creatures variation of response is the rule and its explanation is often hard to fathom. Of all those who contract poliomyelitis less than 1 per cent develop paralysis, and only a fraction of 1 per cent remain crippled. Why one person should respond one way and another another remains obscure. To argue that, because 99 per cent recover, polio is a harmless infection would obviously be absurd. Similarly, in the field under consideration, to argue that because most individuals recover from the effects of a separation or loss these experiences are of no account would be equally absurd.
Nevertheless the problem of differential response remains
-5-
important. Conditions likely to be playing a part can be considered under two main heads: those intrinsic to or closely associated with the separation itself, notably the conditions in which a child is cared for while away from mother;
those present in the child's life over a longer period, notably his relations with parents during the months or years before and after the event.
Here we consider variables in category (a). Those in category (b) are discussed in the later chapters of Part III.
We start by reviewing observations of how children behave when cared for in one of two very different settings. The first is an ordinary residential nursery, in which a child finds himself in a strange place with strange people none of whom is sufficiently available to give him more than very limited mothering. The second is a foster home in which a child receives the fulltime and skilled care of a foster mother with whom he has become in some degree familiar beforehand.
Conditions leading to intense responses
In our early studies children were observed during stays in institutional settings and it was on the basis of these observations that the sequence of responses which we term protest, despair, and detachment came first to be delineated ( Robertson & Bowlby 1952). Since then two further studies have been conducted by colleagues in the Tavistock Child Development Research Unit, the first by Christoph Heinicke ( 1956) and the second by Heinicke & Ilse Westheimer ( 1966). Although in each of these investigations only a handful of children were observed (six in the first and ten in the second), the studies are unique for the care of their design and the amount of systematic observation. Moreover, for each sample of separated children a contrast group was selected and observed: in the first study it was a fairly well- matched group of children observed during their first weeks of attendance at a day nursery; in the second it was a similarly matched group of children observed while living in their own homes. Heinicke & Westheimer treat their data statistically and also describe in some detail the behaviour of individual children.
In the larger investigation ( 1966), work was conducted in three residential nurseries. Arrangements and facilities were fairly similar. In each, a child belonged to a defined group of
16
-6-
children and was cared for mainly by one or two nurses. Ample opportunities were available for free play either in large rooms or outdoors in a garden. Before a child entered the nursery, contact was made with the family by a psychiatric social worker ( Ilse Westheimer), who was also responsible either then or later for collecting full information about the family and the child. Arrival at the nursery was observed; and in the course of his stay a child was observed during free play on six occasions each week. Each of the two observers (one male, Christoph Heinicke, and one female, Elizabeth Wolpert) observed for a period of at least half an hour during each of the three sample periods into which the week was broken (Monday and Tuesday; Wednesday and Thursday; Friday, Saturday, and Sunday). The method used, of categorizing behavioural units in terms of agent, object, relation, mode, and intensity, had been used in the earlier study and had been shown to be reliable.
In addition to the categorized observations of free behaviour, similarly categorized observations were made of every child's behaviour in standardized doll-play sessions; and a number of other records of each child's stay in the nursery were kept.
It was originally intended to select the separated children in accordance with the five criteria used in the first study, namely: (i) that the child had had no previous separations of more than three days, (ii) that he fell within the age-limits of fifteen to thirty months, (iii) that he did not enter the nursery with a sibling, (iv) that he was living with both mother and father at the time the separation occurred, and (v) that there was no evidence that being placed in a nursery indicated a rejection by his parents. Because of the difficulty of obtaining cases, however, these criteria had to be modified to allow greater latitude.
Although most of the children had had either no separations or only very brief ones prior to the one being studied, in one case the length of previous separation was four weeks and in two it was three weeks. The age-range was slightly extended and ran from thirteen to thirty-two months, instead of from fifteen to thirty months. But the most marked departure from the previous criteria was that four of the children entered the residential nursery in the company of a sibling: in three cases this was a four-year-old sibling and in one case the sibling was younger. The remaining two criteria remained unmodified: each of the children was living with both mother and father at the time of separation, and there were no indications that he
-7-
was being rejected by the parents by being placed in the nursery.
The reason that the ten children studied were cared for in a residential nursery was that, in a family emergency, neither relatives nor friends were available to take temporary care of them. In the case of seven families, mother was to be away in hospital having a new baby. In two others mother was to be in hospital for some other form of medical attention. In the tenth case, the family became homeless.
Among much else in their book, Brief Separations ( 1966), Heinicke & Westheimer describe behaviour typical of the ten children during their time in the nursery, and, similarly, behaviour typical of the children after they had returned home. In the paragraphs that follow some of their principal findings are presented. Every one of the patterns reported had also been
17
observed and recorded by Robertson during his earlier, less systematic though more extensive, studies.
Behaviour during Separation
The children arrived at the nursery in the care of one or both parents. Four of them, brought by father, stayed close to him and seemed already subdued and anxious. Some of the others, who had come with mother or both parents, seemed more confident and were ready to explore the new environment. They ventured forth on short or long excursions and then returned.
When the moment came for the parent (s) to depart, crying or screaming was the rule. One child tried to follow her parents, demanding urgently where they were going, and finally had to be pushed back into the room by her mother. Another threw himself on the floor and refused to be comforted. Altogether eight of the children were crying loudly soon after their parents' departure. Bedtime was also an occasion for tears. The two who had not cried earlier screamed when put in a cot and could not be consoled. Some of the others whose initial crying had ceased broke into renewed sobs at bedtime. One little girl, who arrived in the evening and was put straight to bed, insisted on keeping her coat on, clung desperately to her doll, and cried 'at a frightening pitch'. Again and again, having nodded off from sheer fatigue, she awoke screaming for Mummy.
Crying for parents, mainly for mother, was a dominant response especially during the first three days away. Although it decreased thereafter, it was recorded sporadically for each of
-8-
the children for at least the first nine days. It was particularly common at bedtime and during the night. In the early hours of her second day of separation one child, Katie, aged eighteen months, awoke screaming and shouting for Mummy. She remained awake and continued to cry for mother until noon. During the early days away a visit from father led to renewed crying. Another little girl whose father visited her on the third day cried frantically and continuously for twenty minutes after he left.
Searching for mother occurred also and was particularly evident in Katie. After the first week, Katie stopped crying for mother and, instead, seemed content to sit on the nurse's lap watching television. From time to time, however, she demanded that they go upstairs. When asked what she hoped to find there her unhesitating reply was 'Mummy'.
Oriented as they were to their missing parents, these small children were in no mood either to cooperate with the nurses or to accept comfort from them. Initially the children refused to be dressed or undressed, refused to eat, refused the pot. During the first day all but one child, the youngest, refused to be approached, picked up, or comforted. After a day or two resistance abated, but even at the end of two weeks over onethird of the nurses' requests and demands were still being resisted.
Nevertheless, although resistance to the nurses continued to be frequent, the children also began occasionally to seek some sort of reassuring or affectionate response from them. At first these bids for affection were not discriminating but before the end of the second week a few children were beginning to exhibit preferences. For example, one little girl, Gillian, who had refused any dealings with the nurses during the early days, had by the sixth day singled
18
out one nurse and seemed happy sitting on her lap. When the nurse left the room, moreover, Gillian looked longingly at the door. Even so Gillian's feelings for her nurse were not unmixed: when the nurse returned Gillian walked away.
The children's relations with the two research observers were also not unmixed. During the first day most of the children seemed friendly to at least one of the observers. Subsequently they made a point of avoiding the observers by moving away, turning their back, leaving the room, shutting their eyes, or burying their head in a pillow. Especially dramatic were certain occasions when a child broke into a panic the moment
-9-
one of the observers entered the room. On seeing him, or her, a child might scream and run to cling to a nurse. Sometimes a child would show marked relief as soon as an observer had left.
Needless to say, the observers were as unobtrusive as possible. In general their role was not to initiate interaction but to respond in a friendly way whenever a child approached them. Nevertheless, part of the plan was that, fairly late during each of the observation periods, the observer 'actively though cautiously approached the child to see how he would react'. In later chapters of this volume (Chapters 7 and 8) it will be seen that, unwittingly, this plan resulted in conditions that, in combination, are likely to be especially frightening. In some degree, at least, the children's fear of the observers must be attributed to these circumstances.
All but one of the ten children brought with them to the nursery a favourite object from home. For the first three days or so they clung closely to their object and became extremely upset if a nurse, trying to be helpful, happened to take hold of it. Subsequently, however, the children's treatment of their favourite object changed: at one moment they would cling to it, at another throw it away. For instance, one little girl alternated between carrying her rag doll about in her mouth, like a mother cat with a kitten, and flinging it away shouting 'All gone'.
Hostile behaviour, though infrequent, tended to increase during the two weeks of observation. It often took the form of biting another child or ill-treating the favourite object brought from home.
A breakdown in sphincter control was usual. Of the eight children who had attained some degree of control before arriving in the nursery, all but one lost it. The exception, Elizabeth, aged two years eight months, was the oldest of the children.
Although certain kinds of behaviour were common to all or almost all the children, in other ways the children differed. For example, four were constantly active whereas two others preferred to stay in a single spot. A few rocked; others, who seemed constantly on the verge of tears, continually rubbed their eyes.
It will be recalled that four of the children entered the nursery with a sibling, in three cases a four-year-old child and in one case a younger one. As had been expected, the frequency and intensity of the responses typical of children staying in a
-10-
19
residential nursery were much diminished in these children. They cried less and showed fewer outbursts of marked hostility. During the early days especially, siblings constantly sought each other's company, talked and played together. To outsiders they presented a united front, with exclamations such as 'She's not your sister, she's my sister'.
Behaviour at and after Reunion
Inevitably in a situation of this kind the length of time children remain away from home varies. In this study six children were away from twelve to seventeen days; the other four were away for weeks, the periods being seven, ten, twelve, and twenty-one weeks respectively. The ways in which the individual children responded when they returned home differed in many respects; and part of the difference was related to the length of time they had been away, a finding confidently forecast from the results of Robertson's earlier observations.
In this phase of the study two principal lessons learnt from earlier Tavistock studies were applied. The first was to make continuous first-hand observations of how a child responded on first meeting his mother again and during the next few hours. The second was to pay especial attention to a child's response to the visit to his home of an observer he had seen regularly in the nursery. Accordingly, with three research workers available, the following dispositions were made.
One worker, Ilse Westheimer, who had made contact with each of the families before a child went to the nursery, continued contact while the child was away, e. g. by visiting mother in hospital, and was on the spot to make observations when the child was reunited with his parents. Except for one brief visit, she avoided the children while they were in the nursery. In a complementary role, the two observers, Christoph Heinicke and Elizabeth Wolpert, who were responsible for all the observations made on the children in the nursery, took no part in liaising with the families; and they avoided visiting a child at home after he had returned there until a planned visit was made exactly sixteen weeks after his return. 1 (The only exception to this arrangement occurred because Ilse Westheimer was unavailable on one occasion when a child was returning
____________________
1 During the intervening sixteen weeks contact with the home was maintained by Ilse
Westheimer, who also administered doll-play procedures during the sixth and the sixteenth weeks after reunion, and during the equivalent weeks for the control children.
-11-
home, and observation of the return was in this case made by Elizabeth Wolpert. )
In seven cases Ilse Westheimer met mother at the nursery, witnessed the meeting of child and mother, and then drove them home. In three others she met father at the nursery, witnessed the meeting of child and father, and drove the pair home to mother. (In one case they picked mother up on the way, at the hospital in which she had been a patient. )
On meeting mother for the first time after the days or weeks away every one of the ten children showed some degree of detachment. Two seemed not to recognize mother. The other eight turned away or even walked away from her. Most of them either cried or came close to tears; a number alternated between a tearful and an expressionless face.
20
In contrast to these blank, tearful retreats from mother, all but one of the children responded affectionately when they first met father again. Furthermore, five were friendly to Ilse Westheimer as well.
As regards detachment, two findings of earlier studies were clearly confirmed in this one. The first is that detachment is especially characteristic of the way in which a separated child behaves when he meets his mother again, and is much less evident with father; the second is that the duration of a child's detachment from mother correlates highly and significantly with the length of his time away.
In nine cases detachment from mother persisted in some degree throughout almost the first three days of reunion. In five children it was so marked that the mother of each complained, characteristically, that her child treated her as though she were a stranger; none of these children showed any tendency to cling to mother. In the other four, detachment was less pronounced: phases during which they turned away from mother alternated with phases during which they clung to her. Only one child, Elizabeth, who was the oldest and whose separation was among the shortest, was affectionate towards her mother by the end of the first day home. Both Elizabeth and the four who alternated soon showed that they were afraid to be left alone and became far more clinging than they had been before they had gone away.
There is reason to believe that after a very prolonged or repeated separation during the first three years of life detachment can persist indefinitely; the problems to which this gives rise will be discussed in Volume III. After briefer separations
-12-
detachment gives way after a period lasting usually hours or days. It is commonly succeeded by a phase during which a child is markedly ambivalent towards his parents. On the one hand, he is demanding of their presence and cries bitterly if left; on the other, he may become rejecting, hostile, or defiant towards them. Of the ten children studied eight showed a noticeable degree of ambivalence, and in five of them it persisted for not less than twelve weeks. Among determinants of the length of time ambivalence lasts one of the most influential is likely to be the way a mother responds to it.
It will be evident from the descriptions given that when a child returns home after a period away his behaviour presents his parents, especially his mother, with great problems. How a mother responds depends on many factors: for example, the kind of relationship she had with the child before separation and whether she believes that a disturbed and demanding child is better treated with comfort and reassurance or with discipline. A variable to which Westheimer ( 1970) has drawn attention is the way in which a mother's feelings for her child may change in the course of a long separation from him, lasting many weeks or months, during which she does not see him. Warm feelings are apt to cool and family life to become organized on lines that leave no place into which the returning child can fit.
There is abundant evidence that after a child has been away from home in a strange place and in the care of strangers he is liable to be very frightened lest he be taken away again. This was borne in on Robertson during some of his early studies. Children who had been in hospital tended to panic, he found, at the sight of anyone in a white coat or nurse's uniform, and they indicated clearly their fear of being returned to hospital. Several children were apprehensive
21
when Robertson himself visited them at home. They took care to avoid him and, provided they were not in a detached condition, clung close to mother.
In the Heinicke-- Westheimer study one of the two observers who had been present in the nursery visited each child sixteen weeks after his return home. All the children seemed clearly to remember the observer and reacted with strong feeling; all but one made 'a desperate attempt' to avoid the observer. The mothers were much surprised that their child should be so afraid, and affirmed that other strangers who visited the house did not arouse such reactions.
-13-
The case of Josephine, who had been aged two years when she was in the nursery for thirteen days and was now aged two years and four months, illustrates the anxious hostile behaviour seen typically on the occasion of these visits. 1
When CH approached the door of the suburban home, he could hear Josephine making all kinds of excited, joyous noises. As her mother opened the door, however, Josephine at once exclaimed 'No', ran over to the staircase, sat down, ejaculated another 'No', and then picked up the gollywog she had had in the nursery and threw it at the visitor. Mother, observer, and child then went to sit in the garden. Josephine could not sit still, however, and remained excited throughout. She pulled clothes off the line and began to throw them on the grass. Though this seemed deliberately provocative, her mother at first did nothing.
Josephine became more excited, ran about vigorously and repeatedly threw herself in the air and landed on her bottom, but she seemed to ignore any pain she may have inflicted on herself. Later she became aggressive towards her mother, threw herself at her and began to bite, first her mother's arm and then her necklace. Mother was surprised by this behaviour since nothing like it had occurred for some time, and she now restricted it.
Throughout this time Josephine had been afraid of the observer and had assiduously avoided him. When he walked towards her a very worried look came over her face, she cried 'Mummy' and went over to her. Although Josephine continued to react to any approach of the observer by running away, she would try to sneak up to him and to hit him on his back as long as he remained still. Sometimes she ran away and then turned towards him and hit him suddenly. Finally, while the observer sat quietly in one place, Josephine crept close enough to cover him with a small blanket, whereupon she exclaimed 'All gone'. She then uncovered him again.
Mother remarked that the way Josephine had treated the observer was quite different from the way she had treated other strangers, and she expressed surprise that Josephine should so anxiously avoid someone she had not seen for sixteen weeks.
Support for the view that the children were responding in a specially fearful way to the visitor they had known in the nursery comes from a comparison of their reactions with those
____________________
1 The case report is adapted and abridged from that given in Heinicke & Westheimer (
1966). -14-
22
of the matched group of children who had not been away. For the control group, a period of weeks was designated that was treated as the equivalent of the separated children's period of separation, and a succeeding period of weeks was treated as the equivalent of the separated children's period of reunion. During the period equivalent to separation, the control children were visited in their homes by either Christoph Heinicke or Elizabeth Wolpert who administered the same doll-play procedures as they had administered to the children in the nursery. In both samples each child was given either two or three such sessions. In the case of the separated children these were on the third and eleventh days of separation, and, when separation lasted more than three weeks, again a few days before return home. (In the case of all the control children sessions were given on the third, eleventh, and twenty-first days. ) At the end of the sixteenth week of the period equivalent to reunion each of the control children was visited by one of the two workers who had administered the doll-play procedures, just as the separated children were. The response of the control children was quite different from that of the separated children. In every case the control children seemed to recognize the visitor and then approached. 1
At one time it was supposed by critics of our thesis concerning separation that distress seen in a child during a period away from mother, and increased ambivalence and anxiety seen after it, must betoken an unfavourable relationship between child and mother before the event, or reflect perhaps a child's anxiety about his mother's pregnancy or illness. Yet observations of healthy children from thoroughly satisfactory homes,
____________________
1 It was not only at the sixteenth-week visit that the control children behaved differently
towards the observer when compared with the separated children. When the control children were visited in their homes during the period equivalent to separation they treated CH and EW in a far more friendly way than the separated children had done, and nothing resembling panic was seen. It must be remembered, however, that, compared with the separated children, the control children had a very different experience of the observers during the period equivalent to separation. Whereas the separated children were observed six times each week during play in the nursery as well as given two (or three) doll-play sessions, the control children had only the three dollplay sessions. It is possible, therefore, that some part of the difference in the ways in which the two groups of children treated the observers was due to their different experiences of them.
-15-
who are separated from mother for one of many different reasons, show that, whatever contribution other variables may make, when a young child is in a strange place with strange people and with mother absent, protest, despair, and detachment still occur. The only children so far observed in such conditions who appear undisturbed have been those who have never had any figure to whom they can become attached, or who have experienced repeated and prolonged separations and have already become more or less permanently detached. There can be no doubt that a number of variables when combined with absence of mother figure increase the degree of disturbance seen. For example, the more strange the surroundings and the people, or the more painful any medical procedures, the more frightened a child is likely to be and the greater will be his disturbance, both during and after the separation. Again, however, observations of how very differently young children respond to any or all of these conditions when mother is with them make it clear that these conditions of themselves are not sufficient
23
to cause more than transient distress and that, in determining the sequence of protest, despair, and detachment, a key variable is mother's presence or absence.
