Show Summary Details
Page of

(p. 1) John Bowlby and the Tavistock Separation Research Unit 

(p. 1) John Bowlby and the Tavistock Separation Research Unit
(p. 1) John Bowlby and the Tavistock Separation Research Unit

Robbie Duschinsky

Page of

PRINTED FROM OXFORD CLINICAL PSYCHOLOGY ONLINE ( © Oxford University Press, 2022. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Clinical Psychology Online for personal use (for details see Privacy Policy and Legal Notice).

date: 02 July 2022

Biographical sketch

John Bowlby reported that his interest in psychological issues was kindled in 1929 whilst working at a school for troubled children, following studies in natural science at Cambridge. There he had ‘known an adolescent boy who had been thrown out of a public school for repeated stealing. Although socially conforming, he made no friends and seemed emotionally isolated from adults and peers alike. Those in charge attributed his condition to his having never been cared for during his early years by any one motherly person, a result of his illegitimate birth.’1 The previous year Bowlby had read Sigmund Freud’s Introductory Lectures on Psychoanalysis; in light of his experiences at the school, Freud’s ideas ‘came alive for me’, and Bowlby decided to train as a psychoanalyst in London.2 In 1948 he founded the Separation Research Unit at the Tavistock Clinic and appointed James Robertson as an assistant for a study of the effects on young children of hospitalisation with no or minimal visitation from their parents. In 1950 he expanded the research group, appointing Mary Ainsworth as a clinical postdoctoral researcher (Chapter 2). Reflecting on the observations and ideas of this research group, Bowlby developed his novel theory of the nature of the parent–child relationship, of the role of inhibition as a defence against the expectation of rejection, and the form and nature of grief. This theory found its expression first in a series of articles in the International Journal of Psychoanalysis, and then in a trilogy of books: Attachment, Separation, and Loss. Bowlby’s discipline-spanning research made a critical contribution to thinking about human development, offering a model that mitigated important limitations of both behaviourism and psychoanalytic theory. This vital contribution to developmental science, together with his passion for public engagement, has made him one of the most important and influential psychologists of the twentieth century. It has also made him at once one of the most accessible to understand superficially, and one of the most difficult to understand in depth.


In an unpublished article from 1969, Bowlby wrote that ‘an individual holding an idiosyncratic model of the world or of himself is likely to find himself facing the world alone’.3 This expectation that individuality and novel ideas will be met with isolation and rejection had a (p. 2) number of sources.4 Bowlby himself attributed his distrust in others in his professional life in part to the ambush and misunderstanding he received from peers when he first began to present his ideas at the Institute of Psychoanalysis.5 The expectation of misunderstanding and rejection led to sharp distinctions throughout his life in what Bowlby would reveal to different audiences. The result of this was, in a sense, three John Bowlbys, each with somewhat different motivations, and different tones to their writing, reflecting three anticipated audiences.6 The human self is, of course, always a multiple thing, even if there is a certain amount of hierarchical organisation. Nonetheless, consideration of three relatively distinct Bowlbys helps make sense of breakdowns of communication between his writings across different forums and eras, and between the audiences of these texts.

The most well known to the public is the Bowlby available in works written for a general audience during the 1950s. These include the famous 1953 Child Care and the Growth of Love,7 but also his articles in popular women’s magazines8 and presentations to professional organisations.9 Here a tone of sobriety and daring is expressed as authority, even when the evidence was—by his own later admission—‘sketchy’ or ‘inadequate’.10 In these texts from the 1950s, Bowlby argued that a young child needs their mother ‘as an ever-present companion’, providing ‘the provision of constant attention night and day, seven days a week, and 365 days in the year’.11 What he hoped to get across, above all, was that young children should have someone they feel confident turning to when alarmed. Bowlby had been clearly informed by his wife, based on her own experience, that ‘constant attention’ to a child was both an impossible and unhelpful aspiration for mothers.12 And late in life, he acknowledged that he regretted this statement and the implied demand for ever-present care.13 In the 1950s, however, the language of total presence had the advantage for Bowlby of appealing to the obviousness and authority of popular British stereotypes about women and children and nature.14 As has often been remarked, these attitudes were shaped by the context of post-war (p. 3) Britain.15 Bowlby was knowing and explicit that he was drawing on ‘pure prejudice’,16 that in his popular writings he ‘exaggerates everything’.17 Such a strategy helped him get some core ideas heard, even if these were mostly the rind of the views he actually held.

Subsequent attachment researchers have engaged in some whack-a-mole efforts to correct fallacies arising from Bowlby’s populist claims about attachment, for instance his overstrong claims about the influence of early experiences or his polemical claims about the responsibilities of mothers. However, this has been at most partially successful in shaping public perceptions of the implications of attachment theory. Some things are irreversible once they are put out into the world. But additionally, even if they could have been reversed, the second generation of attachment researchers generally did little, especially compared to Bowlby, to speak to a wider public.18 This left Bowlby’s early statements in popular works unqualified, contributing to misalignment between the technical positions of attachment theorists and criticisms levelled at popular representations of the paradigm.

Among commentators who have read his work carefully, there has been growing recognition that there is little evidence that Bowlby believed in the positions with which his name became associated in the public imagination.19 At the same time, it should be recognised that he did little to correct them in his writing for wider publics. In 1954, Donald Winnicott wrote to Bowlby warning of such consequences, and asking Bowlby to issue a public statement (p. 4) revealing his belief that there was, in fact, a shortage of nursery care.20 There is no evidence that Bowlby replied to this letter, and no public statement was issued. Winnicott later reflected on both the advantages and disadvantages of Bowlby’s strategy:

His propaganda for the avoidance of unnecessary breaks in the infant–mother relationship had gone round the world, though I do also feel that the propaganda element necessarily led to a fashion in child care and to the inevitable reactions which follow propaganda.21

A good part of the poor relationship between Bowlby and his critics stemmed from this predicament: in the expectation of misunderstanding and rejection, he had simplified and excluded qualifications from his position in writing populist propaganda in the 1950s for the sake of conveying a message that could travel easily between contexts and down through time; as a consequence, he elicited rejection, and was not able to subsequently make full use of feedback in further amending his views. For example, he made harsh and mocking remarks about his feminist critics, unable or unwilling to take on their concerns, for instance regarding his often underspecified claims about the harms of ‘separations’.22

When asked in 1976 about the discrepancy between statements in his academic writings and those in which he used his position as clinician to speak with authority to the public, Bowlby replied: ‘I would defend this; in different roles one is entitled to speak in different voices.’23 However, at times it was more than a single person speaking in different voices: there seemed to be some cases where communication between the different personae failed or became distorted. For instance, in 1971 he wrote to Michael Rutter ‘I think you misrepresent me and mislead the reader’ when Rutter attributes to Bowlby the oversimplified positions—such as an undifferentiated use of the concept of ‘separation’, or a particular focus on the figure of the mother—that are readily evident from Bowlby’s early popular writings.24 Bowlby appeared genuinely baffled as to why he was being misrepresented, apparently having forgotten his earlier statements. For instance, in an article from 1981, Rutter wrote ‘Bowlby’s argument is that the child’s relationship with mother differs from other relationships specifically with respect to its attachment qualities’, a passage annotated by Bowlby in his private copy of the article ‘Where did he get this idea from?’25 A thoroughly different man is known to those who read Bowlby’s scientific and clinical writings compared to his popular writings.26 This is surely a good part of why his critics and his successors deal essentially (p. 5) with two irreconcilable individuals with the same name, and often seem perplexed by one another. In these scholarly writings, Bowlby presented his ‘idiosyncratic model’ of the world, laying out its logic. Bowlby’s scientific and clinical writings were just as daring as the popular texts, but authority is tempered with sobriety. He was concerned with the particularity of the things he discussed, not the familiar idea of them, even if this demanded dense writing and some loss of ready readability. In general, there is quite a clean division between the lively metaphors and appeal to common sense in Bowlby’s popular writings27 and the somewhat dry prose of his writing for medical audiences and the psychoanalytic community, and in books such as Personality and Mental Illness from 1940 and Attachment, Volume 1 from 1969.28 True, the division between the populariser and the scholar is not complete. For instance, though mild in comparison with the outright homophobia that can be identified in other psychoanalytic texts in the 1960s, Bowlby leans on metaphors of deficit and futility to characterise gay and lesbian sexualities in explaining his new account of human motivation in passages deep inside Attachment, Volume 1.29 However, the fact that even his most hostile critics have never mentioned these passages evidences the division between his popular and scholarly readers, corresponding to his popular and scholarly texts.30

The very content of Bowlby’s claims and his use of familiar words was different between his popular works and his scientific and clinical writings—as was the period of their composition: the vast majority of the popular works were published prior to Attachment, Volume 1 in 1969; the works most read and discussed by subsequent attachment research were published after 1969. In Attachment, Volume 1, Bowlby was absolutely explicit that ‘almost from the first many children have more than one figure towards whom they direct attachment behaviour; these figures are not treated alike; the role of a child’s principal attachment-figure can be filled by others than the natural mother’.31 In dialogue with academic colleagues, Bowlby stated clearly that, for children, ‘provided he sees plenty of his principle figure’, whoever this is, ‘it is an advantage for him to have others as well’.32 He specified that he intended the child’s primary caregiver when he used the term ‘mother’, which was standard practice in the psychoanalytic discourse of his day.33 In his scholarly writings, Bowlby prided himself on a measured and scientific tone largely without feeling, even as he wrote about its causes. He largely avoided polemics, and in his academic writings the characterisation of his London colleagues in the psychoanalytic community was cordial and respectful. Part of Bowlby’s discipline-spanning courage came from his deep acknowledgement of the limited status of (p. 6) theory, as a conjecture to then be appraised against the available evidence.34 In his notes on the philosopher of science Karl Popper, Bowlby wrote: ‘Intelligibility requires the model to be cast in terms mainly of some analogous and better understood set of ideas. Plausibility that it does not affront widely held assumptions.’35 His daring and genre-busting theoretical works, in this sense, were also set up from the start in expectation of being debated, as part of the scientific process.

Bowlby as populariser treated readers as needing to be coaxed to even the most crude points through appeal to their stereotypes and preconceptions. The scientific and clinical commentator held his audience firmly at bay with a carefully orchestrated remoteness. The Bowlby Archive is also a reflection of another kind of reserve and distrust, evident in the hidden array of wayward, profound thoughts that he did not trust to print. Bowlby’s wife described him as ‘completely inarticulate’ when he tried to talk about feelings.36 Yet in his private writings and correspondence available in the Archive, Bowlby is visible as a human: with deep feelings, a personal emotional fragility, frequently thoughtful about his personal limitations. (A volume of writings that reflect this ‘hidden’ Bowlby has recently been published by Duschinsky and White, with the encouragement of the Bowlby family, under the title Trauma and Loss: Key Texts from the John Bowlby Archive.)37 In his notes Bowlby took an interest in monkeys left alone in zoo cages. When angered by visitors, they self-mutilate. Indeed, remarked Bowlby, how like cages are our inhibitions.38 The Archive reveals the engine-room of Bowlby’s thinking behind his inhibitions, in the ordering and transformation of difficult feelings evoked by the world, and by himself, forming the materials for imagination, insight, and propositions. It is this engine that motored his thinking, as he sought sense in the stories he was hearing from veterans from the battlefields of France and the sorrows and symptoms of children.

In the mid-1950s, Bowlby was part of a series of meetings by the ‘Psychobiology of the Child’ study group, organised by Ronald Hargreaves at the World Health Organisation. The discussants included leading researchers from across the world: Jean Piaget, Erik Erikson, Konrad Lorenz, Margaret Mead, and Ludwig von Bertalanffy. A transcription of the discussions was published. In a strangely unguarded moment, the transcript reveals Bowlby reflecting on his own experiences in considering child development and the predicament of patients seen in his clinical work. ‘It seems to me’, Bowlby said, ‘that the main problem with which we are all faced in the process of growing up is that of making a tolerable and compatible synthesis out of a number of manifestly incompatible components.’ Within us are all kinds of positions and elements, including tender feelings and callousness, kindness and cruelty, and ‘these things are literally incompatible’. If an individual wishes to benefit from (p. 7) these different parts, the first thing is that he must ‘own to all these different parts of himself’, rather than pretend that they do not exist. And secondly, he must ‘gradually relate them in some self-balancing unity’.39

There is little unity between Bowlby as populariser, as scholar, and as private thinker through the 1950s and early 1960s. From the late 1960s, the relations between these three personae became more integrated. In this period he came to greater acknowledgement of misunderstandings that had arisen as a result of such divisions, and also to reflect in his theoretical work on the damage that may be done by inhibiting communication between aspects of the self. By the 1970s, with retirement from clinical work, Bowlby further scaled back his activity as a public figure to focus more on his scholarship and his family. He almost exclusively ceased to publish popularizing works after 1969. In discussions with his son, Bowlby explained at the time that he needed to focus on scholarly rather than popular works as he ‘could not afford to be taken as a lightweight’.40 In interview he continued to offer polemical views that played on sexist attitudes towards women.41 But in the final two decades of Bowlby’s life, his central focus was on securing the scholarly standing and clinical relevance of attachment theory—and on pursuing meticulous historical work of the life of Charles Darwin. His desire to influence wider publics was allowed to slide. He made no attempt to update them on developments and qualifications of his position between the 1950s and 1980s, even in late interviews, focusing instead on his classic message of the importance of maternal availability for young children.42

The dominance of the three Bowlbys varied over time, with the populariser quiet after 1969 and the academic scholar in much greater ascendance. However, over the span of Bowlby’s career, a certain generative interplay can be identified. The private thinker fed the scholar experiences and ideas about intimacy and pain, though these were very thoroughly intellectualised by the time they hit the printed page. The private thinker also fed the populariser his passion and his courage—as well as buffering him from the effects of rejection and misunderstanding, which were anticipated in advance. The scholar permitted the private thinker some order and containment, and gave the populariser credibility. And the populariser provided the private thinker both a spur and an outlet, and provided the scholar a wildly increased audience for key claims. There was, however, a personal price to this self-balancing unity. In the World Health Organisation discussions in the mid-1950s, Margaret Mead responded to Bowlby’s remarks by asking a penetrating question: ‘What are the conditions of disassociation in which you do or do not own to this part of your personality?’ Bowlby’s reply (p. 8) is remarkable: ‘I think one could refer to the notions of forgivability and unforgivability.’43 The ultimate question, Bowlby indicates, is whether the parts of oneself can accept one another, can forgive one another, for what may be irreversible.

Bowlby and psychoanalysis

Of special importance for Bowlby’s popularizing, scholarly, and private writings was his training and work as a psychoanalytic clinician. His central ideas emerged from within psychoanalytic theory and clinical work. Not least, Bowlby’s earliest attention to the term ‘attachment’ was in annotations from 1942 on Young Children in War-Time by Dorothy Burlingham and Anna Freud,44 where he marked the term wherever it appeared. Bowlby not only underlined but also highlighted and starred the following passages: ‘Whenever certain essential needs are not fulfilled, lasting psychological malformations will be the consequence; these essential elements are the need for personal attachment, for emotional stability, and for permanency of educational influence’ and ‘It is a known fact that children will cling even to mothers who are continually cross and sometimes cruel to them.’ That there will be some form of ‘attachment of the small child to his mother’ seemed to Burlingham and Freud ‘to a large degree independent of her personal qualities’.45 The latter passage was prominently cited 16 years later in Bowlby’s foundational article on ‘Nature of the child’s tie to his mother’, published in 1958.46

In a late manuscript circulated only to other psychoanalysts,47 Bowlby explained that, as a young psychoanalyst, he envied colleagues (Donald Winnicott is implied) their clinical intuition and grace. And he saw other peers doing good clinical work in spite of the theories to which they subscribed. However, ‘I have not that sort of mind, nor am I strong on intuition. Instead, I tend to apply such theories as I hold in an effort to understand my patient’s problems. This works well when the theories are applicable but can be a big handicap when they are not.’ Without clinical grace or intuition, Bowlby felt that he had no choice but to develop a new theory. Lay and colleagues succinctly identify that this new theory salvaged five aspects of psychoanalytic theory: ‘(1) that infants have a complex social and emotional life, (2) that (p. 9) early experiences can have lifelong implications, (3) that mental representations of early experiences mediate effects on later behavior and development, (4) that defensive processes play a role in affect regulation, and (5) that loss of an attachment figure—at any age—is an emergency and mourning is a process that serves an adaptive affect-regulation function.’48 Yet despite these continuities, Bowlby perceived three major problems with psychoanalytic theory and method: a weak recognition of actual family experiences in shaping child psychology; a mistaken account of the causes of incompatible elements within the human mind; and a conflation of self-preservative actions and sexuality.

Actual experience

The issue of actual experiences was perhaps the earliest of the three central problems Bowlby found with the psychoanalytic theory of his day. Bowlby’s notes from the 1930s and his own retrospective accounts both suggest that Bowlby entered psychoanalytic training in the belief that Freud attributed the emotional problems of his patients to adverse experiences, especially in the context of the family. It was an uncomfortable surprise for him to learn that the psychoanalytic establishment in London, and the work of Melanie Klein in particular, had come to downplay past experience as the basis of symptoms and instead emphasise the role of fantasy.49 In a record of his dreams from the early 1930s, Bowlby reported a dream in which he was imprisoned by the Nazis. Likely with a certain irony, he interpreted the fantasy material as reflecting his anxiety about how to think about fantasy, the unconscious, and sexuality, and about the status of psychoanalytic knowledge in the society around him.50 It was not that Bowlby thought that fantasy was unimportant. But he believed that it had come to be badly overemphasised, at the expense of attention to the effects of biographical experiences.

In private conversations or in interview, when Bowlby was asked to explain the origins of attachment theory, he would frequently refer to clinical cases he saw under the supervision of Melanie Klein. These cases encapsulated key problems Bowlby experienced as a young psychoanalyst. From around 1938, he began treatment of a small boy of about three years old, who was hyperactive, anxious, and aggressive. The boy’s mother was evidently highly troubled, but Melanie Klein insisted that the mother should just sit outside the room. This made Bowlby uncomfortable. He had previous experience of delivering therapy to families within a Child Guidance Clinic, which had convinced him that ‘the problem which as a rule we need to solve is the tension among all the different members of the family’.51 Three or four months into the treatment of the little boy, the mother was admitted as an inpatient to a mental hospital. Bowlby was appalled that Melanie Klein was not interested in the impact of this on the boy, except insofar as it had the practical effect of interrupting the analysis. World (p. 10) War II halted Bowlby’s training as a child analyst: ‘When it was over and I was free to resume my training I could not face doing so however: the absurdity of treating young children and neglecting their parents was too much for me.’52

In staking out the importance of actual experiences for children, Bowlby developed new terminology. Kleinian theorists had used the term ‘internal object’ to refer to the image of the parent held by a child. Bowlby was dissatisfied with this term. He felt that it left the ‘object’ a shimmering abstraction rather than anything concrete. Influenced by the growing interest in cognitive and representational processes of the 1960s,53 Bowlby offered instead the term ‘internal working model’. The term was used by Bowlby to gesture to the cognitive components associated with the attachment behavioural system, as a way to ‘broach the large, difficult, and profound questions of how a child gradually builds up his own “internal world” ’:

Starting, we may suppose, towards the end of his first year, and probably especially actively during his second and third when he acquires the powerful and extraordinary gift of language, a child is busy constructing working models of how the physical world may be expected to behave, how his mother and other significant persons may be expected to behave, how he himself may be expected to behave, and how each interacts with all the others. Within the framework of these working models he evaluates his situation and makes his plans. And within the framework of the working models of his mother and himself he evaluates special aspects of his situation and makes his attachment plans.54

Bowlby’s proposal was that early attachment relationships shape cognitive representations that then inform action. This accounts for continuities between children’s experiences of early care and their later expectations of their caregivers, and subsequently of other people. The idea of the ‘internal working model’ was not, as Bretherton has observed, a fully worked out theory. In part it was a metaphor, one that allowed Bowlby above all to highlight in a general and encompassing way that representations, as ‘models’, should be regarded as tolerably accurate encapsulations of the history of past experiences.55

A clear disadvantage was that the ‘model’ included a huge variety of cognitive content. It is not really clear what the term means, except that it refers to a representation relevant to attachment. At one point or another, most second-generation attachment researchers have stated that Bowlby’s gestural use of the term ‘internal working models’ has made it difficult to generate specific hypotheses using the idea.56 Indeed, part of the ritualised inheritance (p. 11) of Bowlby’s ideas by second-generation attachment researchers was criticism and revision of the concept of the internal working model. It has served as a wide, undefined target onto which each researcher can project his or her own image of what an attachment theory should be. Nonetheless, at the time, there were distinct advantages to the metaphor for Bowlby’s purposes. The idea of the ‘working model’ helped Bowlby to emphasise that expectations about relationships are in constant use, day in and day out, as individuals respond to present demands and forecast future needs.57 It also helped him emphasise that expectations can lead us to search for confirmatory input. Interpretations by later attachment researchers of the term ‘working model’ as suggesting ‘provisional (in the sense of “working” drafts—changeable plans)’58 are a post hoc reconstruction: Bowlby apparently did not intend this connotation.59 His most important emphasis in using the term was to contrast the psychoanalytic emphasis on fantasy with his own emphasis on models as tolerably accurate representations of what actually happened in childhood.

As Peter Fonagy and Morris Eagle have noted, Bowlby’s opposition between the psychoanalytic theory of his day and his own focus on ‘internal working models’ that reflected the role of actual childhood experiences was, for polemical reasons, oversimplified.60 There were analysts like W.R.D. Fairbairn, who held aligned perspectives, as Bowlby himself acknowledged.61 And Anna Freud chided Bowlby for his depiction of her as a secondary-drive theorist, with a position similar to Klein’s; Freud felt that her stance was much closer to Bowlby’s in emphasizing the importance of actual experiences of caregiving for understanding anxious and aggressive behaviours by children.62 Furthermore, it should be highlighted that (p. 12) Bowlby did not assume that all adult recollection reflected historical experience in any simple way. Unpublished clinical cases from the 1930s to the 1960s and published later reflections show Bowlby’s attentiveness in his clinical work to the role of psychological processes that may shape, edit, or distort a patient’s account of their actual experiences.63 In particular, he was interested in the way that early experiences of care could contribute to a tendency to either unrealistically denigrate or idealise attachment figures, and the therapist.64

Nonetheless, it is also true that Bowlby’s emphasis on actual experience represented a shift in clinical technique and in epistemology. In terms of clinical technique, it oriented his discussions with patients and interpretation of their symptoms towards consideration of past experiences.65 And in terms of epistemology, Bowlby regarded psychoanalytic knowledge as compatible with other forms of scientific knowledge and measurement. This distinguished Bowlby from even sympathetic members of the London Institute at the time.66 It was a position that would open the door to both the interdisciplinary synthesis and the focus on concrete observational measures that would lie at the heart of attachment theory. It would also contribute to the backlash of psychoanalysis against attachment theory. In contrast to psychoanalysts who enacted their rebellions more implicitly, the focus on actual experiences at the expense of fantasy has meant that even to this day attachment theory is not formally taught at the British Psychoanalytic Institute.67 This is highly unfortunate. In a longitudinal follow-up study of children treated at the Anna Freud Centre, Target and Fonagy described several cases in which therapists, having been taught that representations of parents represented fantasy, interpreted children’s reports of abuse as such, rather than recognising them as reports of actual events. These children had especially poor long-term outcomes.68

There were, then, several important advantages to Bowlby’s introduction of the idea of ‘internal working models’ in negotiating the relationship between his ideas and psychoanalytic perspectives of the day. However, it should be noted that Bowlby’s use of the concept incorporated several of the same flaws as the psychoanalytic terms, such as ‘object’, it was replacing.69 First, and most significantly, the concept was simply too encompassing, since (p. 13) it seemed to refer to any and all cognitive content about how interactions work within relationships between self and others. For instance, the representational model was situated as both a processor of experience and a repository for experiences, with use of the term often bouncing between these two very different meanings. Second, there was a lack of clarity regarding whether the model was specific to a relationship or general across relationships on the model of early experiences.70 Third, though clearly infants and adults both have expectations about the availability of attachment figures, it is not clear that it makes sense to use the concept of ‘model’ to refer to both the basic non-representational goal-directed expectancies of infants and the elaborate representations held by adults, given all the differences between them. This was something Bowlby himself acknowledged and mused on in his unpublished notes.71 In the 1990s, it became general consensus among attachment researchers that pre-verbal procedural memories of relationship interactions are qualitatively different to the representations informed by memory following verbal development, making an encompassing concept of ‘model’ wholly misleading.72 Each of these issues would leave its legacy for future attachment researchers (Chapters 3 and 5).

Causes of psychological conflict

Bowlby made another important challenge to the dominant psychoanalytic discourses of his day. This was in his thinking about the causes of incompatible elements within the human mind. Bowlby praised Freud for drawing ‘attention to the fact that human beings are organisms which at times are driven by forces within themselves which they cannot easily control. We fall in love, we lose our tempers, we panic, we are possessed by forces which seem alien to ourselves.’73 However, it was especially in his account of the nature of these forces that Bowlby departed from Freud. In the psychoanalytic theory of the 1930s and 1940s, ambivalence was a central theoretical concept, considered an inevitable consequence of the incompatibility of human drives. Most critically, psychoanalytic theory suggested that children feel both love and resentment for their parents. The way that this predicament—the Oedipus complex—is resolved was thought to be of the utmost importance for a child’s later development. During his analytic training, this was a position that Bowlby initially accepted.74

(p. 14) By the 1950s, however, Bowlby had become critical of a tendency in Freud and Klein to presume natural individual differences in ambivalence, rather than examining the social and caregiving context that could intensify such a state. For instance, in the Wolf Man case study, Freud wrote of ‘the patient’s own ambivalence, which he possessed in a high degree of development’.75 Next to this passage, in marginalia from the mid-1950s in his personal copy of Freud’s text, Bowlby wrote: ‘How does Freud explain its genesis?’76 Bowlby’s marginalia on the work of Melanie Klein likewise expresses this concern. In Contributions to Psychoanalysis, Klein had observed that ‘Unpleasant experiences and the lack of enjoyable ones, in the young child, especially lack of happy and close contact with loved people, increase ambivalence, diminish trust and hope and confirm anxieties.’77 However, elsewhere she had argued against the ‘common tendency to over-estimate the importance of unsatisfactory surroundings, in the sense that the internal psychological difficulties, which partly result from the surroundings, are not sufficiently appreciated. It depends, therefore, on the degree of the intrapsychical anxiety, whether or not it will avail much merely to improve the child’s environment.’78 Both passages were underlined and highlighted by Bowlby in his personal copy. In the margins he expressed frustration that according to the second statement, ‘because hatred is primal, the vicious cycle is unavoidable’.79

Bowlby felt that the role of experiences of care in fanning or calming frustration was only inconsistently acknowledged by Klein, who showed insufficient ‘appreciation that hatred itself and therefore intrapsychic conflict can be intensified by bad surroundings’.80 Bowlby anticipated that if a child receives gentle care that acknowledges their feelings, then conflicts can be relatively easily managed. If a child is led to expect rejection or harshness, or experiences long-term separation from their familiar caregiver, then inner conflicts are intensified.81 The Oedipal potential for love and hate of primary caregivers was acknowledged by Bowlby as important, since these emotions would stand in necessary conflict within an intense and important relationship. However, Bowlby regarded this potential as the basis for mental ill health only when a child’s experiences of the parent had been troubled ones, for instance as a result of neglect, a parent’s threats to harm the child, or an inability to keep the child safe.82

In conceptualizing the causes of ambivalence, Bowlby strongly appreciated the emphasis Klein placed on the experience of loss, and how consequential this was for development. However, he was frustrated that Klein predominantly thought about loss as a normal developmental stage within a child–parent relationship. For Klein, every parent would sometimes satisfy a child’s desires and sometimes disappoint them, contributing to a sense of (p. 15) ambivalence and a sense of loss. Her emblem for this predicament was that of the infant wanting to access the mother’s breastmilk, but also feeling envy and destructive feelings towards the breast. Bowlby regarded this account as obscuring the importance of loss as an event that may damage or remove the child–caregiver relationship, rather than solely as taking place inside it. It underplayed the potential causal relationship between experiences of loss and aggression towards caregivers and others.83

In 1956, Bowlby wrote to Klein directly, expressing appreciation for her ‘insistence on the central importance of the conflict of ambivalence for the loved object starting early in life … all my own thinking stems from that’. However, he insisted in this letter that Klein was wrong to regard ambivalence about the maternal breast in particular, and relationships with parents in general, as innate.84 In 1957, he presented ‘The nature of the child’s tie to his mother’ to the British Psychoanalytic Society. A few days prior to the meeting, Bowlby had given a copy of the paper to a colleague. At the meeting, Klein and her followers had clearly received a copy in advance, and were well prepared with an array of hostile remarks. Above all, Bowlby felt that they criticised his emphasis on actual experiences, rather than inherent conflict about feeding, as the origin of feelings of ambivalence within family relationships.85 Bowlby experienced this ambush as incontrovertible evidence for the rest of his career that his ideas would inevitably be met with rejection. As a by-product, it led to a tendency for Bowlby to throw scorn on any proposal, even coming from Ainsworth, that early feeding experiences could be important for emotional development (Chapter 2).


A third central problem for Bowlby with the psychoanalytic theory of innate drives was the status of self-preservatory behaviours. Bowlby’s first exposure to psychoanalysis was reading the Introductory Lectures on Psychoanalysis. In these lectures Freud posited a drive for self-preservation, which would include seeking safety, seeking help, and eating. However, during the period in which Bowlby began training as a psychoanalyst, Freud changed his stance: he posited a sexual drive as primary, which secondarily became altered into a concern with self-preservation through unpleasurable experiences in the world.86 From his earliest writings, Bowlby was not happy with this shift:

It must be realised that Freud equates pleasurable and sexual … Psycho-analytically, the term sexual should be used only to designate pleasurable and should not be used as identical with genital. I think however there is some confusion here. In the act of sucking, Freud distinguishes between the activity of taking nourishment and the pleasure obtained … On (p. 16) the other hand he does not make this distinction in describing the reproductive act … It seems to me that the difficulty has arisen through using the term sexual to equal pleasure in the first instance and then to refer to its reproductive significance in the second.87

Bowlby perceived that Freud’s ‘discovery that symptoms sometimes represented the patient’s sexual activity led him into an over-generalisation … I should like to suggest that symptoms can be excited purely in the interests of avoiding danger-situations.’88 In this, one influence on Bowlby was likely the concern of evolutionary theory with both survival and reproduction, which had been significant in his undergraduate training in natural sciences at Cambridge. Whereas Freud treated self-preservation as an essentially rational tendency adapted to the perception of experienced reality, in his notes in the 1930s Bowlby set out the position that the self-preservative response has its own intrinsic predispositions and preferences: ‘Although the baby neither foresees nor reasons about cold being bad for its health, [this] does not preclude the possibility of its being concerned with an inclination to react to cold with anxiety and to gentle warmth with pleasure. In such a primitive mechanism, I should recognise the gears of the tendency to self-preservation.’89

Further reflections on self-preservation were prompted by Bowlby’s own experiences of fatherhood. John and Ursula Bowlby’s first child, Mary Hamilton, was born in February 1939. Their second, Richard, was born in August 1941; Pia Rose was born in February 1945; finally, Robert was born in April 1948. A traditional upper-class household in this regard, with the additional separations caused by Bowlby’s wartime responsibilities, John had no responsibility for feeding the children. Nonetheless, he recalled observing during the 1940s that his young children would seek him out both for affectionate interaction and when alarmed.90 This ran contrary to the idea, common at the time, that a child’s relationship with their parent developed out of the pleasure that came from feeding (‘cupboard love’). In this seeking of a familiar caregiver when alarmed, Bowlby saw a tendency to self-preservation, which he distinguished from the 1950s onwards from feeding activity in strict terms (indeed overstrict. See Chapter 2).

This position had a number of major practical consequences. One implication was that affection shown to children would not ‘spoil’ them and make them anxious about separation, unable to cope without affection. In fact, Bowlby argued, the opposite would be true. Children who feel confident that they are ‘cherished’ would be less likely to be anxious about separation.91 Another critical implication was that the signs of depression and grief seen by children in the context of long-term separations from their caregivers should, indeed, be recognised as mourning for this relationship.92 From the vantage of the present, this seems like an obvious point. However, it ran against psychoanalytic orthodoxy. Even in 1970, major (p. 17) psychoanalysts were still claiming that ‘mourning as defined by Freud and as observed in the adult is not possible until the detachment from parental figures has taken place in adolescence’.93

Bowlby interpreted childhood loss in terms of adult mourning, and adult bereavement in light of children’s response to separations: ‘Since the evidence makes it clear that at a descriptive level the responses are similar in the two age groups, I believe it to be wiser methodologically to assume that the underlying processes are similar also, and to postulate differences only when there is clear evidence for them’.94 This transposition between childhood and adulthood proved an influential move for later attachment research. It represented the beginnings of a heuristic, or even a method, within subsequent attachment theory and the development of attachment assessments in which adulthood and childhood are interpreted as on analogy with one another. Above all, adult attachment would be interpreted by the second generation of attachment researchers through analogy and extrapolation from individual differences in infant attachment. In the 1980s, this was Mary Main’s approach in developing the Adult Attachment Interview (Chapter 3) and Hazan and Shaver’s in the development of the ‘love quiz’ (Chapter 5). However, more recently, the analogy has been reversed. Researchers such as Fraley and Roisman have deployed ideas about the underlying dimensions of adult attachment as the basis for reimagining individual differences in infant attachment (Chapters 2 and 3).

Following Hinde

Activating and terminating conditions

For Bowlby, the mistaken idea that children cannot mourn their relationships was a consequence of the assumption in psychoanalytic theory that the relationship is secondary, because sexuality is primary. He saw this view as an obstacle to recognition that affectionate bonds are of various kinds and can serve various functions, at an individual level and for humans in general at an evolutionary level. It was in conceptualizing human social responses in the context of research on animal behaviour that Bowlby found strongest support for his thinking about these different kinds and functions of relational behaviour.

In 1951, Bowlby was introduced by Julian Huxley to the work of Lorenz, which revealed an exciting development: that in the study of animal behaviour, careful work was taking place differentiating between behavioural tendencies.95 Niko Tinbergen, Konrad Lorenz, Karl von Frisch, and colleagues, working under the label of ‘ethology’, had developed a revolutionary approach to the study of behaviour. This approach took as its central premise the idea that not just biological structures but also sequences of observable behaviour could be the product of evolution through natural selection, contributing in predictable ways to an (p. 18) individual’s chances of survival and reproduction.96 Ethologists asked four questions of behavioural sequences they observed: How did it develop in the individual? What causes it? What is its function? How did it evolve in the species? When these questions are not adequately distinguished, the ethologists warned, researchers will talk past one another about ‘adaptiveness’, a concept that can refer to any of these levels. They were also worried that the idea of ‘adaptiveness’ can imply the implicit value judgement that a response is warranted or useful for the individual, when in fact this would need to be demonstrated (Chapter 2).97 Indeed, the ethologists argued that responses that support survival and reproduction in general may contribute to highly counterproductive behaviours by an individual, depending on circumstances. Similarly, they observed that the current deployment of a behavioural pattern need not be equated with the function for which it evolved. In fact, an action pattern may become active, reach its predictable outcome, and terminate, all without any direct relationship with the function for which it evolved.

Bowlby was primed to take an interest in ethology by his personal passion for birdwatching. Growing up, Bowlby had quite a formal, distant relationship with his mother; however, she was a passionate and knowledgeable naturalist, with a particular delight in birds. Bowlby learnt that he could retain his mother’s attention by engaging her about his sightings of birds.98 Over time, birdwatching became a firm hobby. However, it should also be emphasised that Bowlby was exhilarated by the quality of the research being done in ethology, especially in contrast to the rudimentary observations of human development available in the 1950s: ‘They were brilliant, first-rate scientists, brilliant observers, and studying family relationships in other species—relationships which were obviously analogous with that of human beings—and doing it so frightfully well. We were fumbling around in the dark; they were already in brilliant sunshine.’99

Critical for Bowlby’s engagement with this literature was his friendship with the ethologist Robert Hinde, who he met in 1954. For the next ten years, Hinde read and gave comments on most of Bowlby’s scholarly writings, and attended Bowlby’s weekly research seminars at the Tavistock.100 Together, Bowlby and Hinde identified several ways in which debates in ethology could advance psychoanalytic theory. The most important point of intersection and difference was in the theory of motivation. Both Freud and Lorenz tended to think about activity as motivated or inhibited by the availability of a somewhat underdefined notion of ‘psychological energy’ or ‘drive’. Hinde addressed this model in his paper on ‘Ethological models and the concept of “drive” ’, published in 1956.101 Here Hinde acknowledged that it was possible in general terms for an organism to become exhausted after various forms of energy expenditure. However, from this he argued that it should not be concluded that all forms of behaviour draw from the same psychological energy or drive. The concepts of ‘energy’ and ‘drive’, Hinde noted, had become nodes within a taken-for-granted theoretical (p. 19) framework; it was time to reappraise the actual behavioural sequences that they were being used to describe and explain. Hinde proposed that, if behaviours are observed closely in comparative perspective across species, different action patterns could be distinguished, along with their activating and terminating conditions. In contrast to the idea of a single reservoir of energy, an advantage of the idea of distinct action patterns was that it was easier to ask the four key ethological questions, considering the development, causation, function, and evolution of behavioural sequences.

Looking back on his career, Bowlby situated Hinde’s 1956 paper on drives as one of the most influential works he ever read. It led directly to the account of motivation at the heart of attachment theory.102 Hinde’s approach lent itself, much more than the psychoanalytic notion of drive, to observational and experimental research to identify the activating and terminating conditions of behavioural responses.103 For instance, Bowlby grudgingly acknowledged, it can happen that babies stop crying when they are exhausted. However, he had no patience with any model of motivation that implied energy or exhaustion. Personally and intellectually, experiences of depletion were generally disavowed by Bowlby, at least until his book on Darwin where they were treated as a symptom of mental illness.104 He emphasised that babies generally stop crying because the terminating conditions for the crying have been met, for instance when they are picked up.105 The specific antecedents, processes, and consequences of the crying response and its termination could be studied empirically much more easily than when crying was theorised in terms of psychological energy or drive.

A second reason that the work of the ethologists appealed to Bowlby was that it distinguished between the infants’ desire for proximity with familiar caregivers and the desire for nutrition. Lorenz had found that exposure to a familiar moving figure was sufficient to elicit a following response in baby geese, a process he termed ‘imprinting’.106 Lorenz’s work was of special interest to Bowlby as it validated his experiences of fatherhood, which suggested that offspring could seek affection and protection from a caregiver based on familiarity, not on the basis of the pleasure of nutrition. The birds studied by Lorenz were not dependent on their parents for food, but could feed themselves by catching insects. Nonetheless, the geese families remained together for at least 12 months.107 Bowlby discussed these findings with Hinde in 1956, who concluded that they indicated the critical importance of the following (p. 20) response for the safety of offspring, since it emerges even before the sequence of behaviours that would allow a gosling to flee from a threat.108 This perspective helped Hinde and Bowlby account for findings by animal researchers such as John Paul Scott and Harry Harlow that non-human animals would continue to show a following response, and indeed might intensify their following response, to a caregiver who was unkind or maltreated them.109

Whereas Lorenz had depicted ‘imprinting’ as a mechanism that was either on or off, Hinde’s experiments showed that the following response, though initially elicited by a wide range of objects and sounds, was elaborated by practice and subsequently organised around the particular objects that elicited the response. In a passage heavily underlined in Bowlby’s personal copy, Hinde made a comparison with the human infant, for whom an incipient following response can often be elicited by even a stranger walking away, a stronger response can be elicited by a sibling, but for whom the full following response will generally occur primarily with a familiar adult caregiver.110 Though the following response might be primed by the nervous system, it required interaction for its elaboration and specification. For Hinde, this implied that not only behaviour but the motivational response itself was integrated and altered on the basis of experience.

An additional source of discussion between Hinde and Bowlby was the observation by ethologists that behavioural sequences might contain components that are unrelated to the function of the whole. This shed light for the vexed problem of ‘infantile sexuality’ that had caused significant consternation and confusion for the psychoanalytic community and its publics, especially during the 1930s when Bowlby was training as an analyst.111 For Bowlby, the work of the ethologists provided an elegant explanation: ‘A great tit whilst still a fledgling may, for example, show isolated fragments of reproductive behaviour—snatches of sub-song, nest-building, and copulatory behaviour—but those fragments appear in contexts quite divorced from the context in which they appear in the adult.’112 Likewise, many of the component sequences of sexual behaviour may well be present in humans already from infancy. And human children can sometimes show these fragments before they become coordinated as the sequences of adult sexual behaviour. However, Bowlby argued, such behaviours do not imply a unitary sexual drive present from birth in all humans.113

At a paper read to the Association for the Study of Animal Behaviour in April 1955, Bowlby praised the ethological research community especially for their combination of close observation and evolutionary theory, which led to conceptually precise distinctions between sequences of child to parent safety-seeking responses, parent to child care-providing (p. 21) responses, and sexuality. Bowlby felt that ‘Freud’s observations that these are apt to become mixed up with each other, although certainly true, does not necessarily mean that there are not three main responses.’114 In a later version of this argument, written up for the British Journal of Medical Psychology in 1957, Bowlby expressed his exhilaration that ethological researchers were finding common aspects to these behaviour patterns across different species, despite all their vast differences. This raised the prospect that human behaviour, too, could be conceptualised in this way.115 Behavioural patterns such as child safety-seeking and following responses, care-providing responses, and sexual behaviour might be discerned that, in human evolutionary history, contributed in predictable ways to individual survival or reproductive fitness. This did not mean that the behaviour as seen in any individual member of a species would be serving this function. However, Bowlby felt that this ethological perspective nonetheless offered the potential to shed light on both typical and atypical forms of development in humans.


One quality of following, care provision, and sexual behaviours was that all three seemed to be directed towards preferred targets, even when suitable alternatives were available. In November 1955, Bowlby wrote to the classicist D.C.H. Rieu to ask for help in describing a quality of the following response: ‘What I am seeking is a term to denote this tendency to restrict these basic social responses to particular individuals.’ He gives the example of the way an infant discerns their mother from among other mothers. Rieu wrote back: ‘I think the word you want is “monotropy”.’116 In retrospect there is an important mismatch here. Bowlby was requesting a word to describe the restriction of following behaviour to ‘particular individuals’—plural. However, Rieu’s term implies the restriction to a single individual—‘mono’—of the tendency to ‘turn to’ the familiar person. It is rather tragic in the fact that Bowlby, whose appeal to ordinary language so often led to misunderstandings of his ideas even as it helped popularise them, unusually sought a new and technical Latinate term on this occasion, but was handed one by Rieu that differed from his request and contributed no less to confusion and polemics.

The term ‘monotropy’ was first used and defined by Bowlby in a 1958 paper as ‘the tendency for instinctual responses to be directed towards a particular individual or group of individuals and not promiscuously towards many’.117 In a footnote, Bowlby clarifies that the meaning is the same as William James’s concept of ‘the law of inhibition of instincts by habits’, which is the observation that ‘when objects of a certain class elicit from an animal a certain sort of reaction, it often happens that the animal becomes partial to the first specimen of the class on which it has reacted’.118 The idea that ‘monotropy’ was intended to mean ‘restriction on the individuals or groups towards whom a response is directed on the basis of (p. 22) experience’ is also evident in Bowlby’s close collaborators, who used the term in this sense.119 With changes in experience, it is possible for the restriction itself to alter, though this becomes more difficult with time; there is no implication that attachment is limited to one person, or that it is fixed regardless of later experiences.120 However, this was certainly not the impression of Bowlby’s readers. The reasons for this seem clear. In Attachment, Volume 1 from 1969, the term appears only once: Bowlby referred his reader back to this 1958 discussion for the meaning of this term, but summarised this earlier account of monotropy briefly—and inaccurately—as ‘the bias of a child to attach himself especially to one figure’.121

It is this latter characterisation that was the public understanding of the term and of Bowlby’s position, supported by the literal implication of the word ‘monotropy’. This implication became a natural rallying point for Bowlby’s critics, as ‘monotropy’ neatly encapsulated within one word the complicity between Bowlby’s dense and scholarly theory and the polemical claims in his popular writings about mothers’ natural responsibilities. Later attachment researchers have worked hard to dispel the idea that attachment theory implies monotropy in this sense,122 but this has been a slow and incomplete process. Some things are, unfortunately, irretrievable. Apparently unaware of his characterisation of the term in Attachment, Volume 1, and of the consequences of his appeals to the natural role of mothers in his popular writings of the 1950s, from the 1970s onwards Bowlby expressed dismay that critics insisted on making him a ‘straw man’, seeking to discredit him through inaccurate characterisation of his discussion of monotropy. He disagreed vehemently with critics who supposed that he introduced the term to imply attachment only to the biological mother, and expressed bafflement in his private notes as to the origin of this view, which he regarded as ‘nonsense’.123

In fact, ‘monotropy’ was intended by Bowlby as a technical term for the way that experience leads a response to become oriented towards particular targets. For Bowlby, ‘monotropy’ was intended to mean a relationship to a particular ‘person or place or thing’ that is personally significant, based on a felt sense of need, and not superficial or interchangeable with other people, places, or things even if they are somewhat similar.124 We cherish certain people, places, or things and not others, and this is part of what it is to cherish. This is in contrast to forms of ‘liking’ where the targets are interchangeable, without hierarchy, but our investment in them is also bloodless. In an unpublished text from 1955 ‘Notes on child attachment and monotropy’, Bowlby wrote: ‘Focusing of instinctive responses on individuals. This is the rule—not the exception and allies to all three basic social responses’—presumably (p. 23) child-to-parent following, parent-to-child care, and sexuality—and ‘it is in the nature of the instinctive response to focus on an individual, though this may be everything from complete to very partial’.125 As such, ‘ “monotropy” is only a special case of discrimination becoming heightened through learning’. The term ‘discrimination’ captures Bowlby’s intention better than monotropy, in fact—and indeed was preferred by Ainsworth.126

In the discussions of the ‘Psychobiology of the Child’ study group at the World Health Organisation in the 1950s, and then later in Separation in 1973, Bowlby argued that the discrimination of particular figures occurs in the case of all the basic behavioural responses. The sexual response, for example, is not evoked indiscriminately in adults by any potential stimulus, but has been trained by experience to become restricted to certain kinds of people or situations.127 There is likewise discrimination in the ‘flight’ response. We have particular people that make us feel safe, who we turn to when worried or scared.128 Bowlby made an argument for ‘the following response’ as another such basic behavioural response, where discrimination occurs regarding its target. Though his focus was on people, Bowlby allowed that objects can also be treated as targets of following and clinging, and sought for ‘the intense sense of reassuring comfort’ that they can, at least in some regards, provide.129 Elsewhere Bowlby added that we also have particular institutions, ideas, and places—such as our nation or our place of prayer130—that are discriminated by the flight and following responses as signifying safety, which we might seek when alarmed.

Tinbergen called attachments to things or locations in the physical environment ‘site attachment’, and Bowlby the ‘personal environment to which we are attached’ in Separation.131 ‘Home’ as an idea and as a location is a clear example. Bowlby acknowledged that though his focus was generally on ‘distress felt and expressed when a person, particularly a child, is separated from his mother figure, distress is felt also on separation from certain familiar objects of other kinds. Attachment to a particular house and environs as home is usual in humans (as it is in animals of other species).’132 Gruneau Brulin and Granqvist have argued (p. 24) that attachment to ‘home’ is simply a secondary effect of its association with family and attachment figures.133 This claim seems overstated: the very concepts of a secure base and safe haven (Chapter 2) are metaphors for territorial movement, away from and back to a base in the context of potential threat.134 Nonetheless, most of Bowlby’s published works focus narrowly on attachment relationships with parents, allowing other attachments such as siblings or to home to slide into the background. Such issues would both have to be rediscovered by later researchers, spurred by the primacy of these concerns in fields such as social work and in ecological approaches.135 Furthermore, Bowlby was unclear regarding what processes exactly led to the discrimination of attachment figures. Main and Fonagy admonished Bowlby for this.136 In the 1990s, they later argued that discrimination and the basis for selective attachments occur especially when someone, or something, is perceived as contingently responsive to us.

Following and attachment

In historical perspective, without access to comparisons with animal behaviour, and to Robert Hinde in particular, it seems likely that there would have been no attachment theory. This was also Bowlby’s own view.137 Though use of the phrase has dropped away since the mid-1990s, in the 1970s and 1980s researchers used ‘ethological attachment theory’ as the official name for the paradigm as a way to ‘commemorate the influence of ethological theory and research on Bowlby’s early thinking’.138 Key strengths of Bowlby’s thinking can be found (p. 25) precisely in the revision of psychoanalytic theory on the basis of ethology: the differentiation of behavioural responses; the care in thinking about value of behaviour for the individual and for the species; recognition that the predisposition for certain responses may come preprogrammed but that their expression requires elaboration in the context of experience and learning; the privilege given to observational methodology; and the specification of the following response in birds and mammals, and in comparison with humans.139 This is a large chunk of attachment theory.

Whereas Bowlby made appeal to the idea of ‘love’ throughout the 1950s in both his popular and scientific works—most importantly in Child Care and the Growth of Love—from 1961 the term was expunged from his vocabulary.140 Hinde’s contribution had made Bowlby acutely aware that the term was too absorptive, that it hid within itself profoundly diverse processes with different causes and consequences. Clinical experience had taught Bowlby time and again that humans can love fiercely even while we fail one another; we can seek comfort even in the absence of love; we can experience affection as claustrophobic and dangerous, as holding us captive.141 It may appear that ‘love’ designates an intense and determinate feeling. In fact, however, the term is more a magnet and placeholder for complex and diverse intensities. Even if appeal to ‘love’ could help Bowlby with intelligibility and plausibility to his readers, Hinde’s precise conceptualisation of motivation had shown the necessity of going beyond evocative euphemism. Michael Rutter later situated the replacement of ‘love’ with ‘attachment’ as one of the most important landmarks in the history of attachment research.142

However, the establishment of attachment theory on the foundation of Hinde’s account of the following response also generated problems. Hinde was discussing a specific behavioural pattern: the following, greeting, and clinging response shown by offspring to their parents. And Bowlby and Ainsworth were both insistent that literal following and bodily clinging were essential referents of the term ‘attachment’.143 However, the concept was fused into the metapsychology of psychoanalytic theory, with its focus on the emotional and symbolic (p. 26) meanings of parent–child relationships.144 On the one hand, this gave the ethological concept of ‘following’ a much deeper emotional resonance. On the other hand, it gave a psychoanalytic model greater behavioural specification. The resulting concept, ‘attachment’, ended up with both narrow and broad meanings (Chapters 2 and 5).145 Narrowly, attachment could mean the following response and related actions that serve to monitor and maintain access to the caregiver; broadly, the same term could mean an emotionally invested relationship, as a symbolic source of comfort and protection. Bowlby shuttled between these distinct meanings, sometimes intending one, sometimes the other, and sometimes both.

Though this certainly contributed to conceptual muddle, the basis for this movement was in Bowlby’s attempts to capture, as best he could, the expression of a cross-species phenomenon in human beings, a species with particular capacities for symbolisation and shared meanings.146 In Bowlby’s unpublished writings and correspondence of the 1950s, his dissatisfaction is visible with this predicament.147 His basic referent for an expression of the attachment system was of infant proximity-seeking in the context of distress as a means of eliciting care. Part of his positioning against psychoanalytic theory was his emphasis on observable interactions. Yet, for example, he wished to maintain that a widow or widower who experiences security and comfort in remembering their spouse is likewise benefiting from a real and ongoing attachment relationship, even if this is now solely at a symbolic level.148 Between these two extremes, he knew that most attachment relationships after toddlerhood would have elements of the basic behavioural system and of its symbolic elaboration: ‘A child separated from his mother comes to crave both for her love and for its accompanying (p. 27) symbols.’149 However, the interrelation between cross-species behavioural system and the symbolic elaboration especially characteristic of humans remained a source of theoretical and terminological problems for him, firmly tangled up within his use of the word ‘attachment’, as well as his difficulties in rooting the mind in the body.150

Another conceptual issue made a potent contribution to this confusion. In Bowlby’s writings, the broad notion of attachment was generally used alongside, specifically, a narrower concept of internal working model—to mean the specific symbolic and affective representations made by humans about attachment figures and their availability, and the value of the self to these attachment figures. By contrast, when Bowlby used the concept of attachment narrowly to mean the specific ethological following response, it was accompanied by a broader concept of internal working model—to mean expectations about the other’s likely availability in response to attachment behaviour.151 This is how infants, puppies, and lambs can have ‘internal working models’ in Attachment, Volume 1 even without the capacity for much representational thought: the concept was being used in this second sense, not the first. This is also why more ethologically oriented attachment researchers, such as Kobak, define internal working models very simply as expectations about attachment figures and how interaction with them will go,152 whereas less ethologically oriented attachment researchers require more elaborate interpretations (Chapter 5).

Already in the 1960s, Hinde was very worried by Bowlby’s misleading use of broad and narrow meanings of the concepts of attachment and internal working model. He felt that ‘behaviour is diverse: an attempt to squeeze it into a system involving only a few explanatory concepts is liable to lead to one of two results—either facts which do not fit will be ignored, or the concepts will be stretched until they become valueless’.153 Hinde tried to warn Bowlby, arguing for greater precision in use of the term ‘attachment’ and what it meant for interpreting observable behaviour. For instance, with early drafts and ideas from Attachment, Volume 1 in hand, Hinde expressed concern with the way Bowlby was conceptualizing the terminating conditions of attachment behaviour, and the idea of internal working models. He wrote to Bowlby in 1967 directly: ‘I think that you sometimes reify the concepts that you are using for explanation as though they were mechanisms.’154

(p. 28) Seeing insufficient change in Bowlby’s stance, in a 1982 chapter Hinde criticised Bowlby on this matter in print. He argued against the implication in Bowlby’s writing that evolution had wired human infants to seek proximity as the sole strategy for achieving the set-goal of the attachment behavioural system. This seemed implausible. Natural selection, Hinde felt, would likely ‘favour individuals with a range of potential styles from which they select appropriately’.155 Whilst direct proximity-seeking might be regarded as the desirable response in many circumstances, Hinde emphasised that survival of infants would be more likely if they could adapt to the conditions of care in which they found themselves. They therefore needed alternative strategies for other conditions. Hinde therefore anticipated that evolution would have given humans a repertoire of ‘conditional strategies’ for responding to caregiving environments where direct proximity-seeking was not possible or effective.156 The availability of conditional strategies could be anticipated to contribute to survival under such circumstances.

The emphasis on proximity as the goal of the infant attachment behavioural system also had discrepancies with Ainsworth’s empirical findings. As Chapter 2 documents, in the 1960s Ainsworth noticed that some infants—those she labelled B1—clearly achieved termination of the attachment behavioural system without proximity. Ainsworth also found other infants—who she labelled C—who did not achieve termination of the attachment behavioural system even when they did gain proximity, and continued to exhibit behaviours seemingly aiming to retain the attention of their caregiver. Citing Ainsworth’s observations as inspiration, across the 1970s and 1980s researchers demonstrated empirically that the caregiver’s physical availability facilitated infant exploration in a novel room less effectively than the caregiver’s emotional and attentional availability.157

Nonetheless, ‘proximity’ remained enshrined in theory as the terminating condition of the attachment behavioural system; it was only late in his career that Bowlby consistently amended his descriptions of the attachment behavioural system in infancy to specify that it could be terminated by the ‘availability’, not simply proximity, of the caregiver.158 However, by this time, a generation of attachment researchers had come of age. For instance (p. 29) ‘disorganised/disoriented attachment’ had already been established by Main and Solomon as behaviour suggesting disruption in an infant’s proximity-seeking (Chapter 3).159 Likewise, severe conceptual difficulties had been experienced as the field attempted to extend Bowlby’s model into adulthood (Chapter 5). For instance, in the 1970s Bob Marvin observed patterns of availability-seeking in otherwise secure-seeming three and four year olds in the Strange Situation that did not rely on proximity. This seemed developmentally expectable. Nonetheless, because proximity was not sought by the children in the Strange Situation, in attempting to square his observations with Bowlby’s stated theory, he was improbably forced to wonder whether these were even attachment relationships.160

As well as his insistence on proximity as the terminating condition for following, a portion of the confusion lay in Bowlby’s use of the term ‘attachment’. As Rutter later observed, the word ‘attachment’ was used by Bowlby ‘to refer to discrete patterns of behaviour (such as proximity-seeking), to a dyadic relationship, to a postulated inbuilt predisposition to develop specific attachments to individuals, and to the hypothesised internal controlling mechanisms for this predisposition’.161 ‘Attachment’ (inbuilt predisposition) means that children show ‘attachment’ (discrete behaviours) within ‘attachments’ (dyadic relationships). How does this happen? Attachment! (a hypothesised internal controlling mechanism). Where the different meanings of the term are not clearly held in mind and distinguished, the result is a recipe for clouded and for circular thinking, and a weak basis for identifying theoretical differences and deviations. Bowlby and subsequent attachment researchers only came to acknowledge the contribution of the sprawling connotations of the term ‘attachment’ to misappropriations of the theoretical tradition in the 1970s. By this time the term, swollen as if infected, had come to define the research paradigm as a whole, and was inextricable.162

Ambiguities about the meaning of ‘attachment’ as the headline concept for Bowlby’s theory have had a widespread legacy. Charles and Alexander expressed concern that Bowlby’s varied and underspecified uses of the term ‘attachment’ have permitted many spurious forms of therapy to claim a basis in the subsequent attachment research paradigm.163 Likewise, where reference to ‘attachment’ appears within social policy, it is common to find circular and confused reasoning, with the term—and the associated research paradigm—invoked for various (p. 30) purposes so long as they can align with the idea of the importance of child–caregiver relationships. So, for instance, in the UK, since the 2010s there have been consistent appeals to Bowlby and the idea of attachment by the political right, who have argued that a policy focus on the early years justifies cuts to other public services, with attachment security presented as an alternative to social security.164 Much the same goes for ‘attachment parenting’ discourses that, in fact, lack an evidence-base or anything but the most selective and strategic relationship with the tradition of attachment research, and that play off cultural stereotypes about motherhood.165 Given the common use of the ‘attachment’ label and the gulf that separates attachment research from these ‘attachment parenting’ discourses, Ross Thompson observed that ‘if you talk to attachment researchers, you will find involuntary wincing—sometimes followed by groaning—when someone brings up attachment parenting’.166


Forty-four thieves

Bowlby’s discussions with Hinde and his emerging model of attachment were fed by and in turn contributed to a concern with the role of long-term separations in childhood for subsequent development. Bowlby regarded himself as predisposed to an interest in major separations by his own childhood experiences, such as being sent to boarding school.167 This interest was then intensified by his early experiences as a clinician. From the late 1930s, Bowlby began work in the London Child Guidance Clinic. Soon after his arrival he saw two cases, one after the other, where the child had been referred for conduct problems. (p. 31) Both had been caught stealing, but more generally were considered rude and disobedient. Bowlby was curious that both of these children had spent nine months in hospital for fever when they were toddlers, during which time they were isolated and separated from their caregivers.

During this period, Bowlby was beginning his training in child psychoanalysis, and reading Klein’s works carefully. He was questioning her description of ‘loss’ only as a normative developmental stage, and her inattention to the possibility of actual separation and loss as consequential experiences for a child. Bowlby found support for his concerns in the fact that, in both of these cases from the Child Guidance Clinic, the parents reported that their children were ‘emotionally remote’ when they returned home from the hospital. Bowlby found it remarkable that ‘these stories were extraordinarily similar. So I generalised from a sample of the two … I found a lot of other cases and the upshot was that I wrote this monograph’, ‘Forty-Four Juvenile Thieves’.168 In this paper, published in 1944, Bowlby examined the records of 44 consecutive cases of children who had been caught stealing, but where the relevant authorities had sent the child to the Child Guidance Clinic rather than to court. Half the referrals had come from schools. These children were compared to unselected control cases from the Clinic with no history of stealing. Fourteen of the 44 thieves he described as ‘affectionless’ types; of these, 12 had experiences of early separation, compared to 10% of the controls. He proposed a general theory that early separation experiences predispose later conduct problems by disrupting the bases of self-worth and capacity for empathy.169 A more specific, and speculative, theory linked early separations specifically to delinquency and criminal behaviour.170

The importance of separation was, as commentators often note, likely made especially salient by Bowlby’s observations of evacuated children during World War II, as part of his work on the Cambridge evacuation survey.171 In unpublished papers written at the time, Bowlby documented the mental health symptoms shown by these children, which (p. 32) included ‘tempers, sullenness, disobedience, stealing, sleeplessness, bed-wetting, timidity, pains of an undefined sort’.172 They also seemed disoriented when they saw their parents again, which Bowlby suspected was a bad sign.173 Bowlby attributed the mental health symptoms of the evacuated children to feelings of being abandoned by their parents. However, he noticed that such symptoms were reduced—though still present—among those children billeted with affectionate foster parents. And he observed that such symptoms were more common when children passed through multiple carers or foster homes with very large carer-to-child ratios.174 Bowlby noted, too, that these symptoms were shown not just by children who missed kind and affectionate parents but also by those children who had cruel or unkind parents. Ultimately, he reflected, ‘it must be remembered that even socially bad homes are nevertheless the child’s only harbours in life. Without a home a child feels lost.’175

An additional factor, rarely mentioned in commentaries on Bowlby, should also be highlighted as contributing to his attention to long-term separations. Examination of the ‘Forty-Four Juvenile Thieves’ case study reveals that some form of child maltreatment appears in around half of the case studies, alongside the separations. Maltreatment is not mentioned in Bowlby’s discussion, however, which focuses solely on separation as the cause of the children’s conduct problems.176 This was done knowingly. Later, in an interview in 1986, Bowlby acknowledged that working with abusive interactions between parents and their children was ‘the run of the mill of what we were doing clinically’ in the Child Guidance Clinic.177 For instance, in an unpublished case history, likely from the early or mid-1960s, Bowlby described the case of Martin, who lived with both parents, but whose experiences of physically abusive punishment from his mother contributed to a tendency to turn to his father for comfort when he was alarmed.178 However, Bowlby reported in the 1986 interview his perception that abusive caregiving could not be turned into a viable research topic with the tools available in the 1940s: ‘My only reason for focusing so exclusively on separation and loss was the fact that one could get comparatively reliable evidence about them whereas during the forties and fifties we had neither manpower nor means of recording less crude variables. As a result of adopting this research strategy, it often appeared that I was unaware of other adverse family events.’179 In particular, he recalled that one thing that his ‘Forty-Four Juvenile Thieves’ paper ‘misses terribly’ was the physical abuse many of the children had also suffered, alongside multiple separations or neglect.180

As well as being difficult to measure with the tools of the time, in the 1940s Bowlby felt that a report on the prevalence of abuse among children referred to the Clinic for conduct (p. 33) problems would raise scandal, and prove unacceptable to public or clinical opinion.181 The overarching issue of the importance of actual childhood experiences for later development would risk getting lost. The initial priority, Bowlby felt, had to be to support the study of child development as a science.182 Bowlby castigated his fellow psychoanalysts in 1943 for their hostility towards scientific methodology, indeed any methodology besides clinical observation. He felt that this stance was rendering psychoanalysis increasingly irrelevant to matters of policy or professional practice: ‘We find ourselves in a rapidly changing world and yet, as a Society, we have done nothing, I repeat nothing, to meet these changes, to influence them or to adapt to them. That is not the reaction of a living organism but of a moribund one. If our Society died of inertia it would only have met the fate that it has invited.’183

To try to be intelligible and credible in context, Bowlby sacrificed reporting the abuse experiences of the children in his clinic in favour of a focus only on documentable separations. This hard decision to focus on loss rather than abuse had to be taken, he believed, so that later researchers, living in a society more frank about family life and with the tools to do more rigorous work, could make child abuse and its prevention the object of scientific measurement. Even with such justifications, it was one of the professional decisions that Bowlby found most difficult to forgive. And in an interview the year he died, Bowlby described that he remained ‘appalled’ at himself and colleagues: ‘Although I was plugging real life events, there were a whole lot of events I didn’t give enough attention to. Sexual abuse is one … Physical abuse is another.’184 In a manuscript circulated only to fellow psychoanalysts late in his career, Bowlby reported that ‘I often shudder’ when he found himself thinking about the accounts by patients of abuse experiences that he ignored.185

The focus on documentable separations was successful in drawing clinical, research, and public attention. However, in certain regards the strategy backfired. In the 1950s Bowlby had something of a tendency to document separations as present or absent, as a crude measure with high reliability. Anyone could check the record and agree that there had been a separation or not. Yet this methodology contributed to a tendency for Bowlby to think and write about separations as merely present or absent, at least until the early 1970s.186 Kinds of separations were not distinguished. Instead in his early work they were all grouped under the label ‘maternal deprivation’, and in his later work discussed as ‘separation’ or ‘lack of continuity’. The concept of ‘maternal deprivation’ had the problem that it implied that the child belonged, specifically, with their mother. Ainsworth repeatedly criticised Bowlby for the fact that at times he used the term ‘maternal deprivation’ when describing children whose predominant experience had actually been cruelty or abuse from their caregiver, with separation (p. 34) only a lesser feature.187 However, an additional problem was that ‘maternal deprivation’, ‘separation’, and ‘lack of continuity’ were all used in an undifferentiated way, and could connote everything from a child sleeping alone in a room, to use of daycare, to child neglect, to institutionalisation in an orphanage.188

In unpublished writings, and in correspondence, Bowlby was quite capable of making these distinctions from the 1940s onwards.189 Writing to Michael Rutter in 1971, he stated: ‘As regards long-term effects of brief experiences, we [Bowlby’s research group] have endeavoured to keep an absolutely open mind. The view I have held for some years is rather like Doll’s view of smoking. Whilst serious effects are found almost always only by prolonged and heavy smoking, even lighter and less prolonged smoking can have adverse effects in some people. Where one draws the line in practice then becomes a matter for private judgement.’190 This more qualified position in private is unfortunately generally absent in Bowlby’s published writings. His tendency to document separation as a binary variable aligned with populist appeal to a stereotyped image, sun-lit and stock, of mother and child as a natural whole. In both his popular and his scholarly writings Bowlby tended to describe separation as a quantitative phenomenon, without qualitative differentiation. In Separation, for instance, the qualifications are removed from the comparison previously offered to Rutter, and the claim is made that more separations of any kind are simply wore: ‘the effects of separation from mother can be likened to the effects of smoking or of radiation. Although the effects of small doses appear negligible, they are cumulative’.191 The crude stance on separations in much of Bowlby’s published work has been considered mistaken and misleading not only by feminist critics and other researchers, but also by his collaborators, followers, and even Bowlby’s own family.192 It has had, additionally, an abiding impact on perceptions of Bowlby and of attachment theory in general.

(p. 35) Prospective studies of separations

The treatment of any form of caregiver–child separation as equivalent was unwarranted, and to an extent polemical. Bowlby had no evidence on which to base such claims. He did have some limited evidence regarding the consequences of relatively long-term separations in childhood. The ‘Forty-Four Juvenile Thieves’ paper looked backwards from clinical cases to find potential pathological causes. However, Bowlby was well aware that this research strategy had significant methodological flaws, not least the problem of confirmation bias for pre-existing theoretical ideas.193 Instead, he advocated a longitudinal methodology, which began by taking children who had experienced a long-term separation and examining its sequelae. Belief in the promise of slow, empirical, longitudinal study of emotional and family life would become a hallmark of attachment research.

In 1946 Bowlby took up a position with clinical and research responsibilities at the Tavistock Clinic in London. There, in 1948, he founded the Separation Research Unit, and hired James Robertson as a research assistant to study children undergoing hospitalisation.194 Mary Boston and Dina Rosenbluth, child psychotherapy trainees at the Tavistock, also contributed to the research.195 At the time, it was common policy to keep parents from visiting their children. For infectious diseases this was partly to prevent the spread of infection. However, it was also general policy for non-infectious diseases as well, since children tended to become distressed and difficult for the hospital staff to manage after their parents had visited. Bowlby had long been interested in the predicament of hospitalised children, which he regarded as a natural experiment to test the effects of separation.196 In the 1930s, he had seen a 15-year-old girl, Joan, in the Child Guidance Clinic:

Joan was examined because of severe headaches, which had begun when she was ten years old. Since no organic basis could be found and psychical factors were obvious she was treated with psychotherapy. After some weeks she described how she suffered from absent periods which were evidently hysterical dream-states. For periods up to two or three hours her head would feel funny and she would be unable to remember the recent past. Things looked different and seemed unreal. If she went to touch a thing she found it was not there. She herself remarked ‘It’s as if I’m in a dream’. The episode ended suddenly and other girls at school would tell her she had been staring curiously, ‘looking beyond usual things’… The episodes began with severe headaches and would come on when she was frightened or anyone was angry. In this case the hysterical headaches and dream-states had begun when she was ten years old after she and her brother had been in hospital with scarlet-fever and diphtheria.197

(p. 36) From 1948 Robertson began making direct observations of a sample of children who had been hospitalised. The first outputs from Robertson’s research were a film and a 1952 paper entitled ‘A two-year-old goes to hospital’, which documented the behaviour of a two-and-a-half-year-old girl, Laura, who was hospitalised for eight days.198 This film helped contribute to recognition of the sorrow major separations can cause children, and to the important movement to change hospital visitation regulations in the 1950s.199

Robertson documented that Laura initially showed a great deal of distress and protest. Her affect then turned towards apparent depression, though accompanied by tic-like stress movements. Three months after she returned home, her mother was away in hospital herself, to have a baby. On her mother’s return, Laura seemed avoidant and somewhat disoriented on reunion:

Laura was very excited and keen to return home. Half an hour later she arrived and the mother could hear her banging on the outside door and calling “Mummy, Mummy”. But when her mother opened the door, Laura looked at her blankly and said, “But I want my Mummy”. For the next two days she did not seem to recognise her mother, and although quite friendly was completely detached. This naturally upset her mother very much. When the father came home an hour or two after Laura’s return, Laura was for a few moments mute toward him but then recovered quickly and was friendly and sure of him.200

Robertson went on to observe 50 cases, although less than half of these were observed intensively, and the context and kind of hospitalisation was highly diverse.201 Reviewing these cases, Robertson and Bowlby came to the conclusion that because the children’s efforts to regain their familiar caregivers had failed both chronically and painfully, they responded by inhibiting their feelings, and especially their yearning for their family. Bowlby termed this ‘detachment’.202 Ainsworth sought, essentially unsuccessfully, to persuade him that the term was misleading, since Bowlby did not intend to suggest that the child was no longer attached to their caregiver, but that an inhibition was observable that blocked intense feeling and its expression.203 Robertson and Bowlby proposed that such inhibition was the cause of the depressed affect, the tic-like tension movements, and also the avoidant or disoriented behaviour on reunion.204 Robertson noted his qualitative impression that children who sustained (p. 37) this inhibition of feeling longer, and who received less comfort from staff during the hospitalisation, were those that showed more psychological disturbance once they returned home. These children displayed more anxiety and aggression, and less affection or help-seeking towards their parents; and these affects were also more likely to occur at odd moments, without apparent reason.205

Bowlby’s confident tone in his academic writing about the impact of the hospitalisations, and crude statements warning about the dangers of separation in his popular writings, led many readers, even sympathetic ones, to assume that he saw a direct relationship between early separations and later behaviour.206 Hazen and Shaver called this, in their assessment, ‘one of the most common misconceptions about attachment theory’.207 This impression was likely reinforced by a strong tendency in Bowlby’s reporting of his own and his analysis or exposition of others’ quantitative findings to neglect attention to moderators and interaction effects. One contributing factor may have been Bowlby’s lack of expertise in using and interpreting statistical tests for interaction effects in empirical research. In fact Bowlby, in his theoretical writing already from the early 1950, explicitly acknowledged two additional factors that need to be included in any causal model, based on Robertson’s observations. The first is that children will have ‘differential susceptibility of an inherited kind’ to the effects of separation.208 The other is that the implications of the separation for a child’s development will be mediated through the consequences of this event and the child’s behaviour for family interactions: ‘Most mothers find their children’s failure to recognise them and to respond, or their outright rejection of them, on reunion extremely hard to bear, and later, when feelings have thawed, their intense possessiveness and whining mummyishness tries patience to the limit. Events such as these set up vicious circles which undoubtedly play an important part in establishing adverse patterns of behaviour in the personality.’209 Rather than treating long-term separation itself as solely directly harmful as implied by Bowlby in popular works and papers reporting empirical findings, in theoretical reflections Bowlby emphasised that harms could be expected to be partially mediated through the vicious cycles that could be expected to be established in family relationships.210

In fact, to Bowlby’s disappointment, the quantitative measures from the Robertson study of the effects of hospitalisation did not supply the evidence he expected of negative (p. 38) consequences of long-term separation.211 The quality of the data was, by his own admission, unusably poor. Ainsworth later commented that Bowlby had been overconfident in his hypothesis, and so did not take sufficient care in choosing his measures: ‘He had expected this to be so conspicuous that he used very crude measures of assessment—teachers’ ratings—and centred in on the IQ. Actually there was nothing in the IQ. The IQs of these children were not lower. And the teachers’ ratings were not sufficiently sensitive really, to turf up any very conspicuous differences.’212 These quantitative results were therefore held back from publication. However, the qualitative descriptions made of the children by Robertson during the hospitalisation and on their return home were very rich and suggestive in their detail.213 Ainsworth later described Robertson’s qualitative descriptions as ‘entrancing’ and ‘deeply impressive’.214 In 1968 she remarked to Bowlby that ‘despite all the lapse of time and subsequent research’ it is ‘still the best’ and most revealing descriptions made by early research on attachment.215 In an oral history conducted in the same period, Ainsworth recalled:

I was tremendously impressed with this material. Jimmy was a social worker at the time but he has since been qualified as an analyst. His observations were the most sensitive direct observations I had ever encountered. I don’t think I have ever encountered anyone who was more perceptive.216

One of the most consequential aspects of Robertson’s work was the distinction between ‘ambivalent’ anxiety/preoccupation and ‘withdrawn’ avoidance. Following separation, these were often—though certainly not always217—observed in sequence, with protest at separation followed by flattened affect over time. Robertson also identified related behaviour upon and following reunion. Anxious preoccupation with the parent was often shown by the formerly hospitalised children.218 And avoidant behaviour was sometimes evident at the moment of reunion with the parents, and could also manifest as withdrawal from the parent in the months after the child returned home.219 In a paper from 1956 reporting on their follow-up study with the hospitalised children, Bowlby, Ainsworth, and colleagues wrote (p. 39) that ‘the personality patterns of children who have experienced long separation tend to fall into one or other of these two opposite classes’: either (i) ‘over-dependent’ and ‘ambivalent’ or (ii) ‘mother-rejecting … having repressed their need for attachment’.220

Ainsworth found it thought-provoking that the two major classes of behaviour appeared to be ‘opposites’. She expressed her fascination with the ‘anxious over-dependence on the one hand, and superficiality and affectionlessness on the other’ in the follow-up study, and the way that this seemed to correspond to ‘the anxious clinging response following reunion after relatively brief or mild separations on the one hand, and the detachment and failure to re-establish affectional relations after long and severe separations on the other’ (Chapter 2).221 Bowlby, Ainsworth, and colleagues discussed this extensively, and they came increasingly to regard the two classes ‘as the prototypes of responses that, when seen in acute and chronic form in older individuals and out of family context, are habitually labelled as psychiatric symptoms. Thus, for example, when the family situation is ignored, it becomes easy to label despair as depression, detachment as psychopathic lack of affect, protest as hysterical.’222

However, these observations of hospitalised children were dismissed as circumstantial evidence by Bowlby’s critics. For instance, the behaviour of hospitalised children could be rejected as unrepresentative of children who were healthy. Experimental research was needed to support the case, but major separations of human children in experimental research was not morally permissible. Bowlby helped Hinde secure the funds to set up a colony of rhesus monkeys at Madingley, Cambridge.223 Rhesus monkeys were chosen on the basis of the intimate relationship between infant and mother during early childhood, which made attachment phenomena especially visible. One of many important findings from Hinde’s Madingley colony of rhesus monkeys from the 1960s was observation of both the ‘classes’ of behaviour following separation that had been identified by Bowlby, Ainsworth, and colleagues in their 1956 paper. Spencer-Booth and Hinde studied the effects of separating rhesus monkey infants from their mothers for six days, too long to be treated as an ordinary separation and so able to offer a parallel to hospitalisation. The researchers observed that ‘the infants’ behaviour during the mothers’ absence can only be described as depressed. They sat in the hunched, passive attitude of a subordinate animal.’224 After reunion, they ‘showed exceptionally intense tantrums when rejected by their mothers, and often flung themselves (p. 40) violently on to their mothers, or sometimes, when the mother had rejected them, on to aunts’:

A further interesting feature was the way in which the infants could change from being relaxed to being very upset and clinging without apparent cause. Thus Tim on Days 11 and 12 was recorded as coming off his mother in an apparently calm fashion, then suddenly panicking and going on her geckering. The most dramatic example was that of Linda, who on Day 16 was playing in a very relaxed fashion for about the first 35 min of the watch, then went on the mother and slept. When she awoke she seemed very upset and, terrified, cringed and would hardly leave her mother.225

Spencer-Booth and Hinde also documented that ‘the deprivation experience accentuated pre-existing individual differences. When the various symptoms were combined to give a distress index, those infants which had been rejected most and/or played the largest role in maintaining proximity before separation were most disturbed after. By contrast, the actual time the infant spent off or at a distance from its mother before separation bore practically no relationship to the disturbance it showed subsequently.’226 Spencer-Booth and Hinde further found that all effects were intensified if they separated the infants from their mothers for two six-day periods or for thirteen days.227 These experimental studies with rhesus monkeys therefore provided further support that depression or detachment could be expected during a major separation from the primary caregiver, and that anxious preoccupation with the caregiver occurred following reunion, especially for those infants who had received the least proactive and nurturing caregiving.


However, the ambivalent and avoidant classes of behaviour were not the only behaviours observed by Robertson that would prove consequential for attachment theory. Robertson identified other anomalous behaviours, especially during separation, but also sometimes on reunion. These were described in detail in an unpublished book of Robertson’s observations, written and re-written with Bowlby over the span of about ten years from the mid-1950s to mid-1960s.228 The central theme that Robertson perceived in these anomalous behaviours was disorientation. Bowlby and Robertson regarded the behaviours as suggesting some disruption of the attachment response. However, what exactly they signified was not clear, and this may have contributed to the decision by Bowlby to hold back the book from publication, alongside Robertson’s gradual departure from the Tavistock in the early 1960s. There was also little academic interest during this period in disoriented behaviour as a mental health symptom.229 Ainsworth regarded the absence of the Robertson and Bowlby book as a major (p. 41) loss to developmental psychology, and a source of personal sadness given the quality of the observations.230 The behaviours only came back into central focus in the work of Mary Main, some decades later (Chapter 3).

One sign of disorientation was that behaviour became unmoored from environmental cues for activation and termination. Robertson noted that some children would swing, seemingly without external prompt, between the ambivalent and avoidant classes of behaviour. For instance: ‘By about the end of her fifth month at home Jacqueline was frequently seeking “baby cuddles” from her mother not only in the evening but also in the daytime. In these brief moments she would curl up and revel in the mutual indulgence she and mother permitted themselves. But between the extremes of being a helpless sensual “baby” and of being detached and independent there was little behaviour of a moderate and quietly affectionate kind.’231

Robertson also documented many signs of disorientation on reunion. He offers a vivid description of Laura’s behaviour during a brief visit by her mother to the hospital:

One of the most striking features about Laura is that, despite being only two years and five months old, she contrived much of the time to control the expression of her grief. Mother announced “I’m going home now”. Laura’s expression was instantly tense and unhappy. Mother insisted “Don’t cry” and pointed an admonitory finger; Laura nodded uncertainly. As her mother left and before she was out of sight Laura turned away with an expression of the deepest misery on her face. The relief of tears, which would have come to most children of that age in that situation, was not available to Laura. As she tried to keep her feelings in check she idly turned the pages of a book, fingered her hair, and both hands fluttered impotently before her face as if she had been momentarily disoriented.232

The withdrawn avoidant class of behaviour was, he thought, especially often accompanied by disorientation in relation to the caregiver. Robertson interpreted both disorientation and avoidance as suggesting the ‘repression of attachment behaviour’.233

A different child showed fear on reunion with her mother. Robertson interpreted this as an effect of disorientation, with the mother misrecognised as a stranger due to repression of the attachment response:

When they reached home and Mary saw her mother for the first time in six weeks, she screamed and refused to go near her. This rejection continued for several days during which she treated her mother so much as though she was a frightening stranger … Only after a week did Mary begin to show a wish to be near her mother. Then her behaviour moved to the other extreme and she followed mother about continuously as if afraid to let her out of her sight. She became increasingly aggressive towards her sisters, and made several vicious attacks on her new baby brother.234

(p. 42) Another behaviour that Robertson described as having a disoriented quality was freezing when alarmed rather than looking for support from a familiar caregiver. For instance, Vicky showed such behaviour soon after returning from the hospital: ‘On the third day her grandmother took her out. Although she had wanted to see the traffic, when she was halfway across a road she suddenly became petrified by fear, and refused to move, while the cars stopped and hooted and people stared.’235

Robertson also described forms of disorientation that led to care-seeking behaviours being directed towards strangers rather than familiar caregivers, what might now be referred to as ‘disinhibited social engagement’. For example, he described the behaviour of Jacqueline: ‘unlike ordinary children of this age, when hurt she did not go to her mother for comfort. Indeed, on an occasion when she bruised a finger quite badly Jacqueline did not turn to her mother but to a visitor who happened to be in the room.’236 Robertson’s impression was that a significant minority of the hospitalised children seemed to show ‘a tendency to make shallow relationships with one and all, and thus to be undiscriminating’ in their attachment behaviour following their hospitalisation.237 He described the behaviour of Stephanie: ‘Although she had not seen me for 18 months, she was immediately friendly and asked for a ride in my car: she was driven around the avenues without showing anxiety, a response mother commented upon. The parents also expressed a related concern—that she was unable to concentrate on anything for more than a very short time. They linked this, with considerable alarm, to the fact that she had been wandering away from home.’

Behavioural systems

The idea of a behavioural system

In the unpublished Robertson and Bowlby book, the authors described many observations relevant to the disruption of the attachment response: ambivalent behaviour, avoidant behaviour, and various forms of what appeared to be disorientation. However, the interpretation of these observations kept running into trouble, across the various drafts and redrafts of the book. ‘Attachment’, ‘detachment’, ‘disorientation’, ‘preoccupation’—all these terms were used at times to refer to observable behaviour and at times to an inferred process at a motivational level. Sometimes they referred to voluntary actions; sometimes to involuntary responses or predispositions. This led the research group in circles at times, as Bowlby became increasingly aware as the 1950s progressed.238 Bowlby’s work on what became attachment theory was driven in part by a desire to develop a conceptual apparatus adequate to the findings of the Separation Research Unit. Indeed, Attachment, Volume 1 began life as a single theoretical chapter for a book reporting on the empirical work with (p. 43) hospitalised children.239 The fundamental concept that Bowlby elaborated was that of a ‘behavioural system’.

The idea of a behavioural system was developed within ethology.240 It referred to ‘systems postulated as controlling a group of behaviour patterns that together serve to achieve a given biological end’.241 For Bowlby, the concept was essentially a metaphor, ‘conceived on the analogy of a physiological system organised homeostatically to ensure that a certain physiological measure, such as body temperature or blood pressure, is held between appropriate limits’.242 If these limits are breached, then steps are taken by the individual to alter the environment or itself to regain them, achieving what Bowlby termed the ‘set-goal’ of the system and re-estabilishing homeostatic equilibrium.243 Each behavioural system can recruit various kinds of resources, most visibly behaviours, to respond flexibly to the environment to achieve the set-goal. The concept of a behavioural system therefore presupposes a hierarchy between an overall goal and the means of achieving this goal, which must be assembled and coordinated together.244

As such, different behaviours commonly associated with the same system may not correlate, or may even negatively correlate with one another, as they are different paths to the same goal. For instance, crying and smiling directed towards the caregiver may be alternative paths to achieving the availability of an attachment figure; freezing and fleeing alternative paths to avoiding danger for the fear system. The concept of behavioural systems allowed Bowlby to clearly distinguish between the behavioural system (the motivation) and behaviour (the observable actions undertaken in the service of the motivation). Admittedly these terms are sufficiently close, and Bowlby was insufficiently careful in his use of the concepts, that this was and has remained an enduring point of confusion for his readers.245 However, the distinction between invisible motivation and observable behaviour is sharp enough when Bowlby is being careful, and it was a consequential distinction for subsequent attachment theory and research.

When activated, behavioural systems such as attachment, caregiving, exploration, fear, and aggression initiate a disposition to try to achieve a set-goal. Where past and present information suggest that the environment will be receptive, the motivation finds expression in behaviour. Indeed, in contexts where the environment may demand the intense activation of a behavioural system, but provides only subtle cues, Bowlby observed in Loss that normal attentional processes may be sharpened and narrowed, as ‘perceptual vigilance’. Such vigilance will lower the threshold at which conditions in the environment stir a slumbering (p. 44) behavioural system to life; it will also increase the intensity of motivation, which will likely be expressed in more elaborate and intense behaviour to achieve the set-goal of the system. This model appeared to Bowlby to account for the clingy behaviour documented by Robertson in children following hospitalisation: the distress of the separation prompted perceptual vigilance regarding proximity to their parents in these children. This led to an intense activation of the attachment system even to minor cues of caregiver unavailability. In turn, the opposite process could explain the ‘detached’ behaviour of the children who had been hospitalised for some time, and on reunion with their caregiver. Where the environment has come to be perceived as unreceptive and unwelcoming for a behavioural system, the activation and expression of a behavioural system can be inhibited.

Ethologists had used the concept of behavioural systems solely to refer to behaviour.246 Bowlby expanded the concept, arguing that a behavioural system also recruits cognitive aspects, most especially expectations based on repeated past experiences and interactions, and perceptions of the present situations.247 In Attachment, Volume 1 he used the term ‘internal working models’ to characterise these cognitive components of a behavioural system. However, too much has often been made of internal working models as an element in Bowlby’s thinking, especially among commentators in the late 1980s and early 1990s who took the internal working model as the ultimate ‘content’ of qualitative differences in attachment, in part based on an interpretation of Main’s work (Chapter 3).248 In fact, Bowlby used the concept relatively sparingly from the 1970s onwards, though he remained keen to emphasise that behavioural systems have cognitive components.249

Furthermore, for Bowlby behavioural systems contain not just behavioural and cognitive components, but also emotional aspects. This aspect of behavioural systems is acknowledged but not detailed in Attachment, Volume 1, which sought only to secure the argument that affects alter our perception of the point that a behaviour system requires activation or termination.250 More than in his books, Bowlby’s key remarks on the emotional components of behavioural system are generally contained in his publications in psychoanalytic journals—which have generally been little read since the publication of Bowlby’s books—as well as in unpublished work. He observed: ‘I conceive overt behaviour to be only one component of a motivational system within the organism, and fantasies, thoughts and affects, conscious and unconscious, to be integral to, and other components of, such systems.’251 Furthermore, some affects, he proposed, accompany the activation of a behavioural system, some to its termination, and others its lack of assuagement. For instance, the sexual system may be accompanied by arousal, its termination by satisfaction and comfort, and its lack of assuagement to yearning and dissatisfaction. The attachment system is accompanied by distress, its termination by comfort, and its lack of assuagement to yearning and dissatisfaction, and eventually despair.252

(p. 45) Groundplans of desire

Another important consequence of the concept of behavioural systems was that it allowed Bowlby to circumvent simplistic notions of nature and nurture. Harkness observed that first among the concerns raised by critics of attachment research since the 1970s has been the question ‘How can attachment be both biologically based and determined by context?’.253 The theory of attachment as a behavioural system was Bowlby’s answer to this question, though it is not an answer that has been widely or well understood. For Bowlby, humans may be primed to develop behavioural systems along certain lines, but the very assemblage of a system depends upon experience. As Bowlby put it in his notes in the 1960s, discussing the attachment behavioural system: ‘What I mean by this is that in an infant’s behavioural equipment there is a built-in bias, genetically determined, that in the ordinarily expectable environment leads to the development of object relations. The built-in bias is there from the first: the actual development of object relations takes time.’254

Attachment behaviour is neither inevitable nor pre-wired, which is why reference by later researchers to the attachment system as a whole using terms such as ‘innate’255 can be misleading. Behavioural systems are a very specific subset of responses that are neither simply innate nor learnt, and that are not primarily expressions of emotion without environmental responsiveness. On the one hand, behavioural systems are neurologically primed at the level of motivation; they represent behaviour towards particular goals that would increase individual survival or reproductive success, at least in the environment within which humans evolved. However, on the other hand, behavioural systems are assembled out of component behaviours and experiences in a way that is flexibly responsive to the environment, with behaviours used interchangeably to achieve the set-goal. Shaver stated that ‘according to Bowlby and Ainsworth, an infant can be viewed as an assembly of behavioural systems’.256 This is essentially incorrect, or at least suggests a broader use of the concept of ‘behavioural system’ by Shaver than that of Bowlby and Ainsworth (Chapter 5). For Bowlby and Ainsworth an infant is much more than the sum of behavioural systems: they observed that infants have some almost entirely innate and pre-given responses, and some entirely learnt responses.257 Additionally, not all behaviours have a set-goal. Some recognisable and patterned responses appeared to Bowlby and Ainsworth to be expressions of overwhelming affect, such as rage and despair, but without being environmentally responsive in a way that recruits component behaviours to achieve a specific end.258

For Bowlby, behavioural systems have specific properties, including that they are neither simply innate nor simply learnt, and that they have a set-goal for their termination. In one (p. 46) of Bowlby’s key metaphors: ‘As in the case of a military operation, the master plan gives only main objective and general strategy; each commander down the hierarchy is then expected to make more detailed plans and to issue more detailed instructions for the execution of his part in the master plan.’259 This phrasing might imply that the activating and terminating points of the attachment system are constant, that only the behaviour is shaped by experience. This was a common misunderstanding of Bowlby’s position among his critics, as Ainsworth identified.260 In fact, Bowlby’s claim was more radical: that the very parameters of the motivational system, though predisposed by evolution, are nonetheless also shaped by our encounters with others, such as experiences of expressing behaviour and how this behaviour is received.261 Not just how, but what and when we desire or do not desire, and what that desire means to us, are inscribed by experience, even if the groundplan of desire may be available from human evolutionary history.

In Attachment, Volume 1, Bowlby proposed that, in humans, ‘most behavioural systems are to some extent environmentally labile—namely, the form they take in an adult turns in some degree on the kind of environment in which that adult is reared. The advantage of this is that the form ultimately to be taken by the system is in some degree left open so that it can, during development, become adapted to the particular environment in which the individual finds himself.’262 The complex cognitive capabilities of humans can also permit this flexibility to extend to individual or collective forms of substitution, so that if a desired set-goal is unattainable, some interim, partial, or symbolic form can nonetheless reduce activation of the behavioural system and offer some portion of the emotional rewards of achieving the set-goal.263

Nonetheless, experiences cumulatively channel this flexibility. ‘Once a human has had experience of reaching a consummatory situation the behaviour that leads to it is likely to become reorganised in terms of a set-goal and a plan hierarchy. That is what appears to occur in sexual behaviour’, where early experiences of relationships in general, and in particular of intimacy lit by intensity, serve to channel the later shape and targets of desire.264 Furthermore, this ontogenetic flexibility is not limitless. A basic quality of behavioural systems is that they require certain kinds of environments for the homeostatic system to operate successfully. A ‘feature of control systems that it is important to note is that they can operate effectively only within certain environmental limits. For example, the system responsible for maintaining body temperature is effective only provided the ambient temperature lies within certain limits. Outside those limits it fails.’265

(p. 47) In his popular and early writings, Bowlby tended to imply that the expected environment for an infant was their primary caregiver, generally the mother. However, over time he came to acknowledge that a context with multiple familiar caregivers simply could not lie outside of the bounds of the attachment system’s responsiveness. His last published work explicitly stated his mature view that the attachment system ‘contributes to the individual’s survival by keeping him or her in touch with one or more caregivers’,266 and he told colleagues that ‘a baby interacting with and forming trusting (secure) relationships (attachments) with a larger number of significant persons will as a child and later as an adult walk more securely in the world’.267 For example, in the environment within which humans evolved it would have been quite possible that grandparents would have been on hand when a baby required care.268 By contrast, he supposed that institutional care, as in orphanages, with a rapid turnover of paid professional caregivers, was likely outside the limits of what the attachment system had evolved to handle. The attachment system is not pre-given, and requires environmental learning and input, wrapping itself around and elaborating itself on the basis of discriminated relationships. Growing up in an institutional context with continual turnover of professional caregivers, it would be much harder for a child to experience any caregiver as familiar enough to entrain the attachment system and its elaboration.

Five behavioural systems

In his scholarly writings, Bowlby gave particular attention to five behavioural systems: the attachment system; the caregiving system; the exploratory system; the fear system; and aggression. This is not an exclusive list—for instance, Bowlby also described a sexual system; he suggests sleep may have qualities of a behavioural system; and he discussed an affiliative system that organises friendly behaviour towards others. However, attachment, caregiving, exploration, fear, and aggression are the most well-characterised instances in his writings, and will be discussed in turn (Table 1.1).

Table 1.1 Behavioural systems and their parameters in infancy in Bowlby’s writing

Behavioural system

Activating condition(s)

Terminating condition

Key component behaviour

Species-level function


Alarming stimuli or the potential for separation from familiar caregivers

The perceived availability of a familiar caregiver

Sucking, clinging, crawling, smiling, distress cries

Protection of self from harm (especially predation)


Perceptions of infant need, especially for retrieval

The perceived cessation of infant signals of need

Retrieval, care, encouragement, protection

Protection and support of others


Novel and/or complex stimuli

Perceived familiarity

Exploratory manipulation



Alarming stimuli or the potential for separation from familiar caregivers

Perceived safety

Startle, wariness, fleeing, hiding, freezing


Aggression (anger of hope)

Frustration of another behavioural system

Satisfaction of the frustrated behavioural system

Threatening, attack


The attachment behavioural system

In Bowlby’s account of behavioural systems, a central place is given to the system’s activating and terminating conditions. Activating conditions trigger a motivation and behaviour to achieve the set-goal; achievement of the set-goal deactivates the motivation and its behaviour. Bowlby theorised that the attachment system in infancy is usually comparatively dormant, tasked primarily with monitoring the caregiver’s whereabouts and checking in from time to time to ensure that a line of retreat to the caregiver remains open. In this dormant state, the system is also engaged with gaining relevant information about the caregiver and (p. 48) the environment, especially information related to the caregiver’s responses to the child’s behaviour and to potential cues for danger. When a child is alarmed, the attachment system will prompt attempts to gain proximity with the caregiver.

From the primate researchers Harry Harlow and Robert Zimmermann, Bowlby took the phrase ‘haven of safety’ to refer to the way that an infant’s alarm and motivation to seek their caregiver was terminated once they achieved proximity with the caregiver.269 However, Bowlby was keen to make clear that the extent of proximity required by the attachment system was flexible, not a biological given. The set-goal of the attachment system may, indeed, abruptly change to specify the degree of proximity more narrowly or more loosely, and this will bring about attachment behaviours of different forms and intensities. Even the same environmental cue, such as a loud and sudden noise, may elicit only a look to the caregiver from an infant when accessibility feels ready and sure, but may elicit swift approach and clinging when the set-goal of the attachment system has been calibrated at full physical contact.

In Attachment, Volume 1 Bowlby highlighted three conditions as of particular importance to the calibration of the set-goal. A first was past experience: for children who have come to expect that their caregiver might not be accessible when the attachment system is activated, a lower threshold would be set for its activation and a higher threshold for termination. By contrast, for children confident in the accessibility of their caregiver, physical proximity-seeking may only occur at a high threshold, and can be more readily terminated.270 A second factor that would influence these thresholds, Bowlby suggested, was the extent of current (p. 49) sources of threat. In environments with alarming stimuli, or cues that an unexpected or major separation might occur, the parameters of the attachment system might well be altered.271 Finally, a third factor that Bowlby thought would especially impact the attachment system’s parameters was human cognitive development. With maturation beyond infancy, we have at our disposal more sophisticated strategies for maintaining contact, including verbal communication and mental representation.

As Ainsworth’s student Bob Marvin documented in his 1972 doctoral thesis, maturation raises the threshold for the activation of the attachment system, and also the intensity of the attachment behaviours that are likely to be expressed.272 Infants may cry and seek to cling to their attachment figure following an unexpected separation. By contrast, preschoolers, with greater tools at their disposal for understanding and achieving the caregiver’s accessibility, may simply offer a look and smile of greeting. Ultimately the set-goal of the attachment system is not a particular inner state or set of external conditions, Bowlby argued, but a point at their intersection: ‘a certain sort of relationship with another specified individual’.273 Until his final writings, Bowlby generally claimed that this relationship was physical accessibility. Later, taking heed of Ainsworth’s observations and arguments, he acknowledged that the ‘certain sort of relationship’ is, in infancy and across childhood, both the physical and psychological ‘availability’ of the attachment figure.

Though this was a point generally missed by his critics, and could have been better highlighted, Bowlby regarded few, if any, of the component behavioural sequences of the attachment system as exclusive to it.274 Behaviours can and often do serve more than one system. The attachment system takes as key early components behavioural sequences such as sucking, clinging, crawling, and smiling.275 However, sucking is also important for what Bowlby called the ‘eating’ or ‘feeding’ behavioural system276 and what Ainsworth termed the ‘food-seeking system’ (Chapter 2). The shared components of the following and the food-seeking systems were critical for the conflation of the two in psychoanalytic theory, under the notion of the ‘oral stage’.277 Smiling is not only used for activating caregiver responses, but is more generally critical for the affiliative system, and can be co-opted by the fear system as a form of placation. Developmental maturation also allows behaviours usually characteristic of other systems to be recruited by the attachment system if its usual repertoire for seeking care has not been successful. As discussed below in the section ‘The caregiving behavioural (p. 50) system’, Bowlby’s particular interest was in the recruitment of the caregiving system in the service of the attachment system. However, he also noted how sexuality could be used in the service of the attachment system when other forms of care-seeking have failed, and ‘one can regard attempted suicide as an aberrant form of care-eliciting behaviour resorted to by only those who have had very unstable relationships in the past, and who have learned that more normal types of care-eliciting behaviour fail to work. As an aberrant form of attachment behaviour, of course, it is not far removed from total despair.’278 Behaviours can also serve more than one system at the same time: for instance, Bretherton and Ainsworth argued that coy and submissive behaviour is coordinated by the fear and affiliative systems together.279 Rough-and-tumble play integrates, stylises, and perhaps ritualises elements from both the exploratory and the anger systems. Close observation of children during such play reveals how aggression and exploration, specifically, can be used in the service of the other.280 For Bowlby:

Affect laden behaviour I tend to view in terms of structures built of component bricks. The bricks are relatively stereotyped behaviour patterns, e.g. bird song or sucking, which, according to the species, may be built in or learnt or a combination of both. The larger structure, e.g. courtship or nest building, is less stereotyped and a complex synthesis of these components. Although in principle any component is available for any synthesis, in practice each synthesis tends to select a particular group of components. None the less it is probably usual for certain component items to be utilised in more than one synthesis.281

For instance, Bowlby discussed the close relationship between the attachment response and the flight response. These may have many common components, since we often flee to those we turn to with the expectation of protection, though there may be occasions when we do not do so, for instance when we do not perceive such individuals as available to us. Bowlby also expected the attachment and fear behavioural systems to share important cognitive components, such as expectations about the effectiveness of previous attempts to gain safety, and present perceptions of sources of danger.

Another pair of behavioural systems with significant shared components are attachment and sexuality. Ainsworth felt that Bowlby had seriously underplayed the importance and complexity of the sexual system, as part of his attempt to pull away from psychoanalytic theory.282 This is certainly true of the 1970s and 1980s. However, during the late 1950s, whilst (p. 51) Ainsworth was relatively out of contact in Uganda (Chapter 2), comparison of the sexual response and the following response was key to discussions between Bowlby and Hinde. Bowlby discussed the sexual system in some detail during the meetings of the WTO Study Group of the Psychobiological Development of the Child in the late 1950s.283 In these discussions Bowlby did not consider the activating and terminating conditions of the sexual system. Instead, his focus was on how the sexual response takes component behaviours from attachment relationships: ‘certain components in the human sexual response are derived from parent–child response’.284 Such common components might include affective and cognitive elements such as trust, affection, gentleness, and expectations around what interactions such as acceptance or rebuff will ultimately mean for a relationship. Common aspects might include behavioural components such as gazing, kissing, and coming into close proximity. This account allowed Bowlby to qualify Freud’s claim that early relationship patterns are organised as a whole into later sexual desire. For Bowlby, it seemed more cautious and more probable that only some components of earlier patterns are integrated into how adults show affection, enact courtship and sexuality, and provide care to their own offspring.285 These components are the thin wires connecting early care to later sexuality; there is no inevitable or single line of continuity.

Even if components of the attachment behavioural system ultimately get recruited and repurposed in adulthood, and even if the threshold for activating the system is increased, Bowlby nonetheless claimed that the attachment system remains available throughout life. In an emergency, or when hurt or scared, we are prompted to seek those we know and trust, and feel the more troubled when they are not available.286 As adults, when alarmed we may seek a spouse, a parent—or, ‘more often than might be supposed’, adults may look to their own child for reassurance.287 However, the surplus of meanings assigned to the term (p. 52) ‘attachment’ caused Bowlby to swing between two very different positions.288 On the one hand, ‘attachment’ was sometimes used broadly to mean early relationships as a whole. With this meaning in mind, Bowlby made strong claims for the influence of attachment on ‘a person’s whole emotional life’, and as the condition of possibility for all later emotional development and mental health.289 In Loss, for instance, he claimed that ‘attachments to other human beings are the hub around which a person’s life revolves … From these intimate attachments a person draws his strength and enjoyment of life.’290 Such claims later contributed to the impression by audiences of attachment theory that individual differences in early attachment are fixed for life (Chapter 2).

On the other hand, when ‘attachment’ was understood narrowly as a specific behavioural system, Bowlby’s stance was rather different. With the narrow ethological meaning in mind, he proposed that influences on this behavioural system in early life could have a pervasive influence on later mental health, but nonetheless tended towards more qualified claims. Human development is very complex, and no single factor is likely to play more than a moderate role. However, for a single behavioural system, Bowlby anticipated that the attachment system would be quite influential and therefore worthy of particular attention. The period of early following and help-seeking has its importance because ‘the period when they are most active is also the period when patterns of control and of regulating conflict are being laid down’. It therefore has a particular significance for socio-emotional development.291 Yet even children who have had ‘ghastly experiences’ may ‘nevertheless develop favourably’, with maybe only 20–30% subsequently showing severe problems. Furthermore, those who show severe problems are often those who have experienced combinations of adversities, rather than disruptions of attachment alone.292 Nonetheless, Bowlby anticipated that individual differences in early attachment experiences would have some implications for later mental health and wellbeing across the spectrum (see also Chapter 4).

In his final years, Bowlby saw the life of Charles Darwin as offering an especially clear case illustrating the importance of the disruption of the attachment system in childhood for later development, and also the emphasis he placed on family interactions for mediating the effects of separation or loss. Especially after seeing visitors or after some other excitement, as an adult Darwin experienced numerous symptoms including gastric pains, vertigo, vomiting, (p. 53) tremors, cramps, headaches, eczema, anxiety, and a compulsion to wander or search for something without knowing what. There remains much speculation about the cause of these symptoms, and the relative contribution of organic and psychological illness.293 In Charles Darwin: A Life, published in 1990, Bowlby argued that several other symptoms could essentially be explained as panic attacks, caused by the poor integration of behavioural, affective, and cognitive components of the attachment system following the loss of Darwin’s mother. The issue was not this disruption of the attachment relationship alone, however. Bowlby instead highlighted that the family insisted that Darwin was not permitted to talk about his mother after she died when he was eight years old. Bowlby speculated that this could have contributed to exciting emotions being misrecognised as threatening ones.294 In turn, he proposed that this misrecognition may have contributed to Darwin’s tendency to hyperventilate, with his body always on the cusp of physiological overarousal.295 In support of his claims, Bowlby offers evidence that Darwin’s symptoms were exacerbated whenever as an adult he felt threatened or experienced a loss.

The caregiving behavioural system

It is sometimes assumed that Bowlby also regarded parental behaviour and feelings as an expression of attachment. This error is more common among Bowlby’s critics, who generally privilege his writings of the 1950s and ignore his later qualifications and clarifications.296 However, such usage can sometime be seen in the work of attachment researchers too, albeit less since the 2000s.297 It follows some confusing and overextended use of the term ‘attachment’ by Bowlby in the early 1950s that did include caregiving behaviour. Though he attempted subsequently to distinguish attachment and caregiving, the early work set up a chain reaction that shaped both interpretation of his theory and measurement tools built from it. For instance, the Maternal–Fetal Attachment Scale was introduced by Cranley in 1981 to measure a mother’s ‘attachment’ (feelings of care for) her unborn baby.298 Such institutionalised uses of ‘attachment’ to mean caregiving have caused even attachment researchers assiduous about terminology, such as van IJzendoorn and Bakermans-Kranenburg, to (p. 54) refer to ‘attachment’ when they mean caregiving.299 Perhaps a source of confusion between caregiving and attachment for Bowlby in his early writings was that there did seem to be common elements. As mentioned earlier, in his mature writings Bowlby identified that some components of the attachment system, like the capacity to communicate about emotion with others, may be recruited when an adult is required to provide care. However, this does not imply that caregiving is a reflection or expression of attachment. The two systems evolved in parallel, and function reciprocally, producing important forms of potential cooperation and friction.300

In Attachment, Volume 1, Bowlby developed an account of the caregiving behavioural system modelled on his existing ideas regarding the attachment behavioural system. The caregiving system was specifically anchored in the retrieval response, as a reciprocal partner to the infant’s following response for maintaining proximity:

Retrieving can be defined as any behaviour of a parent a predictable outcome of which is that the young are brought either into the nest or close to mother, or both … The retrieving behaviour of a primate mother gathers the infant into her arms and holds him there. Having a similar outcome to the attachment behaviour of young, it is probably best understood in similar terms—namely, as being mediated by a number of behavioural systems the predictable outcome of which is maintenance of proximity to the infant … especially when her infant is playing contentedly with other known individuals in the vicinity, a mother may let things be. Yet her tendency to retrieve him is not wholly dormant: she is likely to keep a watchful eye on him.301

Yet just as the attachment behavioural system was anchored in the infant’s following response, but inserted into a broader image of comfort-seeking and relationship, the caregiving system had a parallel hinge. Like the attachment system, then, Bowlby’s notion of the caregiving system sustained an ambiguity between the narrow notion of the retrieval response (and, presumably, related aspects of holding) and a more expansive notion of the provision of ‘encouragement, support, help and protection’.302 In notes from 1978, disliking throughout his career the ambiguity of the word ‘empathy’, he referred to the key emotional component of the caregiving system as ‘concern for the welfare of others’.303 And in a late interview he emphasised the importance of attentional processes that serve as architecture for the caregiving system.304 Yet Bowlby’s written reflections on the set-goal of the caregiving (p. 55) system are scarce. And the conflation of specific and general set-goals of the caregiving system is a major limitation of his remarks, given their significant differences. Not least, retrieval, following, and attachment are all homeostatic systems at root, whereas the provision of care and nurturance in a wider sense aims to support growth, not a return to equilibrium. Retrieval and encouragement/support also seemed to be modulated and their successful achievement met by potentially quite different affects in the caregiver, including quite different senses of commitment and pleasure.305 Indeed, subsequent to Bowlby, ethologists have come to distinguish a consoling system from a caregiving system, given their distinctive behavioural repertoires and conditions of activation and termination.306

Despite conceptual ambiguity in Bowlby’s writings about caregiving, neither the narrow nor the broad notions of the caregiving system implied, as Vicedo has mistakenly claimed, that Bowlby discussed mothers as ‘unthinking and natural’, acting ‘just out of instinct’, in behaviour devoid ‘of rationality, of choice and of moral value’.307 Vicedo’s description is a reductive characterisation of even Bowlby’s populist writings, and as a characterisation of his theoretical texts, it is essentially polemical. Bowlby did not regard parenting behaviours as characteristic only of mothers—though Ainsworth was more critical than Bowlby of the cultural values that assigned responsibility for caregiving to mothers.308 As mentioned in the section ‘The attachment behavioural system’, Bowlby regarded all behavioural systems as neurologically primed, but dependent on experience and contextual support for their elicitation, elaboration, and expression. He stated explicitly in relation to caregiving behaviour that his ‘view of behavioural development contrasts sharply with both of the older paradigms, one of which, invoking instinct, over-emphasised the preprogrammed component and the other of which, reacting against instinct, overemphasised the learned component. Parenting behaviour in humans is certainly not the product of some unvarying parenting instinct.’309 In a presentation of his ideas for radio in 1969, Bowlby drew a comparison between parenting behaviour and language. In both cases, humans are predisposed to pay attention to certain cues, and elaborate systems of meaning and practice. However, the nature of these meanings and practices depends on our opportunities for learning, and our induction into (p. 56) cultural systems. For instance, the development of retrieval response/caregiver behaviour is shaped by what opportunities an adult has had to touch children in the past, and the child’s own reciprocal signals.310

The role of both the motivational and behavioural aspects of the caregiving system, and of the role of learning and culture, can be illustrated by two phenomena that especially concerned Bowlby. A first was the use of threats as part of caregiving, a behaviour that had interested him from his earliest work as a clinician working with families.311 In Separation, threats by caregivers are a central theme. Bowlby acknowledged that threats might well have the ultimate end of retrieval or protection, the two suggested set-goals of the caregiving system depending on whether it was narrowly or broadly construed. For instance, a threat of abandonment by a parent might be used effectively to keep a child nearby or ensure their obedience, where this is considered of special importance.

However, the same behaviour might express the anger behavioural system, or be a compound fed by both the anger and caregiving systems. Bowlby emphasised that threatening as a cultural practice is quite complex for a young child to understand: it is intensely felt and expressed by the parent, but in fact localised and not fully serious. To the degree that the serious-but-not-serious quality of threats can be difficult for children to understand or trust, he worried about the long-term harm verbal threats could have on a child’s perception of caregiver availability in times of need.312 And he argued that verbal threats to hurt or abandon the child by a parent can contribute to conflict between the attachment and fear systems, since threats with a marked aggressive quality by a caregiver would elicit a desire for both approach and avoidance by their child (Chapter 3).313 In a letter to Henry Hansburg in 1979 he observed that parents may ‘insist that no one loves a child who behaves so and so, and that no one will ever love you unless etc. etc. I believe that when a child is exposed to these pressures over many years, the notion of his being unlovable becomes deeply engrained.’314

A second phenomenon that especially interested Bowlby was what he termed ‘compulsive caregiving’ in The Making and Breaking of Affectional Bonds, published in 1979.315 This was the term Bowlby used when caregiving behaviour was shown by an individual as an alternative to seeking care. Children or adults may receive clear signals from their attachment figures that asking for comfort is not permitted or would backfire.316 However, in providing care themselves, they at least gain some closeness with and availability from their attachment figures. The culturally mandated and practical closeness that caregiving permits allows individuals to achieve the set-goal of the attachment system, even if in a contorted and limited way.317 The caregiving system is therefore used in the service of the attachment system, a (p. 57) phenomenon that highlights both that caregiving behaviour in humans is not the product of some unvarying parenting instinct, and that the silent expression of attachment needs may be present even lodged inside child-to-adult caregiving behaviours.

In his book A Secure Base, published in 1988, Bowlby gave particular attention to conditions that may reliably elicit caregiving in the place of attachment behaviour, at least from toddlerhood, and the increased cognitive resources for attending to the mind of the caregiver.318 Child-to-parent caregiving in the place of attachment may be seen when a parent has very severe depression such that the child’s attachment signals elicit little or no response except when the child him-/herself helps the parent to cope, for instance through retrieval behaviours shown towards the parent. Child-to-parent caregiving may additionally be elicited if a parent’s own attachment system is activated and becomes directed, through circumstances or misdirection as a consequence of conflict, towards the child.319 Bowlby observed that child-to-parent caregiving responses may also be elicited, and combined with components of the fear system, in placatory behaviour. In such cases, for instance, ‘children learn early that it is possible to placate a disturbed and potentially violent mother by constant attention to her wishes. Such apparently placatory behavior has been observed in young abused children, some less than two years old.’320

Yet a problem with Bowlby’s characterisation of the caregiving system was the ambiguity between its narrow reference, referring primarily to the retrieval response, and its flung spray of wider meanings.321 Bowlby’s inattention to this system and ambiguities in conceptualizing its set-goal contributed to a neglect of the caregiving behavioural system by attachment researchers until the mid-1990s. A re-examination of Bowlby’s position by George and Solomon, in the context of a rising cultural concern with ‘motherhood as an experience and institution’, led to renewed scientific discussion and debate of Bowlby’s position.322 An important point made by George and Solomon was that the mature balance of the caregiving systems with other behavioural systems is critical for the provision of ‘good enough’ care to a child. They contrasted their account to the image put forward by Bowlby, especially in his (p. 58) early writing, of the caregiving system as the perfect match for the needs of the attachment system and hence the child.323

George and Solomon also argued that economic and social supports are important for the effective elaboration and functioning of the caregiving system. This is, in fact, exactly in line with an argument made by Bowlby himself in his book from 1953, The Roots of Parenthood, which was not published in America and so was unavailable to George and Solomon. There Bowlby emphasised that caregiving is dependent on the material and social resources available to a parent, which support a caregiver’s energy, patience, and courage in the face of the demands of caring for a child. Without support, a caregiver may well ‘give up trying’, no matter that ‘they would like to give their children all that good parents do’.324 Bowlby condemned government inattention to ‘the poverty of mothers with young children’ and called on his readers to ‘campaign unremittingly until it is remedied’.325 This was not a campaign that he, however, would pursue over the coming decades. Instead, in his later work this concern with the conditions of the effective functioning of the caregiving system focused on psychological factors. He recognised the role of despair in hindering caregiving. However, after the early 1950s, he did not spell out poverty and isolation as potential contributors to despair. The traumatic loss of Bowlby’s close friend, the Labour MP Evan Durbin, in 1948 undoubtedly played a role in this transition in Bowlby’s thought.326

Besides George and Solomon, another important appraisal of Bowlby’s concept of the caregiving system was presented by the anthropologist and primatologist Sarah Hrdy. In her 1999 book Mother Nature, Hrdy observed that ‘Bowlby’s ideas will stand among the greatest contributions made by evolutionary-minded psychologists to human wellbeing.’327 However, she questioned his assumption in Attachment, Volume 1 that in the environment within which humans evolved, care would have been provided by a parent alone; the resources required by a human infant are far more than any caregiver could be expected to provide on their own. Hrdy suggested that Bowlby’s own biases and popularizing polemics combined with the available ethological and anthropological evidence of the time to paint a false image of mother–infant care, neglecting the necessity of social and economic supports for caregiving. Additionally, Bowlby drew on baboons and rhesus monkeys as his primate models, when in fact contrary evidence was shown from infant-sharing primates like langurs.328 Whilst infants usually have a relatively small number of primary caregivers, the network of people involved in offering some care to a child may be very wide. Certainly (p. 59) in human evolutionary history, care has not usually been provided by one person alone as Bowlby sometimes implied, especially in his earlier writings.329

Hrdy offered no objection to Bowlby’s claim that the attachment behavioural system discriminates caregivers, and perhaps even has a hierarchy of attachment figures. However, she proposed that humans have evolved to engage in cooperative care, with mothers, fathers, grandmothers, aunts, siblings, and adult friends all involved, depending on circumstances. And when insufficient economic and social resources and supports are available, caregivers can be expected to divest from their children, reducing their availability as sources of protection.330 The attachment system evolved in the context of this threat, and it is part of what makes exploration a dangerous activity for an infant who is not sure of the caregiver’s availability. Likewise, the potential for an attachment figure to divest from their infant if other demands are pressing is part of what makes a network of attachment figures part of the evolutionary expectable conditions of infant survival.331 Dragged towards reductionism at times by the narrow version of caregiving as retrieval, Bowlby did not adequately recognise the extent to which the ordinary functioning of the human caregiving behavioural system entails not only direct care, but also bringing additional supports on line for one we care for. And he did not adequately recognise that an integral part of the functioning of the caregiving system entails protecting the one we care about from ourselves, whether this is failures of patience or the wish to abandon the other when we feel overwhelmed. These are important ‘anti-goals’ of the caregiving behavioural system (Chapter 5).

Hrdy’s critique has been well taken at a theoretical level by attachment researchers. In general, attachment researchers emphasise that young children will likely not be significantly harmed, and may even benefit, from high-quality care from several caregivers who develop personal relationships with the child. The primary qualifications that have been offered are that: (i) the child should not be passed around so many people that it proves difficult to develop at least one stable relationship, as is often the case with institutional care; and (ii) insecure attachment is made more likely if an infant is in very extensive daycare (e.g. over 70 hours a week), or (iii) if nights are spent with relative strangers.332 A primary caregiver by no means needs to be present to the degree that Bowlby claimed in his popular writings of (p. 60) the 1950s. And Bowlby’s use of the term ‘mother’ to mean a primary caregiver—following common practice in psychoanalytic discourse of his day—should not be taken to imply that a primary caregiver needs to be a mother.

It is false to say of the contemporary field of empirical attachment research, as do Keller and Chaudhary, that it ‘aims at demonstrating the uniqueness of the mother–child bond’.333 If this were the aim, then, paradoxically, greater attention would have been paid to a variety of caregivers in order to comparatively demonstrate the particular qualities of maternal care. Other care providers have sometimes been studied by attachment researchers, for instance daycarers and foster parents.334 However, these have hardly been comparisons to demonstrate the uniqueness of maternal care, as Keller and Chaudhary allege. Yet Keller and Chaudhary are undoubtedly correct in their criticism that, in practice, attachment research after Bowlby in the developmental tradition has predominantly focused on practice on mother–infant relationships. Comparatively few researchers have been able to access or prioritise the resources for studies of the contributions of wider caregiving networks using the labour-intensive measures of developmental attachment research,335 though certainly some research has been done on caregiving networks by developmental researchers.336

Nonetheless, it is striking, for instance, that not a single study to date has examined infant–grandparent attachment using the Strange Situation, given the widely acknowledged importance of grandparents as alloparents in both human evolutionary history and contemporary societies.337 The foremost reason why attachment researchers have focused research attention on mothers is that mothers have had primary care responsibilities for infants in America and in Europe, where the majority of attachment researchers are based. However, there are certainly enough infants living with their grandparents for this to have been a viable topic for study at some point since the 1970s. A focus on maternal care may have been partly self-perpetuating for attachment researchers after Bowlby in the developmental tradition, as it made this population the standard for comparability and study of the effects of specific variables on care.338 A positive recent development, however, are plans over the coming years by Or Dagan and Avi Sagi-Schwartz to pool the raw data from all studies of attachment (p. 61) to multiple caregivers using standardised measures, to permit statistical analyses with good depth and breadth, and to reconcile questions of comparability.

The exploratory behavioural system

The exploratory behavioural system was first detailed in Attachment, Volume 1, though the discussion remained relatively cursory.339 Following Hinde’s earlier discussions, Bowlby argued that the activating conditions of the exploratory behavioural system are ‘stimuli that are novel and/or complex’.340 An additional criterion implicit in Bowlby’s account of the exploratory system, but left unstated, is that the stimuli need to have some potential relevance to the individual. In infancy, Bowlby argued that the exploratory system has three elements that often form a sequence: orienting towards the stimulus to gain information and prime responsiveness; approach to the stimulus to examine it; and manipulation or experimentation with the stimulus in order to understand what it can do.341 The characteristic affect that supports and calibrates the expression of the exploratory system is ‘curiosity’.342 Like attachment, exploration was conceptualised as a homeostatic system, since once relevant novelty and complexity had been examined and understood, the system was expected to terminate. Bowlby situated the exploratory system as equivalent to Winnicott’s concept of ‘play’, highlighting its importance as a system through which a child learns that meanings are partly self-created and partly supported by what the world makes available.343 The experience of joy requires a world that can still surprise us with things we cherish, especially when we have attachment figures with whom to share discoveries. So Bowlby regarded the exploratory system as implicated in the pleasures that can come from connecting and reconnecting the (p. 62) self and world, in sustaining their dialogue.344 As the game ‘peek-a-boo’ demonstrates, even the absence of the caregiver may offer the pleasure of manipulation and experimentation if it occurs under the aegis of the exploratory behavioural system, and is brief enough not to threaten the dominance of this system.

Ainsworth qualified Bowlby’s account through her attention to half-hearted forms of exploration. She identified this kind of play in infants showing avoidant behaviour in the Strange Situation Procedure (Chapter 2). She suggested that these infants were using play to distract themselves from attachment-relevant internal and external cues, thereby suppressing the expression of the attachment system.345 Likewise, Main later observed that frame-by-frame video analysis of the reunion episodes of avoidant dyads ‘reveals that the inanimate object generally has far from the infant’s full attention. The infant may, for example, rather frantically turn toward a tale leg and finger it (but with eyes fixed blankly on the wall ahead). Or, in a clumsily decisive move, it may suddenly drop a toy into a box but then close the box on its hand. The general impression is, again, that the infant could not succeed in maintaining its avoidance of the parent without the aid of the object seized.’346 Here exploration is used as somewhere else to go, a projection of attention away from felt concerns and into the environment.

The contrast between the lacklustre and distracted infant exploration seen in avoidant dyads and true play burnished by lively curiosity led Ainsworth to criticise Bowlby for his account of the exploratory system, which only infrequently drew distinctions between these two forms.347 Ainsworth suggested that Bowlby had given a place to exploration, but did not take it seriously enough as a behavioural system. His characterisation had focused too much on its behavioural components, at the expense of affective and cognitive components that lead a child to engage in ‘actively seeking’ information within a social and cultural context.348 She offered the example of an infant encountering a novel and potentially alarming situation. Before the sequence described by Bowlby of orienting, approaching, and experimenting, Ainsworth asserted that a first stage of the exploratory system in infancy will be social referencing of attachment figures. The caregiver helps offer shape and definition to the infant’s world and will often be actively sought to serve as ‘referee’ for the meaning of the stimulus, especially when there is a question whether its novelty should stimulate exploration and (p. 63) affiliation or wariness. For Ainsworth, exploration took part of its significance from the long infancy of human beings, and the dense, multifaceted urgency of the human society that infants must learn to navigate. A behavioural system that makes matters that are novel and/or complex attractive for exploration is therefore a great asset, since, alongside staying safe moment-to-moment, the other great task of children is to encounter and learn from the world around them (Chapter 2).349

The fear behavioural system

In Bowlby’s account, when a stimulus is judged as alarming, two systems may be activated: the attachment system and the fear system. The first third of his book Separation was essentially a theory of the fear behavioural system.350 Bowlby urged recognition that attachment and fear are ‘distinct behavioural systems that (a) have the same function, (b) may be elicited by many of the same conditions, (c) are frequently compatible with each other, but (d) can easily be in conflict’.351 The compatibility of the attachment system and the fear system is that both are oriented towards increasing proximity to safety in contexts of perceived threat. The difference, for Bowlby, is that the attachment system disposes attempts to achieve the availability of attachment figures, whereas the fear system prompts attempts both to escape the threat and to find people or places that have proven safe in the past. Bowlby suggested that the term ‘alarm’ should be used for concern about a threatening stimulus, and ‘anxiety’ should be used for concern about the availability of safety, whether due to the presence of danger or questions about the availability of attachment figures.352 In distinguishing the attachment and fear systems, however, Bowlby left unaddressed whether or how feelings of safety and of security might inflect one another, and how the set-goal of the attachment behavioural system might be conditioned by experiences of the fear behavioural system and its resolution.

(p. 64) Component behaviours of the fear system identified by Bowlby in his notes included the startle response, withdrawal, fleeing, and hiding.353 Later, on the suggestion of Hinde, he added freezing as another behavioural component of the fear system.354 This led him to discuss ‘three distinct kinds of predictable outcome’ of the fear behavioural system ‘(a) immobility, (b) increased distance from one type of object, and (c) increased proximity to another type of object’.355 These responses, he proposed, are highly responsive to context: ‘When a chimpanzee is startled by a sudden noise or movement nearby, its immediate response is to duck its head and to fling one or both arms across its face; alternatively, it may throw both hands in the air. Occasionally these startle reactions are followed by a hitting-away movement with the back of the hand towards the object, at other times by flight. When the alarming object is another and more dominant chimpanzee, flight is accompanied by loud screaming; when it is anything else, flight is quite silent.’356 In an unpublished paper from 1973 entitled ‘Wariness’, Bowlby discussed a continuum of fear responses. At one end was flight and expressions of terror. At the other end was various forms of wary behaviour.357

In Separation, Bowlby proposed that the activating conditions for the fear system are partly learnt on the basis of experience, but are also partly neurologically primed by human evolutionary history. Children are certainly more likely to fear the dark if they have been told that darkness is dangerous, they have seen others respond with caution to darkness, or if they associate darkness with unhappy experiences. However, Bowlby argued that humans are predisposed to have a lower bar for treating certain cues, such as darkness, as threatening. Other such cues potentiated to have lower bars for activation include being cold and loud sudden noises. Perhaps the most important such cue for danger, Bowlby observed, is the feeling of being alone, or of blocked access to attachment figures.358 Furthermore, he argued that the bar for activation of the fear behavioural system would fall dramatically where two or more such cues were present at once, and especially when one of these was feeling alone. Where multiple cues to danger occur at once, most adult humans respond with an activation of the fear system.359

However, especially in childhood, we may not know how to interpret the cause of such an activation: ‘Often, in fact, when we feel impelled to act in a certain way that is readily explicable in terms of biological function, we concoct “reasons” for doing so that bear little or no relation to the causes of our behaviour. For example, a child or adult, who in order to reduce risk is biologically disposed to respond to strange sounds in the dark by seeking his attachment figure, gives as his reason that he is afraid of ghosts.’360 Bowlby contended that children’s actual experiences in situations with such primed cues for danger are likely to be especially important for the later development of mental health problems. For instance, the origins of (p. 65) many phobias and paranoid symptoms can be understood as ‘intelligible, albeit distorted’ responses to historical experiences in contexts with multiple such cues, where a caregiver was not available to help soothe the child and interpret the meaning of their response.361

On the basis of analysis of their activating and terminating conditions, Bowlby was clear that the attachment and fear systems were distinct. With Robertson, he had considered the predicament of hospitalised children, who experienced fear but without being able to turn to their caregivers. Part of Bowlby’s focus on long-term separations stemmed from his sense that the clinical community and public were not able to acknowledge child abuse by parents. Yet in the early 1970s, the tide was turning on the acknowledgement of child physical abuse within Britain, and within the medical community in particular.362 The manuscript history of Separation sees Bowlby edging slowly towards the claim that the attachment and fear behavioural systems could conflict as a result of directly frightening or abusive actions by the attachment figure. In an early draft of material for Separation Bowlby reported having seen a lamb attempting to cross directly into the path of a car, having been startled by the sound of the car and seeing its mother on the other side of the road.363 In correspondence with Ainsworth, and then in the typescript, Bowlby added consideration of the case of an infant who is confronted with a threatening stimulus, such as a ‘barking dog’, that lies in the path towards the attachment figure.364 Reflecting further, in a pencil annotation on the typescript and then in the published version, Bowlby added that ‘a special but not unusual situation in which there is conflict between attachment behaviour and withdrawal is when the attachment figure is also the one who elicits fear, perhaps by threats or violence. In such conditions young creatures, whether human or non-human, are likely to cling to the threatening or hostile figure rather than run away from him or her.’365 A threatening or violent caregiver, Bowlby observed, could be expected to activate the fear and attachment systems, producing conflict between the two behavioural systems (Chapter 3).

The aggression behavioural system

With one exception in Attachment, Volume 1, Bowlby did not generally describe aggression as a behavioural system.366 Yet consideration of the case of the aggression system in closing this section on behavioural systems may help shed light on the boundaries of this concept for Bowlby. He gave aggression only passing and faltering attention.367 Despite the fact that the subtitle of his book Separation is ‘Anxiety and Anger’, in fact Part I focuses on anxiety, Part (p. 66) II focuses on fear, and Part III returns essentially to a consideration of anxiety, with the exception of a dozen pages in Chapter 17 (entitled ‘Anger, Anxiety, and Attachment’).368 Mary Main later observed that Bowlby’s exclusion of aggression as a behavioural system ‘has puzzled many clinicians’, and emphasises the disparity this has left between theory and empirical research on aggression.369 Shaver and Mikulincer expressed disappointment with Bowlby for his poor attention to aggression, since it has left attachment theory without a specified place for anger (Chapter 5).370 And van IJzendoorn and Bakermans-Kranenburg and colleagues felt forced to bypass Bowlby in their development of an intervention with caregivers struggling with aggressive child behaviour.371

It may be speculated that aggression was refused the status of a behavioural system because Bowlby wanted to avoid even the slightest implication of acceptance of Melanie Klein’s position. Whereas Klein regarded aggression as an innate human drive, Bowlby claimed that most aggression is a response to the frustration of other behavioural system.372 Certainly this is a point that he reiterates time and again when aggression is under discussion. Yet, as Ainsworth observed, frustration could still then be a behavioural system: it is neurologically primed as a result of human evolutionary history, but requires frustrating experiences for its elaboration, and works to achieve its goals through various behavioural means.373

An additional factor may have contributed to Bowlby’s wariness in describing aggression as a behavioural system. Bowlby’s younger colleagues at the Tavistock recall that he used to think aloud in seminars and conversations about whether the idea of ‘aggression’ was too much of a catch-all. He was far from certain that it was a single thing.374 In Separation, he described two forms of aggression, only one of which resembled a behavioural system. (p. 67) The first form was what he termed ‘the anger of despair’.375 In this form of aggression the behaviour occurs primarily as the expression of emotion, spinning within its own loops of intensity, without the environmental responsiveness that would help behaviour to achieve a specific set-goal. The anger of despair, Bowlby argued, ‘occurs whenever a person, child or adult, becomes so intensely and/or persistently angry with his partner that the bond between them is weakened, instead of strengthened, and the partner is alienated. Anger with a partner becomes dysfunctional also whenever aggressive thoughts or acts cross the narrow boundary between being deterrent and being revengeful.’376 The anger of despair is a mood that disbelieves hope; it is premised on the assumption that aggression cannot change the environment.

He contrasted the anger of despair with ‘the anger of hope’.377 This form of anger is elicited by frustration of another behavioural system. It may seek removal of the obstacle. More often, though, it acts upon people or the environment in such a way as might coerce them to yield, permitting the satisfaction of the other system. For instance, Robertson documented a great increase in aggressive behaviour in children following their hospital separations. Bowlby felt that this might be conceptualised as a kind of punishment of the caregivers for the separation by the child, signalling that the experience should not be repeated: ‘In its functional form anger is expressed as reproachful and punishing behaviour that has as its set-goals assisting a reunion and discouraging further separation. Therefore, although expressed towards the partner, such anger acts to promote, and not to disrupt, the bond.’378 The anger of hope is a mood that disbelieves despair; it is premised on the wish that aggression may change the environment.

The anger of hope is a dangerous strategy, and may readily backfire by provoking retaliatory aggression or rejection from the partner or caregiver. As Waters and Sroufe later observed, aggression risks (further) disrupting the trust and patience on which any secure attachment relationship must rest.379 However, this risk may be better, or be felt to be better, than alternatives, such as being unnoticed or despairing of achieving co-regulation within a relationship.380 The anger of hope has a good claim to be a behavioural system, especially in light of its clear set-goal. However, perhaps a mark against even the anger of hope as a behavioural system is that it remains the servant of other behavioural systems, and when it achieves independence from them it is as the dysregulated anger of despair. The anger of hope may have seemed to Bowlby as a component behaviour of other behavioural systems, rather than having the status of a behavioural system in itself. However, as Waters and colleagues have argued, such a stance ascribes too little independence to aggressive behaviours, not least since they may be fed by more than one source of frustration.381 Moreover, several second-generation attachment researchers including Shaver and Fonagy later criticised (p. 68) Bowlby on the grounds that aggression can be evoked in the absence of the frustration of a behavioural system.382

The meaning and structures of symptoms


A central motivation—perhaps the central motivation—for Bowlby’s development of a theory of behavioural systems was to account for the development of mental health symptoms in humans. His new model allowed him to consider symptoms that might arise from the interplay or conflict of different systems, from poor alignment between behavioural systems and their social context, or from aspects of the functioning or misfiring of individual systems themselves. Reflecting on his clinical experiences with adults, Bowlby was impressed by the psychoanalytic insight that even an inhibited response not only continues to exist but can develop, ‘putting out derivatives and establishing connections’ even if these may be contorted by the inhibition.383 For instance, inhibited behavioural systems, even if they do not influence our behaviour, may nonetheless exert a powerful influence on our judgement and mood, as well as the activation of memories and forms of imagination or daydream.384 Such processes may influence individuals’ experiences even of their own motivations. In Attachment, Volume 1, Bowlby wrote that ‘reports are usually trustworthy, but psychoanalysts know well that that is not always so. A subject may in fact mis-identify the set-goal of a behavioural system currently active—and such mis-identification may itself be the result of interference by an active system that has a set-goal which is incompatible with the first. This leads to the concept of unconscious wish.’385

However, in discussions of such phenomena, Bowlby felt that the concept in common use in the psychoanalytic community, ‘defence’, was often confusing. The term absorbed a variety of psychological and behavioural activities that aimed to reduce or eliminate experiences that might threaten the integrity or stability of the individual. Bowlby felt that this usage was lazy, and left nameless and soundless many critical psychological processes.386 He addressed (p. 69) these concerns in a short unpublished book from 1962 entitled Defences that Follow Loss: Causation and Function. (‘Defences that Follow Loss’ recently appeared in Trauma and Loss: Key Texts from the John Bowlby Archive, edited by Duschinsky and White.)387 In this work, Bowlby disparaged the concept of ‘defence’ as a confused mix of description and explanation. He felt that it also served to mask the important distinction between the inhibition of a motivation and the inhibition of behaviour, and the distinction between the immediate cause and the ultimate function of inhibition. All too often, Bowlby felt, psychoanalysts discussed ‘defences’ as if they were initiated in order to avoid some foreseeable consequence.388 For instance, psychoanalytic discussions of obsessive-compulsive symptoms at the time tended to situate them as a defence against Oedipal conflicts. Klein argued that ‘obsessional neurosis is an attempt to cure the psychotic conditions which underlie it’, which arise from a predicament of experiencing both love and hate for the mother.389 Bowlby felt that such claims failed to clarify the initiating conditions of obsessional symptoms, the contexts that led them to become enduring, and the evolutionary function obsessive-compulsive behaviours may have served.

In failing to draw such distinctions Bowlby argued that Klein and other contemporary psychoanalysts were unable to distinguish effectively between defensive strategies and strategies for coping.390 Some forms of defence are evidently under our conscious control; others indicate that conscious control is breaking down. Some forms of defence are helpful for individuals in responding to their environment; others are highly destructive. Yet psychoanalysts were making these distinctions remarkably rarely (with a few, inconsistent exceptions like Donald Fairbairn and Joseph Sandler).391 In general, Bowlby felt that in psychoanalytic theory ‘the relation of defence to healthy control, or to coping processes, has never been clarified. Like Melanie Klein, most analysts hold the view that there are no great differences between them.’392 The primary current of psychoanalytic thought directed attention away from the question of which defences were able to contribute to individual coping, for instance through offering short-term adaptation to an adverse environment for an individual.

To address this issue Bowlby appealed to the four questions of ethology: (i) the contribution a behaviour may have for species survival or reproduction; (ii) how the behaviour came about in the course of natural selection; (iii) the behaviour’s mechanism or how it works; and (iv) how it develops in the individual. To Bowlby, the processes and behaviours subsumed under the term ‘defence’ likely have very different answers to these questions. However, what they likely have in common, he argued, is that whilst at a species or population level they may (p. 70) help at times to contribute to survival or reproduction, for a given individual they may be baffling and sometimes counterproductive: ‘defensive processes come into action in certain conditions and that they do so without the individual having any more idea of their biological function than the ordinary man has of the function served by his temperature rising when he contracts an infection’.393 Later, in Loss, Bowlby urged that ‘the effects of defensive activity must be judged on a number of distinct scales. For example we can ask: what are its effects, beneficial or otherwise, on the personality concerned? what are its effects, beneficial or otherwise, on the members of the person’s family? what are its effects, beneficial or otherwise, on the community at large?’394 However, Bowlby felt that the concepts available for even asking these questions were overloaded and confused in psychoanalytic discourse. This prompted his decision to introduce the concept of segregated systems.


Another problem that Bowlby had with the concept of defence as used by the psychoanalytic community of his day was that it seemed to him oriented by the basic image of withdrawal from an aversive source of excitation. However, whereas in some ways the concept of defence was too absorptive, in this regard Bowlby found it too restrictive. Picking apart the concept, Bowlby narrowed in on the idea of the inhibition of behavioural systems.395 In Defences that Follow Loss: Causation and Function in 1962, he introduced for the first time the concept of ‘segregated systems’. This concept would be pulling the strings as the governing concept of Bowlby’s later thought, but in fact was only introduced in print 18 years later in Loss, in a discussion so brief that subsequent researchers have generally not found it clear or usable.396 Defences that Follow Loss, by contrast, engaged in an extended and elaborate discussion of the concept. In this book Bowlby argued that behavioural systems are ‘segregated’ if there is reduced, distorted, or blocked communication with perception, memory, and/or other behavioural systems. Though there was, intentionally, some overlap between Bowlby’s term ‘segregation’ and Freud’s concept of ‘repression’, Bowlby preferred the former term since he felt that ‘repression’ had accrued too much baggage, such as implying processes that keep something unconscious. By contrast, segregation could occur between two conscious systems, or between a conscious system and memory or perception.397 And whereas for Freud, (p. 71) repression worked against the dynamic unconscious as a ‘cauldron full of seething excitations’,398 for Bowlby segregation was an abstraction used to describe any process that resulted in a blockage or inhibition of communication within or between behavioural systems.

It might be thought that segregation would be regarded by Bowlby as a bad thing. And this has been the impression of Bowlby’s readers who have access to only his published discussions of segregation: an apparent denigration of segregation has been part of Bowlby’s legacy to later researchers and clinicians, among the very few who have engaged with the concept.399 Bailey and colleagues, for instance, seem to presume that all segregation must be an effect of trauma.400 However, such denigration ignores the potential for helpful segregation, through processes that buffer or segment and creatively recompose what is taken in from outside. In fact, in Defences that Follow Loss, Bowlby argued in some detail that long-term mental health would be supported by effective communication between mental systems on the basis of relative and flexible forms of segregation, rather than those that were strictly held. He was attentive to the role of segregation within typical development as well as atypical development, and its contribution to mental health, as well as mental ill health.401 He was specifically interested in ‘the functions that in health the segregation processes serve’ and gave careful consideration to the idea that segregation should be regarded ‘in the same light that Claude Bernard taught us to view physiological illness, namely as the outcome of processes that are beneficial in kind but faulty in amount’.402

Bowlby used the term ‘selective exclusion’ to refer to strategies of behavioural and attentional aversion or withdrawal that enact a flexible and relatively minimal segregation. ‘Selective exclusion’ may contribute to the mutual enrichment of behavioural systems. Bowlby gives the example of engrossment: selective exclusion of other thoughts, distractions, and daydreams may insulate and protect the effective operation of the exploratory system. Other cases are easy to imagine. For instance, selective exclusion can help preserve sources of private and sustaining joy, warped and wonderful against the flattening bustle of the world. Or to take another example, selective exclusion may be helpfully deployed to keep worries away during relaxation or sleep. The direction and quality of attention would need to be flexible enough to change once work began again. Where this can be achieved, communication between systems ensures that benefits of physical and attentional rest are transferred in the form of feeling genuinely refreshed.

Segregation essentially means that a system is not accepting information from another source, or potentially from any source. Sometimes, as in the case of selective exclusion, this (p. 72) is a generally beneficial filtering process. However, segregation can also occur because otherwise useful information is experienced as difficult to accept or incompatible with currently held values.403 This is likely the meaning of Bowlby’s claim, cited earlier in the ‘Introduction’, that the conditions of lack of integration between aspects of the self rest on conditions of forgivability or unforgivability: the critical question for the segregation or desegregation of systems is whether they are willing to accept information from one another, from memory and from the world, recognizing it as content available for inclusion within the system. If the information is regarded as unacceptable or dangerous, it may not be reconciled, causing segregation of the behavioural system.

Defensive exclusion

Bowlby was not concerned about segregation and the incompatibility of behavioural systems so long as ‘when they conflict, as habitually they do, regulation is tolerably smooth and efficient’.404 Some segregation is an inevitable part of being human, and localised and controlled incompatibility can provide a foundation of imagination, creativity, and work–life balance. By contrast, Bowlby felt that rigid and intense segregation can lead to very significant problems.405 One fundamental consequence of rigid and/or intense segregation is that the system then does not have access to new perceptions, or to reciprocal development with other behavioural systems. Not only does the cognitive component of a segregated behavioural system not receive the benefits of the learning provided by the exploratory system, or conscious reflection, but also the opportunities for feedback and revision are blocked that would usually occur when a behavioural system is expressed.406 Without such feedback, behavioural systems cannot fully benefit from the opportunities and lessons of the environment. When the behavioural system does get expressed, the result may be too strong or extreme, or clumsily formulated or ineffective.407

Furthermore, in Bowlby’s account, behavioural systems become active only when their activating conditions are met. As a consequence, to the degree that a behavioural system is segregated from information about these conditions, the system will only be partially activated or not activated at all. When strategies of behavioural and attentional aversion or withdrawal are used to inhibit memories that would otherwise activate a (p. 73) behavioural system, Bowlby referred to ‘cognitive disconnection’.408 The main effect, he proposed, was that strong emotions would seem to appear out of nowhere or be associated with inappropriate objects, since they have been disconnected from their original, provoking source. If cognitive disconnection is the segregation of memory, Bowlby offered the term ‘defensive exclusion’ to refer to the segregation of external perception. Cognitive disconnection keeps memory, reflection, and other cognitive processes from causing the activation of a behavioural system. Defensive exclusion keeps new experiences from achieving this same consequence. Both are dynamic properties that block relevant information from having implications for a behavioural system or systems. In turn, ‘segregation’ is the structure that such processes create or hold in place, a state in which information or its meaning is blocked, thinned, or distorted in such a way as to shape the activating or terminating conditions of a behavioural system. As Crittenden and Bretherton both observed, the exact nature of what is being blocked, thinned, or distorted, and by what agency, is not clear in Bowlby’s account of segregated systems.409 It would seem that segregation can be maintained by various processes. Among these, however, Bowlby gave defensive exclusion special importance. External perception offers the potential to gain from the opportunities and lessons of the environment. Without this, segregation becomes rigid, a situation with particular implications for mental health.

In unpublished notes from the time of Defences that Follow Loss: Causation and Function, Bowlby created a table to distinguish between four levels of exclusion (Table 1.2).410

Table 1.2 Selective exclusion of motivation and levels of motivation



Level 1: Low

Slight, e.g. daydreaming, free-association

Level 2: Moderate

Moderate. Some degree of concentration but still possible to attend to other Input

Level 3: High

Considerable. Strong concentration and resulting in exclusion of all other input

Level 4: Very high (or persistence of high)

Erratic. Whilst exclusion continues on a considerable scale, there is also a tendency for responses to a wider class of objects than in other conditions of motivation (cf Hinde)

Source: Theory of Defence, JB notes, 1960–63, PP/Bow/H10.

Level 1 is exclusion with only the barest drop of defence. It may, for instance, be little more than engrossment in something novel and/or complex, where the alternative would be rumination. Or it might be thinking about food or work when we might otherwise find (p. 74) ourselves getting upset, bringing one behavioural system partly online in order to exclude another a little. Or exclusion with a drop of defence might be identified in the absorption facilitative of endurance during exercise or sport. What Bowlby termed ‘the ordinary everyday things of life’,411 its familiar, conventionalised rhythms and language, can also be characterised as Level 1 defensive exclusion. They keep us connected to others but without demanding vulnerability or deeper availability to others—or to ourselves. ‘How are you?’ ‘Fine thanks, how are you?’ Such are the ‘outer rings’ of social stabilisation, Bowlby argued, that keep individual behavioural systems, such as the attachment system, from needing to be activated.412

Level 2 is described by Bowlby as ‘moderate’ exclusion. It is ‘still possible to attend to other input’. There is still some allowance, at least, for untidy, troubling experiences. Laugh-or-you’ll-cry humour might be offered as an illustration of this level of defensive exclusion, since the distress is walled off yet, if incipiently, personally and socially acknowledged. In Level 3 there is ‘exclusion of all other input’. The information permitted through the filter is rigorously and rigidly policed. Finally, at Level 4, ‘whilst exclusion continues on a considerable scale, there is also a tendency for responses [tobe made] to a wider class of objects than in other conditions of motivation’. The exclusion becomes extended beyond its original objects, for instance excluding not just perception but whole related classes of emotions as well. An example of this exclusion of a ‘wider class of objects’ is provided by a patient described by Bowlby in the chapter entitled ‘Psychoanalysis as Art and Science’, who until she was ten years old had ‘been terrified of another separation; but then she had “switched off” her anxiety “like a tap”, as she put it, and with the anxiety had disappeared most of her emotional life as well.’413

In Defences that Follow Loss: Causation and Function, Bowlby theorised that defensive exclusion and cognitive disconnection enact a weakening of the integration between or within behavioural systems. Defensive exclusion and cognitive disconnection may, then, contribute to coping when the alternative might otherwise be a greater or more enduring disruption to the system, or to the psychological apparatus that underpins and orchestrates behavioural responses more generally. This is why children are especially vulnerable to using segregative processes, because their psychological apparatus is still immature and less resilient to disruption. Defensive exclusion and cognitive disconnection, then, permit a certain kind of resilience in the face of disintegrative threats precisely by accepting some determinate and limited degree of segregation.

On this basis, Bowlby distinguished avoidance and dissociation as different intensities and kinds of defensive exclusion. Some kinds of avoidant response may involve little segregation, just perhaps a turning of attention away from a person, situation, or thing that might otherwise activate a behavioural system.414 There is some loss of information from reality, but not (p. 75) a loss of contact with reality. For instance, we may tune out small occurrences that might otherwise generate impatience or irritation, to keep the aggression system from coming online and disrupting our overall plans. This would be Level 1 in Bowlby’s taxonomy. It is also possible to have more thorough-going segregation with avoidance, as for instance when access to anger is muted—the threshold for activating the aggression system is raised—for an individual who was made to feel endangered when they displayed anger as a child. Muted access to anger would be Level 2; if anger is essentially unreachable, this would be Level 3.415

In the 1950s Robertson observed that withdrawn and avoidant behaviours shown by hospitalised children were often accompanied by disorientated behaviours. However, Bowlby treated dissociation as, in contrast to avoidance, a more extreme form of segregation—more of an emergency break.416 At a population level, forms of defensive exclusion such as dissociation may have evolved as a way to shut off and thereby protect behavioural systems or other psychological processes from extreme perceptions or memories that might otherwise risk a greater disruption. Instead of segregating certain information from a behavioural system, perception in general is segregated from other mental processes. He speculated that it would specifically be episodic information about occurrences within time and place that would be lost to dissociative processes; general attitudes might well remain unaffected.417 Bowlby proposed that such a process will produce responses that, unlike avoidance, are not responsive to the environment. This produces the phenomenon of ‘coming to’ following dissociation.418 The generalisation of segregation to all sense perceptions for a time suggests that dissociation would be considered by Bowlby as one kind of Level 4 defensive exclusion.

Displays of conflict

Yet defensive exclusion is not always successful. Bowlby found parallels between psychoanalytic theory and the work of ethologists, who had observed odd behaviour by animals experiencing a conflict of motivations, for instance when cued both to approach and to flee.419 (p. 76) Bowlby was fascinated that Hinde had discerned in his study of chaffinches that ‘paradoxically, strangeness evokes both escape and curiosity, and that there is a complex balance between the two competing response systems’.420 Likewise, Hinde observed that the following response show by young birds towards their parents may be disrupted or discoordinated where this competes with another response system, for instance if the parent is also in some way a cue for danger, not uncommon given ‘the broad range of stimuli eliciting both following and fleeing at this age’.421 In such situations, animals show ‘conflict behaviours’ such as the rapid transition between one tendency and the other, poorly coordinated combinations, freezing in place, misdirected movements, or signs of confusion or tension.422 Bowlby also considered that anger might be evoked by conflict between two behavioural tendencies, presumably since conflict can obstruct the achievement of one or both tendencies, contributing to frustration.423

Bowlby’s interest in conflict behaviour was likely primed by having observed several kinds of conflict response in his clinical work. For instance, as an army psychiatrist in World War II, working with veterans on their return from the front, he observed amnesias, loss of bodily control, misdirected and undirected behaviours, out of context anger or anxiety, signs of confusion, and tic-like behaviours or other signs of tension. Bowlby interpreted these behaviours as reflecting veterans’ experiences of psychological conflict between feelings of duty and lingering fear from their time at the front.424 Robertson also reported many of the same behaviours in his description of children during hospitalisation and on their return home. There were also discussions of conflict behaviour in the psychanalytic community. For instance, in 1956, Anna Freud described the case of a 13-month-old infant whose behaviour towards her mother was disrupted by screaming and states of withdrawal. Since these behaviours were only shown in the relationship with the mother, Freud anticipated that they suggested ‘some traumatic event’ in the history of this relationship, a hypothesis that was later confirmed.425

Bowlby came to believe that his concept of segregated systems could account for some forms of conflict behaviour, including when displayed towards attachment figures. On the one hand, conflict could occur between a behavioural system and its defensive exclusion, especially when information from perception or memory for the activation of a behavioural (p. 77) system is particularly salient. For instance, a child who is hurt may seek to exclude distressing information that might otherwise trigger the attachment system, if they have learnt that their caregiver will likely punish them for being upset or displaying a desire for comfort. They may ignore their injury, pretend it does not cause them distress, or keep their attention away from information about their caregiver’s whereabouts to avoid the desire to seek the caregiver. However, this exclusion will become increasingly precarious to the degree that the child’s injury is especially painful, is experienced as upsetting, or the caregiver is on hand. The result will be conflict between the attachment system and its inhibition, an approach–avoidance conflict, which may be visible in the child’s behaviour.

On the other hand, Bowlby was also interested in the forms of conflict that could arise when two behavioural systems were active at the same time. Often, of course, different behavioural systems are easily compatible. This is why, as Ainsworth repeatedly argued, it is a mistake to study behavioural systems in isolation: ‘In any situation, the extent to which attachment behaviour occurs depends not only on the strength to which it is activated, but also on the strength of activation of other competing or compatible behavioural systems.’426 A favourite example for both Bowlby and Ainsworth was that the fear system and the attachment system are often easily reconciled and integrated, in directing an infant away from a perceived threat and towards a haven of safety.427 Sometimes behavioural systems, whilst somewhat different in their demands, can be combined: for instance where courtship systems and wariness combine to produce flirting behaviour, or the anger and caregiving systems combine to produce harsh forms of discipline.

In other cases, however, Bowlby argued that behavioural systems may come into significant conflict, especially if there has been segregation of one or both, hindering communication and effective compromise.428 For example, conflicts may arise between the attachment and caregiving systems. The injured child discussed above may have caring responsibilities for a sick parent. This may be one of the reasons why the child feels that their attachment system should be inhibited where possible, since they have learnt that their parent is most responsive when they seem happy and available to meet the parent’s wishes.429 In such a case, however, pain or distress associated with the injury would come into conflict with the child’s caregiving system. The child may attempt to achieve the ends of both systems at once. The two systems do have some compatible behaviour, like physical approach to the caregiver. If the child attempts to maintain his caregiving role, this may be undermined or interrupted by distress from the injury. If the child lets the parent know about his injury and conveys his (p. 78) feelings of distress, this may nonetheless be inflected, even distorted, by the child’s habitual caregiving response and relative exclusion of the attachment system.

As a general principle, Bowlby suggested that ‘whenever a system that has been deactivated becomes in some degree active, such behaviour as is then shown is likely to be ill-organised and dysfunctional’.430 The smooth resolution of conflicts depends upon systems being able to communicate and compromise. When one behavioural system is active at the expense of a conflicting one that has been segregated, trapdoors are to be found inside our intentions and actions:

Not only are information and motor response relevant to any one goal narrowly restricted but information and motor responses relevant to some other and perhaps incompatible goal may be allowed through. It is as though an enquiry clerk, when asked about trains to Cornwall, gave information endlessly about the night express to Plymouth, with occasional intrusions about a plane to Rome.431

Psychoanalytic theory proposed that conflict between incompatible psychological demands is part of being human. However, the strength of these demands, the extent of their incompatibility, and other challenges may turn such conflict into a source of symptoms, in which one or both of these demands gets expressed in a distorted form. For instance, in a passage highlighted by Bowlby in his personal copy, Karl Abraham suggested that a conflict between the sexual response and its suppression could find outlet in the intensification of hunger: ‘The great frequency of hunger attacks in frigid women is very striking … Strong libidinal impulses, against the undisguised appearance of which consciousness protects itself, can be unusually well masked by a feeling of hunger. For hunger is a sensation that can be admitted to oneself and to others.’432

Bowlby proposed that conflict behaviours would become stabilised as symptoms of mental ill health to the extent that the situation that evoked the conflict was stable, and especially when it related to a close and enduring relationship. Signs of ‘tension, anxiety and depression’ may be especially expected to the degree that contradictory responses (or conflict between a response and its defensive exclusion) are felt towards an attachment figure.433 When access to attachment figures is uncertain, as may occur in conflict situations, this may evoke tension and anxiety. Conflict situations, when they are sustained, also block the satisfaction of at least one behavioural system (and perhaps more). This predicament can contribute to the onset and/or maintenance of depression by provoking a sense that the achievement of the set-goal is impossible, no matter what strategy is deployed.434

(p. 79) Bowlby anticipated that encounters with attachment figures that evoke conflict are powerful sources of anxiety and depression. He identified several reasons for this. Part of the importance of the attachment system in this regard is that it has implications for the individual’s basic sense of being intrinsically worthy, acceptable, and capable of being cherished. It therefore has a developmental role in informing the components of other behavioural systems. Another reason for the important role of conflict in relation to attachment figures in prompting anxiety and depression is that our attachment figures are often socially difficult to avoid, not only but especially in childhood. In the chapter ‘The making and breaking of affectional bonds’, Bowlby speculated that the intensity of emotion in close relationships may serve especially to activate early behavioural, cognitive, and affective components of behavioural systems, so that forgotten wishes and disappointments from childhood become incorporated into present-day behaviour and expectations.435 This was one aspect of what Bowlby termed the ‘risks of intimacy’.436 It is in good part in response to these risks that closeness gets so thoroughly modulated and ritualised in everyday family life. This reduces occasions that might risk disturbing the others’ composure by activating early and raw components of a behavioural system rather than the more mature and well-conditioned components.

Bowlby’s interest in the contribution of psychological conflict to depression and anxiety made him enthusiastic about the work of Aaron Beck, and the emergent paradigm of cognitive behavioural therapy. In Bowlby’s terms, to the degree that thoughts such as hopelessness in relationships, personal responsibility, and the unacceptability of the self become segregated from feedback, they would in turn reinforce depression and anxiety.437 In print, Bowlby stated that his model of depression and anxiety was ‘cast in the same mould’ as that of Beck, and that ‘his theory is compatible with mine’.438 He was especially impressed by the fact that Beck appeared to distinguish between two forms of depression—one anxious and needy, the other avoidant and with reduced affect—that corresponded to the two classes of response to separation that he and Robertson had identified.439 However, Bowlby felt that, more than Beck, attachment theory highlighted the importance of early experiences in contributing to psychological conflicts. In particular, the theory highlighted the relationship expectations and forms of defensive exclusion that could perpetuate conflicts even after the original situation evoking conflict had ceased.440

(p. 80) In the clinic

Mrs Q

In his last article, written with Mary Ainsworth, Bowlby highlighted especially his work as a clinician with children and families, and additionally the decades in which he ran a mother’s group in a wellbaby clinic, ‘learning much from his informal observations of mother–child interaction there’.441 Bowlby was mindful that his theory addressed general tendencies at the level of populations, and had only a probabilistic relationship with any individual, whereas his clinical practice addressed the concrete dynamics of individual lives, each with their particular vitality and equilibrium.442 In a sense, Bowlby observed in private notes, the clinician experiences the complex human being in front of his as like a work of art: ‘The criteria of a work of art are a) sensory and formal vitality (tension, life); and 2) order (balance, equilibrium, and symmetry) … What is true of a work of art is also true of a human personality.’443 Bowlby emphasised the need to distinguish between a clinical understanding of tension and order in the lives of individuals and the generalisations of diagnostic or research frameworks, and to establish a productive relationship between the two:

In the past there has been a deplorable tendency for the experimentalist to despise the clinician’s lack of precision and the clinician to reciprocate with contempt for the experimentalist’s lack of insight into human nature. Each has stoutly maintained that his own method was the one true way to knowledge. These claims are absurd: each method is indispensable. It is the clinician who usually has the earliest insights, defines the problem, and formulates the first hypotheses. But [through] the detailed minute study of the feelings and motivations of his patients, and the complicated intellectual and emotional repercussions to which they give rise, the clinical worker provides information regarding the relations of psychic and environmental forces which can be obtained in no other way.444

During Bowlby’s lifetime, there was already a good deal of discussion of the clinical implications of attachment theory.445 Since then, commentators have described an ‘explosion’ of texts offering exposition of Bowlby’s ideas for practitioner audiences.446 These works for (p. 81) practitioner audiences offer both exposition and revision of Bowlby’s ideas.447 However, close readings of Bowlby’s clinical remarks have been rare. One contributing factor may have been that the reader of Bowlby’s works is presented with his general reflections on his clinical experience, interspersed only occasionally with a paragraph describing a clinical case. These cases generally describe the childhood factors that predisposed a patient to mental health problems, and the precipitating adult context that led them to enter therapy. Bowlby’s own efforts, approaches, successes, and failures with patients are almost never reported. The reason appears to be, alongside Bowlby’s characteristic reserve, that unlike many psychoanalysts of his day he did not want to give details about his patients in publications without their permission.448 In the Bowlby Archive in the Wellcome Collection, Bowlby’s case files are embargoed until a century after the clinical work took place; as a result, for the most part, his earliest clinical cases will become available in around 2035.449

A marked exception, however, is the case of ‘Mrs Q’, which is discussed repeatedly by Bowlby in six different published treatments, and several unpublished manuscripts, over the span of decades, beginning from Defences that Follow Loss. Given its clear exceptionality, Bowlby presumably sought approval from the patient to discuss the case. In both Separation and Loss the case receives attention on multiple occasions. In lectures on the clinical implications of attachment research, across different countries, he positioned this case as paradigmatic.450 Putting together the jigsaw puzzle of Bowlby’s remarks on the case across his different writings offers an unusual opportunity to see his clinical work with a patient over time.451

Bowlby described his introduction to Mrs Q as follows:

Some years ago the doctor at a maternity and child welfare clinic asked me to see a little boy of 18 months who was not eating and was losing weight. His mother was intensely anxious and depressed and had been so since the boy’s birth. On enquiry I found that she was terrified lest her son die and was therefore pestering him to eat. She also told me that she had (p. 82) sometimes had impulses to throw the baby out of the window and to commit suicide. Only some months later did she tell me that on occasion she became hysterical, smashed the dishes and battered the baby’s pram.452

Bowlby characterised Mrs Q as ‘one of the most anxious patients I have ever treated’.453 Mrs Q described her commitment to giving her son a happy childhood, in contrast to her own. In many respects, Bowlby felt, she succeeded very well. However, Mrs Q was dismayed and confused by her own outbursts of violence, which seemed to coincide especially with occasions when her mother came over to visit.454 Over the first three years of analysis, Bowlby worked with Mrs Q to trace how her experiences as a child had been an important contributory factor to her adult anxiety and aggression. Her parents had been physically violent and threatened to kill one another. Her mother would also, at times, seek to coerce her husband and children by threatening to leave the family unless they did as she said.

Mrs Q experienced feelings of irrational distrust of Bowlby, which puzzled her. In the clinical work, Bowlby used these feelings as a basis for exploring Mrs Q’s experiences of attachment relationships. He gave particular attention to Mrs Q’s experiences of her mother’s threats of suicide: ‘Mrs Q. described how on two occasions she had returned from school to find her mother with her head in the gas oven and how at other times her mother would pretend to have deserted by disappearing for half a day. Naturally, Mrs Q. grew up terrified that if she did anything wrong her mother would go.’455 Additionally, Bowlby explored with Mrs Q how she had become scared of her own capacity for anger, which she had learnt to redirect ‘either towards herself or towards something which, or someone who, could not retaliate. When a child, Mrs Q. recalled, she retreated to her room and bit herself severely or attacked her dolls.’456

The clinical work was slow. Mrs Q ‘claimed for a long while not only that her feelings for her mother were of love, which was true since her mother had many good qualities, but that that must exclude hatred’.457 However, an important point in the therapy came when she acknowledged that, when she became angry with Stephen, ‘she said the most dreadful things, the very same things, in fact, that her mother had said to her when she was a girl’:

Once the facts were known it was possible to arrange some joint sessions with mother and son during which mother, with real regret, acknowledged making the threats and Stephen explained how terrified they made him. Mother assured Stephen that she would never do it really. All was not well thereafter, but recognition that Stephen’s fears were well based and an opening of communication between mother and son eased the situation.458

(p. 83) Over three years of therapeutic work, Bowlby helped Mrs Q come to acknowledge the anger she felt towards her mother. As this became less segregated, the violent outbursts towards Stephen reduced. Mrs Q came to understand that, as a child, there had been good reason for segregating her anger. Expressions of anger could otherwise provoke her mother to threaten suicide or to abandon her. Additionally, Mrs Q was strictly told not to let anyone else know about her parents’ behaviour, which contributed to the segregation by blocking opportunities for feedback and affirmation of her experiences.

Towards the end of the third year of therapeutic work Mrs Q’s father died unexpectedly, following an elective operation for cataracts. In the first years of therapy, Mrs Q’s account of her father had been ‘overtly negative’, hiding the positive feelings that she also felt towards him.459 Following her loss, Mrs Q became intensely depressed, had thoughts of suicide, and also described dissociative symptoms. Among these were anomalous beliefs about her father’s death and its cause:

During the weeks following her father’s death, she now told me, she had lived in the half-held conviction that the hospital had made a mistake in identity and that any day they would phone to say he was alive and ready to return home. Furthermore, she had felt specially angry with me because of a belief that, had I been available, I would have been able to exert an influence on the hospital and so enabled her to recover him.460

She experienced the sense that her father’s home had to be kept exactly as it was because he was very much alive and would be displeased to find anything changed when he returned.461 Bowlby was struck that Mrs Q appeared to have two distinct sets of thoughts and feelings in operation: one that led her to act as if her father was dead, and another that led her to act as if her father was alive.462 Yet neither organisation was strictly unconscious: Mrs Q could discuss her experience of holding both views. In Defences that Follow Loss, Bowlby observed that Mrs Q’s symptoms could not be explained using the conventional psychoanalytic concept of repression, which suggests a division between conscious and unconscious material. ‘Mrs Q. is, however, little different,’ he mused. Like with repression, ‘there is evidence of a psychic system with its accompanying affects and fantasies that is alien to the one with which we as analysts are in communication’. However, in contrast to repression, this system was not unconscious, but nonetheless blocked from communicating with other conscious systems. It is from this set of reflections on the case of Mrs Q that Bowlby, in fact, went on in the manuscript to use the concept of ‘segregated systems’ for the first time.463

Bowlby as therapist

Three key principles can be gleaned from the treatment of Mrs Q and more generally from Bowlby’s late writings about his approach to clinical practice. A first was the importance for Bowlby of the clinical transference: the behaviours, affects, and cognitions displayed by (p. 84) the patient towards the therapist, which may find their origin in the plans developed on the basis of earlier experiences of family and intimate relationships. In the 1960s, Bowlby had assumed that relationships with attachment components would be formed by children solely on the basis of familiarity with the caregiver. By contrast, in his late writings on therapeutic technique, he situated the successful longer-term therapeutic relationship for adult patients as one with attachment components, in the provision of a secure base.464 In this, he seems to acknowledge that attachment components could and should form in the relationship with the therapist, not merely due to familiarity, but due to other factors as well. Bowlby appeared to see something about the intimacy of communication within a therapeutic relationship contributing to its status as a relationship with attachment components. This activated the patient’s expectations about close relationships, and at the same time provided an opportunity to reflect on them.465

In the case of Bowlby’s clinical work with Mrs Q, the transference relationship was used to help the patient notice her difficulties around trust. It was on the basis of this recognition about her relationship with Bowlby as therapist that Mrs Q could then be supported to see that trusting a parental figure was a justifiable fear, given her early experiences. Bowlby stated that ‘my therapeutic approach is far from original,’466 and indeed a special attention to transference phenomena was common to the psychoanalysis of his day, especially for Klein and her followers. However, a significant point of contrast was a second principle of his therapeutic technique. Bowlby believed that the therapist’s task, where possible, is to address stable patterns of interaction within close relationships, rather than solely problems ‘inside’ particular individuals.467 Even though Bowlby tended to treat behavioural systems as properties of individuals, recovery from symptoms of mental ill health is rarely achieved without considering the resourcing of relevant systems, and the contexts and interpersonal meanings that calibrate their circumstances of activation and termination. The case of Mrs Q began with her son, who was not eating. Bowlby sought the origins and eliciting conditions of these symptoms in the meaning of the boy for his mother, Mrs Q, and in her behaviour towards him—and, in turn, the meaning of Mrs Q’s behaviours in her biographical and (p. 85) relational context. From the 1950s until the end of his life, Bowlby referred to himself as a ‘family psychiatrist’, since he put emphasis on seeing and helping the whole family, rather than only the member of the family showing the symptoms of a mental health problem.468

In his attention to the familial context, Bowlby felt that his approach was truly one that avoided blaming parents, but instead considered their actions, thoughts, and feelings in turn in a wider context of predisposing, triggering, and sustaining factors. In ‘The making and breaking of affectional bonds’, Bowlby argued that when connected to the history or situation that has predisposed or elicited it, much parental behaviour that might otherwise seem simply inexplicable comes in fact to make sense—either as a response to a truly distressing situation or as an attempt to avoid reacting to this situation.469 In a letter to Joan Stevenson-Hinde, he expressed his strong disagreement with the idea common in family therapy circles that ‘patterns of interaction has some particular purpose e.g. that of keeping the family together’. Of course, actions to keep a family together may be a conscious intention of individuals. But Bowlby was hostile to interpretations by therapists that symptoms shown by patients could be understood in terms only of their present function. He felt that such accounts were often generated when the therapist did not yet understand enough about the history of the family, its members, and interactions.470

A third principle in Bowlby’s late writings on technique was that no advice or guidance should be offered until it is clear that the patient or family is able to understand this as an attempt to be supportive rather than critical. Early in his career, Bowlby regarded himself as more non-directive as a therapist than his peers. However, in his late writings on clinical technique, he made more of a space for challenging and guiding patients and families. Such directive interventions were only recommended, though, on the condition that the patient understands that the therapist’s ‘concern is to help the patient review his own life, to look at his problems in his own way’.471 Bowlby was worried that unless therapy leads with support, the encounter may in fact harm the patient or family by reinforcing feelings of guilt and despair. Additionally, where the therapist is perceived as showing disregard for the world as the patient sees it, the potential for trust is eroded. The therapist should not be quick to cast their role as the ‘representative of reality’, contradicting patients’ own perspectives.472 As well as potentially harmful, Bowlby felt such approaches were ineffective at helping change to occur. (p. 86) In ‘Constructions in Analysis’, Freud wrote that ‘no damage is done if, for once in a way, we make a mistake and offer the patient a wrong construction as the probably historical truth’. Bowlby, in his marginalia, described Freud’s stance as ‘complacent’.473

Given the importance of laying the ground through support, Bowlby divided therapy into two ‘phases’.474 In the first, the therapist should primarily seek to offer companionable support, combined with open-ended exploratory questions. The focus in this first phase should be on present-day experiences and the wider social context, seeking to identify the kinds of situation that repeatedly tend to be difficult or cause problems for the patient or family. These are likely to be those that led to the initiation of therapy, though the patient may or may not be aware of the pattern. For instance, in cases where a behavioural system has been chronically suppressed (Level 3 defensive exclusion), the first task is for the patient to recover a sense of what it might mean for this system to be activated. Where the attachment system of a patient has been subject to full defensive exclusion, the therapist might explore current experiences outside of therapy and within the transference where this system would otherwise be activated, and the thoughts, feelings, and behaviours that occur instead.475 This entails slow, careful work to support patients to articulate the ecology of their ordinary lives, and the circumstances that prompt the activation or deactivation of behavioural systems and their associated affects.

Both the attachment system and the exploratory system may need to be coaxed online by this combination of support and attention to present-day experiences: the attachment system bringing fruitful material for the transference, and the exploratory system engaging and integrating this material. Past experiences will have already been at least implicit in the first phase, since both the present and our perceptions of it always hold the past within them. Yet, only once the patient or family feel supported to engage the exploratory system to address present-day experiences should the therapist, in the second phase, seek to explore the nature of these feelings in depth, and then consider past experiences that may have contributed to these difficulties. This should include helping the patient or family sort through their feelings about the past and present, to reduce the extent to which these contaminate one another and become confused when decisions are made in the present.476 In the case of Mrs Q, Bowlby described patiently supporting her towards recognition of the anger she felt towards her mother. The aspect of the therapeutic task which ultimately ‘can best wait is consideration of the past since its only relevance lies in the light it throws on the present’:

The sequence may often be for the therapist and patient, working together, first to recognise that the patient tends habitually to respond to a particular type of interpersonal situation in a certain self-defeating way, next to examine what kinds of feeling and expectation such situations commonly arouse in him, and only after that to consider whether he may (p. 87) have had experiences, recent or long past, which have contributed to his responding with those feelings and expectations.477

Bowlby offered a metaphor: ‘If a ball has gone down a dark passage, a child may be frightened to go there and get the ball, but if I say “Look, I will come with you”, he may be quite happy. In psychotherapy we act as a companion to a patient who is too frightened to look at what has happened to him in the past. So we accompany him in the exploration.’478 Bowlby believed that this companion and social point of reference is important because it gives patients the confidence to unlatch their personal hopes and fears from matters of fact, through close scrutiny of both. Bowlby clearly saw the discovery of historical truth in therapy as making some contribution to the reduction in symptoms, since it permitted more effective information-processing in the present.479 However, just as important was his emphasis on therapy as supporting curiosity and courage within and beyond the therapeutic setting.

Attachment disorders

Also relevant to his work as a clinician, Bowlby made a number of important remarks on the meaning and function of diagnosis. His overall position was that diagnosis is a relevant and valuable clinical tool for medical professionals, but that it should not be mistaken for explanation. From early in his career, he emphasised that development was multiply determined, and would therefore often exceed static characterisations that depicted particular symptoms as manifestations of specific disturbances.480 In an undated text, archived with material from around 1939 to 1940, Bowlby reflected carefully on the purposes of psychiatric classification in an attempt to develop diagnostic groupings for children under five:

Classification and diagnosis in child psychiatry is at present in a state of anarchy. Very few of the children seen correspond to any of the classical psychoneuroses or psychoses—the fact unjustly being simply “character-problems”. For these character cases there is as yet no good classification for adults, let alone for children … Even in adults it is sometimes difficult to distinguish clearly between the habitual personality and the particular syndrome of symptoms from which the patient is suffering. This difficulty is increased in childhood. Consequently the classification used here is only provisional.481

(p. 88) In his early book Personality and Mental Illness from 1940, Bowlby acknowledged that ‘few people remain the same throughout their lives’, and he explored this lack of continuity in the case of mental health symptoms. He pointed out that only rarely do individuals exactly fit the criteria of psychiatric classification, and generally only those with relatively less severe symptoms. Among patients with more severe problems, ‘many show at successive periods an unstable personality, symptoms of psycho-neurosis and of psychosis’. Bowlby therefore called it ‘absurd’ to regard such a patient ‘in terms of any one condition. He must be thought of as an individual of certain potentialities, a unity of which the particular traits and symptoms shown at any one moment are but fleeting expression.’482 Symptoms are not, in such an account, the manifest effect of an underlying and discrete disorder. For instance, low mood, sleeplessness, and a lack of energy are not the expressions of a disorder of ‘depression’, as distinct from other mental health problems. Rather, Bowlby’s account was transdiagnostic. He suggested that the individual represents a unity of potentialities, in this case towards mental health or ill health. Particular symptoms, or even clusters of symptoms that can be grouped as mental disorders, are best considered as consequences of this set of potentialities. He argued that, where early childhood experiences have been positive, this set of potentialities will predispose mental health. By contrast, early adverse experiences will support the intensification and expression of potentialities towards mental ill health across the lifespan.

In his mature thought, Bowlby developed these reflections further. In ‘Developmental psychiatry comes of age’, published 1988, he insisted that ‘a sharp distinction must be drawn between current functioning, measured in terms of presence or absence of psychiatric disorder and personality structure, measured in terms of greater or less vulnerability to adverse life events and situations’.483 Furthermore, ‘the features of personality to which we draw attention are different to those that most clinical instruments are designed to measure and not necessarily correlated with them’.484 Bowlby considered factors that might contribute to such transdiagnostic, developmental vulnerability. He was especially concerned with the way that the defensive exclusion of attachment-related motivations may deplete possibilities for experiencing life as ‘emotionally rich and varied’, and contribute variously to distress, aggression, and depression.485 Bowlby’s position implied that attachment processes can readily impact other areas of functioning, but this did not mean that the resulting behaviours are therefore to be considered as attachment.486 Though he was particularly concerned with attachment, Bowlby was clear that defensive exclusion of other behavioural systems could contribute to a vulnerability to mental illness. Defensive exclusion of the exploratory system could make it difficult for an individual to learn from experience, or from therapy.487 And loss of access to the prompts of the fear system could lead an individual to enter cruel and painful relationships. Without itself constituting a form of mental illness, Bowlby anticipated that such processes would contribute to the initiation or the maintenance of various mental health symptoms.

From Separation onwards, Bowlby emphasised that clinicians and researchers should note the importance of ‘developmental pathways’, self-reinforcing patterns in children’s trajectories (p. 89) towards or away from mental health. Bowlby was inspired by Waddington’s description of how cells can initially develop in a variety of different ways, but that once they do begin to develop, they become canalised such that a change from the established pathway requires greater and greater intervention.488 Bowlby reflected that, with human development conceptualised in terms of pathways, we can expect ‘adverse childhood experiences489 [to] have effects of at least two kinds. First they make the individual more vulnerable to later adverse experiences. Secondly they make it more likely that he or she will meet with further such experiences.’490 This pathways metaphor suggested a nuanced model of continuities over time:

All pathways are thought to start close together so that, initially, an individual has access to a large range of pathways along any one of which he might travel. The one chosen, it is held, turns at each and every stage of the journey on an interaction between the organism as it has developed up to that moment and the environment in which it then finds itself. Thus at conception development turns on interaction between the newly formed genome and the intrauterine environment; at birth it turns on interaction between the physiological constitution, including germinal mental structure, of the neonate and the family, or non-family, into which he is born; and at each age successively it turns on the personality structure then present and the family and, later, the wider social environments then current … As development proceeds and structures progressively differentiate, the number of pathways that remain open diminishes.491

This model implied for Bowlby that neither knowledge of general patterns nor knowledge of the specific case should be abandoned. Today, this might be discussed in terms of the distinction between ‘diagnosis’ and ‘formulation’.492 On an individual level, diagnostic (p. 90) classifications tell clinicians about what one child may have in common with other children with similar symptoms. They can help clinicians to develop an integrative account of the child’s difficulties, encourage more consistent care-planning, and encourage professionals to consider the treatments that may help. Formulation, on the other hand, is an individual approach that considers the ways in which the child is unique from others. Within formulation, assessment of the attachment system and its functioning serves as one part of nuanced thinking about a child’s behavioural profile, potential risk factors, and the role of the child’s external environment.

In 1980, the ‘Infancy, Childhood and Adolescent Disorders’ committee of the American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders (III) (DSM-III) introduced the category of ‘reactive attachment disorder in infancy’ as a recognised diagnosis. This diagnosis was an attempt to bring within medical assessment practice observations that had been made of specific forms of behaviour shown by institutionalised and former institutionalised children, influenced by Bowlby’s 1951 report to the World Health Organisation, as well as other reports.493 The disorder was, as it was initially proposed, to be diagnosed on the basis of weak infant physical growth, poor social responsiveness, and emotional apathy, as a consequence of grossly inadequate experiences of caregiving. In 1987, the diagnosis was revised to remove the physical growth criterion and adjust the age criterion, and has since seen other changes.494 It is notable that in the ten years between 1980 and his death in 1990, Bowlby made no public statement discussing this new official diagnosis ostensibly drawing inspiration from his work and using his headline concept of ‘attachment’. Some have treated this as implying that the diagnosis was simply a natural expression of Bowlby’s position and that he generally approved.495 However, Bowlby’s silence on the advent of ‘attachment disorders’ more likely reflected discomfort.

The introduction of the attachment disorder diagnosis as a problem of ‘attachment’ appears to have been based solely on Bowlby’s earliest writings on institutionalisation and his diffuse claims about its socioemotional implications, essentially prior to the development of attachment theory from the late 1950s. Justin Call, the primary member of the ‘Infancy, Childhood and Adolescent Disorders’ committee to discuss Bowlby’s work in print, justified the new diagnostic category in 1982 by arguing that ‘attachment disorders of infancy are characterised by the absence, disruption, or distortion of normally occurring developmental sequences of attachment behaviors’.496 Yet he gave no reference to Bowlby’s work after 1958. (p. 91) This would account for the fact that Call and the DSM interpret the ‘disruption’ or ‘distortion’ of attachment to mean primarily extreme deprivation. Call’s lack of familiarity with Bowlby’s work would also explain why the DSM-III initially limited attachment disorders to infants under eight months—when, in Attachment, Volume 1, Bowlby had argued that the attachment system is not likely formed until at least nine months. The DSM criteria displayed no concern to assess disruption to the attachment system, for instance in distinguishing between behaviour in contexts where the attachment system would be activated and behaviour in other contexts. The criteria also situated ‘attachment disorder’ as a property of the individual child, in contrast to Bowlby’s emphasis on the dyadic status of the attachment system in infancy.497

The DSM-III was a major event and sent shock-waves through the psychiatric establishment. Even if particulars were regarded as arguable, it appeared to offer the basis for a valid and reliable category-based diagnosis system, in which all symptoms find their logic and cause in an underlying disorder.498 However, the ‘attachment disorder’ diagnosis was not a part of this general upsurge of interest and support for diagnostic categorisation. No empirical studies used the category within research until after Bowlby’s death.499 If the idea of an ‘attachment disorder’ diagnosis without consideration of Bowlby’s writings on attachment theory had been put forward in another document, it seems probable that it would have been ignored. Certainly, Call’s chapter ‘Attachment Disorders in Infancy’ in the 1982 Comprehensive Textbook of Psychiatry, written to accompany the new DSM diagnosis, has rarely been cited.500 And in the next edition Call’s chapter was replaced by one that largely assimilated the new attachment disorder category into the frame of infant failure to thrive and psychosocial dwarfism.501

However, the diagnosis of ‘attachment disorder’ was hooked inside one of the single most influential medical documents of the twentieth century. The DSM-III contributed to a transformation in mental health assessment to give prominence to diagnosis and diagnostic pathways.502 Diagnostic categories have been made central to the delivery and administration of clinical services, to clinical resource allocation, and to clinical training. Throughout his career Bowlby was concerned with clean distinctions, where possible, between diagnostic activity and reflection on developmental processes. Furthermore, it is evident that his overall impulse was towards attention to developmental pathways in understanding children’s symptoms, and the use of diagnosis only for specific, limited purposes. Bowlby addressed these matters further in a paper presented to the British Psychological Society three years after the publication of the DSM-III (included as Chapter 4 in Trauma and Loss: Key Texts (p. 92) from the John Bowlby Archive). Bowlby denounced the focus of the psychological establishment on category-centric practice:

The categorists are still searching for diagnostic criteria that distinguish the mentally ill from the normal, though today their search is more likely to be for genetically determined biochemical anomalies than for any behavioural criterion. [On the other hand, there are] those others who, like myself, believe continuity to be a more fruitful perspective.503

Bowlby’s remarks to the British Psychological Society suggest that he would have endorsed Sroufe’s view that ‘the circumscribing of attachment problems to specific disorders reveals a failure to grasp the developmental significance of attachment history and the potential power of a developmental approach to psychopathology in general’ (Chapter 4).504 As we saw in the section ‘Disorientation’, Robertson documented behaviours that resembled the ‘attachment disorders’ in the children he observed during and following long-term hospitalisation. However, Bowlby decided not to publish his book with Robertson containing these observations, and did not integrate them into his subsequent theorizing. He also kept silent when the attachment disorder diagnosis was introduced into the DSM. These choices, in retrospect, appear to have helped initiate a split between widely circulating clinical discourses of attachment disorder and the bulk of empirical and theoretical work on attachment.

In the context of diagnosis-focused clinical practice, and continued observations of behaviours such as disinhibited social engagement by clinicians—particularly among those in adoption and fostering contexts—later psychiatrists interested in infant mental health, such as Charles Zeanah, acted as if that they had to work with and elaborate the diagnosis that was ‘on the books’.505 Likewise, the World Health Organisation’s International Classification of Diseases largely took on the disorder as characterised in the DSM. The Zero to Three manual introduced in 1994 for the assessment of the mental health and development of young children is an exception, with uses of attachment ideas in the manual informed by Bowlby’s later work and subsequent attachment research.506 The Zero to Three manual, however, is grossly subordinate to the DSM in status. It is rarely possible for American clinicians to seek recompense from insurance companies for their work using Zero to Three. The standard remains the DSM diagnoses.

The primary use of the DSM classification in the 1990s was to describe children reared in institutional care, where two profiles could sometimes be seen: withdrawal from all caregivers, and indiscriminate attachment behaviours shown towards caregivers and (p. 93) non-caregivers. Over subsequent years, the latter has been redescribed as disinhibited social engagement. In general, the relationship between the DSM disorder and the broader field of attachment theory and research has remained unclear, except the repeated observation that attachment disorder may be more likely when children have not had an opportunity to form a selective attachment or attachments.507 Some clinicians have argued for the narrowing of the attachment disorder to exclude disinhibited social engagement, since its relationship with the attachment system is unclear.508 Other clinicians have argued for a broader applicability for the concept of attachment disorder.509 However, mostly the established generation of attachment researchers, with a few exceptions such as Zeanah, Lyons-Ruth, and Spangler,510 have largely ignored the clinical diagnosis. Several attachment researchers, for instance Sroufe, van IJzendoorn, and Bakermans-Kranenburg, have expressed exasperation with the attachment disorder diagnosis, given its ill-fit with Bowlby’s attachment theory or the predominant categories used in mainstream attachment research.511 There is a sense in these writings that attachment researchers feel powerless to alter the diagnosis, now that it exists. Yet it is also not clear that they would want a diagnostic entity that better reflected attachment theory. In commentaries on the attachment disorder diagnosis by attachment researchers, attachment is frequently depicted as a dyadic phenomenon, incommensurate with individual-focused diagnostic systems like the DSM.512

The poor integration between clinical discourses of attachment disorder and the predominant focus of researchers on patterns of attachment has had several consequences. One has been the still relatively weak research base on disinhibited social engagement, and especially the correlates and implications of these behaviours when shown by home-reared (p. 94) children.513 A second consequence has been that, in the gap left between the clinical and research communities, the development of treatments for ‘attachment disorder’ have taken on a life of their own, very largely unmoored to research.514 Clinicians wishing to identify insecurity and attachment-related needs within a diagnosis-focused clinical culture have also been drawn to use of the ‘attachment disorder’ label.515 Such clinical uses align with, and may have been supported by, occasional use of the term ‘disordered attachments’ by researchers to mean simply problems in attachment relationships.516 Third, the disjointed relationship between clinical systems of diagnosis and the non-diagnostic categories of attachment has hindered the informed integration of attachment research with diagnosis-focused clinical services and clinical training. It can be especially hard for generalist clinicians to know what meaning attachment should have for their work, given that it figures as both a rare clinical diagnosis and a transdiagnostic developmental perspective.517 Commentators have warned, however, that slippage between the broad and circumscribed uses of the term ‘attachment disorder’ has contributed in some quarters to an overdiagnosis of attachment disorders, misuse of appeal to attachment disorder in psychological assessments for family courts, and neglect of children’s potential other psychological needs.518 The gap between clinical discourses and the research paradigm has also been filled at times by inappropriate uses of the disorganised attachment classification, forced to play the role of a quasi-diagnostic category in child welfare practice (Chapter 3).519

Some remaining questions


Three limitations of Bowlby’s theorizing can be identified as of special significance for subsequent attachment theory. A first limitation is Bowlby’s assumption that early adverse (p. 95) experiences would have strong continuities with later mental health, not only in the extent of symptoms but also in the kinds of symptoms shown. This reflected Bowlby’s residual but strong commitment to a psychoanalytic concept of ‘identification’ in which the parent is set up as a model for the child, together with influence from Bandura’s Social Learning Theory and clinical experiences: ‘I strongly suspect that the particular form of atypical care-eliciting behaviour selected by a patient is greatly influenced by modelling, the term introduced by Bandura, and roughly equivalent to identification, to describe adopting the same behaviour that one has observed engaged in by others … More and more in work with parents I have been struck by the extent to which they have adopted the same disciplinary procedures towards their children as they themselves were subjected to—often despite their wish to behave quite otherwise.’520

Bowlby felt that there are many factors that keep developmental pathways stable, or even intensify them, once they have begun to develop. Some of these are individual: Bowlby was especially interested in the way that children’s expectations, once shaped, guide their subsequent behaviour. He was confident that, shaped by past experiences, ‘present cognitive and behavioural structures determine what is perceived and what [is] ignored, how a new situation is construed, and what plan of action is likely to be constructed to deal with it’.521 Bowlby highlighted the role of social factors in strengthening and steadying developmental pathways: he had theorised that the effects of early long-term separations on later mental health occur at least in part as a result of the fact that the clingy and difficult behaviour children show following such separations may elicit rejecting or hostile responses from caregivers. Yet, beyond their independent influence, Bowlby felt that the reciprocal reinforcement of psychological and social processes was especially critical to the stability of developmental pathways: ‘Whatever family pressures led the development of a child to take the pathway he is now on are likely to persist and so to maintain development on that same pathway’,522 and ‘whatever expectations are developed during those years tend to persist relatively unchanged throughout the rest of life’.523

Throughout his career, Bowlby was attentive to factors that might mediate, buffer, or break continuity between early experiences and later mental health. Sroufe and colleagues are right that ‘Bowlby is not deterministic but, rather probabilistic’ in his claims.524 Nonetheless, Bowlby’s characterisation of how the attachment system changes in the context of maturation or changes in social environment remained underdeveloped, with the amorphous concept of ‘internal working models’ picking up the slack, but often not doing so effectively.525 He profoundly underspecified the relationship between attachment and personality traits: it (p. 96) is not clear whether attachment sets early personality, whether it forms the first point in a loose and interruptible chain of social experiences, whether it moderates the relationship between innate qualities and behaviour, or all three (Chapter 5).526 Implicitly contrasting her position to that of Bowlby, in 1988 Ainsworth reported that ‘increasingly, we are concerning ourselves with increasing our understanding of change, and with defining the conditions under which it takes place’.527 Later attachment researchers would find only partial support for the strength of Bowlby’s emphasis on continuity, and would, as Ainsworth hoped, have much to say about change (Chapter 4).

Furthermore, Bowlby’s overstrict opposition between ‘actual experiences’ and ‘fantasy’ led him to assume that experiences would be mirrored in symptoms, with adults ‘continuing to respond in social situations with the very same patterns of behavior that they had developed during early childhood’.528 So, for instance, in his 1944 ‘Forty four juvenile thieves’ paper he claimed that children who do not receive affection develop an incapacity for affection, which in turn contributes to their delinquency.529 And in ‘Violence in the family’, from 1984, he proposed that children who experience one kind of abuse will then show this same kind of abuse to their spouse and children.530 It is true that later research would support the idea that early experiences of neglect, violence, and highly unstable caregiving arrangements will shape developmental pathways and contribute towards the probability that an individual will develop mental health symptoms. However, Bowlby’s claim for a specific mirror between early experiences and later symptoms has not been supported. His claims in the 1940s linking early separations and stealing have not been supported. And later research has found that early experiences of abuse make later abusive behaviour more likely, but not necessarily the same form of abuse (Chapter 4).


A second consequential limitation of Bowlby’s work also stemmed from his overstrict distinction between ‘actual experience’ and ‘fantasy’: his underelaboration of the emotional (p. 97) components of behavioural systems. In his writings, Bowlby was clear that affects are important components of behavioural systems, and attention to feelings do emerge from time to time in his scholarly writings. He would often repeat the claim that ‘A person’s whole emotional life—the underlying tone of how he or she feels—is determined by the state of these long-term, committed relationships. As long as they are running smoothly the person is content; when they are threatened, he or she is anxious and perhaps angry; when the person has endangered them by his or her own actions the person feels guilty; when they are broken, the person feels sad; and when they are resumed he or she is joyful.’531 Such passages have sometimes been cited, especially by those seeking a model for therapeutic practice in attachment theory, to situate Bowlby as the quintessential theorist of human emotional life. However, such accounts depend upon anachronism. Johnson, for example, describes attachment for Bowlby as oriented by a striving for ‘felt security’ and ‘emotion regulation’, when in fact these represent the ideas of the later attachment researchers such as Sroufe, Waters, and Cassidy (Chapter 2, 3, and 4).532 In fact, Bowlby gives the causal role of emotions little space compared to the behavioural and cognitive aspects of his theory, and as a consequence his account is often oversimple. It is never quite clear what relationship Bowlby perceived between emotion and motivation, perhaps because he was often extrapolating across species from ethological models.533 Compared to Sroufe (Chapter 4), Bowlby generally offered little acknowledgement that moods and other felt states, such as confidence and fatigue, may arise or be held in place by multiple factors.534 The caregiving system is a clear example: it is sometimes hard to recognise the actual dynamics of caregiving in Bowlby’s characterisation of the caregiving behavioural system, for instance in how he neglects to give attention to both the worries and relief of being needed.

Bowlby highlighted the role of emotion only: (i) in modulating the activation and termination of behavioural systems; and (ii) as a consequence of the satisfaction or non-satisfaction of set-goals. Implicit in this is recognition that moods can offer us partially incompatible visions of the world, including of the viability and meaning of other emotions. Yet Bowlby rarely recognised, at least explicitly, the extent to which the fundamental shape and deployment of behavioural systems, including the calibration of their set-goal, can be conditioned by emotions. In his later writings especially, he came to consider the effects of depression and hopelessness. However, the emotional components of behavioural systems make them more varied and fragile than Bowlby generally acknowledged. This fragility can be seen with horrible clarity if we consider how a behavioural system looks when it becomes invested by shame or boredom, two emotions left aside by Bowlby (and, following him, by (p. 98) later empirical attachment research).535 Indeed, except in relation to outright conflict between behavioural systems, in general Bowlby was consistently poor in acknowledging and characterizing states in which motivation becomes half-hearted, bendy, or sleepy, states of velleity rather than concrete intention or outright inhibited action. He also did not consider the way that some affects simultaneously undermine or sustain different behavioural systems. For instance, excitement can facilitate the exploratory system, but qualifies fear. Even during his lifetime, albeit writing in German, Mary Main criticised Bowlby’s account of emotions for treating them as ‘precipitates’: an invisible part of a liquid that becomes a visible solid only after a chemical reaction. The metaphor suggests the role of emotion as a necessary part of behavioural systems, and actually integral to shaping their form—but as only visible, at least to Bowlby, at the end of the reaction as a ‘consequence’ of behavioural systems.536

In undated notes, likely from around 1955, Bowlby asked himself why he continually underplayed fantasy and emotion in his writing. He concluded that the ‘reason is partly because we think psychobiological reactions have been neglected by psychoanalysis & partly because research method has led this way. This is not to say that phantasy is unimportant—obviously it comes in & tends to complicate things still further. Roots of phantasy are in the psychobiological responses.’537 Bowlby evidently had a strong aversion to appeals to the idea of ‘fantasy’. Speaking in an interview in 1986, he observed that ‘nowadays, the word fantasy has become used in the analytic world to mean any cognitive process. It means conscious wish, it means a daydream, it means expectation.’538 He regarded this loose use of the concept as obscuring phenomena that he regarded as both important and distinct. He also did not like the way that the term ‘fantasy’ collapsed distinctions between different ways of relating to the future, such as expectation, wishing, and planning. Such distinctions, he felt, were important for acknowledging the accounts of patients, and maintaining a ‘neutral and empathetic position’ in relation to what they said about their experiences.539 In conversations with junior clinicians at the Tavistock, Bowlby used to ask that conversations of fantasy were ‘parked’, to see how or whether the matter in question could be addressed with greater precision in other ways.540

Bowlby’s opposition between ‘actual experience’ and ‘fantasy’ was criticised by Ainsworth in their correspondence.541 Ainsworth felt that Bowlby neglecting the emotional aspects of (p. 99) behavioural systems, ‘quite deliberately’, in order to avoid ceding ground to those trends in psychoanalytic thought that privileged fantasy over past experiences.542 In the 1980s, this motivation likely aligned with wider trends in cognitive science, in which thought and feeling were opposed, downplaying their interrelation.543 To give an example, in ‘Developmental psychiatry comes of age’ in 1988 Bowlby stated that ‘to Sigmund Freud is due the credit for having emphasised the influence on how people think, feel, and behave that is exerted by their internal world—namely, by the way they perceive, construe, and structure the events and situations they encounter’.544 However, Bowlby was well aware that how people ‘think, feel and behave’ is not equivalent to how people ‘perceive, construe and structure’. He equated them deliberately and strategically, to ensure that the cognitive aspect of behavioural systems, and the role of these in shaping emotions, would not be missed. The price, which appeared to be knowingly accepted by Bowlby, was that the role of emotion within behavioural systems, and the relationship between emotion and motivation, would remain hazy. This would leave problems for later attachment theory, for instance in how exactly to conceptualise affects at odds with expressed attachment behaviour (Chapter 3).

Predation and evolution

Bowlby’s familiarity with ethology and developments in evolutionary biology, by his own admission, dropped away sharply after the publication of Attachment, Volume 1. A consequence was that aspects of his theory remained partially trapped in amber, and especially his knowledge of developments in evolutionary biology.545 His account of attachment behaviour as a repertoire developed in human evolutionary history to sustain survival was a powerful cross-disciplinary integration of forms of knowledge, and an important plank in the emergence of developmental psychology as a subdiscipline informed by evolutionary theory. Yet Bowlby tended to rigidly emphasise that the evolutionary purpose of the attachment system was to save young children from predation by ensuring proximity to at least one adult. Downplayed in such an account, however, were the other evolutionary advantages of being physically close to a caregiver—advantages that Bowlby apparently knew. One such advantage is that closeness with the caregiver helps support psychobiological regulation. Bowlby stated this himself in ‘The nature of the child’s tie to his mother’ in 1958, citing with approval a discussion by Winnicott.546 The notion is also implied in Bowlby’s use of a thermostat as the central model or metaphor for the attachment system. Another evolutionary advantage of the attachment system is that proximity with a caregiver offers opportunities for nurturance (p. 100) and learning. In 1965, Ainsworth wrote to Bowlby arguing against the exclusive focus on predation as the evolutionary function of attachment, and giving the examples of nurturance and learning.547 These points were not taken up by Bowlby, except tacitly in his characterisation of attachment behaviour as directed towards someone ‘stronger and/or wiser’. However, it would appear from correspondence that he regarded ‘wiser’ to be less a description of the actual properties of an attachment figure, and more a cognitive bias in the perception of the attachment figure when the attachment system is activated.548

In 1970, whilst still a doctoral student with Ainsworth, Mary Main nonetheless took Bowlby’s side and argued that a desire for learning could not account for the clinging response as part of the attachment system. Opportunities for learning might be ‘correlates that “ride on” ’ the evolutionary function of proximity for protection.549 However, Main’s support for Bowlby’s position still offers no argument against physiological regulation and nurturance as possible evolutionary advantages of the attachment system. From the 1980s, physiological regulation and nurturance would be discussed by Myron Hofer as two ‘hidden regulators’ missed by Bowlby’s exclusive focus on the attachment system as protection against predation.550 Hofer claimed that Bowlby had ignored the ‘chronic’ advantages that physiological regulation and nurturance offer. The repetitive sequences of physiological regulation and nurturance may even play a role in constituting attachment as a coherently organised system, and infant separation anxiety may reflect a withdrawal response to these sources of regulation as much as a desire for protection. Furthermore, Hofer’s perspective implied that attachments are not formed solely on the basis of familiarity, as Bowlby generally suggested, but with familiar sources of biosocial homeostasis. Even an abusive caregiver can be an attachment figure, if they are familiar and provide sufficient physiological regulation and nurturance to be felt as a source of regulation.

Bowlby entirely accepted Hofer’s claims. In a warm letter to Hofer in 1983 he agreed on all points. Furthermore, he offered the notable claim that in humans as well as animals, ‘internal representations are usually a poor second best’ within the operation of behavioural systems, (p. 101) which likely much more ‘rely on sensorimotor pathways’.551 However, Bowlby did not integrate such insights into his subsequent published writings on the evolutionary function of the attachment behavioural system. A hint of a somewhat qualified position on the function of the attachment system can be seen only in Bowlby’s last publication, and even this is quite limited: ‘It contributes to the individual’s survival by keeping him or her in touch with one or more caregivers, thereby reducing the risk of harm, for example from cold, hunger or drowning and, in the human’s environment of evolutionary adaptedness, especially from predators.’552

Bowlby’s almost exclusive emphasis on proximity and predation would be criticised in turn by nearly all of the second generation of attachment researchers as part of marking their own emerging voices in the field. Despite this, in the decades following Bowlby’s death in 1990, it took a long while for the theory to receive even partial update in light of developments in evolutionary biology.553 It has not helped that ethology, the original alloparent of the attachment paradigm, has gone into decline since the 1990s, as interest shifted from behavioural sequences to other levels of analysis such as sociobiology and, more recently, gene–environment coevolution.554 However, the failure of further engagement with changing paradigms in biology has been fed by an internal dynamic within the field of attachment research: since Bowlby, authority about the biological function of behaviour has been part of leadership of the developmental tradition of attachment research, since in part it defines the nature of attachment as a research object and therefore the questions to be asked. With biological theory mixed up with authority and power, this dynamic has helped insulate attachment theory against a continued update of its models, in Bowlby’s era and subsequently. Bowlby’s ossified account of the relationship between attachment and evolutionary theory contained in the predation model would eventually lead to disagreements between Bowlby and Main, with lasting consequences for attachment as a research programme (Chapter 3). It would later also be significant ammunition for Peter Fonagy and Chloe Campbell’s call for a new paradigm to replace attachment research whilst absorbing its strengths (Chapter 6).


See Table 1.3 in the Appendix to this chapter for consideration of key concepts discussed here.

Table 1.3 Some key concepts in Bowlby’s writings


Mistaken for

Technical meaning

Attachment system

The instinctive relationship with a familiar caregiver

The ‘attachment system’ is a way of describing a form of motivation. The motivation is activated when a person is alarmed. When the person feels that a particular, familiar person—or familiar people—is available and responsive to their concerns, the motivation is reduced. Where the system is strongly activated, some form of contact is generally sought (though this contact may still be verbal rather than touch). This motivation has some basis in evolution, and for this reason is especially easy for humans to develop. However, a great deal about this motivation, including exactly the conditions that prompt and terminate it, are shaped deeply by experiences in relationships. This is why it is misleading to think of attachment as an ‘instinct’.

The attachment system has some characteristic behaviours, but in principle any behaviour can be recruited that helps achieve the goal of attachment figure availability.

Attachment researchers have debated the conditions that lead to the satisfaction of the attachment system and a reduction in display of attachment behaviour. The physical and attentional availability of a familiar caregiver has been emphasised as terminating conditions in infancy; other attachment researchers, since Sroufe and Waters, have emphasised the infant’s ‘felt security’ as the terminating conditions for the system.

In general it is agreed that experience, circumstance, and culture can all shape the conditions under which the attachment system is activated, the forms of behaviour recruited by the attachment system, and how these are expressed.

Attachment behaviour

Pre-set behaviours that express attachment as an instinct

Anything in principle can be an attachment behaviour. All that is required is that the behaviour should be clearly directed towards gaining the availability and responsiveness of a familiar person or familiar people. It is not an ‘instinctual’ pre-set pattern of behaviour.

Attachment behaviours are observable, whereas the motivation they are presumed to express is inferred.

The term ‘attachment behaviour’ was used in two different ways by Bowlby. His most common use of the term was to refer to proximity-seeking and contact-maintaining behaviours, such as smiling, crawling towards the caregiver, clinging, and directed cries to attract the caregiver’s attention. These behaviours were understood as direct expressions of the attachment behavioural system.

Sometimes, however, Bowlby also used the term ‘attachment behaviour’ to refer to any behaviour that occurs in the context of the activation of the attachment system. This could even include withdrawal from the caregiver and attempts at self-reliance by a child who had found that seeking proximity with their caregiver when alarmed or distressed would be counterproductive.

Often these two meanings are aligned. However, some behaviours, such as when a child shows caregiving behaviour towards a parent, are attachment behaviours in the second sense, but not the first sense. At times this has caused a lack of clarity in discussing such behaviour, its eliciting conditions, and its relationship with the attachment behavioural system.

Attachment bond

Parent–child bonding

Bowlby characterised relationships and their qualities as diverse. Attachment dynamics characterised only some relationships and not others. He therefore distinguished the broad class of affectional bonds, in which members are special to one another and seek to remain in contact. Within this broad class are relationships with attachment dynamics. These are characterised by the fact that the other person is taken to be the object of the attachment behavioural system: there is a disposition to seek this person under conditions of alarm, and a sense of security when this person is reliably available and responsive to concerns.

The attachment bond is distinct from ‘parent–child bonding’, the process by which parents develop an affectionate bond with their child and take the child as the familiar target of the caregiving behavioural system.

Attachment relationship

An absolute state characteristic of a child’s relationship with their mother

Being an attachment figure is not a yes/no situation. Bowlby proposed that an attachment relationship is present to the extent that an individual is disposed to seek the availability of a familiar other when alarmed. This disposition may exist even if the other is rejecting or abusive.

Bowlby felt that an individual could have a variety of attachment relationships—including wider kin (e.g. grandparents), divine beings, and also a person’s relationship with their physical home. However, he believed that evolution had primed humans to develop these dynamics especially with our familiar caregivers from childhood. Other relationships would be more contingent in the degree to which these dynamics would be expected.

Major separation

Occasions when the child and parent are not together

Attachment researchers sometimes discuss children experiencing ‘major separations’. This is a technical term, which can easily be confusing. What makes a separation ‘major’ is that the child is alarmed by the absence of their attachment figure, and this alarm continues for long enough that the attachment behavioural system then becomes chronically unresponsive for a long period. In effect, the child appears to give up searching for, calling, or expecting the parent to return. The result is that even when the caregiver is available, the child is not able to use them—at least for a time—to regulate distress. Or, in Bowlby’s terms, the behavioural system becomes chronically unresponsive for a period to cues for its activation and/or termination.

The classic case of a major separation was the long-term hospitalisations observed by James Robertson in the 1950s, in which there were no or few visits to young children over several months.

Attachment researchers do not regard some use of daycare as a ‘major separation’ in the technical sense.


The exclusivity and priority of child–mother attachment

Bowlby introduced the term ‘monotropy’ in 1958 with the intention that it would refer to particular, special relationships, shaped by time and habit. Unfortunately, the literal meaning of the term is ‘mono’ (one) + tropy (turning to). This gave the mistaken impression he meant the exclusive importance of one caregiver for children.

Bowlby later mostly abandoned the term, given the extent of misapprehension of his meaning.

Caregiving system

The natural capacity of parents, especially mothers, to care for their children

The ‘caregiving system’ is a way of describing a kind of motivation. A motivation to help is activated when a child or other person in our care is alarmed, and terminated when we have identified and responded to what we understand to be their concerns.

In his initial description of caregiving as a behavioural system, Bowlby focused on the caregiver’s motivation to retrieve infants who are alarmed or in trouble. However, later in his career he described caregiving as more broadly concerned with encouragement, support, help, and protection.

Effects of early experience

The notion that early social experience can be expected at an individual level to strongly determine later emotional and social experience

In his early writings, Bowlby sometimes made claims that implied that every child who receives poor care or who experiences major separations will develop social and emotional problems. From the 1970s onwards, he was more careful, claiming that—on average—poor care or major separations are likely to increase the chances of later social and emotional problems.

Later attachment researchers have synthesised findings from many studies through meta-analysis, indeed finding that early care does have effects on later socioemotional development, but that early experience does not determine later outcome, and that there are important mediators and moderators.

Internal working model

Representations of caregivers, which become generalised to all relationships with development

This is perhaps the single most confusing concept used by attachment researchers. Bowlby used the term in two different ways.

Firstly, he intended it only to mean that the way the attachment system works depends on expectations based on previous experiences of interaction with caregivers in childhood—and with partners and friends in adulthood. So a synonym for the internal working model, in this sense, in ordinary language is simply ‘expectations’. Bowlby’s point was that expectations about early relationships can play a role in shaping our assumptions about later social relationships and interactions.

Both humans and non-humans will have expectations about our caregivers or partners and their availability. However, humans also develop elaborated cognitive and cultural representations about ourselves and our attachment figures. These include narratives and images about the availability of attachment figures, and how we think they feel about us. A second use of the term ‘internal working model’ by Bowlby was therefore to refer to the specific symbolic and affective representations made by humans about attachment figures and their availability, and the efficacy of attempts to seek them when alarmed.

When Mary Main and colleagues introduced the Adult Attachment Interview (AAI) in the 1980s, they documented individual differences in the coherence of autobiographical accounts by participants of their childhood. Initially she referred to these differences in speakers’ narratives as reflecting differences in ‘internal working models’ about attachment. By the 1990s, she had abandoned and criticised the use of the term ‘internal working model’ to refer to these differences. Given the two different meanings above, she felt that the term was confusing and misleading for describing what the AAI was measuring. Main preferred to characterise individual differences in the AAI as reflecting ‘states of mind regarding attachment’. However, many attachment researchers still refer to the AAI as measuring internal working models.


The inhibition of information; essentially the same as dissociation

In Bowlby’s later writings the term ‘segregation’ is used to refer to a coping strategy in which some information is filtered out of experience. This can be minor and remain flexible: the filter can be raised or dropped as needed. Or the flow of information to or from whole behavioural systems can be blocked over a long period, regardless of the circumstances.

Bowlby distinguished two forms of segregation. A first was ‘defensive exclusion’. Here the filter is placed on perception. So, certain things in the world may not be noticed. Or if noticed, they may not prompt a response. The paradigmatic form of defensive exclusion is the infant in an avoidant attachment relationship, who directs attention away from their caregiver on reunion. In doing so, they filter out information about their situation that might otherwise prompt the activation of the attachment system.

A second form of segregation was ‘cognitive disconnection’. Here the filter is placed on memory. So, certain memories may not be available. Or if available, they may not be tagged with well-defined and accurate meanings.

The segregation of information about attachment figures was anticipated by Bowlby to contribute to an individual holding multiple incompatible perceptions and expectations of these figures.

Bowlby’s primary book on segregation, defensive exclusion, and cognitive disconnection remained unpublished. The terms appear only briefly in his published works. As a result, these terms are only used rarely now by attachment researchers.

Multiple attachments

Bowlby felt that a child should always be cared for by their mother

In his early writings Bowlby sometimes made claims that suggested a child should always be cared for by their mother. However, he subsequently regretted these claims. In his mature writing Bowlby saw value in a child having access to multiple secure bases and safe havens, and did not think that one attachment would be at the expense of another.

Bowlby’s final statement was that the attachment system ‘contributes to the individual’s survival by keeping him or her in touch with one or more caregivers’. The idea of attachment relationships as a network was developed by subsequent researchers such as Avi Sagi-Schwartz and Marinus van IJzendoorn.

Illustrative statement: ‘Bowlby argued that early experiences of care within monotropic attachment relationships, such as potential separations, contribute to the later integration or segregation of mental processes, and to the child’s own caregiving behaviours when they reach adulthood.’

Mistaken for: Bowlby argued that the quality of the mothers’ care, including separations such as maternal employment, will determine children’s later mental health and how they parent their own children.

Technical meaning: Bowlby argued that a child’s experiences in specific relationships with familiar caregivers will shape and calibrate the operation of the attachment behavioural system. This process may be disrupted, however, by major separations such as being hospitalised without visitation. Disruptions to the attachment system may influence the formation and operation of later mental processes, including coping strategies. Where disruptions lead to fixed or extreme distortions or blockages of attachment-relevant information, a predisposition to mental health problems may be anticipated—though this will be seen at a population level rather than in any given individual. Behaviours, affects, and cognitions that form components of the attachment behavioural system may influence the caregiving system when it develops, since elements may be inherited by or inform the latter system (e.g. expectations about what intimacy entails).

(p. 102) Sroufe argued that ‘as is always characteristic of development, whether in an individual or in a scientific field, Bowlby’s work both integrates and transforms what went before creating an alternative way of viewing the world without leaving behind critical insights contained in previous viewpoints’.555 Bowlby described himself as speaking a ‘hybrid, bastard language’, bringing together concepts from ethology and psychoanalysis, as well as terms from ordinary language.556 In creating this hybrid combination of discourses, Bowlby’s theory could benefit from the precision of Hinde’s conceptualisation of behavioural responses, their conditions and their conflicts; the emotional punch and drama of Kleinian theory of intense emotions and defences invested with his own personal and clinical experiences; and the explosive obviousness of the ordinary language connotations of ‘attachment’, ‘separation’, ‘loss’, ‘mother’, ‘love’, and others.557 Attachment theory was able to make an iconic assemblage of the precision, emotional power, and intuitiveness of these elements in ways that quickened the potential of each.558 Yet, in making an identity, it also retained qualities of these constituent differences. Various faces of this assemblage offered the theoretical basis for an empirical research paradigm, an approach to clinical work with children and adults, and a popular discourse regarding child development and family life.

If Bowlby had remained a theorist only attentive to the precisely defined ‘following response’, and had not plugged these ethological observations of behavioural sequences into the evocative world of psychoanalytic concerns, the theory would have been much less rich. Hinde himself would later concede this in reflections many years after Bowlby’s death.559 And if Bowlby had achieved his synthesis of ethology and psychoanalysis, but had communicated only in the language of these two disciplines, his ideas would have had a very limited audience. Bowlby’s appeals to ordinary language and cultural stereotypes in writing for popular forums helped set his theory alight; it glowed to widespread visibility, even as its qualifications and technical subtlety burned away as fuel. The manner of Bowlby’s popular writings helped create what Bourdieu termed ‘allodoxia’, a ‘light’, commodified version of a more complex cultural form, appealing to a wider base of constituents without the tools or means to access the original.560 This is in contrast to forms of popularisation that attempt to convey the intricacy and technical quality of the form, generally by unpacking terms and explaining in longhand or prioritizing the most essential elements, using an accessible style and stories to keep the audience engaged.561 Certainly not all works of popular science are allodoxia. What characterises allodoxia in psychology is the circulation of a simplified account of the human mind as if it had the same meaning as the technical account of empirical researchers.

(p. 103) All science has a ‘price for full entry’ in terms of theoretical and technical competence and commitment.562 This price may be paid by specialists through their training as part of formal qualifications or accreditation. Some or all access may also be gained by non-specialists, for instance through reading, depending on the structural barriers to entry and how much socialisation in tacit skills is required.563 In both cases there remains some recognition of the distinction between technical and non-technical use of concepts. By contrast, the accessibility of elements of Bowlby’s popularizing discourses from the 1950s and 1960s offered cut-price tickets to attachment theory, even though these only granted access to a fairly antiquated portion of the fairground, and—the source of much later trouble—looked much the same as full entry tickets. The cut-price popular discourse of attachment was evocative and underdetermined, as well as having the appearance of scientific credibility. This gave it flexibility, urgency, and reach for these diverse constituents concerned with speaking about the nature of family relationships and child development.564 The language of attachment theory can be experienced as running towards us, calling our name. This was deliberate. Both the circulation of half-truths about attachment and their rhetorical insistence can be regarded, in part, as a by-product of Bowlby’s marketing strategy for his scholarly thinking.565

In one of his final papers, Bowlby argued that science has two stages, drawing ideas from the philosopher of science Karl Popper. In the first stage, the task is to frame hypotheses. Where possible, this process benefits from ‘detailed and first-hand knowledge of the problem … together with a dose of intuition and imagination’.566 Raw materials such as ordinary language and clinical anecdote can be relevant elements for this stage of work, alongside all the other resources of the humanities. Bowlby, in fact, chided behaviourism for failing to see such resources as relevant, making the paradigm rather arid and, at times, ‘barren’ despite its rigour.567 Nonetheless, even the thin theory of behaviourism, he felt, was better than atheoretical research in psychology, since theory is necessary for the effective planning and interpretation of empirical research as a cumulative endeavour, as well as the design of interventions and public health policies.568 Looking back over his long career, Bowlby offered the self-criticism that his work had displayed an ‘absence of a follow-through’, remaining in the first stage of hypothesis-generation.569 It would only be with the work of Bowlby’s colleague Mary Ainsworth that attachment theory would become the basis for attachment research as an empirical research paradigm. Ainsworth’s contribution remained exploratory. It was also work in the context of hypothesis-generation, though it was sometimes misrecognised as hypothesis-testing research. However, Ainsworth’s development of procedures for measuring individual differences in attachment and related constructs would prove the bedrock of a hypothesis-testing tradition of attachment research within developmental science.

(p. 104) Appendix


1 Bowlby, J. (1981) Perspective: a contribution by John Bowlby. Bulletin of the Royal College of Psychiatrists, 5(1), 2–4, p.2.

2 Bowlby, J. (1979) The ten books which have most influenced my thought, 24 October 1979. PP/Bow/A.1/8.

3 Bowlby, J. (1969) Anxiety, Stress and Homeostasis. Unpublished manuscript, April 1969. PP/Bow/H10.

4 Van Dijken, S. (1998) John Bowlby: His Early Life: A Biographical Journey into the Roots of Attachment Theory. London: Free Association Books; Bowlby, U. (1992) A memoir of John. PP/Bow/P.6/3: John believed that ‘his own childhood had been sufficiently unhappy to want him to investigate—but not so unhappy that he had obliterated the subject’.

5 Bowlby, J. (1984) Letter to Phyllis Grosskurth, amending discussions of Bowlby in Grosskurth’s biography of Klein, 10 January 1984. PP/Bow/A.5/7.

6 Bakhtin, M.M. (1981) The Dialogic Imagination: Four Essays by M.M. Bakhtin. Austin: University of Texas Press; Showers, C.J. & Zeigler-Hill, V. (2003) Organization of self-knowledge: features, functions, and flexibility. In M.R. Leary & J.P. Tangney (eds) Handbook of Self and Identity (pp.47–67). New York: Guilford.

7 Bowlby, J. (1953) Child Care and the Growth of Love. Harmondsworth: Pelican.

8 See e.g. Bowlby, J. (1954) Should a baby be left to cry? Parents, March 1954, pp.32–35; Bowlby, J. (1958) Should mothers of young children work? Ladies Home Journal (November) 75, 58–59, 158–61.

9 E.g. Association for Psychiatric Social Workers (1955) Presentation at the Annual General Meeting 1955: Dr John Bowlby on preventative activities. Modern Records Centre Warwick University. MSS.378/APSW/P/16/6/19-20.

10 Bowlby, J. (1986) Interview with the BBC. PP/Bow/F.5/8: ‘I published this report for the World Health Organisation in 1951, Maternal Care and Mental Health … all the evidence was still sketchy, it was inadequate.’

11 Bowlby, J. (1953) Child Care and the Growth of Love. Harmondsworth: Pelican, p.53, 76.

12 Ross, L.R. (2014) Reading Ursula Bowlby’s letters (1939–1940). Journal of the Motherhood Initiative for Research & Community Involvement, 5(1), 67–82.

13 Doyle, C. (1987) A continuing case for keeping children at home. Daily Telegraph, 23 June 1987: ‘He insists that he has never intended to imply that a continuous relationship should mean every minute of the day, and now adds “intermittent” to “warm and continuous” ’ as the qualities of care he wished children to receive from their familiar caregivers.

14 Bowlby less fell into the naturalistic fallacy than emblazoned it on his shield. Bowlby, J. (1990, 2011) John Bowlby: interview by Leonardo Tondo. Clinical Neuropsychiatry, 8(2), 159–71: Tondo—‘Do you agree with the rather simplistic view that the mother may be more important than the father? Bowlby— ‘That view, I think, is well attested by the information we have … This is the way all societies operate. So my concern is always with human nature, about which I am most confident about Western culture. When I teach my students, I say, “Look, the first thing to remember is that Western society is not a human norm.” We behave in a way that human societies have never behaved in the past. If you take human societies over the past hundred thousand years so far as we know and around the world, Western societies are peculiar. We do things in funny ways which may be alright, and it may not. Do not think they are normal. They are not the normal way human beings are meant to behave” … You either go along with human nature or you fight it. If you fight it you get problems. If you don’t fight it life is much more comfortable.’ (164–5)

15 E.g. Franzblau, S.H. (1999) Historicizing attachment theory: binding the ties that bind. Feminism & Psychology, 9(1), 22–31; Vicedo, M. (2011) The social nature of the mother’s tie to her child: John Bowlby’s theory of attachment in post-war America. British Journal for the History of Science, 44(3), 401–26. Van der Horst has also situated Bowlby’s work in the wider context of post-war Europe, especially in the context of his travels in the 1950s. Van der Horst, F.C.P. (2011) John Bowlby—From Psychoanalysis to Ethology. Unravelling the Roots of Attachment Theory. Oxford: Blackwell.

16 Bowlby, J. (1958) Should mothers of young children work? Ladies Home Journal (November) 75, 58–59, 158–61, p.158. LeVine has claimed that Bowlby had little awareness of the extent to which he was engaging with contemporary ideologies of family life. Careful examination of the full range of Bowlby’s public and private writings indicate that the ideological aspects of Bowlby’s popular writings were not simply the result of a lack of self-awareness, but had a decidedly strategic component. LeVine, R.A. (2014) Attachment theory as cultural ideology. In H. Otto & H. Keller (eds) Different Faces of Attachment: Cultural Variations on a Universal Human Need (pp.50–65). Cambridge: Cambridge University Press. In sociological perspective, Bowlby’s appeal to existing concerns and prejudices of his day may be considered in light of Bourdieu’s remark that when attempting to reach a mass market, the ‘more directly and completely’ must cultural producers direct their goods to ‘a pre-existing demand, i.e. to pre-existent interests in established forms’. Bourdieu, P. (1993) The Field of Cultural Production. Cambridge: Polity, p.97.

17 Bowlby, J. (1987) Baby love. Hampstead and Highgate Express, 3 April 1987.

18 For example, Alan Sroufe (Chapter 4) wrote a popular textbook and a handful of newspaper articles primarily focused on the overprescription of ADHD medication. But he did not attempt in a sustained way to speak to wider publics in the manner of Bowlby. In Germany, Klaus and Karin Grossmann were more active in attempting to speak to wider publics and increase public understanding of attachment. See e.g. Grossmann, K. & Grossmann, K. (2011) Das Geflecht des Lebens. DVD, Auditorium Netzwerk: Freigegeben ohne Altersbeschränkung. Perhaps the closest to Bowlby in direct engagement with policy-makers and publics has been Peter Fonagy, e.g. Fonagy, P. (2018) Evidence submitted to the Evidence-Based Early-Years Intervention Inquiry. Science and Technology Committee (Commons).

19 E.g. Riley, D. (1983) War in the Nursery: Theories of the Child and Mother. London: Virago Press; Thomson, M. (2013) Lost Freedom: The Landscape of the Child and the British Post-War Settlement. Oxford: Oxford University Press, p.87.

20 Winnicott, D. (1954, 1987) Letter to Dr J. Bowlby. In F. Robert Rodman (ed.) The Spontaneous Gesture: Selected Letters of D.W. Winnicott. London: Karnac, pp.65–66. See also Lewis, J. (2013) The failure to expand childcare provision and to develop a comprehensive childcare policy in Britain during the 1960s and 1970s. Twentieth Century British History, 24(2), 249–74.

21 Winnicott, D. (1953, 1989) John Bowlby. In C. Winnicott, R. Shepherd, & M. Davis (eds) Psycho-analytic Explorations (pp.423–32). London: Karnac, p.427.

22 E.g. Ringold, E.S. (1965) Bringing up baby in Britain. New York Times, 13 June 1965: “Dr Bowlby, however, sticks to his guns. Interviewed in his office at the Tavistock clinic he said: “Whenever I hear the issue of maternal deprivation being discussed, I find two groups with a vested interest in shooting down the theory. The Communists are one, for the obvious reason that they need their women at work and thus their children must be cared for by others. The professional women are the second group. They have, in fact, neglected their families. But it’s the last thing they want to admit.’

23 Bowlby, J. (1976) Bowlby on latch-key kids: interviews with Dr Nicholas Tucker. Psychology Today, Autumn 1976, 37–41, p.38.

24 Bowlby, J. (1971) Letter to Michael Rutter, 6 October 1971. PP/Bow/J.9/161.

25 Annotations by Bowlby (PP/Bow/J.9/162) on Rutter, M. (1981) Social-emotional consequences of day care for preschool children. American Journal of Orthopsychiatry, 51(1), 4–28.

26 Bowlby’s scientific and clinical writings are placed together since the tone and arguments are quite well integrated; there are fewer disparities than with Bowlby’s popular writings or his private reflections. Bowlby’s routine work as a clinician, and clinical notes, likely represent a further distinct Bowlby. However, there are few available textual traces, except insofar as this work contributed to Bowlby’s scholarship or his private reflections. Most of Bowlby’s clinical notes are closed still for several decades in the Wellcome Collections.

27 Duniec, E. & Raz, M. (2011) Vitamins for the soul: John Bowlby’s thesis of maternal deprivation, biomedical metaphors and the deficiency model of disease. History of Psychiatry, 22(1) 93–107.

28 Bowlby, J. (1940) Personality and Mental Illness. London: Kegan Paul; Bowlby, J. (1957) An ethological approach to research on child development. British Journal of Medical Psychology, 30(4), 230–40; Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin.

29 Ibid. pp.130–31.

30 A partial exception is Vicedo, M. (2011) The social nature of the mother’s tie to her child: John Bowlby’s theory of attachment in post-war America. British Journal for the History of Science, 44(3), pp.401–26, where there is valuable analysis of links between Bowlby’s popular and academic writings of the 1950s.

31 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, pp.303–4.

32 Bowlby, J. (1971) Letter to Michael Rutter, 6 October 1971. PP/Bow/J.9/161.

33 Melanie Klein had set a trend in which the word ‘mother’ (or ‘breast’) was deployed as a synecdoche for the infant’s experience of the caregiving environment in general, rather than referring specifically to the biological mother. See Hinshelwood, R.D. (1989) A Dictionary of Kleinion Thought. London: Free Association Books.

34 Bowlby, J. (1982) A Secure Base. London: Routledge: ‘All knowledge is conjectural and … science progresses through new theories coming to replace older ones when it becomes clear that a new theory is able to make sense of a greater circle of phenomena than are comprehended and explained by an older one and is able to predict new phenomena more accurately.’ (84)

35 Bowlby, J. (c.1982) Popper’s evolutionary epistemology. PP/Bow/H.98. The passage is heavily underlined in red pen for emphasis.

36 Ursula Bowlby, cited in Karen, R. (1998) Becoming Attached: First Relationships and How They Shape Our Capacity to Love. Oxford: Oxford University Press, p.29.

37 Duschinsky, R. & White, K. (eds) (2019) Trauma and Loss: Key Texts from the John Bowlby Archive. London: Routledge.

38 Bowlby, J. (1933–36) Research notes for uncompleted PhD. PP/Bow/D.1/2/13: ‘Zuckerman describes how monkeys will mutilate themselves if they are alone in their cages and visitors make them angry … Although not in visitor-cages, humans are often in cages of inhibition.’

39 Bowlby, cited in Tanner, J.M. & Inhelder, B. (eds) (1958) Discussions on Child Development: Proceedings of the WTO Study Group of the Psychobiological Development of the Child, Vol. 3. London: Tavistock, p.207. Use of the first person plural (‘the main problem with which we are all faced’) is sufficiently unusual to permit a biographical interpretation, especially given the salience of the theme of integration for Bowlby. It should be acknowledged, of course, that this could have simply been a manner of speaking.

40 Richard Bowlby, personal communication, February 2019. An exception is Bowlby, J. (1974) A guide to the perplexed parent. New York Times, 2 March 1974.

41 E.g. Bowlby, J. (1976) Bowlby on latch-key kids: interviews with Dr Nicholas Tucker. Psychology Today, Autumn 1976, 37–41.

42 E.g. Doyle, C. (1987) A continuing case for keeping children at home. Daily Telegraph, 23 June 1987. Careful examination of the interview signals shifts in Bowlby’s thinking—even his regret in his use of the term ‘continuous’ care by mothers in his early writings. Yet the dominant narrative remains remarkably similar to the early popular works, and there is no attempt to explain theoretical developments. Bowlby’s lack of explicit attempt to clarify revisions to his picture of child–caregiver relationships was likely supported by a lack of interest in such developments in media forums, where reference to basic stereotypes about mothers and children made for more accessible reading.

43 Tanner, J.M. & Inhelder, B. (eds) (1958) Discussions on Child Development: Proceedings of the WTO Study Group of the Psychobiological Development of the Child, Vol. 3. London: Tavistock, p.208. Some legacy of Klein may be felt here, in the idea of forgiveness as psychological integration.

44 Burlingham, D. & Freud, A. (1942) Young Children in War-time. Oxford: Allen & Unwin.

45 Annotations by Bowlby dated 1942 on Burlingham, D. & Freud, A. (1942) Young Children in War-time. Oxford: Allen & Unwin, pp.10, 47. Copy held in the library of Human Development Scotland. Earlier than Burlingham and Freud, Ian Suttie had written that ‘instead of an armament of instincts—latent or otherwise—the child is born with a simple attachment-to-mother who is the sole source of food and protection … the need for a mother is primarily presented to the child mind as a need for company and as a discomfort in isolation’. Suttie I. (1935) The Origins of Love and Hate., London: Free Association Books, p.15. It is possible that the use of the term came to Bowlby via Suttie and/or conversations through the Tavistock or the wider London psychoanalytic scene. In any case, no textual record is available of Bowlby having read Suttie until after Anna Freud. On the link between Bowlby and Suttie see van der Horst, F.C. & van der Veer, R. (2010) The ontogeny of an idea: John Bowlby and contemporaries on mother–child separation. History of Psychology, 13(1), 25–45.

46 Bowlby, J. (1958) Nature of the child’s tie to his mother. International Journal of Psycho-Analysis, 39, 350–73. Bowlby’s first published use of the term ‘attachment’ would appear in Bowlby, J., Ainsworth, M., Boston, M., & Rosenbluth, D. (1956) The effects of mother–child separation: a follow-up study. British Journal of Medical Psychology, 29, 211–47, p.237.

47 Bowlby, J. (1985, 1991) The role of the psychotherapist’s personal resources in the treatment situation. Bulletin of the British Psychoanalytic Society, 27(11), 26–30. Published as Chapter 12 in Duschinsky, R. & White, K. (eds) (2019) Trauma and Loss: Key Texts from the John Bowlby Archive. London: Routledge.

48 Lay, K.L., Waters, E., Posada, G., & Ridgeway, D. (1995) Attachment security, affect regulation, and defensive responses to mood induction. Monographs of the Society for Research in Child Development, 60(2–3), 179–96, p.179. On continuities between attachment and psychoanalysis see also Eagle, M. (1995) The developmental perspectives of attachment and psychoanalytic theories. In S. Goldberg, R. Muir, & J. Kerr (eds) Attachment Theory: Social, Developmental & Clinical Perspectives (pp.123–50). Hillsdale, NJ: Analytic Press.

49 Bowlby, J. (1933–36) Research notes for MD thesis: ‘Anxiety— Essays’. PP/Bow/D.2/46/6; Bowlby, J. (1981) Perspective: a contribution by John Bowlby. Bulletin of the Royal College of Psychiatrists, 5(1), 2–4.

50 Bowlby, J. (c.1933) Dreams. PP/Bow/D.2/45/7.

51 Bowlby, J. (1949) The study and reduction of group tensions within the family. Human Relations, 2, 123–28.

52 Bowlby, J. (1984) Letter to Phyllis Grosskurth, amending discussions of Bowlby in Grosskurth’s biography of Klein, 10 January 1984. PP/Bow/A.5/7. Bowlby requested that Grosskurth add to her biography of Klein the following passage: ‘It was fortunate for Bowlby that the war then intervened. He tells people that it saved him from open conflict with Melanie Klein for which he would not then have been ready.’

53 See Gardner, H. (1986) The Mind’s New Science: A History of the Cognitive Revolution. New York: Basic Books.

54 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.353. Internal working models are ‘none other than the internal worlds of traditional psychoanalytic theory seen in a new perspective’ (82).

55 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.82. Bretherton, I. (1998) Internal working models and communication in attachment relationships. In A. Braconnier & J. Sipos (eds) Le Bébé et les Interactions Précoces (pp.79–90). Paris: Presses Universitaires de France: ‘I would urge, however, that the various metaphors Bowlby … used as tools to think about internal working models are not to be taken too literally.’ (79)

56 See e.g. Crittenden, P.M. (1990) Internal representational models of attachment relationships. Infant Mental Health Journal, 11(3), 259–77; Bartholomew, K. (1990) Avoidance of intimacy: an attachment perspective. Journal of Social and Personal relationships, 7(2), 147–78; Main, M. (1991) Metacognitive knowledge, metacognitive monitoring, and singular (coherent) vs. multiple (incoherent) models of attachment: some findings and some directions for future research. In P. Marris, J. Stevenson-Hinde, & C. Parkes (eds) Attachment Across the Life Cycle (pp.127–59). London: Routledge; Collins, N.L. & Read, S.J. (1994) Cognitive representations of attachment: the structure and function of working models. In K. Bartholomew & D. Perlman (eds) Advances in Personal Relationships, Vol. 5 (pp.53–90). London: Jessica Kingsley; Waters, H.S. & Waters, E. (2006) The attachment working models concept: among other things, we build script-like representations of secure base experiences. Attachment & Human Development, 8(3), 185–97.

57 Bowlby, J. (1988) Developmental psychiatry comes of age. American Journal of Psychiatry, 145, 1–10.

58 Mikulincer, M. & Shaver, P.R. (2003) The attachment behavioral system in adulthood: activation, psychodynamics, and interpersonal processes. In M.P. Zanna (ed.) Advances in Experimental Social Psychology, Vol. 35 (pp.53–152). New York: Academic Press, p.60. The idea of working models as changeable, provisional representations is likewise a central theme in many more applied works making appeal to Bowlby, e.g. Johnson, S.M. (2019) Attachment Theory in Practice. New York: Guilford.

59 Alan Sroufe, personal communication, January 2019: ‘Bowlby never used “working model” to mean provisional. I asked him explicitly because I kind of liked that idea. He said no, he didn’t mean that.’

60 Fonagy, P. (1999) Points of contact and divergence between psychoanalytic and attachment theories: is psychoanalytic theory truly different. Psychoanalytic Inquiry, 19(4), 448–80; Eagle, M. (2013) Attachment and Psychoanalysis. New York: Guilford.

61 Scharff, D.E. & Fairbairn Birtles, E. (1997) From instinct to self: the evolution and implications of W.R.D. Fairbairn’s theory of object relations. International Journal of Psycho-Analysis, 78, 1085–103: ‘Bowlby specifically acknowledged his Fairbairnian orientation in the development of attachment theory and the ethological approach to infant development (personal communication)’ (1100). Bowlby’s annotations on Fairbairn’s writings offer testament to his alignment and agreement with the latter’s positions. For instance, he wrote ‘crucial points’ in the margins when Fairbairn emphasized that even a parent perceived as unkind may still be the person a child will want to turn to for comfort. Annotations by Bowlby on Fairbairn, W.R.D. (1944) Endopsychic structure considered in terms of object-relationships. International Journal of Psycho-Analysis, 25, 70–92. Bowlby’s copy of this paper is held by Richard and Xenia Bowlby.

62 Anna Freud urged recognition that the idea of the ‘pleasure principle’ in psychoanalytic theory was not opposed to the idea of attachment; all it posited was that when impulses or responses were activated for a child, a homeostatic response would be initiated to reduce the feeling of tension or motivation, especially in ways that provide pleasure or comfort. This was therefore aligned with, rather than contrary to, the idea of attachment as a homeostatic system. Freud, A. (1958, 1960, 1969) Discussion of John Bowlby’s work on separation, grief and mourning. In A. Freud & D.T. Burlingham (eds) The Writings of Anna Freud, Vol. 5 (pp.167–86). New York: International Universities Press. Press. Blum’s statement that ‘Bowlby’s ideas angered just about everyone he knew. Anna Freud dismissed him outright’ is a thoroughgoing oversimplification, based on selective reading. Blum, D. (2002) Love at Goon Park: Harry Harlow and the Science of Affection. New York: Basic Books, p.59.

63 Bowlby, J. (undated) Maternal behaviour: humans. PP/Bow/H.136; Bowlby, J. (1979) On knowing what you are not supposed to know and feeling what you are not supposed to feel. Canadian Journal of Psychiatry, 24(5), 403–8.

64 This was the topic of his first published paper: Bowlby, J. (1940) The influence of early environment in the development of neurosis and neurotic character. International Journal of Psycho-Analysis, 21, 154–78: ‘Every patient who comes to us has a distorted view of his parents … Some patients will project all that they feel to be bad in themselves on to their parents and blame and hate their parents. Others will project all the good and idolize their parents’ (176). Such reflections on two classes of distortion in adults’ perceptions of their parents align with the later position of Mary Main (Chapter 3).

65 Bowlby, J. (1977, 1979) The making and breaking of affectional bonds. In The Making and Breaking of Affectional Bonds (pp.150–88). London: Routledge.

66 See e.g. Isaacs, S. (1944) Letter to Major Bowlby, 20 October 1944. PP/Bow/J.9/111.

67 Fonagy, P. (2015) Mutual regulation, mentalization, and therapeutic action: a reflection on the contributions of Ed Tronick to developmental and psychotherapeutic thinking. Psychoanalytic Inquiry, 35(4), 355–69.

68 Target, M. & Fonagy, P. (2003) Attachment theory and long-term psychoanalytic outcome: are insecure attachment narratives less accurate? In M. Leuzinger-Bohleber, A.U. Dreher, & J. Canestri (eds) Pluralism and Unity? Methods of Research in Psychoanalysis (pp.149–67). London: International Psychoanalytical Association, p.163.

69 Fonagy attempted to spell out clearly some of the elements Bowlby absorbed within the concept of ‘internal working model’: ‘Four representational systems compose the internal working model (IWM): (1) expectations of interactive attributes of early caregivers created in the first year of life and subsequently elaborated; (2) event representations by which general and specific memories of attachment-related experiences are encoded and retrieved; (3) autobiographical memories by which specific events are conceptually connected because of their relationship to a continuing personal narrative and developing self-understanding; and (4) understanding of the psychological characteristics of other people (inferring and attributing causal motivational mind states, such as desires and emotions, and epistemic mind states, such as intentions and beliefs) and differentiating these states from those of the self.’ Fonagy, P. (2001) The human genome and the representational world: the role of early mother–infant interaction in creating an interpersonal interpretive mechanism. Bulletin of the Menninger Clinic, 65(3), 427–48, p.436. See also Collins, N.L. & Allard, L.M. (2004) Cognitive representations of attachment: the content and function of working models. In M.B. Brewer & M. Hewstone (eds) Perspectives on Social Psychology: Social Cognition (pp. 75–101). Oxford: Blackwell.

70 Kobak, R., Rosenthal, N., & Serwik, A. (2005) The attachment hierarchy in middle childhood: conceptual and methodological issues. In K.A. Kerns & R.A. Richardson (eds) Attachment in Middle Childhood (pp.71–88). New York: Guilford.

71 E.g. Bowlby, J. (not dated, c. 1955) Thought, conceptualization, language, psycho-analysis. PP/BOW/H.115.

72 E.g. Bretherton, I. (1985) Attachment theory: retrospect and prospect. Monographs of the Society for Research in Child Development, 50, 3–35; Crittenden, P.M. (1990) Internal representational models of attachment relationships. Infant Mental Health Journal, 11(3), 259–77; Lyons-Ruth, K. (1999) The two-person unconscious: intersubjective dialogue, enactive relational representation, and the emergence of new forms of relational organization. Psychoanalytic Inquiry, 19(4), 576–617.

73 Bowlby, J., cited in J.M. Tanner & B. Inhelder (eds) (1956) Discussions on Child Development: Proceedings of the WTO Study Group of the Psychobiological Development of the Child, Vol 1. London: Tavistock, pp.182–3.

74 Bowlby, J. (1938, 1950) An examination of the psychological and anthropological evidence. In E.F.M. Durbin & John Bowlby (eds) Personal Aggressiveness and War (pp.51–150). New York: Columbia University Press.

75 Freud, S. (1918, 2001) From the history of an infantile neurosis. In The Standard Edition of the Complete Psychological Works of Sigmund Freud, Vol. 17 (pp.1–124). London: Vintage, p.116.

76 Annotations by Bowlby dated 1956 on The Standard Edition of the Complete Psychological Works of Sigmund Freud, Vol. 17. Copy held in the library of Human Development Scotland.

77 Klein, M. (1948) Contributions to Psychoanalysis. London: The Hogarth Press and the Institute of Psycho-Analysis, p.314.

78 Ibid. p.280

79 Annotations by Bowlby dated 1948 on Klein, M. (1948) Contributions to Psychoanalysis. London: The Hogarth Press and the Institute of Psycho-Analysis. Copy held in the library of Human Development Scotland.

80 Ibid. The debate between Bowlby and Klein on the status of aggression can be situated as one skirmish within a multiparty controversy running over decades on the status of anger within psychoanalytic theory. See Freud, A. (1972) Comments on aggression. International Journal of Psychoanalysis, 53, 163–71.

81 Bowlby, J. (1960) Separation anxiety. International Journal of Psycho-Analysis, 41, 89–113.

82 Bowlby, J. (1953) The Roots of Parenthood. London: National Children’s Home; Bowlby, J. (1973) Separation. London: Pimlico, p.287; Bowlby, J. (1983) Letter to Dr Marco Bacciagaluppi, 6 July 1983. PP/Bow/J.9/11.

83 Bowlby, J. (1960) Separation anxiety. International Journal of Psychoanalysis, 41, 89–113, p.108.

84 Bowlby, J. (1956) Public letter to Melanie Klein, 7 February 1956, following presentation of her paper ‘A study of envy and gratitude’. PP/Bow/G.1/4.

85 Bowlby, J. (1984) Letter to Phyllis Grosskurth, amending discussions of Bowlby in Grosskurth’s biography of Klein, 10 January 1984. PP/Bow/A.5/7.

86 Young-Bruehl, E. & Bethelard, F. (1999) The hidden history of the ego instincts. Psychoanalytic Review, 86(6), 823–51. Bowlby, J. (1933–36) Functional approach to super-ego. PP/Bow/D.2/49: ‘Freud uses the term sex to describe all positive sentiments between two people … These needs clearly are not always sexual—may be nutritive or self-protective e.g. parent–child. The point may be, however, that the sexual impulses are apt to be aroused in such situations.’ This early work by Bowlby appears as Chapter 5 in Duschinsky, R. & White, K. (eds) (2019) Trauma and Loss: Key Texts from the John Bowlby Archive. London: Routledge.

87 Bowlby, J. (1933–36) Research notes for MD thesis. PP/Bow/D.2/44/3.

88 Bowlby, J. (1933–36) Mechanisms—symptom-formation. PP/Bow/D.2/44/1.

89 Bowlby, J. (1933–36) Anxiety: essays. PP/Bow/D.2/46/6.

90 Bowlby, J. (1986) Attachment theory: new directions. ACP-Psychiatric UPDATE, 7(2), panel discussion, Washington 1986. PP/BOW/A.5/1.

91 Bowlby, J. (1956) The growth of independence in the young child. Royal Society of Health Journal, 76, 587–91. The concept of ‘cherishing’ was an irregular but important one for Bowlby. He refers to the concept in various places, e.g. Bowlby, J. & Robertson, J. (1965) Protest, Despair and Detachment. PP/BOW/D.3/38.

92 Bowlby’s attention to the issue of mourning was stimulated by reading Marris, P. (1958) Widows and their Families. London: Routledge. Marris’s book was unusual for the time in giving consideration to typical as well as atypical mourning processes. The idea that adult and child mourning represents the same process was one Bowlby inherited from Melanie Klein.

93 Nagera, H. (1970) Children’s reactions to the death of important objects: A developmental approach. Psychoanalytic Study of the Child, 25(1), 360–400.

94 Bowlby, J. (1963) Pathological mourning and childhood mourning. Journal of the American Psychoanalytic Association, 11(3), 500–41, p.521.

95 Van der Horst, F.C.P. (2011) John Bowlby—From Psychoanalysis to Ethology. Unravelling the Roots of Attachment Theory. Oxford: Blackwell.

96 Burkhardt Jr, R.W. (2014) Tribute to Tinbergen: putting Niko Tinbergen’s ‘Four Questions’ in historical context. Ethology, 120(3), 215–23.

97 Tinbergen, N. (1963) On aims and methods of ethology. Zeitschrift Tierpsychology, 20, 410–33.

98 See Bowlby’s correspondence with his mother, writing from boarding school, 1921–1924. PP/BOW/A.1/17/1.

99 John Bowlby interview with Robert Keren, 14 and 15 January 1989, cited in Karen, R. (1998) Becoming Attached: First Relationships and how They Shape Our Capacity to Love. Oxford: Oxford University Press, p.94.

100 Bowlby, J. (1977–79) Interview with Alice Smuts and Milton J.E. Senn. PP/BOW/A.5/2. The interview has been published as Chapter 11 in Duschinsky, R. & White, K. (eds) (2019) Trauma and Loss: Key Texts from the John Bowlby Archive. London: Routledge.

101 Hinde, R. (1956) Ethological models and the concept of ‘drive’. British Journal for the Philosophy of Science, 6(24), 321–31. See also Hinde, R.A. (1959) Unitary drives. Animal Behaviour, 7(3), 130–41.

102 Bowlby, J. (1979) The ten books which have most influenced my thought, 24 October 1979. PP/Bow/A.1/8: ‘Robert Hinde (1956) Ethological models and the concept of drive. British Journal of the Philosophy of Science. I first met Hinde in 1954 and in the years following read almost all his papers on publication and often before. It was this paper and others of this published around the same time that led me to the concepts of instinctive behaviour presented in Attachment.’

103 Bowlby, J. (1957) An ethological approach to research on child development. British Journal of Medical Psychology, 30(4), 230–40.

104 Bowlby’s niece has pointed to Bowlby’s disidentification with his disabled younger brother as important in shaping this stance. Hopkins, J. (in press) The need to put things right: a response to Bowlby’s chapter ‘Hysteria in Children’. Attachment. On tiredness as one of the symptoms of Darwin’s illness see Bowlby, J. (1990) Charles Darwin: A Life. New York: Norton.

105 A few years later, Ainsworth reported from her Baltimore home observation study that picking a baby up stopped his or her crying 86% of the time: ‘this degree of effectiveness is remarkable when one notes that it occurred irrespective of the conditions that activated crying’. Ainsworth, M., Bell, S., & Stayton, D. (1972) Individual differences in the development of some attachment behaviors. Merrill-Palmer, 18(2), 123–43, p.132.

106 Lorenz, K. (1937) The companion in the bird’s world. Auk, 54, 245–73.

107 Lorenz, K. (1949, 1952) King Solomon’s Ring. London: Methuen & Co. Bowlby, J. (1986) Interview with the BBC. PP/Bow/F.5/8: ‘First of all you see this following response had nothing to do with food because young geese feed themselves on insects … It’s very powerful and these geese families, they stay together for at least 12 months. Very important. So I said well if this is true of some animal species it might be true of humans too.’

108 Bowlby, J. (1956) Sequence in maturation of drives: notes from discussion with Hinde, August 1956. PP/Bow/H146.

109 E.g. Harlow, H.F. (1958) The nature of love. American Psychologist, 13(12), 673–85; Scott, J.P. (1963) The process of primary socialisation. Monographs of the Society for Research in Child Development, 28, 1–47.

110 Annotations by Bowlby (PP/Bow/H.226) on Hinde, R.A. (1961) The establishment of the parent–offspring relation in birds, with some mammalian analogies. In W.H. Thorpe & O.L. Zangwill (eds) Current Problems in Animal Behaviour (pp.175–93). Cambridge: Cambridge University Press.

111 E.g. Ferenczi, S. (1933, 1980) Confusion of tongues between adults and the child: the language of tenderness and passion. In M. Balint (ed.) Final Contributions to the Problems and Methods of Psychoanalysis (pp.156–67). New York: Brunner.

112 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.157.

113 This account would, essentially, supplant the psychoanalytic model within the psychoanalytic community over the subsequent decades. Holmes, J. (1998) The changing aims of psychoanalytic psychotherapy: an integrative perspective. International Journal of Psychoanalysis, 79, 227–40; Zeuthen, K. & Gammelgaard, J. (2010) Infantile sexuality—the concept, its history and place in contemporary psychoanalysis. Scandinavian Psychoanalytic Review, 33(1), 3–12.

114 Bowlby, J. (1955) Paper read to the Association for the Study of Animal Behaviour, April 1955. PP/Bow/H.146.

115 Bowlby, J. (1957) An ethological approach to research on child development. British Journal of Medical Psychology, 30(4), 230–40.

116 Bowlby/Rieu correspondence, November 1955. PP/Bow/H146.

117 Bowlby, J. (1958) The nature of the child’s tie to his mother. International Journal of Psycho-analysis, 39(5), 350–73, p.370.

118 James, W. (1890, 2003) Psychology: The Briefer Course. Toronto: Dover Books, p.266.

119 E.g. Hinde (1986) Ethology. New York: Fontana, p.230.

120 See also Hinde, R.A. (1963) The nature of imprinting. In B.M. Foss (ed.) Determinants of Infant Behaviour, Vol. 2 (pp.227–30). London: Methuen.

121 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.249.

122 See e.g. van IJzendoorn, M.H., Sagi, A., & Lambermon, M.W.E. (1992) The multiple caregiver paradox. Some Dutch and Israeli data. In R.C. Pianta (ed.) New Directions for Child Development, No. 57. Beyond the Parent: The Role of Other Adults in Children’s Lives (pp.5–25). San Francisco: Jossey-Bass.; Cassidy, J. (1999) The nature of the child’s ties. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research, and Clinical Applications (pp.3–21). New York: Guilford.

123 Bowlby’s correspondence with Peter K. Smith, PP/Bow/J.9/184, discussing Smith, P.K. (1980) Shared care of young children: alternative models to monotropism. Merrill-Palmer Quarterly of Behavior and Development, 26(4), 371–89. See also Bowlby’s correspondence with Michael Rutter, PP/Bow/J.9.161-2.

124 Indeed, Bowlby offered the speculative claim that such relationships are the basis for all that constitutes ‘deep feeling’ within human life. Bowlby, J. (1960) Separation anxiety. International Journal of Psycho-Analysis, 41, 89–113: ‘It is because of this marked tendency to monotropy that we are capable of deep feelings; for to have a deep attachment to a person (or place or thing) is to have taken them as the terminating object of our instinctual responses.’ (101)

125 Bowlby, J. (c.1955) Notes on child attachment and monotropy. PP/Bow/H146.

126 E.g. Ainsworth, M. (1964) Patterns of attachment behavior shown by the infant in interaction with his mother. Merrill-Palmer, 10(1), 51–58, p.56. More recently, Lyons-Ruth and the Boston Change Process Study Group have suggested describing these discriminated relationships as ‘charged’. For them, ‘charged’ relationships are characterized by some underlying valuation, their priority over other relationships, and sufficient continuity to scaffold trust. Boston Change Process Study Group (2018) Engagement and the emergence of a charged other. Contemporary Psychoanalysis, 54(3), 540–59.

127 Bowlby cited in Tanner, J.M. & Inhelder, B. (1956) Discussions on Child Development: Proceedings of the WTO Study Group of the Psychobiological Development of the Child, Vol 1. London: Tavistock, pp.184–5; Bowlby, J. (1973) Separation. London: Pimlico. See also Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, pp.137–8.

128 Bowlby, cited in Tanner, J.M. & Inhelder, B. (1956) Discussions on Child Development: Proceedings of the WTO Study Group of the Psychobiological Development of the Child, Vol 1. London: Tavistock, pp.184–5.

129 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books: ‘Laughlin (1956) has proposed a new term “soteria”, as an obverse of phobia, to denote the intense sense of reassuring comfort that a person may get from a “love object”, be it a toy …’ (148).

130 Bowlby, J. (1956) The growth of independence in the young child. Royal Society of Health Journal, 76, 587–91, p.589.

131 Tinbergen, N. (1956) The functions of territory. Bird Study, 4(1), 14–27; Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books: ‘It is still too little realized, perhaps, that the individuals of a species, so far from roaming at random throughout the whole area of the earth’s surface ecologically suitable to them, usually spend the whole of their lives within an extremely restricted segment of it, known as the home range … each individual has its own relatively small and very distinctive personal environment to which it is attached’ (177).

132 Bowlby, J. (undated) Distress at loss of home: Chapter 3. c.1969–1971. PP/Bow/H.55. See also Bowlby, J. (1965) Attachment behaviour: a note after Ciba 1965. Revised Note January 1966. PP/Bow/H.146: ‘Further definition of attachment: equilibrium point is proximity to a certain type of object … Consider also habitat attachment.’ It must be acknowledged, however, that Bowlby was inconsistent on this point. Likely in an attempt to hack away at the thicket of wider connotations of the word ‘attachment’, he came to emphasize that the object of an attachment relationship had to be human. Stevenson-Hinde, J. (2007) Attachment theory and John Bowlby: some reflections. Attachment & Human Development, 9(4), 337–42: ‘During the first conference, I recall John emphatically stating, “We cannot allow ‘attachment to an umbrella’!” He insisted that “attachment” be used to describe an emotional bond to someone (i.e., a person) usually perceived as older or wiser (e.g., mother or father). While other kinds of bonds undoubtedly exist, they should not be called “attachment,” in order to keep some precision in the use of terms’ (338). This, exclamation, of course, produces an excluded middle: our feelings for our home are neither the trivial affection we might feel for a favourite umbrella, nor the feelings we have for a parent.

133 Gruneau Brulin, J. & Granqvist, P. (2018) The place of place within the attachment-religion framework: a commentary on the circle of place spirituality. Research in the Social Scientific Study of Religion, 29, 175–85.

134 The claim that home cannot be a secure base or safe haven would seem to depend on a reification of ‘attachment relationship’ beyond its constituent elements. Gruneau Brulin and Granqvist claim that home is non-individual, which is implausible. They also claim that it cannot be an object of attachment since it is non-reciprocal. However, reciprocity was never part of either Bowlby or Ainsworth’s definition of an attachment relationship.

135 On attachment to place see e.g. Scannell, L. & Gifford, R. (2013) Comparing the theories of interpersonal and place attachment. In L.C. Manzo & P. Devine-Wright (eds) Place Attachment. Advances in Theory, Methods and Applications (pp.23–36). London: Routledge. On attachment to siblings see e.g. Teti, D.M. & Ablard, K.E. (1989) Security of attachment and infant–sibling relationships: a laboratory study. Child Development, 1519–28; Farnfield, S. (2009) A modified strange situation procedure for use in assessing sibling relationships and their attachment to carers. Adoption & Fostering, 33(1), 4–17.

136 Fonagy, P., Steele, M., Steele, H., Higgitt, A., & Target, M. (1994) The Emanuel Miller memorial lecture 1992: the theory and practice of resilience. Journal of Child Psychology and Psychiatry, 35(2), 231–57; Main, M. (1999) Epilogue. Attachment theory: eighteen points with suggestions for future studies. In J. Cassidy & P. Shaver (eds) Handbook of Attachment (pp.845–87). New York: Guilford, p.848.

137 E.g. Bowlby, J. (1991) Ethological light on psychoanalytic problems. In P. Bateson (ed.) The Development and Integration of Behaviour: Essays in Honour of Robert Hinde (pp.301–14). Cambridge: Cambridge University Press: ‘The typed drafts of major parts of the first edition of his [Hinde’s] Animal Behaviour (1966) which he lent me in 1965 when I was starting work on my volume on Attachment (1969). Whatever merits my own volume has owed a tremendous debt to his’ (303).

138 Waters, E., Kondo-Ikemura, K., Posada, G., and Richters, J.E. (1991) Learning to love. In M.R. Gunnar & L.A. Sroufe (eds) Self Processes and Development (pp.217–55). New York: Psychology Press, p.227. Ainsworth, M. & Bowlby, J. (1991) 1989 APA award recipient addresses: an ethological approach to personality development. American Psychologist, 46(4), 333–41: ‘The distinguishing characteristic of the theory of attachment that we have jointly developed is that it is an ethological approach’ (333).

139 See e.g. Bowlby, J. (1963) Remarks at the MRC Ethology Meeting organised by Bowlby, 23 May 1963. PP/Bow/D.6/5: ‘The role of a comparative approach: a) in facilitating observation b) in elucidating the evolution of behaviour c) in taxonomy d) in providing a basis for generalisation e) in leading to an understanding of function’ (3).

140 Bowlby, R. (2017) Growing up with attachment theory—a personal view. Psychodynamic Psychiatry, 45(4), 431–9: ‘When my father sat down he told us that he was looking for a new term to replace [a] child’s “tie.” He said the image of a child being tied to mother or mother being tied to a child had become socially unacceptable, and he was thinking of using the child’s “attachment” instead. We all groaned and said how boring and why couldn’t he use “love” like he had originally? He explained that “love” was not strictly accurate and anyway he had already decided he was going to use attachment from then on’ (436).

141 Bowlby, J. (1985) Letter to John Byng-Hall, 12 April 1985. PP/Bow/J.9/45.

142 Rutter, M., Kreppner, J., & Sonuga-Barke, E. (2009) Emanuel Miller lecture: attachment insecurity, disinhibited attachment, and attachment disorders: where do research findings leave the concepts? Journal of Child Psychology and Psychiatry, 50(5), 529–43. In fact, ‘love’ would always be threatening to return, hammering at the door of attachment research. This was especially the case for the social psychology tradition (Chapter 5), e.g. Hazan, C. & Shaver, P. (1987) Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52(3), 511–24. However, appeal to ‘love’ is also a feature in some work in the developmental tradition of attachment research. For instance, ‘loving’ would be one of the scales of the Adult Attachment Interview, and used in the assessment of transcripts as ‘earned secure’ (Chapter 3).

143 E.g. Ainsworth, M. (1967) Infancy in Uganda. Baltimore: The Johns Hopkins Press: ‘Attachment is more than a discrimination between people and implies something far more active—a literal or figurative seeking out, fastening on’ (440).

144 The centrality of symptom formation through symbolization in psychoanalysis was established in Breuer, J. & Freud, S. (1893–95, 2001) Studies on hysteria. In The Standard Edition of the Complete Psychological Works of Sigmund Freud, Vol. 2 (pp.1–305). London: Vintage. Bowlby was especially interested by Rycroft, C. (1956) Symbolism and its relationship to the primary and secondary processes. International Journal of Psycho-Analysis, 37, 137–46, and had correspondence with the author (PP/Bow/H.116). Part of the significance of Rycroft’s work on symbolization for Bowlby was that it showed that ‘the emphasis placed in Kleinian theory on the fact that “psychical reality” and “external reality” are both subjectively real does, I think, tend to obscure the fact that there are none the less essential differences between them, and that psychical reality is itself divisible into one part which is developmentally bound to external reality and another which has been formed by idealization’ (141).

145 For an example of Bowlby taking away implications for relationships in general from a discussion with Hinde of specific aspects of the following response see e.g. Bowlby, J. (1957) Discussion with Hinde, January 1957. PP/Bow/H.128: ‘Attachment behaviour comprises all those responses which subserve the total task of relating to another human being.’

146 The capacity for symbolization was, in a sense, highlighted by Main and colleagues’ ‘move to the level of representation’ (Chapter 3). However, Main and colleagues did not distinguish the cross-species ethological aspects of attachment from those aspects associated with human symbolic capabilities. Implicit acknowledgement of the issue also appeared in the 1990s in the writings of researchers focused on intersubjectivity and rhythms of parent–infant interaction, e.g. Beebe, B., Lachmann, F., & Jaffe, J. (1997) Mother–infant interaction structures and presymbolic self- and object representations. Psychoanalytic Dialogues, 7(2), 133–82; Trevarthen, C. (1998) The concept and foundations of infant intersubjectivity. In S. Braten (ed.) Intersubjective Communication and Emotion in Early Ontogeny (pp.15–46). Cambridge: Cambridge University Press. However, concepts of intersubjectivity and rhythm emphasize continuities between presymbolic procedural expectations in relationships and the symbolic capacities that emerge from them. The distinction and potential disjuncture between presymbolic and symbolic senses of the concept of ‘attachment’ was not drawn out. In the history of attachment research, this distinction appears to have first been made focally and clearly in the application of attachment theory to religious life: Kirkpatrick, L.A. (1999) Attachment and religious representations and behavior. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research, and Clinical Applications (pp.803–22). New York: Guilford.

147 E.g. Bowlby, J. (not dated, c.1955) Thought, conceptualization, language, psycho-analysis. PP/BOW/H.115; Bowlby, J. (1955) Letter to C.F. Rycroft, 26 April 1955. PP/Bow/H.116.

148 Bowlby, J. (1980) Loss. New York: Basic Books, p.243. The confusion evoked by this stance for later interpretations of Bowlby is detailed in Fraley, C.R. & Shaver, P.R. (2016) Attachment, loss, and grief: Bowlby’s views, new developments, and current controversies. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research, and Clinical Applications (3rd edn, pp.40–62). New York: Guilford.

149 Bowlby, J. (1944) Forty-four juvenile thieves: their characters and home-life (II). International Journal of Psycho-Analysis, 25, 107–28, p.121.

150 On this latter point, criticisms of Bowlby are presented by Fonagy, P. & Target, M. (2007) The rooting of the mind in the body: new links between attachment theory and psychoanalytic thought. Journal of the American Psychoanalytic Association, 55(2), 411–56. However, Fonagy and Target did not adequately recognize the polysemy of Bowlby’s concept of attachment, which hindered their own discussion of attachment and mentalization.

151 Note Bowlby’s wording in his very definition of internal working models as ‘starting, we may suppose, towards the end of his first year’, in a procedural form, and as developing into but conceptually distinguishable from semantically elaborated internal working models ‘during his second and third when [the child] acquires the powerful and extraordinary gift of language’. Bowlby, J. (1969) Attachment. London: Penguin, p.353.

152 Kobak, R. & Esposito, A. (2004) Levels of processing in parent–child relationships: implications for clinical assessment and treatment. In L. Atkinson & S. Goldberg (eds) Attachment Issues in Psychopathology and Interventions (pp. 139–66). Mahwah, NJ: Lawrence Erlbaum, p.140.

153 Hinde, R. (1957) Consequences and goals: some issues raised by Dr Kortland’s paper on aspects and prospects of the concept of instinct. British Journal of Animal Behaviour, 5, 116–18, p.116.

154 Hinde, R. (1967) Letter to John Bowlby, 28 June 1967. PP/Bow/K.4/11. Hinde would later make the observation in print, looking back on Bowlby’s overall contribution to the study of behaviour in Hinde, R. (1991) Relationships, attachment and culture: a tribute to John Bowlby. Infant Mental Health, 12(3), 154–63; and Hinde, R. (1991) Commentary. In P. Bateson (ed.) The Development and Integration of Behaviour: Essays in Honour of Robert Hinde (pp.411–18). Cambridge: Cambridge University Press.

155 Hinde, R.A. (1982) Attachment: some conceptual and biological issues. In C.M. Parkes & J. Stevenson-Hinde (eds) The Place of Attachment in Human Behavior (pp.60–76). London: Tavistock, p.71.

157 Carr, S.J., Dabbs Jr, J.M., & Carr, T.S. (1975) Mother–infant attachment: the importance of the mother’s visual field. Child Development, 46(2), 331–8; Sorce, J.F. & Emde, R.N. (1981) Mother presence is not enough: effect of emotional availability on infant exploration. Developmental Psychology,17(6), 737–45. See also Joffe, L.S., Vaughn, B.E., Barglow, P., & Benveniste, R. (1985) Biobehavioral antecedents in the development of infant–mother attachment. In M. Reite & T. Field (eds) The Psychobiology of Attachment and Separation (pp.323–49). Orlando, FL: Academic Press.

158 A first acknowledgement appears in Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books: ‘Accessibility in itself is not enough. Not only must an attachment figure be accessible but he, or she, must be willing to respond in an appropriate way; in regard to someone who is afraid this means willingness to act as comforter and protector. Only when an attachment figure is both accessible and potentially responsive can he, or she, be said to be truly available. In what follows, therefore, the word ‘available’ is to be understood as implying that an attachment figure is both accessible and responsive’ (234). Throughout the 1970s, Bowlby tended still to refer to proximity as the set-goal of the attachment system. His final word, however, is in Bowlby, J. (1991) Ethological light on psychoanalytic problems. In P. Bateson (ed.) The Development and Integration of Behaviour: Essays in Honour of Robert Hinde (pp.301–14). Cambridge: Cambridge University Press: ‘The goal of attachment behaviour is to maintain certain degrees of proximity to, or of communication with, the discriminated attachment figure(s)’ (306). Some of the tensions in Bowlby’s account of the set-goal of the attachment system had been discussed already by Bretherton, I. (1980) Young children in stressful situations: the supporting role of attachment figures and unfamiliar caregivers. In G.V. Coelho & P. Ahmed (eds) Uprooting and Attachment (pp.179–210). New York: Plenum Press.

159 Still by 1990 Main was adamant that physical touch with the caregiver was ultimately the set-goal of the attachment system in infancy. Main, M. (1990) Parental aversion to infant-initiated contact is correlated with the parent’s own rejection during childhood: the effects of experience on signals of security with respect to attachment. In T.B. Brazelton & K. Barnard (eds) Touch (pp.461–95). New York: International Universities Press.

160 Marvin, R.S. (1977) An ethological–cognitive model for the attenuation of mother–child attachment behavior. In T.M. Alloway, L. Krames, and P. Pliner (eds) Advances in the Study of Communication and Affect, Vol. 3 (pp.25–60). New York: Plenum Press, pp.56–7.

161 Rutter, M. (1995) Clinical implications of attachment concepts: retrospect and prospect. Journal of Child Psychology and Psychiatry, 36(4), 549–71, p.551. See also Stern, D. (1985) The Interpersonal World of the Infant. New York: Basic Books: ‘Attachment is a set of infant behaviours, a motivational system, a relationship between mother and infant, a theoretical construct, and a subjective experience for the infant’ (25).

162 From the 1970s onwards, Bowlby tested out referring to ‘care-seeking’ rather than ‘attachment behaviour’, e.g. Bowlby, J. (1986) Attachment, life-span and old age. In J. Munnichs & B. Miesen (eds) Attachment, Life-Span and Old Age. Utrecht: Van Loghum, p.11.

163 Charles, G. & Alexander, C. (2014) Beyond attachment: mattering and the development of meaningful moments. Relational Child and Youth Care Practice, 27(3), 26–30: ‘Herein lies another problem with “attachment.” It is one of those terms which we all think we understand the meaning of but when we actually examine it we find that it has significantly different meanings for different people … Today the term is often a loose metaphor for a relationship-based intervention, and those using the term do not necessarily have an accurate understanding of the concept. The absence of a precise and universally understood definition has led to a wide variety of interpretations of what is a practical “attachment intervention.” For example, there are a number of controversial “attachment” treatments based on various forms of “therapeutic holding” ’ (27).

164 Duschinsky, R., Greco, M., & Solomon, J. (2015) Wait up! Attachment and sovereign power. International Journal of Politics, Culture, and Society, 28(3), 223–42. The polyvalence of attachment as a political discourse should be highlighted, however. For uses of attachment theory for a social agenda more aligned with the political left see e.g. Kraemer, S. & Roberts, J. (1996) The Politics of Attachment: Towards a Secure Society. London: Free Association Books. There are also a variety of policy texts invoking attachment without a marked political agenda, and instead using the term in a general sense to characterize the value of ‘positive’ parent–child relationships. See e.g. Scottish Government (2012) National Parenting Strategy: Making a Positive Difference to Children and Young People through Parenting. Edinburgh: Scottish Government.

165 ‘Attachment parenting’ is one of the most powerful discourses of intensive parenting. It was introduced by Bill and Martha Sears (1993) in The Baby Book: Everything You Need to Know about your Baby (Boston: Little and Brown). The Sears already had their ideas in place, but initially called them ‘immersion parenting’. Use of the idea of attachment and appeal to Bowlby’s authority was post-hoc and strategic, made available by Bowlby’s ambiguous and overgeneral statements about the dangers of separation and the need for mothers to spend time with their baby: ‘At a talk one time in Pasadena, a grandmother came up to Bill and said she thought the term immersion mothering was a good one, because some moms find themselves “in over their heads.” When he told me of this, I realized we needed to change the term to something more positive, so we came up with AP, since the Attachment Theory literature was so well researched and documented, by John Bowlby and others’ ( On the weak evidence for ‘attachment parenting’ and its lack of a link with attachment research see Chaffin, M. (2006) Report of the APSAC Task Force on Attachment Therapy, Reactive Attachment Disorder, and Attachment Problems. Child Maltreatment, 11(1), 76–89; Fairclough, C. (2013) The problem of ‘attachment’. In E. Lee, J. Bristow, C. Faircloth, & J. Macvarish (eds) Parenting Culture Studies (pp.147–64). London: Palgrave.

166 Keller, H. & Thompson, R. (2018) Attachment theory: past, present & future. Recorded at the 2nd ‘Wilhelm Wundt Dialogue’, 28 November 2018, Leipzig University, hosted by the Leipzig Research Center for Early Child Development (LFE).

167 Bowlby, J. (1990, 2015) John Bowlby: an interview by Virginia Hunter. Attachment, 9, 138–57, p.147. See also Van Dijken, S. (1998) John Bowlby, his Early Life: A Biographical Journey into the Roots of Attachment Theory. New York: Free Association Books.

168 Bowlby, J. (1986) Interview with the BBC. PP/Bow/F.5/8.

169 Bowlby, J. (1944) Forty-four juvenile thieves: their characters and home-life (II). International Journal of Psycho-Analysis, 25, 107–28. An earlier version of the paper, delivered on 30 November 1937, is available as ‘Some unconscious motives in habitual pilfering’, PP/BOW/C.3/10/2.

170 The specific theory linking separations to delinquency has not subsequently been well supported. There are, however, few relevant studies. Ryan and colleagues have reported that group care, as opposed to family or foster care, increases the likelihood of criminal activity: Ryan, J.P., Marshall, J.M., Herz, D., & Hernandez, P.M. (2008) Juvenile delinquency in child welfare: investigating group home effects. Children and Youth Services Review, 30(9), 1088–99. In the attachment literature, Allen and colleagues found a concurrent relationship between criminal activity and insecurity with the Adult Attachment Interview: Allen, J. P., Hauser, S.T., & Borman-Spurrell, E. (1996) Attachment theory as a framework for understanding sequelae of severe adolescent psychopathology: an 11-year follow-up study. Journal of Consulting and Clinical Psychology, 64(2), 254. By contrast, Brennan and Shaver found no association between self-reported attachment style and criminal activity: Brennan, K.A. & Shaver, P.R. (1998) Attachment styles and personality disorders: their connections to each other and to parental divorce, parental death, and perceptions of parental caregiving. Journal of Personality, 66(5), 835–78. For a review see also Schimmenti, A. (2020) The developmental roots of psychopathy: an attachment perspective. In S. Itzkowitz & E.F. Howell (eds) Psychopathy and Human Evil: Psychoanalytic Explorations. London: Routledge.

171 This interpretation of Bowlby was first offered in Chodorow, N. (1978) The Reproduction of Mothering. Berkeley: University of California Press, p.75. As well as the clinical cases seen at the Child Guidance Clinic, another experience that may have been relevant to Bowlby’s attention to the pathogenic role of separations was his training as a child analyst. Whereas at the Child Guidance Clinic Bowlby generally saw children and their primary caregiver or caregivers together, the technique for child analysis in the 1930s was to meet with the child alone. Bowlby would later recall to his former student Victoria Hamilton that he had been upset by the distress he would cause young children, time and time again, as he would separate the children from their caregiver in the waiting room and take them to the consulting room. Hamilton, V. (2007) The nature of a student’s tie to the teacher: reminiscences of training and friendship with John Bowlby. Attachment, 1, 334–47.

172 Bowlby, J. & Fairbairn, C.N. (c.1939–42) The billeting of unaccompanied school children. PP/Bow/C.5/4/1.

173 Bowlby, J., Miller, E., & Winnicott, D. (1939) Evacuation of small children. Letter to the Editor of the British Medical Journal, 16 December 1939. In Winnicott, D. (1984) Deprivation and Delinquency (pp.13–14). London: Tavistock.

174 Bowlby, J. (c.1939–42) Psychological problems of evacuation. PP/Bow/C.5/4/1. This early work by Bowlby appears as Chapter 7 in Duschinsky, R. & White, K. (eds) (2019) Trauma and Loss: Key Texts from the John Bowlby Archive. London: Routledge.

175 Bowlby, J. & Fairbairn, C.N. (c.1939–42) The billeting of unaccompanied school children. PP/Bow/C.5/4/1.

176 Follan, M. & Minnis, H. (2010) Forty-four juvenile thieves revisited: from Bowlby to reactive attachment disorder. Child: Care, Health and Development, 36(5), 639–45.

177 Bowlby, J. (1986) An interview with John Bowlby on the origins and reception of his work. Free Associations, 1, 36–64, p.39.

178 Bowlby, J. (undated) Untitled case history beginning ‘Mrs E. consulted the Clinic about her son Martin’, in the file Maternal Behaviour: Humans. PP/Bow/H.136.

179 Bowlby, J. (1985) Letter to Tirril Harris, 17 September 1985. PP/Bow/J.9/94.

180 Bowlby, J. (1986) Interview with the BBC. PP/Bow/F.5/8.

181 Bowlby, J. (1984) Violence in the family as a disorder of the attachment and caregiving systems. American Journal of Psychoanalysis, 44, 9–27: ‘It was, indeed, largely because the adverse behavior of parents toward their children was such a taboo subject in analytic circles when I was starting my professional work that I decided to focus my research on the effects on children of real-life events of another sort, namely separation’ (10).

182 Bowlby, J. (1990, 2011) John Bowlby: interview by Leonardo Tondo. Clinical Neuropsychiatry, 8(2), 159–71, p.160.

183 Bowlby, J. (1943, 1992) Contribution to business meeting. In P. King & R. Steiner (eds) The Freud–Klein Controversies 1941–45, 2nd edn. London: Routledge, p.369.

184 Bowlby, J. (1990, 2015) John Bowlby: an interview by Virginia Hunter. Attachment, 9, 138–57, p.156.

185 Bowlby, J. (1985, 1991) The role of the psychotherapist’s personal resources in the treatment situation. Bulletin of the British Psychoanalytic Society, 27(11), 26–30, p.29–30. Published as Chapter 12 in Duschinsky, R. & White, K. (eds) (2019) Trauma and Loss: Key Texts from the John Bowlby Archive. London: Routledge.

186 By the 1980s, Bowlby was admonishing colleagues for not being specific enough in their use of the term ‘separation’, e.g. Bowlby, J. (1985) Letter to John Byng-Hall, 12 April 1985. PP/Bow/J.9/45: ‘I think one needs to be a little more precise about lengths of separation—words like moderate and prolonged are obscure. I suggest a week or two instead of “moderate”, and “longer than that” in place of prolonged.’

187 E.g. Ainsworth, M. (1962) The effects of maternal deprivation: a review of findings and controversy in the context of research strategy. In Deprivation of Maternal Care: A Reassessment of its Effects (pp.87–195). Geneva: WHO, p.99. See also Yarrow, L. (1961) Maternal deprivation: toward an empirical and conceptual re-evaluation. Psychological Bulletin, 58, 459–90.

188 Rutter, M. (1972) Maternal Deprivation Reassessed. London: Penguin; Rutter, M. (2002) Maternal deprivation. In M. Bornstein (ed.) Handbook of Parenting, 2nd edn (pp.181–202). Mahwah, NJ: Lawrence Erlbaum.

189 E.g. Bowlby, U. & Bowlby, J. (1940) Difficult children. PP/Bow/H7: ‘The term—broken home covers a multitude of situations—illustrate. Vague, don’t intend to use term but to examine different situations … Broken home does not cause trouble.’

190 E.g. Bowlby, J. (1971) Letter to Michael Rutter, 6 October 1971. PP/Bow/J.9/161.

191 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.96. Another contributing factor appears to have been Bowlby’s initial difficulties in articulating the distinction between observable attachment behaviour and the invisible attachment behavioural system. Bowlby would often make claims urging parents to always do their best to follow the cues of a child’s attachment to ensure their child’s wellbeing. To the degree that this refers to the attachment behavioural system, the claim is clearly overstated, and neglects his friend Robert Hinde’s criticisms that weaning and other requirements on parents mean that the short-term demands of the attachment system should not always be given priority in facilitating children’s long-term security. However, to the degree that Bowlby was implying that parents should follow the dictates of attachment behaviours—such as distress on separation—the result is an even more extreme position. It would imply that any separation is, in itself, potentially harmful. See, for instance, Bowlby, J. (1987) Baby love: an interview. Hampstead and Highgate Express, April 1987. PP/Bow/A.5/19: ‘The more a child’s attachment is respected and responded to, the more he’ll feel secure.’

192 Ainsworth, M. (1962) The effects of maternal deprivation: a review of findings and controversy in the context of research strategy. In Deprivation of Maternal Care: A Reassessment of its Effects (pp.87–195). Geneva: WHO, p.101; Bowlby, R. (2005) Fifty Years of Attachment Theory. London: Karnac; Vicedo, M. (2011) The social nature of the mother’s tie to her child: John Bowlby’s theory of attachment in post-war America. British Journal for the History of Science, 44(3), 401–26.

193 Bowlby, J. (1965) Comments on Joffe and Sandler 1965 ‘Notes on Pain, Depression and Individuation’. PP/Bow/J.9/168-9.

194 Van der Horst, F.C.P. (2011) John Bowlby—From Psychoanalysis to Ethology. Unravelling the Roots of Attachment Theory. Oxford: Blackwell.

195 Rustin, M. (2007) John Bowlby at the Tavistock. Attachment & Human Development, 9(4), 355–9.

196 Bowlby, J. & Robertson, J. (1965) Protest, Despair and Detachment. PP/BOW/D.3/38: ‘Separation can be likened to the natural experiments that are exploited by students of nutrition.’

197 Bowlby, J. (1939) Hysteria in children. In H. Milford (ed.) A Survey of Child Psychiatry (pp.80–94). Oxford: Oxford University Press, p.84.

198 Bowlby J., Robertson, J., & Rosenbluth, D. (1952) A two-year-old goes to hospital. Psychoanalytic Study of the Child, 7, 82–94. During the war, Robertson had worked with Anna Freud at the Hampstead Nurseries with children, many of whom had been evacuated from London or who had no family to care for them. Freud and colleagues documented avoidant behaviour by young children to caregivers, including following reunions. This likely primed Robertson’s interest in the avoidant behaviour shown by Laura and other hospitalized children on reunion. See Burlingham, D. & Freud, A. (1944) Infants without Families. London: Allen and Unwin, p.63.

199 Van der Horst, F.C. & van der Veer, R. (2009) Why we disagree to disagree: a reply to commentaries by Robertson and McGilly, and Lindsay. Attachment & Human Development, 11(6), 569–72.

200 Bowlby J., Robertson, J., & Rosenbluth, D. (1952) A two-year-old goes to hospital. Psychoanalytic Study of the Child, 7, 82–94, p.85.

201 Bowlby, J. & Robertson, J. (1965) Protest, Despair and Detachment. PP/BOW/D.3/38.

202 Bowlby, J. (1960) Separation anxiety. International Journal of Psycho-Analysis, 41, 89–113, p.90. See also Southgate, J. (1998) Attachment, intimacy, autonomy. British Journal of Psychotherapy, 14, 389–93: ‘During my supervision with John … he argued to keep its original meaning on the grounds that once in an attachment space or relational field there is always some form of attachment even if it is disorganized and chaotic’ (390).

203 Ainsworth, M. (1962) Letter to John Bowlby, 11 December 1962. Mary Ainsworth papers, Box M3168, Folder 1. For an example of an otherwise careful reader and friend of Bowlby assuming that ‘detachment’ meant the opposite of attachment see Birtchnell, J. (1987) Attachment—detachment, directiveness—receptiveness: a system for classifying interpersonal attitudes and behaviour. British Journal of Medical Psychology, 60(1), 17–27.

204 Bowlby, J. & Robertson, J. (1965) Protest, Despair and Detachment. PP/BOW/D.3/38.

205 Ibid: ‘It is our impression, and that of others, that children who have reached a detached state … “blow up” more easily and more violently event than the ordinary child at home.’ An example was Bobby: ‘Everyone was punched when necessary—including father—and it was usually possible to detect the immediate reason for it. But his treatment of mother was exceptional in that she was often punched for no apparent reason. At times he would approach her with a bland or smiling face that gave no warning of the severe body blow that was to follow.’ See also Robertson, J. & Robertson, J. (1971) Young children in brief separation: a fresh look. The Psychoanalytic Study of the Child, 26(1), 264–315.

206 E.g. Eagle, M. (2013) Attachment and Psychoanalysis. New York: Guilford.

207 Hazan, C. & Shaver, P.R. (1994) Deeper into attachment theory. Psychological Inquiry, 5(1), 68–79, p.70.

208 Bowlby, J. (1953) Some pathological processes set in train by early mother–child separation. Journal of Mental Science, 9, 265–72, p.271. The exact meaning of the term ‘differential susceptibility’ for Bowlby is unclear. It may or may not necessarily imply that genetic factors may be ‘for better or for worse’, as later for Belsky, van IJzendoorn, and Bakermans-Kranenburg.

210 Bowlby was pleased with later research findings indicating the mediating role of family relationships on the impact of parental loss in childhood on an individual’s later mental health. See PP/BOW/F.4/1:Box 40 and PP/BOW/J.9/19:Box 60 for Bowlby’s reflections on and correspondence with Harris, T., Brown, G.W., & Bifulco, A. (1986) Loss of parent in childhood and adult psychiatric disorder: the role of lack of adequate parental care. Psychological Medicine, 16(3), 641–59.

211 Bowlby, J. (1976) Bowlby on latch-key kids: interviews with Dr Nicholas Tucker. Psychology Today, Autumn 1976, 37–41: ‘I felt that because we had used such very superficial measures we were in no real position to give an adequate account of how these children had developed … I really don’t think the study has much scientific value’ (38).

212 Ainsworth, M. (1969) CPA oral history of psychology in Canada interview. Unpublished.

213 Though a proportion of the most important data was second-hand. Bowlby, J. & Robertson, J. (1965) Protest, Despair and Detachment. PP/BOW/D.3/38: ‘As experience accumulated, it came to be realised that the way a child greets his mother on the occasion of his return home is of great interest. This, however, was not realised in our earlier studies and for this reason in all but a few cases our data referring to this event were obtained second-hand.’

214 Ainsworth, M. (1983, 2013) An autobiographical sketch. Attachment & Human Development, 15(5–6), 448–59, p.454.

215 Ainsworth, M. (1968) Letter to John Bowlby, 2 November 1968. Mary Ainsworth papers, Box M3168, Folder 3.

216 Ainsworth, M. (1969) CPA oral history of psychology in Canada interview. Unpublished.

217 Van der Horst, F.C. & van der Veer, R. (2009) Separation and divergence: the untold story of James Robertson’s and John Bowlby’s theoretical dispute on mother–child separation. Journal of the History of the Behavioral Sciences, 45(3), 236–52.

218 That the preoccupied response was a reflection of concern about the availability of the caregiver was confirmed by Bowlby’s clinical observations of the possessiveness, anger, and jealousy with which toddlers responded to the birth of a new sibling, especially if the toddler already had a troubled relationship with his or her caregiver. Bowlby, J. (1955) New baby jealousy. Parents, December 1955, p.42–4.

219 Bowlby, J. & Robertson, J. (1965) Protest, Despair and Detachment. PP/BOW/D.3/38.

220 Bowlby, J., Ainsworth, M., Boston, M., & Rosenbluth, D. (1956) The effects of mother–child separation: a follow-up study. British Journal of Medical Psychology, 29, 211–47, p.238. On the two classes as ‘opposites’ see also Bowlby, J. (1968, 1970) Disruption of affectional bonds and its effects on behavior. Journal of Contemporary Psychotherapy 2(2), 75–8: ‘In the separated children, two forms of disturbance of affectional behaviour were seen, neither of which were observed in the comparison group of non-separated children. One form is that of emotional detachment; the other, its apparent opposite, i.e., an unrelenting demand to be close to mother’ (82). Both responses would be documented again, decades later, by Stovall-McClough and Dozier, observing children one to two years old in the first months after joining a foster-family. Stovall-McClough, K.C. & Dozier, M. (2004) Forming attachments in foster care: infant attachment behaviors during the first 2 months of placement. Development & Psychopathology, 16(2), 253–71. Stovall-McClough and Dozier observed that these behaviours may mis-cue the foster-parent about the child’s needs. These findings fed into the construction of the Attachment and Biobehavioral Catch-up intervention (Chapter 6).

221 Ainsworth, M. (1962) The effects of maternal deprivation: a review of findings and controversy in the context of research strategy. In Deprivation of Maternal Care: A Reassessment of its Effects (pp.87–195). Geneva: WHO, p.140.

222 Bowlby, J. (1976) Human personality development in an ethological light. In G. Serban & A. Kling (eds) Animal Models in Human Psychobiology (pp.27–36). New York: Plenum Press, p.28.

223 Bateson, P., Stevenson-Hinde, J., & Clutton-Brock, T. (2018) Robert Aubrey Hinde CBE. 26 October 1923–23 December 2016. Royal Society Biographical Memoirs, 65.

224 Spencer-Booth, Y. & Hinde, R. (1967) The effects of separating rhesus monkey infants from their mothers for six days. Journal of Child Psychology & Psychiatry, 7, 179–97, p.187.

225 Ibid. p.190.

226 Hinde, R. & Spencer-Booth, Y. (1970) Individual differences in the responses of rhesus monkeys to a period of separation from their mothers. Journal of Child Psychology & Psychiatry, 11, 159–76, p.174.

227 Hinde, R. & Spencer-Booth, Y. (1971) Effects of brief separation from mothers on rhesus monkeys. Science 173(3992), 111–18.

228 Bowlby, J. & Robertson, J. (1965) Protest, Despair and Detachment. PP/BOW/D.3/38.

229 Ross, C.A. (1996) History, phenomenology, and epidemiology of dissociation. In L.K. Michelson & W.J. Ray (eds) Handbook of Dissociation: Theoretical, Empirical and Clinical Perspectives (pp.3–25). New York: Springer; Van der Hart, O. & Dorahy, M.J. (2009) Dissociation: history of a concept. In P.F. Dell & J. O’Neill (eds) Dissociation and the Dissociative Disorders: DSM-V and Beyond (pp.3–26). London: Routledge.

230 Ainsworth, M. (1970) Letter to John Bowlby, 28 September 1970. Mary Ainsworth papers, Box M3168, Box 3.

231 Bowlby, J. & Robertson, J. (1965) Protest, Despair and Detachment. PP/BOW/D.3/38. Bowlby’s colleagues Heinicke and Westheimer also documented such responses following brief separations. Heinicke, C.M. & Westheimer, I. (1966) Brief Separations. Oxford: International Universities Press.

232 Bowlby, J. & Robertson, J. (1965) Protest, Despair and Detachment. PP/BOW/D.3/38.

235 Bowlby, J. & Robertson, J. (undated, 1950–65?) Cases relating to part II and III. PP/BOW/D.3/11-12.

236 Bowlby, J. & Robertson, J. (1965) Protest, Despair and Detachment. PP/BOW/D.3/38.

238 Bowlby, J. (1960) Separation anxiety. International Journal of Psycho-Analysis, 41, 89–113, p.95.

239 Ainsworth, M. (1966) Letter to John Bowlby, 29 October 1966. Mary Ainsworth papers, Box M3168, Folder 2.

240 Sevenster, P. (1961) A causal analysis of a displacement activity (Fanning in Gasterosteus aculeatus L.). Behaviour, 9, 1–170; Baerends, G.P. (1976) The functional organization of behaviour. Animal Behaviour, 24, 726–38. The movement of the concept from ethology to attachment theory is discussed further in Grossmann, K. & Grossmann, K. (2012) Bindungen—das Gefüge psychischer Sicherheit Gebundenes. Stuttgart: Klett-Cotta.

241 Hinde, R. (1983) Ethology and child development. In P.H. Mussen (ed.) Handbook of Child Psychology (pp.27–93). New York: Wiley, p.57.

242 Bowlby, J. (1982) Attachment and loss: retrospect and prospect. American Journal of Orthopsychiatry, 52(4), 664–78, p.670.

243 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.69.

244 Marvin, R.S., Britner, A.A., & Russell, B.A. (2016) Normative development: the ontogeny of attachment in childhood. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research, and Clinical Applications, 3rd edn (pp.273–90). New York: Guilford.

245 An important case was that of contemporary social learning theorists, e.g. Gewirtz, J.L. (ed.) (1972) Attachment and Dependency. Washington, DC: Winston.

246 E.g. Baerends, G.P. (1976) The functional organization of behaviour. Animal Behaviour, 24, 726–38.

247 Bowlby, J. (1958) The nature of the child’s tie to his mother. International Journal of Psycho-Analysis, 39, 350–73, pp.365–6; Bowlby, J. (1980) Loss. London: Pimlico, p.348.

248 E.g. Zeanah, C.H. & Anders, T.F. (1987) Subjectivity in parent–infant relationships: a discussion of internal working models. Infant of Mental Health Journal, 8(3), 237–50.

249 E.g. Bowlby, J. (1982) Outline of attachment theory. PP/Bow/H.260.

250 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin: ‘The process of categorising parts of the environment in terms of fitness to elicit a particular class of behaviour is itself experienced as coloured by the appropriate emotion’ (114).

251 Bowlby, J. (1962) Defences that Follow Loss: Causation and Function. PP/Bow/D.3/78.

252 Bowlby, J. (1960) Separation anxiety. International Journal of Psycho-Analysis, 41, 89–113, p.96.

253 Harkness, S. (2015) The strange situation of attachment research: a review of three books. Reviews in Anthropology, 44(3), 178–97, p.179.

254 Bowlby, J. (1965) Comments on Joffe and Sandler 1965 ‘Notes on Pain, Depression and Individuation’. PP/Bow/J.9/168–9.

255 E.g. Mikulincer, M. & Shaver, P. (2008) Contributions of attachment theory and research to motivation science. In J.Y. Shah & W.L. Gardner (eds) Handbook of Motivational Science (pp.201–16). New York: Guilford, p.204.

256 Koski, L.R. & Shaver, P.R. (1997) Attachment and relationship satisfaction across the lifespan. In R.J. Sternberg & M. Hojjat (eds) Satisfaction in Close Relationships (pp.26–55). New York: Guilford, p.27.

257 The correspondence between Bowlby and Rene Spitz on smiling addressed just this issue (see PP/Bow/H.169). As an example of a largely pre-given response requiring only sufficient maturation and the absence of a grossly inhibiting environment, Bowlby offered the example of infant sucking. As an example of an essentially learnt response, Bowlby gave the example of throwing a ball.

258 Bowlby, J. (1973) Separation. New York: Basic Books, p.288.

259 Bowlby, J. (1969, 1982) Attachment, Volume 1, London: Penguin, p.78.

260 Ainsworth, M., Blehar, M., Waters, E., & Wall, S. (1978, 2015) Patterns of Attachment: A Psychological Study of the Strange Situation. Bristol: Psychology Press, p.11.

261 In a remarkable statement, Bowlby observed that this gives attachment behaviour a certain developmental and causal priority over the attachment relationship itself: ‘Attachment behaviour leads to the development of affectional bonds or attachments, initially between child and parent.’ Bowlby, J. (1980) Loss. London: Pimlico, p.39.

262 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.129. Bowlby later reflected that this flexibility is especially potent for humans: ‘the longer-lived an individual the more necessary is ontogenetic flexibility to enable it to adapt to changes in the environment’. Bowlby, J. (1982) Evolution theory. PP/Bow/H102.

263 Bowlby, J. (1957) An ethological approach to research on child development. British Journal of Medical Psychology, 30(4), 230–40, p.239.

264 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.160.

265 Bowlby, J. (1976) Human personality development in an ethological light. In G. Serban & A. Kling (eds) Animal Models in Human Psychobiology (pp.27–36). New York: Plenum Press, p.31. This idea has its origins for Bowlby not only in control systems theory, but also in his reading of Freud. Annotations by Bowlby dated 1960 on the ‘Project for a Scientific Psychology’ in Sigmund Freud: The Origins of Psychoanalysis. Copy held in the library of Human Development Scotland. Bowlby highlighted and starred the passage: ‘Every contrivance of a biological nature has limits to its efficiency, beyond which it fails. Such failures exhibit themselves as phenomena bordering on the pathological’ (368).

266 Bowlby, J. (1991) Ethological light on psychoanalytic problems. In P. Bateson (ed.) The Development and Integration of Behaviour: Essays in Honour of Robert Hinde (pp.301–14). Cambridge: Cambridge University Press, p.306.

267 Bowlby 1984 personal communication, cited in Harwood, I. (2003) Creative use of gender while addressing early attachment, trauma, and cross-cultural issues in a cotherapy group. Psychoanalytic Inquiry, 23, 697–712.

268 Bowlby, J. (1986) Interview with the BBC. PP/Bow/F.5/8.

269 Bowlby, J. (1958) Nature of the child’s tie to his mother. International Journal of Psycho-Analysis, 39, 350–73; Harlow, H.F. & Zimmermann, R.R. (1958) The development of affectional responses in infant monkeys. Proceedings of the American Philosophical Society, 102(5), 501–9.

270 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.228–9.

271 Bowlby, J. (1979) By ethology out of psychoanalysis: an experiment in interbreeding. (The Niko Tinbergen Lecture.) Animal Behaviour, 28(3), 649–56, p.651.

272 Marvin, R.S. (1972) Attachment, exploratory and communicative behavior of two, three and four year old children. Unpublished doctoral dissertation, University of Chicago. Discussed in Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.70.

273 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.140.

274 The point would need to be clarified repeatedly by early members of the second generation of attachment researchers. The most important contribution on this score was that of Everett Waters, who also demonstrated the claim empirically (Chapter 2).

275 Bowlby, J. (1958) The nature of the child’s tie to his mother. International Journal of Psycho-Analysis, 39, 350–73, p.351.

276 Bowlby, J. (1980, 1988) Caring for children. In A Secure Base (pp.1–21). London: Routledge, p.6; Bowlby, J. (1973) Letter to Scott Henderson, 30 July 1973. PP/Bow/J.9/98.

277 Bowlby, J. (1960) Grief and mourning in infancy and early childhood. Psychoanalytic Study of the Child, 15, 9–52: ‘I believe that the hypothesis now advanced would have been advocated earlier had it not happened that the phase of attachment to a mother figure was so late in being recognized and had not theory become preoccupied instead on the one hand with primary narcissism and on the other with orality’ (14).

278 Bowlby, J. (1979) Letter to Professor K.S. Adam, 12 February 1979. PP/Bow/J.9.2.

279 Bretherton, I. & Ainsworth, M.D.S. (1974) Responses of one-year-olds to a stranger in a strange situation. In M. Lewis & L.A. Rosenblum (eds) The Origin of Fear (pp.131–64). New York: Wiley. This was also discussed by Tony Ambrose, working within Bowlby’s group: Ambrose, T. (1960) The smiling and related responses in early human infancy: an experimental and theoretical study of their course and significance. Unpublished PhD thesis, Birkbeck College, London.

280 See Attili, G. & Hinde, R.A. (1986) Categories of aggression and their motivational heterogeneity. Evolution and Human Behavior, 7(1), 17–27.

281 Bowlby cited in Tanner, J.M. & Inhelder, B. (1960) Discussions on Child Development: Proceedings of the WTO Study Group of the Psychobiological Development of the Child, Vol. 4. London: Tavistock, p.40.

282 Ainsworth, M. (1997) Peter L. Rudnytsky—the personal origins of attachment theory: an interview with Mary Salter Ainsworth. Psychoanalytic Study of the Child, 52, 386–405: ‘PLR: Is there any sense in which your view and Bowlby’s diverge at all, or is there a complete meeting of the minds? MSA: Well, I think there’s more to the oedipal situation than he does. Bowlby just doesn’t talk about it’ (399); ‘In the oedipal situation, there’s no question in my mind that in the parents’ dynamics there is a lot of sexuality’ (402).

283 Ainsworth does not cite the transcript of these discussions, and she may not have owned a copy.

284 Tanner, J.M. & Inhelder, B. (1956) Discussions on Child Development: Proceedings of the WTO Study Group of the Psychobiological Development of the Child, Vol 1. London: Tavistock, p.184–5. For later considerations of the relationship between the attachment and sexual behavioural systems see e.g. Crittenden, P.M. (1998) Patterns of attachment and sexual behavior: risk of dysfunction versus opportunity for creative integration. In L. Atkinson & K.J. Zucker (eds) Attachment and Psychopathology (pp.47–93). New York: Guilford; Diamond, L.M. (2003) What does sexual orientation orient? A biobehavioral model distinguishing romantic love and sexual desire. Psychological Review, 110(1), 173–92.

285 Bowlby cited in Tanner, J.M. & Inhelder, B. (1960) Discussions on Child Development: Proceedings of the WTO Study Group of the Psychobiological Development of the Child, Vol 4. London: Tavistock, p.41.

286 Bowlby, J. (1969) Affectional bonds: their nature and origin. In H. Freeman (ed.) Progress in Mental Health (pp.319–27). London: J. & A. Churchill, p.323.

287 Bowlby, J. (1977, 1979) The making and breaking of affectional bonds. In The Making and Breaking of Affectional Bonds (pp.150–88). London: Routledge, p.157. The use of a child as an attachment figure by adults appears to have been deliberately neglected by subsequent attachment research, in part to maintain—against misunderstanding—the characterization of attachment as something a child shows to their adult caregiver. See e.g. Ainsworth, M.D.S. (1991) Attachments and other affectional bonds across the life cycle. In C.M. Parkes, J. Stevenson-Hinde, & P. Marris (eds) Attachment Across the Life Cycle (pp. 33–51). London: Routledge: ‘We talk of the bond of a mother to her child … a mother does not normally base her security on her relationship with her child, however eager she may be to give care and nurturance’ (40). The primary exception is Doherty, N.A. & Feeney, J.A. (2004) The composition of attachment networks throughout the adult years. Personal Relationships, 11(4), 469–88. Forty percent of participants with children reported using their child as a safe haven, as a secure base, and experiencing separation anxiety. However, both the attachment and caregiving systems can prompt separation anxiety and proximity-maintenance. This difficulty disentangling these systems may be a secondary reason why attachment researchers have neglected the phenomenon of use of children as attachment figures by adults.

288 This hinge seems to have been primarily the result of conceptual imprecision. It may at times have served as a ‘motte and bailey’ rhetorical strategy, with a poorly defendable but expansive outer area and a more defendable but narrower inner position. However, the fact that the term’s meaning slides around just as much in Bowlby’s private notes suggests that such a strategy was not his intent. On the ‘motte and bailey’ rhetorical strategy see Shackel, N. (2005) The vacuity of postmodernist methodology. Metaphilosophy, 36(3), 295–320.

289 E.g. Bowlby, J. (1984) Violence in the family as a disorder of the attachment and caregiving systems. American Journal of Psychoanalysis, 44, 9–27, p.11.

290 Bowlby, J. (1980) Loss. London: Pimlico, p.442.

291 Bowlby, J. (1960) Separation anxiety. International Journal of Psycho-Analysis, 41, 89–113, p.105; see also Steele, H. & Steele, M. (1998) Response to Cassidy, Lyons-Ruth and Bretherton: a return to exploration. Social Development, 7(1), 137–41.

292 Bowlby, J. (1986) Interview with the BBC. PP/Bow/F.5/8: ‘I’m often accused of exaggerating and various people point out that not every child who has had these sorts of experience comes to grief … First of all, maybe he’ll be vulnerable if things go wrong in his life … Another point is that, supposing it’s true that some children having had some pretty ghastly experiences nonetheless develop favourably … But then the question arises, what percentage of people who contract polio are left with long term paralysis? The answer is less than 1 per cent. The fact is 99 per cent get by. Now in the case of severe maternal deprivation—first of all I think that much more than 1 per cent suffer. I think the percentage is way up in the 20s or 30s. And the other thing is that crippling of personality is much more serious’. See also Bowlby, J. (1971) Letter to Michael Rutter, 6 October 1971. PP/Bow/J.9/161: ‘The more serious effects of experiences of the type we are discussing are when there are certain special combinations of variables present.’

293 Finsterer, J. & Hayman, J. (2014) Mitochondrial disorder caused Charles Darwin’s cyclic vomiting syndrome. International Journal of General Medicine, 7, 59–70.

294 Bowlby, J. (1987–90) Papers on Charles Darwin. Cambridge University Library. MS Add. 8884, marginalia on Liotti, G. (1986) Structural cognitive therapy. In Dryden, W. & Golden, W. (eds) Cognitive-Behavioural Approaches to Psychotherapy. Milton Keynes: Open University Press, p.125.

295 Bowlby, J. (1990) Charles Darwin: A Life. New York: Norton, p.6.

296 Two clear cases are Vicedo, M. (2013) The Nature and Nurture of Love: From Imprinting to Attachment in Cold War America. Chicago: University of Chicago Press, and Gottlieb, A. (2014) Is it time to detach from attachment theory? Perspectives from the West African rainforest. In H. Otto & H. Keller (eds) Different Faces of Attachment: Cultural Variations on a Universal Human Need (pp.187–214). Cambridge: Cambridge University Press.

297 Bretherton, I., Biringen, Z., Ridgeway, D., Maslin, C., & Sherman, M. (1989) Attachment: the parental perspective. Infant Mental Health Journal, 10(3), 203–21; Waters, E. & Cummings, E.M. (2000) A secure base from which to explore close relationships. Child Development, 71(1), 164–72: ‘To examine relations between early attachment experience and both secure base use and secure base support skills later in life and to examine them across contexts such as marriage, parenting, caring for adult parents, and requesting care from others. As currently formulated, attachment theory suggests that these are all organized by the same attachment control system’ (171).

298 Cranley, M.S. (1981) Development of a tool for the measurement of maternal attachment during pregnancy. Nursing Research, 30(5), 281–4.

299 Huffmeijer, R., van IJzendoorn, M.H., & Bakermans-Kranenburg, M.J. (2013) Ageing and oxytocin: a call for extending human oxytocin research to ageing populations—a mini-review. Gerontology, 59(1), 32–9: ‘In (soon-to-be) mothers, increases in plasma oxytocin concentrations over the course of pregnancy have been found to predict greater attachment to the unborn baby’ (33).

300 Bowlby, J. (1986) Attachment, Life-Span and Old Age. Eds J. Munnichs & B. Miesen. Utrecht: Van Loghum, p.11. See also Cassidy, J. (2000) The complexity of the caregiving system: a perspective from attachment theory. Psychological Inquiry, 11(2), 86–91, p.88.

301 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.240.

302 Bowlby, J. (1973) Letter to Scott Henderson, 30 July 1973. PP/Bow/J.9/98: the caregiving provided by the caregiving system ‘includes rather more’ than the bodily contact of the retrieval response and the affects of ‘interest, esteem and affection’. In particular, Bowlby urged Henderson to note that caregiving also included ‘encouragement, support, help and protection’. On the ambiguity between the ‘retrieval’ and ‘nurturance’ models of caregiving in Bowlby see Bell, D.C. & Richard, A.J. (2000) Caregiving: the forgotten element in attachment. Psychological Inquiry, 11(2), 69–83.

303 Bowlby, J. (1978) Caregiving. In ‘Emotion and feeling’. PP/Bow/H.5.

304 Bowlby, J. (1990) Interviewed in The Nuts and Bolts of Ben Bowlby (Channel 4).

305 On the important role of the affect ‘delight’ in the functioning of the caregiving system see e.g. Bernard, K. & Dozier, M. (2011) This is my baby: foster parents’ feelings of commitment and displays of delight. Infant Mental Health Journal, 32(2), 251–62. Delight and other positive communicated affect may also be a significant relay between the caregiving and attachment behavioural systems. Dozier, M. & Bernard, K. (2019) Coaching Parents of Vulnerable Infants: The Attachment and Biobehavioral Catch-up Approach. New York: Guilford: ‘In a personal communication (August 2001) Mary Main suggested that parental delight communicates to children how important they are to their parents, a sentiment she indicated Mary Ainsworth shared’ (57).

306 E.g. Burkett, J.P., Andari, E., Johnson, Z.V., Curry, D.C., de Waal, F.B., & Young, L.J. (2016) Oxytocin-dependent consolation behavior in rodents. Science, 351(6271), 375–8.

307 Vicedo, M. (2011) The social nature of the mother’s tie to her child: John Bowlby’s theory of attachment in post-war America. British Journal for the History of Science, 44(3), 401–26, p.423. For an interesting illustration of how the concept of ‘maternal instinct’ is now treated as a cultural form rather than part of human nature by attachment researchers see Murphy, A., Steele, M., & Steele, H. (2013) From out of sight, out of mind to in sight and in mind: enhancing reflective capacities in a group attachment-based intervention. In J.E. Bettmann & D.D. Friedman (eds) Attachment-Based Clinical Work with Children and Adolescents (pp.237–57). New York: Springer. For a plea for a more measured and less polemical discussion of attachment research, and its heterogeneity see Duschinsky, R., van IJzendoorn, M., Foster, S., Reijman, S., & Lionetti, F. (2019) Attachment histories and futures: reply to Vicedo’s ‘Putting attachment in its place’. European Journal of Developmental Psychology, 17(1).

308 Ainsworth, M.D.S. (1991) Attachments and other affectional bonds across the life cycle. In: C.M. Parkes, J. Stevenson-Hinde, & P. Marris (eds) Attachment Across the Life Cycle (pp. 33–51). London: Routledge: ‘Nowadays in Western societies these traditional roles are being challenged, and many couples are experimenting with alternative ways of providing adequate care to their infants and young children. The more successful solutions seem to involve the male taking more responsibility for direct caregiving to the children’ (42).

309 Bowlby, J. (1980, 1988) Caring for children. In A Secure Base (pp.1–21). London: Routledge, p.5.

310 Bowlby, J. (1969) Ape and apex. BBC Radio, recorded 16 October 1969. PP/Bow/F.5/5: ‘Our capacity to learn diversity of behaviour is itself genetically determined e.g. speech and language … Development of maternal behaviour dependent on appropriate experience, cf. Hinde on false dichotomy. Development of mothering greatly helped by 1. Opportunity to touch, examine infant; 2. Smiling of infant.’

311 Bowlby, J. (1933–36) Anxiety, guilt, etc: old papers. PP/Bow/D.1/2/13

312 Bowlby, J. (1973) Separation. London: Penguin, pp.231–2.

313 Ibid. p.117. On threatening caregiver behaviour as a predictor of approach/avoidance conflict in infants see Jacobvitz, D., Leon, K., & Hazen, N. (2006) Does expectant mothers’ unresolved trauma predict frightened/frightening maternal behavior? Risk and protective factors. Development & Psychopathology, 18(2), 363–79.

314 Bowlby, J. (1979) Letter to Henry Hansburg, 21 February 1979. PP/Bow/J.9/90.

315 Bowlby, J. (1977, 1979) The making and breaking of affectional bonds. In The Making and Breaking of Affectional Bonds (pp.150–88). London: Routledge.

316 Bowlby, J. (1986) Letter to John Birtchnell, 14 October 1986. PP/Bow/J.9/22: ‘Compulsive caregiving is based mainly on a fear of seeking love for fear of being rejected.’

317 Bowlby, J. (1980) Loss. London: Penguin, p.411.

318 Bowlby, J. (1988) A Secure Base. New York: Basic Books.

319 Personal communication by John Bowlby to Scott Henderson, cited in Henderson, S. (1974) Care-eliciting behavior in man. Journal of Nervous and Mental Disease, 159(3), 172–81.

320 Bowlby, J. (1984) Violence in the family as a disorder of the attachment and caregiving systems. American Journal of Psychoanalysis, 44, 9–27, p.18. Crittenden and DiLalla would later distinguish placatory behaviour from caregiving, as different but related behavioural repertoires: Crittenden, P.M. & DiLalla, D.L. (1988) Compulsive compliance: the development of an inhibitory coping strategy in infancy. Journal of Abnormal Child Psychology, 16(5), 585–99.

321 In the 2000s, social psychologists would highlight the significant limitations of Bowlby’s image in Attachment, Volume 1 of retrieval as the central form and symbol of the caregiving system. Mikulincer, M., Shaver, P.R., Gillath, O., & Nitzberg, R.A. (2005) Attachment, caregiving, and altruism: boosting attachment security increases compassion and helping. Journal of Personality and Social Psychology, 89(5), 817–39, p.818; Collins, N.L., Guichard, A.C., Ford, M.B., & Feeney, B.C. (2006) Responding to need in intimate relationships: normative processes and individual differences. In M. Mikulincer & G. S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.149–89). New York: Guilford.

322 Solomon, J. & George, C. (1996) Defining the caregiving system: toward a theory of caregiving. Infant Mental Health, 17(3), 183–97. See also Stern, D. (1995) The Motherhood Constellation. New York: Basic Books; Lieberman, A. (1996) Aggression and sexuality in relation to toddler attachment: implications for the caregiving system. Infant Mental Health, 17(3), 276–92; Heard, D. & Lake, B. (1997) The Challenge of Attachment for Caregiving. London: Routledge; Feeney, B.C. & Collins, N.L. (2001) Predictors of caregiving in adult intimate relationships: an attachment theoretical perspective. Journal of Personality and Social Psychology, 80, 972–94. On changes in public attention to the mother as ‘subject’ see e.g. Rich, A. (1995) Of Woman Born: Motherhood as Experience and Institution. New York: Norton.

323 George, C. & Solomon, J. (1999) The development of caregiving: a comparison of attachment theory and psychoanalytic approaches to mothering. Psychoanalytic Inquiry, 19(4), 618–46. Seemingly unaware of earlier critiques of Bowlby by Mary Main (Chapter 3) and Solomon and George, anthropologists have developed this point in detail especially over the past decade, e.g. Carlson, V.J. & Harwood, R.L (2014) The precursors of attachment security: behavioral systems and culture. In H. Otto & H. Keller (eds) Different Faces of Attachment. Cambridge: Cambridge University Press, 278–303.

324 Bowlby, J. (1953) The Roots of Parenthood. London: National Children’s Home, p.16.

325 Ibid. p.14.

326 Mayhew, B. (2006) Between love and aggression: the politics of John Bowlby. History of the Human Sciences, 19(4), 19–35.

327 Hrdy, S.B. (1999) Mother Nature: A History of Mothers, Infants and Natural Selection. New York: Pantheon, p.xiii.

328 Hrdy, S.B. (2005) Evolutionary context of human development: the cooperative breeding model. In L.A.C.S. Carter, K.E. Grossmann, S.B. Hrdy, M.E. Lamb, S.W. Porges, & N. Sachser (eds) Attachment and Bonding: A New Synthesis (pp. 9–32). Cambridge, MA: MIT Press. More recently, see Myowa, M. & Butler, D.L. (2017) The evolution of primate attachment: beyond Bowlby’s rhesus macaques. In H. Keller & K.A. Bard (eds) The Cultural Nature of Attachment. Contextualizing Relationships and Development (pp.53–68). Cambridge, MA: MIT Press.

329 Hrdy, S.B. (1999) Mother Nature: A History of Mothers, Infants and Natural Selection. New York: Pantheon, p.495, discussing Tronick, E.Z., Winn, S., & Morelli, G.A. (1985) Multiple caretaking in the context of human evolution: why don’t the Efé know the western prescription for child care? In M. Reite & T. Field (eds) The Psychobiology of Attachment and Separation (pp.293–322). Orlando, FL: Academic Press. For a more recent discussion of multiple caregiving see Meehan, C.L. & Hawks, S. (2013) Cooperative breeding and attachment among the Aka foragers. In N. Quinn & J.M. Mageo (eds) Attachment Reconsidered: Cultural Perspectives on a Western Theory. London: Palgrave.

330 Hrdy appears to have been, understandably, unaware of Bowlby who made exactly this point in his 1953 book The Roots of Parenthood. The book was not distributed in the USA. Bowlby, J. (1953) The Roots of Parenthood. London: National Children’s Home.

331 Hrdy, S.B. (1999) Mother Nature: A History of Mothers, Infants and Natural Selection. New York: Pantheon, p.536.

332 Hazen, N.L., Allen, S.D., Christopher, C.H., Umemura, T., & Jacobvitz, D.B. (2015) Very extensive nonmaternal care predicts mother–infant attachment disorganization: convergent evidence from two samples. Development & Psychopathology, 27(3), 649–61; van IJzendoorn, M.H., Bakermans-Kranenburg, M.J., Duschinsky, R., & Skinner, G.C.M. (2019) Legislation in search of ‘good-enough’ care arrangements for the child: a quest for continuity of care. In J. Dwyer (eds) Handbook of Children and the Law. Oxford: Oxford University Press. Belsky and colleagues have, however, found that the extent of non-familial care is associated with a small increase in child conduct problems and impulsivity, with maternal sensitivity as a moderator: e.g. Burchinal, M.R., Lowe Vandell, D., & Belsky, J. (2014) Is the prediction of adolescent outcomes from early child care moderated by later maternal sensitivity? Results from the NICHD study of early child care and youth development. Developmental Psychology, 50(2), 542–53.

333 Keller, H. & Chaudhary, N. (2017) Is the mother essential for attachment? Models of care in different cultures. In H. Keller & K.A. Bard (eds) The Cultural Nature of Attachment (pp.109–37). Cambridge, MA: MIT Press, p.112. This claim rests on a characteristic rhetorical strategy among many critics, in which early statements by Bowlby are used to characterize all subsequent empirical attachment research (Chapter 2).

334 E.g. Sagi, A., van IJzendoorn, M.H., Aviezer, O., Donnell, F., Koren-Karie, N., Joels, T., & Harl, Y. (1995) Attachments in a multiple-caregiver and multiple-infant environment: the case of the Israeli kibbutzim. Monographs of the Society for Research in Child Development, 60(2–3), 71–91.

335 Keller, H. & Chaudhary, N. (2017) Is the mother essential for attachment? Models of care in different cultures. In H. Keller & K.A. Bard (eds) The Cultural Nature of Attachment (pp.109–37). Cambridge, MA: MIT Press.

336 E.g. van IJzendoorn, M.H. (2005) Attachment in social networks: toward an evolutionary social network model. Human Development, 48(1–2), 85–8; De Schipper, J.C., Stolk, J., & Schuengel, C. (2006) Professional caretakers as attachment figures in day care centers for children with intellectual disability and behavior problems. Research in Developmental Disabilities, 27, 203–16. There has been a somewhat greater focus on wider networks among attachment researchers in the social psychological tradition, but not with respect to caregiving. Gillath, O., Karantzas, G.C., & Lee, J. (2019) Attachment and social networks. Current Opinion in Psychology, 25, 21–5.

337 Gibson, M.A. & Mace, R. (2005) Helpful grandmothers in rural Ethiopia: a study of the effect of kin on child survival and growth. Evolution and Human Behavior, 26(6), 469–82.

338 However, there has been a growing literature on fathers and on co-parenting (Chapter 4). An important development has been the inclusion of fathers within research on attachment-based interventions: e.g. Iles, J. E., Rosan, C., Wilkinson, E., & Ramchandani, P.G. (2017) Adapting and developing a video-feedback intervention for co-parents of infants at risk of externalising behaviour problems (VIPP-Co): a feasibility study. Clinical Child Psychology and Psychiatry, 22(3), 483–99.

339 The limited acknowledgement of the exploratory system in Bowlby’s work might be linked to the scholarly context of the 1960s. Social learning was a pivotal discourse on child development, and one with which Bowlby had some sympathies. Both Bowlby and social learning theorists were concerned with the consistency and responsiveness of caregiving. However, he was keen to distinguish attachment behaviour from the effect of behaviourist conditioning, which was the dominant frame at the time for conceptualizing social learning. A more detailed elaboration of the exploratory system would have confronted Bowlby with the need to tackle social learning theory head-on, which would have been a demanding task, as demonstrated when it eventually fell to Ainsworth to clarify the differences between the attachment and social learning paradigms. See Bowlby, J. (1961) Comment on paper by Dr Gewirtz. In B.M. Foss (ed.) Determinants of Infant Behaviour (pp.301–304). London: Methuen; Ainsworth, M. (1972) Attachment and dependency: a comparison. In J. Gewirtz (ed.) Attachment and Dependency (pp. 97–137). Washington, DC: Winston. An additional contemporary dynamic was that creativity and play research were just getting off the ground. Attention to the exploratory system may well have been fed by this source, as shown by Bowlby’s appeal to Winnicott’s concept of creative play. However, theories of creativity and play were still weak when Bowlby was working on Attachment, Volume 1.

340 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.238, discussing Hinde, R.A. (1954) Factors governing the changes in strength of a partially inborn response. 1. The nature of the response, and an examination of its course. Proceedings of the Royal Society B, 142, 306–31. Other important sources for Bowlby were Berlyne, D.E. (1950) Novelty and curiosity as determinants of exploratory behaviour. British Journal of Psychology, 41, 68–80; Rheingold, H.L. (1963) Controlling the infant’s exploratory behavior. In B.M. Foss (ed.) Determinants of Infant Behaviour, Vol. 2 (pp.171–217). London: Methuen.

341 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, pp.237–38.

342 Ibid. p.197. See also Marvin, R.S., Britner, A.A., & Russell, B.A. (2016) Normative development: the ontogeny of attachment in childhood. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research, and Clinical Applications 3rd edn (pp. 273–90). New York: Guilford.

343 Bowlby, J. (1977, 1979) The making and breaking of affectional bonds. In The Making and Breaking of Affectional Bonds (pp.150–88). London: Routledge, p.183–4; cf. Winnicott, D.W. (1967) The location of cultural experience. International Journal of Psychoanalysis, 48, 368–72.

344 Some of Bowlby’s examples are quite culturally specific, e.g. on novelty and pleasure in family holidays, Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.41. Whether exploration of complex/novel stimuli can be interpreted as equivalent to Winnicott’s concept of ‘play’ depends on what is understood by the latter term, an understanding shaped in important ways by culture. Lancy, D.F. (2007) Accounting for variability in mother–child play. American Anthropologist, 109(2), 273–84.

345 Ainsworth, M. (1984) Attachment. In N.S. Endler & J. McVicker Hunt (eds) Personality and the Behavioral Disorders. New York: Wiley, 559–602, p.565. See also Grossmann, K., Grossmann, K.E., Kindler, H., & Zimmermann, P. (2008) A wider view of attachment and exploration: the influence of mothers and fathers on the development of psychological security from infancy to young adulthood. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research, and Clinical Applications, 2nd edn (pp.857–79). New York: Guilford.

346 Main, M. (1981) Avoidance in the service of proximity: a working paper. In K. Immelmann, G.W. Barlow, L. Petrinovich, M. Biggar Main (eds) Behavioral Development: The Bielefeld Interdisciplinary Project (pp.651–93). Cambridge: Cambridge University Press, p.554–5.

347 E.g. Bowlby, J., Robertson, J., & Rosenbluth, D. (1952) A two-year-old goes to hospital. Psychoanalytic Study of the Child 7, 82–94, p.83. He also discussed the use of distraction activities in adults: Bowlby, J. (1980) Loss. London: Pimlico: ‘One or more behavioural systems within a person may be deactivated, partially or completely. When that occurs one or more other activities may come to monopolize the person’s time and attention, acting apparently as diversions’ (64).

348 Ainsworth, M. (1992) A consideration of social referencing in the context of attachment theory and research. In S. Feinman (ed.) Social Referencing and the Construction of Reality in Infancy (pp.349–67). New York: Plenum Press, p.361.

349 Ainsworth, M. & Bell, S. (1970) Attachment, exploration, and separation: illustrated by the behavior of one-year-olds in a strange situation. Child Development, 41(1), 49–67, p.51. The underdevelopment of the idea of the exploratory system in Bowlby’s work is widely acknowledged, e.g. Gullestad, S.E. (2001) Attachment theory and psychoanalysis: controversial issues. Scandinavian Psychoanalytic Review, 24, 3–16. Subsequent to Ainsworth, the major sustained interest in the relationship between exploration and attachment has been among clinically-focused commentators, who have been concerned with what attachment research can suggest about moderators of our capacity to explore the minds of others and learn from experience e.g. Fonagy, P., Luyten, P., Allison, E., & Campbell, C. (2017) What we have changed our minds about: part 2. Borderline personality disorder, epistemic trust and the developmental significance of social communication. Borderline Personality Disorder and Emotion Dysregulation, 4, 9; Golding, K. & Hughes, D. (2012) Creating Loving Attachments: Parenting with PACE to Nurture Confidence and Security in the Troubled Child. London: Jessica Kingsley; Powell, B., Cooper, G., Hoffman, K., & Marvin, B. (2016) The Circle of Security Intervention. New York: Guilford.

350 Bowlby had also been very interested in human fear behaviour since his work as an army psychiatrist. However, the proximal cause of Bowlby’s attention to the topic in the early 1970s appears to have been the proposal of fear as a distinct behavioural system by the American developmentalist Gordon Bronson. Bronson, G.W. (1968) The development of fear in man and other animals. Child Development, 39(2), 409–31. Bronson observed that the threshold for activation of the fear behavioural system seemed to be lower for institutionally reared human children and monkeys, and that fear of strangers seemed to develop at around the same time as discrimination of attachment figures. However, Bowlby felt that Bronson’s distinctions were not adequately sharp between fear, wariness, and anxiety. In a wider context, cultural discourses appealing to and distinguishing ‘fear’ and ‘anxiety’ were gaining ground from the late 1960s, and may have contributed to the salience of the topic for Bronson and Bowlby. See Jenkins, P. (2006) Decade of Nightmares: The End of the Sixties and the Making of Eighties America. Oxford: Oxford University Press; Bourke, J. (2006) Fear: A Cultural History. Emeryville, CA: Shoemaker & Hoard.

351 Bowlby, J. (1973) Separation. London: Penguin, p.117.

352 Ibid. Appendix III.

353 Bowlby, J. (1968) Types of fear response. PP/Bow/H.209.

354 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books: ‘Hinde (1970) reports a finding by Hogan that, in chicks, withdrawal occurs from stimuli at high intensity (and some others) whereas freezing is elicited by stimuli that are strange, novel, or surprising’ (154).

355 Ibid. p.114.

356 Ibid. p.157.

357 Bowlby, J. (1973) Wariness. Unpublished manuscript, June 1973. PP/BOW/J.9/39. The same distinction was used by Bretherton, I. & Ainsworth, M.D.S. (1974) Responses of one-year-olds to a stranger in a strange situation. In M. Lewis & L.A. Rosenblum (eds) The Origin of Fear (pp.131–64). New York: Wiley.

358 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.119.

359 Bowlby, J. (1970, 1979) Self-reliance and some conditions that promote it. In The Making and Breaking of Affectional Bonds (pp.124–49). London: Routledge, p.147.

360 Bowlby, J. (1981) Psychoanalysis as a natural science. International Review of Psycho-Analysis, 8, 243–56, p.248.

361 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.210.

362 Parton, N. (1985) The Politics of Child Abuse. Basingstoke: Macmillan; Hacking, I. (1991) The making and molding of child abuse. Critical Inquiry, 17(2), 253–88; Ferguson, H. (2004) Protecting Children in Time: Child Abuse, Child Protection and the Consequences of Modernity. London: Palgrave.

363 Bowlby, J. (1973) Draft material towards Separation. PP/Bow/K.5./17.

364 Bowlby, J. (1973) Letter to Ainsworth, 29 October, 1973. PP/Bow/J.1/33; Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books: ‘Conflict can easily occur, for example, whenever a stimulus situation that elicits both escape and attachment behaviour in an individual happens to be situated between that individual and his attachment figure; a familiar instance is when a barking dog comes between a child and his mother’ (116). The typescript is available at PP/Bow/K.5./17.

365 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.117. The typescript is available at PP/Bow/K.5./17.

366 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin: ‘In Tom, it can be said, there is a tendency to appraise certain situations in such a way that a behavioural system is activated that results in his attacking his little sister and biting her’ (118).

367 Bowlby’s most sustained attention to aggression appears in Durbin, E.F.M. & Bowlby, J. (1939) Personal Aggressiveness and War. London: Kegan Paul. Little from this book fed through into his later reflections on aggression in Separation. Those who knew Bowlby personally have often remarked that he did not seem at all comfortable with aggression, which may have contributed to a theoretical antipathy.

368 This emphasis has been echoed by later attachment researchers: for instance, Cassidy’s influential introduction to the Handbook of Attachment discusses the attachment, caregiving, exploratory, sociable, and fear systems, but does not discuss anger as a behavioural system. Cassidy, J. (1999) The nature of the child’s ties. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research, and Clinical Applications (pp.3–21). New York: Guilford. This selection of behavioural systems was maintained by Cassidy in the second (2008) and third (2016) editions.

369 Main, M. (1993) Les bébés et leurs colères. In M.C. Busnel (eds) Le Langage des Bébés, Savons-nous Entendre? (pp.17–91). Paris: Grancher; Main, M. (1995) Recent studies in attachment: overview, with selected implications for clinical work. In S. Goldberg, R. Muir, & J. Kerr (eds) Attachment Theory: Social, Developmental and Clinical Perspectives (pp.407–70). Hillsdale, NJ: Analytic Press, p.460. In work with Jude Cassidy, Main describes the role of aggressive behaviour within displays of dominance from children towards their attachment figures (Chapter 3), but without ever referring this back to Bowlby’s discussion of aggression. An important disjuncture lay in the fact that Bowlby’s accounts of aggression characterized primarily frustration and protest in the service of attachment, whereas what Cassidy and Main were seeing was more like use of aggression in the service of dominance in the service of attachment.

370 Shaver, P.R., Segev, M., & Mikulincer, M. (2011) A behavioral systems perspective on power and aggression. In P.R. Shaver & M. Mikulincer (eds) Human Aggression and Violence: Causes, Manifestations, and Consequences (pp.71–87). Washington, DC: American Psychological Association, p.71.

371 Van Zeijl, J., Mesman, J., van IJzendoorn, M.H., et al. (2006) Attachment-based intervention for enhancing sensitive discipline in mothers of 1- to 3-year-old children at risk for externalizing behavior problems: a randomized controlled trial. Journal of Consulting and Clinical Psychology, 74(6), 994–1005.

372 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.319.

373 Ainsworth, M. (1997) Peter L. Rudnytsky—the personal origins of attachment theory: an interview with Mary Salter Ainsworth. Psychoanalytic Study of the Child, 52, 386–405, p.403.

374 E.g. Issroff, J. (2005) Donald Winnicott and John Bowlby: Personal and Professional Perspectives. London: Karnac: ‘Bowlby disliked what he called sometimes “portmanteau” and sometimes “umbrella” or “omnibus” words like “aggression”, which covered too many different broad possibilities and created confusion in the way they were used’ (56).

375 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.285.

376 Ibid. p.288.

377 Ibid. p.285.

378 Ibid. p.287.

379 Waters, E. & Sroufe, L.A. (1983) A road careened into the woods: comments on Dr Morrison’s commentary. Developmental Review, 3(1), 108–14, p.223.

380 Knox, J. (2005) Sex, shame and the transcendent function: the function of fantasy in self development. Journal of Analytical Psychology, 50, 617–39: ‘I was fortunate enough to hear John Bowlby lecture on one occasion and one comment he made, almost in passing, remains imprinted on my mind—it was that children can survive the experience that their hate may drive a parent away but to have one’s love rejected is intolerable’ (625).

381 Waters, E., Posada, G., Crowell, J., & Lay, K.L. (1993) Is attachment theory ready to contribute to our understanding of disruptive behavior problems? Development & Psychopathology, 5(1–2), 215–24, p.220.

382 Shaver, P., Schwartz, J., Kirson, D., & O’Connor, C. (1987) Emotion knowledge: further exploration of a prototype approach. Journal of Personality and Social Psychology, 52(6), 1061–86, p.1078; Fonagy, P. (2003) The violence in our schools: what can a psychoanalytically informed approach contribute? Journal of Applied Psychoanalytic Studies, 5(2), 223–38, p.230.

383 Annotations by Bowlby dated 1958 on Freud’s essay ‘Repression’ in The Standard Edition of the Complete Psychological Works of Sigmund Freud, Vol. 14. Copy held in the library of Human Development Scotland. Highlighted and underlined: ‘Repression does not hinder the instinctual representative from continuing to exist in the unconscious, from organising itself further, putting out derivatives and establishing connections’ (p.149). Bowlby marginalia: ‘Very much my position.’

384 Bowlby, J. (1980) Loss. London: Pimlico: Information about experience ‘can be retained long enough outside consciousness in a temporary buffer store for it to influence judgement, autonomic responses and, I believe, mood’ (49).

385 Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, p.138.

386 Bowlby’s concerns resembled closely those of Laplanche and Pontalis: ‘When operations as diverse as, say, rationalisation, which brings complex intellectual mechanisms into play, and turning against the self, which is a “vicissitude” of the instinctual aim, are attributed to a single function, and when the same term “defence” connotes such a truly compulsive operation as “undoing what has been done” as well as the search for a form of “working-off” after the fashion of certain kinds of sublimation, then it may well be asked whether the concept in question is a really operational one.’ Laplanche, J. & Pontalis, J.B. (1973) The Language of Psycho-Analysis, trans. D. Nicholson-Smith. London: Karnac, p.109.

387 Duschinsky, R. & White, K. (eds) (2019) Trauma and Loss: Key Texts from the John Bowlby Archive. London: Routledge.

388 Bowlby, J. (1962) Defences that Follow Loss: Causation and Function. PP/Bow/D.3/78.

389 Klein, M. (1932) The Psychoanalysis of Children. London: Hogarth, p.226.

390 The concept of coping strategies was just entering the academic literature in the early 1960s. E.g. Murphy, L.B. (1960) Coping devices and defense mechanisms in relation to autonomous ego functions. Bulletin of the Menninger Clinic, 24(3), 144–53. The concept would gain ground in academic psychology, until it detonated into widespread use with Lazarus. R.S. & Folkman, S. (1984) Coping and adaptation. In W.D. Gentry (ed.) The Handbook of Behavioral Medicine (pp.282–325). New York: Guilford.

391 Fairbairn, W.R.D. (1929, 1994) Dissociation and repression. In E. Fairbain & D.E. Scharff (eds) From Instinct to Self: Selected Papers of W.R.D. Fairbairn (Vol. II). New York: Aronson Publishing; Sandler, J., Joffe, W.G., Baker, S., & Burgner, M. (1965) Notes on obsessional manifestations in children. The Psychoanalytic Study of the Child, 20(1), 425–38. Bowlby’s annotations on and correspondence with Sandler are of special interested in this regard. Bowlby was intrigued by the fact that Sandler was not always able to hold on to the distinction between cause and function even despite his best efforts. See Bowlby, J. (1965) Comments on Joffe and Sandler 1965 ‘Notes on Pain, Depression and Individuation’. PP/Bow/J.9/168-9.

392 Bowlby, J. (1962) Defences that Follow Loss: Causation and Function. PP/Bow/D.3/78.

394 Bowlby, J. (1980) Loss. London: Pimlico, p.67.

395 Bowlby, J. (1961) Letter to Dr Robert Hinde, 24 January 1961. PP/Bow/H224: ‘The whole psychoanalytic notion of defence is confused and I want to explore the notion that at the time of onset of what is later called repression one motivational system is blocked of expression and another, incompatible with it, evoked: both continue active but out of communication with each other. In the human, one of them is likely to be more accessible to consciousness than the other … Supposing this is a correct picture of events, the problem now becomes to define the conditions that give rise to this state of affairs and their mode of operation.’

396 Among the few attachment researchers to have applied the concept are Bretherton, and Solomon and George. It is revealing about the underdeveloped status of the concept in Bowlby’s published writings that their treatments differ vastly from one another, and that Bretherton expressed hesitancy as to whether she fully grasped Bowlby’s meaning. See Bretherton, I. (2005) In pursuit of the internal working model construct and its relevance to attachment relationships. In K.E. Grossmann, K. Grossmann, & E. Waters (eds) Attachment from Infancy to Adulthood: The Major Longitudinal Studies (pp.13–47). New York: Guilford; Solomon, J. & George, C. (2011) Disorganization of maternal caregiving across two generations. In J. Solomon & C. George (eds) Disorganized Attachment & Caregiving (pp. 25–51). New York: Guilford.

397 Bowlby, J. (1962) Defences that Follow Loss: Causation and Function. PP/Bow/D.3/78; cf. Bowlby, J. (1956) Annotations on Charles Brenner ‘The nature and development of the concept of repression in Freud’s writings’. PP/BOW/J.9/33.

398 Freud, S. (1933, 2001) New introductory lectures on psychoanalysis. In The Standard Edition of the Complete Psychological Works of Sigmund Freud, Vol. 22 (pp.1–182). London: Vintage, p.73. In Loss, Bowlby acknowledged some parallels between his concept of segregation and Freud’s concept of ‘splitting’, though he felt that the latter term had accrued too much baggage to offer much clarity. ‘Splitting’ would later feature within Crittenden’s reinterpretation of Bowlby’s information processing model: Crittenden, P.M. (2016) Raising Parents: Attachment, Representation, and Treatment (2nd edn). London: Routledge.

399 See e.g. Lemma, A., Target, M., & Fonagy, P. (2011) Brief Dynamic Interpersonal Therapy: A Clinician’s Guide. Oxford: Oxford University Press: ‘Although there is never a direct correspondence between external and internal, as what is internal reflects the operation of defensive processes that distort what is taken in from the outside’ (95).

400 Bailey, H.N., Redden, E., Pederson, D.R., & Moran, G. (2016) Parental disavowal of relationship difficulties fosters the development of insecure attachment. Revue Canadienne des Sciences du Comportement, 48(1), 49–59, p.50.

401 In this regard, Defences that Follow Loss is a forerunner of the priorities and concerns of developmental psychopathology (Chapter 4).

402 Bowlby, J. (1962) Defences that Follow Loss: Causation and Function. PP/Bow/D.3/78. On Bernard see Gross, C.G. (1998) Claude Bernard and the constancy of the internal environment. The Neuroscientist, 4(5), 380–85.

403 Main would later criticize Bowlby on this point, suggesting that he placed too much emphasis on information that was difficult to accept because integration would be painful, and not enough emphasis on information that was difficult to accept simply as a result of human cognitive biases and developmental processes, such as the difficulties of a three year old in holding in mind contradictory qualities of a single person. Main, M. (1991) Metacognitive knowledge, metacognitive monitoring, and singular (coherent) vs. multiple (incoherent) models of attachment: some findings and some directions for future research. In P. Marris, J. Stevenson-Hinde, & C. Parkes (eds) Attachment Across the Life Cycle (pp.127–59). New York: Routledge, p.138–9.

404 Bowlby, J. (1962) Defences that Follow Loss: Causation and Function. PP/Bow/D.3/78.

405 Indeed, the strength of his emphasis on segregation as a problem in his book on Loss would begin a divergence between two rather different trends in subsequent attachment theory: one that focused on whether internal working models were positive, and one that focused on whether behavioural systems were flexibly used and open to revision. The first to call attention to this ambiguity was van IJzendoorn, M.H., Tavecchio, L.W.C., Goossens, F.A., & Vergeer, M.M. (1982) Opvoeden in geborgenheid: Een kritische analyse van Bowlby’s attachmenttheorie. Amsterdam: Van Loghum Slaterus, pp.61–2.

406 Bowlby, J. (1964) Segregation of psychic systems. PP/Bow/H10. See also Reisz, S., Duschinsky, R., & Siegel, D.J. (2018) Disorganized attachment and defense: exploring John Bowlby’s unpublished reflections. Attachment & Human Development, 20(2), 107–34.

407 Bowlby, J. (1980) Loss. London: Pimlico, p.348.

408 Ibid. p.66.

409 Crittenden, P.M. (1997) Truth, error, omission, distortion, and deception: the application of attachment theory to the assessment and treatment of psychological disorder. In S.M. Clancy Dollinger & L.F. DiLalla (eds) Assessment and Intervention Issues Across the Lifespan (pp.35–76). London: Lawrence Erlbaum; Bretherton, I. (2005) In pursuit of the internal working model construct and its relevance to attachment relationships. In K.E. Grossmann, K. Grossmann, & E. Waters (eds) Attachment from Infancy to Adulthood: The Major Longitudinal Studies (pp.13–47). New York: Guilford.

410 Bowlby, J. (1960) Theory of Defence. JB notes, 1960-63. PP/Bow/H10.

411 Bowlby, J. (1940) The influence of early environment in the development of neurosis and neurotic character. International Journal of Psycho-Analysis, 21, 154–78, p.173.

412 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books: ‘the regulatory systems that maintain a steady relationship between an individual and his familiar environment can be regarded as an “outer ring” of life-maintaining systems complementary to the “inner ring” of systems that maintain physiological homeostasis’ (180).

413 Bowlby, J. (1978, 1988) Psychoanalysis as art and science. In A Secure Base (pp.43–64). London: Routledge, p.58.

414 Bowlby, J. (1960) Theory of Defence. JB notes, 1960-63. PP/Bow/H10: ‘Hypothetical function of narrowing of attention. Narrowing of attention is achieved by reducing input, but sensory and cognitive. The advantages are: a) to cut-out irrelevant and confusing input; b) to cut-out input that in fact evokes other motivational systems because such system would distract the organism from the task in hand by reducing relevant input; c) to cut-out input that might require abandonment of current organisation of a motivational system, and consequently reorganisation with its attendant layer of inability to act—disorganisation.’ Some commentators have assumed that all ‘detached’ behaviours observed by Robertson represented dissociative phenomena, e.g. Barach, P. (1991) Multiple personality disorder as an attachment disorder. Dissociation, 4, 117–23. Whilst avoidance may perhaps have relevant prospective links with dissociation (Chapter 4), Bowlby’s thinking about different kinds of defensive exclusion acknowledges their differences.

415 A remark by Ursula Bowlby offers an illustration. She described her husband as, for biographical reasons, incapable of feeling the emotion of fear. She considered this an aspect of Bowlby’s general inhibition of negative feelings, and a contributing factor to his indomitable courage as a public intellectual and theoretician. Again, this would be Level 3 defensive exclusion. Ursula Bowlby interview with Robert Keren, cited in Karen, R. (1998) Becoming Attached: First Relationships and how they Shape our Capacity to Love. Oxford: Oxford University Press, p.469.

416 Bowlby, J. (1962) Defences that Follow Loss: Causation and Function. PP/Bow/D.3/78: ‘I am introducing the generic term “to segregate” and “segregated process”; they denote any process that creates barriers to communication and interaction between one psychic system and another … Other additional terms are required for the many other particular sorts of segregating process.’ Handwritten marginal note: ‘dissociation’. See also Bowlby, J. (1961) Processes of mourning. International Journal of Psycho-Analysis, 42, 317–40, p.336.

417 Bowlby, J. (1982) Outline of attachment theory. PP/Bow/H.260.

418 E.g. Bowlby J., Robertson, J., & Rosenbluth, D. (1952) A two-year-old goes to hospital. Psychoanalytic Study of the Child 7, 82–94: ‘During the hospital experience these splits were relatively brief; after a few minutes of blankness she “came to” and responded to her real mother … This, perhaps, helps us to understand how during longer experiences of separation this split can develop to a point where integration on reunion with the mother is no longer automatic and the child is unable to link his need for a good mother and his hatred of a frustrating one to an individual woman’ (86).

419 Bowlby, J. (1976) Human personality development in an ethological light. In G. Serban & A. Kling (eds) Animal Models in Human Psychobiology (pp.27–36). New York: Plenum Press: Ethology and psychoanalysis fit so well together because ‘both were interested in the effect of early experience on later development’ and ‘in conflict arising from social situations’ (28). As well as Robert Hinde, another important influence on Bowlby’s reflections on conflict behaviour was von Holst, E. & von Saint Paul, V. (1963) On the functional organisation of drives. Animal Behaviour, 11, 1–20. They detail seven kinds of response to the conflict of behavioural responses: display of one toned by the other; averaging of the responses; alternation between the two; each cancelling the other out; the production of a third rather different kind of response (e.g. pecking and feeing becomes threat screeching); the dominance of one which masks weak expressions of the other; display of only one or the other.

420 Bowlby, J. (1960) Separation anxiety. International Journal of Psycho-Analysis, 41, 89–113, p.97.

421 Hinde, R.A. (1961) The establishment of the parent–offspring relation in birds, with some mammalian analogies. In W.H. Thorpe & O.L. Zangwill (eds) Current Problems in Animal Behaviour (pp.175–93). Cambridge: Cambridge University Press, p.185. Passage underlined as of particular note in Bowlby’s personal copy of the text. PP/Bow/H.226.

422 Hinde, R. (1970) Animal Behavior, 2nd edn. New York: McGraw-Hill.

423 Bowlby, J. (1965) Motivation. PP/Bow/H.128.

424 Bowlby, J. & Soddy, K. (1940) War neurosis memorandum. PP/Bow/C.5/1; Bowlby, J. (1942) Selection and diagnosis in Army: notes for a talk. PP/BOW/C.5/2/3.

425 Freud, A. (1956, 1969) The assessment of borderline cases. In The Writings of Anna Freud, Vol. 5 (pp.301–14). New York: International Universities Press, p.310.

426 Ainsworth, M.D.S. (1977) Attachment theory and its utility in cross-cultural research. In: P.H. Leiderman, S.R. Tulkin, & A. Rosenfeld (eds) Culture and Infancy: Variations in the Human Experience (pp. 49–67). New York: Academic Press, p.59.

427 Bretherton, I. & Ainsworth, M.D.S. (1974) Responses of one-year-olds to a stranger in a strange situation. In M. Lewis & L.A. Rosenblum (eds) The Origin of Fear (pp. 131–64). New York: Wiley.

428 This is discussed in the section ‘Incompatible behavioural systems: results of simultaneous activation’ of Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin, pp.97–101.

429 Bowlby, J. (1986) Transcript of interview by Lange. Wellcome Trust Library Archive. PP/BOW/A.5/16: ‘Anxiously avoidant children who have a parent who tends towards rejection … They are really caught in a classic approach–avoidance conflict. On the one hand they really would love to have affection and support and comfort. They would love to have that sort of relationship. But their attempts to develop such a relationship were met with so many painful rejections in the past that they dare not attempt it again. And they pretend, sometimes convince themselves and convince everyone else, that they get along very well, thank you, without any of that sentimental nonsense … They are incredibly out of touch with their feelings and out of touch with the situations which evoke their feelings.’

430 Bowlby, J. (1980) Loss. London: Pimlico, p.346.

431 Bowlby, J. (1962) Defences that Follow Loss: Causation and Function. PP/Bow/D.3/78.

432 Annotations by Bowlby on Abraham, K. (1927) Selected Papers. London: Hogarth Press, p.263; not dated but handwriting and reference on p.74 suggest annotations are from 1933. Copy held in the library of Human Development Scotland.

433 Bowlby cited in Tanner, J.M. & Inhelder, B. (1956) Discussions on Child Development: Proceedings of the WTO Study Group of the Psychobiological Development of the Child, Vol. 1. London: Tavistock, p.186.

434 Bowlby, J. (1974) Psychological processes evoked by a major psychosocial transition. Presented to Tavistock Research Workshop, March 1974. PP/Bow/F.3/90; Bowlby, J. (1989) Foreword to Emmy Gut’s Productive and Unproductive Depression (pp.xiii–xv), London: Routledge.

435 Bowlby, J. (1977, 1979) The making and breaking of affectional bonds. In The Making and Breaking of Affectional Bonds (pp.150–88). London: Routledge: ‘The stronger the emotions aroused in a relationship the more likely are the earlier and less conscious models to become dominant’ (168).

436 Bowlby, J. (1985) Letter to John Byng-Hall, 12 April 1985. PP/Bow/J.9/45: ‘You need to clarify what the risks of intimacy are. One risk, which may motivate either child or mother, is fear of rejection; another risk, which may also motivate either partner, is fear of being held captive by the intense attachment behaviour of the other.’

437 This point is elaborated with case studies in Bowlby, J. (1980) The place of defensive exclusion in depressive disorders. PP/Bow/K.7/94; see also Bowlby, J. (1987) Notes on depression, towards correspondence with Emmy Gut. PP/Bow/B.3/15.

438 Bowlby, J. (1980) Loss. London: Pimlico, p.249–50. For a recent discussion see Bosmans, G. (2016) Cognitive behaviour therapy for children and adolescents: can attachment theory contribute to its efficacy? Clinical Child and Family Psychology Review, 19(4), 310–28.

439 Bowlby, J. (1981) Letter to Aaron Beck, 8 October 1981. PP/Bow/J.9/16. See Beck, A. (1983) Cognitive therapy of depression: new perspectives. In P.J. Clayton & J.E. Barrett (eds) Treatment of Depression: Old Controversies and New Approaches (pp.265–90). New York: Raven.

440 Bowlby, J. (1990, 2011) John Bowlby: interview by Leonardo Tondo. Clinical Neuropsychiatry, 8(2), 159–71, p.167.

441 Ainsworth, M. & Bowlby, J. (1991) 1989 APA award recipient addresses: an ethological approach to personality development. American Psychologist, 46(4), 333–41, p.336. The Bowlby Archive contains a wonderful description of the clinic, entitled ‘Mothers Discussion Group’ from 1967 (PP/Bow/C.6/3). In justifying the benefits of the group Bowlby wrote: ‘It is from other mothers that a beginner can learn most. Professional people can add information about different sorts of behaviour and the range of ages at which different developments are likely to occur—but how to cope best is learned from others who are actually confronted with the job.’

442 Bowlby, J. (1981) Psychoanalysis as a natural science. International Review of Psycho-Analysis, 8, 243–56, p.253.

443 Bowlby, J. (1957–59) Untitled notes responding to the work of John Alford. PP/Bow/B.3/19.

444 Bowlby, J. (1951) Maternal Care and Mental Health. Geneva: World Health Organisation, p.61.

445 E.g. Belsky, J. & Nezworski, T. (eds) (1988) Clinical Implications of Attachment. Hillsdale, NJ: Lawrence Erlbaum; West, M., Sheldon, A., & Reiffer, L. (1989) Attachment theory and brief psychotherapy: applying current research to clinical interventions. Canadian Journal of Psychiatry, 34, 369–75; Byng-Hall, J. (1990) Attachment theory and family therapy: a clinical view. Infant Mental Health Journal, 11, 228–37.

446 An important recent contribution to and review of this literature is Slade, A. & Holmes, J. (2017) Attachment in Therapeutic Practice. London: SAGE. For a thematic review, calling for a shift in emphasis from theory- to research-based guidance for attachment-inspired therapies see Berry, K. & Danquah, A. (2016) Attachment-informed therapy for adults: towards a unifying perspective on practice. Psychology and Psychotherapy: Theory, Research and Practice, 89(1), 15–32. For descriptions of the boom in therapeutic approaches claiming genesis in attachment theory as an ‘explosion’ see Magnavita, J.J. & Anchin, J.C. (2013) Unifying Psychotherapy: Principles, Methods, and Evidence from Clinical Science. New York: Springer, p.67; Johnson, S.M. (2019) Attachment Theory in Practice. New York: Guilford, p.22.

447 E.g. McCluskey, U. (2005) To Be Met as a Person: The Dynamics of Attachment in Professional Encounters. London: Karnac Books; Lyons-Ruth, K. (2007) The interface between attachment and intersubjectivity: perspective from the longitudinal study of disorganized attachment. Psychoanalytic Inquiry, 26(4), 595–616; Crittenden, P.M., Dallos, R., Landini, A., & Kozlowska, K. (2014) Attachment and Systemic Family Therapy. London: McGraw-Hill.

448 For instance, Bowlby’s request of Emmy Gut to use anonymized material from her case, discussed and appraised in Ross, L.R. (2006) Talking theory, talking therapy: Emmy Gut and John Bowlby. Issues in Mental Health Nursing, 27(5), 475–97.

449 There is one early case which is, however, available: Bowlby’s clinical notes of a case from 1938 to 1939 (PP/Bow/C.4/23). The reason why these are available appears to be the lack of identifying details in these notes. There are also a series of undated short case histories under the title ‘Maternal Behaviour: Humans’ (PP/Bow/H.136) in the Bowlby Archive. However, these contain few details about Bowlby’s therapeutic practice, and remained unpublished. The case histories were likely written in the early 1960s, and describe cases seen by Bowlby at the Child Guidance Clinic.

450 Bowlby, J. (1981) Clinical applications: material for lectures. PP/Bow/F.3/103.

451 The first discussion of ‘Mrs Q’ is in a 1962 manuscript. The first published mention is 1963, discussing an analysis that has lasted three years. This suggests that the patient is the ‘Mrs K’ with whom Bowlby conducted detailed clinical interviews between 1960 and 1964, keeping these interviews for reference. These are the only clinical notes Bowlby kept in this way. The ‘reports of interviews with Mrs K (mother) (Mother and Child Welfare Clinic)’ (PP/BOW/C.6/6-8) are embargoed until the 2060s.

452 Bowlby, J. (1979) By ethology out of psychoanalysis: an experiment in interbreeding. (The Niko Tinbergen Lecture.) Animal Behaviour, 28(3), 649–56, p.653.

453 Bowlby, J. (1969) Psychopathology of anxiety: the role of affectional bonds. British Journal of Psychiatry, 3, 80–86, p.85.

454 Bowlby, J. (1984) Violence in the family as a disorder of the attachment and caregiving systems. American Journal of Psychoanalysis, 44, 9–27, p.17.

455 Bowlby, J. (1979) By ethology out of psychoanalysis: an experiment in interbreeding. (The Niko Tinbergen Lecture.) Animal Behaviour, 28(3), 649–56, p.653–4.

457 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.269.

458 Ibid. p.270.

459 Bowlby, J. (1963) Pathological mourning and childhood mourning. Journal of the American Psychoanalytic Association, 11(3), 500–41, p.517.

460 Ibid. p.517.

461 Bowlby, J. (1980) Loss. London: Pimlico, p.151.

462 Ibid. pp.348–9.

463 Bowlby, J. (1962) Defences that Follow Loss: Causation and Function. PP/Bow/D.3/78.

464 Whether provision of a relationship with attachment components should be the goal of shorter-term work by helping professionals has been debated, e.g. Charles, G. & Alexander, C. (2014) Beyond attachment: mattering and the development of meaningful moments. Relational Child and Youth Care Practice, 27(3), 26–30.

465 Bowlby’s copy of The Standard Edition of the Writings of Sigmund Freud still retains his bookmark, in Volume 11, marking the following underlined passage: the patient’s ‘symptoms, to take an analogy from chemistry, are precipitates of earlier experiences in the sphere of love (in the widest sense of the word), and it is only in the raised temperature of his experience of the transference that they can be resolved and reduced to other psychical products.’ Freud, S. (1910) Five lectures on psycho-analysis. In The Standard Edition of the Complete Psychological Works of Sigmund Freud, Vol. 11 (pp.1–56). London: Hogarth, p.51; Bowlby’s copy is held in the library of Human Development Scotland.

466 Bowlby, J. (1979, 1988) On knowing what you are not supposed to know and feeling what you are not supposed to feel. In A Secure Base. London: Routledge, pp.111–33, p.132.

467 Bowlby, J. (1949) The study and reduction of group tensions within the family. Human Relations 2, 123–8, p.123; Bowlby, J. (1969, 1982) Attachment, Volume 1. London: Penguin: ‘Nothing in child psychiatry has been more significant in recent years than the increasing recognition that the problems its practitioners are called upon to treat are not often problems confined to individuals but are usually problems arising from stable interactional patterns that have developed between two, and more often several, members of a family’ (349). Potential links between attachment theory and systemic family therapy have been widely discussed. See e.g. Cowan, P.A. (1997) Beyond meta-analysis: a plea for a family systems view of attachment. Child Development, 68(4), 601–603; Akister, J. & Reibstein, J. (2004) Links between attachment theory and systemic practice: some proposals. Journal of Family Therapy, 26(1), 2–16; Crittenden, P.M., Dallos, R., Landini, A., & Kozlowska, K. (2014) Attachment and Systemic Family Therapy. London: McGraw-Hill; Vetere, A. (2016) Systemic theory and narratives of attachment: integration, formulation and development over time. In M. Borcsa & P. Stratton (eds) Origins and Originality in Family Therapy and Systemic Practice (pp.129–39). New York: Springer.

468 Bowlby, J. (1986) ‘Attachment Theory: New Directions’. ACP-Psychiatric UPDATE 7(2), panel discussion, Washington. PP/BOW/A.5/1. On Bowlby’s position in relation to the history of family therapy and systems approaches to family relations see Byng-Hall, J. (1991) An appreciation of John Bowlby: his significance for family therapy. Journal of Family Therapy, 13(1), 5–16. For a discussion of societal factors contributing to ‘the family’ as the unit of concern and intervention see Weinstein, D. (2013) The Pathological Family: Postwar America and the Rise of Family Therapy. Ithaca, NY: Cornell University Press.

469 Bowlby, J. (1977, 1979) The making and breaking of affectional bonds. In The Making and Breaking of Affectional Bonds (pp.150–88). London: Routledge, p.170.

470 Bowlby, J. (1990) Letter to Joan Stevenson-Hinde, 26 March 1990. PP/Bow/J.9/186.

471 Bowlby, J. (1990, 2011) John Bowlby: interview by Leonardo Tondo. Clinical Neuropsychiatry, 8(2), 159–71, p.169. Bowlby was reacting to the emphasis in the Kleinian tradition on transference interpretations. He gave particular emphasis to the importance of support, and helping the patient or family get to a point where they can make sense and use of interpretations—much like the concept of ‘developmental help’ within the Anna Freudian tradition. See Edgcumbe, R. (1995) The history of Anna Freud’s thinking on developmental disturbances. Bulletin of the Anna Freud Centre, 18(1), 21–34. Bowlby’s emphasis, however, led some second-generation attachment researchers to assume that he was arguing only for support, without challenge, in therapy. See e.g. Bretherton, I. (1991) Pouring new wine into old bottles: the social self as internal working model. In M.R. Gunnar and L.A. Sroufe (eds) Self-Processes and Development: The Minnesota Symposia on Child Development, Vol. 23 (pp.1–41). Hillsdale, NJ: Lawrence Erlbaum, p.32.

472 E.g. Bowlby, J. & Parkes, C.M. (1970, 1979) Separation and loss within the family. In The Making and Breaking of Affectional Bonds (pp.99–123). London: Routledge, p.114.

473 Annotations by Bowlby dated 1964 on The Standard Edition of the Complete Psychological Works of Sigmund Freud, Vol. 23, p.261. Copy held in the library of Human Development Scotland.

474 Bowlby, J. (1973) Separation: Anger and Anxiety. New York: Basic Books, p.354.

475 Bowlby, J. (1981) Psychoanalysis as a natural science. International Review of Psycho-Analysis, 8, 243–56, p.251.

476 See the three unpublished case histories in ‘Maternal behaviour: humans’ (PP/Bow/H.136), where this is the common central feature of his therapeutic approach. See also Bowlby, J. (1971) Letter to Graham Davies, 15 April 1971. PP/Bow/H.5: ‘I see psychoanalytic therapy as an attempt to help a patient explore his own motives, his own model of the world and himself in it, and also to reconsider the validity of that model. Often I believe it is more valuable to raise questions for a patient than to attempt to inform him by means of interpretations … In helping a patient make these reappraisals, I believe it useful for us to have some knowledge of the conditions that commonly lead an individual to grow up in ignorance of his motives and with more or less distorted models of his world and himself.’

477 Bowlby, J. (1977, 1979) The making and breaking of affectional bonds. In The Making and Breaking of Affectional Bonds (pp.150–88). London: Routledge, p.173.

478 Bowlby, J. (1990, 2011) John Bowlby: interview by Leonardo Tondo. Clinical Neuropsychiatry, 8(2), 159–71, p.163.

479 Fonagy has regularly criticized strands of psychoanalytic theory for treating the discovery of historical truth as the mechanism and target of therapeutic action, e.g. Fonagy, P. & Tallandini-Shallice, M. (1993) On some problems of psychoanalytic research in practice. Bulletin of the Anna Freud Centre, 16, 5–22. Bowlby’s emphasis on the contribution of defensive exclusion to mental illness, and the curative powers of the reintegration of reality within information processing, do at times veer in this direction. However, suggesting that this may be a structural issue faced by psychotherapy as an activity, in his more recent work on therapeutic technique Fonagy has himself at times become tangled in the same problem, e.g. Fonagy, P. (2016) The role of attachment, epistemic trust and resilience in personality disorder: a trans-theoretical reformulation. DMM News, 22 September 2016. ‘Feeling recognized opens the epistemic path necessary to update the neural nets which in turn enable accurate (resilient) interpretation of reality’ (6).

480 Bowlby, J. (1954) The diagnosis and treatment of psychological disorders in childhood. Health Education Journal, 12(2), 59–68, p.62.

481 Bowlby, J. (undated, c.1939) Diagnostic Groupings for Children Over 5 Years. PP/Bow/C3/9.

482 Bowlby, J. (1940) Personality and Mental Illness. London: Kegan Paul, p.187.

483 Bowlby, J. (1988) Developmental psychiatry comes of age. American Journal of Psychiatry, 145, 1–10, p.6.

484 Bowlby, J. (1980) Loss. London: Pimlico, p.202–203.

485 Bowlby, J. (1988) Developmental psychiatry comes of age. American Journal of Psychiatry, 145, 1–10, p.6.

486 For a later statement well aligned with this position see Waters, E., Posada, G., Crowell, J., & Lay, K.L. (1993) Is attachment theory ready to contribute to our understanding of disruptive behavior problems? Development & Psychopathology, 5, 215–24.

487 See also Fonagy, P. & Allison, E. (2014) The role of mentalizing and epistemic trust in the therapeutic relationship. Psychotherapy, 51(3), 372–80.

488 Waddington, C.H. (1957) The Strategy of Genes. London: Allen & Unwin.

489 This term would later become the label for an influential self-report measure of adversity and trauma. Felitti, V.J., Anda, R.F., Nordenberg, D., et al. (1998) Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: the Adverse Childhood Experiences (ACE) Study. American Journal of Preventive Medicine, 14(4), 245–58. If there is a debt to Bowlby’s earlier use, it is not acknowledged by Felitti and colleagues. This is discussed by Partridge, S. (2019) Review of ‘The Deepest Well: Healing the Long-term Effects of Childhood Adversity’ by Nadine Burke Harris. Attachment, 13(1).

490 Bowlby, J. (1981, 1988) The origins of attachment theory. In A Secure Base (pp.22–42). London: Routledge.

491 Bowlby, J. (1973) Separation. London: Pimlico, p.412. The importance of the concept of developmental pathways for Bowlby’s late thought cannot be overestimated. See for instance the prominent place it receives in his plan for the unpublished book, representing his final position (on the model of Freud’s ‘Outline of Psychoanalysis’), Bowlby, J. (1982) Outline of attachment theory. PP/Bow/H.260:

  • Chapter 1. Historical. A way of conceptualising family influence.

  • Chapter 2. Main features: developmental pathways; description of attachment, caregiving, exploration; abuse

  • Chapter 3. Cognitive models

  • Part II: Pathways of development

  • Chapter 4. Pathways to health, including healthy parenting

  • Chapter 5. Pathways to anxious attachment & phobia

  • Chapter 6. Pathways to depressive disorders. Suicide?

  • Chapter 7. Pathways to false self, masked depression

  • Chapter 8. Pathways to delinquency & psychosis

  • Chapter 9.

  • Part III: Applications

  • Chapter 10. Prevention and crisis intervention

  • Chapter 11. Treatment

  • Part IV: Problems of theory

  • Chapter 12. Evolutionary theory, control systems, activation, termination function

  • Chapter 13. More re cognitive models and defence. Consciousness and unconsciousness

  • Chapter 14. Attachment and science’

492 Johnstone, L. & Dallos, R. (2013) Formulation in Psychology and Psychotherapy: Making Sense of People’s Problems. London: Routledge.

493 Bakwin, H. (1949) Emotional deprivation in infants. Journal of Pediatrics 35, 512‒21; Provence, S. & Lipton, R. (1962) Infants in Institutions. New York: International Universities Press; Tizard, R. & Rees, J. (1975) The effect of early institutional rearing on the behaviour problems and affectional relationships of four year old children. Journal of Child Psychology and Psychiatry, 16, 61–73.

494 There has been a growing body of research on attachment disorders since 1998, and especially since the increased prominence and detail of the diagnosis in DSM-IV in 1994 and DSM-5 in 2013. These developments in diagnostic practice have been influenced by the pivotal series of studies of children adopted from Eastern European orphanages. See Zeanah, C.H., Smyke, A.T., & Settles, L.D. (2006) Orphanages as a developmental context for early childhood. In K. McCartney & D. Phillips (eds) Blackwell Handbook of Early Childhood Development (pp.424–54). Hoboken, NJ: Wiley-Blackwell; Zeanah, C.H., Chesher, T., Boris, N.W., & American Academy of Child & and Adolescent Psychiatry (2016) Practice parameter for the assessment and treatment of children and adolescents with reactive attachment disorder and disinhibited social engagement disorder. Journal of the American Academy of Child & Adolescent Psychiatry, 55, 990–1003.

495 E.g. Kanieski, M.A. (2010) Securing attachment: the shifting medicalisation of attachment and attachment disorders. Health, Risk & Society, 12(4), 335–44.

496 Call, J. (1982) Attachment disorders of infancy. In H.I. Kaplan, A.M. Freedman, & Z.B.J. Sadock (eds) Comprehensive Textbook of Psychiatry, Vol. 3 (pp.230–68). Baltimore: Williams and Wilkins, p.2587. Repeated in Call, J.D. (1984) Child abuse and neglect in infancy: sources of hostility within the parent–infant dyad and disorders of attachment in infancy. Child Abuse & Neglect, 8(2), 185–202, p.190.

497 Additionally, the Robertson and Bowlby book discussing observations in the 1950s of anomalous and apparently dissociated behaviours of institutionalized children remained unpublished, so Call and colleagues on the ‘Infancy, Childhood and Adolescent Disorders’ committee did not have access to this subterranean current of attachment theory, which would find full recognition only in the 1980s in the work of Mary Main and Judith Solomon on the ‘disorganized/disoriented’ attachment classification (Chapter 3).

498 Mayes, R. & Horwitz, A.V. (2005) DSM-III and the revolution in the classification of mental illness. Journal of the History of the Behavioral Sciences, 41(3), 249–67.

499 Richters, M. & Volkmar, F. (1994) Reactive attachment disorder of infancy or early childhood. Journal of the American Academy of Child and Adolescent Psychiatry, 33, 328–32.

500 Call, J. (1982) Attachment disorders of infancy. In H.I. Kaplan, A.M. Freedman, & Z.B.J. Sadock (eds) Comprehensive Textbook of Psychiatry, Vol. 3 (pp.230–68). Baltimore: Williams and Wilkins.

501 Green, W. (1985) Attachment disorders of infancy and early childhood. In H.I. Kaplan & Z.B.J. Sadock (eds) Comprehensive Textbook of Psychiatry, Vol. 4 (pp.1722–31). Baltimore: Williams and Wilkins.

502 Greco, M. (2016) What is the DSM? Diagnostic manual, cultural icon, political battleground: an overview with suggestions for a critical research agenda. Psychology & Sexuality, 7(1), 6–22.

503 Bowlby, J. (1983) Darwin: Psychiatry and Developmental Psychology. Contribution to a Symposium on Darwin and Psychology held at the conference of the British Psychological Society, London, December 1983. PP/BOW/F.3/132.

504 Sroufe, L.A. (1997) Psychopathology as an outcome of development. Development & Psychopathology, 9(2), 251–68, p.263. Bowlby’s position would perhaps align with developments in the UK, where clinical guidelines now advise professionals to consider a broader range of problems described as ‘attachment difficulties’ for children who are adopted or are at risk of going into care, outside of the constraints of the ‘attachment disorder’ diagnoses. National Institute for Health & Care Excellence (2015) Children’s attachment: attachment in children and young people who are adopted from care, in care or at high risk of going into care. NICE Guideline (NG26).

505 E.g. Zeanah, C.H., Mammen, O., & Lieberman, A. (1993) Disorders of attachment. In C.H. Zeanah (ed.) Handbook of Infant Mental Health (pp.332–49). New York, Guilford.

506 Zero to Three/National Center for Clinical Infant Programs (1994) Diagnostic Classification of Mental Health and Developmental Disorders of Infancy and Early Childhood (Diagnostic Classification: 0-3). Washington, DC: Zero to Three; Lyons-Ruth, K., Zeanah, C.H., & Benoit, D. (1996) Disorder and risk for disorder during infancy and toddlerhood. In E.J. Mash & R.A. Barkley (eds) Child Psychopathology (pp.457–91). New York: Guilford.

507 Atkinson, L. (2019) Reactive attachment disorder and attachment theory from infancy to adolescence: review, integration, and expansion. Attachment & Human Development, 21, 205–17.

508 Lyons-Ruth, K., Zeanah, C.H., & Gleason, M.M. (2015) Commentary: should we move away from an attachment framework for understanding disinhibited social engagement disorder (DSED)? A commentary on Zeanah and Gleason (2015) Journal of Child Psychology and Psychiatry, 56(3), 223–7.

509 Minnis, H., Marwick, H., Arthur, J., & McLaughlin, A. (2006) Reactive attachment disorder—a theoretical model beyond attachment. European Child & Adolescent Psychiatry, 15(6), 336–42.

510 E.g. Zeanah, C.H., Smyke, A.T., Koga, S.F., Carlson, E., & Bucharest Early Intervention Project Core Group (2005) Attachment in institutionalized and community children in Romania. Child Development, 76(5), 1015–28; Lyons-Ruth, K., Zeanah, C.H., & Gleason, M.M. (2015) Commentary: should we move away from an attachment framework for understanding disinhibited social engagement disorder (DSED)? A commentary on Zeanah and Gleason (2015) Journal of Child Psychology and Psychiatry, 56(3), 223–7; Lyons-Ruth, K., Riley, C., Patrick, M.P., & Hobson, R.P. (2019) Disinhibited attachment behavior among infants of mothers with borderline personality disorder, depression, and no diagnosis. Personality Disorders, 10(2), 163–72; Spangler, G., Bovenschen, I., Jorjadze, N., et al. (2019) Inhibited symptoms of attachment disorder in children from institutional and foster care samples. Attachment & Human Development, 21(2), 132–51. Spangler, like van IJzendoorn and Cassidy, is perhaps best considered an intermediate case for the heuristic contrast between second- and third-generation attachment researchers.

511 E.g. Sroufe, L.A., Duggal, S., Weinfield, N., & Carlson, E. (2000) Relationships, development, and psychopathology. In A.J. Sameroff, M. Lewis, & S.M. Miller (eds) Handbook of Developmental Psychopathology (pp. 75–91). New York: Kluwer Academic/Plenum Press, p.83; van IJzendoorn, M.H. & Bakermans-Kranenburg, M.J. (2003) Attachment disorders and disorganized attachment: similar and different. Attachment & Human Development, 5(3), 313–20. See also Allen, B. (2016) A RADical idea: a call to eliminate attachment disorder and attachment therapy from the clinical lexicon. Evidence-based Practice in Child and Adolescent Mental Health, 1, 60–71; Lyons-Ruth, K., Bureau, J.F., Riley, C.D., & Atlas-Corbett, A.F. (2009) Socially indiscriminate attachment behavior in the Strange Situation: convergent and discriminant validity in relation to caregiving risk, later behavior problems, and attachment insecurity. Development & Psychopathology, 21(2), 355–72; Granqvist, P., Sroufe, L.A., Dozier, M., et al. (2017) Disorganized attachment in infancy: a review of the phenomenon and its implications for clinicians and policy-makers. Attachment & Human Development, 19(6), 534–58.

512 In this, researchers commenting on the attachment disorder diagnosis rather neglect the question of whether, by late childhood and adolescence, expectations about the availability of attachment figures become a partial property of individuals in much the same way as mental health symptoms.

513 Though see Scheper, F.Y., Groot, C.R., de Vries, A.L., Doreleijers, T.A., Jansen, L.M., & Schuengel, C. (2019) Course of disinhibited social engagement behavior in clinically referred home-reared preschool children. Journal of Child Psychology and Psychiatry, 60(5), 555–65.

514 Chaffin, M., Hanson, R., Saunders, B.E., et al. (2006) Report of the APSAC task force on attachment therapy, reactive attachment disorder, and attachment problems. Child Maltreatment, 11(1), 76–89. Allen and Mercer have criticized attachment disorder and interventions to resolve it as, at present, more pseudoscience than science: Allen, B. (2018) Misperceptions of reactive attachment disorders persist: poor methods and unsupported conclusions. Research in Developmental Disabilities, 77, 24–9; Mercer, J. (2019) Conventional and unconventional perspectives on attachment and attachment problems: comparisons and implications, 2006–2016. Child and Adolescent Social Work Journal, 36(2), 81–95. Nonetheless, perhaps it can be said that ‘attachment disorder’ has been integrated into attachment theory as a characterization of children who have had little chance to form attachments due to neglect or institutionalization.

515 Barnes, G.L., Woolgar, M., Beckwith, H., & Duschinsky, R. (2018) John Bowlby and contemporary issues of clinical diagnosis. Attachment, 12(1), 35–47.

516 E.g. Fonagy, P., Target, M., Steele, M., et al. (1997) Morality, disruptive behavior, borderline personality disorder, crime, and their relationships to security of attachment. In L. Atkinson & K.J. Zucker (eds) Attachment and Psychopathology (pp. 223–74). New York: Guilford, p.224.

517 Turner, M., Beckwith, H., Duschinsky, R., et al. (2019) Attachment difficulties and disorders. InnovAiT, 12(4), 173–9.

518 Woolgar, M. & Baldock, E. (2015) Attachment disorders versus more common problems in looked after and adopted children: comparing community and expert assessments. Child and Adolescent Mental Health, 20(1), 34–40; White, S., Gibson, M., Wastell, D., & Walsh, P. (2019) Reassessing Attachment Theory in Child Welfare. Bristol: Psychology Press.

519 Reijman, S., Foster, S., & Duschinsky, R. (2018) The infant disorganised attachment classification: ‘Patterning within the disturbance of coherence’. Social Science & Medicine, 200, 52–8.

520 Bowlby, J. (1973) Letter to Scott Henderson, 30 July 1973. PP/Bow/J.9/98. See also Bowlby’s undated notes on the concept of identification, PP/BOW/H.117-121. For instance, the note ‘Identification—Klein’, in which Bowlby writes ‘Mrs K. assumes that the copying follows and is a manifestation of introjection. It seems more likely that the identification is in fact the resultant of copying’ PP/Bow/H.118 (parentheses suppressed). For Bandura and the concept of modelling see Bandura A. & Menlove F.L. (1968) Factors determining vicarious extinction of avoidance behavior through symbolic modeling. Journal of Personality & Social Psychology, 8, 99–108.

521 Bowlby, J. (1973) Separation: Anger and Anxiety. New York: Basic Books, p.417.

523 Ibid. p.235; see also Bowlby, J. (1984) No such thing as a baby, March 1984. PP/Bow/F.3/136: ‘How to explain persistence: a) Parents commonly continue in the way they started; b) virtuous or vicious circles pattern perpetuates itself.’

524 Sroufe, L.A., Duggal, S., Weinfield, N.S., & Carlson, E. (2000) Relationships, development, and psychopathology. In M. Lewis & A. Sameroff (eds) Handbook of Developmental Psychopathology, 2nd edn (pp.75–92). New York: Kluwer Academic/Plenum Press, p.87.

525 Gilmore, K. (1990) A secure base. Parent–child attachment and healthy human development: by John Bowlby. Psychoanalytic Quarterly, 59, 494–8; Crittenden, P.M. (1997) Truth, error, omission, distortion, and deception: the application of attachment theory to the assessment and treatment of psychological disorder. In S.M. Clancy Dollinger & L.F. DiLalla (eds) Assessment and Intervention Issues Across the Lifespan (pp.35–76). London: Erlbaum; Bretherton, I. (2005) In pursuit of the internal working model construct and its relevance to attachment relationships. In K.E. Grossmann, K. Grossmann, & E. Waters (eds) Attachment from Infancy to Adulthood: The Major Longitudinal Studies (pp.13–47). New York: Guilford.

526 This concern has also been discussed by Weinfield, N.S., Whaley, G. & Egeland, B. (2004) Continuity, discontinuity, and coherence in attachment from infancy to late adolescence: sequelae of organization and disorganization. Attachment & Human Development, 6(1), 73–97; Fraley, R.C. & Shaver, P.R. (2008) Attachment theory and its place in contemporary personality research. In O. John & R.W. Robins (eds) Handbook of Personality: Theory and Research, 3rd edn (pp. 518–41). New York: Guilford. The problem stems in part from the fact that much of Bowlby’s early thinking in the 1940s had taken ‘personality’ as its object; and this thinking was then transferred to ‘attachment’, without full elaboration of the relationship between the two constructs. On the concept of personality in Bowlby’s early work see Bowlby, J. (1940) Personality and Mental Illness. London: Kegan Paul.

527 Ainsworth, M.D.S. (1988) Security. Unpublished discussion paper prepared for the Foundations of Attachment Theory Workshop, convened for the New York Attachment Consortium by G. Cox-Steiner & E. Waters, Port Jefferson, NY.

528 Bowlby, J. (1984) Violence in the family as a disorder of the attachment and caregiving systems. American Journal of Psychoanalysis, 44, 9–27, p.21. An early critic of Bowlby on this point was Rutter, M. (1972) Maternal Deprivation Reassessed. London: Penguin.

529 Bowlby, J. (1944) Forty-four juvenile thieves: their characters and home-life (II). International Journal of Psycho-Analysis, 25, 107–28.

530 Bowlby, J. (1984) Violence in the family as a disorder of the attachment and caregiving systems. American Journal of Psychoanalysis, 44, 9–27, p.21.

531 Ibid. p.11.

532 Johnson, S.M. (2019) Attachment Theory in Practice. New York: Guilford, Chapter 2. See for instance the anachronistic attribution of a focus on ‘felt sense’ of emotions such as security to Bowlby, rather than Sroufe and Waters, on p.34.

533 Bell, D.C. & Richard, A.J. (2000) Caregiving: the forgotten element in attachment. Psychological Inquiry, 11(2), 69–83, p.73. It is interesting to watch Bowlby ponder the relationship between motivation and emotion in his notes on ‘Emotion and feeling’, PP/Bow/H.5. These include reflections from the 1970s on emotion theorists such as Tomkins and Izard. His most determinate statement on the relationship between motivation and emotion in these notes seems to be the claim that the two are distinct, and that feelings contribute to the rise of motivations: ‘The feeling of being oriented & drawn toward some end gives rise to the sensation of desiring or wanting to achieve that end.’ However, he criticized Izard for insufficient attention to ‘causal factors’, seemingly besides feelings, ‘that activate an actual sequence of behaviour’.

534 One of Bowlby’s most sustained considerations of mood is actually in his notes on history-taking from the mid-1930s: see ‘History taking; methods of examining’, PP/BOW/D.2/13. However, there he was interested in the role of moods as indices of personality, rather than in their cause or effects.

535 By contrast, shame has been a central theme in clinical elaborations of Bowlby’s ideas by Allan and Judy Schore and by Dan Hughes and colleagues: Schore, A.N. (1991) Early superego development: the emergence of shame and narcissistic affect regulation in the practicing period. Psychoanalysis and Contemporary Thought, 14(2), 187–250; Hughes, D., Golding, K.S., & Hudson, J. (2015) Dyadic developmental psychotherapy (DDP): the development of the theory, practice and research base. Adoption & Fostering, 39(4), 356–65.

536 Main, M. (1977) Sicherheit und wissen. In K.E. Grossmann (ed.) Entwicklung der Lernfahigheit in der sozialen Umwelt (pp.47–95). Munich: Kindler.

537 Bowlby, J. (undated) Unpublished notes from his filing cabinets. c. 1955. PP/Bow/H115.

538 Bowlby, J. (1986) An interview with John Bowlby on the origins and reception of his work. Free Associations, 1, 36–64, p.42.

539 Bowlby, J. & Parkes, C.M. (1970, 1979) Separation and loss within the family. In The Making and Breaking of Affectional Bonds (pp.99–123). London: Routledge, p.123.

540 This is described amusingly in Issroff, J. (2005) Donald Winnicott and John Bowlby: Personal and Professional Perspectives. London: Karnac: ‘Bowlby … “parked” the term “fantasy” in what I always imagine to be a quite capacious parking lot in his mind’ (56).

541 Ainsworth, M. (1971) Comments on the manuscript of Attachment Volume 2. Letter to Bowlby, 16 August 1971. PP/BOW/K.5/62: ‘You yourself, naturally, push your own view. And yet I really am convinced (and I am sure that you basically agree) that in many cases both kinds of “dynamics” are at work … It would take only a few alterations of turns of phrase to leave the classical Freudians and Kleinians a little more opportunity to feel that you are asking them to extend their view rather than to abandon it.’

542 Ainsworth, M. (1997) Peter L. Rudnytsky—the personal origins of attachment theory: an interview with Mary Salter Ainsworth. Psychoanalytic Study of the Child, 52, 386–405, p.398.

543 Waters, E., Corcoran, D., & Anafarta, M. (2005) Attachment, other relationships, and the theory that all good things go together. Human Development, 48(1–2), 80–84: ‘Bowlby’s attachment theory remains a work in progress … his notion of attachment working models was limited to what could be done with the cognitive psychology of the day’ (82).

544 Bowlby, J. (1988) Developmental psychiatry comes of age. American Journal of Psychiatry, 145, 1–10, p.1.

545 Bowlby, J. (1990, 2011) John Bowlby: interview by Leonardo Tondo. Clinical Neuropsychiatry, 8(2), 159–71: ‘The main thing about the monkey work has been that, with fairly rigorous experimental designs and methods, they have demonstrated the ill effects of separation and its obvious consequences (Hinde 1966). It is a huge literary reserve which I was fairly familiar with in the 1960s because it was very dramatic but I haven’t kept up’ (166).

546 Bowlby, J. (1958) Nature of the child’s tie to his mother. International Journal of Psycho-Analysis, 39, 350–73, p.356, citing Winnicott, D.W. (1948) Pediatrics and psychiatry. British Journal of Medical Psychology 21, 229–40.

547 Ainsworth, M. (1965) Letter to John Bowlby, 7 October 1965. Mary Ainsworth papers, Box M3168, Folder 2: ‘I was very pleased with your chapter on Instinct Theory … There was one point at which I was jerked out of my “yes” attitude. On page 60 you state that the main function of attachment behaviour is to ensure safety of the young—safety from predators. Of course, you will develop this point fully in your revision of the “Child’s Tie” paper. I did not say “no” to your statement, but I did think about functions of maternal behaviour other than protection from predators—specifically nurturance of the young, but also training.’ Several of Ainsworth’s students have followed her lead on this matter rather than Bowlby’s position. See e.g. Waters, E. (2008) Live long and prosper: a note on attachment and evolution.

548 Bowlby, J. (1988) A Secure Base. London: Routledge p.121. Bowlby, J. (1985) Letter to Malcom West, 13 November 1985. PP/Bow/J.9/201: ‘I suspect that, in the specific situations in which A’s attachment behaviour is activated, A always regards the partner B as definitely the more competent to deal with them. And also that, in other situations, the roles of the partners are reversed. Thus, whilst overall each may regard the other as no more competence, etc. than the self, whenever attachment behaviour is activated the roles become complementary rather than reciprocal.’

549 Main, M. (1970) Infant play and maternal sensitivity in primate evolution. PP/Bow/J.4/1: ‘Alternate biological functions have been proposed, chief among them that mother and infant will come together in order that the infant may learn from the mother the behaviours which will promote its survival. The conditions which activate and terminate attachment behaviour and caretaking behaviour, such as alarming events in the environment and the actual attainment of contact or proximity, indicate that Bowlby’s explanation is preferable.’

550 Hofer, M.A. (1983) On the relation between attachment and separation processes in infancy. In R. Plutnik (ed.) Emotion: Theory, Research and Experience, Vol. 2 (pp.199–219). New York: Academic Press; Hofer, M.A. (1984) Relationships as regulators: a psychobiologic perspective on bereavement. Psychosomatic Medicine, 46(3), 183–97. For a review of the literature on hidden regulators and attachment see Polan, H.J. & Hofer, M. (2016) Psychobiological origins of infant attachment and its role in development. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment, 3rd edn (pp.117–32). New York: Guilford.

551 Bowlby, J. (1983) Letter to Myron Hofer, 29 July 1983. PP/Bow/J.9/102. This correspondence supports the claim of Cassidy and colleagues that Bowlby downplayed non-representational affective and physiological processes as part of his focus on cognitive processes: Cassidy, J., Ehrlich, K.B., & Sherman, L.J. (2013) Child–parent attachment and response to threat: a move from the level of representation. In M. Mikulincer & P.R. Shaver (eds) Nature and Development of Social Connections: From Brain to Group (pp.125–44). Washington, DC: American Psychological Association.

552 Bowlby, J. (1991) Ethological light on psychoanalytic problems. In P. Bateson (ed.) The Development and Integration of Behaviour: Essays in Honour of Robert Hinde (pp.301–14). Cambridge: Cambridge University Press, p.306.

553 An update in light of evolutionary biology would partially take place from the late 1990s especially through the work of Jay Belsky. Belsky, J. (1999) Modern evolutionary theory and patterns of attachment. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment (pp.141–61). New York: Guilford.

554 Griffiths, P.E. (2008) Ethology, sociobiology, and evolutionary psychology. In S. Sarkar & A. Plutynski (eds) A Companion to the Philosophy of Biology (pp.393–414). Oxford: Blackwell. There remain some ethologists who have continued to conduct empirical studies with a focus on sequences of behaviour in the manner of Lorenz and Tinbergen. See e.g. Suomi, S. (2016) Attachment in rhesus monkeys. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment (pp.133–54), 3rd edn. New York: Guilford; Polanco, A., Díez-León, M., & Mason, G. (2018) Stereotypic behaviours are heterogeneous in their triggers and treatments in the American mink, Neovison vison, a model carnivore. Animal Behaviour, 141, 105–14.

555 Sroufe, L.A. (1986) Bowlby’s contribution to psychoanalytic theory and developmental psychopathology. Journal of Child Psychology and Psychiatry, 27, 841–9, p.841.

556 Bowlby, J. cited in J.M. Tanner & B. Inhelder (eds) (1956) Discussions on Child Development: Proceedings of the WTO Study Group of the Psychobiological Development of the Child, Vol. 1. London: Tavistock, p.184.

557 Bowlby’s passionate focus on what an assemblage can do, and not of the price of its formation or use, can be seen clearly in a late interview: Bowlby, J. (1986) An interview with John Bowlby on the origins and reception of his work. Free Associations, 1, 36–64: ‘What’s important about a person’s work is what they have contributed. I don’t care two pins about their mistakes or their shortcomings or their omissions—what have they contributed?’ (63).

558 Cf. Bartmanski, D. (2012) How to become an iconic social thinker: the intellectual pursuits of Malinowski and Foucault. European Journal of Social Theory, 15(4), 427–53.

559 Hinde, R. (2005) Ethology and attachment theory. In K. Grossman, E. Waters, & K. Grossman (eds) Attachment from Infancy to Adulthood: The Major Longitudinal Studies (pp.1–12). New York: Guilford, p.9. See also Petters, D.D. (2019) The attachment control system and computational modeling: origins and prospects. Developmental Psychology, 55(2), 227–39.

560 Bourdieu, P. (1979, 1984) Distinction, trans. R. Nice. Cambridge, MA: Harvard University Press.

561 Bucchi, M. (2013) Style in science communication. Public Understanding of Science, 22(8), 904–15.

562 Bourdieu, P. (2004) The Science of Science and Reflexivity, trans. R. Nice. Chicago: University of Chicago Press, pp.53–4.

563 Collins, H. & Evans, R. (2018) A sociological/philosophical perspective on expertise: the acquisition of expertise through socialization. In K. Anders Ericsson, R.R. Hoffman, A. Kozbelt, & A.M. Williams (eds) The Cambridge Handbook of Expertise and Expert Performance (pp.21–32). Cambridge: Cambridge University Press.

564 On the practical logic of underdetermined discourses: de Certeau, M. (1984) The Practice of Everyday Life, trans. S.F. Rendall. Berkeley: University of California Press; Bourdieu, P. (2000) Pascalian Meditations. Stanford: Stanford University Press; Mercer, J. (2015) Revisiting an article about dyadic developmental psychotherapy: the life cycle of a Woozle. Child and Adolescent Social Work Journal, 32(5), 397–404; Alexander, J.C. (2016) Dramatic intellectuals. International Journal of Politics, Culture, and Society, 29(4), 341–58; Pettit, M. & Young, J.L. (2017) Psychology and its publics. History of the Human Sciences, 30(4), 3–10.

565 See also Duniec, E. & Raz, M. (2011) Vitamins for the soul: John Bowlby’s thesis of maternal deprivation, biomedical metaphors and the deficiency model of disease. History of Psychiatry, 22(1), 93–107.

566 Bowlby, J. (1988) Where science and humanism meet. Group Analysis, 21, 81–8, p.81.

568 Bowlby, J. (1982) Attachment and loss: retrospect and prospect. American Journal of Orthopsychiatry, 52(4), 664–78, p.676.

569 Bowlby, J. (1988) Where science and humanism meet. Group Analysis, 21, 81–8, p.81.