John Bowlby

Introduction

John Bowlby (26 February 1907 – 2 September 1990) argued that a child's need for a caregiver is not a habit formed by feeding and not a sentimental invention of the middle classes, but a biological requirement — a behavioural system with the same standing as hunger or sex, evolved because an infant who stays near a protector is more likely to survive.

That claim sounds obvious now, which is a measure of how completely it won. When Bowlby made it in the 1950s it was heretical on two sides at once. Behaviourism and psychoanalysis both held that the infant's tie to the mother was a secondary drive — the child loves the mother because she feeds it, and the bond is a by-product of drive reduction. Bowlby argued the opposite: the bond is primary, feeding is one of the things that happens within it, and a child can be perfectly nourished and still be starved of contact.

He spent forty years building the case, drawing on ethology, control theory, developmental psychology and psychoanalysis — and he broke with the psychoanalytic establishment that trained him, because its theory located the important events inside the child's fantasy rather than in the child's actual experience.

He also produced one of the most consequential and most criticised claims in twentieth-century psychology: that prolonged separation from a primary caregiver in early childhood does lasting damage. That claim was overstated, and the correction of it — largely by Michael Rutter — is as much a part of the story as the original.

Early Life and Education

Bowlby was born in London into a comfortable upper-middle-class family. His father, Anthony Bowlby, was a prominent surgeon who later became Surgeon to the King; his mother, May Brignal Mostyn, was the daughter of a bishop.

The household ran on the standard pattern of its class and time. The children were looked after by a nanny and saw their parents briefly and formally — Bowlby later recalled seeing his mother for perhaps an hour a day. His older brother died when Bowlby was five. When he was four, the nanny who had been his principal caregiver left.

He was sent to Lindisfarne, a preparatory school, and then to the Royal Naval College at Dartmouth, on the assumption that he would follow a family pattern of naval and medical service. He did not want it, and left after a year.

At Trinity College, Cambridge, he read psychology and the preclinical sciences, and graduated in 1928. The psychology course there was one of the few in Britain and it gave him the thing he used for the rest of his life: a habit of treating a psychological question as a biological one.

He then spent a year teaching at Priory Gate, a school for maladjusted children, and this is the origin of everything. He worked with two disturbed boys: one a thief, incapable of affection and of guilt; the other withdrawn and frozen. Both, he found, had suffered severe loss or separation in early childhood. The coincidence was not statistical, but it was enough to fix the question.

He then trained in medicine at University College Hospital, qualifying in 1933; in psychiatry at the Maudsley under Aubrey Lewis; and in psychoanalysis at the British Psychoanalytical Institute, where his training analyst was Joan Riviere, a close collaborator of Melanie Klein.

Academic and Professional Career

The break with psychoanalysis

Bowlby qualified as an analyst in 1937 and began working at the London Child Guidance Clinic in 1936. He was, formally, a Kleinian — but he had come to think that the Kleinian emphasis on unconscious phantasy got the causality backwards. The analytic orthodoxy held that what mattered about a child's history was how the child had fantasised it. Bowlby's clinical experience was that what mattered was what had happened: whether the mother had actually gone, for how long, and who had looked after the child while she was away.

The consequence was a long and bitter estrangement. His 1940 paper "The Influence of Early Environment in the Development of Neurosis and Neurotic Character" argued the point directly, and the British Psychoanalytical Society treated him as having left the fold. He did not formally break with it, but he stopped being part of it.

The war, and the juvenile thieves

During the Second World War Bowlby served in the Royal Army Medical Corps, working on officer selection for the War Office Selection Boards — an experience that taught him how to design studies of real people in real institutions.

In 1944 he published "Forty-Four Juvenile Thieves: Their Characters and Home-Life", comparing forty-four thieves referred to a child guidance clinic with forty-four non-delinquent children referred to the same clinic. Of the thieves, seventeen had experienced prolonged separation from their mothers before the age of five; of the controls, two. And fourteen of the thieves were what Bowlby called affectionless characters — children who could not form ties, lied without shame, and showed no guilt.

The study was correlational and the sample tiny, and Bowlby knew it. But it gave him the material for the argument he made for the next thirty years.

The Tavistock years

After the war Bowlby joined the Tavistock Clinic, becoming deputy director in 1946 and head of the Children and Parents Department in 1950. He built a research unit there and hired Mary Ainsworth in 1950, and the collaboration that followed produced the empirical apparatus his theory needed.

In 1951 the World Health Organization commissioned him to write a report on the mental health of homeless children, published as Maternal Care and Mental Health. Its central sentence — that "the infant and young child should experience a warm, intimate, and continuous relationship with his mother (or permanent mother-substitute) in which both find satisfaction and enjoyment" — became one of the most quoted and most misused sentences in post-war social policy.

The report did not say that mothers should not work, and Bowlby said so repeatedly. But the language is insistently maternal, the framing implies a single irreplaceable figure, and the political use of it in the 1950s — as scientific backing for keeping women out of the workforce — was real and was not something he successfully prevented.

Ethology, and the trilogy

In the 1950s Bowlby found the theoretical resources he had been missing. He encountered ethology — the study of animal behaviour in natural conditions — and specifically Konrad Lorenz's work on imprinting, in which a gosling forms an irreversible attachment to the first moving object it sees, without being fed by it. Here was a demonstrable attachment that had nothing to do with feeding. He also drew on Niko Tinbergen, on Robert Hinde, and on the control-systems language of cybernetics.

The experimental confirmation came from Harry Harlow, whose infant monkeys, given the choice between a wire mother that fed them and a cloth mother that did not, clung to the cloth one and went to the wire one only to feed. Contact comfort, not feeding, was the basis of the bond. Bowlby's theory and Harlow's data together finished the secondary-drive account.

Bowlby's major statement is the trilogy Attachment and Loss: Attachment (1969), Separation: Anxiety and Anger (1972), and Loss: Sadness and Depression (1980). He was made CBE in 1972. He died at his home on the Isle of Skye on 2 September 1990.

Major Contributions to Psychology

Attachment as a behavioural system

Bowlby's theoretical claim is that attachment is a behavioural system with a biological function — protection — and a predictable structure. The system has a set-goal (maintaining proximity to the attachment figure), is activated by particular conditions (alarm, fatigue, illness, unfamiliarity), and is terminated by a predictable outcome (contact). Crying, clinging, following, smiling and calling are the system's components, and they are coordinated rather than separate.

The infant is not passive in this. Attachment behaviours are what the infant does to bring about the outcome, and they are organised from the beginning toward a figure who is identified, not merely whoever happens to be feeding.

The secure base is the other half. Attachment is not clinging — the whole point of a secure attachment is that it licenses exploration. A securely attached child uses the caregiver as a base from which to move out and to which to return, and the quality of the attachment shows up less in how much contact the child seeks than in how confidently they leave.

Internal working models

Bowlby borrowed the concept of the internal working model from the psychologist Kenneth Craik's 1943 book The Nature of Explanation, and it is the part of the theory that has proved most generative. The child builds models of the attachment figure (is she available when I need her?) and of the self (am I the kind of person who is worth responding to?). The models are largely unconscious, they are built from actual experience, and — crucially — they are used to anticipate and interpret subsequent relationships without being revised, because they operate outside awareness.

This is what makes continuity possible without making it inevitable. The model shapes what the person expects and how they read other people's behaviour, which tends to bring about the outcome the model predicted.

The sequence of separation

Working with James Robertson, a psychiatric social worker at the Tavistock who filmed children in hospital and residential care, Bowlby described the standard sequence following a prolonged separation from the attachment figure:

  1. Protest — the child cries, searches, actively attempts to recover the figure, and rejects substitute caregivers
  2. Despair — the child becomes quiet, withdrawn, apathetic, and may appear to settle
  3. Detachment — if the figure returns, the child does not greet them; they are superficially settled but the tie has been defended against

The third stage is the dangerous one, because it looks like recovery to an adult who does not know the sequence. Bowlby and Robertson's work — particularly the 1952 film A Two-Year-Old Goes to Hospital — changed hospital policy in Britain and elsewhere, leading to unrestricted parental visiting and rooming-in. That is a reform attributable fairly directly to one research programme.

The stages of attachment

Bowlby and Ainsworth proposed a developmental sequence:

  • Pre-attachment (birth to about 6 weeks) — the infant signals to anyone and discriminates little
  • Attachment in the making (6 weeks to 6–8 months) — the infant responds preferentially to familiar figures
  • Clear-cut attachment (6–8 months to 18–24 months) — the infant actively seeks the specific figure, shows separation anxiety, and uses them as a secure base
  • Goal-corrected partnership (from about 24 months) — the child can now take the caregiver's plans and feelings into account and negotiate

The fourth stage is Bowlby's point about what the system is for: it is not a state the child has but a relationship that is continually renegotiated.

Major Publications

  • "Forty-Four Juvenile Thieves: Their Characters and Home-Life" (1944) — the study that started the programme
  • Maternal Care and Mental Health (1951) — the WHO report, and the most consequential thing he wrote
  • Attachment and Loss, Vol. 1: Attachment (1969) — the theory stated in full, with ethology and control theory
  • Attachment and Loss, Vol. 2: Separation: Anxiety and Anger (1972) — separation responses
  • Attachment and Loss, Vol. 3: Loss: Sadness and Depression (1980) — bereavement, drawing on Freud's mourning and on Darwin
  • Charles Darwin: A New Life (1990) — his last book, on the man he regarded as the founder of the ethological approach

Impact and Legacy

Attachment theory is one of the few large frameworks in developmental psychology that has grown rather than shrunk. It generated Ainsworth's Strange Situation and the classification of attachment patterns; it was extended to adult romantic relationships by Cindy Hazan and Phillip Shaver in 1987; it produced the Adult Attachment Interview and the whole tradition of attachment-informed clinical work; and it now connects to work on stress regulation and the developing neurobiology of the stress response.

Its practical effects are unusually direct. Hospital visiting policy, the design of residential care, adoption and fostering practice, and the assessment of children in care proceedings all changed because of it. Bowlby's argument that a child's behaviour cannot be understood apart from their actual history is now standard in child protection work.

And the theory supplied the concepts that later work operates with: secure base, internal working model, separation anxiety, the idea that the infant is an active agent in forming the relationship. See Attachment Theory.

Criticism

Maternal deprivation was overstated, and Rutter corrected it. This is the main revision. Bowlby used "maternal deprivation" to cover several things that need separating: separation (losing a figure to whom one is attached), deprivation (having a figure but of poor quality), and privation (never having had one). Michael Rutter's 1972 Maternal Deprivation Reassessed and his 1981 reconsideration argued that the evidence did not support the strong claim that early separation alone produces later psychopathology — that the outcomes Bowlby described are mediated by the quality of care after the separation, by family conflict, and by the whole pattern of the child's environment, and that Bowlby had underweighted the role of subsequent experience. Rutter's formulation is the one now standard: early separation is a risk factor that operates through its effects on later relationships, not a direct cause with a specific outcome.

The Romanian adoptee studies are the decisive test, and they cut both ways. The English and Romanian Adoptees study followed children adopted out of the Ceaușescu-era orphanages, where children were severely deprived rather than merely separated. Children adopted before about six months showed substantial and often near-complete recovery. Children adopted after six months showed persistent difficulties — cognitive impairment, quasi-autistic features, inattention and overactivity, and difficulties with peer relationships — that had not resolved years later. That is a genuine vindication of Bowlby's claim that there is a period in which the damage is done, and a genuine correction of his implication that the damage is uniform and largely irreversible.

The politics of the WHO report are a real charge. Bowlby's language was consistently about "the mother", and whatever he said about substitutes, the report was used in the 1950s and 1960s as scientific warrant for a domestic arrangement in which women stayed home. Feminist critics from Betty Friedan onward have treated that as a substantive harm rather than a misreading, and it is difficult to separate the theory's content from its use.

Attachment is not destiny, and the effect sizes are modest. Predictions from early attachment to later outcomes are real but far weaker than the theory's framing implies; the correlations between infant attachment classification and later adjustment are modest, and later experience revises early patterns substantially. Bowlby himself was more careful about this than his followers have been, and the popular version — that the first three years determine the person — is not supported.

And the theory is culturally specific. Attachment theory was built on a mid-century, Western, middle-class nuclear family with one principal caregiver. In societies where care is distributed among many people, the model of a single figure to whom the child is monotropically attached does not describe the arrangement — and cross-cultural work on the Strange Situation found, for example, high rates of "avoidant" classification in German samples and high rates of "resistant" classification in Japanese samples, which reflect different childrearing conventions rather than different rates of insecurity. The classification is not culture-free, and applying it elsewhere without adjustment has produced misleading results.

Monotropy is not his term and overstates his position. The emphasis on a principal attachment figure is Bowlby's, but he accepted that children have multiple figures organised hierarchically, and the word often attached to his view was not one he used.

The theory's use in family courts has been formally challenged by attachment researchers themselves. This is the most substantial recent development, and it is the direct descendant of the WHO report problem: attachment theory was once again pressed into institutional service, and once again the institution used a version of it that the research does not support. In 2021 seventy attachment researchers and practitioners, including Marinus van IJzendoorn, Pehr Granqvist, Michael Lamb, Dante Cicchetti, Glenn Roisman, Jude Cassidy, Charles Zeanah, Mary Dozier, Ruth Feldman and Robbie Duschinsky, published a consensus statement in Attachment & Human Development reviewing how the theory is applied in child protection and custody proceedings. Their conclusions were unusually blunt for a document written by the field's own specialists. Attachment quality, they wrote, is not the same as relationship quality, and it cannot be judged from isolated behaviours. The theory is routinely misread as prescribing one "psychological parent" — usually the mother — with the corollary that a relationship with one caregiver comes at the expense of others, when children can and do maintain secure attachments to several figures at once, and a network of relationships is itself protective. And crucially, attachment measures validated at group level have insufficient precision for individual-level prediction: the authors concluded that such assessments are suitable for directing supportive interventions, that they should never be used in isolation, that they should be carried out only by formally trained observers using standardised protocols, and that their members remained in disagreement about whether they should inform custody decisions at all.

The statement also noted that attachment theory has been held to categorically prescribe joint physical custody with equal time allocation regardless of the child's age, including daily overnight transitions — a claim for which the authors found a notable scarcity of empirical research. What makes the episode worth recording alongside the WHO controversy is the pattern rather than the content: in both cases a psychological theory was adopted by an institution that needed a scientific-looking basis for decisions it had to make anyway, and in both cases the theory's originators were not in a position to control the use. In 1951 Bowlby said the report did not mean mothers should not work; in 2021 the researchers said the measures do not do what courts want them to do. Neither disclaimer is usually what gets cited.

Associated Figures

  • Mary Ainsworth — who joined his Tavistock unit in 1950 and supplied the empirical apparatus
  • James Robertson — whose observational films of separated children produced the protest–despair–detachment sequence and changed hospital policy
  • Harry Harlow — whose monkey experiments demolished the feeding account of attachment
  • Konrad Lorenz and Niko Tinbergen — the ethologists whose work supplied the theoretical framework
  • Michael Rutter — who corrected the deprivation hypothesis and established the current evidence-based position
  • Joan Riviere — his training analyst; a Kleinian, and the tradition he broke from

Related Tests

Learn More

References

  • Bowlby, J. (1969). Attachment and Loss: Vol. 1. Attachment. Basic Books.

Last updated: 2026-10-20

Menu
Loading