Attachment Theory

In brief

John Bowlby (1907 to 1990), a British psychiatrist and psychoanalyst, proposed in a 1958 paper and across the three volumes of Attachment and Loss (1969, 1973, 1980) that infants carry an evolved behavioral system for staying close to a protective caregiver, as biologically real as hunger or fear, not a byproduct of being fed. His collaborator Mary Ainsworth (1913 to 1999) gave the idea a method: the Strange Situation, a short procedure sorting infants into distinct patterns by how they used a caregiver as a base for exploring and a refuge in distress. Around the same time, working independently, Harry Harlow (1905 to 1981) showed that infant rhesus monkeys raised with two artificial mothers, one wire and one cloth, overwhelmingly preferred the soft one even when only the wire one fed them. Later work complicated the picture: inborn temperament shapes a child's response to separation and reunion, so the same behavior is not always a clean readout of caregiving history.

The full treatment

The problem it answers

Before Bowlby, the dominant explanations of why infants love their mothers were what he called "cupboard love" theories. Freudian psychoanalysis held attachment a secondary result of oral gratification, and behaviorism held the same thing in different language: the mother becomes a conditioned reinforcer because she delivers food. Bowlby, who had worked with delinquent and institutionalized children and wrote a 1951 report for the World Health Organization on maternal deprivation, found the drive-reduction account inadequate to what he saw. Children separated from a caregiver for extended periods, even when fed and physically cared for, showed grief and lasting difficulty in later relationships. He wanted a theory treating the bond to a caregiver as a primary system, not a side effect of feeding.

How it works

Trained partly in ethology and influenced by Konrad Lorenz's studies of imprinting in geese and Robert Hinde's work on primate behavior, Bowlby proposed that natural selection built infants with a repertoire of signals, crying, clinging, following, smiling, whose function was to keep a caregiver near, since an infant separated from protection faced serious risk. He modeled attachment as a control system, closer to a thermostat than a drive: when the caregiver is near and responsive, the infant treats her as a "secure base" and explores; when she is unavailable, or a threat appears, attachment behaviors switch on until proximity is restored. Out of repeated cycles, the infant builds what Bowlby called an "internal working model", an assumption about whether a caregiver can be counted on, carried into later relationships and, on his view, adulthood.

What it claims

The theory separates into distinct claims. Attachment is a primary motivational system, not derived from feeding. A window in the first year is especially open to being shaped. What varies across children is not whether they attach, nearly all do, but the security of that attachment, which tracks caregiver sensitivity rather than time spent together. The internal working model, once formed, resists change and colors relationship expectations into adulthood, though it remains revisable given a different pattern of relationships later.

The key study or demonstration

Ainsworth turned the theory into an instrument. After fieldwork observing mother-infant pairs in Uganda in the mid-1950s, she designed the Strange Situation: an eight-episode, roughly twenty-minute procedure in which a one-year-old is observed with its mother, then a stranger, through two brief separations and reunions. What mattered was not distress at separation but behavior at reunion. In Patterns of Attachment (1978), written with Mary Blehar, Everett Waters, and Sally Wall, Ainsworth identified three patterns from her Baltimore sample: secure infants (Type B) sought contact at reunion and were readily comforted; avoidant infants (Type A) ignored the returning caregiver; resistant infants (Type C) sought contact but resisted it angrily, unable to settle. She linked these patterns to how sensitively each mother had responded to the infant's signals across the preceding year. A fourth pattern, disorganized (Type D), marked by contradictory or frozen behavior at reunion, was identified later by Mary Main and Judith Solomon in 1986, and linked since to frightening caregiving.

Harlow's experiments at the University of Wisconsin, reported in "The Nature of Love" (1958), tested the cupboard-love hypothesis directly. Infant rhesus macaques were separated from their mothers and raised with two surrogates: a bare wire frame fitted with a feeding bottle, and a frame padded in soft cloth that provided no food. If attachment followed the feeding drive, the infants should have preferred the wire mother. Instead they clung to the cloth mother most of the day, ran to her when frightened by a strange object, and used her as a base for exploring even though only the wire mother fed them. Harlow's later, more troubling isolation experiments, in which motherless monkeys developed severe, often permanent social disturbance, made the same point from the opposite direction and helped drive later reform of animal research ethics.

Related distinctions

The most persistent complication is temperament. Jerome Kagan argued, most fully in The Nature of the Child (1984), that an infant's inborn reactivity, how intensely it responds to novelty and stress, shapes its Strange Situation behavior independently of caregiving, so a classification read as a history of care might partly be reading disposition instead. The strongest reply is that classifications are largely specific to a relationship: the same infant is often classified securely with one parent and insecurely with the other, hard to explain by temperament alone, since temperament travels with the child, not the relationship. Most researchers now treat the two as interacting, with temperament shaping the flavor of an insecure pattern more than whether attachment is secure at all.

Lineage

Bowlby's synthesis drew on ethology, evolutionary theory, and the object-relations wing of psychoanalysis associated with Melanie Klein, from which he eventually broke. Ainsworth, who had worked alongside Bowlby at London's Tavistock Clinic, supplied the observational method that made his framework testable. The theory displaced an earlier climate shaped by behaviorist child-rearing advice, associated with John B. Watson, warning against too much affection toward infants. Its descendants include Cindy Hazan and Phillip Shaver's 1987 extension of attachment categories to adult romantic love, Mary Main's Adult Attachment Interview, and Peter Fonagy's mentalization-based clinical work.

The strongest case for it

The theory has been tested across cultures and decades and has largely held up. A 1988 meta-analysis by Marinus van IJzendoorn and Pieternel Kroonenberg, pooling Strange Situation studies from eight countries, found secure attachment the most common pattern everywhere, with most variation occurring within cultures rather than between them. The Minnesota Longitudinal Study of Risk and Adaptation, led by Alan Sroufe and Byron Egeland, tracked children from infancy into adulthood and found early attachment security a real, if moderate, predictor of later social competence. The theory also has integrative power: it drew clinical observations of institutionalized children, cross-species animal evidence, and evolutionary reasoning into one falsifiable framework, generating a research program still active and still capable of surprising its own practitioners.

The strongest case against it

Beyond Kagan's temperament argument, Michael Rutter mounted an influential critique of Bowlby's boldest claims. In Maternal Deprivation Reassessed (1972), Rutter argued that Bowlby's 1951 report had bundled several distinct risks under one label, "maternal deprivation": lack of a consistent caregiver, understimulation, family discord, and institutional neglect, each with different effects, and overstated how irreversible early harm was. The English and Romanian Adoptees study, led by Rutter's team from the late 1990s, followed children adopted out of severely depriving Romanian orphanages and found substantial, though incomplete, recovery when placed in responsive families.

A separate criticism questions whether the Strange Situation measures the same thing across cultures. Keiko Takahashi's 1990 study in Japan found a much higher rate of resistant classifications among Japanese infants, plausibly because those infants, raised with less practice at brief separation, found the procedure more distressing for reasons unrelated to the underlying relationship. That raises a real question about how far the procedure encodes one local pattern of caregiving as the norm against which other ways of raising children read as insecure.

Where it stands now

Attachment theory remains the standard framework in developmental psychology for early social and emotional development, taught in every introductory course and used in child welfare and adoption practice worldwide. It is no longer presented as a single, sufficient explanation. Researchers treat attachment security as one important, partly independent contributor to later development, alongside temperament and family environment, rather than a template fixed permanently in the first year. The disorganized category identified by Main and Solomon is now the pattern of greatest clinical concern, linked to higher risk for later difficulty. The theory survived its critics by narrowing its claims, which is usually how a strong idea ages.

Test yourself

Think about how you behave when a close relationship feels threatened: whether you move toward the other person, pull away, or do both at once. Ask whether you can trace that pattern to how reliably you could count on someone in your first years, then ask honestly whether some of it is simply how you have always reacted to stress. Attachment history and temperament rarely arrive separately enough to tell apart with certainty.