Food Was Never the Answer

Eight infant rhesus monkeys, two dummies: one of bare wire, one covered in terry cloth. In Wisconsin, Harry Harlow separated the young animals from their mothers and counterbalanced the feeding — for four of them the wire mother gave the milk, for the other four the cloth mother. The ruling doctrine had a clear prediction: the infant belongs to the milk source. The infants spent most of the day on the cloth mother — including those fed by the wire mother, which they visited only to drink. Frightened, they fled without exception to the terry cloth, not to the milk. Harlow reported the experiment in 1958 in the American Psychologist, under the title The Nature of Love. The counterbalancing is the decisive trick: without it, the clinging to the cloth mother could still be read as a side effect of the feeding. Contact comforts; food does not.

What fell with it was a piece of bookkeeping on which the field’s two great schools had, for once, agreed around 1950. For learning theory the mother was a conditioned reinforcer — she stands next to the food, so the food rubs off on her. For psychoanalysis she was the object of oral drive satisfaction: another language, the same sum. In both versions affection was a detour of nutrition. That this sum collapsed within twenty years is one of the cleanest rebuilding stories in the field, and it ran through three stations.

The ethical price of the monkey experiments

Harlow’s rooms are the place where the knowledge and the ethics of this subject cannot be divided from each other. In follow-up experiments he reared animals in complete isolation and showed that without an early relationship, gravely disturbed adults grow up. These are findings whose ethical price is hard to bear today and which would no longer be approved in this form. This archive’s research-ethics note addresses that explicitly; the effect of the experiments on the theory is a separate matter, and it was enormous. The same isolation experiments that refuted the nutritional theory produced gravely disturbed animals; later primate research learned from both — from the findings and from the price. Whoever argues with Harlow today argues with results the field has forbidden itself to repeat.

Bowlby’s break with interpretation

In London, John Bowlby had the same suspicion at the same time, from different material. His study of forty-four juvenile thieves, published in 1944, had counted instead of interpreted; his WHO report of 1951 carried the separation question around the world. He fetched his explanation from Lorenz and Tinbergen rather than from Freud: attachment as an independent biological system, evolved because closeness to a protector meant survival. Not a by-product of milk or drive but a control loop of its own with expressions of its own — crying, clinging, following, protest at separation. The British Psycho-Analytical Society treated him thereafter as a renegade. Harlow’s monkeys and James Robertson’s hospital films, among them A Two-Year-Old Goes to Hospital of 1952, supplied the vividness his theory lacked. By the time the first volume of his trilogy appeared in 1969, the heresy had become a research programme.

What Ainsworth made measurable

What the programme lacked was an instrument. It came from Mary Ainsworth, who had worked with Bowlby before she sat in living rooms in Uganda and Baltimore observing mothers with their children, for hundreds of hours. Her Strange Situation of 1969/70 condensed those observations into twenty minutes of laboratory: eight episodes of rising strain, in which what is coded is not the crying but the behaviour at reunion. Securely attached children seek closeness, let themselves be comforted, go back to playing; avoidant ones ignore the return; ambivalent ones cling and resist at the same time. A fourth pattern was added only decades later. In the eighties Mary Main and Judith Solomon described the disorganised children, whose behaviour at reunion collapses in on itself; Ainsworth’s three-way division had left exactly those pairs over as unclassifiable. The theoretically most important finding comes not from the laboratory but from the home visits of the Baltimore study, published in 1978 as Patterns of Attachment: the mother’s everyday sensitivity predicted the later classification. And the most important clarification stands in this archive’s Ainsworth file: the categories are strategies, not diagnoses — the avoidant child too has learned something sensible, namely that showing distress helps little with this particular caregiver.

The detour through Uganda belongs in this story because it explains why the instrument turned out so well. Ainsworth did not go into the field with a questionnaire but with notebooks. From 1954 onwards she observed twenty-six families in the area around Kampala, each of them over nine months, every fortnight for two hours, with an interpreter. She described when infants used their mothers as a secure base — the term that later carried the theory comes from those living rooms. The Strange Situation was meant as a supplement to the home visits, not as their replacement; that the twenty-minute laboratory displaced the fieldwork it grew out of is an irony Ainsworth herself lamented. Whoever understands the categories without knowing the observational work behind them easily takes them for a sorting scheme. They were a summary of thousands of hours of observation.

What the tests left standing

Attachment research has been through every mill since. The Strange Situation became a world standard with a strict coding discipline — and with the price of small, expensive samples. The meta-analysis by Marinus van IJzendoorn and Pieter Kroonenberg pooled thirty-two samples from eight countries in 1988, N = 1990. On average it found a distribution close to Ainsworth’s Baltimore result: around 65 % secure, 21 % avoidant, 14 % ambivalent. The distributions differ between countries, though by less than samples within a single country differ from one another; individual samples resembled foreign ones more than their own compatriots. Which of the two spreads carries the cultural reading is disputed to this day.

In 1972, in Maternal Deprivation Reassessed, Michael Rutter took apart monotropy — the claim that only a single attachment figure counts — and separated separation from deprivation. The Romanian institutionalised children after 1989 became the test nobody had planned. The English-Romanian adoption study followed them into young adulthood: severe early deprivation leaves measurable traces there, and catch-up curves show at the same time how much recovery is possible. The predictive power of the early classification itself is real and modest — it shifts probabilities, it seals no lives. Whoever reads slogans like “attachment style” as a verdict on a life is reading against the data.

What remains

The core is secure: closeness and comfort are a need in their own right, not a side effect of feeding. It is equally secure that an adult’s sensitivity is connected with the child’s later behaviour in the Strange Situation — and that it can be worked on. Short, video-based programmes show parents how their child sends signals and what a prompt, fitting answer looks like. In randomised trials they improve parental behaviour and the attachment distribution, with effect sizes around d = 0.33 for sensitivity and d = 0.20 for attachment security. That is less than the advice literature promises, and more than most of the theories in this archive have ever brought into a randomised test.

What remains contested is the reach: how much the classification says about later relationships, and how culturally the distributions are to be read. The archive therefore lists the stations of this story with different weights — Ainsworth’s Patterns of Attachment as replicated, Bowlby’s thieves study of 1944 as only partly covered, since Rutter took its inferences apart. In everyday life the difference between a finding and a slogan shows in one feature: the research describes a relationship between two people, the advice books a label for one. At the end stands a point this archive is glad to record: the school that expelled Bowlby today cites attachment research as its best empirical support. The road there led over a terry-cloth mother, a paperback report and a coding system that cannot be learned on the side.