John B. Watson

The question of 1920 was whether fear can be conditioned in a human being the way salivation is in a dog. Watson and Rosalie Rayner wanted to show that an emotional reaction need not be an innate instinct but can be learned — and that this can be manufactured in the laboratory. The stakes were higher than the report’s four pages suggest: Watson had promised the field in his manifesto seven years earlier that behaviourism would explain emotional life too, and the instinct lists of the competition — William McDougall was then running dozens of innate propensities — were to shrink to three inborn reactions: fear, rage, love. Everything else, ran the claim, is conditioning. Little Albert was the evidence meant to carry that claim, and he carries more textbook pages to this day than any other single child in the history of science.

The subject was an infant, called Albert B. in the report, eight months and 26 days old at the first test — son of a wet nurse who lived and worked at the Harriet Lane Home of Johns Hopkins University. He was described as calm and unusually fearless; at a white rat, a rabbit, a dog, cotton wool and burning newspaper he at first reacted with curiosity. Exactly this stability was the selection criterion: Watson and Rayner wrote that with so robust a child the experiment could do “relatively little harm” — a sentence proving that the question of harm was raised and set aside.

The unconditioned stimulus was a noise: a steel bar four feet long was struck with a hammer behind Albert’s back. At this the child startled and cried — one of the three inborn reactions Watson had catalogued in preliminary studies of hundreds of infants. In the conditioning trials the rat appeared, and at the moment Albert touched it, the hammer fell. After a total of seven such pairings, spread over two sessions, he cried as soon as the rat appeared alone, and crawled away from it.

Five days later they tested generalisation. Albert reacted to a rabbit, a dog, a fur coat, cotton wool and a Santa Claus mask with a beard. The fear had spread to similar things. A month later he still showed the reactions, weakened but present — the report scored that as a finding too: learned fear endures.

![Still from Watson’s 1920 film](../../../../assets/figuren/watson-albert.jpg "Fig. 2 — Still from the 1920 film: testing generalisation — “Now he fears even Santa Claus.” Public domain · Wikimedia Commons”)

The textbook version ends here, often with the addition that Watson afterwards trained the fear away again. That is false. The report itself records that a deconditioning was planned and did not happen: Albert left the hospital before it could begin. Watson knew the date in advance. The confusion has a real core, and it belongs to the reception history: the untraining existed — but on another child. In 1924 Mary Cover Jones, advised by Watson, treated three-year-old Peter, who came into her study with a fear of rabbits, and dismantled the fear step by step by bringing the animal a little closer at every meal. The case of Peter is the counterpart that the textbooks narrate into the case of Albert — and at the same time the first documented behaviour therapy: Joseph Wolpe, who developed systematic desensitisation, later called Cover Jones the mother of the method.

The series was run in the winter of 1919–20 and published in February 1920 in the Journal of Experimental Psychology — four pages, without statistics, in the tone of a workshop report. Watson also had the sessions filmed and cut them into a teaching film on his infant research: Little Albert is thus among the first experimental subjects in psychology whose behaviour was not only recorded in writing but captured on film — the beginning of a form of evidence that occupies the field to this day, because films show but do not measure.

Methodologically the study does not bear weight by today’s standards. Sample: one child. No control condition. The fear reaction was not measured but judged by the experimenters — the same people who had framed the hypothesis and wielded the hammer. The surviving film footage shows reactions that can be read otherwise: a tired, beleaguered child turning away. Replication attempts by other laboratories in the 1930s — Bregman, for instance, failed to condition fear onto geometric objects — came out inconsistent and suggested that not every stimulus links to fear equally well. Half a century later Martin Seligman turned exactly this unevenness into a concept: preparedness — evolution has pre-paved some connections; rats and spiders are learned more easily than building blocks. The finding that survived Albert is thus smaller and more interesting than Watson’s thesis: fear is conditionable, but not arbitrarily.

The study’s career in the textbooks has itself become a research subject. In 1979, under the title ”Whatever Happened to Little Albert?”, Ben Harris went through the printed versions and catalogued the deviations: textbooks had Albert react to teddy bears, cats, white gloves or his mother’s fur hat — stimuli that do not occur in the report —, shifted his age, invented the cure, and rounded the story until it carried its moral cleanly. Hardly any classic of the field has been misretold so often, and the file does not record this as a curiosity: a discipline that makes its founding stories prettier than its records allow has a replication problem already in the library.

Rosalie Rayner, whose name stands on the report and regularly disappears from the retellings, was a graduate student at the time of the study; she married Watson in 1921, later published with him on child-rearing, and died in 1935 at 36. Her share in design and execution can no longer be separated out of the sources — that too a file note on the division of labour of those years.

Ethically the study became the field’s precedent — the case with which introductory lectures still explain why ethics committees exist. An infant who cannot consent; a mother in a dependent position whose documented consent is missing; a planned harm without a planned repair. The American Psychological Association gave itself its first ethics code in 1953; that Little Albert belongs in its prehistory is a consensus of the discipline’s history.

In substance the conditioning explanation of fear has held only in part. Ask people with pronounced animal phobias and a considerable share can recall no triggering experience, while conversely the majority of those who have suffered a dog bite or a fall develop no phobia. Later learning theory made three routes of this: one’s own experience, observation of others, and mere information — fear can also be taken over and told. To this comes the observation that some stimuli attach to fear more readily than others: snakes and spiders yes, electrical sockets and cars hardly, although the latter are incomparably more dangerous. What Watson demonstrated on one child is thus one route among several, and one with a built-in bias.

A repetition of the study has been out of the question for decades — no ethics committee in the world would approve it. Little Albert thus belongs to the findings that have become unreplicable in principle: what is true in them can only be tested indirectly, through fear conditioning with more harmless stimuli and through what therapy has made of it.

Unresolved to this day is who Albert was, and the file holds both candidates. Hall Beck and colleagues identified him in 2009, after years of archival work, as Douglas Merritte, son of a hospital wet nurse, who died in 1925 of hydrocephalus — which would mean Watson described a neurologically impaired child as “healthy”; the authors around Alan Fridlund sharpened the suspicion in 2012 into the charge that Watson knew the child’s condition and concealed it. Russ Powell and colleagues disagreed in 2014 and produced evidence for William Albert Barger — born in the same building, at the fitting time, son of another wet nurse — who died in 2007 at the age of 87 and whose relatives reported that he avoided animals all his life without ever learning why anyone asked. The question is open, and it is more than a footnote: depending on who Albert was, the same study is a rough classic or a research scandal. That the field only began to settle the difference ninety years later says more about its record-keeping than about Watson.