Stimulus and Response

The first organised dissent came not from cognitive psychology but from Berlin: the Gestalt school disputed, as early as 1912, that perceiving and thinking can be assembled from stimulus-response building blocks. Wertheimer’s apparent motion showed an impression contained in no single stimulus, and Köhler’s chimpanzees solved tasks for which no reinforcement history held the ready answer — insight as restructuring rather than chaining. Behaviourism never refuted this objection, only outvoted it; the cognitive turn inherited it.

For forty years psychology was in large part a field that kept quiet about the inside. In 1913 Watson demanded that introspection be given up and that the discipline confine itself to observable behaviour; the aim was to be prediction and control, not understanding. The renunciation was not a dogma without a reason. What two observers can count independently of each other can be replicated; what a person reports about their own experience cannot.

The gain from that renunciation is still in use. Pavlov’s conditioned reflex is among the best-evidenced phenomena in the field and is stable across species, kinds of stimulus and decades. Skinner’s operant chamber made behaviour continuously measurable rather than countable in trials, and his schedules of reinforcement describe a pattern that sits inside every slot machine and every feed today: irregular reward binds more strongly than reliable reward. Both findings stand independently of the theory that produced them.

The price was reach. The programme explains well how a response becomes more or less frequent; it explains badly how someone forms a sentence they have never heard. Chomsky’s 1959 review of Skinner’s Verbal Behavior went precisely there, and the review’s effect on the field was greater than that of the book it reviewed. The question was not whether reinforcement works, but whether it is enough.

The end came not as a refutation but as a shift of interest. Neisser’s Cognitive Psychology gave a name in 1967 to a field that wanted to describe inner processes again — as information processing, in stages between stimulus and response. What behaviourism had forbidden was not thereby permitted again but stated differently: not as a report about experience, but as a model from which measurable predictions can be derived.

The files under this topic show the transition in its weaknesses too. Watson’s best-known experiment has a sample of one, no control condition and an open ending; Skinner’s reading of his pigeons' superstition was replaced in 1971 by Staddon and Simmelhag with a more economical explanation. And Neisser charged his own field, nine years after founding it, with having retreated into artificial laboratory tasks — the same charge behaviourism had had to listen to, newly cast.