The Cognitive Revolution

In brief

For roughly half a century, American academic psychology had ruled the contents of the mind out of scientific bounds. B. F. Skinner (1904 to 1990) and the behaviorists he led held that only observable behavior, and the stimuli and reinforcement histories that shaped it, counted as data; talk of memory, thought, or meaning was unverifiable mentalism dressed up as science. Between 1956 and 1967, that consensus broke. George Miller's 1956 paper on the limits of short-term memory showed that an internal mental capacity could be measured like any other variable. Noam Chomsky's 1959 review of Skinner's Verbal Behavior argued that language could not be explained by reinforcement at all. Ulric Neisser's 1967 textbook, Cognitive Psychology, gave the resulting field its name and its organizing metaphor: the mind as a system that takes in, stores, transforms, and retrieves information, on the model of a computer. Psychology did not abandon rigor; it relocated it to the inside of the head.

The full treatment

The problem it answers

Behaviorism had won its dominance honestly. John B. Watson's 1913 manifesto and later Skinner's operant conditioning replaced the murky introspection of earlier psychology with objective, replicable procedures: present a stimulus, record a response, manipulate reinforcement, chart the result. It worked well for simple learning in pigeons and rats. But by the 1950s it strained against phenomena it could not accommodate. Children produce and understand sentences they have never heard, with no plausible history of reinforcement for each one. People can hold a phone number in mind only as long as it takes to dial it, then lose it, a precise limit no stimulus-response account predicted. A pianist or a fluent speaker executes long, structured sequences of action that do not look like chains of individually reinforced responses. Something had to be organizing the sequence in advance. The cognitive revolution answered by proposing that the organizing structure is real, internal, and open to scientific study.

How it worked

The new approach kept behaviorism's insistence on objective evidence but aimed it inward. Rather than asking subjects to introspect on their own minds, as Wilhelm Wundt's nineteenth-century laboratory had done, cognitive psychologists inferred internal structure from patterns in behavior itself: reaction times, error types, memory span, the sentences a grammar could and could not generate. The mind's hidden architecture, on this approach, could be reconstructed the way a physicist infers an atom's structure from how it scatters particles. The controlling metaphor came from a machine barely known outside laboratories, the digital computer: if it takes input, encodes it, stores it, and produces output by program, perhaps a mind could be described the same way, attention as a limited channel, short-term memory as a small buffer, long-term memory as a vast store.

What it claims

The central claim is representational and computational: mental states are internal representations of information, and cognition is the manipulation of those representations by rule, whether the rules of a grammar, a memory-encoding process, or a decision procedure. This breaks from behaviorism, which treated any inner state as an unneeded fiction, and from introspectionism, which took conscious report as direct data: the states are real and causally responsible for behavior, but must be inferred indirectly, never simply reported.

The key studies and demonstrations

Three moments anchor the story. George Miller's "The Magical Number Seven, Plus or Minus Two" (Psychological Review, 1956) showed that people can reliably hold about seven, plus or minus two, discrete chunks of information in immediate memory, whether digits, letters, or larger grouped units, a sharp numerical limit on an internal capacity that behaviorism had no vocabulary to describe. Noam Chomsky's review of Skinner's Verbal Behavior (Language, 1959) argued that Skinner's operant concepts, stimulus, response, reinforcement, became empty when stretched to cover language: no child is reinforced for each of the infinite new sentences she eventually produces and understands, an argument later called the poverty of the stimulus. Chomsky's alternative, laid out in Syntactic Structures (1957), proposed an internal generative grammar, a finite rule system capable of producing infinite sentences, built into the mind's native equipment. Ulric Neisser's Cognitive Psychology (1967) drew these and other threads, on perception, attention, and memory, into a single textbook, giving the emerging field its name and unifying picture of mind as an information-processing system. Miller later dated the field's symbolic birth to a 1956 symposium at MIT where he presented his memory paper, Chomsky sketched an early version of generative grammar, and Allen Newell and Herbert Simon demonstrated a program that proved mathematical theorems, an early claim that machines, and minds, could reason.

Related distinctions

The cognitive revolution should be distinguished from cognitive science, the broader field it founded, formalized around the Cognitive Science Society in 1979 and spanning psychology, linguistics, philosophy, and computer science. It should also be distinguished from the computational theory of mind specifically, the claim that cognition just is symbol manipulation on the model of a computer. That was cognitivism's dominant early form, later challenged from within cognitive science itself, as the next section shows.

Lineage

The revolution's deepest debt, and its sharpest point of departure, is Wilhelm Wundt (1832 to 1920), whose Leipzig laboratory first treated conscious experience as an object of controlled study, a project behaviorism spent decades trying to bury. Karl Lashley's paper "The Problem of Serial Order in Behavior," delivered in 1948 and published in 1951, was an earlier crack in behaviorist orthodoxy, arguing that sequences like speech or piano playing require an internal plan, not a chain of triggered responses. Technical scaffolding came from outside psychology: Claude Shannon's information theory (1948), Norbert Wiener's cybernetics (1948), and Alan Turing's 1950 proposal that a machine might be judged intelligent by its behavior, all supplying vocabulary, information, feedback, computation, that cognitive psychology borrowed wholesale. Downstream, the revolution fed into artificial intelligence, cognitive neuroscience once brain imaging arrived, and Daniel Kahneman and Amos Tversky's research on judgment, which read the mind's errors as further evidence of internal architecture rather than noise.

The strongest case for it

The revolution earned its dominance by explaining what behaviorism could not, and by generating falsifiable predictions about unobservable structure. Miller's memory-span limit has been replicated and refined for decades. Chomsky's poverty-of-the-stimulus argument remains a live challenge that any purely associationist account of language must still answer. The information-processing framework proved fertile well beyond its origin, producing artificial intelligence, cognitive neuroscience, and behavioral economics as working fields rather than speculative programs. Few scientific reorientations have been this productive this fast.

The strongest case against it

The behaviorists did not concede quietly. Kenneth MacCorquodale, in "On Chomsky's Review of Skinner's Verbal Behavior" (1970), argued that Chomsky attacked a straw version of Skinner's terms, importing "stimulus" and "response" from their precise, laboratory-defined meanings, then ridiculing the loose, everyday sense he substituted for them. On this view, Chomsky's review won by rhetoric and timing more than decisive refutation, and behaviorism's retreat from language was never actually forced by disconfirming evidence. Skinner himself, in "Selection by Consequences" (1981), argued that cognitivism's inner grammars and representations were homunculi, little agents that themselves need explaining, trading one mystery for another rather than solving it.

A second attack came from within cognitive science itself. David Rumelhart and James McClelland's Parallel Distributed Processing volumes (1986) showed that phenomena taken as evidence for explicit symbolic rules, such as how children learn English past-tense verbs, could emerge instead from a connectionist network with no rules and no symbols, undercutting the computer-as-rule-follower metaphor at the revolution's core. James J. Gibson's ecological approach to perception (1979) went further, arguing that perceiving the world requires no internal representation at all, since the relevant information sits directly in the structure of light. Hubert Dreyfus, in What Computers Can't Do (1972), pressed a related objection: human understanding is embodied and contextual in ways that resist discrete symbol manipulation.

Where it stands now

Cognitive science is now an unremarkable, fully institutionalized field, with its own society, journals, and departments. Its central premise, that internal, unobservable information-processing states are legitimate objects of psychological science, is accepted even by many who reject the specific computer metaphor that first carried it. That metaphor has since been challenged and supplemented from within, by connectionist and deep-learning models, by embodied and ecological approaches, and by cognitive neuroscience's direct look at the brain doing the processing. Behaviorist methods have not vanished either: applied behavior analysis remains a serious clinical practice, and reinforcement learning, a descendant of operant conditioning, is central to artificial intelligence. The revolution's lasting achievement was not a final theory of mind but a permanent shift in what counts as a legitimate question: not whether the mind can be studied, but how.

Test yourself

Think of a sentence you have never heard or read before, one you just produced in your own head, and notice that you understood it instantly and could judge whether it was grammatical. No one reinforced you for that specific sentence. Ask what, in you, made that possible, and notice that the honest answer requires you to say something about your mind's internal structure, not only about your history of reward.