The Review That Changed a Field
Kenny Easwaran walks readers through one of the most consequential book reviews in intellectual history: Noam Chomsky's 1959 demolition of B.F. Skinner's Verbal Behavior. The review did not merely critique a book. It helped inaugurate cognitive science as a discipline, dethroning behaviorism's grip on psychology and linguistics in the process. Easwaran, a philosopher at Texas A&M, uses the review as a teaching text, and his commentary captures both the force of Chomsky's arguments and the rhetorical savagery with which they were delivered.
Chomsky's central charge is that Skinner's behaviorist vocabulary -- stimulus, response, reinforcement -- functions coherently only within the narrow confines of laboratory experiments with rats and pigeons. The moment Skinner attempts to apply these terms to human language, they lose all objective content and become, in Chomsky's words, "mere paraphrases for the popular vocabulary commonly used to describe behavior." The technical apparatus creates an illusion of scientific precision while explaining nothing that ordinary language does not already cover.
The Dilemma of Stimulus and Response
Easwaran highlights the structural elegance of Chomsky's critique. Chomsky presents Skinner with a dilemma: either define "stimulus" and "response" broadly -- in which case nearly all behavior falls outside any lawful relationship and the theory explains almost nothing -- or define them narrowly, restricting them only to cases where behavior and environment show smooth, reproducible curves. The narrow definition makes the theory true by definition but confines it to trivially simple cases like bar-pressing in a Skinner box.
Skinner does not consistently adopt either course. He utilizes the experimental results as evidence for the scientific character of his system of behavior, and analogic guesses formulated in terms of a metaphoric extension of the technical vocabulary of the laboratory as evidence for its scope.
This is the fatal move Chomsky identifies: Skinner trades on the prestige of laboratory results while stretching his concepts so far beyond the lab that they retain none of their original rigor. When Skinner claims that seeing a painting controls the verbal response "Dutch," Chomsky observes that a viewer might equally say "clashes with the wallpaper," "I never saw this one before," "beautiful," "hideous," or "remember a camping trip last summer." If every possible response is under the control of some "property" of the stimulus, the concept of stimulus control explains everything and therefore nothing.
Reinforcement as Ritual Invocation
The critique of Skinner's use of "reinforcement" is even more devastating. In the laboratory, reinforcement has a clear operational meaning: a food pellet drops into a tray when the rat presses the bar, and the rate of bar-pressing increases. But as Easwaran traces through Chomsky's examples, Skinner extends the term to cover situations where no identifiable reinforcing stimulus exists at all. A person talks to himself "because of the reinforcement he receives." A child is "automatically reinforced" when duplicating the sound of airplanes. A writer is "reinforced by the fact that his verbal behavior may reach over centuries or to thousands of listeners."
From this sample it can be seen that the notion of reinforcement has totally lost whatever objective meaning it may ever have had... the phrase "X is reinforced by Y" is being used as a cover term for "X wants Y," "X likes Y," "X wishes that Y were the case," etc.
Chomsky's point is not merely that Skinner is vague. It is that the vagueness is systematic and self-serving. By calling everything "reinforcement," Skinner obscures genuine distinctions among wanting, liking, hoping, wishing, and believing -- distinctions that a cognitive approach would take as its starting point. The word "reinforcement" performs, as Chomsky puts it, "a purely ritual function."
The Poverty of the Stimulus
Having dismantled the behaviorist framework, Chomsky offers his alternative. The final section of the review sketches the program that would dominate linguistics for the next half-century. Language, Chomsky argues, cannot be explained by general-purpose learning mechanisms because the data available to the child is radically insufficient to account for what the child learns. Every normal child, regardless of intelligence, acquires an enormously complex grammar in just a few years -- a grammar capable of generating and recognizing an infinite number of sentences, most of which the child has never heard.
The fact that all normal children acquire essentially comparable grammars of great complexity with remarkable rapidity suggests that human beings are somehow specially designed to do this, with data-handling or hypothesis-formulating ability of unknown character and complexity.
This is the poverty of the stimulus argument that became Chomsky's signature contribution. The child does not learn language the way a rat learns to press a bar. Something innate -- a language faculty, a universal grammar -- must be doing much of the work. Easwaran notes that Chomsky situates this debate within the centuries-old philosophical contest between empiricism and rationalism, placing himself explicitly in the rationalist tradition of Descartes against the empiricist tradition of Locke and, by extension, Skinner.
Where the Counterarguments Land
Easwaran is not uncritical of Chomsky's performance. He notes that "some defenders of Skinner have argued that the view that it's criticizing is a caricature of Skinner rather than the actual Skinnerian view," and he links to published defenses of Skinner against the review. The tone is, as Easwaran acknowledges, "incredibly scathing," and in places Chomsky may be uncharitable. When Chomsky mocks Skinner's method for evoking the response "pencil" -- asking someone to say "pencil," or holding a gun to their head -- the satire lands, but it elides whatever subtlety Skinner may have intended.
More significantly, Easwaran raises a point that Chomsky could not have anticipated in 1959: the rise of deep learning. Modern neural networks do not operate on anything as crude as Skinnerian stimulus-response chains, but they do rely on domain-general learning mechanisms rather than the specialized, innate modules Chomsky posited. Large language models acquire linguistic competence that would have seemed impossible to explain without innate structure, yet they do so through statistical patterns in data -- a result that sits uncomfortably with the poverty of the stimulus argument. Easwaran observes that "in some ways" these systems suggest "reinforcement learning may be sufficient to get a much larger amount of mental behavior than Chomsky and the cognitivists would have thought."
This is a genuinely interesting complication. Chomsky's critique of Skinner remains logically sound on its own terms -- Skinner's vocabulary really does collapse into mentalistic paraphrase when applied to human language. But the deeper claim, that no empiricist approach could ever account for language acquisition, looks less certain in an era when transformers trained on text corpora produce grammatical sentences, resolve ambiguities, and generalize to novel constructions. The question is no longer whether domain-general learning can produce complex linguistic behavior, but what kind of architecture and training signal it requires -- and whether that architecture constitutes, in some meaningful sense, the innate structure Chomsky predicted.
Bottom Line
Easwaran's walkthrough of Chomsky's review is a valuable guide to one of the twentieth century's most important intellectual confrontations. Chomsky did not simply win an argument about a book; he redrew the boundaries of what counts as a scientific explanation of human behavior. The review's lasting contribution is its insistence that the complexity of language demands a theory of internal structure, not merely a catalog of input-output correlations. Whether that internal structure is best described as an innate universal grammar or as something more like the learned representations of a neural network remains an open and increasingly urgent question. What Skinner got wrong was not the impulse to look for general principles of learning, but the refusal to acknowledge that what is being learned -- human language -- has a structure so rich that any adequate theory must reckon with it on its own terms.