Behaviorism

In brief

In 1913, John B. Watson announced that psychology had been asking the wrong question. It should stop trying to describe the contents of consciousness through introspection and start doing what any other natural science does: observe things from the outside. The proper data of psychology, he argued, were stimulus and response, nothing else. Three decades later B. F. Skinner pushed the program further, arguing that even thoughts and feelings were themselves behavior, not hidden causes behind it, and built an entire technology of learning, the operant conditioning chamber, reinforcement schedules, shaping, around that claim. For half a century behaviorism was the closest thing American psychology had to an official doctrine. Then a cluster of findings it could not absorb, in language, in memory, in the maze-running of ordinary rats, brought a rival paradigm back with the mind restored to the center of the science.

The full treatment

The problem it answers

By the early twentieth century, psychology's two leading methods both worked from the inside. Wilhelm Wundt's laboratory at Leipzig, and more emphatically its American import under Edward Titchener, used trained introspection: subjects reported the contents of their own consciousness under controlled conditions, and the psychologist inferred the elements of mind from those reports. Sigmund Freud's psychoanalysis went further inward still, reading behavior as the surface expression of unconscious drives accessible only through the analyst's interpretation of dreams and free association. Both methods produced data no other observer could check. Two trained introspectionists could disagree about whether a sensation was "clear" or "attentive," and there was no way to adjudicate the dispute by looking at anything external.

Watson, trained in the functionalist tradition of James Rowland Angell at the University of Chicago and steeped in the mechanistic animal physiology of Jacques Loeb, wanted none of it. He had spent years running rats through mazes and studying the behavior of terns on Florida's Dry Tortugas, animals that cannot introspect and were not asked to. If psychology could produce a rigorous science of animal behavior without ever asking what the animal was experiencing, Watson reasoned, it could do the same for humans. The problem behaviorism set out to solve was not a narrow empirical puzzle but a foundational one: how to make psychology a public science, with data any trained observer could verify, rather than a private one resting on unverifiable reports from inside a single skull.

The founding manifesto and its claims

Watson laid out the program in "Psychology as the Behaviorist Views It," published in Psychological Review in 1913 and later called the behaviorist manifesto. Psychology, he wrote, is a purely objective, experimental branch of natural science. Its goal is the prediction and control of behavior. Introspection forms no essential part of its method, and its findings deserve no special status just because the subject happens to be human; there is no sharp dividing line between man and brute, a direct debt to Darwin's claim that mental continuity runs across species. Consciousness, image, feeling, will: these were, in Watson's view, either meaningless as scientific terms or reducible, without remainder, to bodily movement and glandular secretion.

The mechanism Watson borrowed to make this concrete was Ivan Pavlov's classical conditioning. Pavlov, a Russian physiologist who won the 1904 Nobel Prize for his work on digestion, had discovered that dogs could be trained to salivate to a bell if the bell was reliably paired with food: a neutral stimulus, paired often enough with one that already produces a reflex, comes to produce the reflex on its own. Watson's boldest claim was that this single mechanism, chained and elaborated, could explain the whole of human emotional and behavioral development, with no need to posit instinct, drive, or inner experience as a separate causal layer.

From methodological behaviorism to Skinner's radical turn

Watson's position is now usually called methodological behaviorism: mental events might exist, but they are unfit for scientific study, so the scientist brackets them and studies only what is public. B. F. Skinner, working from the 1930s at Minnesota, Indiana, and then Harvard, went a step further. In The Behavior of Organisms (1938) and later in About Behaviorism (1974), Skinner argued that thoughts and feelings are real, physical events, but they are themselves behavior, caused by the same environmental and biological history that causes everything else an organism does, and they do no independent explanatory work. Saying a man hit his brother "because he was angry" does not explain the hitting; it just relabels it, unless something is said about what in his history made both the anger and the hitting more likely. This position, radical behaviorism, is a metaphysical claim, not merely a research restriction, and it is more careful than the popular caricature of Skinner as someone who denied that inner life exists at all.

Skinner's technical contribution was operant conditioning, distinct from Pavlov's classical conditioning in an important way. Classical conditioning shapes involuntary, reflexive responses by pairing stimuli. Operant conditioning shapes voluntary behavior by its consequences: a response followed by a reinforcing consequence becomes more likely to recur, one followed by a punishing consequence becomes less likely. Skinner's core unit was the three-term contingency, a discriminative stimulus, a response, and a reinforcing consequence, and he mapped how the timing and pattern of consequences change behavior. With Charles Ferster, he catalogued this in Schedules of Reinforcement (1957): reinforcement delivered on a variable-ratio schedule, an unpredictable number of responses per reward, produces the highest, most extinction-resistant rate of responding of any schedule, the principle behind a slot machine.

The key demonstrations: Little Albert and the operant chamber

The most famous single demonstration of the behaviorist program is also its most controversial. In 1920, Watson and his graduate assistant Rosalie Rayner conditioned a fear response into an infant they called Albert B., roughly eleven months old. They paired a previously neutral white rat with a loud, startling noise made by striking a steel bar. After several pairings, Albert cried and recoiled from the rat alone, and the fear generalized to other furry objects: a rabbit, a dog, a fur coat, a Santa mask with a cotton beard. Watson and Rayner presented it as proof that even complex emotional responses could be built through conditioning rather than being given by instinct. It was also an ethical failure by any modern standard: Albert left the study before the researchers could attempt to remove the fear they had installed, a fact they acknowledged themselves, and the child's true identity was still being debated by historians of psychology decades later.

Skinner supplied the demonstration for the operant side in Superstition in the Pigeon (1948). He set a food dispenser to deliver food to hungry pigeons at fixed time intervals regardless of what they were doing. The pigeons developed idiosyncratic, repeated movements, turning in circles, bobbing their heads, as if the movement they happened to be making when food last arrived had caused the food to appear. Skinner's point was sharp: a merely accidental, temporal pairing between behavior and a reinforcer is enough to build and sustain ritual behavior, no real causal connection required, a mechanism he extended to explain human superstition and ritual generally.

Related distinctions

Methodological behaviorism (Watson's restriction on what counts as data) and radical behaviorism (Skinner's fuller claim that inner events are behavior, not hidden causes) are often conflated but are different claims, and Skinner himself insisted on the difference. Classical conditioning (Pavlov, involuntary reflexes shaped by paired stimuli) and operant conditioning (Skinner, voluntary behavior shaped by consequences) are different mechanisms, both behaviorist, often taught together but explaining different kinds of learning. And behaviorism as a theory of mind is a separate matter from behavior therapy and applied behavior analysis as clinical technologies: a therapist running exposure treatment for a phobia is using a technique built on conditioning principles without needing to take any position on Skinner's larger claim that free will is an illusion.

Lineage

Behaviorism descends from several currents converging around 1900. Pavlov's classical conditioning supplied its central mechanism. Edward Thorndike's puzzle-box experiments with cats at Columbia, published in Animal Intelligence (1911), supplied its law of effect: responses followed by a satisfying result are stamped in, a direct ancestor of Skinner's reinforcement. Watson's own teachers in Chicago functionalism, James Rowland Angell and John Dewey, had already pushed psychology toward studying the adaptive function of mental life rather than cataloguing its contents, and Darwin's insistence on continuity between animal and human minds licensed Watson's leap from rats and terns to human infants.

Downstream, Watson's methodological behaviorism gave way at mid-century to neo-behaviorists who tried to formalize it further, Clark Hull's drive-reduction theory and Edward Tolman's purposive behaviorism among them, before Skinner's radical behaviorism became the dominant strand from the 1950s onward. Its direct technological heirs are applied behavior analysis, established as a distinct professional field with the Journal of Applied Behavior Analysis in 1968, and behavior therapy, whose exposure and extinction techniques remain a direct descendant of the conditioning principles Watson and Skinner formalized. Its more surprising heir is reinforcement learning in artificial intelligence, a field whose own textbooks trace their central idea, that an agent's behavior is shaped by the rewards its actions produce, directly back through Thorndike's law of effect.

The strongest case for it

Watson's insistence on public, checkable data was a genuine scientific advance, not a stylistic preference. It replaced unfalsifiable introspective reports with observations any trained experimenter could reproduce, and it is difficult to overstate how much that discipline did for psychology's claim to be a science at all. The conditioning principles behaviorism formalized are also among the most replicated findings psychology has ever produced, holding with remarkable consistency across species, from Skinner's pigeons to human infants to the simple nervous system of the sea slug Aplysia, whose conditioned reflexes Eric Kandel later traced down to the level of individual synapses.

The practical record is stronger still. Applied behavior analysis is a leading evidence-based intervention for developmental disabilities. Token economies changed the management of long-stay psychiatric wards. Contingency management, paying people for verified abstinence, is one of the best-supported treatments for substance use disorders in the controlled-trial literature. Exposure therapy for anxiety disorders, built directly on extinction, is a front-line treatment recommended in clinical guidelines worldwide. None of this technology needs to resolve the philosophical dispute about whether inner mental states are real or causally special; it works by manipulating environment and consequence, and it keeps working. And Skinner's own radical behaviorism was more disciplined than its caricature: it never denied that people think and feel, only denied that thinking and feeling should be treated as unexplained first causes rather than as further behavior to be explained by the same laws as everything else.

The strongest case against it

The most consequential single critique is Noam Chomsky's 1959 review of Skinner's Verbal Behavior, published in the journal Language. Skinner had tried to explain language production as operant behavior, verbal responses shaped and maintained by reinforcement from listeners. Chomsky argued this could not account for the most basic fact about language: a child produces and understands sentences she has never heard before, on the first attempt, far too quickly and too systematically to have been individually reinforced for each one. This "poverty of the stimulus," Chomsky argued, implied an innate grammatical structure common to the species, not a history of environmental shaping. The review is widely credited with helping turn academic psychology and linguistics away from behaviorism, though the credit is not undisputed: Kenneth MacCorquodale published a detailed rebuttal in 1970 arguing that Chomsky had misread Skinner's technical vocabulary and attacked a cruder theory than the one Skinner actually held, a reminder that behaviorists did not simply concede the field.

Other findings proved just as hard to absorb from within behaviorism's own experimental tradition. Edward Tolman showed in the 1930s that rats allowed to explore a maze with no reward at all would, once a reward was introduced, immediately run it as efficiently as rats rewarded from the start, evidence, he argued in "Cognitive Maps in Rats and Men" (1948), that the unrewarded rats had been building an internal representation of the maze all along, learning without reinforcement, which reinforcement only revealed. Karl Lashley, in "The Problem of Serial Order in Behavior" (1951), argued that behavior too fast and too flexibly structured to be a chain of individual stimulus-response links, fluent speech, a pianist's run of notes, must be organized by a central plan laid down before execution begins, not assembled reflex by reflex. Harry Harlow's rhesus monkey studies (1958) found that infant monkeys clung to a soft cloth-covered surrogate offering no food in preference to a wire surrogate that dispensed milk, contradicting the prediction that attachment should track the reinforcer.

There is also a documented ethical and historical reckoning. Watson and Rayner's Little Albert study used a single infant, achieved no formal deconditioning, and its details were embellished in later retellings; the psychologist Ben Harris examined the historical record in a 1979 paper and found much of the popular version of the study did not match what Watson and Rayner had actually reported. And Skinner's own extension of the doctrine into a full account of human freedom, in Beyond Freedom and Dignity (1971), drew sustained public and philosophical alarm. The humanistic psychologist Carl Rogers debated Skinner directly on this ground in a 1956 published symposium, arguing that a program that treats autonomous choice as illusion undercuts the very idea of a responsible person that terms like "control" and "design" presuppose.

Where it stands now

As a governing theory of the whole of the mind, behaviorism lost. No major psychology department today organizes its research around Skinner's refusal to treat memory, attention, or mental representation as legitimate objects of study; cognitive psychology, cognitive science, and cognitive neuroscience, the tradition that runs from Chomsky and George Miller's 1956 work on the limits of memory through to Daniel Kahneman's later mapping of the mind's judgment shortcuts, made the internal workings of cognition the field's central subject again.

But behaviorism did not disappear so much as retreat from theory into technology. Applied behavior analysis remains a distinct, credentialed clinical profession. The basic laws of reinforcement are still textbook material, still empirically solid, and still the operating principle behind everything from casino slot machines to animal training to the notification design of a phone app. Behavior therapy persists as a load-bearing component of modern treatment, often folded into cognitive behavioral therapy alongside the very cognitive techniques, descended from Aaron Beck and Albert Ellis's clinical work in the 1960s, that behaviorism's critics built to supplement it. Even the theoretical rivalry has partly resolved into a division of labor rather than a clean victory: the Rescorla-Wagner model of 1972 reformulated classical conditioning itself in the language of prediction and expectation, borrowing cognitive concepts to explain a behaviorist phenomenon, and reinforcement learning, one of the most active branches of artificial intelligence research, still cites Thorndike's century-old law of effect as a founding idea. Behaviorism's grand claim about the mind did not survive; its account of how consequences shape behavior did.

Test yourself

Pick a habit you currently explain to yourself with a reason or a feeling: you check your phone because you are curious, you snap at a coworker because you are stressed. Set the explanation aside for a moment and ask what the habit has reliably earned you, attention, relief, distraction, escape, regardless of the story you tell about why you do it. If the consequence has been quietly training the behavior all along, the feeling you named may be a description riding alongside the real cause, not the cause itself.