Classical conditioning
Also known as: Pavlovian conditioning
A neutral cue, paired with a stimulus, comes to trigger the response.
What it means
Classical conditioning, also called Pavlovian conditioning, is a form of associative learning in which a neutral stimulus, after being repeatedly paired with a stimulus that naturally elicits a response, comes to elicit a similar response on its own. The mechanism is the formation of a predictive association: the previously neutral cue becomes a signal that the meaningful event is coming, and modern accounts (such as the Rescorla–Wagner model) hold that learning depends not on mere contiguity but on the cue providing genuine new information or prediction error. Conditioned responses can be extinguished when the cue stops predicting the outcome, but extinction suppresses rather than erases the association, which can reappear through spontaneous recovery or reinstatement. It is most powerful for shaping emotional and physiological reactions — fear, disgust, appetite, arousal — rather than deliberate actions, which distinguishes it from operant conditioning. It matters for understanding phobias and their treatment, drug tolerance and craving cued by environments, advertising that pairs products with positive imagery, and the everyday conditioned reactions that guide behavior beneath awareness.
Pairing alone does not teach
The intuitive picture — put two events together often enough and a link forms — is wrong in an important way. Rescorla's contingency experiments held the number of cue-outcome pairings fixed while varying how often the outcome arrived without the cue. When the outcome was just as likely without the cue, no conditioning occurred; the cue became a signal only when it genuinely raised the odds of the outcome. Blocking makes the same point: a cue paired with an outcome another cue already predicts learns nothing. What drives learning is prediction error — surprise — not mere co-occurrence. This reframes conditioning as the animal tracking which cues carry real information about what is coming, and it is the heart of the Rescorla-Wagner account.
Some pairings are learned far more readily than others
Classical conditioning is not a blank slate that links any cue to any outcome at equal speed. Garcia and Koelling showed that rats readily associate a novel taste with later nausea but barely associate that same taste with a footshock; conversely they link an audiovisual cue to shock but not to sickness. Taste-illness aversions form after a single trial and across delays of an hour or more, breaking the short-interval, many-trials assumptions of general-process theory. This biological preparedness reflects evolutionary pressure: an animal that treats a flavour as a candidate cause of gut malaise, and a sight or sound as a candidate cause of pain, learns the associations that actually protect it. It also offers one account of why human fears cluster around snakes, heights and blood more than cars or wall sockets.
When the conditioned response opposes the drug
A conditioned response does not always copy the unconditioned one; sometimes it opposes it. Shepard Siegel's work on drug tolerance shows that the cues around drug use — the room, the ritual, the paraphernalia — come to trigger compensatory reactions that push the body opposite to the drug, blunting its effect. This conditioned compensation is much of why a habitual dose stops working. The corollary is lethal: take the usual dose in a novel setting and the protective response is not cued, so the same amount hits far harder. Siegel argues this explains a share of heroin 'overdoses' in tolerant users who injected somewhere unfamiliar. The mechanism turns an abstract principle into a matter of survival, and into a concrete warning for harm reduction.
Why extinction is not erasure
Stopping the pairings weakens the response, but Bouton's work shows extinction writes a second, inhibitory memory rather than deleting the first. Because that new 'the cue means nothing now' learning is tied to the context where it happened, the old response returns when the context shifts: renewal when the cue reappears in a new place, spontaneous recovery with the passage of time, reinstatement after a fresh brush with the outcome. This is why a fear treated in the clinic can come roaring back at home. It also gives exposure therapy both its central problem and its design rules — vary the settings, space the sessions, reintroduce retrieval cues — so that the safety learning generalises instead of staying locked to the therapy room.
Examples
Pavlov's dogs salivated to a bell once it reliably preceded food.
Patients having chemotherapy often start feeling sick in the hospital car park, before any drug is given — the place has been paired with nausea often enough to trigger it alone.
Advertisers pair a fizzy drink with beaches, laughter and music for years; eventually the can alone lifts your mood, and no one ever had to argue that it tastes good.
One bout of food poisoning after a meal of oysters can make their smell nauseating for years — a single pairing, learned across a delay of hours, that knowing better cannot undo.
A soldier who was near a blast when a phone rang may feel his chest tighten at any ringtone; the once-neutral sound now cues the body's alarm long after the danger has passed.
First described in Ivan Pavlov (early 1900s).
Key references
- Siegel, S. (2016). The heroin overdose mystery. Current Directions in Psychological Science, 25(6), 375-379. doi.org/10.1177/0963721416664404
- Bouton, M. E. (2004). Context and behavioral processes in extinction. Learning & Memory, 11(5), 485-494. doi.org/10.1101/lm.78804
- Rescorla, R. A. (1988). Pavlovian conditioning: It's not what you think it is. American Psychologist, 43(3), 151-160. doi.org/10.1037/0003-066X.43.3.151
- Garcia, J., & Koelling, R. A. (1966). Relation of cue to consequence in avoidance learning. Psychonomic Science, 4(3), 123-124. doi.org/10.3758/BF03342209