Behavioural Science & Sociocultural Psychiatry · Paper A
Classical conditioning
Classical conditioning is cue-learninga neutral signal becomes meaningful because it reliably arrives with something that already produces an automatic response. For example, a bell that repeatedly precedes food eventually makes a dog salivate before the food appears. In practice, the exam trick is to ask whether the response is elicited by a cue, rather than chosen because of a consequence.
Start with an unconditioned stimulus, such as food, naturally producing an unconditioned response, such as salivation. Pair a neutral stimulus, such as a bell, with that food often enough, and the bell becomes the conditioned stimulus; salivation to the bell alone becomes the conditioned response. Acquisition strengthens the association. Extinction happens when the conditioned stimulus is repeatedly presented without the unconditioned stimulus, so the response falls. Modern learning accounts treat extinction as new inhibitory learning, not simple erasure, which explains why spontaneous recovery can appear after a rest period. Generalisation spreads the response to similar cues; discrimination narrows it to the exact predictive cue.
Everyone remembers Pavlov's dog salivating to a bell, but the exam rarely tests the story itself: it tests the specific named phenomena that describe what happens as the learned link forms, weakens and sometimes reappears. Getting those names precisely right, and telling them apart from each other, is the actual skill this topic builds.
Classical, or Pavlovian, conditioning is a form of learning in which a neutral event comes to trigger an automatic response by being repeatedly paired with something that already triggers that response naturally. An unconditioned stimulus (US, something that naturally triggers a response without any prior learning), such as food, naturally produces an unconditioned response (UR, the automatic reaction itself), such as salivation. Pairing a neutral cue, such as a bell, with the US repeatedly turns the bell into a conditioned stimulus (CS, a previously neutral cue that now triggers the response through learning), and salivation to the bell alone becomes the conditioned response (CR, the learned version of that same reaction).
Most candidates assume that once a learned response fades, the learning itself has been erased. It has not. Extinction, the fading of a conditioned response after repeated presentation of the CS without the US, is NEW inhibitory learning laid down on top of the original memory, not deletion of that memory. This is exactly why the response can reappear later, and that reappearance is itself a specifically named phenomenon.
Imagine a bell that has always preceded feeding for a dog. Before any pairing, the bell means nothing and produces no salivation; food alone produces salivation automatically. After many pairings of bell-then-food, the dog starts salivating to the bell alone, even before any food appears: the strengthening of this association across repeated pairings is called acquisition. If the bell then rings many times with no food following, salivation to the bell gradually fades: that is extinction. But because extinction only suppresses the response rather than erasing the original bell-food memory, a rest period allows the faded response to partially reappear when the bell is next presented; this is spontaneous recovery, direct proof that extinction did not erase the original learning.
Acquisition: the initial strengthening of the CS-CR link across repeated CS-US pairings.
Extinction: repeated presentation of the CS without the US, causing the learned response to reduce; this is new inhibitory learning, not erasure of the original association.
Spontaneous recovery: the partial reappearance of an extinguished conditioned response after a rest period following extinction, without any further pairing.
Generalisation: the tendency for the conditioned response to spread to stimuli that resemble the original CS. For example, a dog conditioned to salivate to one specific tone also salivates, somewhat, to a similar-pitched tone it has never heard before.
Discrimination: the opposite process, in which the response becomes restricted to only the specific CS that reliably predicts the US, and does not occur to merely similar stimuli.
Higher-order conditioning: a second neutral stimulus paired with an already-established CS can itself become a new, weaker conditioned stimulus, even though it was never directly paired with the original US.
Classical conditioning is distinguished from operant conditioning by what drives the response: classical conditioning pairs two stimuli to produce an involuntary, elicited response, while operant conditioning uses the CONSEQUENCES of a voluntary action to strengthen or weaken that action. Habituation, a simple reduction in response to one repeated stimulus with no pairing involved at all, is distinguished from classical extinction, which specifically requires a previously learned CS-US pairing to have existed first. Operant extinction, in which a reinforcement is withdrawn, is distinguished from classical extinction, in which the US is simply omitted after the CS.
In practice, classical conditioning underlies conditioned fear responses, such as a fear reaction to a hospital setting after a painful medical procedure there, which can fade with repeated non-painful visits yet reappear after a long gap without any new bad experience: exactly the pattern predicted by extinction being inhibitory learning, not erasure. It also underlies some clinical exposure-based treatments, which deliberately use repeated CS presentation without the feared US to build extinction learning, while accepting that spontaneous recovery is expected and is not evidence that the treatment has failed.
Classical conditioning itself is a learning-theory mechanism, not a diagnosable illness, so it has no diagnostic criteria of its own: you do not "diagnose" classical conditioning any more than you diagnose gravity. But when conditioning produces a clinical problem, most commonly a specific phobia (an excessive, persistent fear of a specific object or situation, out of proportion to any real danger) or a conditioned anxiety response, that downstream problem does have a real diagnostic framework: ICD-11 criteria for phobic anxiety disorders require a fear that is consistently triggered by the specific object or situation (the conditioned stimulus), is out of proportion to the actual danger, is persistent, and leads the person to avoid the trigger or endure it with intense distress. Recognising the conditioning history behind the presentation (a clear CS that reliably predicts the feared reaction, often traceable to an original pairing with a genuinely frightening event) is what distinguishes a classically-conditioned phobia from anxiety with a vaguer or more generalised trigger.
Because classical conditioning is a mechanism rather than an illness, "treating" it usually means deliberately re-running the learning process in reverse. Graded exposure therapy is the first-line management approach recommended by NICE for phobias and related anxiety disorders: the patient is exposed to the conditioned stimulus in a controlled, gradually escalating way, with the feared unconditioned outcome absent, which deliberately builds new extinction learning on top of the old association rather than trying to erase it. This is why relapse after exposure therapy is expected to look like spontaneous recovery rather than treatment failure: a patient can do very well in the clinic, then show some return of fear after a long gap, and that reappearance does not mean the therapy did not work, it means the extinction learning is doing exactly what extinction learning always does. Systematic desensitisation, an older but related technique pairing gradual exposure with relaxation, rests on the same acquisition and extinction logic built up earlier on this page.
Whenever a vignette describes a learned response that faded and then reappeared after a break with no new pairing, that is spontaneous recovery, not evidence that the original extinction training failed or was reversed.
Extinction weakens a conditioned response through new inhibitory learning rather than erasing it, which is exactly why spontaneous recovery can occur later.