Pavlovian Conditioning
A neutral cue paired with a meaningful event can come to trigger the response on its own.
Author
Ivan Pavlov (1902–1927); extended by Rescorla & Wagner, Kamin
Model type
About
Pavlovian (classical) conditioning explains how associations form: a neutral stimulus becomes a conditioned stimulus (CS) by being paired with an unconditioned stimulus (US) that naturally elicits an unconditioned response (UR). After enough pairings, the CS alone evokes a conditioned response (CR). Learning depends on contiguity (timing) and, crucially, contingency/prediction error—the CS must carry information that the US is coming.
How it works
Core terms
- US → UR: food → salivation; shock → startle
- CS (neutral at first) + US repeatedly → CR to the CS alone
Acquisition – CS precedes US by a short interval (best learning with forward, short-delay pairing).
Extinction – CS without US reduces CR; spontaneous recovery can occur after a pause.
Generalisation / discrimination – responses spread to similar cues; training can narrow them.
Blocking & overshadowing – a well-established CS can block learning about a new cue presented alongside it; more salient cues overshadow weaker ones.
Prediction error – learning tracks “surprise” (Rescorla–Wagner): big updates when outcomes differ from expectation.
Second-order conditioning – a new neutral cue paired with an established CS can also elicit the CR.
Use-cases
Habits & behaviour change – anchor routines to stable cues; pair desired actions with immediate rewards.
Marketing/brand – sonic logos, packaging and contexts that evoke positive affect or appetite.
UX/product – notification sounds and micro-rewards that condition checking or contribution.
Training & safety – alarms that trigger rehearsed responses; animal training basics.
Therapy/health – exposure and counterconditioning for phobias; managing anticipatory nausea in chemo (conditioned responses).
Operations – consistent cues for readiness states (shift bells, kanban signals).
Step-by-step
Define the map – US (meaningful outcome) → UR (target response). Choose a CS you can present reliably.
Pair with timing discipline – use short, forward pairings (CS just before US); keep early sessions dense and consistent.
Control context – train in the real performance environment if you want the response there; vary later to improve transfer.
Measure acquisition – track CR strength (latency, magnitude, frequency) per session.
Stabilise or fade – to maintain, occasionally reinforce (partial schedules). To remove, run extinction (CS without US) or countercondition (CS + relaxing/positive US).
Guard against generalisation – if only specific cues should trigger the response, train discrimination with contrasts.
Mind blocking/overshadowing – avoid piling cues together; teach one salient CS at a time.
Pitfalls & Cautions
Inconsistent pairing → weak learning or extinction.
Wrong timing – backward or long-delay pairings rarely work.
Unintended conditioning – nausea, fear or device anxiety can be conditioned by accident; watch contexts.
Ethics – avoid manipulative or aversive conditioning with people; get informed consent and offer opt-outs.
Context specificity – CR may fail to transfer across settings without deliberate variation.
Relapse – spontaneous recovery and renewal; plan boosters or extinction in multiple contexts.
Related Mental Models
Click below to learn other mental models
The two Arrows
Focus where importance intersects with control. Act directly on what matters and you can change; influence or ignore the rest.
Principal-Agent Problem
When decision rights are delegated, agents optimise their own payoff under information asymmetry. Without smart contracts and governance, effort, risk and horizon drift away from the principal’s goals.
State → Story → Strategy
Align your state, story and strategy; order matters.