Perruchet Effect¶
A conditioning dissociation in which recent trial runs drive conscious expectancy and conditioned responding in opposite directions.
Core Idea¶
The Perruchet effect is a sequential-conditioning dissociation in which conscious expectancy for an event and conditioned-response strength or speed change in opposite directions across recent trial runs. The joint pattern, not either measure alone, is used to test automatic associative-strength and propositional expectancy accounts. In the original eyeblink paradigm, a tone was sometimes presented alone and sometimes paired with an air puff under randomized equal probabilities.
Scope of Application¶
The effect is studied in paradigms that jointly measure sequential expectancy and conditioned performance. The effect applies only to paradigms that manipulate trial history and independently measure expectation and conditioned performance.
- Eyeblink conditioning. Tone–air-puff sequences compare blink responses with explicit expectation.
- Electrodermal conditioning. Skin-conductance responses test whether awareness is necessary.
- Reaction-time tasks. Go/no-go and simple-response speed can oppose reported expectancy.
- Learning theory. The dissociation tests automatic-association and propositional accounts.
- Dual-process evidence. Opposing measures motivate, but do not settle, multiple-process explanations.
Clarity¶
Report trial probabilities, sequence generation, run definition, conditioned and unconditioned stimuli, expectancy scale, behavioral measure, and both trend functions. Distinguish the observed dissociation from the theory used to explain it. Test whether response recency, sensitization, or demand characteristics reproduce the pattern. The closest near miss sets the boundary: A simple expectancy–response correlation is the closest near miss; the effect requires a systematic dissociation, commonly opposite trends, not merely imperfect correspondence.
Manages Complexity¶
The paradigm turns a broad implicit-versus-explicit debate into two simultaneously observed curves over trial history. It isolates where expectation and performance separate while retaining enough design detail to compare associative, propositional, and non-associative accounts. The central association strength–propositional expectancy tradeoff is this: The paradigm is designed so recent pairings and conscious sequence inference predict opposite behavior. A second dissociation evidence–process inference tension matters because Two curves can separate without proving two independent cognitive systems.
Abstract Reasoning¶
Use three linked moves: construct randomized trial runs with known contingencies and define run-length categories in advance; measure explicit next-event expectancy and conditioned performance independently on the same sequence; estimate each measure as a function of recent run type and length. As a collapse test, the case exits when the two measures do not dissociate or when artifacts such as sensitization fully explain the response trend. A fourth check is to test for opposite trends and interaction rather than separate significance alone.
Knowledge Transfer¶
The effect transfers from eyeblink to electrodermal or reaction-time settings only when the same sequential dissociation is operationalized. Everyday conflicts between expectation and behavior are analogies, not literal Perruchet effects. Dissociation carries the broader structure while the named phenomenon remains experimental and conditioning-specific. No canonical parent prime is currently asserted; broader structural comparisons remain related-prime analogies until separately adjudicated in the DAG. Opposing measures motivate separation of underlying mechanisms.
Neighborhood in Abstraction Space¶
Perruchet Effect sits in a moderately populated region (43rd percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.
Family — Named Cognitive & Behavioral Effects (32 abstractions)
Nearest neighbors
- Crespi Effect — 0.88
- Stimulus Control — 0.87
- Motivated forgetting — 0.87
- Procedural memory — 0.87
- Extinction (psychology) — 0.87
Computed from structural-signature embeddings · 2026-10-08