Neisser checked John Dean's Watergate testimony against the Nixon tapes and found the details wrong but the gist right — coining 'repisodic' memory

In "John Dean's Memory: A Case Study" (Cognition 9(1):1-22, 1981), the cognitive psychologist Ulric Neisser exploited a rare natural experiment: John Dean testified before the Senate Watergate Committee from memory about conversations that — unknown to Dean at the time — had been secretly tape-recorded in the Nixon White House. Neisser could therefore compare sworn recollection against a verbatim transcript.

His reported finding was two-layered. Many specific conversational details in Dean's testimony were wrong, and the errors were not random: Dean tended to inflate his own centrality, remembering himself as more present and pivotal than the tapes showed. Yet the gist held — that a cover-up existed, and roughly who was involved. Neisser coined "repisodic" memory for this pattern: a recollection that is not a faithful record of a single episode but a compressed, functionally accurate blend of a repeated series of similar events.

The study was a deliberate instrument of Neisser's broader program. He was an outspoken critic of laboratory memory research's ecological validity, arguing psychologists had "hardly ever studied" the memory phenomena that matter in real life; the Dean case was a real-world test chosen precisely to answer that complaint rather than a lab word-list. This places the note in the reconstructive-memory lineage running back to Bartlett (question-bartlett-distortion-typology-maps-to-model-collapse-failure-modes) and forward into how AI systems model recall.

The "wrong in the details, right in the shape" structure matches what the oral historian Alessandro Portelli reported the same year from testimony with no tape to check against (claim-portelli-oral-testimony-relocated-trastulli-death-to-1953); the independent same-year convergence is treated in observation-1981-convergence-testimony-wrong-in-detail-right-in-shape. It also bears on the older question of how much weight a lone piece of testimony deserves (claim-roman-canon-law-rated-one-witness-equal-to-a-private-document).

The specific findings and the "repisodic" coinage rest on secondary paraphrase — the paper's full text was paywalled at capture time. Verification against the primary is routed to question-verify-neisser-john-dean-cognition-1981-primary.