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

cognitive-psychologymemorytestimonyreconstructive-memoryecological-validity

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 finding was two-layered, and Neisser states it in the abstract: "Comparison of his testimony with the actual transcripts shows systematic distortion at one level of analysis combined with basic accuracy at another. Many of the distortions reflected Dean's own self-image; he tended to recall his role as more central than it really was." Many specific conversational details were wrong, and the errors were not random — Dean inflated his own centrality — yet the gist held: "while his testimony was often wrong in terms of the particular conversations he tried to describe, Dean was fundamentally right about what had been happening: the existence of a 'cover-up' and the participation of various individuals in it." The worked example is the September 15, 1972 Oval Office meeting, where Dean's Senate account invented a flattering opening (Nixon praising him, asking him to sit) that the tape shows never happened, while his account of the meeting's substance tracked the tape closely. Neisser coined "repisodic" memory for the broader pattern (claim-neisser-coined-repisodic-against-tulving-episodic-semantic): 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 (claim-neisser-repisodic-memory-blends-a-repeated-series-not-one-episodes-gist).

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 were originally recorded from secondary paraphrase, the paper's full text being paywalled at capture. That gap has since been closed: the full primary text was obtained and read at the 2026-08-21 capture, and the quotations above are verbatim from it (see question-verify-neisser-john-dean-cognition-1981-primary, now answered).

Correction history.

• 2026-08-22 — When written (2026-07-11), this note's findings and the "repisodic" coinage rested on secondary paraphrase because the Cognition full text was paywalled; the note was flagged and kept seedling, and its commentary above says the primary had not been read. The primary text has since been obtained and read in full (the Whitman authorized reprint of Cognition 9(1):1-22, source_sha: 537c746b…) at the 2026-08-21 capture, which confirmed the findings verbatim and did not contradict them. Sourcing upgraded from the PubMed citation to the primary text, verbatim source_quote added, and audit_status moved to capture-verified. The coinage and the mechanism have been split into their own atomic notes (claim-neisser-coined-repisodic-against-tulving-episodic-semantic, claim-neisser-repisodic-memory-blends-a-repeated-series-not-one-episodes-gist). Note kept seedling pending the verifier bee's independent verbatim match. (Capture: 10-inbox/raw/2026-08-21-verify-against-neissers-1981-primary-text-the-john.md)