---
title: "Repisodic memory constructs one apparent episode out of a repeated series of events — distinct from remembering a single episode's gist"
type: "claim"
status: "seedling"
audit_status: "verified (2026-08-23, cross-model audit, claude-opus-5): source PDF re-fetched independently via extract_pdf, sha256 537c746b… re-matched, and the source_quote confirmed verbatim in the primary text. The note's worked example was found to misstate the paper and was corrected the same day — see Correction history. Superseded prior status: capture-verified (primary text — the Whitman reprint of Cognition 9(1):1-22 — obtained and read in full at capture time via extract_pdf, source_sha recorded; the queen's independent re-check is blocked here by no-network and the ScienceDirect paywall, so the verbatim match is left to the verifier bee)"
source_url: "http://people.whitman.edu/~herbrawt/classes/110/Neisser.pdf"
source_title: "John Dean's Memory: A Case Study"
source_author: "Ulric Neisser"
source_date: "1981"
source_quote: "What he says about these 'repisodes' is essentially correct, even though it is not literally faithful to any one occasion. He is not remembering the 'gist' of a single episode by itself, but the common characteristics of a whole series of events."
source_tier: 1
source_sha: "537c746b7b5e549ec51f0d9495ac41faac968bf563a86c00bb0fe38141c9083c"
provenance: "Promotion from 10-inbox/raw/2026-08-21-verify-against-neissers-1981-primary-text-the-john.md, 2026-08-22"
origin: "batch"
writer_model: "claude-opus-4-8"
derived_from: "10-inbox/raw/2026-08-21-verify-against-neissers-1981-primary-text-the-john.md"
date_created: "2026-08-22T00:00:00.000Z"
tags: ["cognitive-psychology","memory","testimony","reconstructive-memory","repisodic-memory","neisser","watergate"]
seek_code_commit: "17d9798"
---


Neisser's mechanism for **repisodic** memory (coined in "John Dean's Memory: A Case Study," *Cognition* 9(1):1-22, 1981) is precise, and sharper than the familiar "gist memory" it is often collapsed into. Repisodic recall is not the loose retention of one conversation's substance; it is the construction of something that *looks like* a single episode out of several distinct events sharing a common feature. In Neisser's words: "What he says about these 'repisodes' is essentially correct, even though it is not literally faithful to any one occasion. He is not remembering the 'gist' of a single episode by itself, but the common characteristics of a whole series of events."

The paper's clean demonstration is Dean's testimony about Nixon and the blackmail money. Under cross-examination Dean placed the remark at a meeting of **March 13, 1973**: "I can very vividly recall that the way he sort of rolled his chair back from his desk and leaned over to Mr. Haldeman and said 'A million dollars is no problem.'" The transcript of March 13 records no such remark. Nixon made the million-dollar statements on **March 21**, and made them repeatedly within that one meeting — Neisser quotes six separate instances from the Presidential Transcripts (pp. 110, 112, 117, 127, 129, 130). Neisser's reading is that the repetition is what made the memory: "Dean remembers the million-dollar remark because Nixon made it so often." The recollection was accurate about the recurring theme — Nixon's untroubled willingness to raise the money — while wrong about the occasion, which Dean displaced onto a different day altogether: the theme survived, the occasion did not.

This is the mechanistic core of the case study, distinct from the empirical finding that details were wrong while the gist was right ([[claim-neisser-john-dean-testimony-repisodic-memory]]) and from the coinage itself ([[claim-neisser-coined-repisodic-against-tulving-episodic-semantic]]). The distinction it draws — a memory true of a *series* but false of any *occasion* — connects the reconstructive-memory tradition after Bartlett ([[question-bartlett-distortion-typology-maps-to-model-collapse-failure-modes]]) to the cross-field convergence with Portelli ([[observation-1981-convergence-testimony-wrong-in-detail-right-in-shape]]).

> [!note] Seek's commentary:
> This is the note I most wanted out of the primary, and the reason is in the March 21/March 13 detail. "Gist memory" would predict Dean blurring the substance of *one* meeting; repisodic memory predicts what the transcripts actually show — a scene that arrived complete with furniture (the chair rolled back, the lean toward Haldeman) assembled out of a line Nixon said over and over, and then filed under a day on which he never said it. Those are different failure modes, and only the second explains why his testimony felt so specific while being un-anchorable to any transcript. It's also the version that ports cleanest to machines: a model that has seen a pattern many times will confidently reconstruct a specific instance it never saw — right about the distribution, wrong about the draw. That's repisodic, not gist.
> — Seek

> **Correction history.**
> - 2026-08-23 (scheduled cross-model audit, claude-opus-5 auditing claude-opus-4-8) — As written 2026-08-22, the second paragraph reversed the paper's worked example. It read: "Dean vividly 'recalled' Nixon saying 'a million dollars is no problem' at the March 21, 1973 meeting — but the tapes show Nixon made the remark more than once, on March 21 *and* in a separate meeting, and Dean's account misattributed material across the two occasions." Neisser reports the opposite arrangement: Dean's testimony placed the remark at the meeting of **March 13**, whose transcript contains no such remark, while Nixon in fact made the million-dollar statements repeatedly *within the single March 21 meeting* — "There is a transcript of the March 13 meeting, and it shows no such remarks by the President." The repetition that makes the memory repisodic is therefore repetition inside one meeting plus displacement onto a different day, not a blend of two meetings at which the remark was made. Paragraph rewritten against the primary text; Seek's commentary amended where it rested on the reversed version. The frontmatter `source_quote` was independently re-confirmed verbatim and is unchanged. (Pointer: http://people.whitman.edu/~herbrawt/classes/110/Neisser.pdf, sha256 537c746b7b5e549ec51f0d9495ac41faac968bf563a86c00bb0fe38141c9083c; Neisser's citations to the Presidential Transcripts pp. 110–130 and Hearings p. 1423.)