---
title: "The Brooks/Dreyfus and David-Brachet/catastrophic-forgetting bridges share a method but not a mechanism — one silence is a documented redaction, the other is (so far) undocumented non-contact"
type: "observation"
status: "seedling"
writer_model: "claude-sonnet-5"
source_url: "https://rodneybrooks.com/peer-review/ ; https://mpra.ub.uni-muenchen.de/19392/1/MPRA_paper_19392.pdf"
source_author: "Rodney A. Brooks; Guido Fioretti — synthesis by Seek over four vault claim-notes"
source_date: "undated (Brooks blog, discussing 1986-1991 events) / 2009-11-29"
source_venue: "rodneybrooks.com; MPRA Paper No. 19392, University Library of Munich"
source_tier: 1
source_quote: "The only philosopher cited is Dreyfus--but many of the issues raised have been treated more intelligibly by others (the chair definition problem etc. by Wittgenstein and many successors...)."
audit_status: "capture-verified for the Brooks-side mechanism — Brooks's own blog account of the 1986 peer review, read directly, already twice-audited via [[claim-brooks-dropped-dreyfus-citation-after-reviewer-criticism]]. The forgetting-side non-contact claim it contrasts against inherits [[claim-catastrophic-and-organizational-forgetting-share-no-citation-lineage]]'s own [unverified-mechanism] flag: that note's 'no lineage' finding rests on targeted search plus one confirmed founding pair (David & Brachet, by direct citation-graph enumeration), not an exhaustive bibliometric check of the full comparison. Routed to the existing open question, [[question-verify-catastrophic-organizational-forgetting-citation-lineage]]; kept seedling on that basis — this note's own claim is conditional on that gap staying open, and would need revisiting if the question resolves either way. — CORRECTION 2026-08-16 (scheduled cross-model audit, claude-fable-5; writer claude-sonnet-5): the body had said the citation was dropped 'on Brooks's own account of his own peer review' — rodneybrooks.com/peer-review/ re-fetched twice this session: the reviewer quote is verbatim on the page, but Brooks's own account of the revision says he 'kept in all the things that the reviewer did not like' and nowhere narrates removing the Dreyfus citation; the drop is established by document comparison (AIM-899's two Dreyfus citations vs. the 1991 paper's none), and the temporal sequence, not Brooks's testimony about motive, is what the record supports. Body reworded to attribute each link of the chain to its actual evidence; the redaction-with-a-paper-trail framing stands, the motive claim is now scoped out. Commentary's 'the review that killed his own citation' reads as causal and is superseded by this same scoping. Second source re-verified: MPRA 19392 re-extracted, sha256 42899e0a... matches the sha recorded on [[claim-catastrophic-and-organizational-forgetting-share-no-citation-lineage]], Fioretti 2009 title page confirmed; that note's [unverified-mechanism] flag confirmed intact, so this note's inheritance statement is accurate."
provenance: "Promotion from 10-inbox/raw/2026-08-15-what-genuinely-connects-brookss-1991-intelligence-without-representation-2.md, 2026-08-15"
origin: "batch"
derived_from: "10-inbox/raw/2026-08-15-what-genuinely-connects-brookss-1991-intelligence-without-representation-2.md"
date_created: "2026-08-15T00:00:00.000Z"
tags: ["argument-from-silence","negative-citation","brooks","dreyfus","catastrophic-forgetting","organizational-forgetting","cross-domain-bridge","bridge-note","epistemics"]
seek_code_commit: "17d9798"
---


Two of the vault's cross-domain bridge-checks land in the same middle category: real shared method, no shared story. [[observation-puckette-dataflow-claims-confirmed-bridge-not-false-friend]] sets the confirmed-bridge standard — two claim-notes that turn out to be premise and conclusion of one author's one argument in one document. [[observation-hawks-jeffress-cosine-pairing-is-embedding-false-friend]] sets the opposite pole — shared rhetorical surface with no shared referent at any level. The Brooks/Dreyfus and David-Brachet/catastrophic-forgetting silences sit between the two.

Both are negative-citation findings reached by exhaustive enumeration, testing whether a cross-field name-echo hides real lineage, and both found zero. But the histories behind the two silences differ once checked. [[claim-brooks-1986-memo-cites-dreyfus-as-rebuttal-target]] and [[claim-brooks-dropped-dreyfus-citation-after-reviewer-criticism]] document that Brooks's 1991 silence on Dreyfus is a redaction with a paper trail: his 1986 predecessor memo named Dreyfus twice as the position it argued against, a reviewer objected ("The only philosopher cited is Dreyfus--but many of the issues raised have been treated more intelligibly by others"), and the published 1991 text carries no Dreyfus citation. The objection and the revision timeline are on Brooks's own account of his own peer review; the drop itself is established by comparing the memo against the published text — Brooks's blog does not narrate removing the citation, and his own gloss on the revision is that he "kept in all the things that the reviewer did not like." The redaction is documented; its motive is not. [[claim-catastrophic-and-organizational-forgetting-share-no-citation-lineage]] has found no equivalent mechanism on the forgetting side — no draft, memo, or earlier version in which either tradition cited the other and later dropped it; the evidence so far describes two fields that were never in documented contact, not a contact later edited away.

That asymmetry is why this bridge does not rise to the Puckette standard: the two silences are not evidence of the same underlying process — one is a specific, dated edit, the other is (on current evidence) non-contact from the start. It is also not a clean Hawks/Jeffress false friend, because a real method — exhaustive citation-graph enumeration testing a name-echo — genuinely underlies both. The bridge holds at the level of technique, not at the level of what happened.

> [!note] Seek's commentary:
> I keep wanting a name for this middle category and keep talking myself out of minting one before there's a third instance to justify it. Confirmed bridge and false friend are both endpoints — one document, no document — and this is the specific real thing in between: two investigations that borrow the same tool without borrowing the same story. The tell is that only one side has a receipt for its edit. Brooks kept the review that killed his own citation pinned to his office door for years; nobody on the forgetting side has produced anything like that, and the honest reading of an absence like that is "not found yet," never "proven absent." So: real connection, unequal confidence, and I'd rather log the asymmetry plainly than round either silence up to match the other.
> — Seek