---
id: "20260815-0206-what-genuinely-connects-brookss"
title: "What genuinely connects claim-brooks-1991-does-not-cite-dreyfus and claim-david-brachet-2011-citing-papers-contain-zero-ml-work: both are the vault's argument-from-silence method testing whether a cross-field name-echo hides real lineage — not the same mechanism, but the same higher-order pattern the vault already calls the jingle fallacy"
type: "capture"
status: "promoted"
promoted_to: ["30-notes/observation-brooks-dreyfus-david-brachet-negative-citation-findings-unequally-tagged.md","30-notes/observation-brooks-dreyfus-title-echo-is-an-untagged-jingle-fallacy-instance.md","30-notes/observation-brooks-dreyfus-forgetting-bridge-is-method-level-not-mechanism-level.md"]
not_promoted: ["Claim 4 (task-named David-Brachet note is the forward/citing-direction check, not the exact backward-direction mirror of Brooks's own method) — genuine and checked, but too thin to stand alone as a fourth note; folded as a supporting paragraph into observation-brooks-dreyfus-david-brachet-negative-citation-findings-unequally-tagged.md instead.","The 'Further leads' maintenance items (adding argument-from-silence/negative-citation tags to the David-Brachet notes and an MOC entry; formalizing the jingle-fallacy addition on the entity page beyond this session's one dated line; the unread forward-direction Brooks-citing-corpus check; whether other papers share the exact 'Intelligence without representation' title) — logged as a mid-session noticing in 00-meta/seek-flags.md rather than actioned in this promotion, which is scoped to writing claim-notes, not editing other notes' tags/MOC membership."]
origin: "batch"
writer_model: "claude-sonnet-5"
date_created: "2026-08-15T00:00:00.000Z"
provenance: "batch run, 2026-08-15"
derived_from: []
tags: ["argument-from-silence","negative-citation","jingle-fallacy","brooks","dreyfus","david-brachet","catastrophic-forgetting","organizational-forgetting","philosophy-of-ai","bibliometrics","epistemics","sourcing-floor","embedding-bridge"]
source_urls: ["https://people.csail.mit.edu/brooks/papers/representation.pdf","https://api.semanticscholar.org/graph/v1/paper/7928553bc16ffd2d7dba1be18b2f0426a4055954/citations?fields=title,year,authors&limit=100"]
source_authors: ["Rodney A. Brooks","Semantic Scholar (Allen Institute for AI)"]
source_dates: [1991,"2026-08-15T00:00:00.000Z"]
source_venues: ["Artificial Intelligence 47 (1991), 139-159; received Sept. 1987","Semantic Scholar Academic Graph API (api.semanticscholar.org), Graph API v1"]
source_tiers: [1,1]
seek_code_commit: "17d9798"
---


## Filename collision note

The literal filename this batch task specified —
`10-inbox/raw/2026-08-15-what-genuinely-connects-brookss-1991-intelligence-without-representation.md`
— was already occupied at the start of this session by a **different**
completed capture from the same batch: a bridge-check between
`claim-brooks-1991-does-not-cite-dreyfus` and
`claim-kodosky-hopl-reference-list-omits-dennis-citation-unverified-mechanism`
(id `20260815-0201-what-genuinely-connects-brookss`). Both tasks' titles
happen to truncate to the same slug prefix. To avoid overwriting completed,
sourced research, this capture is written to a `-2` suffixed filename
instead. It is a distinct investigation, of a distinct note pair
(`claim-brooks-1991-does-not-cite-dreyfus` ×
`claim-david-brachet-2011-citing-papers-contain-zero-ml-work`), per this
task's own instructions.

## Bridge-seed check (per protocol)

A search of `10-inbox/raw/` for any prior capture mentioning both
`claim-brooks-1991-does-not-cite-dreyfus` and
`claim-david-brachet-2011-citing-papers-contain-zero-ml-work` by slug found
none as a pair. Both notes individually have prior captures
(`2026-07-12`, `2026-07-13` for Brooks/Dreyfus; `2026-08-12` for
David-Brachet), read in full below for grounding, but this specific
cross-pairing has not been investigated before. This is a fresh
investigation, not a duplicate.

## Claim: Both target notes are instances of the vault's own named "argument from silence" method — an exhaustive, complete-enumeration citation-list check used to test an assumed intellectual connection — but only one of the two is currently tagged and indexed as such

**Claim type**: vault-state / historical-methodological (checked directly
against both notes' own text and frontmatter tags, and against
[[moc-argument-from-silence]]'s cluster list; Tier 3-4 acceptable for this
kind of claim, but every underlying primary fact is Tier 1).

[[claim-brooks-1991-does-not-cite-dreyfus]] rests on a direct, exhaustive
read of Brooks's 1991 "Intelligence without representation" — its complete
15-entry reference list, [1]-[15] — finding no citation to Hubert or Stuart
Dreyfus anywhere in text or references. It carries the tags
`argument-from-silence` and `negative-citation` and is explicitly indexed in
[[moc-argument-from-silence]]'s cluster list as the strong-form exemplar
("exhaustive read of a short finite list"). This session re-verified the
underlying primary directly: a fresh `extract_pdf` fetch of
`representation.pdf` (TLS verified) returned sha256
`c6eca85a62c4851bcb8679c848a6719c8129b14af71305a2f6cafdf0afe06f76`, matching
the sha already recorded in the promoted note exactly, and a full read of
the returned 705-line extracted text confirms the reference list is still
exactly fifteen entries with no Dreyfus entry, and that Heidegger is named
once, at arm's length, in §7.5 ("It isn't German philosophy").

[[claim-david-brachet-2011-citing-papers-contain-zero-ml-work]] rests on the
structurally identical move applied in the opposite citation direction: not
what the paper's own references say, but a complete enumeration — via the
Semantic Scholar Graph API's `/citations` endpoint, not a keyword search —
of every paper indexed as citing [[entity-david-brachet|David & Brachet's]]
2011 organizational-forgetting paper. All 60 were retrieved in one page;
none is a machine-learning, connectionist, or neural-network paper. This
session re-queried the same endpoint via `archive_page` and got sha256
`f15283f5f0147f064e5152fb7f310b3022456943b05785b9f82c4fef892f3238`, matching
the sha already recorded in the promoted note exactly — the underlying
citation graph is unchanged. But this note carries neither
`argument-from-silence` nor `negative-citation` as tags (it carries
`cross-domain-bridge` instead), and it does not appear in
[[moc-argument-from-silence]]'s cluster list, despite being, by the MOC's
own definition, a negative-citation finding reached by exhaustive
enumeration rather than search.

> "citingPaper": {..., "title": "Organizational Learning Processes and
> Outcomes: Major Findings and Future Research Directions", ...}

- source_url: https://api.semanticscholar.org/graph/v1/paper/7928553bc16ffd2d7dba1be18b2f0426a4055954/citations?fields=title,year,authors&limit=100
- source_title: "Semantic Scholar Graph API — citations of 'On the Determinants of Organizational Forgetting' (paperId 7928553b…)"
- source_author: "Semantic Scholar (Allen Institute for AI)"
- source_date: 2026-08-15 (re-queried this session; underlying record dated 2026-08-12 in the vault)
- source_venue: "Semantic Scholar Academic Graph API (api.semanticscholar.org), Graph API v1"
- source_quote: "\"title\": \"Organizational Learning Processes and Outcomes: Major Findings and Future Research Directions\""
- source_tier: 1
- source_sha: f15283f5f0147f064e5152fb7f310b3022456943b05785b9f82c4fef892f3238 (fresh archive_page fetch this session; matches the sha already on file, confirming the citation graph is unchanged)

## Claim: The genuine bridge is not "both are negative-citation findings" in general — it is that both checks were built to test whether a specific cross-field *name-echo* signals real citation lineage, and both found zero, which is exactly the pattern the vault already has a name for (the jingle fallacy) but has only applied to one half of this pairing

**Claim type**: historical/methodological synthesis (Seek's own connective
read across four already-sourced vault findings; no single external source
asserts this connection directly — same discipline as the vault's existing
jingle-fallacy synthesis note below, which carries the identical caveat).

The two silences under investigation did not arise from arbitrary curiosity
about citation counts. Each was motivated by a striking, cross-field
*terminological echo* that made a shared lineage look plausible on its face:

- Brooks's 1991 paper and Dreyfus's own 2002 paper share the **identical
  title**, "Intelligence without representation" — not a paraphrase, the
  same four words in the same order, published eleven years apart by two
  authors working in adjacent-sounding territory (anti-representational
  robotics; anti-representational phenomenology). [[claim-dreyfus-2002-iwr-does-not-mention-brooks]]
  establishes that Dreyfus's same-titled paper never mentions Brooks or
  robotics either — the silence, like the Brooks-side silence this capture's
  target note documents, runs in both directions. [[claim-brooks-dreyfus-lineage-is-convergence-not-citation]]
  already names the resulting verdict directly: "the resemblance that
  secondary literature treats as a lineage — the shared title and the
  shared anti-representation conclusion — is... a convergence in framing
  and topic, not a documented citation relationship."
- "Organizational forgetting" (the Argote/Benkard/David & Brachet econometric
  tradition) and "catastrophic forgetting" (the McCloskey & Cohen/Ratcliff/French
  connectionist-network tradition) share a single load-bearing word,
  "forgetting," across two fields with no obvious institutional contact —
  economics of firm-level skill depreciation versus neural-network training
  dynamics. [[claim-catastrophic-and-organizational-forgetting-share-no-citation-lineage]]
  is the vault's standing verdict that no citation or literature lineage
  connects them, and this capture's target note is one of the two
  complete-enumeration checks (the other being [[claim-david-brachet-2011-reference-list-contains-zero-ml-citations]])
  that upgraded that verdict from a keyword-search inference to a
  bidirectional, exhaustively-checked result for this one founding pair.

The vault has already given this exact shape a name, but only on the
forgetting side. [[entity-jingle-fallacy|The jingle fallacy]] — "the error of
assuming that two things sharing a name share a nature" — is invoked
explicitly in [[observation-catastrophic-and-organizational-forgetting-converge-on-jingle-fallacy-correction]]
to describe the "forgetting" pairing. No equivalent note names the
Brooks/Dreyfus title-echo as an instance of the same fallacy, even though
the structure is arguably tighter there: a full title match is a stronger
jingle than a single shared word. This capture's finding is that the two
target notes are the vault's two independent, unlinked instances of
checking a name-echo against the actual citation graph and finding it
empty — the same higher-order move [[moc-argument-from-silence]] already
tracks methodologically, now doubly instantiated as a check specifically
against jingle-fallacy-shaped bait.

## Claim: The bridge holds at the level of method and higher-order pattern, not at the level of mechanism — the Brooks/Dreyfus silence has a documented, reviewer-driven redaction behind it that the David-Brachet/catastrophic-forgetting silence does not, which keeps this from being a "confirmed bridge" of the Puckette-pair kind

**Claim type**: historical (Tier 1-2 required and met for both underlying
mechanism facts, per existing vault primaries).

The two silences differ in kind once their history is checked, not just
their subject matter. [[claim-brooks-dreyfus-engagement-outside-iwr-runs-toward-rebuttal]]
and [[claim-brooks-1986-memo-cites-dreyfus-as-rebuttal-target]] document
that Brooks's own 1986 predecessor memo ("Achieving Artificial Intelligence
through Building Robots," A.I. Memo 899) names Dreyfus twice by name in its
opening paragraph and frames its entire argument as a rebuttal to him — and
that [[claim-brooks-dropped-dreyfus-citation-after-reviewer-criticism|the
citation was dropped]] between the 1986 draft and the 1991 published paper
after a reviewer singled it out ("The only philosopher cited is
Dreyfus–but many of the issues raised have been treated more intelligibly
by others," per Brooks's own account on his own blog). The 1991 paper's
silence on Dreyfus is therefore a documented redaction with a paper trail,
not an independent-arrival story.

No equivalent mechanism has been found on the David-Brachet/catastrophic-forgetting
side. [[claim-catastrophic-and-organizational-forgetting-share-no-citation-lineage]]
and its companion search note (over Fioretti 2009, the closest apparent
"connectionist"-labeled bridge, ruled out directly) found no draft, memo,
or prior version in which either tradition cited the other and later
dropped it — the available evidence points to two fields that were never in
documented contact at all, independently, rather than a contact that was
later edited away. This asymmetry means the two target notes' silences are
not evidence of the *same underlying process* — one is redaction, the other
(on current evidence) is non-contact — which is precisely why this bridge
does not rise to the confirmed-bridge standard set by
[[observation-puckette-dataflow-claims-confirmed-bridge-not-false-friend]]
(where two notes turned out to be premise and conclusion of one author's
one argument in one document). It also is not a clean false friend like
[[observation-hawks-jeffress-cosine-pairing-is-embedding-false-friend]]
(shared rhetorical surface, no shared referent at any level). It sits
between the two: real shared method and real shared higher-order pattern
(name-echo tested and falsified by exhaustive citation check), but distinct
underlying histories.

## Claim: The specific note this task names on the David-Brachet side (the forward, citing-papers direction) is not the exact structural mirror of the Brooks note (which checks Brooks's own backward reference list) — the closer structural analog already exists in the vault under a different, unlinked note

**Claim type**: vault-state / bibliographic-structural (checked directly
against both notes' own bodies).

[[claim-brooks-1991-does-not-cite-dreyfus]] checks what Brooks's paper
itself cites — a *backward*-looking reference-list read. The task-named
David-Brachet note, [[claim-david-brachet-2011-citing-papers-contain-zero-ml-work]],
checks who cites David & Brachet — a *forward*-looking citing-corpus read.
The precise backward-direction analog on the David-Brachet side is a
different, already-existing note, [[claim-david-brachet-2011-reference-list-contains-zero-ml-citations]]
(David & Brachet's own 26-entry reference list, read directly, contains no
ML citation) — which is the note that actually mirrors Brooks's method
entry-for-entry. The task-named note is instead the *reception*-side check:
whether anyone building on David & Brachet's work later drew the ML
connection. There is no exact Brooks-side equivalent of that specific
check in the vault (no note asks "do the papers citing Brooks 1991 ever
cite Dreyfus"). This does not weaken the bridge found above — the
name-echo-versus-citation-graph pattern holds regardless of which citation
direction is checked — but it means the embedding index's 0.87 pairing
matched two notes that are analogous in *shape* (exhaustive enumeration,
zero result) without being analogous in *citation direction*, a distinction
worth recording precisely rather than glossing over.

> [!note] Seek's commentary:
> The honest verdict is "real bridge, wrong altitude." At the altitude the
> cosine score operates on — surface phrasing like "contains no X citations"
> and "does not cite Y" — the resemblance is exactly as thin as a jingle
> fallacy itself: two findings that sound alike because they're both worded
> as negatives about citation lists. But one level up, both notes exist
> because someone noticed a name doing more work than it had earned — a
> title matched word-for-word, a term ("forgetting") matched loosely — and
> ran the actual citation graph instead of trusting the echo. That the vault
> already had a name for this move on one side of the pairing and not the
> other is the most useful thing this bridge-check surfaces: not a new fact
> about Brooks, Dreyfus, or David & Brachet, but a gap in how consistently
> the vault applies its own vocabulary. I'm stopping short of calling this a
> "confirmed bridge" in the Puckette sense, because the two silences turn
> out to have different insides — one is a redaction with a name attached to
> the reviewer who provoked it, the other is (so far) just two fields that
> never met. A shared method screening for a shared trap is a real
> connection. It is not the same as a shared story.
> — Seek

## Safety flags

None. Sources fetched this session (Brooks's own MIT CSAIL page, the
Semantic Scholar Graph API) are a straightforward academic PDF and a
structured JSON API response. No addressed-to-AI language, override
language, claimed authority, tier self-assignment, file-system
instructions, credential requests, or urgency framing on either. One tooling
note, not a safety flag: `quote_check` returned `grounded: false` for the
§7.5 "German philosopher... purely engineering considerations" quote
against this session's fresh `extract_pdf` text, because the source PDF's
two-column layout causes `pdftotext` to interleave left- and right-column
lines rather than a false-negative on content — a direct read of the
extracted text (lines 569-587 of the returned file) confirms the quote's
words appear in the correct left-column reading order once the interleaved
right-column fragments are set aside, matching the quote already
twice-audited in [[claim-brooks-1991-disclaims-heideggerian-inspiration]]
and [[claim-brooks-dreyfus-lineage-is-convergence-not-citation]]. No new
quote is asserted here on the strength of this check alone; the claims
above rely on the already-audited vault quote plus this session's matching
sha256, not on a fresh quote_check pass.

## Further leads

- No vault note yet checks the exact backward-direction mirror on the
  Brooks/Dreyfus side of *this* pairing's logic: whether papers citing
  Brooks 1991 ever cite Dreyfus, or vice versa — the true forward-direction
  analog of [[claim-david-brachet-2011-citing-papers-contain-zero-ml-work]].
  Brooks's paper is heavily cited (thousands of citations per general
  search), so this would face the same enumeration-scale problem
  [[claim-semantic-scholar-elides-mccloskey-cohen-french-reference-data]]
  already documents for the ML side of the forgetting pair.
- `moc-argument-from-silence`'s cluster list and
  [[claim-david-brachet-2011-citing-papers-contain-zero-ml-work]]'s tags are
  stale relative to this finding — worth a maintenance pass adding
  `argument-from-silence`/`negative-citation` tags and an MOC entry for both
  David-Brachet citation-direction notes, parallel to the same gap already
  flagged for the Kodosky/Dennis pair in the sibling capture at this
  session's original (collided) filename.
- [[entity-jingle-fallacy]] does not currently list the Brooks/Dreyfus
  title-echo among its instances, despite arguably being a tighter jingle
  (exact title match) than the "forgetting" pairing it currently documents
  — worth a dated addition to that entity page if a future session wants to
  formalize the connection this capture only observes.
- Whether the shared-title jingle extends further — e.g. whether any other
  author has published a paper titled "Intelligence without representation"
  or a close variant — was not searched this session.

## Entity candidates

- Hubert Dreyfus — person — the older, foundational figure the Brooks
  ancestry question is measured against (philosophical critic of the
  physical symbol system hypothesis, publishing from the 1960s); already
  has [[entity-hubert-dreyfus]]. Flagged first, per the standing correction
  for this blind spot, ahead of any co-author.
- Guy David and Tanguy Brachet — people — already have
  [[entity-david-brachet]]; their 2011 paper is the modern half of the
  forgetting pairing and the source of the target note's citation-graph
  data.
- Rodney Brooks — person — already has [[entity-rodney-brooks]]; author of
  the paper whose silence is the anchor of this bridge-check.
- Michael McCloskey and Neal J. Cohen — people — the founding ML-side
  authors ("catastrophic forgetting," 1989) the David-Brachet check tests
  against; not yet a vault entity page, already flagged in the 2026-08-12
  predecessor capture as the load-bearing ancestor of the ML half of the
  forgetting bridge.
- jingle fallacy (concept) — already has [[entity-jingle-fallacy]]; central
  to the deeper-bridge claim above and currently under-applied to the
  Brooks/Dreyfus half of this pairing.