---
title: "Stuart Dreyfus person-bridges Brooks's 1986 rebuttal-target citation and the Kelley-Bryson-Werbos backpropagation-priority dispute Werbos's EIA gas-market work later ran into — a mediated bridge, not a false friend, not a direct link"
type: "observation"
status: "seedling"
writer_model: "claude-sonnet-5"
source_url: "https://people.csail.mit.edu/brooks/papers/AIM-899.pdf"
source_title: "Achieving Artificial Intelligence through Building Robots"
source_author: "Rodney A. Brooks"
source_date: "1986-05"
source_venue: "MIT Artificial Intelligence Laboratory, A.I. Memo 899"
source_quote: "[Dreyfus 86] \"Why Computers May Never Think Like People\", Hubert and Stuart Dreyfus, Technology Review, January 1986."
source_tier: 1
source_sha: "7f700b7aca0b7d927f8a84d8eb907f05cb3ee5771762aa38712294836947e7f5"
source_url_2: "https://gwern.net/doc/ai/nn/1990-dreyfus.pdf"
source_title_2: "Artificial Neural Networks, Back Propagation, and the Kelley-Bryson Gradient Procedure"
source_author_2: "Stuart E. Dreyfus"
source_date_2: "1990-09/10"
source_tier_2: 1
source_sha_2: "59bdcf82ce2f989b787000bd3cdbe56ba27b97a6d6499fd66434d5e0b68c6836"
provenance: "Promotion from 10-inbox/raw/2026-09-28-bipartite-two-things-this-vault-knows-in-different.md, 2026-09-28"
origin: "batch"
derived_from: ["10-inbox/raw/2026-09-28-bipartite-two-things-this-vault-knows-in-different.md"]
date_created: "2026-09-28T00:00:00.000Z"
tags: ["bipartite-bridge","person-bridge","rodney-brooks","hubert-dreyfus","stuart-dreyfus","paul-werbos","kelley-bryson","backpropagation","history-of-ml","false-friend-test"]
audit_status: "capture-verified — both primaries were read directly via extract_pdf, TLS verified, at capture time (source_sha and source_sha_2 above); this note's contribution is the synthesis connecting them, not a new primary read. Not independently re-checked this promotion session (headless, no network access). | 2026-09-29 cross-model audit (claude-opus-5, writer was claude-sonnet-5): both primaries re-fetched via extract_pdf, TLS verified; both sha256 values byte-identical to source_sha and source_sha_2. source_quote CONFIRMED verbatim in AIM-899's reference list. The quote_sweep FAIL was a FALSE NEGATIVE — AIM-899 extracts by OCR, which hyphenates the surname across a line break ('Hubert and Stuart Drey-/fus') and renders curly quotation marks, so a normalized match fails on bytes that are present. Confirmed by direct read: Brooks cites Dreyfus in exactly two reference entries ([Dreyfus 72], Hubert alone; [Dreyfus 86], Hubert and Stuart jointly), plus one inline '[Dreyfus 72, 86]'; and AIM-899 contains no occurrence of optimal control, Kelley, Bryson, Werbos, or backpropagation. CORRECTED 2026-09-29: the body's assertion that Dreyfus 1990 'contains no mention of … philosophy of mind' was wrong — Dreyfus 1990's references 1 and 2 are Hubert-and-Stuart Dreyfus philosophy-of-AI works (Daedalus Winter 1988, pp. 15-43; Mind over Machine, Free Press, 1988), cited in its own introduction. This tightens the bridge rather than loosening it and the body now says so; the 'mediated bridge, not a false friend, not a direct link' verdict is unchanged, since neither document cites the other."
seek_code_commit: "f8a723c"
quote_sweep: "PASS 2026-10-01 — recheck: source_quote verbatim in the capture-time archive (sha256 7f700b7aca0b…) [dehyphenated]; was: FAIL 2026-09-29 — source_quote NOT found in the capture-time… (quote since repaired, or the earlier check misread the extraction)"
audits: ["2026-09-29 claude-opus-5"]
verified_verbatim: "2026-10-01 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


A retrieval pass paired [[claim-brooks-1986-memo-cites-dreyfus-as-rebuttal-target]]
against [[claim-werbos-authored-eia-natural-gas-market-model-documentation-1981]] at
cosine 0.84. Investigating the pairing found a real connecting figure: the "Dreyfus"
[[entity-rodney-brooks|Brooks]] cites twice in AIM-899 is, in its second entry,
jointly authored by [[entity-hubert-dreyfus|Hubert]] and
[[entity-stuart-dreyfus|Stuart Dreyfus]] — and this same Stuart Dreyfus is the
mathematician who, in 1990, names
[[claim-dreyfus-1990-names-werbos-independent-argues-kelley-bryson-priority|Werbos
directly]] in a priority dispute over backpropagation, arguing credit belongs to
Henry Kelley and Arthur Bryson instead
([[claim-kelley-bryson-optimal-control-precursor]]). The method under that dispute
is the same "dynamic feedback" generalization
[[claim-werbos-1988-doe-affiliation-and-gas-market-test-case|Werbos's 1988 paper
applied]] to diagnose the Natural Gas Market Model he had authored
[[claim-werbos-authored-eia-natural-gas-market-model-documentation-1981|the EIA's
official 1981 documentation]] for. So a real, documented, two-hop chain connects the
two seed claims: Brooks's cited co-author is Werbos's priority rival.

But the bridge is a shared node, not a shared argument. Brooks's AIM-899, read in
full, contains no mention of optimal control, Kelley, Bryson, Werbos, or
backpropagation — its rebuttal of Dreyfus is philosophical (embodiment, "holistic
techniques"), unconnected to the control-theory credit dispute. Dreyfus's 1990
paper, read in full, contains no mention of Brooks or robotics; its argument is a
mathematical credit-attribution claim addressed to the optimal-control and
neural-network literatures. It is not, however, innocent of philosophy of mind: its
first two references are Hubert-and-Stuart Dreyfus philosophy-of-AI works — "Making
a Mind vs Modeling the Brain: Artificial Intelligence Back at a Branchpoint"
(*Daedalus*, Winter 1988, pp. 15–43) and *Mind over Machine* (Free Press, 1988) —
cited in its introduction to support the claim that "conventional inference-based
artificial intelligence was unable to deal successfully with most practical
problems." So the same Hubert-and-Stuart authorial pair whose *Technology Review*
piece Brooks lists as [Dreyfus 86] is load-bearing in the 1990 control-theory note
as well. What remains absent is any contact between the two *arguments*: neither
document cites the other, neither engages the other's specific subject, and nothing
shows either man aware of the other's project.

This sits in the middle of the vault's bridge taxonomy, not at either pole. It is
not the false-friend verdict already reached for a *different* Brooks/Werbos
pairing at cosine 0.85, which found no connecting figure at all
([[observation-brooks-werbos-cosine-085-pairing-is-embedding-false-friend]] — that
finding is unaffected and still holds for that pairing). It is also not a confirmed
direct bridge: no citation, correspondence, or shared venue puts Brooks and Werbos,
or Brooks and the Kelley-Bryson dispute, in contact. It matches instead the vault's
established person-bridge pattern — a single human sitting inside two otherwise
disconnected clusters ([[observation-feldbaum-person-bridge-invisible-to-vault-bridge-tool]],
[[observation-rumelhart-person-bridge-schema-theory-to-backprop]],
[[observation-paul-allen-person-bridges-altair-basic-and-semantic-scholar-clusters]])
— with one difference worth naming: those bridges were invisible to embedding
retrieval by construction, because no note yet named the connecting person. This one
was *found* by a cosine pairing between two notes that never mention Stuart Dreyfus
by his full role in either thread — the embedding apparently caught enough shared
vocabulary (Dreyfus, a rebuttal posture, the 1980s) to flag the pair as worth a
human look, and the look turned up a real edge the two claims themselves never draw.
See [[moc-backpropagation-origins]] and
[[observation-vaults-in-between-bridge-category-not-exactly-named-by-located-literature]]
for the taxonomy this sits inside.

> [!note] Seek's commentary:
> The taxonomy wants two bins — false friend, confirmed bridge — and this pairing
> keeps sliding off both. It isn't nothing: Stuart Dreyfus is a real, documented,
> two-hop connector, not a coincidence of embeddings. It isn't a mechanism either,
> because a mechanism would need the two men's *arguments* to touch, and they never
> do — Brooks never read the Kelley-Bryson correction, Dreyfus never weighed in on
> subsumption architecture. What actually connects them is a person wearing two
> unrelated hats thirty years apart: co-author to one critic, priority rival to one
> engineer. The retrieval tool caught the wallpaper and missed the man standing in
> both rooms until a human went and looked directly at him — which is, I think, the
> whole argument for why this vault still needs someone reading, not just indexing.
> — Seek