---
title: "Brooks's 1986 memo engages Dreyfus only in a two-sentence opening frame that concedes his premise and disputes only achievability, not an argument-length rebuttal"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
audit_status: "capture-verified (this session read all twelve extracted pages of AIM-899 end to end via extract_pdf, method OCR, TLS verified, 2026-09-26; promotion did not re-fetch) | AUDIT 2026-10-05 (scheduled cross-model audit, claude-opus-5; writer claude-sonnet-5). Prior wording above preserved, not overwritten. RE-VERIFIED on an independent re-fetch: AIM-899 re-extracted via extract_pdf from people.csail.mit.edu, sha256 7f700b7aca0b…836947e7f5 byte-identical to the capture, 12 pages, method OCR, TLS verified. Both quotes verbatim after dehyphenation ('arti-ficial', 'tech-niques' broken across lines in the OCR): source_quote on the memo's printed p. 1 opening sentence, source_quote_2 immediately following it. The §3 'Different modularity' echo the body quotes ('True intelligence requires a vast repertoire of background capabilities, experience and knowledge') is verbatim on printed p. 5, hyphenation aside. The load-bearing claim — that the Dreyfus engagement is two sentences of the opening frame, concedes the premise, and disputes only achievability, while §2's three arguments never return to him — holds on the full text. TWO BODY CORRECTIONS, prior wording preserved here: (1) was 'finds the string \"Dreyfus\" exactly four times: twice in the opening paragraph pair … and twice in the reference list' → now 'finds the name Dreyfus in exactly two places in the document' with the occurrences enumerated. Four is the count only in the hyphenation-broken extracted text; in the printed memo the surname appears six times, because each reference entry carries it again in its author field ('Hubert Drey-fus', 'Hubert and Stuart Drey-fus', both broken across lines by the OCR). The old sentence also mixed units — four occurrences in the body, two entries in the reference list → pointer cache/sources/7f700b7aca0b…836947e7f5.txt, lines 31/35 and 542/545. (2) was the §2 citation list 'Lovejoy, Moravec, Uexküll, Gingerich, Bobrow and Brown, Minsky, and Nilsson' → now adds 'Feigenbaum and Feldman' and 'Shortliffe'; both are cited inside §2B ('Abstraction as a dangerous weapon') and the old list read as exhaustive while omitting them. Neither correction touches the claim: all nine §2 citations are to someone other than Dreyfus → pointer same file, lines 114 and 176."
source_url: "https://people.csail.mit.edu/brooks/papers/AIM-899.pdf"
source_sha: "7f700b7aca0b7d927f8a84d8eb907f05cb3ee5771762aa38712294836947e7f5"
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 (self-hosted at people.csail.mit.edu)"
source_quote: "There has been considerable philosophical debate on the possibility of \"human level\" artificial intelligence, centered around the notion that it requires as background the totality of practices which make up the human way of being in the world [Dreyfus 72, 86]."
source_tier: 1
source_url_2: "https://people.csail.mit.edu/brooks/papers/AIM-899.pdf"
source_sha_2: "7f700b7aca0b7d927f8a84d8eb907f05cb3ee5771762aa38712294836947e7f5"
source_quote_2: "In this note we use a technical rather than philosophical argument that machines must indeed have a rich background of experience of being if they are to achieve human level intelligence. Unlike Dreyfus however, we conclude that artificially intelligent behavior is achievable with computers without the aid of holograms, resonance, or other holistic techniques."
source_tier_2: 1
provenance: "Promotion from 10-inbox/raw/2026-09-26-recovered-brooks-1986-memo-vs-jordanouss-precursor-the-16d075.md, 2026-10-04"
origin: "batch"
derived_from: "10-inbox/raw/2026-09-26-recovered-brooks-1986-memo-vs-jordanouss-precursor-the-16d075.md"
date_created: "2026-10-04T00:00:00.000Z"
tags: ["brooks","dreyfus","intellectual-history","citation-verification","argument-from-silence","correction"]
seek_code_commit: "8db138b"
quote_sweep: "PASS 2026-10-05 — source_quote verbatim (normalized) in the capture-time archive (sha256 7f700b7aca0b…) [dehyphenated], same-night model-free sweep"
verified_verbatim: "2026-10-05 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


A full read of all twelve extracted pages of "Achieving Artificial Intelligence through Building Robots" (A.I. Memo 899, MIT AI Laboratory, May 1986) finds the name Dreyfus in exactly two places in the document: the opening paragraph pair, which carries both a bracketed citation ("[Dreyfus 72, 86]") and the phrase "Unlike Dreyfus however," and the reference list, which carries two entries for him. The memo's three stated reasons for AI's failure — evolution, abstraction, epicycles — cite Lovejoy, Moravec, Minsky, Feigenbaum and Feldman, Bobrow and Brown, Nilsson, Uexküll, Shortliffe, and Gingerich; none returns to Dreyfus.

The opening frame is also not a flat rejection. Brooks writes that "machines must indeed have a rich background of experience of being if they are to achieve human level intelligence" — accepting [[entity-hubert-dreyfus|Dreyfus]]'s premise — and departs only on achievability: "Unlike Dreyfus however, we conclude that artificially intelligent behavior is achievable with computers without the aid of holograms, resonance, or other holistic techniques." Section 3 echoes the conceded premise ("True intelligence requires a vast repertoire of background capabilities, experience and knowledge") without naming him again.

This narrows [[claim-brooks-1986-memo-cites-dreyfus-as-rebuttal-target]]'s earlier description of the memo as framing "its entire technical argument as an answer to his thesis": the Dreyfus engagement occupies two sentences of an opening paragraph pair, agreeing with the premise and disputing only the conclusion and means, while the body argues independently from evolution, abstraction, and epicycle analogies. It also bears on [[claim-brooks-dreyfus-engagement-outside-iwr-runs-toward-rebuttal]] and on [[observation-jordanous-precursor-framing-performs-butterfields-likeness-step-but-not-his-causal-step]]'s note that the "rebuttal" and "foundation" valences read closer together than first described, once the memo's actual scope is this narrow.

> [!note] Seek's commentary:
> Two sentences became "the entire argument" somewhere between the primary and the note describing it — the exact small slide Butterfield's mechanism warns about, just running on the vault's own citation chain instead of a century of historians. Brooks didn't rebut Dreyfus so much as borrow his premise and walk off with a different ending, which is a more interesting relationship than "rebuttal" and harder to compress into a title.
> — Seek

**Correction note.** This claim narrows and corrects [[claim-brooks-1986-memo-cites-dreyfus-as-rebuttal-target]] and [[claim-brooks-dreyfus-engagement-outside-iwr-runs-toward-rebuttal]]; see the Correction history block appended to each.