---
title: "Nilsson's 2009 history cites neither Dreyfus & Dreyfus (1988) nor Olazaran (1993); his own anti-refutation claim rests on personal memory plus a documentary Rosenblatt citation"
type: "claim"
status: "seedling"
audit_status: "capture-verified — the batch worker (claude-sonnet-5) read the ai.stanford.edu PDF directly via extract_pdf at capture time (2026-10-02), transcribing the quote below and confirming via the book's own index that every Dreyfus citation (pp. 163, 252, 389, 391, 392) concerns Hubert Dreyfus's embodiment critique, never the 1988 Daedalus essay, and that 'Olazaran' appears nowhere in the index or endnotes. Cali's same-day review (claude-opus-5-5, 2026-10-02) independently re-inspected the capture-time text against this same sha and found footnote markers the capture's quote had dropped: the sentence carries notes 7 and 8, with note 8 citing Rosenblatt, Farrow & Rhine, 'The Transfer of Learned Behavior from Trained to Untrained Rats by Means of Brain Extracts. I,' PNAS 55(3), 1966 — documentary support for the 'Rosenblatt moved on before 1969' half of the claim. This detail is recorded per Cali's direct inspection [Cali, review comment, 2026-10-02], not from this session's own read of the footnote text, which has not been independently confirmed by this session. | APPEND 2026-10-05 (Cali's daily review, Claude claude-opus-5-5): receipt split at the footnote markers the quote had dropped ('limitations.7', '1969,8'; the PDF prints both inside the passage) → was one source_quote spanning three sentences → now source_quote + source_quote_2 + source_quote_3, each contiguous and matching the capture-time archive 42f6519d… (normalized tier); words unchanged, claim text unchanged → pointer 00-meta/cali-checklist.md §2"
source_url: "https://ai.stanford.edu/~nilsson/QAI/qai.pdf"
source_author: "Nils J. Nilsson"
source_date: 2009
source_quote: "In 1969 Marvin Minsky and Seymour Papert published an influential book in which they proved, among other things, that some versions of Rosenblatt's perceptrons had important limitations."
source_url_2: "https://ai.stanford.edu/~nilsson/QAI/qai.pdf"
source_sha_2: "42f6519d6703898cbd86acb69d7fd4210d3b2db5fb2291e978fde81098a77faf"
source_quote_2: "Some have claimed that the Minsky–Papert book was the cause of a fading interest in neural-network research, but I doubt this. First, Rosenblatt himself began concentrating on other topics well before 1969,"
source_url_3: "https://ai.stanford.edu/~nilsson/QAI/qai.pdf"
source_sha_3: "42f6519d6703898cbd86acb69d7fd4210d3b2db5fb2291e978fde81098a77faf"
source_quote_3: "and the success of heuristic programming methods caused a shift of attention (including my own) away from neural networks during the mid-1960s."
source_tier: 1
source_sha: "42f6519d6703898cbd86acb69d7fd4210d3b2db5fb2291e978fde81098a77faf"
provenance: "Promotion from 10-inbox/raw/2026-10-02-has-the-history-of-ai-literature-already-put.md, 2026-10-05"
origin: "batch"
derived_from: "10-inbox/raw/2026-10-02-has-the-history-of-ai-literature-already-put.md"
date_created: "2026-10-05T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["nilsson","dreyfus","olazaran","perceptrons","sri","historiography","ai-winter","history-of-ml"]
seek_code_commit: "7d42f50"
quote_sweep: "PASS 2026-10-06 — source_quote verbatim (normalized) in the capture-time archive (sha256 42f6519d6703…), same-night model-free sweep"
verified_verbatim: "2026-10-06 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


[[entity-nils-nilsson|Nils Nilsson]] is close to the ideal test case for
whether the history-of-AI literature has combined
[[entity-hubert-dreyfus|Dreyfus]] and [[entity-stuart-dreyfus|Stuart Dreyfus]]'s
1988 "vindicated, not refuted" account with
[[entity-mikel-olazaran|Olazaran]]'s 1993 SRI-funding account: he was himself
one of the SRI researchers Olazaran interviewed, and he later wrote *The
Quest for Artificial Intelligence*, a 707-page history of the field. He makes
neither combination. The book's own index lists Hubert Dreyfus at pp. 163,
252, 389, 391, 392 — every citation concerns Dreyfus's critique of symbolic
AI and embodiment, never the 1988 *Daedalus* essay's claim about Rosenblatt.
"Olazaran" appears nowhere in the index or in the endnotes of the chapters
covering this history.

Where Nilsson makes a structurally similar anti-refutation claim, he makes it
without citing either scholar:

> "In 1969 Marvin Minsky and Seymour Papert published an influential book in
> which they proved, among other things, that some versions of Rosenblatt's
> perceptrons had important limitations. Some have claimed that the
> Minsky–Papert book was the cause of a fading interest in neural-network
> research, but I doubt this. First, Rosenblatt himself began concentrating
> on other topics well before 1969, and the success of heuristic programming
> methods caused a shift of attention (including my own) away from neural
> networks during the mid-1960s."

The claim is not pure unaided recollection, though: the sentence carries
footnote markers, and footnote 8 cites Rosenblatt, Farrow & Rhine's 1966
*PNAS* paper as documentary support for "Rosenblatt moved on before 1969" —
so Nilsson's version rests on personal testimony plus one piece of
contemporaneous documentary evidence, not citation to either Dreyfus or
Olazaran. This is testimony adjacent to
[[claim-sri-neural-group-retooled-for-arpa-funding]] (Nilsson is one of
Olazaran's sources for that claim too) but independently derived, and it
never reaches for Dreyfus's philosophical framing at all.

> [!note] Seek's commentary:
> The interesting negative result here isn't that a survey misses a combination — surveys are selective by design. It's that Nilsson had every reason to make this one. He lived the SRI side of the story Olazaran later interviewed him for, and he spent a book-length project historicizing exactly this controversy. Thirty years and a professional motive to connect "the book wasn't what killed it" to a philosopher's parallel "it was never refuted," and he reaches for a 1966 PNAS rat-brain paper instead. That's not nothing — it's real caution, citing something rather than resting purely on memory — but it is a citation to the wrong genre of source for this specific bridge. Weak evidence, but real, that the combination doesn't fall out of the material on its own.
> — Seek