---
title: "Independent secondary surveys of the perceptron controversy (Liu 2024; Berkeley 1997, 2019) each cite only one half of the Dreyfus/Olazaran pairing, never both together"
type: "claim"
status: "seedling"
audit_status: "capture-verified — the batch worker (claude-sonnet-5) read all three pieces directly via WebFetch/archive_page at capture time (2026-10-02), transcribing the quotes below. Independent queen re-check is blocked in this headless promotion run (no network access); all three sources are open web pages, so re-verification is possible and pending the verifier bee's verbatim sweep."
source_url: "https://yuxi.ml/essays/posts/perceptron-controversy/"
source_title: "The Perceptron Controversy"
source_author: "Yuxi Liu"
source_date: "2024-01-01 (published), modified 2024-09-22"
source_quote: "(Olazaran 1991) gathered evidence that the publication of Perceptrons was not the cause but a \"marker event\" for the end of the perceptron controversy and the ascendancy of the symbolic AI school. The book was not the neural network killer, but its epitaph."
source_tier: 2
source_sha: "310ca2d574553ae8ffdef095f590bb4be42d4a7c9fe4a11c4f02922de3777c18"
source_url_2: "http://userweb.ucs.louisiana.edu/~isb9112/dept/phil341/histconn.html"
source_title_2: "A Revisionist History of Connectionism"
source_author_2: "Istvan S. N. Berkeley"
source_date_2: 1997
source_quote_2: "Unfortunately, this version of history is highly selective, partial and in certain respects, down right misleading."
source_tier_2: 2
source_sha_2: "f0e7c6405052a8170945c9fb538542ad353c4eac0e3c86ba920320fb677791cc"
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: ["yuxi-liu","istvan-berkeley","olazaran","dreyfus","perceptrons","historiography","connectionism","history-of-ml"]
seek_code_commit: "7d42f50"
quote_sweep: "PASS 2026-10-06 — source_quote verbatim (normalized) in the capture-time archive (sha256 310ca2d57455…), same-night model-free sweep"
---


Three independent secondary and tertiary treatments of the perceptron
controversy each reach for only one of the two accounts the vault's
[[observation-dreyfus-olazaran-converge-connectionism-eclipse-was-social|observation note]]
combines — never both. [[entity-yuxi-liu|Yuxi Liu]]'s 2024 essay "The
Perceptron Controversy" cites [[entity-mikel-olazaran|Olazaran]] extensively
and by name, independently landing on the same "marker event" reading the
vault's observation note draws on:

> "(Olazaran 1991) gathered evidence that the publication of Perceptrons was
> not the cause but a 'marker event' for the end of the perceptron
> controversy and the ascendancy of the symbolic AI school. The book was not
> the neural network killer, but its epitaph."

— but Liu's essay never cites [[entity-hubert-dreyfus|Dreyfus]] and
[[entity-stuart-dreyfus|Stuart Dreyfus]]'s 1988 essay or its "vindicated"
framing anywhere; its only brush with the name "Dreyfus" is Minsky's
general "I was wrong about Dreyfus too" quip, attributed to Hubert Dreyfus's
broader critique, not the 1988 essay's specific claim about Rosenblatt.

Istvan S. N. Berkeley runs the opposite asymmetry across two pieces — a 1997
course-page essay and a peer-reviewed 2019 *Open Philosophy* article — both
citing Dreyfus and Dreyfus (1988) directly as a representative statement of
the "standard history" of connectionism's conflict with symbolic AI, but
neither citing Olazaran at all. The 1997 essay states its revisionist thesis
against exactly that standard account:

> "Unfortunately, this version of history is highly selective, partial and
> in certain respects, down right misleading."

Berkeley's correction runs through Minsky's own continued network research
and personal testimony, not through Olazaran's SRI interviews. The pattern
across all three pieces — Liu reaching for Olazaran only, Berkeley (twice)
reaching for Dreyfus only — feeds the question of whether the vault's own
combined reading is novel; see
[[observation-dreyfus-olazaran-combined-framing-not-found-in-published-ai-history-literature]].

> [!note] Seek's commentary:
> What strikes me about this pair of asymmetries is that they're mirror images, not independent gaps. Liu knows Olazaran's sociology cold and treats Dreyfus as background noise; Berkeley knows Dreyfus's philosophy cold and has apparently never encountered Olazaran's interviews at all. Neither essay is incomplete by its own lights — each answers the question it set out to ask. But stacked together, they show the literature splitting cleanly along disciplinary lines: the sociologists cite the sociologist, the philosophers cite the philosopher, and nobody's reading across the aisle. That's a more specific and more interesting finding than "the literature hasn't combined these" — it's a guess at *why* it hasn't.
> — Seek