No published history-of-AI source checked combines Dreyfus 1988's 'vindicated, not refuted' account with Olazaran 1993's SRI-funding account — the vault's own pairing appears to be its own synthesis

dreyfusolazaranhistoriographysociology-of-scienceperceptronsconnectionismai-winterhistory-of-mlmeta-vault

The vault's earlier observation note reads Dreyfus and Stuart Dreyfus's 1988 Daedalus essay and Olazaran's 1993 sociological interviews as two independently-sourced accounts that converge on the same verdict: connectionism's mid-1960s eclipse was social closure, not scientific refutation. This note asks the narrower question of whether any published history-of-AI source has run that same combination as a single argument against "Perceptrons killed connectionism" — and finds, across five sources checked directly, that none has.

Olazaran's own chapter cites Dreyfus 1988 in-text, but only for a narrow dating point, never paired with his own SRI evidence against the refutation framing (claim-olazaran-1993-cites-dreyfus-1988-for-chronology-not-combined-framing). Nils Nilsson — uniquely positioned as both one of Olazaran's SRI interview subjects and a published AI historian — cites neither scholar when making a structurally similar anti-refutation claim of his own (claim-nilsson-2009-cites-neither-dreyfus-nor-olazaran-anti-refutation). And three further independent secondary surveys (Yuxi Liu 2024; Istvan Berkeley 1997 and 2019) split cleanly along disciplinary lines, each citing only the half of the pairing native to its own method (claim-perceptron-controversy-secondary-surveys-draw-only-one-half-of-dreyfus-olazaran-pairing).

No universal negative can be proven by a finite search, and two identified candidates — Margaret Boden's Mind as Machine (2006) and Anderson & Rosenfeld's Talking Nets (1998) — were not read this session (see question-corroborate-boden-anderson-rosenfeld-dreyfus-olazaran-combination). This claim is marked [unverified — convergent negative evidence across five independent sources, but exhaustive disconfirmation is not possible], leaning toward "this vault's own synthesis."

• 2026-10-17: sixth data point added — claim-eby-2020-kernelled-connections-cites-boden-olazaran-not-dreyfus. Michael Eby's 2020 &&&-journal essay cites both of this note's two identified-but-unread candidate books (Boden's Mind as Machine, Anderson & Rosenfeld's Talking Nets) alongside Olazaran's sociology of the controversy, and still never invokes Dreyfus. Strengthens the negative without proving it; the books' own full texts remain unread.