---
title: "Olazaran's 1993 chapter applies Merton's multiple-discovery thesis to Werbos's backpropagation priority claim, even though Harry Collins' controversy-closure framework — not Merton's — governs its main Minsky–Papert–Rosenblatt analysis"
type: "claim"
status: "seedling"
audit_status: "capture-verified, corrected — the originating capture session read only Olazaran's concluding section via extract_pdf and characterized Merton's 1961/1973 citations as appearing 'only in the reference list, nothing in the chapter's argument built on Merton's account.' Cali's 2026-10-06 daily review of that capture read the same archived primary (source_sha c95e7a32…) at p.396 directly and found Merton's multiple-discovery thesis quoted and footnoted there, applied in-argument to Werbos's priority claim — contradicting the capture's 'peripheral, not load-bearing' verdict. This note promotes the corrected finding in place of the capture's as-drafted claim, which is withheld (see the capture's not_promoted list). Two further corrections from the same review folded in: Collins' Empirical Programme of Relativism is the Bath School strand of the sociology of scientific knowledge, not Edinburgh (Edinburgh denotes the separate Strong Programme, Barnes/Bloor); and the capture's unsourced aside asserting that SSK historically developed as an alternative to Mertonian sociology has been dropped rather than carried forward without a citation. Independent re-fetch of the primary not performed in this headless promotion session (no network, by design); both the original read and the correcting read are grounded in the same archived sha."
source_url: "https://gwern.net/doc/ai/nn/1993-olazaran.pdf"
source_sha: "c95e7a3295751ea378d5ae510fee887f402d9dc1fe9365c48f6102073af6e150"
source_title: "A Sociological History of the Neural Network Controversy"
source_author: "Mikel Olazaran"
source_date: 1993
source_venue: "Advances in Computers, vol. 37 (1993), pp. 335-425"
source_quote: "the controversy-closure scheme from the sociology of science (Collins, 1981, 1985) can help interpret those developments"
source_quote_2: "claimed that all scientific discoveries are multiple"
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-10-06-bipartite-ml-ai-epistemology-no-standalone-1969-first-edition-scan.md, 2026-10-10 (headless); corrected per Cali's 2026-10-06 daily-review callout embedded in that capture"
origin: "batch"
derived_from: "10-inbox/raw/2026-10-06-bipartite-ml-ai-epistemology-no-standalone-1969-first-edition-scan.md"
date_created: "2026-10-10T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["perceptrons","olazaran","merton","harry-collins","multiple-discovery","werbos","backpropagation","priority-dispute","controversy-closure","sociology-of-science","history-of-ml"]
seek_code_commit: "d40ee39"
quote_sweep: "PASS 2026-10-11 — source_quote verbatim (normalized) in the capture-time archive (sha256 c95e7a329575…), same-night model-free sweep"
verified_verbatim: "2026-10-13 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


Olazaran's own conclusion names one governing framework for the chapter's
central narrative:

> "the controversy-closure scheme from the sociology of science (Collins,
> 1981, 1985) can help interpret those developments"

— citing [[entity-harry-collins|Harry Collins]]'s "Stages in the Empirical
Programme of Relativism" (1981) and *Changing Order* (1985), the founding
texts of the Bath School's Empirical Programme of Relativism (EPOR) strand
of the sociology of scientific knowledge — distinct from the separate Strong
Programme associated with Edinburgh. Collins' "interpretative flexibility" /
controversy-closure model is the tool Olazaran uses throughout to analyze
how the Minsky–Papert critique was read differently by rival neural-network
research groups, and how that flexibility closed.

But Merton's presence in the chapter is not confined to its reference list.
On p.396, discussing [[entity-paul-werbos|Paul Werbos]]'s priority claim on
[[entity-backpropagation|backpropagation]], Olazaran writes that
[[entity-robert-merton|Robert Merton]] "claimed that all scientific
discoveries are multiple," footnoting Merton's 1961 "Singletons and multiples
in science" (p.356, fn.44) — in the same footnote cluster that cites
Hagstrom (1965) and Merton (1973) specifically for Werbos's case. That is
Merton's multiple-discovery thesis invoked by name and page, in-argument, to
characterize a priority dispute the vault already tracks as a
Merton-shaped multiple-discovery story in its own right
([[myth-werbos-invented-backpropagation]],
[[claim-reverse-mode-multiple-independent-discovery]],
[[claim-merton-multiple-discovery-is-sciences-dominant-pattern]]).

So the one document runs two separable analytic moves on two separable
subjects: Collins' controversy-closure scheme governs the chapter's main
subject, the Minsky–Papert–Rosenblatt dispute; Merton's multiple-discovery
thesis is invoked — not merely listed — for a different episode within the
same chapter, Werbos's backpropagation priority claim. This settles, within
a document the vault had already read for an unrelated reason, a question
the vault had left open: whether Merton's multiple-discovery concept has
ever been applied in print to a specific episode of AI-history priority
adjudication. It has — just not to the episode the capture that first found
this bibliography collision was actually investigating.

> [!note] Seek's commentary:
> The first pass through this chapter stopped at the table of contents of
> its own argument — "Collins governs the conclusion, Merton sits in the
> bibliography" — and that was wrong by one footnote. Cali caught it by
> doing the thing the first pass should have done: reading page 396, not
> just the last page. The correction is more interesting than what it
> replaces. A chapter can run two sociological frameworks at once, one per
> dispute, and the only way to know that is to stop trusting a document's
> stated methodology to describe everything it does.
> — Seek