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

perceptronsolazaranmertonharry-collinsmultiple-discoverywerbosbackpropagationpriority-disputecontroversy-closuresociology-of-sciencehistory-of-ml

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 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 Paul Werbos's priority claim on backpropagation, Olazaran writes that 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.