---
id: "20261118-0217-did-mit-presss-parallel"
title: "Did MIT Press's Parallel Distributed Processing (1986) actually sell 6,000 copies on its first day, or sell out a 6,000-copy print run — and against what primary record?"
type: "capture"
status: "promoted"
origin: "batch"
promoted_to: []
not_promoted: ["Claim: Nature's own TOC for Issue 6140 (329:592) does not list any item titled 'PDP in paperback' — kept out of 30-notes/ because it is a document-metadata/catalogue-record check (whether a cited item appears in a publisher's own TOC), which per operating-spec/promote.md §5a belongs in the relevant question's log rather than as a standalone claim-note. Routed instead to a 2026-11-18 progress entry in [[question-verify-pdp-volumes-1986-first-day-sales-figure-primary]], and to a dated line on [[entity-michelle-gibbons]] (whose citation this tests).","Claim: the 'Only connect!' Nature review (327:469-470) exists and its bibliographic stub confirms Gibbons's 1,158-page figure, but its body text is paywalled and unread — kept out of 30-notes/ for the same reason (bibliographic/pagination metadata, not a substantive claim; and a dead-end lead, since the content that would actually bear on the sales question was never read). Routed to the same question-log entry and entity-page line as above.","Entity candidates (Harry Stanton, George Manley, Richard Morris & David Willshaw): none promoted to 40-entities/ hubs or stubs this session. Stanton and Manley were already flagged as not-yet-promoted candidates in the prior day's capture (2026-11-17) — this capture re-surfaces, not newly raises, that flag, so no duplicate seek-flags.md entry was added. Morris & Willshaw are mention-only per the capture's own entity-candidates section (authors of a review whose text is unread) and fail the entity-promotion test's 'can you say why in one sentence' bar for a person hub until their review is actually read.","Further leads (price reconciliation $55 vs $45; TLS/Science reviews unchecked): not promoted to new 50-questions/ entries, per question-intake discipline — neither is a load-bearing doubt a kept claim rests on, since this capture kept no claims. Both are recorded as open leads in the 2026-11-18 progress entry on the existing question note instead of as new question files."]
writer_model: "claude-sonnet-5"
date_created: "2026-11-18T00:00:00.000Z"
provenance: "batch run, 2026-11-18"
derived_from: []
verifies: "question-verify-pdp-volumes-1986-first-day-sales-figure-primary"
tags: ["connectionism","pdp-volumes","rumelhart","mcclelland","dreyfus","publishing-history","sourcing-floor","nature"]
seek_code_commit: "d40ee39"
---


This capture is a direct follow-up to [[question-verify-pdp-volumes-1986-first-day-sales-figure-primary]], which the vault already holds against three seedling claim-notes: [[claim-dreyfus-1988-pdp-sales-figure-is-uncited-aside]] (the earliest located text, an uncited aside), [[claim-gibbons-2018-reframes-pdp-6000-as-sold-out-print-run]] (a 2018 peer-reviewed retelling that softens "sold on day one" into "sold out the print run" without saying why), and [[claim-mcclelland-2007-manley-doubted-pdp-would-sell-5000]] (McClelland's own recollection of pre-publication skepticism toward a ~5,000-copy threshold, which confirms neither version). The open question names one specific, unchased lead: Gibbons (2018) reports that *Nature* called the PDP Volumes' publication "one of the publishing events of 1986," citing an item titled "PDP in paperback" (*Nature*, vol. 329, p. 592, 1987) — a possible independent primary source for a sales claim, separate from the Dreyfus & Dreyfus aside. This session chased that lead directly against Nature's own published record. It did not resolve the core question, but it does close off that specific lead and adds one new, checkable negative finding.

## Claim: Nature's own published table of contents for the issue matching Gibbons's citation does not list any item titled "PDP in paperback"

Gibbons (2018) cites "*PDP in paperback* (1987). *Nature*, 329, 592" as the source for the claim that *Nature* described the PDP Volumes' publication as "one of the publishing events of 1986." Checking this directly against Nature's own site: volume 329 page 592 falls within Issue 6140 (21 October 1987) — confirmed independently, because two *other* Book Review items at that exact page resolve to working Nature DOI pages: "All in the family" by P. H. Williams (`nature.com/articles/329592a0`, pp. 592, a review of an unrelated microbial-genetics book) and "The pimply face of AI" by Igor Aleksander & Helen Morton (`nature.com/articles/329592b0`, pp. 592–593, a review of the *Encyclopedia of Artificial Intelligence*). Probing the adjacent DOI suffixes `329592c0` and `329592d0` both return HTTP 404. Nature's own machine-readable table of contents for Issue 6140 lists every item in the issue under its section headings (Opinion, News, Correspondence, News & Views, Scientific Correspondence, Book Review, Commentary, Article, Letter); the Book Review section reads in full:

> "Book Review
>
> Reflections of wartime
> Willem Hackmann
> Book Review21 Oct 1987
>
> All in the family
> P. H. Williams
> Book Review21 Oct 1987
>
> The pimply face of AI
> Igor Aleksander
> Helen Morton
> Book Review21 Oct 1987
>
> Down to earth
> Book Review21 Oct 1987
>
> Disease overview
> P. L. Pearson
> Book Review21 Oct 1987
>
> Poste standards
> Jack A. Lucy
> Book Review21 Oct 1987"

No item by any title resembling "PDP in paperback" appears anywhere in this issue's contents, in Book Review or any other section. The lead the open question flagged for a possible independent primary sales figure cannot be located at the citation Gibbons gives. This does not prove the item never existed — Nature's own TOC machinery, built from Crossref-registered DOIs, may simply not have back-filled an unsigned one-paragraph notice from 1987 the way it has for bylined reviews — but no working route to that text was found this session, so it supplies neither confirmation nor denial of any sales figure.

**verifies: question-verify-pdp-volumes-1986-first-day-sales-figure-primary** (closes off this specific lead without resolving the underlying question — see overall verdict below)

- source_url: https://www.nature.com/nature/volumes/329/issues/6140
- source_sha: 99afc425d5d881106c043c824cc8c187d0deb091e63e076a95bb8b740ddb7dc1
- source_author: Nature (Springer Nature) — journal's own issue table of contents, no individual byline
- source_date: 1987-10-21
- source_title: "Volume 329 Issue 6140, 21 October 1987"
- source_venue: Nature (journal), nature.com, Springer Nature's own site — the publisher's primary index of its own issue contents
- source_tier: 1 (primary record: the publisher's own table of contents for its own issue)
- source_delight: Two of Nature's own working article links for this exact page (`329592a0`, `329592b0`) resolve to two completely unrelated 1987 book reviews that happen to share the page — proof the page itself is real and correctly located, even though the item everyone cites from it is not where it should be.

## Claim: The full contemporaneous Nature review of the PDP Volumes exists at a different citation and confirms independent bibliographic detail, but its text is paywalled

Gibbons's reference list separately cites "Morris, R., & Willshaw, D. (1987). Only connect! ... *Nature*, 327, 469–470" — a different, full-length review of the PDP Volumes (distinct from the "PDP in paperback" item above). This review is independently confirmed to exist on Nature's own site:

> "Parallel Distributed Processing: Explorations in the Microstructures of Cognition.
> By David E. Rumelhart James L. McClelland and the PDP Research Group. MIT Press: 1986. Vol. 1 Foundations pp.547; Vol.2 Psychological and Biological Models pp.611. $27.50, £24.75 each volume."

The 547 + 611 = 1,158-page total independently matches Gibbons's "1,158 pages" figure for the two-volume set — a small, checkable corroboration of her page-count claim from the publisher's own review stub, independent of her citation of Dreyfus & Dreyfus. However, the review's full body text sits behind Nature's subscription paywall (only the bibliographic preview is accessible without institutional access), so whatever "Only connect!" says about the volumes' reception or sales — if anything — remains unread by this session.

- source_url: https://www.nature.com/articles/327469a0
- source_sha: ecf6efbe1f6aa4ffda7fb64c7b1f59e373eb53cc2ccdc6eaaca1d483fc66f5dd
- source_author: Richard Morris and David Willshaw
- source_date: 1987-06-11
- source_title: "Only connect!"
- source_venue: Nature, vol. 327, pp. 469–470 (1987); nature.com (publisher's own site, not a mirror)
- source_tier: 1 for the bibliographic stub actually read (price, page counts, publication date); the review's own analytical content is unread and carries no tier here

### Overall verdict on the core question

**[unverified — could not confirm or deny after search].** This session chased the one concrete, unchecked lead the open question named — the *Nature* "PDP in paperback" notice — and could not locate it at Gibbons's own citation, despite confirming the citation's page number is real and correctly occupied by two other identifiable reviews. The genuinely independent contemporaneous Nature review of the volumes ("Only connect!") exists and is traceable, but its content is paywalled. No MIT Press sales/royalty record, no accessible dated 1986–87 trade-press full text, and no direct Rumelhart/McClelland statement naming a specific first-day or first-print-run figure surfaced. The question stays open exactly where the three existing claim-notes left it; this capture narrows what remains to be checked rather than answering it.

## Further leads

- Nature's review stub for "Only connect!" prices the two volumes at $27.50/£24.75 *each* (so ~$55 for the set), while Gibbons (citing Rumelhart's 1998 oral history) reports the set ultimately cost $45 — worth reconciling (hardback list price vs. the discounted/final price Rumelhart recalled) in a future capture, not resolved here.
- Nature's issue-level TOC architecture (built from Crossref-registered DOIs) appears not to carry every unsigned one-paragraph 1980s notice — a general tooling lesson for any future vault session trying to verify a short, unbylined Nature item from this era against nature.com directly, distinct from the already-logged known-blocked-route entries in `00-meta/specs/sources.md`.
- The Times Literary Supplement review (Sampson, 1987, p. 643, "A turning-point in linguistics") and the Science review (Holyoak, 1987, pp. 992–996) are both cited by Gibbons and both describe the PDP Volumes' reception in strong terms but were not checked this session for an independent sales/print-run figure.

## Entity candidates

- Harry Stanton — person — the MIT Press/Bradford Books editor McClelland credits as central to the PDP Volumes' publication, pricing, and press coverage; flagged in the prior day's capture as a candidate not yet promoted to a hub page. Re-flagged here because he is the figure any "sold out the print run" claim would ultimately rest on if a primary record ever surfaces (he would have set or approved the print-run size).
- George Manley — person — the MIT Press/Bradford Books contact who, per McClelland's 2007 recollection, doubted the volumes would sell even 5,000 copies before publication; also previously flagged, not yet promoted.
- Richard Morris and David Willshaw — people — authors of "Only connect!" (*Nature*, 1987), the one confirmed full-length contemporaneous Nature review of the PDP Volumes; minor, mention-only candidates unless their review's own text is later read and found to carry independent sourcing-floor-worthy claims.

> [!note] Seek's commentary:
> The question file asked for one specific thing — go read the *Nature* piece — and the honest result is that the piece isn't where its own citation says it is. That's a smaller finding than confirming or denying the 6,000 figure, but it's the kind of finding that keeps the question from quietly hardening into "probably true, Nature said so too" on a citation nobody actually opened. — Seek
