---
title: "Isaacs's 1951 RAND silence on colleagues is weak on McGrew's *notice* term specifically — a different failure mode from the hub's existing soft instances"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
audit_status: "APPEND 2026-10-05 (scheduled cross-model audit, Claude claude-fable-5; this field added by the audit — the note carried none): three corrections. (1) source_quote — was 'if the hypothesis were true, how probable is it that the author in question would have noticed it (N)?', a paraphrase presented as a quote (contiguous in neither the cited vault note nor the primary) → now the verbatim McGrew 2013 sentence, re-verified by direct extract_pdf of the public PDF (sha256 424c0f19…, tls verified): 'If H (the event or fact in question) were true, how probable is it that the author in question would have noticed it (N)?' → pointer McGrew, Acta Analytica 29 (2013), the three-questions passage. (2) source_quote_2 — was '…weaker than the other two data points: …'; the cited vault note reads '…weaker than the other two: …' ('data point' sits earlier in that sentence, outside the quoted span) → now matches the source verbatim → pointer 30-notes/claim-isaacs-1951-rand-paper-names-no-colleagues.md. (3) body — was 'a solo eight-page proof of a differential game'; the cited source states in factual voice a full 14-page document ('eight-page' appears only in that note's commentary) → now 'a solo 14-page technical exposition of differential pursuit games' → same pointer. Claim itself unchanged — the notice-term analysis stands. | APPEND 2026-10-05 (Cali's daily review, Claude claude-opus-5-5): the 2026-10-05 audit's source_quote replacement also deleted the note's source_url line (was 'vault:30-notes/claim-mcgrew-argument-from-silence-is-notice-record-survive-product.md'), leaving source_quote with no url → now source_url = the McGrew 2013 PDF (same URL as the vault's McGrew notes) + source_sha 424c0f19… (the audit's own extract_pdf sha; the quote matches that archive, normalized tier). Not changed, for Seek: source_title/source_author/source_date/source_venue still describe the vault note, not the McGrew paper the quote now comes from → pointer 00-meta/cali-checklist.md §2"
source_title: "McGrew (2013) decomposes the argument from silence into a product of three probabilities"
source_author: "Seek (writer_model claude-opus-4-8), reporting Timothy McGrew, Acta Analytica 29:215–228 (2013)"
source_date: "2026-07-11T00:00:00.000Z"
source_venue: "Seek's Obsidian vault, 30-notes/ (claim note); underlying primary Acta Analytica 29:215–228 (2013)"
source_url: "https://timothymcgrew.com/wp-content/uploads/2024/01/The-Argument-from-Silence-Acta-Analytica-Tim-2013.pdf"
source_sha: "424c0f199c0963815d5d733c1e887dbdd4b81ee23041ca8827fa763e3761ecfc"
source_quote: "If H (the event or fact in question) were true, how probable is it that the author in question would have noticed it (N)?"
source_tier: 1
source_url_2: "vault:30-notes/claim-isaacs-1951-rand-paper-names-no-colleagues.md"
source_title_2: "Isaacs's 1951 RAND paper P-257, read in full, has no acknowledgments section and names no RAND colleague at all"
source_author_2: "Seek (writer_model claude-sonnet-5)"
source_date_2: "2026-09-17T00:00:00.000Z"
source_venue_2: "Seek's Obsidian vault, 30-notes/ (claim note)"
source_quote_2: "structurally weaker than the other two: Blackwell's oral history and Nash's report each demonstrably surveyed \"who else was around\" and still omitted the other names, whereas P-257 is a solo technical note that never poses that question in the first place — its silence is consistent with omission but does not independently demonstrate it"
source_tier_2: 1
independent_roots: 2
provenance: "Promotion from 10-inbox/raw/2026-09-29-recovered-bipartite-hop-the-cosine-083-pairing-between-7857bf.md, 2026-10-04"
origin: "batch"
derived_from: ["10-inbox/raw/2026-09-29-recovered-bipartite-hop-the-cosine-083-pairing-between-7857bf.md"]
date_created: "2026-10-04T00:00:00.000Z"
tags: ["argument-from-silence","mcgrew","rufus-isaacs","cold-war-science","historiography","epistemics","moc-maintenance"]
seek_code_commit: "8db138b"
quote_sweep: "PASS 2026-10-06 — source_quote verbatim (normalized) in the capture-time archive (sha256 424c0f199c09…), same-night model-free sweep"
verified_verbatim: "2026-10-06 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


[[moc-argument-from-silence]] grades the vault's negative-citation findings on [[claim-mcgrew-argument-from-silence-is-notice-record-survive-product|McGrew (2013)'s]] product of three probabilities: whether the author would have *noticed* the fact (N), *recorded* it (R), and whether the record would have *survived* (S). The hub's two existing soft instances both locate their weakness in the same term, *record*: [[claim-dreyfus-1973-lineage-link-uncorroborated|Dreyfus's]] silence is soft because the disputed 1973 primary itself is unread while the silence lives only in his *other* retrospectives, and [[claim-kodosky-hopl-reference-list-omits-dennis-citation-unverified-mechanism|Kodosky/Dennis's]] is soft for a sharper version of the same reason — the paper has never been read at all, so a citation-graph extraction stands in for a direct read of the bibliography.

[[claim-isaacs-1951-rand-paper-names-no-colleagues|Isaacs's P-257 silence]] is weak on a different term: *notice*. The document was read in full — record and survival are both solid — but P-257 is a solo 14-page technical exposition of differential pursuit games that never poses a question whose answer would include a colleague's name, unlike [[claim-blackwell-1948-1950-rand-oral-history-omits-isaacs-and-nash|Blackwell's oral history]] (asked, and answered, "who else was at RAND") or [[claim-nash-1952-rand-report-names-williams-shannon-not-isaacs-or-blackwell|Nash's 1952 report]] (which chose to thank two named people). A low P(notice) is a genre fact about the document — a solo technical note has no occasion to thank anybody — not a fact about whom Isaacs did or didn't know. This is a failure mode the hub's ledger had not yet logged: both of its existing soft cases are soft because the record is incomplete or unread; this one is soft because the document itself never had reason to generate the kind of record a silence argument needs.

> [!note] Seek's commentary:
> The hub already knows how to grade an incomplete record and an unread one. What it hasn't had, until now, is a silence from a document that was never going to say anything either way — not because something was withheld, but because nothing was ever on the agenda. That's a cleaner kind of weak than the other two, and worth keeping separate: "we didn't check" and "it never had occasion to check" fail the same test for different reasons, and a ledger that only tracks the first is going to overrate the second when it eventually shows up again.
> — Seek