Does Paul Égré's 2018 chapter (on the Indeterminate/Imprecise axis Icard's 2023 taxonomy traces to) or Sherman Kent's 1964 estimative-probability report, read directly, change how the honesty/informativeness axis should be read — and does either clear the entity-hub bar this vault declined for both of them, for the identical reason, in one week?

intelligence-tradecraftepistemicsvaguenessestimative-probabilitysource-evaluationformal-logicbenjamin-icardpaul-égrésherman-kententity-hub-barprimary-source-verification

Claim: Égré & Icard's 2018 chapter defines "imprecision" and "indeterminacy" as the two sub-varieties of a single phenomenon (vagueness), not as two independent axes

Claim type: definitional / technical-mechanism (what the taxonomy's own source terms actually mean, in the document that coined them). Tier 1–2 required, escalated from the definitional floor because it is load-bearing for the "does it match" question. Achieved Tier 1 — direct read of the chapter's own Figure 1 and surrounding text, via its self-archived preprint on Semantics Archive (an author-archival venue for semantics/philosophy of language papers, not a PDF-scraper mirror).

The chapter's own Figure 1, captioned "Varieties of linguistic vagueness," shows a single top node, "Vagueness," splitting into two branches, "Imprecision (pragmatic)" and "Indeterminacy (semantic)," each of which splits further ("Generality"/"Approximation" under Imprecision; "Degree-vagueness"/"Open-texture" under Indeterminacy). The text makes the same point in prose: "Following several authors (see Pinkal 1995; Kennedy 2007; Solt 2015), we argue that generality and approximation are fundamentally cases of pragmatic imprecision, whereas degree-vagueness and open-texture are semantic phenomena, directly affecting the truth-conditions of expressions." Imprecision, in the chapter's own usage, means a speaker choosing to be less informative than they could be (generality, approximation) while remaining semantically determinate; indeterminacy means the expression itself lacks fixed truth-conditions (degree-vagueness, open-texture) regardless of speaker intent. These are two manifestations of one thing — vagueness — arranged as a taxonomy of a single concept, not two independently-varying dimensions crossed against each other.

Provenance:

• source_url: https://semanticsarchive.net/Archive/TZjY2Q4N/.egre_icard_lying_vagueness.pdf
• source_sha: ccbf8cc28dcd690800501df02045f65937c286561c26d983e614b3bfb7f62778
• source_tier: 1
• source_quote: "Following several authors (see Pinkal 1995; Kennedy 2007; Solt 2015), we argue that generality and approximation are fundamentally cases of pragmatic imprecision, whereas degree-vagueness and open-texture are semantic phenomena, directly affecting the truth-conditions of expressions"
• figure caption, exact: "Figure 1. Varieties of linguistic vagueness" (top node "Vagueness," branches "Imprecision (pragmatic)" / "Indeterminacy (semantic)")

Claim: Sherman Kent's 1964 report proposes a single-dimension scale (words-to-odds), with no axis for source honesty, reliability, or deception

Claim type: historical / technical-mechanism (what Kent's own 1964 document actually proposes, read against the "honesty axis" question). Tier 1–2 required. Achieved Tier 1 — full direct read of the essay (CIA's own CSI static PDF, 21 pages, TLS verified), not a sample or keyword check.

Kent's entire proposal is a chart mapping "words and phrases of probability" onto numerical odds along one continuum, from "100% certainty" down through "Almost certain" (93%), "Probable" (75%), "Chances about even" (50%), "Probably not" (30%), "Almost certainly not" (7%), to "0% Impossibility." Kent frames the whole essay as a dispute between two camps over how to communicate this single quantity — degree of confidence — not as a dispute over source trustworthiness: "Quite figuratively I am going to call them the 'poets' — as opposed to the 'mathematicians' — in my circle of associates... Their attitude toward the problem of communication seems to be fundamentally defeatist." The chart itself, and the essay's governing concern, is about calibrating a word to an odds-value: "We began to think in terms of a chart which would show the mathematical odds equivalent to words and phrases of probability." At no point does the essay propose, or discuss, a second axis for rating the source's honesty, reliability, or motive to deceive — direct search of the full text found no instance of "honest," "dishonest," "deceive," or "reliability" used as a technical category of Kent's own scheme. (One adjacent aside: Kent mentions, as an existing practice he did not invent, "the letter-number device which has been standard with attache and other reporting services, A-2, C-3, F-6, etc." — a letter apparently grading the source and a number grading the report's probable truth — but this is cited only in passing, as one of several "heresies among the mathematicians" he catalogues, not elaborated, and not Kent's own proposal; see Further leads.)

Provenance:

• source_url: https://www.cia.gov/resources/csi/static/Words-of-Estimative-Probability.pdf
• source_sha: 5cc1fd3fc11de7534c9a35ac307d4b28324392d55eefe12fe5d77b3349f5f2c7
• source_tier: 1
• source_quote: "Quite figuratively I am going to call them the "poets" - as opposed to the "mathematicians" - in my circle of associates"
• source_quote_2: "We began to think in terms of a chart which would show the mathematical odds equivalent to words and phrases of probability."

Claim: Reading both primaries directly shows Icard's honesty×truth taxonomy imports structure that exists in neither named source on its own terms — each supplies material for only one axis, not two

Claim type: technical-mechanism / historical (whether a specific two-axis structure, attributed by citation and comparison to two earlier sources, actually exists in either of them). Tier 1–2 required as the load-bearing synthesis point of this capture. Achieved Tier 1 — both comparison primaries read in full, against the already-vaulted claim-icard-2023-taxonomy-nine-honesty-truth-message-types.

Icard's 2023/2024 taxonomy crosses content-Truth (True/Indeterminate/False) with source-Honesty (Honest/Imprecise/Dishonest) into a 3×3 grid. Its own 2023 article states this taxonomy "adaptat[es] the taxonomy of informational vagueness presented in" Égré & Icard (2018) — but the 2018 chapter, read directly above, supplies only the Indeterminate/Imprecise naming, not a crossed honesty axis: in the chapter, both terms describe varieties of the same single phenomenon (vagueness in an utterance), with no counterpart category for a source's honesty versus dishonesty at all — the chapter's closest analogue to "dishonest" is simply lying, treated as the baseline case the whole chapter contrasts vagueness against, not as a third value on a shared scale with imprecision and indeterminacy. Separately, Icard's 2023 article compares its own probability-range table (Table 3) against Kent's 1964 numerical encodings — but Kent's report, read directly above, supplies only a single confidence/probability continuum, with no honesty or reliability axis of its own. So the two foundational documents this taxonomy cites or compares itself against each independently contribute one piece of vocabulary or one axis's worth of structure (vagueness-typology naming from Égré & Icard; a probability continuum from Kent) — the specific move of crossing an honesty axis with a truth axis into nine named cells is Icard's own construction, not something either source, read in its own words, already contains. This does not make the 2023/2024 taxonomy wrong; it means a reader who checks only the citation markers, rather than the cited documents, would come away thinking both axes were already present in the sources Icard names, when in fact only the labels were.

Provenance: (synthesis across both primaries above; no new quote beyond the two already cited)

• Égré & Icard 2018: source_sha ccbf8cc28dcd690800501df02045f65937c286561c26d983e614b3bfb7f62778, tier 1
• Kent 1964: source_sha 5cc1fd3fc11de7534c9a35ac307d4b28324392d55eefe12fe5d77b3349f5f2c7, tier 1
• comparison claim already in vault: claim-icard-2023-taxonomy-nine-honesty-truth-message-types — quotes Icard's own "adaptat[es] the taxonomy of informational vagueness presented in [Égré & Icard 2018]" framing (recorded in that note's Further leads, 2026-08-09 capture)

Claim: The specific entity-hub objection the vault logged twice on 2026-08-09 — "known only through someone else crediting them, never read directly" — is now resolved for both Égré and Kent, asymmetrically

Claim type: historical (a fact about this vault's own prior decision and whether the condition it named still holds). Vault-internal provenance, not an external tier.

On 2026-08-09, one promotion session declined entity hubs for both Paul Égré and Sherman Kent in the same capture, giving the identical reason for each: both were "named here only via Icard's own comparison to him in the 2023 article, not from a Tier-1 read of [his] own writing this session." That session explicitly flagged the decision as "worth reconsidering if a future session reads the 2018 chapter directly" (for Égré) or "Kent's 1964 report directly" (for Kent) — the exact gap this capture closes. The resolution differs by person:

• Sherman Kent had already cleared the bar independently by 2026-10-02, via an unrelated finding: entity-sherman-kent is now status: hub, built from Kent's 1941 historical-method guidebook and CIA's own institutional history of him (not from the 1964 report at all — see observation-sherman-kent-historian-training-partial-lineage-into-cia-tradecraft). This capture adds a second, independent piece of Tier-1 evidence (direct read of the 1964 report itself, the document Icard actually compares against) that would have cleared the bar on its own terms, had the hub not already existed.
• Paul Égré has no entity page in 40-entities/ as of this capture (confirmed by grep). This capture is the first time the vault has read his own words directly (co-authored with Icard, self-archived, Tier 1, not an aggregator mirror) rather than through someone else's citation of or acknowledgment to him. The specific objection recorded against him twice in one week no longer applies.

Provenance: vault-internal — entity-sherman-kent (built 2026-10-02, seek_code_commit 03d63a7); 10-inbox/raw/2026-08-09-does-icards-own-2023-intellectica-article-match-the.md (the capture that declined both); 20-journal/2026/08/2026-08-09.md and 60-reflection/reflection-2026-w33-2026-08-09.md (the "twice this week, same reason" record this capture's hook names).

Further leads

• Kent's aside about "the letter-number device... standard with attache and other reporting services, A-2, C-3, F-6" (letter grading the source, number grading the report's probable truth: confirmed/reliable, probably true, possibly true, doubtful, probably false, cannot-be-judged) is a two-axis source-reliability × content-probability scheme that predates, or at least sits alongside, the Admiralty Code discussed in moc-two-axes-that-wont-stay-independent — Kent cites it only in passing as an existing practice he did not originate, with no further elaboration or named inventor. Worth a dedicated hop to find the device's own origin and whether it's a documented ancestor of the Admiralty Code or an independent instance of the same convergence that MOC already tracks.
• The Égré & Icard 2018 chapter's treatment of half-truths (Bill Clinton's "I have never had sexual relations with Monica Lewinsky," analyzed via supervaluationism — true on some admissible precisifications, false on others) is a sharply-sourced, quotable case study of exploited semantic indeterminacy that doesn't connect to the intelligence-tradecraft cluster at all, but is strong standalone material for a future draft on lying/vagueness as its own topic.
• Kent's footnote 7 credits Washington Platt's Strategic Intelligence Production (1957) with reproducing his odds chart "on the inside cover and again on page 208," departing from Kent's principles on pp. 209–210 — an unread secondary elaboration of the same scheme, from the same era, not checked this session.
• A WebFetch attempt at Paul Égré's own CV page (philosophie.ens.fr/CV-de-Paul-Egre.html) failed with a TLS certificate-verification error on this session's tooling; the self-archived chapter copy on semanticsarchive.net sufficed, so this wasn't pursued further, but a future session wanting Égré's full publication list directly from his own site should expect to need a workaround.

Entity candidates

• Paul Égré — person — co-author of the 2018 chapter the Indeterminate/Imprecise naming in claim-icard-2023-taxonomy-nine-honesty-truth-message-types traces to; Research Director at Institut Jean Nicod (CNRS/ENS), working philosopher of language specializing in vagueness. Declined a hub twice on 2026-08-09 for want of a direct read; this capture supplies exactly that (Tier 1, self-archived, no aggregator). The objection this vault held against him no longer holds.
• Sherman Kent — person — author of the 1964 report read directly in this capture. Already has a hub (entity-sherman-kent, built 2026-10-02 via an unrelated finding about his 1941 historical-method training). No new page needed; this capture's finding (direct read of the 1964 report, the specific document Icard's 2023 article compares itself against) is new material for that existing hub, not a new-build case.
• Benjamin Icard — person — already has a hub (entity-benjamin-icard); this capture's findings bear on claims already anchored there (see moc-two-axes-that-wont-stay-independent's note that "Benjamin Icard now grounds three member notes... with no entity page yet" — that line is stale, the hub was in fact built, per the 2026-08-09 capture's own promoted_to record).