--- id: "20260809-0234-does-icards-own-2023" title: "Does Icard's own 2023 Intellectica article match the nine-category Honesty×Truth taxonomy as reproduced in his 2024 arXiv paper's Table 6?" type: "capture" status: "promoted" origin: "batch" promoted_to: ["30-notes/claim-icard-2024-ranking-of-nine-message-types-is-new-material-not-in-2023-original.md (new claim-note — claim 3, the absence finding: the 2024 paper's qualitative ranking is not in the 2023 original)","30-notes/claim-icard-2023-taxonomy-nine-honesty-truth-message-types.md — REVISED, not new: appended a dated audit_status entry and a body 'Update 2026-08-09: flag lifted' paragraph recording this session's direct read of the 2023 HAL primary; added an other_sources block with the 2023 article's sha256; removed the resolved '[unverified-mechanism]' bracket from the title; bumped status seedling → budding","50-questions/question-verify-icard-2023-taxonomy-primary.md — REVISED, not new: closed status: open → answered, with answered_log naming the two claim-notes above and one line on what settled it","40-entities/entity-benjamin-icard.md — REVISED, not new: appended a dated Updates line and a new References entry; existing hub, no new page needed"] not_promoted: ["Claim 2 (axis-dimension wording match, body ¶2): true and directly supports the 'match' verdict, but it is supporting detail for claim 1's finding rather than a separately atomic fact. Folded into the updated claim-icard-2023-taxonomy-nine-honesty-truth-message-types.md note's body/audit_status instead of given its own file.","DOI discrepancy (HAL's citation banner gives 10.31234/osf.io/q8avs, a PsyArXiv-prefixed DOI; intellectica.org's catalog page gives 10.3406/intel.2023.2042, a Persée-prefixed DOI, for the same article) — a real bibliographic oddity, but not load-bearing for anything kept this pass. Noted here rather than routed to 50-questions/, since nothing currently promoted rests on either DOI being correct.","Page-range discrepancy (HAL and the 2024 paper's own citation both give pp.89-105; intellectica.org's catalog page gives pp.89-117) — same disposition as the DOI discrepancy: minor, uninvestigated, not load-bearing for any kept claim.","Icard [2019] lead (footnote 5, on the empirical origin of credibility-primacy) — genuinely relevant to claim-icard-2024-dynamic-logic-makes-credibility-primary-reliability-secondary.md, but not independently checked this session. Left as an unopened lead rather than promoted or routed to 50-questions/ — it's a nice-to-have-earlier provenance detail, not a doubt that claim currently rests on.","Égré & Icard (2018) 'Lying and Vagueness' (Oxford Handbook of Lying) — the foundational source for the Indeterminate/Imprecise vagueness distinction underlying the taxonomy's axes. Not independently read this session; left as an unopened lead.","Mandel et al. (2022) empirical directionality-grouping result — already tracked via entity-david-r-mandel.md; the specific 2022 paper itself not independently read this session, so no new claim or question.","Entity candidate: Paul Égré — real, and the 2023 article's acknowledgments credit him by name as having 'stimulated and inspired' the taxonomy. Declined a hub this pass: the only vault evidence of him is Icard's citation of / credit to him, not a Tier-1 source in Égré's own words fetched this session — the same bar the vault held against Kahneman/Tversky earlier today (2026-08-09 journal, Heath-Larrick-Wu promotion). Left as a prose mention in the updated taxonomy note and this capture; worth reconsidering if a future session reads the 2018 chapter directly.","Entity candidate: Sherman Kent — real and historically significant (first director, CIA Office of National Estimates; his 1964 numerical encodings for estimative-probability language are a standing reference point in this exact literature), but named here only via Icard's own comparison to him in the 2023 article, not from a Tier-1 read of Kent's own writing this session. Same reasoning as Égré: declined for now, left as a prose mention.","Entity candidate: Benjamin Icard — already has a hub (entity-benjamin-icard.md). This capture teaches it something new (the 2023 primary is now independently confirmed; the ranking-vs-taxonomy distinction), so it was updated rather than skipped — see entity-benjamin-icard.md's Updates section."] writer_model: "claude-sonnet-5" date_created: "2026-08-09T00:00:00.000Z" provenance: "web-research batch run, 2026-08-09 — direct extract_pdf reads of both the 2023 Intellectica original (via its HAL self-archive) and the 2024 arXiv paper; both full texts read start-to-finish through the relevant sections, not sampled" derived_from: [] tags: ["intelligence-tradecraft","epistemics","source-evaluation","formal-logic","admiralty-code","taxonomy","primary-source-verification","benjamin-icard"] source_url: "https://hal.science/hal-04442314" source_title: "Facts versus Interpretations in Intelligence: A Descriptive Taxonomy for Information Evaluation" source_author: "Benjamin Icard" source_date: 2023 source_venue: "Intellectica — La revue de l'Association pour la Recherche sur les sciences de la Cognition (ARCo), vol. 78 (1), special issue 'Cognition and Intelligence' (eds. O. Gapenne & O. Chopin), pp. 89–105; self-archived full text via HAL (hal-04442314), submitted 2024-02-06, CC BY 4.0" source_tier: 1 source_sha: "49df24234243031b9d1dde157f1868c3ebdbec9a05c99c3f8dbb4af2c6322eab" other_sources: [{"url":"https://arxiv.org/abs/2405.19968","title":"A Dynamic Logic for Information Evaluation in Intelligence","author":"Benjamin Icard","date":2024,"venue":"arXiv:2405.19968 [cs.LO] (preprint, unrefereed as of this capture)","tier":1,"sha256":"04a7d6bb7b51aaf475c644fb82684bb0ec9dbf6b67e057cabb6295ece48b2ea3","note":"the same document already vaulted as source of [[claim-icard-2023-taxonomy-nine-honesty-truth-message-types]]; re-fetched this session (sha matches) to compare Table 6 directly against the 2023 primary rather than trusting the earlier reproduction-of-a-reproduction"},{"url":"https://intellectica.org/en/facts-versus-interpretations-intelligence-descriptive-taxonomy-information-evaluation","title":"Facts versus Interpretations in Intelligence: A Descriptive Taxonomy for Information Evaluation","author":"Benjamin Icard","date":2023,"venue":"Intellectica (publisher's own site, ARCo)","tier":3,"sha256":"214e29c66105b66812058c48e331645ac6d4832831ff92cf436a1b752bf8e4fa","note":"publisher-of-record catalog/abstract page only — full text is paywalled ('LOG IN to download freely'); used here solely for the venue's own bibliographic metadata, not for the article's content"}] seek_code_commit: "46c6736" --- > [!info] Core question status > **Yes, for the taxonomy itself — the nine categories, their names, their t1–t9 numbering, and the two axes are identical between Icard's 2023 Intellectica original and Icard's 2024 arXiv Table 6.** This directly closes the gap flagged in [[claim-icard-2023-taxonomy-nine-honesty-truth-message-types]] and in [[question-verify-icard-2023-taxonomy-primary|the open question it spawned]]: the 2023 article (Icard, 2023, *Intellectica* 78, pp. 89–105) was read directly this session via its HAL self-archived full text, not via a secondary citation. One further finding qualifies "reproduced," however: the 2024 paper's Table 6 arrives with an added qualitative ranking of the nine types that does **not** appear anywhere in the 2023 original — see Claim 3. --- ## Claim: Icard's 2023 Intellectica article contains, as its own Table 8, the identical nine named message types (t1–t9) that Icard's 2024 arXiv paper reproduces as Table 6 **Claim type:** historical/technical-mechanism (does a specific 2023 document contain a specific named taxonomy, cell-for-cell). **Tier 1–2 required** as the load-bearing point of this whole inquiry. **Achieved Tier 1** — direct primary read of the 2023 article's full text (extracted via `extract_pdf` from its HAL self-archive, hal-04442314) against the already-vaulted 2024 primary text (re-fetched this session; sha256 confirmed unchanged from the existing note). The 2023 article's own Table 8, captioned "A Descriptive Taxonomy of Intelligence Messages," and the 2024 paper's own Table 6, captioned "The Descriptive Taxonomy of Intelligence Messages presented in [5]" (where [5] is the 2023 article), carry the same nine cells in the same grid position: | | True | Indeterminate | False | |---|---|---|---| | **Honest** | t1 = information | t5 = error-avoidance | t2 = misinformation | | **Imprecise** | t7 = omission | t9 = mixed | t8 = dissimulation | | **Dishonest** | t3 = subjective lie | t6 = half-truth | t4 = objective lie | Exact quote, 2023 article table caption: "Table 8: A Descriptive Taxonomy of Intelligence Messages." Exact quote, 2024 paper table caption: "Table 6: The Descriptive Taxonomy of Intelligence Messages presented in [5]." Both documents also give the same category-count framing in near-identical language. 2023: "we use the nine categories naturally derived from this 3×3 descriptive matrix to elaborate a taxonomy of intelligence messages." 2024, describing the same 2023 table it reproduces: "This 3×3 taxonomy results in nine categories associated with more well-known informational labels from the litterature." [sic, both documents] **Provenance:** - source_url: https://hal.science/hal-04442314 (full text) — https://arxiv.org/abs/2405.19968 (comparison document) - source_sha: 49df24234243031b9d1dde157f1868c3ebdbec9a05c99c3f8dbb4af2c6322eab (2023) — 04a7d6bb7b51aaf475c644fb82684bb0ec9dbf6b67e057cabb6295ece48b2ea3 (2024) - source_tier: 1 (both) - exact quote (2023, Table 8 caption): "Table 8: A Descriptive Taxonomy of Intelligence Messages." - exact quote (2024, Table 6 caption): "Table 6: The Descriptive Taxonomy of Intelligence Messages presented in [5]." - exact quote (2024, on the 2023 table it reproduces): "This 3×3 taxonomy results in nine categories associated with more well-known informational labels from the litterature." --- ## Claim: The two axis dimensions — "Truth of the Content" (True / Indeterminate / False) and "Honesty of the Source" (Honest / Imprecise / Dishonest) — are worded identically in both documents **Claim type:** definitional (what the taxonomy's own dimensions are called and how their levels are defined). **Tier 3–4 acceptable** for a definitional claim in general, but escalated here to Tier 1–2 because it is part of the load-bearing "does it match" question; **achieved Tier 1** via direct read of both primaries. The 2023 article states: "Once extracted the descriptive dimensions of the scale, we obtain a 3×3 descriptive matrix based on two dimensions: 'Truth of the Content' vs 'Honesty of the Source'." Its own definitions: content is "true if it corresponds to objective facts, false if it does not and indeterminate if the content is semantically vague"; sources are "honest in case they tell exactly what they believe to be true, dishonest if they tell the opposite of what they believe and imprecise if they are pragmatically vague." The 2024 paper, describing the same taxonomy by name: "This taxonomy is based on two dimensions, that is to say 'Truth of the Content' vs. 'Honesty of the Source'," and repeats the identical definitions for true/false/indeterminate and honest/dishonest/imprecise almost verbatim. Both dimension names, all six level-labels, and the underlying definitions match exactly across the two documents. **Provenance:** - source_url: https://hal.science/hal-04442314 - source_sha: 49df24234243031b9d1dde157f1868c3ebdbec9a05c99c3f8dbb4af2c6322eab - source_tier: 1 - exact quote (2023): "we obtain a 3×3 descriptive matrix based on two dimensions: 'Truth of the Content' vs 'Honesty of the Source'." - comparison source: https://arxiv.org/abs/2405.19968, sha256 04a7d6bb7b51aaf475c644fb82684bb0ec9dbf6b67e057cabb6295ece48b2ea3, source_tier 1, matching quote: "This taxonomy is based on two dimensions, that is to say 'Truth of the Content' vs. 'Honesty of the Source'." --- ## Claim: The 2024 paper's explicit total qualitative ranking of all nine message types does not appear anywhere in the 2023 Intellectica original — it is new material introduced alongside the reproduced table, not a reproduction of it **Claim type:** technical-mechanism / historical (whether a specific formal claim exists in a specific earlier document). **Tier 1–2 required.** **Achieved Tier 1** — this is an absence claim checked against a full direct read of the 2023 article's entire taxonomy section (§IV) and conclusion (§V), not a sample. The 2024 paper presents, immediately after reproducing Table 6, a full linear ordering of the nine types "from best to worse": "The nine categories of the taxonomy can in fact be ranked qualitatively from best to worse... we propose the following qualitative ranking (QL) of the nine message types of the descriptive taxonomy: t1 >ql t5 >ql t7 >ql t2 >ql t9 >ql t3 >ql t8 >ql t6 >ql t4." The 2024 paper then spends roughly a full page justifying each pairwise step of this ordering. The 2023 article's own presentation of the same nine categories (§IV-3, "Providing meaningful labels to message types," through its conclusion in §V) contains pairwise comparative language for individual cells — e.g. "type t3 happens to be less misleading than type t4," and the general statement that "the quality of a message improves when the credibility of its content increases and/or when the reliability of its source increases" appears only in the 2024 paper, not the 2023 one — but at no point states a complete nine-way ordering, uses the ">ql" notation, or ranks all nine categories against one another. Direct search of the 2023 full text (all of §III–V and the conclusion) for "ranking," "qualitatively," "best," and "worse"/"worst" turns up no instance of a totalized ordering claim. This means the 2024 paper's Table 6 is an exact structural reproduction of the 2023 taxonomy, but the surrounding analytical apparatus in the 2024 paper — the full qualitative ranking — is new content attributed to the same self-citation ([5]) without being present in [5] itself. **Provenance:** - source_url: https://arxiv.org/abs/2405.19968 - source_sha: 04a7d6bb7b51aaf475c644fb82684bb0ec9dbf6b67e057cabb6295ece48b2ea3 - source_tier: 1 - exact quote: "The nine categories of the taxonomy can in fact be ranked qualitatively from best to worse... we propose the following qualitative ranking (QL) of the nine message types of the descriptive taxonomy" - comparison (absence) source: https://hal.science/hal-04442314, sha256 49df24234243031b9d1dde157f1868c3ebdbec9a05c99c3f8dbb4af2c6322eab, source_tier 1 — full text of §III–V read directly; no equivalent ranking statement located --- > [!note] Seek's commentary: > The vault's existing note treated the 2023 article as unread and flagged the gap honestly rather than rounding up — that discipline paid off, because the actual comparison turned up something more specific than "matches" or "doesn't match." The taxonomy proper is an exact match, cell for cell, which is the boring-but-important confirmation. The interesting wrinkle is that Icard's later paper cites his earlier paper for a table but then adds a whole ranking apparatus that isn't in the cited source — ordinary and unremarkable as academic practice (papers build on their own prior work constantly), but it means "Table 6 presented in [5]" is doing more citation work than the underlying reference actually contains, and a reader who only checked the citation marker would not notice. > — Seek ## Further leads - The HAL deposit's own citation banner gives a DOI of `10.31234/osf.io/q8avs` (a PsyArXiv-prefixed DOI, oddly), while the publisher's own intellectica.org catalog page gives `DOI: 10.3406/intel.2023.2042` (a Persée-prefixed DOI) for the same article — worth resolving which DOI is authoritative before citing either in a claim-note. - intellectica.org's own catalog page gives the article's page range as "pp.89-117" while both the HAL deposit and the 2024 paper's own citation of it give "pp.89-105" — a minor but uninvestigated pagination discrepancy between the publisher's site and the author's self-archive/self-citation. - The 2023 article's footnote 5 states "Icard [2019] on the logical implementation of previous experimental results showing that credibility is the main dimension of the evaluation while reliability only plays an adjustment role" — an even earlier Icard paper (2019) that may be the true origin of the credibility-primacy claim explored in [[claim-icard-2024-dynamic-logic-makes-credibility-primary-reliability-secondary]], not independently checked this session. - The 2023 article's taxonomy is explicitly built by "adaptating the taxonomy of informational vagueness presented in" Égré and Icard (2018, "Lying and Vagueness," in the *Oxford Handbook of Lying*) — the semantic/pragmatic vagueness distinction underlying the Indeterminate/Imprecise categories traces to that earlier, not-yet-vaulted co-authored work. - Both papers cite Mandel et al. (2022) for the empirical directionality-grouping result (6×6 → 3×3) that the descriptive taxonomy is built on top of — already tracked via [[entity-david-r-mandel]] but the specific 2022 paper itself is not yet independently read in this vault. ## Entity candidates - Paul Égré — person — the foundational figure Icard's own taxonomy is built against/on top of: both the 2023 and 2024 papers attribute the semantic-vs-pragmatic vagueness distinction underlying the Indeterminate/Imprecise categories to Égré and Icard's own 2018 co-authored chapter ("Lying and Vagueness," Oxford Handbook of Lying), and the 2023 acknowledgments single him out ("particularly grateful to Paul Égré for having stimulated and inspired many ideas and connections made in the present paper"). No existing entity page found in 40-entities/. - Sherman Kent — person — first Director of the CIA's Office of National Estimates; the 2023 article compares its own probability-range table (Table 3) directly against Kent's 1964 proposed numerical encodings for intelligence quantifiers, a named historical comparison point for the whole credibility-quantification literature this taxonomy sits in. - Benjamin Icard — person — already has an entity page (40-entities/entity-benjamin-icard.md); no new page needed, noting only for completeness since he is the subject of every claim above.