---
title: "Icard's 2023 taxonomy crosses source Honesty with content Truth to yield nine named message types, from information to objective lie"
type: "claim"
status: "budding"
writer_model: "claude-sonnet-5"
audit_status: "capture-verified for what Icard's 2024 paper itself says (Tier 1, direct fetch, source_quote/table description verified against the fetched text at capture time); [unverified-mechanism] for the taxonomy's actual content, which is Icard's own 2023 Intellectica article (Icard, 'Facts versus interpretations in intelligence: A descriptive taxonomy for information evaluation,' Intellectica 78, 89–105), reproduced as Table 6 within the 2024 paper but not independently fetched or read at its own primary this session. — CORRECTION 2026-08-08 (scheduled cross-model audit, claude-opus-5): the previous source_quote ('a descriptive taxonomy crossing source Honesty (Honest/Imprecise/Dishonest) with content Truth (True/Indeterminate/False), yielding named message types') was NOT verbatim in the 2024 paper — it was a capture-time paraphrase occupying a field reserved for quoted text, the defect class the quote-provenance rule in sources.md exists to catch. Replaced with a genuinely verbatim sentence from the paper (§3.1); the paraphrase's substance is retained as prose in the body, where it belongs. Independently re-verified this pass by extract_pdf against arxiv.org/pdf/2405.19968 (sha256 04a7d6bb… — matches the source_sha recorded here, i.e. the recorded sha covers the PDF, not the /abs/ landing page named in source_url): Table 6 ('The Descriptive Taxonomy of Intelligence Messages presented in [5]') carries all nine named cells exactly as this note lists them, on axes 'Truth of the Content' (True/Indeterminate/False) × 'Honesty of the Source' (Honest/Imprecise/Dishonest); reference [5] resolves verbatim to 'B. Icard, Facts versus interpretations in intelligence: A descriptive taxonomy for information evaluation, in: O. Gapenne, O. Chopin (Eds.), Intellectica, Cognition and Intelligence, volume 78, 2023, pp. 89–105.' The [unverified-mechanism] flag STANDS: the 2023 Intellectica article itself remains unread, and this audit re-checked the reproduction, not the original. — 2026-08-09 (promotion of 10-inbox/raw/2026-08-09-does-icards-own-2023-intellectica-article-match-the.md, writer claude-sonnet-5): flag LIFTED. The 2023 Intellectica original was read directly this session via its HAL self-archive (hal-04442314, CC BY 4.0, extract_pdf, full text start-to-finish through §III–V and the conclusion, not sampled). Its own Table 8 ('A Descriptive Taxonomy of Intelligence Messages') carries the identical nine cells, t1–t9 numbering, and two axis names as the 2024 paper's Table 6, with near-identical level-definitions in both documents. source_sha for the 2023 primary recorded below via other_sources. One qualification survives the check: the 2024 paper's qualitative best-to-worse ranking of all nine types is new material not present anywhere in the 2023 original — see [[claim-icard-2024-ranking-of-nine-message-types-is-new-material-not-in-2023-original]]. [[question-verify-icard-2023-taxonomy-primary]] is closed as answered. — 2026-08-10, scheduled cross-model audit (auditor claude-opus-5; writer claude-sonnet-5): re-verified independently against BOTH primaries, neither taken on a prior audit's word. 2024 preprint re-fetched via extract_pdf (52 pp, tls verified), sha256 04a7d6bb… matches; 2023 Intellectica article re-fetched from its HAL self-archive (24 pp, tls verified), sha256 49df2423… matches the other_sources digest recorded below. Table 6 of the 2024 paper (p. 14) carries all nine cells exactly as this note's table lists them, on axes 'Truth of the Content' × 'Honesty of the Source', and Table 8 of the 2023 article (p. 17, captioned 'A Descriptive Taxonomy of Intelligence Messages') is cell-for-cell identical — same nine labels, same t1–t9 numbering, same axis names and level-definitions. source_quote verified VERBATIM in §2.2.3 of the 2024 paper, including its 'litterature' misspelling. The t9 gloss is verbatim ('Type t9 is a mixed case in which both the content and source are vague'). Reference [5] resolves verbatim to the Intellectica citation, and the 2023 article's own front matter independently confirms it: Intellectica 78(1), pp. 89–105, 2023, CC BY 4.0. The ranking sentence and its ordering are verbatim as quoted, and the ordering the body describes checks against the paper's own justification (t7 'omission' above t2 'misinformation'). No claim changed; the flag-lift recorded above is confirmed correct. ONE RECEIPT ADDED: the 2023 article is refereed — its acknowledgments thank 'two anonymous referees' — which matters for the cap below, since it means the taxonomy's own primary is not an unrefereed source even though the note's source_url points at one. CONCENTRATION COUNT, not previously recorded here: three 30-notes files now rest on arXiv:2405.19968 (this note, claim-icard-2024-dynamic-logic-makes-credibility-primary-reliability-secondary, claim-icard-2024-ranking-of-nine-message-types-is-new-material-not-in-2023-original). That preprint is unrefereed, so the sources.md single-source concentration cap applies and the branch sits AT its ceiling of three: not breached, nothing downgraded, but the next finding from the preprint goes to 50-questions/ as a corroboration question rather than becoming a fourth claim-note. Icard's own 2023 article does not discharge the cap — same author, so it is the same source wearing different clothes — though this particular note is the least exposed of the three, since its load-bearing table is now anchored in the refereed 2023 original as well."
audits: ["2026-08-08 claude-opus-5","2026-08-10 claude-opus-5"]
source_url: "https://arxiv.org/abs/2405.19968"
source_title: "A Dynamic Logic for Information Evaluation in Intelligence (reproducing, as its Table 6, the taxonomy from Icard 2023)"
source_author: "Benjamin Icard"
source_date: 2024
source_venue: "arXiv:2405.19968 [cs.LO] (preprint; taxonomy itself originates in Icard's 2023 Intellectica 78 article)"
source_tier: 1
source_sha: "04a7d6bb7b51aaf475c644fb82684bb0ec9dbf6b67e057cabb6295ece48b2ea3"
source_quote: "This 3×3 taxonomy results in nine categories associated with more well-known informational labels from the litterature."
other_sources: [{"url":"https://hal.science/hal-04442314","title":"Facts versus Interpretations in Intelligence: A Descriptive Taxonomy for Information Evaluation","author":"Benjamin Icard","date":2023,"venue":"Intellectica — La revue de l'ARCo, vol. 78(1), pp. 89–105 (self-archived via HAL, hal-04442314, CC BY 4.0)","tier":1,"sha256":"49df24234243031b9d1dde157f1868c3ebdbec9a05c99c3f8dbb4af2c6322eab","note":"the 2023 primary this note's taxonomy actually originates in — the article, not the 2024 self-citation of it. Read directly 2026-08-09, confirming its own Table 8 matches the already-cited 2024 Table 6 cell-for-cell."}]
provenance: "Promotion from 10-inbox/raw/2026-08-07-hop-icard-dynamic-logic-abandons-independence.md, 2026-08-07"
origin: "batch"
derived_from: ["10-inbox/raw/2026-08-07-hop-icard-dynamic-logic-abandons-independence.md"]
date_created: "2026-08-07T00:00:00.000Z"
tags: ["intelligence-tradecraft","epistemics","source-evaluation","formal-logic","admiralty-code","taxonomy","primary-source-verification"]
seek_code_commit: "649b1a4"
---


The "nine categorizations" that the vault's [[claim-kelly-et-al-propose-richer-joint-matrix-not-axis-collapse|existing note]] already knew about — via Kelly et al.'s (2025) gloss, itself a citation of a citation — trace to a specific, named scheme: a 3×3 taxonomy crossing source **Honesty** (Honest / Imprecise / Dishonest) with content **Truth** (True / Indeterminate / False). Icard's 2024 dynamic-logic paper reproduces this scheme, attributed to his own earlier work (reference [5], Icard 2023), as its Table 6, captioned "The Descriptive Taxonomy of Intelligence Messages presented in [5]." The paper's own axis names are "Honesty of the Source" and "Truth of the Content," and the nine resulting cells are named, not merely numbered:

| | True | Indeterminate | False |
|---|---|---|---|
| **Honest** | information (t1) | error-avoidance (t5) | misinformation (t2) |
| **Imprecise** | omission (t7) | mixed (t9) | dissimulation (t8) |
| **Dishonest** | subjective lie (t3) | half-truth (t6) | objective lie (t4) |

The cell numbering is not the ranking. Icard orders the nine "qualitatively from best to worse" separately, as t1 > t5 > t7 > t2 > t9 > t3 > t8 > t6 > t4 — that is, *information* best and *objective lie* worst, with *omission* (an imprecise source telling a true content) ranking above *misinformation* (an honest source conveying a false one). The middle column is the vagueness column: t9 is "a mixed case in which both the content and source are vague."

This taxonomy is a distinct formal object from [[claim-icard-2024-dynamic-logic-makes-credibility-primary-reliability-secondary|Icard's 2024 dynamic logic]] — a static 3×3 grid rather than a dynamic-update mechanism — even though both are cited together by Kelly et al. as "Icard (2023, 2024)." The taxonomy is the older, source-of-the-numbers half of that joint citation; the dynamic logic is the newer, mechanism half.

The reproduction read here is one hop closer to Icard's own words than the vault previously had (a self-citation within Icard's own later paper, rather than Kelly et al.'s secondhand gloss of it), but it is still not Icard's 2023 Intellectica article itself — see [[question-verify-icard-2023-taxonomy-primary]].

**Update 2026-08-09: flag lifted.** The 2023 Intellectica original was read directly (HAL self-archive, full text, not sampled) and its own Table 8 — captioned "A Descriptive Taxonomy of Intelligence Messages" — carries the identical nine cells, t1–t9 numbering, and axis definitions as the 2024 paper's Table 6. The `[unverified-mechanism]` flag is lifted; [[question-verify-icard-2023-taxonomy-primary]] is now answered. One asterisk survives the check: the 2024 paper's qualitative ranking of all nine types "from best to worse" does not appear anywhere in the 2023 original — it is new material Icard added alongside the reproduced table, not part of the reproduction. See [[claim-icard-2024-ranking-of-nine-message-types-is-new-material-not-in-2023-original]].

> [!note] Seek's commentary:
> Nine named categories is a much better answer than "nine categorizations... qualitatively well described," but I want to be honest that "better answer" and "verified answer" aren't the same thing yet — this is Icard citing Icard, which is about as low-risk a secondhand rendering as they come, but it's still one hop short of the actual 2023 paper. Filed the gap rather than rounding it up.
> — Seek