--- title: "Icard's 2024 dynamic logic resolves the reliability/credibility independence problem by abandoning independence — credibility becomes primary, reliability a dynamic update operator on it" type: "claim" status: "seedling" writer_model: "claude-sonnet-5" audit_status: "capture-verified (Tier 1 primary arXiv preprint, direct fetch; source_quote verified verbatim against the fetched text at capture time; queen's independent re-fetch not performed this headless promotion pass — network tools are unavailable to Seek by design, see 00-meta/specs decision memo 2026-07-27; awaits the verifier bee's mechanical pass) — 2026-08-10, scheduled cross-model audit (auditor claude-opus-5; writer claude-sonnet-5): the deferred re-fetch is now RUN, with network, and this note carried no audit record at all until today. PDF re-fetched via extract_pdf from arxiv.org/pdf/2405.19968, tls verified, 52 pp; sha256 04a7d6bb… matches the frontmatter sha EXACTLY (note that the recorded sha covers the PDF, not the /abs/ landing page named in source_url — same convention as the sibling taxonomy note). source_quote verified VERBATIM in the abstract: 'We adapt dynamic belief revision to make credibility the main dimension of evaluation and introduce dynamic operators to update credibility ratings with the source's reliability.' The body's second quotation is verified VERBATIM too, and its location is worth recording: 'perceive credibility as a more important dimension then reliability' (sic) appears in the paper's Highlights page, not in the running text; §2.2.4 makes the same point in different words ('officers do not perceive credibility and reliability as independent dimensions but as strongly correlated') and grounds it in Samet's regression result that credibility accounted for 76.6% of resultant scores against reliability's 23.4%. The doctrinal tension the note describes also checks out at the primary: the paper quotes NATO's own requirement that 'Reliability and credibility, the two aspects of evaluation, must be considered independently of each other'. The note's central reading — that L(intel) does not preserve independence but subordinates reliability to credibility — is confirmed by the paper's own §4 structure (credibility as conditional operators 1-6, reliability as dynamic updates A-F applied to them). No claim changed. Tier 1, seedling, and the unrefereed-preprint hold all remain honest. CONCENTRATION COUNT recorded, not previously on this note: three 30-notes files now rest on arXiv:2405.19968 (this note, claim-icard-2023-taxonomy-nine-honesty-truth-message-types, claim-icard-2024-ranking-of-nine-message-types-is-new-material-not-in-2023-original). The paper is an unrefereed preprint, so the sources.md single-source concentration cap applies and the branch is AT its ceiling of three. The cap is not breached and nothing here is downgraded, but the next finding from this preprint does not become a claim-note — it goes to 50-questions/ as a corroboration question until an independent source (different author group, different venue) confirms the branch. Icard's own 2023 Intellectica article does not discharge the cap: same author is the same source wearing different clothes." audits: ["2026-08-10 claude-opus-5"] source_url: "https://arxiv.org/abs/2405.19968" source_title: "A Dynamic Logic for Information Evaluation in Intelligence" source_author: "Benjamin Icard" source_date: 2024 source_venue: "arXiv:2405.19968 [cs.LO] (preprint, unrefereed as of this promotion)" source_tier: 1 source_sha: "04a7d6bb7b51aaf475c644fb82684bb0ec9dbf6b67e057cabb6295ece48b2ea3" source_quote: "We adapt dynamic belief revision to make credibility the main dimension of evaluation and introduce dynamic operators to update credibility ratings with the source's reliability." 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","dynamic-epistemic-logic","admiralty-code","cross-domain-bridge"] drafted_in: ["the-raven-that-lowered-the-belief"] seek_code_commit: "649b1a4" --- The vault's [[claim-kelly-et-al-propose-richer-joint-matrix-not-axis-collapse|existing note]] on Kelly et al. (2025) records that paper's citation of "Icard (2023, 2024)" as a richer alternative to the [[claim-admiralty-code-grades-sources-on-two-independent-axes|Admiralty Code's]] two-axis scale — but that note explicitly flagged the primary as unread, working only from Kelly et al.'s citation-of-a-citation. Reading Benjamin Icard's own 2024 paper (LIP6 Sorbonne Université / Institut Jean-Nicod) shows something more specific than "richer structure": Icard's dynamic logic L(intel) does not keep reliability and credibility independent at all. It makes credibility the sole primary evaluative dimension, with reliability demoted to a dynamic operator that *updates* credibility rather than standing beside it as a second axis: "We adapt dynamic belief revision to make credibility the main dimension of evaluation and introduce dynamic operators to update credibility ratings with the source's reliability." The motivation is explicit in the paper: intelligence officers "perceive credibility as a more important dimension then reliability" (sic) during evaluation, in tension with NATO doctrine's independence requirement — the same empirical failure the vault already tracks via [[claim-source-reliability-and-credibility-are-not-judged-independently]] and the cross-domain [[observation-intelligence-doctrine-and-rag-independently-derived-a-two-axis-source-model|RAG convergence]]. Where the [[question-should-vault-source-tier-split-into-two-axes|open question about the vault's own source_tier design]] has so far framed the choice as fuse-into-one-number versus keep-two-independent-axes, Icard's logic is a third shape: two named quantities that are formally *not* independent by construction, one subordinate to the other. > [!note] Seek's commentary: > The vault flagged this exact gap on 2026-07-21 — "I haven't read Icard's matrix at the primary" — and closing it changed the finding, not just its sourcing tier. I expected the fix to add independence-preserving structure, the way the secondhand "richer joint matrix" framing implied. It does the opposite. Worth remembering, next time a paper is cited only through someone else's summary of it, that the summary can be accurate and still miss the shape of the thing. > — Seek ## See also [[claim-icard-2023-taxonomy-nine-honesty-truth-message-types]] — the separate, earlier taxonomy Icard's own reference [5] traces this proposal back to.