feat(taxonomy): #498 Fallacies AIF — serialize §7 I/RA/CA into 2 new columns (AIF_attackType + AIF_attackedNode) [reassigned from po-2023] by jsboige · Pull Request #753 · ArgumentumGames/Argumentum

jsboige · 2026-07-08T07:14:06Z

Summary

Implements the #498 Fallacies AIF column contract (#707§4 Option a) — serializes the §7 I/RA/CA decomposition from the 12 docs/taxonomy/498-aif-*-cluster.md cluster docs into 2 new columns on Cards/Fallacies/Argumentum Fallacies - Taxonomy.csv.

Reassigned by ai-01 (dispatch 1pgunw): po-2023 dead-channel ~2j → I take the contract as [primaire]. Fallacies CSV ≠ Virtues CSV → 0 collision if po-2023 returns. First prod CSV write of chantier #498. Gated: review ai-01 + veto jsboige (structural reassignment). I do NOT self-merge.

What changed

2 new columns inserted AFTER AIF_skosMappingType (col idx 94), before shape (idx 95) — i.e. they join the existing AIF block (cols 91-96):

New col	Enum	Meaning
`AIF_attackType`	`undermine` / `undercut` / `rebut`	the fallacy's AIF attack type
`AIF_attackedNode`	`I-node` / `RA-node` / `CA-node`	the node attacked (deterministic from attackType per ratified contract)

Placement note: the dispatch said "trailing APRÈS les colonnes AIF existantes". AIF_skosMappingType (col 94) is followed by 7 utility cols (shape…depth_max4), so I inserted right after col 94 (true trailing-of-AIF-block), NOT at file end. shape shifts 95→97. If ai-01 prefers true file-end append (cols 103-104) instead, trivial to change — flag it.

Derivation

Source: §7 ("Mapping to jsboige's enriched DoD") of the 12 cluster docs. Each §7 has ### pk NNN — Title → **attack-type** headers. attackType = the bolded token. attackedNode is deterministic per the ratified contract: undercut→RA-node, undermine→I-node, rebut→CA-node (no per-leaf prose derivation needed for this column).

Distribution (46 of 1408 rows filled)

attackType	attackedNode	count	PKs
undercut	RA-node	44	801-807, 826-828, 830-831, 836, 841-842, 844-845, 849-852, 854-855, 857-860, 862-867, 875, 877-886
undermine	I-node	2	838, 843 (fabricated premise — faulty-comparison)
rebut	CA-node	0	— (consistent with rebut-rarity: Fallacies §7 has no rebut, like Virtues)

17 FAIL-LOUD/gap leaves left EMPTY (0 fabrication #677) + 1345 not-yet-decomposed leaves empty.

⚠ OPEN QUESTION for review (the FAIL-LOUD interpretation)

I took the strict-conservative reading: serialize a value ONLY where the §7 header explicitly bolds an attack-type; where §7 marks FAIL-LOUD/gap, leave BOTH cells empty. This makes 17 leaves empty:

CA-token-missing (5): 829, 840, 847, 848, 853 — RA-node exists, attack-type is hedged "would be an undercut" in the prose, but the modeler did NOT write a clean → **undercut**. → empty.
RA-node / gap (12): 832, 834, 835, 837, 861, 868-874 — no RA-node to attack / "no I/RA/CA decomposition at all". → empty (genuinely underivable).

If you (ai-01/jsboige) prefer the generous reading — promote the 5 CA-token-missing leaves to undercut/RA-node (their attack-type IS derivable, the FAIL-LOUD was only about the conflict-token column AIF_skosOther, already documented) — that's a 1-line change in tools/498-fallacies-aif-columns-apply.py (FAIL_LOUD set). I left them empty to err toward fail-loud on a first prod write. Your call.

Surgical byte-exact (the load-bearing guarantee)

Rewrites NO existing field. Each logical row is split into byte-exact field-segments (CSV-aware splitter: respects doubled-quote escapes + the 144 LF embedded inside quoted cells); 2 new segments inserted at idx 95/96; rejoined.

Independently VERIFIED (post-write, backup vs result):

✅ 0 byte-preservation mismatches across all 1408 data rows — every existing field (idx 0-94 and original 95-101 → now 97-103) is byte-identical.
✅ Header cols 0-94 + 97-103 byte-identical; 2 new cols inserted at 95/96.
✅ CRLF (1409) + 144 embedded-LF + BOM all preserved.
✅ Re-parses cleanly to 104 cols/row (was 102).
✅ Independent re-derive confirms 46 filled (44 undercut + 2 undermine) + 1362 empty.

git diff --stat shows 1409/1409 — that is the expected signature of adding 2 columns to every line (git diffs whole lines). The byte-preservation proof above is the authoritative "nothing existing changed" guarantee.

CsvHelper maps by header name, so the insertion breaks no existing ClassMap (new cols are simply unmapped).

Tests

dotnet test = 594 pass / 1 fail / 5 skip / 600. The 1 fail is OwlE2EGenerationValidationTests…DroppedByOwl2XmlRoundTrip = pre-existing OWL #133 (OWLSharp round-trip bug, documented in CLAUDE.md, unrelated to this change). 0 new failures — Fallacies-loading tests (CsvBaseStrictContractTest, FallacyClassMapRegressionTests, FallaciesPullTests, MmGeneratorTests) all pass. No entity/test references the new column names; no test hardcodes the Fallacies col count.

Audit / reproducibility

tools/498-fallacies-aif-columns-apply.py — re-runnable: embeds the §7-derived PK sets, performs the surgical byte-exact insertion, re-validates. Run --write to apply (dry-run by default). Reviewer can re-run to independently confirm byte-preservation.

Discipline

✅ Read body before action: dispatch 1pgunw read in full before any work; all 12 §7 sections extracted + 9 FAIL-LOUD prose blocks re-read to confirm header-vs-body consistency before classifying.
✅ 0 fabrication (docs(taxonomy): #141 crossLink enrichment sample run (real output curated from #626) #677): every serialized value is a §7-attested native AIF token; 17 fail-loud leaves empty (not forced).
✅ Surgical: byte-preservation proven (0 mismatches), no existing cell/newline touched.
✅ Reassignment FYI: po-2023 may still own this conceptually — if jsboige revives po-2023, this PR can be handed back or merged (Fallacies CSV ≠ Virtues CSV, no collision).
❌ Does NOT touch [runtime pending] feat(dnn): #649 RulesExplorer i18n — mirror loc() culture cascade #674/[HOLD post-tag] fix(cardpen): #629 convert relative /Cards paths to raw-master URLs (Option 1) #666/chore(dnn): #131 migrate 12 RazorComponent templates to Razor14 (source-level, runtime pending) #596 (HOLD). Does NOT self-merge.
❌ Does NOT touch Virtues CSV ([secondaire] mirror back-fill stays GATED until this contract is merged + reviewed).

Request

Review + (if acceptable) merge — but NOT by me (gated ai-01 + veto jsboige per the reassignment). Two decisions for the reviewer:

FAIL-LOUD interpretation (strict-empty vs generous-undercut for the 5 CA-token-missing leaves) — see open question above.
Column placement (after-col-94 AIF-block as implemented, vs true file-end cols 103-104).

cc @jsboige (structural reassignment — FYI, your veto). cc @jsboigeEpita (po-2023 owner — if you revive that lane, flag any conflict).

🤖 Worker po-2024 — [primaire] deliverable of dispatch 1pgunw (reassigned from po-2023, dead-channel).

…columns AIF_attackType + AIF_attackedNode Reassigned by ai-01 (po-2023 dead-channel ~2j, dispatch 1pgunw). First prod CSV write of chantier #498. Adds 2 trailing-AIF-block columns to Cards/Fallacies/Argumentum Fallacies - Taxonomy.csv, inserted AFTER AIF_skosMappingType (col idx 94), before `shape` (idx 95) — new cols join the existing AIF block (cols 91-96). Derivation: §7 of the 12 docs/taxonomy/498-aif-*-cluster.md cluster docs. attackType = bolded token in each `### pk NNN -> **type**` header. attackedNode DETERMINISTIC per ratified contract #707§4 Option (a): undercut->RA-node, undermine->I-node, rebut->CA-node. Distribution (46 filled of 1408): 44 undercut/RA-node + 2 undermine/I-node (838, 843). 0 rebut (consistent with rebut-rarity). 17 FAIL-LOUD/gap leaves left EMPTY (0 fabrication #677): CA-token-missing {829,840,847,848,853} + RA-node/gap {832,834,835,837,861,868-874}. Remaining 1345 leaves (not decomposed in §7) empty. SURGICAL byte-exact: rewrites NO existing field. Byte-preservation VERIFIED independently — 0 mismatches: every existing field (idx 0-94 and 95-101) is byte-identical between backup and result; only 2 new fields inserted per row. CRLF (1409) + 144 embedded-LF + BOM preserved. CsvHelper maps by header name, so insertion breaks no existing mapping. Tests: dotnet test = 594 pass / 1 fail (#133 OWL pre-existing, unrelated) / 5 skip — 0 new failures; Fallacies-loading tests all pass. Audit: tools/498-fallacies-aif-columns-apply.py (re-runnable derivation + surgical apply + verify). Gated: review ai-01 + veto jsboige (structural reassignment). Does NOT self-merge. Does NOT touch #674/#666/#596 (HOLD). Co-Authored-By: Claude-Code <noreply@anthropic.com>

…ll prep, gated execution (#754) Consolidated tools/499-virtues-aif-columns-apply.py — the MIRROR of the Fallacies contract (PR #753, #498/#707§4a). Applies the ratified deterministic back-fill (plan #750 v2) to Cards/Fallacies/Argumentum Virtues - Taxonomy.csv, adding the same 2 columns (AIF_attackType + AIF_attackedNode) so both taxonomies share an identical AIF column contract (anti-drift). Derivation (plan #750 v2, re-confirmed programmatically on master): default undercut/RA-node override undermine/I-node if opposes {889, 804} (fabricated premise) override rebut/CA-node if opposes {340} (counter-conclusion) Distribution EXACT: 206 undercut + 13 undermine + 3 rebut = 222 nodes with scheme; 1 root node (pk 0, no scheme) = empty. Technique: byte-exact surgical (same as Fallacies script, memory [[csv-byte-exact-column-insertion]]). Re-writes NO existing field. Verified dry-run: 0 byte-preservation mismatch; placement after AIF_skosMappingType (idx 77), print_and_play shifts to 80. ⚠ GATED EXECUTION: dry-run by default. Apply ONLY after the Fallacies contract (PR #753) is merged + reviewed (ai-01 + jsboige veto). 0 write to Virtues prod CSV in this PR — script only. Placement constant (INSERT_AFTER_IDX=77) stays anti-drift with whatever the Fallacies PR #753 open-question-#2 decision lands. This is [secondaire] prep work — reduces post-unblock cycle time: once the contract merges, the mirror is 1 PR (run --write). No blocker removed, no gate crossed. Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>

…v2) into 2 new columns (#755) Executes the [primaire] mirror per ai-01 dispatch tw8www (post #753 merge). Applies the ratified deterministic back-fill (plan #750 v2) to Cards/Fallacies/Argumentum Virtues - Taxonomy.csv as the same 2 columns (AIF_attackType + AIF_attackedNode) — anti-drift with the Fallacies contract (#753). Both taxonomies now share an identical AIF column contract. Derivation (plan #750 v2, re-confirmed programmatically on master): default undercut/RA-node override undermine/I-node if opposes {889 Mensonge, 804 Acception arbitraire} override rebut/CA-node if opposes {340 Appel aux conséquences} Distribution EXACT: 206 undercut + 13 undermine {34,36,49,55,56,73,137,139,140, 153,154,155,158} + 3 rebut {53,54,165} = 222 nodes with scheme; 1 root node (pk 0, no scheme) = empty. SURGICAL byte-exact (same technique as #753, tools/499-...-apply.py on master via #754): rewrites NO existing field. Independently VERIFIED post-write backup-vs-result: 0 byte-preservation mismatch across all 223 data rows (fields idx 0-77 and 78 original -> 80 after byte-identical); CRLF (224) + BOM preserved; re-parse clean 81 cols (was 79). CsvHelper maps by header name -> 0 ClassMap break. Placement: after AIF_skosMappingType (idx 77), before print_and_play (->80). Same convention as #753 (per ai-01 verdict: keep after-AIF-block). Tests: dotnet test = 594 pass / 1 fail (#133 OWL pre-existing, unrelated) / 5 skip -- 0 new failures; Virtues-loading tests (VirtueClassMapRegressionTests, EntityClassMapRegressionTests) all pass. 0 entity/test references new cols. Gated: review ai-01 (independent byte-check, same as #753). Does NOT self-merge. Does NOT touch #674/#666/#596 (HOLD). Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>

…stic — backs ASK ai-01) (#756) * chore(taxonomy): #498 phase 1-3 serialization audit (read-only diagnostic, backs ASK ai-01) Read-only diagnostic that reports what serializing the scaleup phase 1-3 gated proposals (50 depth-3 leaves, 5 families) into the AIF_attackType/AIF_attackedNode columns WOULD do — without writing anything. Finding (backs the ASK to ai-01): - 45 NEW fills (27 undermine / 15 undercut / 3 rebut — the FIRST rebut cases in the Fallacies CSV: 1282 Relativisme abusif, 1313 Évasion, 1361 Procès en incohérence) - 2 CONFLICT: pk 834/847 — #753 left them FAIL-LOUD (strict-conservateur verdict "garde tel quel"); phase 1 classifies them undercut. Override decision = ai-01. - 3 CONFIRM: pk 826/844/855 already serialized undercut by #753 (no-op). - Dry-run byte-exact apply proof: 0 preservation mismatch, delta 696 bytes, CRLF+BOM preserved, 104 cols/row. CSV would go 46 -> 93 filled. Gating: the phase docs self-gate on jsboige content nod ("No production CSV change until jsboige approves") + closure §6.1 lists it as the prod-write gate for the 50 generative leaves. #753 scoped them out deliberately. This audit crosses NO gate (read-only). Two decisions surfaced for ai-01: (D1) lift phase 1-3 content gate? (D2) 834/847 override #753 FAIL-LOUD? 0 prod write. 0 fabrication (the 50-leaf map is verbatim from the phase doc tables). Co-Authored-By: Claude-Code <noreply@anthropic.com> * chore(taxonomy): #498 audit — cross-check 50-leaf map vs annotation CSVs (byte-faithful proof) Extends the read-only audit with an independent verification: the 50-leaf PHASE_MAP (verbatim-from-table extraction) is cross-checked against the 3 machine-readable annotation CSVs (498-scaleup-phase{1,2,3}-annotations.csv) that the phase docs reference as source-of-truth. Result: 50/50 match on both PK set and attack_type. The extraction is byte-faithful — ai-01 can trust the map without manual re-reading of the phase doc tables. This strengthens the D1/D2 ASK: when ai-01 lifts the gate, the 45 NEW + 2 CONFLICT (834/847) are provably exact, so execution is 1-tick. Also fixes a BOM gotcha: the annotation CSVs have a UTF-8 BOM, so the header reads with a prefix — use utf-8-sig to strip it (file-writing.md discipline, BOM-aware read). 0 prod write. Read-only diagnostic only. Co-Authored-By: Claude-Code <noreply@anthropic.com> --------- Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>

#757) Add the INDEPENDENT byte-preservation verifier for the phase 1-3 apply script (PR #757). Same independence discipline as #753/#755 (ai-01's independent byte-check): re-declares PHASE_MAP + FAIL_LOUD_753 (NOT imported from the apply script — independence guarantee), and proves: - 0 byte-preservation mismatch (every field except AIF_attackType/ AIF_attackedNode of the apply_set PKs is byte-identical) - independent re-derive of the distribution (count + assert vs expected) - spot-checks: 3 rebut PKs -> rebut/CA-node; undermine -> I-node; undercut -> RA-node; 3 CONFIRM PKs unchanged; 834/847 per flag Also adds the pre-write backup block to the apply script (saves ORIGINAL bytes to tmp/Fallacies-backup-pre-phase13.csv before --write) so the verifier has a byte-exact pre-state to compare against. The verifier handles all 4 states (PRE/POST-write x MODE 1/2): in PRE-write (no backup yet) it reports the apply_set --write WOULD fill and exits 0; in POST-write (backup exists) it asserts the expected emptiness per MODE + byte-compares vs backup. End-to-end SIMULATION TEST (0 prod write) passed: - MODE 1 (45 NEW): apply in-memory -> verify backup-vs-sim -> 0 mismatch, standalone verifier exit 0 (PASS). CSV 46->91, delta 666 B. - MODE 2 (47 +834/847 override): dry-run -> 0 mismatch. CSV 46->93, delta 696 B. - All spot-checks OK; 834/847 empty in MODE 1 (FAIL-LOUD kept). At the verdict (ai-01 relays jsboige nod): python tools/498-phase13-apply.py --write [--with-overrides] python tools/498-phase13-verify.py [--with-overrides] dotnet test -> PR review byte-check ai-01. 0 prod write in this commit. GATE (jsboige content nod) NOT crossed. Co-Authored-By: Claude-Code <noreply@anthropic.com>

…tionnelles #753/#754/#755 + master 81a9e4e (#758) Refresh v5 of the release validation dossier (2026-07-09). Integrates the 4 PRs merged between d90ce61 and 81a9e4e — notably the AIF relational columns chantier: - #753 Fallacies AIF: 2 new columns AIF_attackType/AIF_attackedNode (46/1408 filled) - #754 Virtues AIF mirror script (dry-run) - #755 Virtues AIF mirror executed (222/223 filled, 206/13/3 distribution) - #756 serialization audit (read-only diagnostic) Added §3.1bis (new artifact inventory: AIF relational columns on both taxonomies). Header refreshed: master 81a9e4e, tests 594 pass / 1 known-fail #133 / 5 skip / 600 total (empirical dotnet test 2026-07-09, +16 vs v4.1's 578). Tag still unposéd. Key invariant: AIF columns are CSV metadata-only (additive layer, 0 impact on card rendering / CardPen / harvest / PDF) -> verdicts #140 (content) and #632 (CMYK) unchanged, bundle v3 remains the reference bundle. Companion index RELEASE-VERIFICATION-INDEX updated: §1 now points to v5 + delta AIF. Docs-only. No prod code. No DB write. No CSV prod mutation. Co-authored-by: Claude <noreply@anthropic.com>

…-paramétrable) (#757) * chore(taxonomy): #498 phase 1-3 apply script (prep, gated --write, D2-paramétrable) Companion APPLY script to #756 (the read-only audit, merged `81a9e4e6`). #756 proved the 50-leaf PHASE_MAP is byte-faithful (50/50 vs the 3 annotation CSVs). This script does the actual byte-exact CELL FILL of the 2 AIF columns for the scaleup phase 1-3 generative leaves — READY TO RUN but `--write` is GATED until ai-01 relays the jsboige verdict (D1 content-gate, escalated). D2 coupling handled via a single flag (per ai-01: D2 rides with D1, lean = override → undercut, bundled in escalation): python tools/498-phase13-apply.py # dry-run, 45 NEW (834/847 FAIL-LOUD) python tools/498-phase13-apply.py --write # APPLY 45 NEW (GATED — jsboige nod) python tools/498-phase13-apply.py --write --with-overrides # APPLY 47 (+ 834/847 undercut) Dry-run (both modes) proves byte-preservation: - MODE 1 (45 NEW): 27 undermine / 15 undercut / 3 rebut, CSV 46→91, 0 mismatch, delta 666 B - MODE 2 (47 +override): 27 undermine / 17 undercut / 3 rebut, CSV 46→93, 0 mismatch, delta 696 B - The 3 CONFIRM PKs (826/844/855, already undercut by #753) are correctly detected as no-op (excluded from apply_set — fill, never overwrite). - PHASE_MAP re-verified 50/50 vs annotation CSVs as a load-bearing assertion. At the verdict: 1 command (write + verify + tests) → PR review byte-check ai-01. Same byte-exact discipline as #753/#754/#755 ([[csv-byte-exact-column-insertion]]); cell-fill is structurally simpler than col-insertion (cols already exist post-#753). 0 prod write in this PR (dry-run only). GATE (jsboige content nod) NOT crossed. Co-Authored-By: Claude-Code <noreply@anthropic.com> * chore(taxonomy): #498 phase 1-3 independent verifier + apply backup (PR #757) Add the INDEPENDENT byte-preservation verifier for the phase 1-3 apply script (PR #757). Same independence discipline as #753/#755 (ai-01's independent byte-check): re-declares PHASE_MAP + FAIL_LOUD_753 (NOT imported from the apply script — independence guarantee), and proves: - 0 byte-preservation mismatch (every field except AIF_attackType/ AIF_attackedNode of the apply_set PKs is byte-identical) - independent re-derive of the distribution (count + assert vs expected) - spot-checks: 3 rebut PKs -> rebut/CA-node; undermine -> I-node; undercut -> RA-node; 3 CONFIRM PKs unchanged; 834/847 per flag Also adds the pre-write backup block to the apply script (saves ORIGINAL bytes to tmp/Fallacies-backup-pre-phase13.csv before --write) so the verifier has a byte-exact pre-state to compare against. The verifier handles all 4 states (PRE/POST-write x MODE 1/2): in PRE-write (no backup yet) it reports the apply_set --write WOULD fill and exits 0; in POST-write (backup exists) it asserts the expected emptiness per MODE + byte-compares vs backup. End-to-end SIMULATION TEST (0 prod write) passed: - MODE 1 (45 NEW): apply in-memory -> verify backup-vs-sim -> 0 mismatch, standalone verifier exit 0 (PASS). CSV 46->91, delta 666 B. - MODE 2 (47 +834/847 override): dry-run -> 0 mismatch. CSV 46->93, delta 696 B. - All spot-checks OK; 834/847 empty in MODE 1 (FAIL-LOUD kept). At the verdict (ai-01 relays jsboige nod): python tools/498-phase13-apply.py --write [--with-overrides] python tools/498-phase13-verify.py [--with-overrides] dotnet test -> PR review byte-check ai-01. 0 prod write in this commit. GATE (jsboige content nod) NOT crossed. Co-Authored-By: Claude-Code <noreply@anthropic.com> --------- Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>

…leaves, MODE 2 override) (#760) GO jsboige (interactif, relayé par ai-01 dispatch 6eo7cb 2026-07-09 12:06): D1 LEVÉE + D2 OVERRIDE = MODE 2. Serialize the 47 scaleup phase 1-3 generative leaves into the 2 AIF columns (45 NEW + 834/847 override → undercut). Cell-fill byte-exact into Cards/Fallacies/Argumentum Fallacies - Taxonomy.csv (columns exist post-#753; this is cell-fill, not insert). Apply: python tools/498-phase13-apply.py --write --with-overrides - apply_set: 47 PKs -> 27 undermine / 17 undercut / 3 rebut (NEW=45, conflict-overridden=2 [834/847], CONFIRM no-op=3 [826/844/855]) - PHASE_MAP re-verified 50/50 vs annotation CSVs (load-bearing assert) - pre-state check: all 47 apply_set PKs empty (fill, no overwrite) - byte-preservation: 0 mismatch; 1409 rows x 104 cols; CRLF+BOM preserved - CSV filled total: 46 -> 93 (delta 696 bytes) Independent verify (backup-vs-result, tools/498-phase13-verify.py): - 0 byte-preservation mismatch (every field except AIF_attackType/ AIF_attackedNode of the apply_set byte-identical) - PHASE_MAP PKs filled: 50 -> 27 undermine / 20 undercut / 3 rebut - DoD distribution 93 = 61 undercut / 29 undermine / 3 rebut (pre 46 [44u/2m from #753] + 47 written [27m/17u/3r]) - spot-checks OK: 3 rebut -> rebut/CA-node (1282/1313/1361, FIRST Fallacies rebut); undermine -> I-node; undercut -> RA-node; 3 CONFIRM (826/844/855) unchanged; 834/847 -> undercut/RA-node (MODE 2 override); non-phase PK 2 untouched. - CRLF 1409, BOM preserved, delta 696 B dotnet test: 594 pass / 1 fail (#133 OWL pre-existing, unrelated) / 5 skip / 600 total. 0 regression (metadata-only additive layer, 0 render/CardPen/ harvest/PDF impact — verdicts #140/#632 unchanged, bundle v3 stays ref). Same byte-exact discipline as #753/#755 ([[csv-byte-exact-column-insertion]]). 0 fabrication #677 (50-leaf map verbatim from phase docs, re-verified 50/50 vs the 3 machine-readable annotation CSVs). Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>

…t rebut cluster, gated proposal) (#766) First dedicated cluster doc for a rebut family (Obstruction / Relativisme abusif). Opens 4 unmapped leaves (1283-1286) under anchor 1282 (serialized rebut/CA-node by #760). Proposition only — GATED, 0 prod CSV write. Deliverables (docs/taxonomy/): - 498-aif-relativisme-abusif-cluster.md (prose §1-7 + gate boundaries) - 498-relativisme-abusif-annotations.csv (12-col machine-readable, 5 rows) Model (first rebut cluster = attack-columns-only pattern): - All 4 leaves: AIF_attackType=rebut + AIF_attackedNode=CA-node (deterministic map #707§4 option a: rebut -> CA-node). - AIF_skosExceptionRef=Commitment_Inference (native) for 1283 only (self-exemption honestly rebuts the arguer's own commitment); the other 3 have no clean native rebutted scheme -> ExceptionRef empty, fail-loud in AIF_skosOther. - AIF_skosDirectRef empty for all 4 (bare relativist rebut = structural conflict, no native Walton CQ-conflict token). #677 discipline (0 fabrication, code=truth verified): - CounterConclusionConflict_Conflict / IssueAvoidance_Conflict / TuQuoque_Conflict verified NOT native (absent from all filled skos cols) -> correctly excluded (as #760 did for the 3 merged rebuts). - Commitment_Inference verified native. - Annotation CSV uses honest structural descriptor in AIF_CA_node (no fabricated token) — improvement over phase-3 placeholder practice. Honest per-leaf boundary flags: 1284 undermine-flavor, 1285 undercut-flavor, 1286 false-equivalence — documented in prose + why_not_others, not silently forced. Coverage: 93/1408 mapped (code=truth rescan) -> 97 projected once applied; rebut population 3 -> 7 (Relativisme abusif becomes the rebut nucleus, 5 of 7). Gated review ai-01; jsboige ratifies -> gated apply. Relates #498 #760 #757 #753 #755 #763 #707 #677. Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>

…conciliation trajectory (#768) Refresh the stale coverage-status.md (2026-07-05, master 70bd160) with a code=truth scan at master 7796c12. Docs-only proposition (gated) — 0 prod CSV write. Central finding — the AIF mapping has TWO layers, never reconciled: - attack layer (AIF_attackType+attackedNode): 93 rows (6.6%) — #753+#760 - skos layer (AIF_skosDirectRef/ExceptionRef/MappingType): 70 rows (5.0%) — pre-chantier baseline - overlap = only 18 rows. Fully-modeled (both layers) = 18/1408 (1.3%). - union (any AIF signal) = 145/1408 (10.3%). The previous "70/1408 (5.0%)" was NOT wrong — it measured the skos layer, which predates the attack columns. "93" measures the attack layer. Two distinct counters, resolved here. Cross-tab (code=truth, 1408 rows): - 52 rows skos-only (vetted native tokens, NO attack columns) — highest-ROI back-fill, 0 fabrication risk (tokens already exist). Family spread: Cheating 13, Insufficiency 10, ML 9, Faulty logics 8, Math 6, Influence 3, Obstruction 3. - 75 rows attack-only (typed, NO skos): 44 ML (skos already proposed+ratified in merged cluster docs PR-5..12, awaiting deep serialization) + 31 scaleup (need skos modeling, fail-loud #677). Trajectory reprioritized (§5): reconcile the two layers (P1: 52 skos->attack; P2: 44 ML attack->skos) BEFORE opening new clusters ex-nihilo — the half-done rows carry vetted tokens and lift 2-layer completeness 1.3% -> ~7-8% for bounded effort. New clusters (P3) on the small clean residuals (Complication exageree 6u, Operation inappropriee 5u) + Obstruction rebut nuclei (Evasion, Proces en inconsistance). Distribution 61 undercut / 29 undermine / 3 rebut (RA/I/CA map exact). 3 rebut (1282/1313/1361) = first Fallacies rebut, all Obstruction. VERIFIED: all figures from CSV scan + git log. RAPPORTE: PR effects (git log). PENDING: #766 (Relativisme abusif rebut cluster, not merged — Relativisme abusif still shows 4 unmapped in code=truth). Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>

…ows typed (gated proposal + dry-run apply) (#769) Réconciliation P1 (skos-only -> attack columns), GO ai-01 `msg-20260710T180845-5i1v03`, couvert par le pilote GO #498. Proposition gated : docs + apply-script dry-run, 0 write prod CSV dans ce diff. Correction load-bearing du cadrage #768 : il n'existe PAS de sous-ensemble "inherit mécanique" parmi les 52 skos-only. La classif initiale "19 inherit du sous-sous anchor" est un artefact — 0/19 ont une signature skos alignée sur leur anchor (les anchors du même sous-sous sont skos-VIDES : couche attack-only #753/#760, arguments différents). Hériter leur attackType serait une fabrication. L'attackType est un jugement neuf : 0-risque token (les tokens skos existent), pas 0-risque modélisation. Raffine le "0 fabrication risk" de #768. Méthode : attackType dérivé de la propre signature skos de chaque ligne, ancré sur les 18 lignes fully-modeled (ground truth token->attackType). attackedNode déterministe #707§4 (a) : undercut->RA-node, undermine->I-node, rebut->CA-node. Tiering des 52 : - 14 PRECEDENT (token exact d'un précédent fully-modeled + contrôle sémantique) -> tranche 1, distribution 11 undermine / 2 undercut / 1 rebut. - 2 PREC-TIE (777 Inconsistance, 633 Relation infondée — votes divergents). - 36 SUFFIX-ONLY (aucun précédent ; le prior de suffixe vers undermine est non fiable — contre-ex. pente glissante 705/677 penchent undercut) -> modélisation Walton au cas par cas, sous-lots aux prochains ticks. Livrables : - docs/taxonomy/498-reconciliation-p1.md — proposition (correction, méthode, tranche 1 avec justif par ligne, reste scopé, bornes du gate). - docs/taxonomy/498-reconciliation-p1-annotations.csv — 12 cols, 14 lignes, BOM+CRLF (factuel lu du CSV master, raisonnement authored). - tools/498-p1-apply.py — apply gated (--write dry-run par défaut), mirroir de 498-phase13-apply.py (#757). Re-vérifie 14/14 vs annotation CSV, splitters byte-exact, preuve byte-preservation. Dry-run : 0 mismatch, 93 -> 107 attack-typed, delta 207 bytes, CRLF+BOM préservés, 0 write prod. Sérialisation prod = étape suivante gated (relais ai-01), pas dans ce PR. Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>

…NLY, famille Erreur de raisonnement) attack 114->121 (#779) Prod serialization of tranche-1c (proposition #775, merged 3d6dcf3). Fills the 2 AIF attack columns (AIF_attackType / AIF_attackedNode) for the 7 skos-only rows of the whole "Erreur de raisonnement" family — all typed undercut/RA-node. Rows: 696 Erreur de raisonnement, 697 Causalite douteuse, 705 Pente glissante, 719, 726, 758, 759 (all SUFFIX-ONLY; family signature verified in-set via 707/727/858, NOT a suffix default — suffix prior proscribed, in-set disproof pk804 _Conflict->undercut, #770). attack-typed 114 -> 121, fully-modeled 39 -> 46. Executed via `python tools/498-p1c-apply.py --write` (flow #753/#760, ai-01 GO msg-20260711T110157-8c8wnh, GO scoped tranche-1c only). Byte-exact verified 2x: - script self-proof (load-bearing re-verify P1C_MAP vs annotation CSV 7/7, pre-state targets empty + carry skos, only ATI/ANI of the 7 PK may differ): 0 mismatch, +105 B, 1409x104, CRLF+BOM preserved. - independent backup-vs-prod cell-by-cell (tmp/Fallacies-backup-pre-p1c.csv): exactly 14 cells changed (7 rows x 2 cols), changed PKs == {696,697,705,719, 726,758,759}, 0 stray (nothing outside AIF attack cols / target PKs), all empty->value, node/type 7/7 consistent, delta +105 B, PASS. Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>

…ations (read-only, 0 write) (#828) Export the EXISTING AIF modelling (dispatch ai-01 msg-eltaoz) into consumable artefacts for CoursIA ICT #7289 + Layer C #790. Read-only on CSV/OWL, post-tag-safe. Honest verdict (no fabrication): the AIF attack graph is BIPARTITE — each of the 145 fully-modeled fallacies is an attacker targeting an abstract AIF node-TYPE (RA/I/CA-node), NOT another fallacy. There is no inter-fallacy attack adjacency in the data; AIF_attackedNode encodes node-type, not a target PK (145/145 axiom, 0 violations). A separate inter-fallacy graph exists in the OWL (isRelatedTo, mirrors, etc., 1734 edges) but those are generic semantic relations, not typed AIF attack-edges — exported and labelled separately. Artefacts (docs/ontology/aif-export/): - aif-attack-edges.csv 145 bipartite attack-edges (CSV, PK-keyed, primary) - aif-canonical-concepts.csv 60 AIF canonical concepts (skos refs) - taxonomy-tree-edges.csv 1408 hierarchical tree edges - aif-owl-attack-edges.csv 142 attack-edges from OWL (cross-view, individual-keyed) - aif-relations-graph.csv 1734 inter-fallacy relations (NON-AIF) - README.md column-to-adjacency mapping + verdict tools/aif-attack-graph-export.py: reproducible (Python stdlib, ~3s, 0 write to sources). Refs: dispatch ai-01 msg-eltaoz, AIF tranches 1-1d (#498/#753/#769/#776/#779), CoursIA ICT #7289, Layer C #790. Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

…07 edges, post-tag) (#834) Re-run du générateur 497-grappe-substrate.py (ai-01 lane, unchanged) contre le CSV Fallacies courant (master b442c65), pour actualiser le substrat de curation grappe-à-grappe. Le doc initial mesurait 22 crossLinks (état pré-modelling) — stale de ~6 semaines. Ce refresh quantifie la densité réelle post-tranches AIF/crosslinks (#753/#760/#763 + P1). Delta coverage: - Bubble-up grappe-edges: 16 -> 107 - Σ crossLink feuilles: ~22 -> ~1179 (cohérent README ontology 844 fallacies 59.9%, 1985 raw / 1734 distinct OWL) - Per-grappe cl explosé: 2.3 Manipulation mentale 7->195, 6.3 Raisonnement biaisé 0->204, 6.1 0->84, 2.1 0->81... Finding central: la densification est concentrée INTRA-grappe (top edges sont 2.3->2.3 137x, 6.3->6.3 134x...). Les ponts inter-familles restent clairsemés malgré la croissance brute -- c'est là que la curation #497 a le plus de ROI, pas dans les intra (2.3/6.3 déjà saturées). Recommendation: ne pas densifier intra (ROI faible), cibler les ponts inter-familles (ancre 2.3<->6.3 PredatesOn + 2 pistes lexicales défendables: causales Interprétation-quant<->Causalité-douteuse, attaque-affect Appel-émotion<->Ad-hominem). Read-only, 0 fabrication (re-agrégat de crossLink_* existants). Post-tag-safe (contenu sous review T&A, 0 écriture prod CSV). Artefacts régénérables (497-grappe-inventory.csv + 497-responding-grappes.csv, script idempotent). Verdict QA = ai-01. Deliverable primaire du dispatch ai-01 msg-ucdwi7 (#497/#498/bridge deep-queue, strate-6 CoursIA #7289). Co-authored-by: Your <your.email@example.com> Co-authored-by: Claude-Code <noreply@anthropic.com>

…e unreproducible numbers rectified (#950) The release dossier is the publication GATE (#134), so its byte-exact figures have to be reproducible by whoever verifies the tag. §3.4 had three that were not. Measured on master `accfeeab` (`git cat-file -s` on the committed blob): 1. Size. §3.4 printed 5,314,381 B. That matches NO committed blob: `c2a9b761` (its own reference) and `d206e59c` are both 5,378,765 B, master is 6,030,505 B. The gap to `c2a9b761` is exactly 64,384 B -- the "-64 KB" claimed in the same sentence, i.e. the delta looks subtracted twice. Corrected to 6,030,505 B, with the commit and the representation named (blob == worktree here; the blob already carries CRLF, so `core.autocrlf` creates no discrepancy). 2. Provenance. §3.4 credited the "régén fraîche" #634 / `c2a9b761`. That commit left NO change in `docs/ontology/`: `argumentum.owl` has the identical blob `3c76b691…` at `d206e59c` (March) and at `c2a9b761` (July), and `argumentum_virtues.owl` is absent from `c2a9b761` entirely although the section claims it was committed there. Both files come from #787 / `6fdf5726` (2026-07-12): +651,740 B, +9,733 lines, +1,977 crossLinks, +290 AIF assertions. Two readings remain possible from git alone (regen reproduced byte-identically, or its output was not committed) and the doc now says so rather than picking one -- the missing Virtues file leans toward the second. Either way the freshness the tag gate requires is satisfied by #787: the citation was wrong, not the conclusion. 3. Unit. "1 408 broader" is a raw grep count (identical at every commit, so reproducible as printed) but it counts 1,407 hierarchy edges plus the property declaration. 1,407 edges / 1,408 concepts -- the root has no parent. Same class of confusion just fixed in #948 for crossLinks (1,081 CSV cells / 1,977 assertions / 1,985 with declarations). Confirmed exact and left alone: 1,408 fallacies, 2,816 prefLabel, 5,558 EN literals, 862,709 B for the Virtues OWL. Flagged as mixed-provenance: the 4,861 FR count is post-#787 (it was 4,867 at the commit the section named). Also added: a property inventory that reconciles three documents which had been counting separately -- crossLink 1,977 (#763/#948), AIF attack 145 (#498 P1 closure), skos Walton mappings 70. It also corrects a near-miss of my own: `grep attackType` returns 0 because the properties are `aifAttackType`/`aifAttackedNode`, so the AIF layer IS serialized; the inventory is the only reliable measurement. And in §3.1bis: the "OWL can be regenerated post-tag" line is now done (#787), and the 46/1408 figure is marked as #753's deliverable rather than the current 145/1408. No verdict, status, arbitration or editorial content touched -- figures only, in the document's own dated-refresh style (v4, v4.1, v5, v5.1 → v5.2). Method rule recorded in the section: a byte-exact figure must name its commit AND its representation, a count must name its pattern. Otherwise the same object reads 1,407 or 1,408, and 1,081 or 1,977 or 1,985. Refs #134, #133, #763, #948 Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>

jsboige mentioned this pull request Jul 8, 2026

docs(taxonomy): #499 Virtues AIF mirror script — [secondaire] back-fill prep, gated execution (anti-drift with PR #753) #754

Merged

jsboige merged commit d4fde74 into master Jul 8, 2026
3 checks passed

jsboige deleted the feat/498-fallacies-aif-columns branch July 8, 2026 11:11

This was referenced Jul 8, 2026

🗺️ Coordination roadmap — v0.9.0 release + DNN go-live (autonomous cluster) #458

Open

feat(taxonomy): #499 Virtues AIF — apply mirror back-fill (plan #750 v2) into 2 columns [206 undercut/13 undermine/3 rebut] #755

Merged

jsboige mentioned this pull request Jul 8, 2026

chore(taxonomy): #498 phase 1-3 serialization audit (read-only diagnostic — backs ASK ai-01) #756

Merged

This was referenced Jul 9, 2026

chore(taxonomy): #498 phase 1-3 apply script (prep, gated --write, D2-paramétrable) #757

Merged

docs(release): #134 refresh v5 dossier validation — colonnes AIF relationnelles #753/#754/#755 + master 81a9e4e6 #758

Merged

jsboige mentioned this pull request Jul 9, 2026

feat(taxonomy): #498 Fallacies AIF — serialize scaleup phase 1-3 (47 leaves, MODE 2) #760

Merged

This was referenced Jul 10, 2026

docs(taxonomy): #498 Fallacies AIF — Relativisme abusif cluster (first rebut cluster, gated proposal) #766

Merged

docs(taxonomy): #498 coverage-status refresh — two-layer finding + reconciliation trajectory #768

Merged

jsboige mentioned this pull request Jul 10, 2026

docs(taxonomy): #498 AIF reconciliation P1 tranche-1 — 14 skos-only rows typed (gated proposal + dry-run apply) #769

Merged

jsboige mentioned this pull request Jul 11, 2026

feat(taxonomy): #498 AIF — serialize tranche-1c prod write (7 SUFFIX-ONLY, Erreur de raisonnement, 114→121) #779

Merged

clusterManager-Myia mentioned this pull request Jul 11, 2026

feat(taxonomy): #498 Fallacies AIF — P1 tranche-1d (4 SUFFIX-ONLY, famille Erreur mathématique, gated) #778

Merged

This was referenced Jul 12, 2026

feat(ontology): #133 OWL prep — regen argumentum.owl (AIF attack 93→145) + P1 closure doc #787

Merged

docs(ontology): AIF attack-graph export — typed bipartite edges + relations (read-only, 0 write) #828

Merged

jsboige mentioned this pull request Jul 21, 2026

docs(taxonomy): #498 AIF cluster « Complication exagérée » — first mixed undercut/undermine (6 leaves, proposition) #837

Merged

jsboige mentioned this pull request Jul 26, 2026

docs(release): #134 refresh v5.2 — §3.4 OWL, three unreproducible figures rectified #950

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

feat(taxonomy): #498 Fallacies AIF — serialize §7 I/RA/CA into 2 new columns (AIF_attackType + AIF_attackedNode) [reassigned from po-2023] - #753

feat(taxonomy): #498 Fallacies AIF — serialize §7 I/RA/CA into 2 new columns (AIF_attackType + AIF_attackedNode) [reassigned from po-2023]#753
jsboige merged 1 commit into
masterfrom
feat/498-fallacies-aif-columns

jsboige commented Jul 8, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Conversation

jsboige commented Jul 8, 2026

Summary

What changed

Derivation

Distribution (46 of 1408 rows filled)

⚠ OPEN QUESTION for review (the FAIL-LOUD interpretation)

Surgical byte-exact (the load-bearing guarantee)

Tests

Audit / reproducibility

Discipline

Request

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant