docs(taxonomy): #499 Virtues AIF PR-3 — Échange 7.3 Respect cluster + rebut-rarity finding

[jsboige]

Summary

3rd cluster of Virtues AIF I/RA/CA decomposition (schema #707§4 Option (a) ratified). Covers family 7 Échange enrichissant / subfamily 7.3 Respect de la personne (3 leaves pk 208/212/216, scheme: Argument from Commitment).

Follows PR #741 (PR-1, 4.1 Causalités, uniform undercut) and PR #742 (PR-2, 6.1 Fidélité, mixed undermine+undercut). Same method, new family. Docs-only, no code/CSV/DB/OWL writes.

What this PR adds

• Cluster 7.3 modeled: 3 leaves, 6 fallacy-instances opposed, all undercut/RA-node (interpersonal fallacies short-circuit the commitment→evaluation/dialogue inference).
• Mirror Fallacies: Obstruction 7.x primary (1398 ad hominem, 1371 genetic, 1352 poisoning-the-well) + cross-family (322 repoussoir/Influence, 942/Tricherie).
• Honest scheme-divergence (docs(taxonomy): #141 crossLink enrichment sample run (real output curated from #626) #677): path-prefix 7.3↔7.3 is PARTIAL (4/6 Obstruction) — not forced into 1:1 like PR-1/PR-2.
• PK 942 cross-reference PR-2: resisted by BOTH 155 (epistemic credibility) and 212 (dialogic loyalty) — 2 distinct CQs, modeled as 2 arcs, not deduplicated.

🔑 Key cumulative finding (load-bearing for Option decision)

Across PR-1 + PR-2 + PR-3 = 3 families, 3 subfamilies, 9 leaves, ~14 fallacy-instances:

attack-type	count
undercut	14
undermine	1
rebut	0

rebut/CA-node was NEVER cleanly instantiated. Structural reason: fallacies are inference-errors (undercut) or false-premises (undermine), rarely independent counter-arguments. PK 1398 (ad hominem) is defined in the taxonomy itself as « au lieu de réfuter » — the explicit negation of rebuttal.

Implication for Option A: the AIF_attackType column will be undercut+undermine-dominated, with rebut as a sporadic tail case. Recommend not over-investing rebut coverage in validator tooling — its absence is the domain's real structure, not a modeling gap. This reinforces Option A: « good holding of a scheme » = resistance to inference-breakdown + premise-falsity.

Limit: 3 clusters don't cover all 8 families. rebut may surface in unsampled families (1.x Argument pertinent, 2.x Présentation). Finding is directional — recommend 1-2 more varied-scheme clusters before locking the Option decision.

Discipline

• ✅ Read body before action: existing phase-2 relational annotations (echange-b) reused; Fallacies descriptions read (322/1398/1371/1352/942/1297/3/33) before deriving attack-type; anti-duplication check confirmed phase-2 docs are relational-only (no I/RA/CA decomposition — that's the additive layer here).
• ✅ 0 fabrication (docs(taxonomy): #141 crossLink enrichment sample run (real output curated from #626) #677): undercut/undermine/RA-node/I-node/Argument from Commitment = all native AIF/Walton tokens.
• ✅ 0 code change, 0 write Cards/, 0 regen, 0 CSV/DB/OWL. Docs-only (+204 lines).
• ✅ Option A/B/C unchanged — doc written under A; only 2 final columns are Option-sensitive. The I/RA/CA decomposition + rebut-rarity finding are Option-independent.
• ❌ 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).

Test impact

None — docs-only. No build/test counter change.

Request

Review + (if acceptable) hold for jsboige Option A/B/C decision before back-fill. This PR + #741 + #742 give jsboige 3 contrast points (uniform undercut / mixed undermine+undercut / rebut-absent) to decide. cc @jsboige

🤖 Worker po-2024 — cluster 3/3 of dispatch ompwhx (idle-de-secours mandate: continue #499 cluster-by-cluster).