feat(taxonomy): #499 Virtues Phase 2 batch 7b — Raisonnement valide pks 105-127 (23 leaves, depth-6/7)

[jsboige]

#499 Virtues Phase 2 — Batch 7b: Raisonnement valide, sub-batch B (pks 105–127)

GATED PROPOSAL — ai-01 reviews structure, jsboige validates content. No production Virtues CSV change until jsboige approves.

Scope

Second sub-batch of the last and largest Phase-2 family: Raisonnement valide (segment 4, 51 leaves). This batch = 7b (23 leaves, pks 105–127): the syllogistic-modes cluster — 4 depth-6 figure anchors + 19 depth-7 modes (Barbara, Celarent, Darii, Ferio, Cesare, Camestres, Festino, Baroco, Darapti, Felapton, Disamis, Datisi, Bocardo, Ferison, Camenes, Dimatis, Fesapo, Fresison, Bamalip). This reaches depth-7 — the deepest layer of the entire Virtues taxonomy. Mirrors Erreur de raisonnement (Fallacy family PK 696).

Sub-batch	pks	leaves	content
7a	81–104	22	premises · causality · formal validity · inference rules (PR #540)
7b (this PR)	105–127	23	syllogistic modes — 19 depth-7 (Barbara/Celarent/…)
7c (next)	128–133	6	informal/analytical reasoning

Result

• 23/23 annotated, 0 violations across all 3 anti-fab layers (catalog membership, ground-truth vs the real 1408-row Fallacies CSV [PK/text_fr], mirror consistency = 696).
• Anti-fab validator tools(taxonomy): reusable anti-fab validator for scale-up CSVs #518 (tools/validate_taxonomy_annotations.py, kind=virtues): ✓ CLEAN — 23 rows, 0 HARD, 0 WARN.
• Scheme distribution: Rule ×23 (100%) — total single-scheme dominance, the strongest of all batches (exceeds 7a's 17/22). Every syllogistic mode is, by definition, the good holding of a Rule scheme of syllogistic form.
• Syllogisme invalide 784 appears in all 23 rows (100%): each valid mode is the precise inverse of an invalid syllogism. Secondary opponent Erreur de quantification 735 (×16) names the distribution errors; Conclusion hâtive 759 (×4) and Erreur de logique propositionnelle 727 (×3) catch existential-import / compound-form errors.
• Per-mode CQ specificity at depth-7: each CQ names the exact premise quantity/quality of that mode (Barbara = two universal affirmatives → universal affirmative conclusion; Celarent = universal negative + universal affirmative → universal exclusion). gpt-5.5 reproduces the classical syllogistic mnemonics with correct figure and mood — semantic fidelity at the deepest technical layer of the whole taxonomy.

Files (2 — proposal only, no prod data)

• docs/taxonomy/499-virtues-scaleup-phase2-raisonnement-b.md — proposal document
• docs/taxonomy/499-scaleup-phase2-raisonnement-b-annotations.csv — 23-row annotation CSV (10-col presentation schema, UTF-8 BOM)

What this does NOT do

• ❌ Does not edit the production Argumentum Virtues - Taxonomy.csv — awaiting jsboige approval.
• ❌ Does not touch OWL, EPITA consumer, cards, or mindmaps.
• ❌ Does not complete the family — 7c (pks 128–133, 6 leaves) follows to finish Phase 2 (194/194 leaves), then the gated prod write (12-col population) proceeds on jsboige's approval.

Method (unchanged since batch 1)

gpt-5.5 via /v1/responses (reasoning.effort=low, max_output_tokens=7000, no temperature), grounded on the 7 Fallacy family PKs + 63 named depth-3 fallacies (reused verbatim from Phase 1). The prompt forbids referencing any PK outside this catalog; every opposed PK + family PK is re-verified character-for-character against the real Fallacies corpus.

🤖 Worker myia-po-2024 · taxonomy lane · #499