Skip to content

docs(taxonomy): #499 Virtues phase 2 batch 3 — Honnêteté intellectuelle leaves (gated) - #534

Merged
jsboige merged 1 commit into
masterfrom
feat/499-virtues-scaleup-phase2-honnetete
Jun 18, 2026
Merged

docs(taxonomy): #499 Virtues phase 2 batch 3 — Honnêteté intellectuelle leaves (gated)#534
jsboige merged 1 commit into
masterfrom
feat/499-virtues-scaleup-phase2-honnetete

Conversation

@jsboige

@jsboige jsboige commented Jun 18, 2026

Copy link
Copy Markdown
Contributor

#499 — Virtues Scale-up, Phase 2, batch 3: Honnêteté intellectuelle (gated proposal)

No production CSV change. Gated proposal — ai-01 reviews structure, jsboige validates content. This PR adds the Phase-2 depth-3..7 leaf annotation for the Honnêteté intellectuelle family only.

Scope

Batch 3 of Phase 2: Honnêteté intellectuelle (segment 6, 23 leaves: 9 depth-3 + 12 depth-4 + 2 depth-5), mirroring Tricherie (Fallacy family 887).

Method (identical to batches 1 & 2)

  • Model: gpt-5.5 (/v1/responses, reasoning.effort=low, no temperature).
  • Grounding catalog: 7 Fallacy families + 63 named depth-3 fallacies (reused verbatim from Phase 1). Prompt forbids any PK outside the catalog.
  • Family mirror hard constraint: prevented_family_pk = 887 for every row.

Anti-fabrication guarantee

Three independent layers, all re-checked against the REAL corpus (not the prompt catalog):

  1. Catalog membership — every opposed PK ∈ the 63 depth-3 set.
  2. Ground-truth — every PK re-verified against the real 1408-row Fallacies CSV (text_fr), opposed-PK↔name cross-checked character-for-character.
  3. Mirror consistency — every row's prevented_family_pk = 887.

Result: 23/23 annotated, 0 violations across all three layers. [PASS]

Anti-fab validator #518 (tools/validate_taxonomy_annotations.py, kind=virtues): ✓ CLEAN — 23 rows, 0 HARD, 0 WARN.

Semantic check

Scheme distribution is Bias-dominant (11/23) — the expected profile for the intellectual-honesty family whose opposite is cognitive bias and selective attention. This contrasts with batch 1 (Langage exact → Verbal Classification dominant) and batch 2 (Rigueur math → Sign dominant): the scheme shifts per family, confirming semantic fidelity rather than generic pattern-matching.

Files

  • docs/taxonomy/499-virtues-scaleup-phase2-honnetete.md — proposal
  • docs/taxonomy/499-scaleup-phase2-honnetete-annotations.csv — 23-row annotation (10-col presentation schema)

Generation scripts / raw model output kept ephemeral in tmp/ (not committed).

Phase 2 progress

Family leaves status
Langage exact 14 ✅ batch 1 merged (fc2013fc)
Rigueur mathématique 16 ⏳ batch 2 (PR #532)
Honnêteté intellectuelle 23 this PR
Raisonnement valide 51 next (sub-batched)
Échange enrichissant 40 next (sub-batched)
Argument pertinent 29 next
Présentation intègre 21 next

Cumulative: 53/194 leaves (27%). No prod write until jsboige approves the full Phase-2 set.

Closes nothing standalone — part of #499.

🤖 Generated with Claude Code

…ellectuelle depth-3..5 leaves (gated)

Gated proposal, no production CSV change. Batch 3 of Phase 2 depth-3..7 leaf
layer: Honnêteté intellectuelle family (segment 6, 23 nodes: 9 d3 + 12 d4 +
2 d5), mirror Tricherie (Fallacy family 887).

- 23/23 annotated via gpt-5.5 (/v1/responses, effort=low), 2 depth-3
  fallacies opposed per leaf, Walton scheme + CQ restored.
- 3-layer anti-fab verify: catalog membership + ground-truth vs real 1408-row
  Fallacies CSV (text_fr) + mirror consistency — 0 violations.
- Anti-fab validator #518 (kind=virtues): CLEAN — 23 rows, 0 HARD, 0 WARN.
- Scheme distribution Bias-dominant (11/23), expected for the honesty family
  opposite cognitive bias — distinct from batch 1 (Verbal Classification) and
  batch 2 (Sign), confirming per-family semantic fidelity.

Files: docs/taxonomy/499-virtues-scaleup-phase2-honnetete.md + CSV only.
Generation scripts/raw output kept ephemeral in tmp/ (not committed).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

@clusterManager-Myia clusterManager-Myia left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

[NanoClaw]

LGTM. Batch 3 Honnetete intellectuelle — 23 depth-3/4/5 leaves sous Tricherie (887). Meme methode/schema/anti-fab que batches 1-2, validator CLEAN sur 23 rows, gated proposal. Coherent avec #530 (batch 1) et #532 (batch 2). Ready pour jsboige validation.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants