Skip to content

docs(taxonomy): #498 AIF scale-up phase 3 — Influence + Tricherie + Obstruction (27 fallacies, triple-AIF, rebut-type appears) - #546

Merged
jsboige merged 1 commit into
masterfrom
feat/498-aif-scaleup-phase3
Jun 19, 2026
Merged

docs(taxonomy): #498 AIF scale-up phase 3 — Influence + Tricherie + Obstruction (27 fallacies, triple-AIF, rebut-type appears)#546
jsboige merged 1 commit into
masterfrom
feat/498-aif-scaleup-phase3

Conversation

@jsboige

@jsboige jsboige commented Jun 19, 2026

Copy link
Copy Markdown
Contributor

#498 AIF Scale-up — Phase 3: Influence + Tricherie + Obstruction (27 fallacies, triple-AIF)

GATED PROPOSALdocs/taxonomy/ only (md + annotation CSV). No production Fallacies CSV change. Bases on master, independent of phase 2 (#542) merge — reviewable on its own merits. Awaiting ai-01 structural review + jsboige content approval.

Scope

27 depth-3 fallacies = 9 leaves × 3 families (Influence pks 177–511, Tricherie pks 889–1242, Obstruction pks 1282–1398). These three families were entirely unannotated before this batch. With phases 1+2 (23 fallacies) + phase 3 (27), AIF depth-3 coverage reaches 50/63 (79%).

The rebut type finally appears (the headline)

Phases 1–2 (evidence/premise + rule families) produced rebut 0/23 — every fallacy was a bad premise (undermine) or inapplicable rule (undercut). Phase 3 introduces the rebut attack-type (3 cases), exactly where the rhetorical/divergence families license it:

pk Fallacy Type CA_node
1282 Relativisme abusif rebut → conclusion CounterConclusionConflict
1313 Évasion rebut → conclusion IssueAvoidance
1361 Procès en incohérence rebut → conclusion TuQuoque

Distribution: undermine 20 / undercut 4 / rebut 3

Family undermine undercut rebut signature
Influence (9) 9 0 0 pure persuasion = loaded premise (schemes disperse across persuasion catalog)
Tricherie (9) 6 3 0 deception = false premise + double-standard rule shifts (special-pleading trio)
Obstruction (9) 5 1 3 derailment = source attacks + counter-claims/tu-quoque

Verification (3 independent layers, re-checked vs real 1408-row Fallacies CSV)

  • 27/27 annotated, 0 violations. Ground-truth: all 27 present, unique, depth=3, correct family.
  • ASPIC+ coherence (attack_type ↔ attacked_component): PASS.
  • Anti-fab validator tools(taxonomy): reusable anti-fab validator for scale-up CSVs #518: ✓ CLEAN — 27 rows, 0 HARD, 0 WARN (kind=aif-scaleup).
  • md ↔ CSV char-for-char: 0 issues (table built from CSV ground truth, curly apostrophes preserved — same paraphrase-bug guard as batches 7a/6b).

Files

  • docs/taxonomy/498-aif-scaleup-phase3.md (proposal)
  • docs/taxonomy/498-scaleup-phase3-annotations.csv (27-row annotation)

What this does NOT do

  • ❌ Does not edit production Argumentum Fallacies - Taxonomy.csv.
  • ❌ Does not complete depth-3 coverage — 13 stragglers remain (phase 4 mopup: 4 Abus langage + 3 Err raisonn + 3 Err math + 3 Insuffisance).

Phase progression (ASPIC+ prediction confirmed)

  • Phase 1: undercut-dominant (rule/language families)
  • Phase 2: undermine 6 / undercut 6 / rebut 0 (evidence/premise families)
  • Phase 3: undermine 20 / undercut 4 / rebut 3 (rhetorical/divergence families) ← rebut appears here

🤖 Worker myia-po-2024

…bstruction (27 fallacies, triple-AIF)

Phase-3 triple-AIF: the rhetorical/divergence families (Influence + Tricherie +
Obstruction, 27 depth-3 fallacies, 9 each). These three families were entirely
unannotated before this batch. With phases 1+2 (23 fallacies) and phase 3 (27),
AIF depth-3 coverage reaches 50/63 (79%).

Distribution: undermine 20 / undercut 4 / rebut 3. Phase 3 is where the rebut
attack-type (absent in phases 1-2: 0/23) finally appears — exactly where the
divergence families license it: Relativisme abusif 1282 (competing-truth
counter-claim), Évasion 1313 (issue substitution), Procès en incohérence 1361
(tu-quoque counter). The undermine dominance (20/27) is the Tricherie/Obstruction
signature — false/biased premises, credibility attacks, bare assertions.

Family signatures:
- Influence (9): pure undermine — every fallacy counterfeits a persuasion scheme
  via a loaded/emotion-laden premise (schemes disperse across the persuasion
  catalog: Verbal Classification, Popular Opinion, Values x2, Bias, Consequences,
  Sign x2, Commitment).
- Tricherie (6 undermine / 3 undercut): deception splits into false/biased
  premises (Mensonge, fausse attribution, attention sélective, biais x3) and
  double-standard rule shifts (Exigence renforcée/relâchée, Vouloir le beurre) —
  the undercut cluster is the special-pleading trio.
- Obstruction (5 undermine / 3 rebut / 1 undercut): the only rebut-bearing
  family. 3 rebut = the genuine counter-conclusion fallacies; undermine =
  credibility attacks (attaque personnelle, sophisme génétique, empoisonnement
  du puits) + bare assertion.

Verification (3 independent layers, re-checked vs real 1408-row Fallacies CSV):
27/27 annotated, 0 violations. Ground-truth: all 27 present, unique, depth=3,
correct family. ASPIC+ coherence (attack_type <-> attacked_component) PASS.
Anti-fab validator #518: CLEAN — 27 rows, 0 HARD, 0 WARN (kind=aif-scaleup).
md <-> CSV char-for-char consistency: 0 issues (table built from CSV ground truth).

GATED PROPOSAL — no production Fallacies CSV change. docs/taxonomy/ proposition only
(md + annotation CSV). Bases on master, independent of phase 2 merge. Awaiting
ai-01 structural review + jsboige content approval.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@jsboige
jsboige merged commit 4dd3c6b into master Jun 19, 2026
3 checks passed
@jsboige
jsboige deleted the feat/498-aif-scaleup-phase3 branch June 19, 2026 11:07
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.

1 participant