docs(taxonomy): #498 PR-12 Equivoque / reification & magic-of-words cluster (2 modeled + 7 FAIL-LOUD) — completes sub-sub

[jsboige]

⚠ STATUS = PROPOSITION ONLY. No prod CSV write, no DB write, no OWL regen. jsboige ratifies → worker applies CSV edits in a follow-up PR (gated).

Scope

Twelfth cluster of the #498 AIF chantier. Third and final installment of the Equivoque sub-sub (Ambiguïté subfamily): the reification / magic-of-words / level-confusion mechanism (pk 867-875, 9 leaves) — the largest remaining Equivoque block.

Triggered by ai-01 dispatch 5ymo70 (tertiary: continue clusters in autonomy). Base bdba45d8. Completes the Equivoque sub-sub (after PR-10 #717 polysemy, PR-11 #718 residual).

Result

9 rows: 2 fully-modeled + 7 FAIL-LOUD, 0 fabrication. The reification block is the chantier's most gap-dense cluster (7/9 leaves).

pk	leaf	scheme	status
867	Confusion type-exemple	Example_Inference	direct-conflict (first DirectRef use of native token)
875	Motte castrale	VagueVerbalClassification_Inference	direct-conflict (PR-10 polysemy reuse)
868-874	Idola fori, Reification, referential, magic-of-words, map/territory, use/mention, simulacre	FAIL-LOUD	7th ontological gap (cluster)

The 7-leaf level-confusion family = 7th ontological gap

All 7 leaves commit a semantic-ontological category error (language/reality, abstract/concrete, word/object, word-exists/referent-exists, model/reality, use/mention, real/idealized). The candidate native schemes (VerbalClassification_Inference family) model classification-via-term defects; reification is an existence/reference claim — stretching a classification scheme over it violates #677. The gap is argumentative (genuine fallacies, unlike PR-11's pre-argumentative mondegreen). AIF has no scheme for the "language-creates-reality" pattern. Reification_Inference/Reference_Inference/UseMention_Inference flagged as hypothetical ontology-extension candidates.

Findings

• Largest single gap cluster (7 leaves) — more than all 5 prior gaps combined. The canonical "abuse of language" mechanism (Bacon's Idola fori) is absent from AIF.
• Motte-and-bailey = PR-10 polysemy specialization (875). Second cross-PR-10 reuse.
• First Example_Inference DirectRef use (867) — flagged for audit.
• Refines gap typology: first mechanism-wide argumentative gap.

Accounting — uniform-strict headline (po-2024 #716)

+2 strict this PR, 111 cumulative. 17 partial/gap tracked: 5 CA-missing (116 potential), 4 RA-missing (deferred), 8 FAIL-LOUD (861 perceptual + 868-874 reification).

Equivoque sub-sub COMPLETE (21 rows)

PR-10 (polysemy 5) + PR-11 (residual 4+1 gap) + PR-12 (reification 2+7 gaps) = 11 modeled + 8 gaps + 2 templates. Third fully-accounted subfamily-level block; largest single block closed.

Gate boundaries (HARD)

• No prod CSV/DB/OWL write, no aif: assertion. No [runtime pending] feat(dnn): #649 RulesExplorer i18n — mirror loc() culture cascade #674/chore(dnn): #131 migrate 12 RazorComponent templates to Razor14 (source-level, runtime pending) #596/[HOLD post-tag] fix(cardpen): #629 convert relative /Cards paths to raw-master URLs (Option 1) #666/feat(mindmap): Virtues 8-lang parity — #665 code wiring + fr/en/ru/pt/es Batik SVGs (#636) #686 merge.
• 0 fabrication — both proposed tokens native; 7 FAIL-LOUD documented with rejected weak alternatives.
• In-sub-sub scheme reuse, independent of I/RA/CA serialization decision (docs(taxonomy): #498 AIF chantier — coverage-status synthesis for jsboige ratification #707 §4 ratified option (a), post-tag).

Relates: #498, PR-1..PR-11, #707, #709 #716, #712, #677, #133/#130, #499, #458.

Generated with Claude Code