From c96051b0a7c7d408bc1b56e4e2453c03c7c3f40a Mon Sep 17 00:00:00 2001 From: jsboige Date: Sun, 5 Jul 2026 08:56:25 +0200 Subject: [PATCH 1/4] =?UTF-8?q?docs(taxonomy):=20#498=20AIF=20chantier=20P?= =?UTF-8?q?R-1=20=E2=80=94=20Fallacious=20comparison=20cluster=20(proposit?= =?UTF-8?q?ion,=20gated)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit First cluster of the #498 chantier (ai-01 dispatch yj7u3j primary). Proposition only — GATED, 0 write to prod CSV. Multi-session chantier; this PR advances one coherent cluster. Cluster: 'Fallacious comparison' subfamily (Misleading language family), code=truth from Cards/Fallacies/Argumentum Fallacies - Taxonomy.csv: - pk 839 (Fausse analogie) = MAPPED anchor = the cluster template: Analogy_Inference scheme defeated by DifferencesUndermineSimilarity_Conflict CQ. - 5 unmapped leaves (838, 840-843) modeled as defeasible exceptions to legitimate schemes, in the Walton/AIF paradigm #498 asks for. Proposals (per leaf, scheme + exception/CQ + MappingType): - 841 (Analogie etendue): Analogy_Inference + DifferencesUndermineSimilarity (transitive-chain variant), narrowMatch. - 842 (similarite fallacieuse): Analogy_Inference + DifferencesUndermineSimilarity (mere-resemblance), broadMatch. - 843 (Fausse equivalence): Analogy_Inference + DifferencesUndermineSimilarity (argument-level variant), closeMatch. - 838 (Distinction sans difference): HONEST scheme-divergence — actually a classification fallacy (ArbitraryVerbalClassification_Inference + PropertyNotExistant_Conflict), broadMatch. - 840 (Petition analogique): FAIL LOUD — no AIF native circularity-CQ for analogy exists; documented in AIF_skosOther, no fabricated Conflict token. Coverage: 70 -> 75 mapped (cluster-level). Leaf-level 100% explicitly not the goal (#498 DoD). OWL regen deferred (jsboige ratifies -> CSV edits -> OWL #133). No prod CSV write, no DB write, no regen. Rappels HARD tenus. Co-Authored-By: Claude-Code --- .../498-aif-fallacious-comparison-cluster.md | 197 ++++++++++++++++++ 1 file changed, 197 insertions(+) create mode 100644 docs/taxonomy/498-aif-fallacious-comparison-cluster.md diff --git a/docs/taxonomy/498-aif-fallacious-comparison-cluster.md b/docs/taxonomy/498-aif-fallacious-comparison-cluster.md new file mode 100644 index 00000000..96657b31 --- /dev/null +++ b/docs/taxonomy/498-aif-fallacious-comparison-cluster.md @@ -0,0 +1,197 @@ +# 2026-07-05 — #498 AIF chantier, PR-1 : cluster « Fallacious comparison » (Misleading language) + +**Scope**: first cluster of the #498 chantier — model the **Fallacious comparison** subtree (Misleading +language family) as **defeasible exceptions to legitimate argument schemes**, in the Walton/AIF +paradigm. **Proposition only — GATED, 0 write to prod CSV.** Triggered by ai-01 dispatch `yj7u3j` +(primary). This is a multi-session chantier; this PR advances one coherent cluster. + +**Repo reference**: master `dc02e847`. Issue: #498 (reformulated, GO jsboige 2026-06-17 verified). +Linked: #133/#130 (existing OWL ontology, regenerated to reflect new structures — not created). + +> **⚠ STATUS = PROPOSITION ONLY.** No prod CSV write, no DB write, no OWL regen. This doc proposes +> the AIF structure (scheme + exception/CQ) for the cluster's unmapped leaves, derived code=truth +> from the CSV. jsboige ratifies → worker applies the CSV edits in a follow-up PR (gated). + +--- + +## TL;DR + +The **Fallacious comparison** subfamily (`Subfamily=Fallacious comparison`, Misleading language +family) has **1 mapped parent** (pk 839, "Fausse analogie") and **5 unmapped leaves** (pk 838, 840, +841, 842, 843). The parent establishes the cluster's pattern — **exception to `Analogy_Inference` +via the `DifferencesUndermineSimilarity_Conflict` critical question** — and the leaves are +specialized variants of the same defeasible structure. This PR proposes the AIF exception-modeling +for the 5 leaves, reusing the parent's scheme where honest, and **failing loud** where a leaf does +not honestly fit the Analogy_Inference scheme ( documenting it as such rather than fabricating). + +--- + +## 1. Cluster state (code=truth) + +From `Cards/Fallacies/Argumentum Fallacies - Taxonomy.csv`, family **Misleading language**, subfamily +**Fallacious comparison**: + +| pk | text_fr | Latin | DirectRef | ExceptionRef | MappingType | State | +|----|---------|-------|-----------|--------------|-------------|-------| +| 834 | Comparaison abusive (depth-3 parent) | — | — | — | — | unmapped (subfamily root) | +| 835 | Comparaison incomplète | — | — | — | — | unmapped (Faulty comparison sub-sub) | +| 836 | Classification non exclusive | — | — | — | — | unmapped | +| 837 | Comparaison incohérente | — | — | — | — | unmapped | +| 838 | Distinction sans différence | — | — | — | — | unmapped | +| **839** | **Fausse analogie** (depth-3) | — | `DifferencesUndermineSimilarity_Conflict` | `Analogy_Inference` | `skos:closeMatch` | **✅ MAPPED (the cluster model)** | +| 840 | Pétition de principe analogique | — | — | — | — | unmapped (False analogy sub-sub) | +| 841 | Analogie étendue | — | — | — | — | unmapped | +| 842 | Argument de la similarité fallacieuse | — | — | — | — | unmapped | +| 843 | Fausse équivalence | — | — | — | — | unmapped | + +The mapped parent (839) is the **template**: a fallacious comparison = an `Analogy_Inference` scheme +**defeated** because the critical question "do the differences between the compared items undermine +the similarity?" (`DifferencesUndermineSimilarity_Conflict`) is violated. The leaves are specialized +ways that analogy goes wrong. + +**This PR scopes to the `False analogy` sub-subfamily (pk 838-843)** — the tightest coherent cluster +with a mapped anchor. The `Faulty comparison` sub-subfamily (835-838) is a candidate for PR-2. + +--- + +## 2. The parent model (839, the template) + +**Legitimate scheme**: `Analogy_Inference` (AIF native) — "A is like B in property P; B has property +Q; therefore A has property Q." (Walton's argument from analogy.) + +**Critical question violated** (the defeaser): `DifferencesUndermineSimilarity_Conflict` — "Are there +relevant differences between A and B that undermine the transfer of Q?" The fallacy ignores these. + +**Mapping type**: `skos:closeMatch` — "Fausse analogie" is the informal-logic name for the violated-CQ +case of `Analogy_Inference`. + +This is **exactly the paradigm #498 asks for**: a fallacy = a defeasible exception to a legitimate +scheme, named by the violated critical question, not by a Latin label alone. + +--- + +## 3. Proposed AIF structure for the unmapped leaves + +For each leaf, the proposal: **(a) legitimate scheme targeted, (b) exception/CQ that defeats it, +(c) `AIF_skosMappingType`**. Reuse the parent's scheme where the leaf is a specialized analogy-failure; +**fail loud** where a leaf targets a different scheme honestly. + +### pk 840 — Pétition de principe analogique (Analogical begging the question) +- **desc_fr**: "Vous utilisez une analogie qui suppose déjà la conclusion que vous cherchez à + démontrer, en assimilant de manière discutable deux situations." +- **Legitimate scheme**: `Analogy_Inference` +- **Exception/CQ**: the analogy **begs the question** — the disputed conclusion (A and B are + assimilable) is smuggled into the analogy's premise. This is **not** the same CQ as 839 + (differences undermining similarity) — it's a circularity defeater. The closest AIF native conflict + node is a **question-begging / circularity** CQ. +- **Proposal**: `ExceptionRef=Analogy_Inference`, `DirectRef=` *(no clean AIF native circularity + Conflict node exists for analogy)* → **FAIL LOUD**: document that no AIF native CQ cleanly + captures "the analogy presupposes its conclusion"; propose `AIF_skosOther="Circularity: + analogy presupposes the disputed assimilation (no native AIF Conflict node)"`, + `MappingType=skos:closeMatch`. Do **not** fabricate a `*_Conflict` token. + +### pk 841 — Analogie étendue (Extended/Transitive analogy) +- **desc_fr**: "À partir d'une chose qui ressemble à une deuxième et partage une propriété avec une + troisième, vous concluez que la deuxième et la troisième se ressemblent." +- **Legitimate scheme**: `Analogy_Inference` (transitive chain: A~B, A~C ⟹ B~C) +- **Exception/CQ**: transitivity of similarity is not guaranteed — similarity is not transitive. + This **is** a clean AIF conflict: the differences between B and C undermine the inferred similarity. +- **Proposal**: `ExceptionRef=Analogy_Inference`, `DirectRef=DifferencesUndermineSimilarity_Conflict` + (same CQ as 839, applied to the B-C link rather than A-B), `MappingType=skos:narrowMatch` (a + narrower, transitive-chain variant of 839). + +### pk 842 — Argument de la similarité fallacieuse (Fallacious similarity) +- **desc_fr**: "Vous considérez que deux choses qui se ressemblent sont liées ou similaires." +- **Legitimate scheme**: `Analogy_Inference` (or more loosely `Sign_Inference` — resemblance as a + sign of relatedness) +- **Exception/CQ**: superficial resemblance does not establish a real link/causal relation. The CQ: + "is the resemblance merely superficial, without an underlying connection?" Closest AIF: + `DifferencesUndermineSimilarity_Conflict` (the resemblance is too thin to transfer any property). +- **Proposal**: `ExceptionRef=Analogy_Inference`, `DirectRef=DifferencesUndermineSimilarity_Conflict`, + `MappingType=skos:broadMatch` (the fallacy is broader — covers mere-resemblance-without-property- + transfer, a looser case than 839's property-transfer analogy). + +### pk 843 — Fausse équivalence (False equivalence) +- **desc_fr**: "Vous faites un parallèle impropre entre deux arguments distincts." +- **Legitimate scheme**: `Analogy_Inference` (two arguments treated as equivalent) +- **Exception/CQ**: the two arguments differ in a **relevant** respect (strength, structure, + evidence) that defeats their equivalence. This is again `DifferencesUndermineSimilarity_Conflict`, + but applied to **arguments** rather than objects — a well-known specialized case. +- **Proposal**: `ExceptionRef=Analogy_Inference`, `DirectRef=DifferencesUndermineSimilarity_Conflict`, + `MappingType=skos:closeMatch` (a direct, argument-level variant of 839). + +### pk 838 — Distinction sans différence (Distinction without difference) +- **desc_fr**: "Vous distinguez deux éléments qui sont en réalité semblables ou qui ne présentent + aucune différence significative." +- **⚠ Inverse of 839**: this is the **mirror fallacy** — claiming a distinction where none exists + (the opposite error from false analogy, which claims similarity where differences exist). It is + **not** an exception to `Analogy_Inference` — it's a failure of **differentiation**. +- **Legitimate scheme**: `Classification_Inference` / verbal-classification (claiming two items are + in different classes when they share the relevant properties). +- **Exception/CQ**: the classification is arbitrary — the distinguishing property is not relevant to + the classification. Closest AIF: `ArbitraryVerbalClassification_Inference` (already used elsewhere) + with conflict `PropertyNotExistant_Conflict` (the distinguishing property doesn't obtain). +- **Proposal**: `ExceptionRef=ArbitraryVerbalClassification_Inference`, + `DirectRef=PropertyNotExistant_Conflict`, `MappingType=skos:broadMatch`. **This is the honest + scheme** — 838 is a classification fallacy, not an analogy fallacy, despite sitting in the + comparison subfamily. + +--- + +## 4. Cluster summary (proposed) + +| pk | text_fr | Proposed ExceptionRef | Proposed DirectRef | MappingType | Honest? | +|----|---------|----------------------|-------------------|-------------|---------| +| 839 | Fausse analogie | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | `skos:closeMatch` | ✅ (existing) | +| 840 | Pétition de principe analogique | `Analogy_Inference` | *(Other: circularity — no native CQ)* | `skos:closeMatch` | ⚠ FAIL LOUD (no native node) | +| 841 | Analogie étendue | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | `skos:narrowMatch` | ✅ | +| 842 | Argument de la similarité fallacieuse | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | `skos:broadMatch` | ✅ | +| 843 | Fausse équivalence | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | `skos:closeMatch` | ✅ | +| 838 | Distinction sans différence | `ArbitraryVerbalClassification_Inference` | `PropertyNotExistant_Conflict` | `skos:broadMatch` | ✅ (different scheme — honest) | + +**5 leaves proposed, 1 FAIL-LOUD (840)** — documents that AIF has no native circularity-CQ for +analogy rather than fabricating one. **1 honest scheme-divergence (838)** — different scheme than +the parent, documented. + +--- + +## 5. Method notes (for the chantier's后续 PRs) + +- **Cluster = a mapped parent + its unmapped sibling leaves** (depth-3 anchor + depth-4 leaves). This + PR's cluster (False analogy, 839 + 4 leaves + 1 mirror) is the unit of work per PR. +- **Reuse the parent's scheme + CQ** where the leaf is a specialized violation of the same scheme; + vary only the `MappingType` (`narrowMatch` = more specific, `broadMatch` = looser, `closeMatch` = + direct variant). +- **Fail loud** when no native AIF Conflict node captures the leaf's defeater (840) — use + `AIF_skosOther` to document, never fabricate a `*_Conflict` token. +- **Honest scheme-divergence** (838) — when a leaf in the comparison subfamily is actually a + classification fallacy, say so and use the right scheme; don't force `Analogy_Inference`. +- **Coverage accounting** (#498 DoD): this PR adds **5 mapped leaves** to the 70 existing → 75/1408. + At cluster-level (depth 2-3), the False-analogy sub-subfamily goes from 1/5 mapped to 5/5 (+1 + honest scheme-divergence documented). Effective cluster coverage of the comparison subfamily + improves; leaf-level "100%" is explicitly **not** the goal. + +--- + +## 6. DoD status + +| DoD item (#498) | Status | +|-----------------|--------| +| Effective ~31% cluster-level coverage documented (not "100% leaves") | ✅ §5 (70→75 mapped, cluster-level) | +| Per unmapped leaf: legitimate scheme + exception/CQ (not Latin alone) | ✅ §3 (5 leaves, scheme + CQ each) | +| `AIF_skosMappingType` coherent (broadMatch when narrower) | ✅ §4 (varied per leaf) | +| Fail-loud when no honest scheme fits | ✅ §3 pk 840 (no native circularity CQ — documented, not fabricated) | +| OWL regen reflects new structures | ⏳ Deferred — jsboige ratifies this proposition first, then CSV edits applied, then OWL regen (#133) — not in this PR | + +--- + +## Gate boundaries (HARD — proposition only) + +- ❌ No prod CSV write, no DB write, no OWL regen, no `aif:` assertion emitted. +- ❌ No #674/#596 merge, no régén launch. +- ✅ Proposition derived code=truth from the taxonomy CSV (cluster + AIF columns) + AIF scheme + semantics (Walton/`Analogy_Inference` + `DifferencesUndermineSimilarity_Conflict` are AIF native). +- ✅ Fail-loud on 840 (no fabricated Conflict node); honest scheme-divergence on 838. + +Relates: dispatch `yj7u3j` (primary), #498 (chantier), #133/#130 (existing OWL), #499 (inverse: +virtue = good tenor of a scheme), #192 (terminology), #458. From 94e55c8ca3e1ca37932a6a1a7cf4832f80bf87ad Mon Sep 17 00:00:00 2001 From: jsboige Date: Sun, 5 Jul 2026 10:59:31 +0200 Subject: [PATCH 2/4] =?UTF-8?q?fix(taxonomy):=20#498=20PR-1=20correct=20cl?= =?UTF-8?q?uster=20boundary=20=E2=80=94=20pk=20838=20is=20Faulty=20compari?= =?UTF-8?q?son,=20not=20False=20analogy?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Code=truth re-verification against CSV `Soussousfamille` column revealed 3 factual errors in the PR-1 proposition (caught before merge, still in review): 1. pk 838 ("Distinction sans différence") was placed in the False analogy sub-sub — it actually belongs to Faulty comparison. Its honest scheme-divergence modeling (classification fallacy, ArbitraryVerbalClassification_Inference) moves to PR-2. 2. The subfamily root was labeled pk 834 (d3) — the true d2 root is pk 833 "Comparaison fallacieuse" (MAPPED: BiasedClassification_Conflict directRef, skos:broadMatch), which was entirely missing. 3. False analogy leaf count was 5 — it is 4 (840-843). Coverage delta 70→74, not 70→75. Corrections: - TL;DR + §1 table rewritten with exact sub-sub assignments (13 rows, 3 sub-subs, 2 mapped nodes). - §1 adds an explicit "cluster-boundary correction" callout documenting the earlier error. - §3 removes the pk 838 subsection (deferred to PR-2); adds a pointer note. - §4 summary table reduced to 4 leaves (839 anchor + 840-843). - §5/§6 counters reconciled (70→74, 4 leaves, 1/5→5/5 sub-sub). - §5 method notes updated: cluster shape can be anchored by parent d2/d3 when no in-sub-sub anchor. Net effect: PR-1 methodology is strengthened — all 4 False-analogy leaves now honestly reuse Analogy_Inference with no in-cluster scheme-divergence. Gate unchanged: proposition only, 0 prod CSV write. Co-Authored-By: Claude-Code --- .../498-aif-fallacious-comparison-cluster.md | 128 ++++++++++-------- 1 file changed, 72 insertions(+), 56 deletions(-) diff --git a/docs/taxonomy/498-aif-fallacious-comparison-cluster.md b/docs/taxonomy/498-aif-fallacious-comparison-cluster.md index 96657b31..1bf915d9 100644 --- a/docs/taxonomy/498-aif-fallacious-comparison-cluster.md +++ b/docs/taxonomy/498-aif-fallacious-comparison-cluster.md @@ -17,12 +17,27 @@ Linked: #133/#130 (existing OWL ontology, regenerated to reflect new structures ## TL;DR The **Fallacious comparison** subfamily (`Subfamily=Fallacious comparison`, Misleading language -family) has **1 mapped parent** (pk 839, "Fausse analogie") and **5 unmapped leaves** (pk 838, 840, -841, 842, 843). The parent establishes the cluster's pattern — **exception to `Analogy_Inference` -via the `DifferencesUndermineSimilarity_Conflict` critical question** — and the leaves are -specialized variants of the same defeasible structure. This PR proposes the AIF exception-modeling -for the 5 leaves, reusing the parent's scheme where honest, and **failing loud** where a leaf does -not honestly fit the Analogy_Inference scheme ( documenting it as such rather than fabricating). +family) has **2 mapped nodes** — pk 833 (d2 subfamily root "Comparaison fallacieuse", +`BiasedClassification_Conflict` directRef, `skos:broadMatch`) and pk 839 (d3 "Fausse analogie", +`Analogy_Inference` defeated by `DifferencesUndermineSimilarity_Conflict`, `skos:closeMatch`) — +across 3 sub-sub clusters: **Faulty comparison** (834-838), **False analogy** (839-843), and +**Association fallacy** (844-845). **This PR scopes to the False analogy sub-sub**: anchor pk 839 + +**4 unmapped leaves** (pk 840, 841, 842, 843). The anchor establishes the cluster's pattern — +**exception to `Analogy_Inference` via the `DifferencesUndermineSimilarity_Conflict` critical +question** — and the 4 leaves are specialized variants of the same defeasible structure. This PR +proposes the AIF exception-modeling for the 4 leaves, reusing the anchor's scheme, and **failing +loud** where a leaf does not honestly fit the Analogy_Inference scheme (documenting it as such +rather than fabricating). + +> **⚠ Cluster-boundary correction (code=truth re-verification).** An earlier draft of this +> proposition incorrectly placed pk 838 ("Distinction sans différence") in the False analogy +> sub-sub and labeled pk 834 as the subfamily root. Re-verification against the CSV `Soussousfamille` +> column shows: (a) the true subfamily root is **pk 833** (d2, MAPPED — missing from the earlier +> draft); (b) **pk 838 belongs to Faulty comparison**, not False analogy. The False analogy sub-sub +> is strictly pk 839-843 (anchor + 4 leaves). pk 838 (and its honest scheme-divergence modeling) is +> deferred to **PR-2** (Faulty comparison cluster). This correction strengthens PR-1's methodology: +> all 4 False-analogy leaves now honestly reuse `Analogy_Inference`, with no scheme-divergence +> needed in this cluster. --- @@ -31,26 +46,31 @@ not honestly fit the Analogy_Inference scheme ( documenting it as such rather th From `Cards/Fallacies/Argumentum Fallacies - Taxonomy.csv`, family **Misleading language**, subfamily **Fallacious comparison**: -| pk | text_fr | Latin | DirectRef | ExceptionRef | MappingType | State | -|----|---------|-------|-----------|--------------|-------------|-------| -| 834 | Comparaison abusive (depth-3 parent) | — | — | — | — | unmapped (subfamily root) | -| 835 | Comparaison incomplète | — | — | — | — | unmapped (Faulty comparison sub-sub) | -| 836 | Classification non exclusive | — | — | — | — | unmapped | -| 837 | Comparaison incohérente | — | — | — | — | unmapped | -| 838 | Distinction sans différence | — | — | — | — | unmapped | -| **839** | **Fausse analogie** (depth-3) | — | `DifferencesUndermineSimilarity_Conflict` | `Analogy_Inference` | `skos:closeMatch` | **✅ MAPPED (the cluster model)** | -| 840 | Pétition de principe analogique | — | — | — | — | unmapped (False analogy sub-sub) | -| 841 | Analogie étendue | — | — | — | — | unmapped | -| 842 | Argument de la similarité fallacieuse | — | — | — | — | unmapped | -| 843 | Fausse équivalence | — | — | — | — | unmapped | - -The mapped parent (839) is the **template**: a fallacious comparison = an `Analogy_Inference` scheme +| pk | text_fr | Subsubfamily | DirectRef | ExceptionRef | MappingType | State | +|----|---------|--------------|-----------|--------------|-------------|-------| +| **833** | **Comparaison fallacieuse** (d2) | *(root)* | `BiasedClassification_Conflict` | — | `skos:broadMatch` | **✅ MAPPED (subfamily root)** | +| 834 | Comparaison abusive (d3) | Faulty comparison | — | — | — | unmapped (d3 anchor → PR-2) | +| 835 | Comparaison incomplète | Faulty comparison | — | — | — | unmapped (→ PR-2) | +| 836 | Classification non exclusive | Faulty comparison | — | — | — | unmapped (→ PR-2) | +| 837 | Comparaison incohérente | Faulty comparison | — | — | — | unmapped (→ PR-2) | +| 838 | Distinction sans différence | Faulty comparison | — | — | — | unmapped (→ PR-2) | +| **839** | **Fausse analogie** (d3) | False analogy | `DifferencesUndermineSimilarity_Conflict` | `Analogy_Inference` | `skos:closeMatch` | **✅ MAPPED (PR-1 anchor)** | +| 840 | Pétition de principe analogique | False analogy | — | — | — | unmapped (**PR-1**) | +| 841 | Analogie étendue | False analogy | — | — | — | unmapped (**PR-1**) | +| 842 | Argument de la similarité fallacieuse | False analogy | — | — | — | unmapped (**PR-1**) | +| 843 | Fausse équivalence | False analogy | — | — | — | unmapped (**PR-1**) | +| 844 | Sophisme d'association (d3) | Association fallacy | — | — | — | unmapped (d3 anchor → PR-3) | +| 845 | Amalgame | Association fallacy | — | — | — | unmapped (→ PR-3) | + +The mapped anchor (839) is the **template**: a fallacious comparison = an `Analogy_Inference` scheme **defeated** because the critical question "do the differences between the compared items undermine -the similarity?" (`DifferencesUndermineSimilarity_Conflict`) is violated. The leaves are specialized -ways that analogy goes wrong. +the similarity?" (`DifferencesUndermineSimilarity_Conflict`) is violated. The 4 leaves are +specialized ways that analogy goes wrong. -**This PR scopes to the `False analogy` sub-subfamily (pk 838-843)** — the tightest coherent cluster -with a mapped anchor. The `Faulty comparison` sub-subfamily (835-838) is a candidate for PR-2. +**This PR scopes to the `False analogy` sub-subfamily (pk 839-843)** — anchor 839 + 4 leaves — the +tightest coherent cluster with an in-sub-sub mapped anchor. The `Faulty comparison` sub-subfamily +(834-838, no in-sub-sub anchor → anchored by the d2 root 833) is **PR-2**; `Association fallacy` +(844-845) is **PR-3**. --- @@ -120,21 +140,12 @@ For each leaf, the proposal: **(a) legitimate scheme targeted, (b) exception/CQ - **Proposal**: `ExceptionRef=Analogy_Inference`, `DirectRef=DifferencesUndermineSimilarity_Conflict`, `MappingType=skos:closeMatch` (a direct, argument-level variant of 839). -### pk 838 — Distinction sans différence (Distinction without difference) -- **desc_fr**: "Vous distinguez deux éléments qui sont en réalité semblables ou qui ne présentent - aucune différence significative." -- **⚠ Inverse of 839**: this is the **mirror fallacy** — claiming a distinction where none exists - (the opposite error from false analogy, which claims similarity where differences exist). It is - **not** an exception to `Analogy_Inference` — it's a failure of **differentiation**. -- **Legitimate scheme**: `Classification_Inference` / verbal-classification (claiming two items are - in different classes when they share the relevant properties). -- **Exception/CQ**: the classification is arbitrary — the distinguishing property is not relevant to - the classification. Closest AIF: `ArbitraryVerbalClassification_Inference` (already used elsewhere) - with conflict `PropertyNotExistant_Conflict` (the distinguishing property doesn't obtain). -- **Proposal**: `ExceptionRef=ArbitraryVerbalClassification_Inference`, - `DirectRef=PropertyNotExistant_Conflict`, `MappingType=skos:broadMatch`. **This is the honest - scheme** — 838 is a classification fallacy, not an analogy fallacy, despite sitting in the - comparison subfamily. +> **Note on pk 838** ("Distinction sans différence"): this leaf sits in the **Faulty comparison** +> sub-sub (`Soussousfamille=Faulty comparison`), **not** False analogy — see §1 boundary correction. +> It is the mirror fallacy (claiming a distinction where none exists) and honestly targets a +> **classification** scheme (`ArbitraryVerbalClassification_Inference` + +> `PropertyNotExistant_Conflict`) rather than `Analogy_Inference`. Its full modeling is deferred to +> **PR-2** (Faulty comparison cluster), to keep this PR scoped to a single coherent sub-sub. --- @@ -142,34 +153,38 @@ For each leaf, the proposal: **(a) legitimate scheme targeted, (b) exception/CQ | pk | text_fr | Proposed ExceptionRef | Proposed DirectRef | MappingType | Honest? | |----|---------|----------------------|-------------------|-------------|---------| -| 839 | Fausse analogie | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | `skos:closeMatch` | ✅ (existing) | +| 839 | Fausse analogie | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | `skos:closeMatch` | ✅ (existing anchor) | | 840 | Pétition de principe analogique | `Analogy_Inference` | *(Other: circularity — no native CQ)* | `skos:closeMatch` | ⚠ FAIL LOUD (no native node) | | 841 | Analogie étendue | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | `skos:narrowMatch` | ✅ | | 842 | Argument de la similarité fallacieuse | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | `skos:broadMatch` | ✅ | | 843 | Fausse équivalence | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | `skos:closeMatch` | ✅ | -| 838 | Distinction sans différence | `ArbitraryVerbalClassification_Inference` | `PropertyNotExistant_Conflict` | `skos:broadMatch` | ✅ (different scheme — honest) | -**5 leaves proposed, 1 FAIL-LOUD (840)** — documents that AIF has no native circularity-CQ for -analogy rather than fabricating one. **1 honest scheme-divergence (838)** — different scheme than -the parent, documented. +**4 leaves proposed (840-843), 1 FAIL-LOUD (840)** — documents that AIF has no native circularity-CQ +for analogy rather than fabricating one. All 4 leaves honestly reuse the anchor's +`Analogy_Inference` scheme; the only variation is the violated-CQ specialization (840) and the +`MappingType` granularity. The honest scheme-divergence case (pk 838, classification fallacy) +moves to **PR-2** (Faulty comparison cluster), keeping this PR a single coherent sub-sub. --- ## 5. Method notes (for the chantier's后续 PRs) -- **Cluster = a mapped parent + its unmapped sibling leaves** (depth-3 anchor + depth-4 leaves). This - PR's cluster (False analogy, 839 + 4 leaves + 1 mirror) is the unit of work per PR. -- **Reuse the parent's scheme + CQ** where the leaf is a specialized violation of the same scheme; +- **Cluster = a sub-sub-family with an in-sub-sub mapped anchor + its unmapped sibling leaves.** + This PR's cluster (False analogy, anchor 839 + 4 leaves) is the unit of work per PR. Where a + sub-sub has no in-sub-sub anchor (e.g. Faulty comparison 834-838), it is anchored by the parent + d2/d3 node — that is PR-2's shape. +- **Reuse the anchor's scheme + CQ** where the leaf is a specialized violation of the same scheme; vary only the `MappingType` (`narrowMatch` = more specific, `broadMatch` = looser, `closeMatch` = direct variant). - **Fail loud** when no native AIF Conflict node captures the leaf's defeater (840) — use `AIF_skosOther` to document, never fabricate a `*_Conflict` token. -- **Honest scheme-divergence** (838) — when a leaf in the comparison subfamily is actually a - classification fallacy, say so and use the right scheme; don't force `Analogy_Inference`. -- **Coverage accounting** (#498 DoD): this PR adds **5 mapped leaves** to the 70 existing → 75/1408. - At cluster-level (depth 2-3), the False-analogy sub-subfamily goes from 1/5 mapped to 5/5 (+1 - honest scheme-divergence documented). Effective cluster coverage of the comparison subfamily - improves; leaf-level "100%" is explicitly **not** the goal. +- **Honest scheme-divergence** — when a leaf sits in a subfamily but actually targets a different + scheme (pk 838: a comparison-subfamily leaf that is genuinely a *classification* fallacy), say so + and use the right scheme; don't force the anchor's scheme. (838 lives in the Faulty comparison + sub-sub → its modeling is in PR-2, not here.) +- **Coverage accounting** (#498 DoD): this PR adds **4 mapped leaves** to the 70 existing → 74/1408. + At cluster-level, the False-analogy sub-sub goes from 1/5 mapped to 5/5. Effective cluster + coverage of the comparison subfamily improves; leaf-level "100%" is explicitly **not** the goal. --- @@ -177,8 +192,8 @@ the parent, documented. | DoD item (#498) | Status | |-----------------|--------| -| Effective ~31% cluster-level coverage documented (not "100% leaves") | ✅ §5 (70→75 mapped, cluster-level) | -| Per unmapped leaf: legitimate scheme + exception/CQ (not Latin alone) | ✅ §3 (5 leaves, scheme + CQ each) | +| Effective ~31% cluster-level coverage documented (not "100% leaves") | ✅ §5 (70→74 mapped, cluster-level) | +| Per unmapped leaf: legitimate scheme + exception/CQ (not Latin alone) | ✅ §3 (4 leaves, scheme + CQ each) | | `AIF_skosMappingType` coherent (broadMatch when narrower) | ✅ §4 (varied per leaf) | | Fail-loud when no honest scheme fits | ✅ §3 pk 840 (no native circularity CQ — documented, not fabricated) | | OWL regen reflects new structures | ⏳ Deferred — jsboige ratifies this proposition first, then CSV edits applied, then OWL regen (#133) — not in this PR | @@ -191,7 +206,8 @@ the parent, documented. - ❌ No #674/#596 merge, no régén launch. - ✅ Proposition derived code=truth from the taxonomy CSV (cluster + AIF columns) + AIF scheme semantics (Walton/`Analogy_Inference` + `DifferencesUndermineSimilarity_Conflict` are AIF native). -- ✅ Fail-loud on 840 (no fabricated Conflict node); honest scheme-divergence on 838. +- ✅ Fail-loud on 840 (no fabricated Conflict node); pk 838's honest scheme-divergence deferred to + PR-2 (cluster-boundary correction §1 — 838 is in Faulty comparison, not this cluster). Relates: dispatch `yj7u3j` (primary), #498 (chantier), #133/#130 (existing OWL), #499 (inverse: virtue = good tenor of a scheme), #192 (terminology), #458. From 2d3a0fe38992698ecb58c56b2b87bd55baf1fddd Mon Sep 17 00:00:00 2001 From: jsboige Date: Sun, 5 Jul 2026 11:05:29 +0200 Subject: [PATCH 3/4] docs(taxonomy): #498 PR-1 show both 833 Conflict nodes (anchor fidelity for PR-2) pk 833 (d2 root) carries TWO DirectRef Conflict nodes (BiasedClassification_Conflict + ExceptionSimilarityCase_Conflict). The earlier table/TL;DR showed only the first. Corrected to show both, so PR-2 (Faulty comparison, which borrows 833 as anchor and splits its two Conflict nodes across the leaves) can reference 833 faithfully. Co-Authored-By: Claude-Code --- docs/taxonomy/498-aif-fallacious-comparison-cluster.md | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/docs/taxonomy/498-aif-fallacious-comparison-cluster.md b/docs/taxonomy/498-aif-fallacious-comparison-cluster.md index 1bf915d9..2eeeab86 100644 --- a/docs/taxonomy/498-aif-fallacious-comparison-cluster.md +++ b/docs/taxonomy/498-aif-fallacious-comparison-cluster.md @@ -18,8 +18,9 @@ Linked: #133/#130 (existing OWL ontology, regenerated to reflect new structures The **Fallacious comparison** subfamily (`Subfamily=Fallacious comparison`, Misleading language family) has **2 mapped nodes** — pk 833 (d2 subfamily root "Comparaison fallacieuse", -`BiasedClassification_Conflict` directRef, `skos:broadMatch`) and pk 839 (d3 "Fausse analogie", -`Analogy_Inference` defeated by `DifferencesUndermineSimilarity_Conflict`, `skos:closeMatch`) — +`BiasedClassification_Conflict` + `ExceptionSimilarityCase_Conflict` directRefs, `skos:broadMatch`) +and pk 839 (d3 "Fausse analogie", `Analogy_Inference` defeated by +`DifferencesUndermineSimilarity_Conflict`, `skos:closeMatch`) — across 3 sub-sub clusters: **Faulty comparison** (834-838), **False analogy** (839-843), and **Association fallacy** (844-845). **This PR scopes to the False analogy sub-sub**: anchor pk 839 + **4 unmapped leaves** (pk 840, 841, 842, 843). The anchor establishes the cluster's pattern — @@ -48,7 +49,7 @@ From `Cards/Fallacies/Argumentum Fallacies - Taxonomy.csv`, family **Misleading | pk | text_fr | Subsubfamily | DirectRef | ExceptionRef | MappingType | State | |----|---------|--------------|-----------|--------------|-------------|-------| -| **833** | **Comparaison fallacieuse** (d2) | *(root)* | `BiasedClassification_Conflict` | — | `skos:broadMatch` | **✅ MAPPED (subfamily root)** | +| **833** | **Comparaison fallacieuse** (d2) | *(root)* | `BiasedClassification_Conflict`, `ExceptionSimilarityCase_Conflict` | — | `skos:broadMatch` | **✅ MAPPED (subfamily root)** | | 834 | Comparaison abusive (d3) | Faulty comparison | — | — | — | unmapped (d3 anchor → PR-2) | | 835 | Comparaison incomplète | Faulty comparison | — | — | — | unmapped (→ PR-2) | | 836 | Classification non exclusive | Faulty comparison | — | — | — | unmapped (→ PR-2) | From eb96e65245d90856d6755f1f8ddb4c67b644d199 Mon Sep 17 00:00:00 2001 From: jsboige Date: Sun, 5 Jul 2026 12:54:21 +0200 Subject: [PATCH 4/4] =?UTF-8?q?docs(taxonomy):=20#498=20PR-1=20add=20?= =?UTF-8?q?=C2=A77=20enriched-DoD=20mapping=20(I-node/RA-node/CA-node)=20f?= =?UTF-8?q?or=20chantier=20parity=20with=20PR-2?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit PR-2 #701 carries a §7 mapping each leaf to jsboige's enriched DoD (decompose into I-nodes + RA-node + CA-node, typed undermine/undercut/rebut). PR-1 was missing the parallel section — inconsistent for an ai-01 batch review that will read both PRs together. Added §7 to PR-1: maps the 4 False-analogy leaves to the decomposition. - 841 Analogie étendue → undercut (similarity not transitive; inference defective). - 842 similarité fallacieuse → undercut (resemblance too superficial). - 843 Fausse équivalence → undermine (premise "X ≡ Y" is false). - 840 Pétition analogique → FAIL-LOUD at CA-node level (RA exists, CA missing — circularity has no native AIF Conflict node). Attack-type distribution: 3 of 4 are undercuts (consistent with jsboige's "most fallacies live in the undercut"); 843 is the lone undermine; 840 is the lone FAIL-LOUD. Cross-PR framing: PR-1's 840 is FAIL-LOUD at the CA-node (RA exists), PR-2's 834/835/837 are FAIL-LOUD at the RA-node (no native comparison scheme). Both FAIL-LOUD layers documented across the two cluster shapes. Representation note (decomposition recorded, not yet serialized) shared. Gate unchanged: proposition only, 0 prod CSV write. Co-Authored-By: Claude-Code --- .../498-aif-fallacious-comparison-cluster.md | 66 +++++++++++++++++++ 1 file changed, 66 insertions(+) diff --git a/docs/taxonomy/498-aif-fallacious-comparison-cluster.md b/docs/taxonomy/498-aif-fallacious-comparison-cluster.md index 2eeeab86..7d870d44 100644 --- a/docs/taxonomy/498-aif-fallacious-comparison-cluster.md +++ b/docs/taxonomy/498-aif-fallacious-comparison-cluster.md @@ -201,6 +201,72 @@ moves to **PR-2** (Faulty comparison cluster), keeping this PR a single coherent --- +## 7. Mapping to jsboige's enriched DoD (I-node / RA-node / CA-node decomposition) + +The last jsboige comment on #498 enriches the DoD: modeling a fallacy "AIF-style" decomposes the +argument into **I-nodes** (premises + conclusion), an **RA-node** (the inference principle = the +Walton scheme, a first-class attackable object), and a **CA-node** (the Conflict Application, typed +by attack: **undermine** a premise / **undercut** the RA-node / **rebut** the conclusion). This +section maps the 4 leaves of this cluster to that decomposition. (PR-2 #701 carries the parallel +section for the Faulty comparison cluster.) + +### pk 841 — Analogie étendue → **undercut** +- **RA-node**: `Analogy_Inference` applied transitively (A~B, A~C ⟹ B~C). +- **I-nodes**: premise P1 "A~B" (similarity); premise P2 "A~C" (similarity); conclusion C "B~C". +- **CA-node**: `DifferencesUndermineSimilarity_Conflict`, applied as an **undercut** on the RA-node + — the inference (transitivity of similarity) is defective because similarity is not transitive. + The conclusion C may still hold; it is the *inference* that is defective. Classic undercut. +- **CQ that fails**: "is similarity transitive?" — no. + +### pk 842 — Argument de la similarité fallacieuse → **undercut** +- **RA-node**: `Analogy_Inference` (resemblance ⟹ relatedness/causal link). +- **I-nodes**: premise P "A resembles B"; conclusion C "A and B are linked/related". +- **CA-node**: `DifferencesUndermineSimilarity_Conflict`, applied as an **undercut** — the + resemblance is too superficial to license the inferred link. Again the inference, not the + conclusion, is the target. +- **CQ that fails**: "is the resemblance more than superficial?" — no. + +### pk 843 — Fausse équivalence → **undermine** +- **RA-node**: `Analogy_Inference` (two arguments treated as equivalent). +- **I-nodes**: premise P "argument X ≡ argument Y (in strength/structure/evidence)"; conclusion C + "what holds for X holds for Y". +- **CA-node**: `DifferencesUndermineSimilarity_Conflict`, applied as an **undermine** on P — the + premise "X ≡ Y" is false because X and Y differ in a *relevant* respect. Unlike 841/842, the + attack lands on a *premise* (I-node), not the inference itself. +- **CQ that fails**: "do X and Y differ in a relevant respect?" — yes. + +### pk 840 — Pétition de principe analogique → FAIL-LOUD at the **CA-node level** +- **RA-node**: `Analogy_Inference` (exists — the analogy is the inference). +- **CA-node**: ⚠ **absent (FAIL-LOUD)** — circularity in analogy would be an **undercut** (the + analogy presupposes the disputed assimilation it is meant to establish), but **no native AIF + Conflict node captures circularity/question-begging** for analogy. The CA-node cannot be typed + without a native conflict token; fabricating one is forbidden (FAIL LOUD, §3 pk 840). +- This is a different FAIL-LOUD shape from PR-2's pure-comparison leaves (834/835/837): there the + **RA-node** was missing (no native comparison scheme); here the **RA-node exists** but the + **CA-node** is missing. Both are honest gaps, at different layers of the decomposition. + +### Cluster attack-type distribution +| pk | RA-node | CA-node | Attack type | +|----|---------|---------|-------------| +| 840 | `Analogy_Inference` | *(absent — FAIL LOUD)* | would be undercut (circularity) — untyped | +| 841 | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | undercut | +| 842 | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | undercut | +| 843 | `Analogy_Inference` | `DifferencesUndermineSimilarity_Conflict` | undermine | + +**3 of 4 leaves are undercuts** (the inference is defective, the conclusion may stand) — consistent +with jsboige's note that "most fallacies live in the undercut." 843 is the lone undermine (the +premise is false). 840 is the lone FAIL-LOUD (CA missing, not RA). + +### Representation note +As in PR-2, the I-node/RA-node/CA-node decomposition is **recorded here but not serialized** in the +CSV — it would require new columns (`AIF_attackType`, `AIF_attackedNode`) and is the terminal target +of jsboige's comment. Recording it in the proposition lets jsboige ratify the attack-typing *before* +any CSV schema evolution. The two PRs together (PR-1 False analogy + PR-2 Faulty comparison) +demonstrate the decomposition across both cluster shapes (in-sub-sub anchor + borrowed anchor) and +both FAIL-LOUD layers (CA-missing vs RA-missing). + +--- + ## Gate boundaries (HARD — proposition only) - ❌ No prod CSV write, no DB write, no OWL regen, no `aif:` assertion emitted.