Horizontal Validation — 31 claims across the full taxonomy

Before running the quantitative experiment, the taxonomy needed to demonstrate that it could handle the full range of claim types — not just the three selected for replication. The horizontal validation corpus does exactly this.

Thirty-one claims were selected and validated, spanning eleven D1 categories: from historical (D1.1) and statistical-probabilistic (D1.6) to normative (D1.7), predictive (D1.8), analytical (D1.9), and metaphysical-ontological (D1.10, D1.14). The corpus includes claims that are unambiguously false (“Vaccines cause autism”), claims that are framework-dependent (“Vegetarianism is ethically superior”), claims that are analytically true (“A triangle has three sides”), and claims that are in principle undecidable (“Consciousness is irreducible to matter”).

The corpus is not a sample — it is a proof of coverage. Its function is to demonstrate that the taxonomy handles the full ontological range without category collapse or forced misclassification.

Each validation follows the locked format: PRE-STEP 0 → PROSPETTO (six rows: Ontologia / Logica / Decidibilità / Esito / Pre-Step 0 / Sintesi rapida) → [EMPATIA] → [LOGICA] → [INFORMAZIONE] → [EPISTEMIC RESPONSIBILITY CHECK] → [SINTESI ESTESA] → FONTI.

Document: archivio_31_validazioni_v4.docx — February 2026. 31 claims, 11 D1 categories.