Blind Validation of the Indus AUF/AUQ Micro-Grammar: A Controlled Follow-Up to the MASI-77 Structural Model

Le résumé fourni par la source

This study reports a controlled blind validation of the AUF/AUQ micro-grammar previously proposed from Mahadevan's MASI-77. The validation separates neutral structural extraction from paper-derived sign identities, freezes the blind comparison before source audit, and preserves all residual positions without retrospective correction. Across 128 evaluated positions, 112 corresponded in the frozen blind comparison, giving an agreement rate of 87.50%. Sixteen residual positions were subsequently subjected to source-level audit using the Sigilith, TraceGuard, and AXIOMOS workflow. The post-audit ledger classified seven positions as source-morphology verified, two as source-audited distinct, two as normalization discrepancies, three as unresolved neutral identities, one as morphology-review retained, and one as a true paper/source conflict. The frozen 112/128 result remains unchanged. The findings support the narrower claim that substantial structural organization within the bounded AUF/AUQ cluster is recoverable from a neutralized representation, while making no claim of linguistic decipherment.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.