When Physical Applicability Fails: A Token-Level Verdict Protocol for Physical Claims

Abstract

A physical claim can be mathematically well formed, empirically non-trivial, and still leave open whether its asserted physical content is licensed by the commitments that carry it. This paper supplies the token-level verdict layer for an applicability framework developed in two preceding steps: P1 fixes applicability-condition types, while P2 identifies transition-relative profiles and bridge burdens without issuing a verdict. P3 freezes a versioned claim, evidence boundary, preservation set, completion class, practice time, and any warranted tolerance; adjudicates token-level satisfaction; filters ordinary empirical defeat through an applicability-relevance gate; and evaluates claim-preserving closure. The claim-level statuses are PASS, OPEN, and Loss of Applicability, with distinct transition-triggered and closure-obstruction routes to the latter. Crucially, absence of a known or executed repair never licenses Loss of Applicability: token-triggered loss requires practice coverage and an independent certificate of universal non-closure. Three externally sourced case complexes—quantum contextuality, BICEP2/Planck, and muon g−2—produce localized, auditable OPEN and PASS outcomes while preserving claim scope, comparator identity, and temporal evidence boundaries. The BICEP2 case further shows why a documented but unrealized completion cannot be treated as either successful or unsuccessful in the earlier run. No external case in the present sample reaches either Loss-of-Applicability branch. The protocol is truth-neutral: it evaluates the licence of a physical claim at a specified time, not whether the claim is true.