The recoverable-continuity framework, expressed by C 2 = C, proposes that identity, reference,action, consequence, law, and domain expression remain coherently attributable through transformationonly while the connecting relation remains recoverable. Three companion experimental programs nowgive this framework distinct empirical contact: a physical program testing matched-endpoint historyresiduals across quantum transformation, irreversibility, and gravitation; a biological program testinglineage, organization, regeneration, reprogramming, and evolution; and an awareness-locked differentialgravimetry program testing a bounded consciousness–gravity bridge hypothesis. The existence of severaltests, however, creates a new problem. A positive or null result in one pillar cannot be interpretedresponsibly without rules specifying which claim was tested, what alternative explanations remain, howevidence propagates across levels, and what combination of outcomes would support, constrain, revise,or falsify the framework. This paper supplies that architecture. It separates the foundational structuralcondition from domain bridge hypotheses, operational models, and individual protocols; defines a commonmatched-closure formalism; introduces a three-component evidence vector rather than an artificial universalscore; establishes a closure-escalation procedure for distinguishing omitted conventional variables froma recoverable-continuity residual; and provides an eight-outcome cross-domain decision matrix. Fourpropositions govern interpretation: finite confirmations cannot prove universality; empirical bridges areseparable; isolated anomalies do not validate upstream ontology; and independent replicated residualsacross physics and biology provide stronger evidence for a domain-general empirical role than eitherdomain alone. The consciousness–gravity hypothesis is treated as a specific downstream bridge whosefailure need not invalidate the structural framework and whose success would require gravity-class behaviorand independent replication before ontological interpretation. The paper concludes with preregistration,blind-analysis, replication, stopping, disclosure, and public-claim rules. The result is a unified falsificationgrammar for the C2 = C research program: every claim is assigned a level, every experiment a target,every result a bounded interpretation, and every failure a specified consequence.