Question Lineages A Cross-Disciplinary Theory of Identity, Persistence, and Realization

Questions are routinely treated as contents, utterances, cognitive acts, logical objects, measurement interventions, or organizational records. These approaches explain what questions mean, how they are posed and resolved, or how agents and institutions take stances toward them. They do not provide a general identity law for deciding when differently worded, differently borne, differently recorded, or differently resolved question-states belong to one continuing question. This paper proposes the Lineage Criterion, a candidate law of question identity. A question-state is represented as q_t = (A_t, S_t, D_t, R_t), where A is an anchor structure, S an interrogative-slot structure, D a set of outstanding resolution tokens, and R an ordered provenance-bearing record. Continuation requires accountable inheritance of resolution, total disposition of predecessor deficit, anchor continuity, slot lineage, and a non-vacuous continuity core. Physical admissibility requires a bounded lineage variable whose values are distinguishable, permutation-closed, source-preservingly clonable, re-encodable, and forgery-detectable on a bounded physical substrate. Distributed records are classified by contextual gluing into empty, singleton, or multiple global-section sets; multiplicity supports a supervalued invariant core. The criterion was derived through an adversarial, published-source-constrained examination of scientonomy, philosophy of inquiry, erotetic logic, inquisitive semantics, physics of information, sheaf theory, quantum cognition, and issue-based systems. A proprietary bounded reference implementation has completed a frozen constructive-witness campaign: two independent clean-environment executions produced identical canonical evidence outputs across deterministic transition verification, durable persistence and record mortality, the C6 constructor-theoretic operations on transient and durable substrates, all three C7 classifications, and cross-representation reconstruction, with all negative controls failing as specified. This establishes an internally reproduced constructive witness for a finite domain, not confirmation of the ontological hypothesis. A preregistered cognitive compatibility classifier and a cross-schema/method-variance organizational study remain the decisive empirical obligations.