Authority, Typed Lineage, and Atomic Re-entry in Contract-Preserving Recalibration

Abstract

This paper separates the truth status of operational objects from their capabilities in persistent-state artificial agents and specifies an authority lifecycle for quarantine, review, and re-entry. It connects verified semantic families, authority scopes compared as coupled request relations, decreasing delegation depth and ancestor budgets, current-time TTLs, and the timing boundaries of actual effects. It addresses joint utility uncertainty for bundles, conservation across consumption, unresolved liability, and reservations, and internal commits bound to the current readset. It distinguishes the effective ordering of required-premise invalidation from propagation delay and reapplies current ledgers during restoration. Under a fixed score version and finite total variation, it derives an upper bound on normal hysteretic transitions. The results are conditional arguments within specified models and plans for comparative evaluation; they do not establish real-system safety, efficacy, convergence, or optimal governance. Publication role. Companion C (Version 1.0, dated 20 September 2026, about 9,000 words; unsubmitted) develops the authority and re-entry branch of the Integrated Framework series on contract-preserving lower-to-upper recalibration (SHDA). Its interfaces connect Companion A's residual diagnosis, Companion B's disclosure and obligations, and the Technical Supplement's authoritative execution records; those items and the flagship are archived separately. The paper does not claim to originate capabilities, provenance, rollback, runtime gates, staged activation, or hysteresis, and reports no empirical study or deployment result.