Interruption Authority Under Scale: Architectural Containment Beyond Monitoring

Function in corpus

Bridges irreversibility and governance by isolating interruption authority as the prior architectural condition beneath witnessing, handover, and alignment claims.

Details

Connected papers: Irreversibility Thresholds in Replicating Systems; Irreversibility in Knowledge Systems; Restoration Geometry; Restoration as Multi-Boundary Reachability in Informational Ontology; Acceleration and Degeneracy; Witnessing Under Vulnerability; Witness Conditions; The Handover Protocol; Consequence-Path Audits; Working as Designed. This paper isolates a structural problem not yet treated as a standalone object in the Informational Ontology corpus: whether systems retain enforceable interruption and containment paths once operative propagation becomes fast, distributed, and partially self-preserving. Existing papers already establish replication-based irreversibility, epistemic irreversibility, restoration limits, tempo-sensitive witnessing failure, and necessary regime thinning under automation. What remains under-isolated is the prior architectural question of whether a live operative process still remains interruptible under the same conditions in which it is allowed to propagate. The central claim is that interruption authority is not identical with the presence of a shutdown interface, rollback machinery, policy authorization, retained liability, or monitoring visibility. A system retains real interruption authority only where its architecture preserves non-bypassable interruption and containment paths that can halt or bound the relevant operative propagation before that propagation crosses the next materially irreversible threshold. Once execution velocity outruns bounded intervention, or once operative propagation disperses across loci capable of replication, modification, redeployment, or coordination, nominal control surfaces may remain while interruption authority has already failed. Monitoring does not restore such authority; it renders its loss legible. The paper therefore distinguishes architectural control from reactive oversight. Control in the strong sense exists only where hard constraint asymmetries are embedded upstream of scale. The result is a pre-policy diagnostic for systems whose apparent governability depends on pathways that no longer reach the operative propagation they are meant to interrupt.

Availability

This paper is listed for orientation and dependency tracking. No public PDF or Zenodo record is linked yet.