The Closed-World Bargain: Why Bounded Authorization Does Not Establish Authorization Infrastructure

FERZ Doctrine Article · Version 1.1 · August 2026 The Closed-World Bargain: Why Bounded Pre-Execution Authorization Does Not Establish Authorization Infrastructure A bounded authorization gate can provide genuine pre-execution control within a defined workflow. It does so, however, by holding six conditions sufficiently fixed: action vocabulary, meaning, policy, authority, state, and execution topology. This article names that trade the closed-world bargain and examines what changes when an authorization system must operate beyond those fixed assumptions. The article distinguishes integrity verification from an independently reconstructable authorization verdict, defines reconstruction under both replay modes of the Five Tests Standard, State-Replay and Protocol-Replay, and treats provenance, the established origin of the inputs grounding a verdict, as the depth of authorization rather than a layer above it. Against that doctrinal background, it introduces the Composition Test as a falsifiable method for evaluating whether an architecture carries a common authorization and evidence discipline across changing assurance objects: two assurance functions must govern different assurance objects; the validated content of the first artifact must materially condition whether the second function may return ALLOW; the verdict must be independently reconstructable under a declared replay mode; and fail-closed behavior must be demonstrated directly across a declared execution-path scope. The analysis does not deny that a bounded gate can authorize correctly. It addresses the narrower architectural question of what separates valid pre-execution authorization within one closed world from reusable authorization infrastructure. The Composition Test is a scoped demonstration rather than a universal security proof, and it applies to every architecture in the category, including that of FERZ, Inc., which has committed its own demonstration program to the test and has not yet passed it. Version history. v1.1 (August 2026): defines reconstruction under both 5TS replay modes, State-Replay and Protocol-Replay; requires material consumption of the first artifact in the Composition Test; requires fail-closed demonstration across a declared execution-path scope; and states the test's scope limitation. v1.0 (July 2026): initial publication; version DOI 10.5281/zenodo.21643659. Citation and license. Copyright © 2026 FERZ, Inc. This work is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). Attribution should cite FERZ, Inc. and concept DOI 10.5281/zenodo.21643658; each deposited version carries its own version DOI. This Zenodo record is the timestamped record of publication. Full corpus at ferz.ai and the FERZ community on Zenodo.