Premise Currency: A Governance Framework for Policy-as-Code Rules with No Recorded, Lost, or Expired Founding Basis

Every policy-as-code (PaC) rule that governs an autonomous system was authored on a founding premise: an assumption about the operating conditions, regulatory context, or risk profile that justified the rule at the time it was written. That premise is the rule’s claim to authority. This paper observes that a governing rule can continue to execute flawlessly long after its founding premise has become unrecorded, unrecoverable, or invalid — and that when it does, the organisation is enforcing a constraint it can no longer justify. We name the property whose absence produces this condition premise currency: the property that a governing rule’s founding premise is on record, still holds, and is still current. The paper makes three contributions. First, it formalises three distinct failure states of premise currency — Premise-Absent, where no founding premise was ever recorded; Premise-Lost, where a premise existed but is unrecoverable; and Expired, where a renewed justification has itself lapsed — and shows that these are governed by different disposition rules because they have different recovery conditions. Second, it specifies a disposition framework in which a rule that cannot be justified against a baseline defaults to Fail-Safe rather than continued operation, in which the choice between reconstruction, re-founding, and retirement is determined by premise provenance, and in which reconstruction deadlines scale with the autonomy of the governed system. Third, it treats re-attestation — the act of a responsible officer accepting continued operation on a stated basis — as itself a premise, subject to expiry and basis-invalidation, and shows that this terminates cleanly rather than regressing infinitely. The paper argues that existing governance standards and recent runtime-governance systems govern what a rule does — whether an action is permitted, whether behaviour has drifted, whether an output is safe — but not whether the rule’s own founding justification still holds. This is a distinct governance gap, and it is structural rather than incidental: a control layer that treats each rule as a fixed constraint has no place to record why the constraint exists, and therefore no way to detect when the reason has lapsed. Premise currency is proposed as the governance property that closes it.