Glassbox vs. Blackbox Governance — A Canonical Distinction for Legitimate AI Systems

This paper defines the canonical distinction between blackbox explainability and glassbox governance. It establishes glassbox governance as an ex-ante architectural property in which authorization state, epistemic maturity, and semantic coherence are observable prior to inference. The paper positions glassbox governance as a prerequisite for auditability, accountability, and insurability of autonomous systems. The paper is part of the canonical A7SEM / ASOSE / Pre-Inference Governance / JAQ framework family. No implementation instructions or operational guidance are provided. Commercial or operational use requires a separate written license from the author.