The Governing Question
Annotatsiya
Technical disciplines are usually compared by their tools. This essay argues that they are better distinguished by the questions they were built to answer, and applies that distinction to the current convergence of cloud infrastructure, identity and access management, observability, compliance, model risk governance, and alignment research on AI governance. Each discipline supplies an accurate partial account of the problem, and each is optimized for the question its architecture was designed to answer. The difficulty is not that any of them is mistaken. It is that accurate partial descriptions are routinely offered, and accepted, as complete ones. Disciplines are distinguished by the questions they were built to answer. The essay then isolates a question none of those disciplines was built to answer: whether a specific proposed action should execute, before it executes. It traces that question to a structural change, the shift from governing models to governing actions, and argues that the two are not successive versions of one problem. While a person stood between output and effect, governing the model could function as a practical proxy for governing the outcome. When a system can both propose and execute an action, a human is no longer the necessary boundary between the two, and the proxy no longer holds. The action-level question is distinguished from classical capability authorization by decision scope and governed object rather than by contextuality. A contextual access-control decision can evaluate the operation, the resource, and the surrounding attributes at the moment of the call, and in that use the governed object remains whether the actor may invoke the operation. Whether this specific proposed act should proceed, with these parameters and on this evidence, is a determination about the action. Conditionality, whether the determination actually governs execution, is treated as a separate inquiry. The essay closes with a two-step evaluation procedure and three diagnostic questions applicable to any architecture: which question does this component actually answer, where does it sit relative to the effect, and what happens when it is not there. No implementation is proposed. Contribution. The essay introduces the term Governing Question and the associated evaluation procedure, defined and maintained at https://ferz.ai/governance/concepts/governing-question. What this record is. This is the archival edition of an essay whose canonical web version is maintained at https://ferz.ai/articles/the-governing-question. It is an essay rather than a formal research paper: it offers analytical framing rather than empirical results, formal validation, or a comprehensive review of the literature. The deposit provides a durable, independently timestamped record of publication and attribution. Citation. Cite the concept DOI for the work and the version DOI for this specific edition. Related work. Meyman, E. (2026). Authority versus Authorization: A Definitional Framework for AI Governance. FERZ, Inc. https://doi.org/10.5281/zenodo.21341907 Meyman, E. (2026). A Taxonomy of AI Governance Approaches: Distinguishing Visibility, Alignment, and Authorization. Version 1.7. FERZ, Inc. https://doi.org/10.5281/zenodo.18275969 © 2026 FERZ, Inc. Licensed under CC BY 4.0.
Hali tarjima qilinmagan