The Empty Occupant: Answerability, Testimony, and the Limits of Machine-Mediated Authority

The Empty Occupant: Answerability, Testimony, and the Limits of Machine-Mediated Authority is a philosophical monograph examining a fundamental problem of institutional artificial intelligence: an institution can identify who is responsible for an automated decision while lacking the capacity to meaningfully justify, reconsider, correct, or repair it. As automation increases the scale of authoritative decisions, the capacity to answer for their consequences may fail to keep pace. Situated at the intersection of philosophy of language, social epistemology, social ontology, normative theory, AI governance, and administrative law, the book investigates the conditions under which machine-mediated communication constitutes genuine institutional participation, when automated outputs may legitimately acquire authority, and when failures of answerability undermine that authority. The monograph's original contribution is an integrated, substrate-neutral theory of Discursive Ontology connecting three questions ordinarily examined separately: who owns an institutional commitment, what warrants its adoption, and whether the resulting obligations can be sustainably discharged at scale. The book develops this framework through five principal theoretical contributions: Discursive Ontology and Answerability: Establishes four jointly necessary relations—attribution continuity, challenge-responsive reconsideration, repair competence, and standing-consequence—and distinguishes Actual, Residual, Prospective, and Empty Occupancy. Its governing axiom is that missing bearership is not defective ownership. Testimony Without a Testifier: Introduces a two-axis distinction between epistemic value and commitment ownership, demonstrating why reliable, fluent, or knowledge-supporting machine output does not independently establish an answerable speaker. Normative Sponsorship: Develops five conditions for warranted institutional adoption of machine-generated commitments, including a contestation-grain opacity threshold and a distinction between technical opacity and organizational capture. Deontic Debt and Institutional Insolvency: Introduces a formal account of accumulated, inadequately discharged obligations of answerability and identifies when structural incapacity to service those obligations defeats an institution's presumptive authority to continue issuing adverse decisions at scale. Discursive Stewardship: Develops a governance framework linking permissible automated use, meaningful reconsideration, effective repair, institutional capacity, and the restoration of sustainable authority. The theoretical framework is examined through comparative documentary analyses of Australia's Robodebt scheme, Moffatt v. Air Canada, the United States' TennCare system, the Post Office Horizon scandal, and the Dutch childcare-benefits affair, alongside a bounded generative-AI example. A worked obligation ledger, formal serviceability model, conceptual glossary, and proposed audit specification extend the analysis towards practical institutional assessment. The documentary cases illustrate the framework rather than constitute empirical validation of its diagnostic criteria. The book's central thesis is that presumptive institutional authority requires effective Answerability at the level at which that authority is exercised. Neither conversational sophistication, formal attribution, nominal human oversight, nor technical accuracy can substitute for a continuing bearer whose commitments can be challenged and whose obligations can be meaningfully discharged. The monograph contributes a unified philosophical account of commitment ownership, warranted institutional adoption, and sustainable machine-mediated authority, offering a conceptual foundation for research into artificial agency, institutional accountability, automated administration, and the normative limits of scalable AI governance. Ultimately, the volume offers researchers a vocabulary for examining whether institutional reasons can actually govern what automated systems are permitted to do. -- Earlier Preprints and Development of the Framework Earlier formulations of the arguments developed here have circulated as preprints on Zenodo and PhilArchive. This monograph integrates, revises, and extends those arguments within a unified framework. The earlier papers document the development of the individual arguments: Cott, N. (2026). Deontic Debt and Institutional Insolvency: Answerability and the Limits of Scalable Authority. Zenodo. https://doi.org/10.5281/zenodo.23175523 Cott, N. (2026). Normative Sponsorship Under Opacity: When Institutions May Own Machine-Mediated Commitments. Zenodo. https://doi.org/10.5281/zenodo.23153361 Cott, N. (2026). Testimony Without a Testifier: Epistemic Value and Commitment Ownership in Institutional AI. Zenodo. https://doi.org/10.5281/zenodo.23153358 Cott, N. (2026). Answerability and Institutional Commitment: Why Missing Bearership Is Not Defective Ownership. Zenodo. https://doi.org/10.5281/zenodo.23153297 Cott, N. (2026). Empty Occupancy: Answerability and the Ownership of Machine-Mediated Institutional Commitments. Zenodo. https://doi.org/10.5281/zenodo.23151642