Epistemic Agency and the Ontological Continuity Condition: A Constraint on When Knowledge Must Be Owned

It is uncontroversial that many systems possess knowledge without being conscious: biological subsystems retain information, procedural skills guide action, and artificial systems learn and deploy complex representations. What remains insufficiently explained is why epistemic agency must arise at all, rather than how it is merely attributed once present. This article argues that epistemic agency—the capacity to hold, revise, and act upon knowledge as the same agent across time—presupposes consciousness. Building on the ontological continuity condition, it draws a principled distinction between knowledge that merely informs behaviour and knowledge that must be owned, revised, and answered for under long-horizon uncertainty. Where a system must arbitrate between incompatible action trajectories in order to preserve itself as a persisting agent, consciousness becomes functionally necessary on the present account. Epistemic agency is thus characterised not as a metaphysical power, but as a structural role emerging from unified arbitration under identity risk. The article shows why many adaptive systems, including contemporary artificial intelligences, can instantiate knowledge without qualifying as epistemic agents, and clarifies the structural conditions under which artificial epistemic agency might arise in principle. Consciousness is thereby reframed as a threshold for epistemic responsibility rather than for intelligence or knowledge as such.