The Necessity of Bounded Agency: Why Absolute Free Will is Impossible, But Relative Freedom is Inevitable

The free will debate suffers from a persistent categorical error: conflating absolute, uncaused freedom with bounded, computationally determined agency. This paper reframes the discourse by mapping agency onto a topological continuum parameterized by x—the degree of independence from prior causes. By the Law of Excluded Middle, x \0, 1\ or x (0, 1). We establish three axioms: (1) all knowledge is acquired, (2) all rationality rests on non-rational foundations, (3) the experiential log is the boundary of the will. From these, we derive that Path A (x = 0, eliminativism) renders humans indistinguishable from AI, collapsing moral responsibility; Path C (x = 1, absolute freedom) violates thermodynamics and information conservation. Path B (0 < x < 1) emerges as a structural necessity: bounded agency via computational self-blindness. We demonstrate that major objections—non-reductive physicalism, property dualism, existentialism—logically collapse into Paths A or C. Relative free will is not a compromise; it is the inescapable intersection of physics, computation, and ethics.