The classical problem of free will is traditionally framed as a dichotomy between determinism and indeterminism. In deterministic systems, all future states are uniquely determined by initial conditions, leaving no room for genuine agency. In indeterministic frameworks, outcomes are fundamentally random, undermining the notion of controlled choice. Both perspectives fail to provide a satisfactory account of free will. In this work, we propose a structural reformulation of the problem using the TNA (Theory of Necessary Axioms) framework. We argue that free will emerges in dynamical regimes characterized by non-uniqueness of solutions, where identical initial conditions admit multiple admissible trajectories (> 1). In such regimes, the system does not violate physical laws but operates within regions where those laws fail to enforce a unique continuation. We introduce the concept of free will as a selection process over structurally admissible trajectories and distinguish this from both deterministic evolution and ontological randomness. We further argue that non-computability plays a critical role in preventing the reduction of such selection processes to algorithmic procedures. This framework provides a unified interpretation of agency compatible with physical law, while avoiding both strict determinism and pure randomness.
Claudio Bresciano (2026) studied this question.