SUIS-FREEWILL

Natural systems evolve according to coherence constraints defined by invariants, tensions, and multiscale couplings, ensuring that their transitions remain structurally determined by the geometry of the coherence space that governs their dynamics . Human cognition appears to constitute a singular exception: while many aspects of cognition follow predictive coherence, volun- tary disruption appears unique to human-like metacognitive flexibility (contra strict determinism in simpler systems). This work introduces a structural and mathematically grounded account of such transitions within the SIUC (Systemic Integration Unified Coherence) architecture. We define free choice as a non-determined transition in a coherence space C, formalized through a coherence-rupture operator Afree and a residual term R(t0, t1) that quantifies departures from the natural flow Pt. Computational simulations with a GRU predictor show systematic, detectable residuals for Afree-induced transitions, distinguishable from noise via Kolmogorov-Smirnov tests (p < 10−13 in typical cases, up to p < 10−45 in high-probability scenarios). The framework yields empirically testable predictions across behavioral, computational, and neural domains, including attractor violations, predictive-model failures, and transient disruptions in multiscale neural coherence (contra causal interpretations of readiness potentials [ 17 ]). By grounding free choice in the geometry of coherence rather than in metaphysical assumptions, SIUC provides a unified, falsifiable, and structurally explicit approach to a long-standing question in cognitive science.