Thought Formation as Metastable Inference

This preprint proposes a dynamical framework for real-time thought formation. Thought formation is modeled as a metastable inference episode arising through phase-gated coordination dynamics. The framework integrates Bayesian inference with concepts from complexity science and coordination dynamics, offering testable predictions for decision-making, cognition, and psychopathology. This work is intended as a conceptual foundation for subsequent empirical and formal studies.