Pure awareness, entropy, and the foundation of perception

Abstract

Minimal Phenomenal Experiences (MPEs) represent states of consciousness reduced to their most fundamental elements, posing a unique challenge and opportunity for modeling consciousness. This paper introduces a novel computational framework based on Bayesian and active inference to model MPEs. We propose that MPEs arise when precision weighting shifts predominantly to the lower levels of a hierarchical inferential system, leading to a perceptual state characterized by increased entropy and reduced representational (conceptual) content. Crucially, awareness of this simplified state is maintained through epistemic depth: the reflexive sharing of the organism's reality model with itself. Therefore, although the contents of consciousness are exceptionally quiet, a reflexive knowing of the empty field of experience remains. We also conduct an in silico simulation to test the relationship between precision distribution and entropy, showing how this model can generate synthetic EEG data to empirically validate the theoretical framework. By advancing our understanding of pure awareness through a computational neurophenomenology approach, we provide a foundation for future research into the mechanisms underlying various altered states of consciousness, contributing to a more comprehensive understanding of the full spectrum of conscious experience.