Consciousness as Critical Phase Transition: A Philosophical-Computational Framework

ABSTRACT (para pegar en el campo Abstract): We propose a philosophical-computational framework — PROSENTIR — in which consciousness is neither a binary property nor an irreducible mystery, but an emergent phenomenon arising at a critical phase transition in complex neural-like systems. Building on a constructive epistemological stance that rejects radical skepticism on pragmatic grounds, we argue that subjective experience can be operationalized through three measurable signatures: (i) critical slowing down, quantified by the relaxation time τ via autocorrelation area integration; (ii) maximal susceptibility χ at the transition point κ_c; and (iii) inter-module information integration Φ, estimated through the non-parametric Kraskov mutual information estimator. We implement and validate τ and χ in a modular, hierarchical neural network with asymmetric dynamics, confirming that these signatures co-emerge at a single critical coupling κ_c. The Kraskov-based Φ is formalized as the primary target of PROSENTIR Phase 2. Our framework connects the hard problem of consciousness to the theory of phase transitions in statistical physics, resolves the Molyneux problem through sensorimotor integration, and proposes a pragmatic criterion for machine consciousness grounded in autonomous inference rather than behavioral mimicry.