Consciousness via Phase Transition of Learning: Qualia as a Property of Concepts

Abstract

Building on recent advances in understanding of consciousness as constituted by a set of concepts, this work proposes a formal description of this set as a space of consciousness. Geometrically, this is an N-dimensional vector space, where each of the N concepts corresponds to a basis vector. Algebraically, it is a space of N continuous functions, each corresponding to a concept. The concepts and the space of consciousness they constitute emerge via a phase transition of learning (or training) that is analogous to the discrete-to-continuum transition (or clustering transition) in theoretical physics. A concept is a dual elementary entity that combines meaning and quale as its intrinsic properties. Qualia are therefore fundamental intrinsic properties of the space of consciousness and serve as mediators between the concepts and their physical substrates (neurons or neural networks). Since concepts emerge via learning, both their meanings and qualia can be modied through retraining. The role of correlators of concepts, or associations, is discussed. They are the reason for the non-Abelian character of information and therefore for the emergence and irreversibility of phenomenological time.