Time and Becoming in Artificial and Natural Cognitive Systems

Abstract

In my paper I review the main strategies adopted by two leading schools in cognitive science, symbolic artificial intelligence (AI) and connectionism, in modeling time-dependent phenomena such as learning. In particular, I briefly mention shortcomings of non-monotonic logic approach that dominates symbolic AI. I also discuss the problems that beset the recurrent networks approach advocated by connectionists (e.g., S. Grossberg) and philosophers (e.g., P. Churchland), who focus their attention on oscillatory behavior of such networks. I point out that neither approach adequately captures essential features of the dynamics of human temporal awareness. Finally, I develop certain ideas borrowed from neurophysiologists (e.g., W. Freeman) and system theorists and describe the brain-mind, the entity that is capable of producing temporal awareness, as a self-organizing system that exhibits chaotic dynamics and generates a dynamic structure that resembles the chaotic hierarchy o Rossler.