Marc D. Lewis
University of Toronto, Toronto, Canada
The computer metaphor for human cognition has been in decline—albeit a slow decline—since the early 1980s. Cognitive scientists are no longer content to characterize thought as a step-by-step computational sequence. In place of computation, current approaches highlight the dynamic, distributed, and non-linear aspects of a thought process that is rooted in context and embodied in a biological system (Port & van Gelder, 1995; Varela, Thompson & Rosch, 1991). Traditional AI has been all but replaced by neural network models, and the structure and function of the nervous system have become rigorous criteria for plausibility. From this perspective, cognition builds on itself, biasing its own outcomes, and changes unpredictably and unevenly from moment to moment. In short, cognition is becoming recognized as a process of self-organization in a complex dynamical system.
Self-organization means the spontaneous emergence of orderly patterns out of recurring interactions among lower-order elements. The emergence of new forms in embryogenesis, evolution, ecology and social organization exemplifies biological self-organization. Similarly, the emergence of complex, orderly patterns of cognition exemplifies psychological self-organization. At the cognitive level, coherent images, ideas and plans spontaneously assemble out of multiple associations. At the underlying neural level, the phasing or synchronization of brain