Methods for Analyzing Cognitive Architecture

Abstract

We articulate aspects of our methods for analyzing human cognitive architecture. We focus on two approaches with a linked example. One approach is to analyze human cognitive abilities to isolate functions that have particular use in some processing domains (like language), with potential re-use in other domains. Here we focus on symmetry perception and production as may be realized in the function, expect a copy or a specialization of that function, expect a complemented copy. We illustrate the benefits of assuming such an operator is available to language processing by demonstrating that it allows reconciling empirical facts of human language processing with the natural predictions of formal language theory. The other approach is in simulation systems, parameterized to afford exploration of the effects of assuming that posited operators behave in particular ways. We link this to the expect a copy function through this being fundamental to habituation, which is requisite in simulation systems that model evolving behaviours and social constructs, like natural language.