An "Artificial Life" approach to the origin of mental complexity in autonomous agents

1. Aim The central goal of the research is to understand how mental complexity may originate in autonomous agents. The emphasis will be on pre-rational, subsymbolic intelligence. We will test a number of hypotheses: (1) an architecture based on behavior systems as the fundamental unit, (2) a dynamics approach for the definition of behavior systems, (3) on-line evolutionary techniques for generating new behavioral complexity. The first two hypotheses have already been worked out in detail by the group. The last hypothesis is the ma]or focus of the project;