Behavior and Cognition as a Complex Adaptive System: Insights from Robotic Experiments

Publisher Summary This chapter focuses on the evidence collected in robotic research with particular reference to results obtained in experiments in which the robots develop their skill autonomously while they interact with the external environment through an adaptive process. In particular, it demonstrates how the behavioral and cognitive skills developed by the robots can be properly characterized as complex adaptive systems which: arise from the interactions between the brain of the robots, their body, and the environment and eventually between the dynamical process occurring within the robot and within the environment; display a multi-level and a multi-scale organization in which the interaction between behavioral and cognitive properties at a certain level of organization lead to higher-level properties and in which higher-level properties later affect the lower-level properties. The complex system nature of behavior and cognition has important consequences both from an engineering and a modeling point of view. From the point of view of developing effective robotic artifacts, it implies the need to rely on “design for emergence” techniques, i.e. techniques allowing the development of robots which are able to exploit useful emergent properties. From the point of view of modeling biological systems, it implies the need to conceptualize behavior and cognition as dynamical processes which unfold in time while the organism interacts with the environment.