An Evolutionary Robotics Simulation of Human Minimal Social Interaction

摘要

This abstract introduces an evolutionary robotics simulation of experimental results on minimal human social interaction. Such simulation experiments play an important role in a minimal enactive approach that aims to combine experiential, empirical and theoretical results to generate an explanation for cognitive behaviour without reducing behavioural phenomena (macro level) to constituent parts of the agent–environment system (micro level). The results of this study are interesting in three different ways: Firstly, such an enactive account of social interaction that does not focus on individual capacities, but on the dynamical interaction process, is able to account for emergent phenomena that are difficult to understand otherwise. Secondly, the evolutionary robotics simulation uncovers a number of surprising aspects of the task, which allow a different view on the data. Thirdly, the presented work is an example of how the gap between minimal artificial life simulations and the empirical study of human level cognition, involving human conscious experience, can be bridged – not by scaling up the complexity of robotics models, but by scaling down the complexity of the aspects of human behaviour under investigation.