A transition model for cognitions about agency

Abstract

Recent research in a range of fields has explored people's concepts about agency, and this issue is clearly important for understanding the conceptual basis of human-robot interaction. This research takes a wide range of approaches, but no systematic model of reasoning about agency has combined the concepts and processes involved agency-reasoning comprehensively enough to support research exploring issues such as conceptual change in reasoning about agents, and the interaction between concepts about agents and visual attention. Our goal in this paper is to develop a transition model of reasoning about agency that achieves three important goals. First, we aim to specify the different kinds of knowledge that is likely to be accessed when people reason about agents. Second, we specify the circumstances under which these different kinds of knowledge might be accessed and be changed. Finally, we discuss how this knowledge might affect basic psychological processes of attention and memory. Our approach will be to first describe the transition model, then to discuss how it might be applied in two specific domains: computer interfaces that allow a single operator to track multiple robots, and a teachable agent system currently in use assisting primary and middle school students in learning natural science concepts.