Fixed Pattern Deviation Hypothesis of Intention Attribution

A large variety of behavioral cues which invoke intention attribution to inanimate entities are reported. Although agents, including software agents and physical robots, have already possessed these features, many people still can not feel mind to them. We hypothesize that the inability of attributing intention toward agents is caused by interpreting agent’s behavior as fixed pattern. In other words, a human attributes intention to an entity when it deviates fixed pattern of behavior in an efficient way (fixed pattern deviation hypothesis). In the present study we theoretically derived the fixed pattern deviation hypothesis and tested the hypothesis with human participants. We prepared an on-line experimental system in which a participant and an agent play a repeated pennymatching game with a bonus round. We then conducted experiments in which different opponent agents (human or robot) change their strategy during the game. The experimental results indicated that, as expected, adaptation is faster when a human is competing with robot than with another human. This implies that a human expect intentional deviation of fixed pattern of behavior against an human and expect deterministic and algorithmic behavior against robot.