The Vagueness of Robot Emotions

As the field of robotics matures, robots will need some method of displaying and modeling emotions. One way of doing this is to use a human-like face on which the robot can make facial expressions that correspond to its internal emotional state. Yet the connection between a robot’s emotional state and its physical facial expression is not an obvious one; in particular, there is no principled method of mapping changes in emotional states to changes in facial expressions. We give a philosophical analysis of the problem and show that it is rooted in the vagueness of robot emotions. We then outline several methods that have been used in the philosophical literature to model vagueness and propose an experiment that uses our humanoid robot head to determine which philosophical theory is best suited to the task.