Perceptual expertise and the autonomy of perception.

Abstract

According to the received theory of perception, very little information enters the perceptual system as input. Beliefs or theories have to supplement the information to form the perceptual output. This view is the basis for the inferential model of perception and the thesis that perception is theory-laden. While such a theory can account for a range of phenomena, it cannot account for perceptual expertise. I offer an alternative theory, which I call the "exploration model of perception." According to such a view, the inferential model of perception is an artifact of experimental design in psychology and a certain conceptualization of the input. First, the experimental design inspired by the received view impoverishes the amount of perceptual information available to the subject--forcing the perceiver to rely on beliefs or theories. For instance, in the standard experiments, the temporal duration of the stimulus is extremely short, compelling the perceiver to rely on beliefs. Second, the interpretation of the experiments misleads us into thinking that all perception is belief-driven. For example, much of contemporary psychology relies on introspective reports as opposed to skilled motor performances. Recent work shows that introspection has access to less perceptual information than our skilled motor performances, so our reports make it appear that the perceiver receives less perceptual information than she actually does, motivating the view that perception is theory-laden or inferential. By describing those who are in informationally impoverished settings, the standard psychology of perception has produced theories that are not generally true of perception and, in particular, not true of expert perception. It is a psychology of novice perception. I argue that such a psychology is a product of an epistemological foundationalist tradition. The alternative model of perception treats the perceiver as an active agent in the world. The receptors of a moving agent receive information of a different kind than the receptors of a stationary agent. Instead of detecting distal properties (e.g., edge or table), the perceiving agent is monitoring her relationship to the environment (e.g., obstacle to motion). I show how the current empirical literature supports such a view. Perceptual expertise is explained not by the possession of a theory, but by the ability of the agent to exploit the available information. Perceptual expertise is possible because only a small percentage of the available information is being utilized by most perceivers. Finally, I raise some questions concerning how such a theory of perception impacts our epistemology and philosophy of science.