Category Learning as Schema Induction

Categories are essential for practically all aspects of human cognition, and an enormous amount of research has been devoted to understanding how they are learned and represented in memory (see, e.g., Murphy, 2002; Smith & Medin, 1981). In this chapter, I describe a program of research on category learning that Gordon Bower and I began a number of years ago, when I was a graduate student in his laboratory at Stanford University, and which I have continued to extend and develop over the succeeding years. This research began by investigating how people use category knowledge (schemas) to guide attention and organize memory, and later focused on the basic mechanisms by which such categories are discovered and learned. Although category learning has been a traditional focus of research within cognitive psychology, we have taken a rather non-traditional approach in our own investigations of this area. My primary goal in this chapter (besides paying tribute to Gordon) will be to convince readers of the value of this non-traditional approach. It is important to note that most category learning research has been carried out within a standard conceptual framework or paradigm that describes what categories are, what they are for, and the kinds of situations in which they are normally acquired. Within this paradigm, category learning typically is regarded as a form of