Categorization, Recognition, and Unsupervised Learning

In this chapter, we review several strands of research that share one common theme: the close conceptual and empirical links between categorization and recognition memory. We demonstrate that these links are strong, and often surprising. We explore the relation between recognition and unsupervised learning, and we investigate how perceptual processes can provide a unifying link between categorization and recognition. In Gordon Bower’s laboratory in the late 1980s and early 1990s, unsupervised learning was in the air. Several of his students were working on categorization as well as memory, and it appeared that dominant ideas in categorization research were being overturned. In particular, the traditional supervised-learning paradigm