ABSTRACT
The position taken in the review of experiments on concept formation is a general one, intended more to comment on the field as a whole than on any one position in particular. When individual items are stored in memory, the concept formation process can abstract information from them, rather than only from items as they are presented. The computational theory of concept formation begins by assuming that episodes are stored in memory, having been encoded by whatever schemata exist for processing the episodes. The purpose of this chapter was to present a summary of the major ideas about concept formation. The second purpose of the chapter was to present a computational theory of concept formation that, the people believe, can serve to integrate the facts from the different research areas. Some general comments were made to illustrate how the computational model would account for some of the results discussed in the part of the chapter.
