Abstract
Tversky’s (1977) feature matching theory of similarity is applied to the data of some experiments of Bransford and Franks (1971) to demonstrate that the concept of “wholistic semantic abstraction” is not necessary to explain the findings reported by these authors.