The concept of prototype (clear category member) and the fuzzy-set form of cognitive categorization that it entails (Rosch, 1973) has been shown to be useful as an organizational principle for guiding research and assessment in personality psychology. With the aid of a prototype framework for conducting research, personality assessment instruments have been refined (Broughton, 1984), predictable behavioural acts have been brought into focus (Buss & Craik, 1981), abnormal diagnosis has improved (Horowitz, French, Lapid, & Weckler, 1982; Wiggins 1980), and person perception studied (Cantor & Mischel, 1979). These (and other) authors argue that prototypes "work" to the extent they bring us closer to the way in which people actually think about person variables and behaviour. In this article some of the past payoffs of a prototype approach in the field of personality psychology are reviewed, and a multidimensional scaling model of personality assessment, involving a similarity-to-the-prototype rating task, is introduced and its merits discussed.