Abstract
Some common clinical statements about patients with personality disorder are examined for their implications in the choice of category concepts for classifying personality. Clinicians seem to assume that personality categories are imprecise and that membership is probabilistic. These assumptions run counter to those of current classifications. The fact that most cases of personality disorder show features of more than condition creates difficulties for categorical systems. Dimensional systems are frequently proposed as a solution to this problem. Such an approach is rejected at the present time since there is no consensus regarding basic dimensions. Polythetic and prototype categories are examined as alternatives. It is suggested that prototype categorization provides an approach that is more in accord with clinical judgement. This approach, based upon Wittgenstein's concept of family resemblances, is better able to accommodate the imprecise or fuzzy nature of personality categories, although it does require some modification to the way categories are defined, and to the way diagnostic reliability is assessed.