The diagnostic overlap among the personality disorders is quite large. A central issue is to find a conceptual model that can account for this overlap. Lilienfeld, et al. argued that the use of the term “comorbldtty” suggests a disease model. However, a high degree of co-morbidity is not plausible in the context of a disease model. Instead, Lilienfeld et al. argued in favor of a prototype model. In this paper, we show that proto-typic cases of the personality disorders are very infrequent. If a patient meets most or all of the diagnostic criteria for any single personality disorder, then that patient is likely to meet numerous criteria for other personality disorders. The prototype model is no more consistent with the diagnostic overlap data than is the disease model.