Personality disorders, severe and otherwise, constitute what one might call afuzzy set, after the theory developed by Lotfi Zadeh (1987), and expanded by Bart Kosko (1993). An analogous concept is that of “warmth” as applied to the ambient temperature: there are numbers below which almost no one would consider the temperature “warm,” and other readings, say – above 122 °F/50 °C – that would almost universally be experienced astoo warm. As the temperature approached 50 °C, gradually increasing percentages of people would conclude it was “too hot.” This gradual change, which would speed up as one got very near to 50 °C, is the fuzzy set. As Kosko points out, the termlife is itself fuzzy (p. 242). When it begins (when the sperm meets the egg? at the blastula stage? later?) is a matter of shading; a matter of degree and debate. Questions concerning fuzzy sets are decided often by expert opinion, not precise scientific measurement. In the domain of medical diagnosis, models that use cluster analysis may begin with performing a clustering algorithm on a set of patients – “by examining (a) the similarity of the presence and (b) the severity of symptom patterns exhibited by each” (Klir and Folger, 1988, p. 252). The authors mention that the similarity measure is usually computed between the symptoms of the patient in question and the symptoms of a patient possessing (by experts' agreement) the prototypical symptom pattern for each possible disease or condition.