This paper tries to make three points. First, current constructs in personality and psychopathology are based on the restrictive evidence contained in self-reports. As a result, heterogeneous categories of individuals are assigned to the same category. Second, it is suggested that when different sources of evidence are included, theoretically distinct groups will be detected within the prior heterogeneous category. Third, the authors argue that physiological information has the potential to parse individuals with similar phenotypes on self-report data into distinct groups that reveal the temperamental origins of their phenotype.