If people's behaviors reflect their personalities, and if their personalities remain relatively unchanged across situations, then it should be possible to predict what a person is like in one situation from knowing what she or he is like in another. But research shows this is often very difficult to do, despite the concerted effort by many social and personality psychologists. Does personality exist? Is it not stable? Research in my lab to date suggests that hidden in the seemingly random variation of individuals_ behavior across situations is a pattern that is stable and distinctive for each individual. The behavior itself varies, but there is stability in how each individual_s behavior varies from one situation to another. These stable and distinctive patterns, which we call the behavioral signatures of individuals, suggest the existence of a higher order consistency and that an intuitive belief in personality may in fact be based on consistency of such a kind. To account for such higher-order consistency (i.e., consistency in the pattern of variatons), we have been conceptualizing personality as a distinctive and stable network of automatic associations of thoughts and feelings. The thoughts and feelings that are activated at any given moment may change. But the associative network that guides and constrains their activation itself may be stable and distinctive for each individual. Computer simulations have confirmed this prediction. Our research now focuses on developing empirical methodologies for the assessment of automatic (i.e., not consciously controlled) associations among specific cognitions and affects, as well as a method for analyzing situations in terms of a set of psychological features that activate these cognitions and affects.