Conventional models of persistent pain have tended to be dichotomous in nature, with pain viewed as either physically or psychologically based. Inadequacies inherent in both of these views have resulted in alternative conceptualizations that focus on the integration of biomedical with cognitive, affective, and behavioral factors. During the past decade there has been a proliferation of research designed to examine the relative contributions of individuals' attitudes, beliefs, appraisals, self-perceptions, and coping strategies to the perception, experience, and response to noxious sensations as well as treatment, and how these are modified as a result of treatment. In this paper a cognitive-behavioral conceptualization of persistent pain is described and contrasted with sensory, psychogenic, motivational, and operant conditioning models. A number of cognitive assessment procedures and recent research on the role of cognitive schemata, cognitive processes, and ongoing cognition in chronic pain are briefly summarized. The central importance of negative cognition - "catastrophizing" - is emphasized. Once pandora's cognitive box has been opened, a range of important issues must be addressed or one may be consumed by unbridled enthusiasm for the development of instruments and correlational research. Several caveats regarding current research on cognitive mediators are raised, namely, confounds among the cognitive measures that have proliferated and between cognitive measures and measures of mood states, generalizability of results based on pain clinic samples, and adherence to "patient uniformity myths."