While this clinical inference is straightforward, there is a great deal of evidence to suggest that it is only partially correct, if not actually wrong. The critical evidence comes from a large body of experimental data from situations where people with schizophrenia are presented with a wide array of evocative stimuli (foods, video clips, visual images, words, etc.) and are asked to rate how these stimuli make them feel. More often than not, patients' self-reports of emotional responses closely resemble those of healthy volunteers (1, 2). This is a simple and direct test of hedonic capacity, and the evidence for normative experience is robust. How then to reconcile this body of work on evoked experience with the fact that people with schizophrenia typically report less subjective pleasure on the Chapman anhedonia scales (3) and are routinely rated as being anhedonic on the basis of a clinical interview? In both of these cases, the person is asked to recall past experiences of emotion or make judgments about his or her emotional responses to different situations. In both cases, it is necessary to respond on the basis of a representation of an experience rather than the presentation of an emotionally evocative stimulus. So might the problem be in forming, maintaining, and retrieving affective representations rather than "in the moment" experiences of pleasure? And further, might people with schizophrenia therefore have difficulty initiating and sustaining goal-directed behavior because of a problem in representing the anticipated value of stimuli? If so, they might have difficulty selecting responses that are closely correlated with the value of the desired outcome (4, 5).