Several studies have suggested that both affective valence and arousal affect the perception of time. However, in previous experiments, the two affective dimensions have not been systematically controlled. In this study, standardized photographic slides rated for emotional valence and arousal were projected to two groups of subjects for 2, 4, and 6 sec. One group of subjects estimated the projection duration on an analog scale, whereas the second group of subjects reproduced the intervals by pushing a button. Heart rate and skin conductance responses were also recorded during stimulus presentation as indices of attention and arousal. Time estimation results showed neither a main effect of valence nor a main effect of arousal. A highly significant valence X arousal interaction affected duration judgments. For low-arousal stimuli, the duration of negative slides was judged relatively shorter than the duration of positive slides. For high-arousal stimuli, the duration of negative slides was judged longer than the duration of positive slides. The same interaction pattern was observed across judgment modalities. These results are interpreted in terms of a model of action tendency, in which the level of arousal controls two different motivational mechanisms, one emotional and the other attentional.In everyday life, human beings are continually engaged in emotionally driven behaviors. Such behaviors are so highly pervasive that the recent psychological literature has pointed out the centrality ofemotional factors in cognitive processes such as learning a second language (Schumann, 1990(Schumann, , 1994. Damasio (1994) has recently documented a remarkable body of neuropsychological evidence supporting the assumption that emotions are involved in most, if not all, cognitive processes.Although an increasing number of studies have investigated the role of emotions in cognitive activity, only a limited number of studies have analyzed the relationship between emotional states and estimation of time durations. Furthermore, these studies have typically yielded inconclusive results regarding the precise nature of the relationship between emotions and time perception. These inconsistent findings may originate from the use of nonstandardized emotional manipulations that make the quantification and replication of the results rather problematic.A leading theoretical approach in the current literature involves the dimensional analysis of emotions. Dimensional theories of emotion differ from basic emotions theories (Argyle, 1975;Chance, 1980;Plutchik, 1962) in that they do not classify emotions on the basis ofthe presence or absence of independent and specific emotional states (e.g., fear, anger, or joy). Instead, they assume that emotions can be represented in a multidimensional space We wish to thank G. B. Vicario, P. 1. Lang, L. Stegagno, D. Palomba, P. Bressan, R. Mucha, and L. Krueger (associate editor), and the referees for their helpful suggestions and comments. Correspondence should be addressed to Alessandro Angrilli, Antone...