In his target article, Weiner has selected three relatively common occurrences-failure at a task, a request for help, and commission of a harmful act-and shown that in each case the reaction of an observer will be influenced by his or her attributions of cause to the event. Furthermore, as a consequence of these attributions, the observer will assign some degree of personal responsibility to the actor. Much, but not all, of the influence that belief in the actor's responsibility will later have on the observer's behavior toward the actor will be mediated by emotions. This commentary therefore accomplishes two important purposes. First, it describes how objective events are encoded into cognitive representations that guide and activate subsequent responses. Second, it places a heavy emphasis on affect in the connection between the event-as-encoded and the subsequent response. Both of these points are necessary in theories that explain motivation as a process involving both cognitive and affective components. By searching for processes that generalize across different social situations, Weiner has taken a step toward formulating a general theory of social motivation, a worthy goal indeed. In reading the analysis of task failure (prosocial behavior) and aggression presented in this commentary, the reader can easily imagine how the same intervening processes of attribution, assignment of responsibility, and affect would occur in other social contexts. Weiner suggests affiliation as an instance. Social anxiety and social influence are other topics that would be amenable to similar analysis. My comments on this interesting and well-designed commentary pertain to ways in which Weiner's insights may contribute to a larger theory of motivation. As some commentators have noted (e.g., Brody, 1980), the study of motivation is no longer guided by large-scale theories that organize all facets of the phenomenon and to which empirical discoveries can be assimilated. If theory building on such a scale is desirable-and Weiner obviously thinks that it is-then speculation on how the insights of this commentary contribute to broader approaches may be worthwhile. Weiner's observations point to three issues: (a) the role of emotion as a mediator of the relation between attributions of responsibility and subsequent action, (b) the nature of links between attributions and action when emotion is not a mediator, and (c) the possible role played by motives, defined as specific motivational dispositions, in both the affect-mediated and the unmediated processes. Consider first the role played by emotions as mediators of behavior. Clearly, Weiner believes that emotions are powerful mediators of the effects of attributions of responsibility. However, he leaves open the possibility of additional unmediated effects and cites a recent study by Zucker and Weiner (1993) in which the latter were found. This finding poses a question: Under what conditions are the effects of attributions mediated by emotion and under what conditions are they not? Zucker and Weiner (1993) found that the arousal of affect was a function of the degree of closeness to the person seeking help. In such cases, perceived controllability and blame were both correlated with anger. However, in a sample of students anger was negatively correlated with willingness to help the poor only when the help was said to require interaction with the recipient, and in a nonstudent sample anger and willingness to give such personal help were not correlated. In neither the student sample nor a sample of nonstudents was anger related to expressions of support for government welfare to the poor. Expressed support for welfare was negatively correlated in both samples with both political conservatism and attributions of personal responsibility for poverty. People appear to base their opinions about institutionalized welfare not on feelings of anger over people who deserve to be poor, but rather on ideology and perceptions of what constitutes social justice, that is, grounds that are more cognitive than affective. I will return to this distinction later. For the present, however, let us consider another of Weiner's examples, aggression, to which a similar sort of analysis may be applied. Most approaches to human aggression distinguish between affective, or angry, and forms of the behavior; in the former case the main incentive for the act is the expression of angry and hostile sentiments whereas in the latter it is the attainment of some other goal. Descriptions of instrumental aggression by North American psychologists usually include such examples as self-defense, mercenary violence, or the socially sanctioned acts of persons such as soldiers in combat or police officers. But aggression can serve the pursuit of ends other than these and still not necessarily involve the expression of anger. Such aggression should also be characterized as