“On the Problem of Action” contrasts two conceptions of the task of action theory: the dominant conception, which I call the decompositional approach, and an alternative, non-decompositional approach that is implicit in the tradition of action theory descending from Aristotle. Decompositionalists seek to characterize intentional action as a composite of something inward and something outward, bound together by some generic kind of causal relation. I show that this approach is committed to characterizing action in terms that treat the agent’s own standpoint on her action as a separable factor, not integral to the worldly happening that constitutes her action proper, and I argue that this commitment leads decompositionalists to focus their theorizing not on actions-in-progress, but on cases of completed action. I then show how the neo-Aristotelian approach to understanding action contrasts with the decompositional approach in each of these respects: it seeks, not to explain what intentional action is by decomposing action into several not-intrinsically-agential factors, but rather to characterize the understanding of what it is to act implicit in the agent’s knowledge of her own action-in-progress. The main aim of this paper is simply to show that there is an alternative to the decompositional project, one that resists the demand for an explanation of action in non-agential terms, while not simply treating the notion of intentional agency as an unexplained primitive. To see the possibility of this sort of understanding of action is, I argue, to see how action could turn out to be, not a worldly event with certain psychological causes, but a distinctive form of material process, one that is not simply caused by an exercise of reason but is itself a (productive) exercise of reason, it is itself the exercise of a power of practical cognition.