In this thesis, I explore the concepts of willpower and self-control from the perspective of their psychological underpinnings, their rational standing, and their value. Negatively, I criticize the views on which our capacity to overcome conflicting motivations exclusively depends on cognitive mental processes and I reject the idea that we should distinguish between "synchronic" and "diachronic" self-control. Positively, I argue that self-control is best construed as a skill that involves the flexible interplay of cognitive, affective and subpersonal processes. Self-control and willpower, I also insist, are distinctively normative concepts in the following sense: to exhibit self-control is to act under the guidance of one's perception of reasons.

Drawing from this view, I turn to the opposite notion, weakness of will. I defend a unified account of weakness of will that reconciles the traditional Aristotelian-Davidsonian view, on which weak-willed agency is agency that runs against one's better judgment, and the revisionary Holtonian view, on which weakness of will is a kind of irresoluteness. I also provide sustained criticism of the now-popular view that weakness of will can be rational.

Finally, I tackle the question of what makes strength of will a valuable skill. I argue that in addition to its much-discussed pragmatic and ethical benefits, strength of will yields epistemic benefits when stably manifested over time, to the extent that it enables us to acquaint ourselves, through temporally extended agency, with the (dis)value of ends and our personal reasons to pursue them or not. The resulting picture is one where practical reason has a distinctively experimental dimension.