It is widely acknowledged that an agent is doing A intentionally only if she knows she is doing A. It has proved difficult, however, to reconcile two natural thoughts about this knowledge. On the one hand, the agent seems to know what she is doing immediately, simply by doing it. Her knowledge seems to rely upon no evidence, and indeed to rest upon no specifiable epistemic basis at all. On the other hand, the agent can be wrong about what she is doing; she is fallible. The difficulty is to see how an agent can be wrong about her action if her knowledge of it is immediate. My dissertation provides an account of the agent’s knowledge of her own actions that reconciles these natural, but apparently conflicting thoughts. In the face of this difficulty, many philosophers distinguish two objects of knowledge in action: the object of immediate knowledge, which is supposed to be something interior, and what the agent actually does, which is known only mediately. I argue that this two-factor framework is unacceptable, since it cannot account for the insight which motivated the study of intentional action via the agent’s knowledge of these actions: that it is in virtue of this specific way of knowing that the agent is the agent of her intentional actions. Instead, I defend a view on which acting intentionally itself, with no need for further epistemic work, is a way of knowing what actually happens. This account of knowledge in action also allows me to clarify how this knowledge is necessarily related to our capacity for agency. I argue that the rational capacities that are drawn on in figuring out what to do here and now are the very source of both the action’s taking place, and the agent’s knowledge of her actions without evidence. Since the agent’s knowledge is the result of the very same reasoning that brings about the action, it is practical, and the agent’s having it is the mark of her practical rationality at work and her being the knowingly efficacious author of the action. / Philosophy