In this paper, I consider intellectualist and anti-intellectualist approaches to knowledge-how and propose a third solution: a virtue-based account of knowledge-how. I sketch the advantages of a virtue-based account of knowledge-how and consider whether we should prefer a reliabilist or a responsibilist virtue-account of knowledge-how. I argue that only a responsibilist account will maintain the crucial distinction between knowing how to do something and merely being able to do it. Such an account, I hold, must incorporate "learning how to do something" as an essential part. Drawing on an argument by Craig, I hold that the function of the concept of knowing how is to mark out practical experts whom one can trust either to (1) perform an action on one's behalf, or (2) to teach one how to perform that action oneself. The best way to identify practical experts, I hold, is to discover what they have learned about performing the action in question. In arguing for the importance of the concept "learning" to the field of knowledge-how, I argue for a new connection between the philosophy of education and virtue epistemology.