In this dissertation I argue that Wittgenstein’s discussions of simple forms of intelligent behaviour such as reading or sign-following provide the basis for a novel response to the question ‘what is it to act for a reason?’. Responses to this question in the contemporary literature focus on explanations of action, and aim to show in general and abstract terms why we count what is represented by such explanations as instances of ‘acting for a reason’. Building on ideas from the Philosophical Investigations, I show why accounts that take this form must fail in their attempts to explain what it is to act for a reason. In place of their abstract and top-down approach, I argue that we must begin by showing why we count particular and concrete examples of intelligent behaviour such as those described above as instances of ‘acting for a reason’. Focus on these cases also helps show that—contrary to some contemporary criticisms of his work—Wittgenstein gives us the resources to understand how explanations that represent our reasons for action are related to explanations that represent the causes of our action.