From its inception, the philosophy of action has sought to account for action in terms of an associated kind of explanation. The alternative to this approach was noticed, but not adopted, by G.E.M. Anscombe. Anscombe observed that a series of answers to the reason-requesting question ‘Why?’ may be read in reverse order as a series of answers to the question ‘How?’ Unlike answers to the question ‘Why?’, answers to the question ‘How?’ are not explanatory of what they are about: they reveal, not reasons for doing something, but ways of doing something, and they have the form of what Aristotle called a practical syllogism. The alternative to theorizing action in terms of explanation, is, thus, to theorize it in terms calculation. In exploring this alternative, I argue for three main theses: first, that (pace Anscombe) it is not a matter of indifference whether we theorize action in terms of the question ‘Why?’ or in terms of the question ‘How?’; second, that the question ‘Why?’ is a question for an observer of action, whereas the question ‘How?’ is a question for the agent; and finally, that the standpoint of the agent, revealed by the question ‘How?’, is prior to that of an observer, revealed by the question ‘Why?’.