Hume’s view of reason is notoriously hard to pin down, not least because of the apparently contradictory positions which he appears to adopt in different places. The problem is perhaps most clear in his writings concerning induction - in his famous argument of Treatise I iii 6 and Enquiry IV, on the one hand, he seems to conclude that “probable inference” has no rational basis, while elsewhere, for example in much of his writing on natural theology, he seems happy to acknowledge that such inference is not only reasonable, but is even a paradigm of reasoning against which the theistic arguments must be judged. In the face of this apparent contradiction, many recent commentators have proferred “non-sceptical” interpretations of Hume’s argument concerning induction, but in this paper I sketch an alternative and perhaps less radical method of resolving the problem, by identifying a major threefold ambiguity in Hume’s use of the word “reason”. On this interpretation, Hume indeed sees induction as a paradigm of reasonableness in what is arguably the most important sense, but he nevertheless believes induction to be entirely non-reasonable in another sense, which though less important in common life is nevertheless very significant philosophically. A comparison with Locke can help to illuminate Hume’s position, which though indeed not entirely sceptical about induction, is by no means entirely non-sceptical either.