Much work in epistemology, old and new, is devoted to explaining – or explaining away – the appeal of skeptical arguments (that is, arguments designed to impugn most, if not all, of our claims to know), as well as the fact that it also often seems as if we use knowledge-attributing sentences to express truths. More recently, philosophers have sought to explain other kinds of variability in our use of knowledge-attributing sentences besides, such as the kinds of variability that occur as a result of changing what is at stake or the salience of alternative possibilities, as in scenarios such as Keith DeRose's “bank cases.” These cases, and others like them, pose a dilemma very much like that posed by skeptical arguments. For they appear to suggest that we are wrong about something. Either we speak falsely when we attribute knowledge in the situations under consideration, or we speak falsely when we deny it. Or so it would seem. This paper argues that an alternative approach to explaining the relevant variability in our use of knowledge-attributing sentences – namely, a fictionalist one – merits serious consideration.