ABSTRACT: Epistemological contextualism states that propositions about knowledge, expressed in sentences like “S knows that P,” are context-sensitive. Schaffer (2005) examines whether one of Lewis’ (1996), Cohen’s (1988) and DeRose’s (1995) influential contextualist accounts is preferable to the others. According to Schaffer, Lewis’ theory of relevant alternatives succeeds as a linguistic basis for contextualism and as an explanation of what the parameter that shifts with context is, while Cohen’s theory of thresholds and DeRose’s theory of standards fail. This paper argues that Schaffer’s analysis is unsatisfactory since it fails to show that thresholds and standards cannot cope with skepticism, as it is ultimately the conversation participants who control how the conversation plays out. Moreover, Schaffer fails to show that gradability is of no importance in inquiries.