In "Elusive Knowledge" (1996), David Lewis advocates epistemic contextualism on the basis of an analysis of the nature of knowledge. For Lewis, the context-sensitivity of knowledge depends on the fact that 'knowledge that p' implies the elimination of all the possibilities in which 'similar to p'. But since all is context-sensitive, knowledge is also context-sensitive. In contrast to Lewis, in Knowledge and Practical Interests (2005), Jason Stanley argues that since all context-sensitive expressions can have different interpretations within the same discourse, contextualists cannot consistently embrace the following two claims: (i) knowledge functions like a quantifier and (ii) distinct occurrences of knowledge within the same discourse must be associated with the same standard. In response to Stanley, I argue that (i) and (ii) are both true. More specifically, I argue that with the help of global domains, we can overcome Stanley's objections to Lewis and, accordingly, provide the linguistic basis that epistemic contextualism needs.