Abstract

Epistemological contextualism is often defended by appealing to the context sensitivity of our intuitions about knowledge ascriptions. A popular invariantist response is to explain this feature by an appeal to pragmatic implicature. In this paper I argue that this rejoinder faces a hitherto underestimated problem relating to the fact that such supposed implicatures do not appear cancellable, contrary to what we should expect. I defend contextualism by demonstrating that the current invariantist explanation of this lack of cancellability is unsuccessful, owing to a failure to appreciate the diversity of circumstances in which knowledge ascriptions may be used