I’ve just posted a shortish paper entitled “DeRose on Truth-Value Gaps for ‘Knows.'” The title is a bit misleading–I’m not so much discussing Keith’s actual view as a view that Keith might have. Keith holds that the standards for ‘know’ in a certain context are the standards that the speaker chooses to use, rather than the ones that would be reasonable for her to use. If two participants in a conversation insist on different standards, then knowledge ascriptions that meet only one of the standards have no truth-value. I argue that, even if Keith held that the standards in a context were the standards that are reasonable to use, there’d still be truth value gaps in a lot of situations; and that even given his actual views, there are truth-value gaps in more situations than you might think.
Anyway, I’d greatly appreciate any comments anyone has.