Abstract
In this paper I defend two claims regarding complex demonstratives. On the one hand I argue that, when one of these expressions misdescribes the referred object, the right semantic analysis is to treat the expression as uninterpretable. On the other, I claim that the uninterpretability thesis finds itself in trouble when it comes to dealing with non-deictic uses of complex demonstratives, i.e., uses of “that F” in which the speaker has no acquaintance or direct perception of any object that could satisfy the predicative material. In order to make these two claims compatible, I set out to modify the way the uninterpretability thesis has been traditionally formulated.