Proper names have exercised analytic philosophy since its inception. This is no accident – ever since Frege and Russell made their seminal contributions, the semantics of singular terms has been intimately linked to the nature of the proposition and the propositional attitudes, and thus to the nature of thought about things in the the world and its very possibility. But lately, the debate has been in a bit of a rut. The anti-descriptivism that swept the field in the wake of Kripke's Naming and Necessity has been more or less deadlocked with various forms of descriptivism for quite a while. That might be about to change, however: Now, both predicativists and variabilists argue that their views allow for breaking the deadlock by explaining expanded, more varied sets of data than those traditionally considered. This is exciting, but there are some quite fundamental questions that we must not lose sight of. Some of these concern the relevance of data: What kind of data hold lessons specific to names? Others concern consequences for the nature of thought: Are the semantic values we work with suitable as contents for the propositional attitudes? In this talk, I shall argue for a fairly conservative take on the data. Breaking the deadlock is possible nevertheless; I'll tell you how.