Ordinary agents are subject to changes in their position in the world: for example, their position in time. What is the appropriate cognitive response to these changes? I argue that, to answer this question, we should recognize a new kind of cognitive operation, which I dub `cognitive relocation'. My account of this operation revolves around two main ideas. First, relocation manipulates a special kind of information, de re information. Second, differently from processes of learning, relocation is essentially a way of retaining rather than gaining information. I go on to consider some philosophical consequences. In particular, I show how I reconcile intuitions about a well-known case, the Sleeping Beauty scenario, with a plausible but problematic principle about context-dependent information, the Relevance-Limiting Thesis.