Abstract

How does one so much as attempt to think new thoughts about the same old things? When one recognizes a place or a person, when one expects an item to have some feature and sees the expectation fulfilled or not, when one reasons about a thing by considering its properties in turn, one's concern is for the same individual encountered or thought of before. More is required to bring off these cognitive feats than simply identifying an individual successfully on distinct occasions. One must also think of the individual as being the same, or re-identify it. Being able to re-identify individuals puts one in a position to bring facts together in useful ways, to make warranted inferences, and to navigate through a changing landscape. ;Important as this capacity is, until now we have lacked an adequate theory of re-identification. All the models currently on offer, I argue, fail to capture the phenomenon of re-identificatory thought Specifically, extant theories fail to explain re-identification as the act of a subject whose aims involve coping with temporal change, and whose resources mid goals vary with respect to the things identified. Most importantly, none of the existing models makes sense of how re-identification warrants thinkers in making inferences about the stable objects of their thought. Fundamental self-knowledge claims cannot be warranted on such views. ;Taking account of these features, I contend, requires understanding re-identification as springing from a thinker's ability to institute traditions and policies of thought. On my account, re-identification is essentially policy-governed cognition. I show how we can capture the notion of maintaining cognitive policies in terms of maintaining a file about an individual. A file, I argue, can be understood as a cognitive particular with functional-historical identity conditions, and file-management can be understood in terms of information-handling dispositions. So specific feats of file-management constitute the institution of the cognitive policies which in turn constitute re-identificatory thinking. Finally, I investigate the nature of the inferential warrant that derives from policy-governed cognition. I claim that one's warrant for inferences about individuals derives from one's own epistemic vigilance in maintaining a file of information on an individual. Such warrant turns out to be less than foolproof, since one can only be so vigilant. Nevertheless, such warrant is sufficient, I argue, to ground fundamental claims of self-knowledge