In this paper, I want to argue-with certain qualifications-that there cannot be any vague identities, and to outline reasons for scepticism about the view that the world contains vague objects. I also argue that, even if there were vague identities, this would lend no support to the vague-objects view. 1. What would constitute a defence of the vague-identity thesis? It would be an example in which a sentence of numerical identity is indeterminate in truth-value (i.e., neither true nor false), where the indeterminacy is due to vagueness. (Thus we are not concerned with identity sentences whose indeterminacy is due, e.g., to referencefailure or to cross-category identification.) It seems clear that there are such examples: where the vagueness of an identity is a consequence of the vagueness of one or both of its singular terms. For example, the singular term 'the world's greatest ruler' is vague because of the vagueness of the predicate '. . . great ruler'. This predicate is vague, not because it lacks sharp boundaries, but because of its multi-criterial application conditions. Many different factors contribute to the greatness of a ruler-wisdom, fortitude, diplomacy, prudence, etc.,-and the rules of our language do not fix in advance what weight to assign to each factor. Because of this vagueness, the singular term 'the world's greatest ruler' has no determinate reference: it is vague which person it singles out. (Though, as Wiggins has emphasised, from the fact that it is vague which object a term singles out, it does not follow that it singles out something vague.)' Consequently, an utterance of, e.g., 'the world's greatest ruler was the world's wisest ruler' is a plausible example of a vague identity.2