Abstract

Peter Unger's puzzle, the problem of the many, is an argument for the conclusion that we are grossly mistaken about what kinds of objects are in our immediate surroundings. But it is not clear what we should make of Unger's argument. There is an epistemic view which says that the argument shows that we don't know which objects are the referents of singular terms in our language. There is a linguistic view which says that Unger's puzzle shows that ordinary singular terms and count nouns are vague. Finally, there is an ontological view which says that the puzzle shows that there are vague objects. ;The epistemic view offers the simplest solution to the problem of the many, but runs foul of a different problem, the problem of vague reference. The problem of vague reference is that given the presuppositions of the epistemic view there are too many too similar objects that might be the reference of a name such as 'Kilimanjaro' for it to be plausible that the name has a determinate reference. The linguistic view, spelled out in terms of semantic indecision and supervaluation, offers the same solution to the problem of the many and to the problem of vague reference. But it leaves no room for de re beliefs about ordinary material object. The ontological view offers a solution to the problem of the many that avoids the problem of vague reference and the problem of de re beliefs. For these reasons it is preferable to the other two. ;However, ontological vagueness has met strong objections. It has been argued that it is a fallacy of verbalism, that it is inconsistent and that once formulated in a consistent way it is not distinguishable from the linguistic view. These objections can be met, but not without cost. To avoid the charge of being inconsistent, friends of the ontological view have to give up the law of excluded middle. ;A positive account of vague parthood has two parts. First, parthood is not primitive but dependent on other primitive facts. The most important of the primitive facts are about to what kinds objects belong and how objects are casually related. Second, sometimes the primitive facts fail do determine of two objects whether one is part of the other. Given a notion of vague parthood, a notion of vague object can be defined roughly in the following way: An object O is vague if there is an object a such that it is indeterminate whether a is part of O