Part of Frege's concern about whether number words are properties of objects was that if they could be construed as such it would lend support to the view that truths of arithmetic were empirical truths. Such concern is ill-founded. Even if number words do apply to objects as predicates, this does not entail that numerical truths would be empirical, any more than the fact that ‘bachelor’ and ‘unmarried’ are predicates of objects entails that their relationship is an empirical one. The (...) account of number words as predicates given here, consequently, should not be taken as supportive of any one particular theory concerning arithmetic truth rather than another.The view of number words that emerges above from the criticism of Frege's arguments is that sentences containing number words can and do say something about objects in a collective and syncategorematic way. That is, they say something about classes of objects, when it is clear what is to count as an object. Whether this view commits us to accepting certain ontological entities, classes of individuals, above and beyond individuals themselves, I am not prepared to say, and at any rate this question falls beyond the intended scope of this essay. To those who would eschew abstract entities, however, it may be noted that the ontological problems that arise are not peculiar to this view; they seem no more or less severe than those that arise in connection with other collective predicate statements, as, for example, “Blue whales are becoming extinct.”The view presented here also does not pretend to be a fatal criticism of the logicist program. Taking our cue from the above noted fact that syncategorematic predicates share in a common ‘partial’ abstract sense, it seems reasonable to say that it is just the set of such partial senses of number words, and the relationships that hold between such senses, that the logicist and the pure number theorist investigate. (shrink)