4.7 The Necessity of Ontological Categories
Let’s close this chapter with a brief discussion of whether what ontological categories
there are is necessary or contingent. Westerhoff (2005) is committed to contingency;
the account I’ve offered is consistent with both.
In order to make the discussion more tractable, let’sfirst take modes of being to be
higher-order properties and let’s assume that properties have their grade of natural-
ness essentially. Given these assumptions, ontological categories are essentially
ontological categories.^70 But it is still an open question whether ontological categories
necessarily exist.
If properties in general necessarily exist, then ontological categories necessarily
exist even if nothing is a member of them. But, if properties exist only if they have
instances, then an ontological category exists only if some first-order property
exemplifies it. And this in turn is the case only if some member of that ontological
category exemplifies thatfirst-order property. Perhaps there are some ontological
categories that are necessarily non-empty; the putative category of mathematical
entities is plausibly such a category. But it is also plausible that some actual category
could have been empty; both the category of physical object and the putative category
of tropes are plausibly possibly empty.
As discussed in chapter 2, our concept of existence—assuming that we have a
unique concept of existence—is at best an analogous concept. (We will further
complicate our picture of existence in chapter 5.) The feature that it represents is
less natural than the various modes of being that are its analogue instances. Analo-
gous properties are akin to disjunctive properties but more natural than merely
disjunctive properties. But perhaps thinkingfirst about disjunctive properties will
be illuminating. Does a disjunctive property exist only if its disjuncts exist? If a
disjunctive property is literally composed of its parts, then, if mereological essential-
ism holds for properties, the answer is“yes”; otherwise perhaps not. Another
possibility is that a disjunctive property exists in any world in which at least one of
its disjuncts exist. Fine (1994b) recognizes a mode of composition that functions in
this way. And given Lewis’s (1986) view that a property is a set of actual and possible
instances, a disjunctive property is simply a set that is the union of its disjuncts. On
Lewis’s view, a disjunctive property can have an instance at a possible world even if
one of its disjuncts does not. (In general, on Lewis’s view, a set can have an instance at
a world even if not all of its instances are to be found at that world.)
So it is hard to settle the question of the modal status of existence without taking a
stand on the nature of properties in general. What we can see though is that, on one
theory of properties, existence exists only if each of its analogue instances exist; and
hence there are worlds in which some things enjoy a mode of being without enjoying
existence. This is a strange but coherent view. But I think the right view is that a
(^70) We will reconsider whether properties have their degree of naturalness essentially in section 9.7.