is a constraint on our account of ontological categories, neither Sommer’s Law nor
the existence of a topmost category can also be constraints. I suspect that neutrality is
overrated, but I also suspect that there are some categorial schemes inconsistent with
Sommer’s Law and that it is more probable that one of these is true than that each of
them is false.
There are plausible ontological systems that do not obey Sommer’s Law. Consider
the following system of ontological categories:the actual,the merely possible,univer-
sals, andparticulars. These ontological categories can cut across each other: there are
merely possible particulars as well as actual particulars, and there are merely possible
universals as well as actual ones. Smith (2004: 253–4) discusses a Cartesian onto-
logical scheme that recognizes two pairs of crisscrossing categories:substanceand
attribute, andphysicalandmental. It is not difficult to envision arguments that could
motivate these systems. Perhaps even Aristotle toyed with an ontology inconsistent
with Sommer’s Law; recall that Aristotle discusses the possibility that some qualities
might also be“relatives”in theCategories.^16 And Brower (2014: 208–11) argues that
Aquinas recognizes at least ten modes of being but only three non-overlapping
classes of (created) things that participate in them. That Sommer’s Law rules out
these ontologies byfiat is a reason to be cautious about Sommer’s Law and any
account of ontological categories that implies it. (A view of ontological categories that
implies it is not terribly ontologically neutral.)
The bulk of Westerhoff’s case for his account of ontological categories consists in
showing that competing accounts fail. Westerhoff (2005: ch. 2) provides a nice survey
of accounts of ontological categories. He dismisses“accounts”of ontological cat-
egories that in effect take the notion as a primitive, or merely produce examples of
schemes of ontological categories.^17 Westerhoff discusses (and argues against)gen-
erality accounts(2005: 25–40),intersubstitutability accounts (2005: 40–55), and
identity accounts(2005: 59–65).
In reverse order, identity accounts appeal to the notion of an identity condition,
which is closely tied to the notion of ade reessential property. On identity accounts,
things belong tokindsand part of what it is to belong to a kind is to have a certain
kind of modal profile: in virtue of belonging to kind K, certain changes across time or
worlds are not possible. Ceasing to be an instance of K is one of the changes that is
impossible; the identity conditions of most kinds are typically richer. Say that a kind
K1includesa kind K2 just in case any change across times or worlds permitted by K2
is permitted by K1, but some possible change permitted by K1 is forbidden by K2.
(^16) Aristotle’sCategories11a37 (1984a: 17). See Frede (1987: 13) for commentary on this passage.
(^17) One philosopher who refrains from defining“ontological category”is Roderick Chisholm (1996: 1),
who writes,“What a category is may be shown by depicting the table of categories that is defended here.”
Chisholm never attempts to define“ontological category.”It would be disappointing to learn that what is
meant by“ontological category”can only be shown and never said.