186 Marzia Soavi
contrasted with nominal kinds. These are kinds that collect objects that do not necessarily
share any common nature, again bachelor, vixen, and widow are traditional examples of
nominal kinds. For any metaphysics that accepts the existence of real kinds, if an object
o is individuated as an object of a kind K then o is a real object only if K is a real kind.
For example, we can individuate the same portion of matter as a certain amount of clay
or as a statue, but if we do not admit that the kind of statue is among real kinds then the
statue is not a real object. If we acknowledge the kind of amount of clay, then the amount
of clay is a real object. If we acknowledge that both the kind of statue and the kind of
amount of clay exist and that criteria of identity for statues and amounts of clay do not
coincide, then we are bound to admit that there are possibly two objects occupying the
same space at the same time.
If we accept the idea that when natural kinds strongly support induction they cannot do
it by accident but by selecting precisely those objects that share the same nature, then it
is easy to see why the distinction between natural and artifi cial kinds is assimilated in the
distinction between real and nominal kinds. Natural kinds strongly support induction
because they individuate real kinds, that is, they collect objects with the same nature and,
ideally, the nearer they get to the individuation of some real kind the stronger their support
of induction will be.
Artifi cial kinds, by contrast, are nominal kinds; they do not necessarily collect objects
that share the same nature and they trace distinctions simply according to our needs,
beliefs, or linguistic practices.
This relation has sometimes caused confl ict between the epistemological and the onto-
logical distinction, leading to the use of natural kinds as a synonym for real kinds. In such
cases the qualifi cation “natural” has nothing to do with the notion of “natural” that is nor-
mally contrasted with that of “artifact.” The distinction between artifacts and natural
objects is controversial and highly problematic. Generally speaking, artifacts can be con-
sidered man-made objects, mostly those produced to perform a certain function, while
natural objects are not man-made. Indeed there are objects that are deemed to be natural
objects that are intentionally produced, but I am not interested in the defense of this dis-
tinction here, nor am I interested in a refi ned version of it. What is important is to keep
the two notions of “nature”—one that is compared to the notion of “artifact” and one that
is compared to the notion of “artifi cial kinds”—clearly distinct, otherwise the expression
“natural kinds” seems to lead to the trivial conclusion that artifact kinds are not real kinds
while, conversely, there is nothing in the idea that an object is intentionally produced that
can allow us to infer that there is no real or natural kind to which such objects belong.
That is, there is nothing in the assertion that o is an artifact that can allow us to infer the
assertion that o is not a real object.
Devitt claims that artifact kinds are not entitled to be real kinds because they are meta-
physically unnecessary—artifacts already fall under common physical kinds (unfortu-
nately he does not provide any example of such common physical kinds). According to