408 Marc Ereshefsky
lieves that some species taxa are spatiotemporally continuous entities, while other
species taxa are spatiotemporally scattered groups whose members share struc-
tural similarities. Such species are defined by their members having theoretically
important genetic, chromosomal, and developmental similarities. Thus, Kitcher
believes that there are two types of species: “historical” species and “structural”
species. Historical species are defined by genealogy; structural species are defined
by significant similarity. A problem with Kitcher’s account of species is that it is
out of step with contemporary biological practice. Biologists do not posit struc-
tural species. A quick glance at a biology text reveals that historical approaches to
species, not structural approaches, are the going concern in biology. The question
of whether species are kinds or individuals is a question about how the theoretical
term ‘species’ functions in biology. Kitcher’s argument against all species being
individuals posits an account of species that is outside of contemporary biology.
2.3 Homeostatic Property Cluster Kinds
Another response to the species are individuals thesis is offered by proponents
of an alternative approach to natural kinds. According to Boyd [1999a; 1999b],
Griffiths [1999], Wilson [1999], and Millikan [1999], species are natural kinds on
a proper conception of natural kinds. These authors adopt Boyd’s Homeostatic
Property Cluster (HPC) theory of natural kinds. HPC theory assumes that natural
kinds are groups of entities that share stable similarities. HPC theory does not
require that species are defined by traditional essential properties. The members
ofCanis familaris, for example, tend to share a number of common properties
— having four legs, two eyes, and so on, but given the forces of evolution, no
biological property is essential for membership in that species. For HPC theory,
the similarities among the members of a kind must be stable enough to allow
better than chance prediction about various properties of a kind. Given that we
know that Sparky is a dog, we can predict with greater than chance probability
that Sparky will have four legs.
HPC kinds are more than groups of entities that share stable clusters of sim-
ilarities. HPC kinds also contain “homeostatic causal mechanisms” that are re-
sponsible for the similarities found among the members of a kind. The members
of a biological species interbreed, share common developmental programs, and are
exposed to common selection regimes. These “homeostatic mechanisms” cause the
members of a species to have similar features. Dogs, for instance, tend to have four
legs and two eyes because they share genetic material and are exposed to common
environmental pressures. An HPC kind consists of entities that share similarities
induced by that kind’s homeostatic mechanisms. According to Boyd, species are
HPC kinds and thus natural kinds because “species are defined...by...
shared properties and by the mechanisms (including both “external” mechanisms
and genetic transmission) which sustain their homeostasis” [1999b, 81].
HPC theory provides a more promising account of species as natural kinds
than essentialism. HPC kinds need not have a common essential property, so