It's wrong to think of Homeostatic Property Clusters as a definition of natural kinds.
The Homeostatic Property Cluster (HPC) account of natural kinds has become popular since it was proposed by Richard Boyd in the late 1980s. Although it is often taken as a defining natural kinds as such, it is easy enough to see that something's being a natural kind is neither necessary nor sufficient for its being an HPC. This paper argues that it is better not to understand HPCs as defining what it is to be a natural kind but instead as providing the ontological realization of (some) natural kinds.
@ARTICLE(Magnus2014,
AUTHOR = {P.D. Magnus},
TITLE = {NK$\ne$HPC},
JOURNAL = {Philosophical Quarterly},
YEAR = {2014},
MONTH = jul,
VOLUME = {64},
NUMBER = {256},
PAGES = {471--477},
DOI = {10.1093/pq/pqu010}
)
The first on-line draft of this paper was posted 28nov2013.