Notes

[1] Duprè (1993 Dupré, J. 1993. The disorder of things: Metaphysical foundations of the disunity of science, Cambridge, MA: Harvard University Press.  [Google Scholar]) argues for “promiscuous realism,” the view that there are many inconsistent yet legitimate ways of classifying things into “natural” kinds. Duprè's promiscuity thesis outruns anything considered here. Here, I grant that natural kinds are in some way restricted to those that figure in “important” causal generalizations and that are underwritten by mechanisms. Boyd (1999 Boyd, R. 1999. “Homeostasis, species, and higher taxa”. In Species, Edited by: Wilson, RA. MIT Press: Cambridge.  [Google Scholar], p. 159–162) largely embraces Dupre's promiscuous realism, holding, for example that “lily” is a legitimate horticultural/landscaping kind, even if it is not a kind recognized by contemporary biological taxonomy. (See Wilson, 1999 Wilson, RA. 1999. “Realism, essence, and kind: Resuscitating species essentialism?”. In Species, Edited by: Wilson, RA. Cambridge, MA: MIT Press.  [Google Scholar], for an effort to defend a stronger form of realism about HPC kinds). The present challenge is that this restriction to causally important kinds (accommodation to the instrumental demands of some disciplinary matrix or another) is consistent with a level of promiscuity incommensurate with the normative ambitions for the HPC view as an arbitrator among taxonomies of kinds in the special sciences.

[2] Jaegwon Kim (1992) gives the example of jade, a conventional kind that includes both jadeite and nephrite. Membership in the kind ‘jade’ does not guarantee that a particular sample will exhibit the same physical and chemical properties, respond in like ways to similar interventions, or figure in the same explanatorily and causally significant generalizations.

[3] Keil's project is to develop a psychological theory of how organisms chunk the world into kinds, not (primarily) to develop a normative account of how they ought to, which is one of the objectives of philosophical defenders of the HPC.

[4] This manner of speaking is admittedly awkward. Boyd developed the HPC view in order to argue that species are natural kinds. He does not consider the kinds of physiology and psychology, such as the parts of animals and their characteristic behaviors, though it is clear that he intends the account to have such application. Other defenders of the HPC view have made this extension explicit (Wilson et al., forthcoming).

[5] Thanks to Ken Kendler for this example.

[6] This problem cannot be blocked by requiring, in addition, that there must be a difference in the memory phenomenon itself. Part of characterizing the phenomenon is to characterize the conditions of breakdown and intervention. This is why such research strategies are useful for shaping the space of possible mechanisms (see Craver, 2007 Craver, CF. 2007. Explaining the brain: Mechanisms and the mosaic unity of neuroscience, Oxford: Oxford University Press. [Crossref] , [Google Scholar], chapter 4). As noted above, different mechanisms, if detectably different in their entities, activities, and organizational features, will respond differently to some interventions.

[7] For an account of this constitutive variety of explanatory relevance, see Craver, 2007 Craver, CF. 2007. Explaining the brain: Mechanisms and the mosaic unity of neuroscience, Oxford: Oxford University Press. [Crossref] , [Google Scholar], chapter 4.