HOMEOSTATIC PROPERTY CLUSTER ANALYSES
Let us examine a final case in which naturalistic grounds for advocating a particular thesis on the ontological status of species do not clearly carry the day. In the case of SAI, we considered concerns over a positive metaphysical account despite its being naturalistically motivated. This may be disappointing for naturalists, but perhaps it simply represents an overreach. It thus may be instructive to consider a more modest critical argument against a particular metaphysical perspective on species that proceeds on naturalistic grounds but which also (I think) fails.
The case involves Boyd’s Homeostatic Property Cluster (HPC) account of natural kinds (Boyd 1999a), an account that many philosophers of biology argue evades the criticisms levied against essentialistic conceptions of natural kinds (see, e.g., Rieppel 2007; Wilson, Barker, and Brigandt 2007). Species might be natural kinds, on this view, in virtue of sharing a cluster of properties, perhaps imperfectly, whose broad stability/sociability is maintained via the operation of certain homeostatic mechanisms. In my view, the HPC account faces a number of serious theoretical issues and problems for its application to species taxa; I am not an advocate (for criticism, see, e.g., Haggqvist 2005; Reydon 2009; Slater 2013, §6.2.2; 2015). But it has recently come under some criticism that I believe represents naturalist-overreach.
A chief critic has been Marc Ereshefsky, who has argued in several papers that HPC theory offends from prevailing views in biological systematics. Naturalistic concerns are clearly front and center in his argument; he writes (with
Mohan Matthen): “The HPC account of biological taxa should be informed by, or at least consistent with, biological systematics. But that does not seem to be the case: the historical essentialism of HPC theory is not historical enough for contemporary biological systematics” (2005, 17). Why not? The answer stems from some critical comments Boyd levies against constraints certain systematists insist upon but that Boyd finds to be unmotivated (1999a, 180-82; 2010)— in particular, he allows that HPC kinds might be para- or polyphyletic. Here Ereshefsky objects, “This should not rest easily with those who believe that [species or] higher taxa must be monophyletic” (2007, 296).
But the issue here is a bit more subtle than Ereshefsky allows. We can agree to the insistence that HPC theory ought to be informed by and consistent with empirical results in biology; but it is not clear that it must be just as demanding as a particular school of systematics that not all biologists happen to advocate. This would be a very strong and implausible interpretation of consistency. Does consistency with biology require that biologists be disposed to adopt (or at least not reject) the theory in question? That is not obvious either. I argued above that SAI’s popularity among biologists was not particularly probative; likewise, the fact that some biologists may be expected to reject HPC theory does not, in and of itself, show that it is inconsistent with biology. After all, biologists—even those with a phylogenetic orientation—disagree with each other about how to organize taxa (de Queiroz 1999; Wheeler and Meier 2000). Consistency with biological systematics will remain elusive as long as extant theories of biological systematics are not consistent with each other.
But set this biological conflict to one side. Suppose (what is radically contrary to fact), that biologists were united in advocating monophyly as a constraint on how species and higher taxa ought to be individuated. Would this not reinstate the force of the naturalist criticism of HPC? Perhaps it would make the case stronger, but again I do not believe that HPCers ought to feel forced to capitulate. For just as we are entitled to differentiate between the epistemic worth of particular empirical theses, we should take seriously the potential for acknowledging epistemic differences between such theses and prevailing classificatory practices. While it is tempting to place all aspects of scientific practice onto the same epistemic pedestal as its acknowledged fruit, we should recognize the possibility that such practices do not contribute equally to the epistemic worth of science and so do not clearly merit trumping force when apparently in conflict with metaphysical theses. Similar considerations can be applied to naturalistic arguments stemming from the ways in which a particular community of scientists talk. If there are other possible ways of classifying, investigating, and talking that would lead scientists to comparative success, we ought to be cautious about reading off conclusions from these considerations.
I will return to some consequences of this stance in the final section. But before we move away from our final case study, it is worth noting two further contrasts between the (in my view) successful anti-essentialist case and this unsuccessful anti- HPC case. In the former case, essentialists were making a claim with content that conflicted with a well-established empirical result in biology—for example, that species taxa share a common “genetic structure” that makes them the species they are. That seems not to be true (nor should we expect it to be true, given other things we know about evolution, development, and the sources of variation). In the latter case, it is less clear that the conflict between HPC and cladists has the same empirical character. Cladists may say “biological taxa are monophyletic,” but it is not evident that this is an empirical claim in the same way that “biological taxa are genetically heterogeneous” as much as it is a normative claim about how we ought to group organisms together—that is, “biological taxa should be grouped according to mono- phyly.” Such a claim is not, as it were, read off from the data. Granted: it is an oversimplification to suppose any scientific claim can be so read, or regarded as “true on inspection,” but arguably there is a qualitative difference here.
There is more to say here, of course, about classificatory norms—how they interact with empirical content of scientific theories, how they come to be, how they are embedded in scientific practice more generally. My discussion has so far remained rather general (and so it must remain in this context). It is premised on an undefended but I think plausible claim: that classificatory practice, like any component of scientific practice, results from a complex interaction between contingent normative commitments about values, aims, and significance-ascriptions and empirical facts (which are in turn partially conditioned by those norms); such complex interactions will involve trade-offs and thus tend to demand a pluralistic outlook in those cases. But there is no similar call to be pluralistic about claims involving the degree of genetic homogeneity we find in certain kinds of biological taxa (so long as one fixes on a particular understanding of the taxa involved).
The second contrast with the anti-essentialism case is the way in which essen- tialism and views like HPC may be applied to biological taxa. Devitt’s “Intrinsic Biological Essentialism," for example, is the thesis that “Linnaean taxa have essences that are, at least partly, intrinsic underlying properties" (2008, 346). This could be interpreted either as an empirical claim about the genetic structure of already identified (and identifiable) taxa or as a normative constraint on how taxa should be identified (or, if need be, revised). On either interpretation, there is a conflict with our best science—with what it takes to be true or how it goes about systematics. HPC, on the other hand, drops the former empirical claim and offers a far more flexible framework for recognizing “natural kind phenomena." In my view, treating species as HPC kinds need not be taken as a constraint upon scientific practice (cf. Boyd 1999b, 162-63). Ereshefsky’s criticism is that HPC takes the wrong orientation to identifying taxa. While he allows that “there may be many HPC kinds in the world," he seems to deny that any biological taxa are among them: “HPC’s emphasis on similarity is at odds with phylogenetic approaches to taxonomy. HPC kinds are fundamentally similarity classes, whereas taxa are fundamentally genealogical entities" (2007, 296). But this ignores the possibility that “genealogical entities" may also have members (or parts, if you insist) that share certain clusters of properties in common due to the activity of certain homeostatic mechanisms. The fact that being an HPC kind involves no genealogical qualifications clearly does not mean that things that do meet such qualifications cannot be HPC kinds; a metaphysics of species need not be the sole source of factual claims about species—it had better not be if we are to remain in generally naturalistic territory.
-  A quick and dirty primer on some of these terms: taxa are monophyletic if they consist of all and only thedescendants of a common ancestor; advocates of the monophyly requirement are sometimes called “cladists,”though (confusingly) this term is also sometimes used interchangeably with more general “phylogenetic”approaches to systematics that see historical approaches to taxonomy as fundamental. Different phylogeneticapproaches might recognize paraphyletic taxa, which contain only descendants of a common ancestor, but notall of them (it might leave a branch off) or polyphyletic taxa, dropping the “only” requirement as well: theymay comprise some descendants of multiple lineages.
-  2 Note also that even if it turned out that species are substantially artifacts of our classificatory whims, the empirical content of these claims would still differ.
-  In other work (Slater, n.d.), I have argued that much of classificatory practice is guided (and not determined)by various classificatory norms that, while responsive to empirical facts, are not fixed by them.
-  To be fair to Ereshefsky, it sometimes seems that Boyd is using HPC to criticize the monophyly requirementimposed by some schools of systematics. I do not have the space to evaluate this framing of the dispute; theissues here are complex.