310 Specieslost or amalgamated into other species, and varietal or subspecific names are equally
unstable. In the Phocidae (earless or “true” seals) there are eight cases of genera
being reassigned or created from existing genera out of the ten current genera (most
of which are formed from Phoca, a genus assigned by Linnaeus himself). Therefore,
there is a problem of metaphysics in terms of how the species category changes under
evolutionary assumptions, and one of conventional and operational considerations
forced by the nomenclatural instability due to splitting and lumping.
We should distinguish diagnosis from causation; the epistemological from the eti-
ological aspects of species and other taxa. In short, the appropriate way to approach
the problem is to look at species as the results of history rather than of investigation;
the “history, not characters” approach. If species are the results of a physical—that is
to say, biological and causal—historical sequence, and it turns out that no empirically-
based discipline can fully and reliably recognize them or talk about them, so much
the worse for the science. That would be a fact about biologists and their capacities
rather than biological organisms. But there is no need to be an epistemic nihilist. So
far, the science does very well, and the prospect of imminent catastrophic failure is
remote. The point is, however, that the species problem cannot be resolved by con-
vention, fiat, or practical decisions, for it is a problem about the organic world more
than it is about the scientific community, even though there is an element of social
construction in any discussion of a scientific concept (because, in order to discuss
it, we must enter into the scientific community to some degree). How we resolve
this depends a lot on what we individually conceive the relationship to be between
scientists and their study objects. But we need not be scientific realists in order to
think that a scientific term should have its meaning fixed by features of the objects
to which it refers. All that is required is a strong empiricism, and experimentalism—
the fitting of models to data, and not data to models. More abstract questions con-
cerning the status of such concepts as “cause,” “truth,” “reference,” and “meaning”
can be bracketed off from any discussion of the causal role and referential status of
a term like “species” and deferred to a more competent forum, since whatever is the
case with “species,” “gene,” or any other disputed or accepted general scientific term
is true of all scientific terms or concepts. If “species” is theory-dependent, then so is
“organism.” If either are empirically dependent, all scientific terms may be. If one
can refer or fail to refer, all of them can.
However, terms do not exist in Plato’s heaven—at least, not those derived from
science. They exist in the practices of scientists in the scientific community, in sci-
entific communication. Moreover, they do not remain constant; concepts evolve.
They can have quite different meanings at one time or at another. Agassiz “refuted”
Darwinism by effectively arguing out of a dictionary: species are by definition static
entities, and hence they could not evolve; Darwin could not therefore account for
species. To which Darwin reacted with a redefinition: species are temporary things.
Argumentum ad lexicon can be countered by a different translation manual.^81 Are
Darwin’s temporary species “the same” as Agassiz’s permanent ones, or is Darwin just
changing the subject? This could be an example of taxonomic incommensurability,^82
(^81) I owe the translation manual point to Paul Griffiths.
(^82) Sankey 1998.