The Development of the Philosophy of Species 315
So, the FRP, while true, is no real help in specifying which groups of organ-
isms are distinct unless you already know which characters to use. When I asked
one specialist who uses cladistic methods how he selected the right characters to
use, his reply was that he should hope that he knew his organisms well enough by
now to choose correctly. Again, we see the reciprocal illumination of which Hennig
spoke—knowing something about one’s group of organisms enables one to choose
the characters that will uncover the relationships that allow one to test which charac-
ters one should use. The circularity is not vicious—each step acts to further test the
next and to iteratively refine the overall picture.
Not surprisingly, many of the algorithms used to analyze the character sets by
pheneticists have been incorporated into cladistic analysis, only now they are used to
cluster basal taxa (species, as represented by a specimen, or by a “wild-type” set of
characters which are known or suspected to be invariant across the species, and which
are unique to it—autapomorphies, in other words). Cladistics is often accused of being
typological for this reason. There is a sense in which this is harmlessly true, and a sense
in which it is malign, but false. The type that is usually implicit in cladistic analyses is
a modal type;^92 it is the mode that most, if not all, of the members of the species bear
at the same stage in their lifecycle. This is gained empirically by field work, large-scale
sampling, and biogeographical methods. Such typology is universal in alpha taxonomy
and studies of organisms. If we had perfect knowledge of a species, then we could draw
the distribution curves for each trait and specify the modal value for use in the data
matrices used in taxonomic analysis. But the “essentialistic” sense that derives from
logic, that a species is to be defined by the presence or absence of some character/s, is
no more a part of cladism than it is of any other approach which is not “Aristotelian.”
Indeed, Aristotle’s empirical work, where he recognizes “the more and the less” in
species essences, is not, in the popular understanding of the term, Aristotelian, either.
The Qua Problem
Species are thought by some^93 to be named in a causal account similar to the Kripkean
baptismal notion of general names. We first encounter a Tiger, and give it a name
and a kind to match. When we meet other tigers, we learn what can vary and what
is common to all. In this way, the first tiger acts as a kind of baptismal “type speci-
men” for subsequent uses of the name. Now this goes to a matter in the philosophy of
language over meaning and (semantic) essentialism that to most if not all biologists
is rather recondite, but, like essentialism itself, it has direct implications in biology.
One of these is the issue of essentialism as a way of defining kinds, of course, but
another derives from an extension of a problem raised first by Willard Van Orman
Quine in his Word and Object, known as the Gavagai Problem.^94
(^92) In the statistical sense of a mode, not the philosophical sense of a modal operator. The discovery of
modal distributions of homological characters that are not homoplasious in other ways is independent
of phylogenetic analysis, and we might be able to do a cluster analysis of populations to distinguish
species, but there can be no prior values that will settle the issue, because species can be polytypic,
and the variation within them can be greater than the variation between them.
(^93) For example, Hull 1976, Hull 1984; see Kitts and Kitts 1979.
(^94) Quine 1960, 29.