474 Robin O. Andreasen
tionary trajectories but exist today as part of the larger population.Though this objection applies to Kitcher’s account as well as to the cladistic
account, one might argue that it is particularly problematic for the cladistic race
concept. One reason why cladistic classification is typically applied at or above
the species level is because monophyly requires evolutionary branching without
reticulation. Since significant gene flow prevents branching, it would thus impede
the formation of cladistic races [Zack, 2002; Gannett, 2004].
This is probably the strongest objection to phylogenetic conceptions of human
race — yet, it is not fatal. Part of the reason is that, as noted above, there is some
disagreement over the significance of Lewontin’s data. Edwards [2003], for exam-
ple, questions Lewontin’s analysis of his data on the grounds that it fails to take
gene correlations into consideration. In addition, he adds that once gene correla-
tions are taken into consideration, it is possible to derive a convincing branching
diagram of human evolutionary history. Like Edwards, many other human ge-
neticists accept Lewontin’s analysis of his data, but nonetheless maintain that the
allele frequencies differences that exist are highly structured and are useful for
identifying distinct lineages [Prichardet al., 2000; Risch, 2002; Rosenberg, 2002;
Bamshadet al., 2003]. Taking a different (but consistent) approach, Andreasen
[1998] has suggested another reading of Lewontin’s data. On her view, the data
are consistent with the possibility that, while there may be significant outbreed-
ing today, there was once enough reproductive isolation for distinct lineages to
have evolved. Indeed, this is the main reason why she suggest the possibility that
human races once existed, but are now anastamosing.
That being said, I would like to add that Andreasen and Kitcher each maintain
that their thesis is best understood as a conditional claim: If there is, or has been,
reasonable reproductive isolation among human breeding populations, then it is
possible to provide a biologically significant definition of human race. Both are
aware of competing models of human evolution — such as the trellis model or the
multiregional evolution hypothesis — which assume that there is, and always has
been, too much gene flow among human populations for phylogenetic differentia-
tion to have occurred [Wolpoffet al., 1988; Wolpoff, 1989; Thorne and Wolpoff,
1992; Templeton, 1999]. Furthermore, both acknowledge that if such a model were
to be empirically vindicated, then human races don’t exist and never have existed.
A second objection to phylogenetic conceptions of race has been advanced by
Lisa Gannett [2004]. Gannett’s objection has three-parts.^29 Her main concern is
that the branching diagrams (constructed by defenders of candelabra models of
human evolution) upon which phylogenetic conceptions rest are not empirically
well supported. They are not empirically well supported, on her view, because
they lack independent confirmation that the breeding populations used in tree
(^29) Gannett’s objection focuses largely on the cladistic race concept, but since much of what she
says applies to other phylogenetic conceptions as well, I will present her objection as being more
general.