536 Tim Lewens
in vision explains the presence of lenses, it explains the presence of earlobes, too.
Why not say, then, that earlobes have helping to see as one of their functions,
hence that all earlobes are malfunctioning? We need to clarify the SE account. A
token trait of type T has function F just in casepast tokens of the same type T
contributed to fitness by producing effect F, and thereby explained the proliferation
of tokens of type T. Earlobes don’t have helping to see as a function, because the
past token traits of that same type did not contribute to fitness by assisting in
vision. But if we are to say this, we need some way of picking out trait types
independently of how they have contributed to fitness. The obvious way to do this
is by appeal to homology. So, on this view, we sort tokens into types using the
concept of homology, and we then assign functions to trait tokens according to
how earlier homologous tokens contributed to fitness. Any classification of traits
by function is parasitic on a prior classification by homology.
5 MALFUNCTIONThis discussion of homology sets up useful background for a more serious problem
that faces the selected effects account. Elsewhere I have argued that the SE
account is not, in fact, uniquely able to meet the desiderata usually accepted for
a theory of functions [Lewens, 2004]. Rather than repeat those arguments here,
I want to ask why an account of evolutionary function needs to make sense of
malfunction at all. (My arguments here draw on those by [Davies, 2000].) My
attack will be two-pronged. On the one hand, I want to downplay the theoretical
interest in the category of malfunction from the perspective of evolutionary biology.
On the other hand, I want to suggest that what the selected effects account says
about malfunction is rather awkward. The result is that I am sceptical of whether
the ability of an account of biological functions to ground claims about malfunction
should really be viewed as a desideratum for such accounts.
Let us begin with the question of the role of malfunction in evolutionary theory.
Variation is a feature of all biological populations. What is the theoretical interest
in marking some of this variation asmalfunctioning? The evolutionary biologist is
primarily interested in explaining the changes of frequencies of various traits in a
population. Now we can assign functions to traits of various types according to the
effects that they have which account for their representation in the population as
a whole. Here we are, in effect, treating the population as a system, and treating
classes of resembling traits as parts of that system whose frequencies in the system
stand in need of explanation. Within this type of explanatory project there is no
need to posit the category of malfunction as a correlate of function: this kind of
project is a Cummins-style functional analysis of a population, and we have seen
that such an analysis stops short at a purely descriptive account of the systems
whose capacities it explains.
Some traits that arise as single mutational oddities in a large population —
whether fitter than the average or less fit — may be so rare that they can be
ignored in the analysis of that population. These, however, can be viewed as