32 Mark Perlman
acknowledging the validity and usefulness of Cummins-style systematic functions, and
argued for a kind of pluralism about functions. Others have advocated pluralism as well,
such as Peter Godfrey-Smith (1993), Beth Preston (1998), and Peter Schwartz (2004). But
I think we should also not view the various kinds of functions theories as merely expound-
ing technical notions (as Millikan urges in LTOBC). If any of these varieties of theories
of function is to help us explain events, objects, and structures in the world around us,
then I would think the functions of which they speak must be real parts or aspects of the
world. To play instrumentalist about these notions of functions would deprive them of real
application to the world. But this is not to say that any or all of them are the meaning
behind a single intuitive notion of “function.”
A pluralist approach does not mean we must beg off of all the debates and adopt a kind
of tentative relativism. For instance, Cummins (2002) gives powerful and compelling
arguments against Millikan’s view of about function and natural selection. As he points
out, natural selection doesn’t select between winged and wingless sparrows, but between
sparrows with better and worse wing designs, where both the good and the bad wings have
the same proper function—enabling fl ight. So Cummins concludes that selection is not
sensitive to function, and does not pick out proper functions. Nothing about teleo-pluralism
prevents us from endorsing some conclusions about what certain kinds of functions can
and cannot explain. We should recognize that each of the various notions of “function”
has different strengths and weaknesses, and will be appropriate in explaining different
things. I argue elsewhere (Perlman 2000, 2002) against teleosemantics, the view that
teleofunction can explain the mental representation and misrepresentation. The problems
of indeterminacy of mental content reappear in the teleofunctions that are supposed to
explain content, and thus functions can’t hold the weight of explaining content.^10 So to be
a pluralist about functions is not to be a noncommittal relativist. And yet there is still room
for those who (like Millikan) think functions can be determinate enough to ground mental
content to argue their case.
While various people have tried to unify the notions of “function” into one (Griffi ths
1993; Buller 1998; Kitcher 1993, with his notion of “design”; Walsh and Ariew 1996, with
the relational theory), I think we shouldn’t focus the argument on what the “real” defi nition
of function is, but rather on which conception of “function” will adequately explain which
phenomena. That being said, we should also not say that the different notions of “function”
are restricted to different fi elds within biology. The difference has to do with kinds of
explanation, and what is being explained, not the fi eld of investigation itself. In some cases
a focus on Cummins-style systematic functions will give us the best explanation of the
phenomenon in question (many of the issues in biochemistry, neuroscience, or develop-
mental biology). In other cases, an evolutionary account such as Millikan’s account will
be preferable. In still others, a focus on recent-past (goal-contribution) will do better than
focus on distant evolutionary past (especially for talking about the designer’s intention as
a source of the function of a human artifact). In still other cases, future-looking propensity