Biological and Cultural Proper Functions in Comparative Perspective 41
they still have winnowing as a proper function, even though that is not what they are cur-
rently being selected for.
So it looks like cultural selection will have all the same problems as natural selection
with regard to picking out proper functions. So far the problem has been that although
natural selection does succeed in picking out most proper functions, there are some func-
tions we want to call proper functions where natural selection is not operative. But Robert
Cummins (2002) has reentered the function theory fray with a much more fundamental
critique. On his view, natural selection does not pick out proper functions even in the
favorable cases in which it does occur. The target of Cummins’s argument is a widely held
view he calls neo-teleology. It claims, fi rst, that biological organisms have the traits they
have because of the functions those traits fulfi ll (e.g., vertebrates have hearts because
hearts circulate blood), and second, that natural selection supplies the connection between
function and the existence of traits by selecting traits for their functions (hearts have been
selected because they circulate blood, and this accounts for their existence and ubiquity).
On this view, natural selection accounts, fi rst, for the very existence of traits because it
“builds” them incrementally, and second, for their spread through populations of individual
organisms. With regard to the “building” of traits, Cummins agrees that natural selection
is largely responsible for this, but points out that it works by means of a piecemeal, long-
term process that is entirely insensitive to the ultimate proper function of the trait (2002:
168–169). With regard to the spread of traits, Cummins again agrees that this is often
accomplished by natural selection, which in this case is sensitive to function—but only to
the relative success with which a function is performed by the current variants (2002:
164–165). For example, natural selection does not select between winged pileated wood-
peckers and wingless pileated woodpeckers, but between pileated woodpeckers with better
and worse wing designs. And the crucial point here is that both better and worse wings
already have the proper function of enabling fl ight. In short, in the case of both the build-
ing and the spread of traits, what natural selection selects for does not correspond to the
proper functions of these traits. So natural selection does not pick out proper functional
performances.
The question then is whether an analogous argument goes through for material culture.
Let us fi rst consider spread. Like natural selection, cultural selection also appears to spread
types and traits of artifacts only by selecting among better and worse variants with the
same proper function. For example, quill pens and fountain pens both have the proper
function of writing, but fountain pens are more effi cient because they have a larger ink
reservoir and a more durable writing point. So they proliferated while quill pens fell into
disuse. Nor does the ink reservoir—which, as a component part, is more nearly analogous
to the wings of birds in our biological example—represent a new proper function. The
hollow shaft of a quill pen is also an ink reservoir but an ineffi ciently small one in com-
parison to that of the fountain pen. While there may be some rare cases in which a com-
pletely novel artifact capable of a novel performance is introduced, this is the vanishingly