Functions and Norms 95
thumping sounds, the heart nonetheless was selected only for pumping, not for thumping.
It is supposed to pump, not to thump.
There are however three serious objections to this view: 1) Nature does not in fact select
traits or organs for their functions in the same way that a watchmaker selects his gears
and springs for their functions. Nature cannot build organisms out of selected traits; it
selects organisms for their traits, and this results in the proliferation of those traits or the
production of new traits. 2) If we were to say that the function of a trait is to do what
caused its production or proliferation, then many traits would have the function of being
linked to useful traits. 3) Even if natural selection can explain the origin and proliferation
of the traits, it still might not be able to explain why the traits have functions. It is not
immediately evident why the causal past of a trait should determine its normative future.
Thus the etiological approach still has to legitimate the normativity of function ascriptions.
Causal history seems no more normatively binding than intention.
On the other hand, the causal-role or dispositional view of functions (where the
function of an item is its causal role in the performance of some specifi ed activity of
its containing system) is thought, even by its adherents, to have diffi culty in coping
with malfunction: If the function of an item in a larger system lies in its contribution
to the performance of some action of that system, then if it makes no contribution, it
has no function. A piece of steel that cannot regulate steam quantity is just not a governor,
one could say. It does not malfunction—it simply does not have the function. Similarly,
one could say that a piece of tissue that cannot regulate circadian rhythms does not
malfunction—it just has no function. However, let us examine this organ that looks
like a pineal gland, that is located where the pineal gland is normally found, that
arises embryologically just the way the pineal gland arises, but cannot regulate circadian
rhythms. We would not say that such an organ is not in fact a malfunctioning pineal gland
but rather not a pineal gland at all, or that it is a pineal gland without a function. Something
is wrong with this argument: we don’t normally identify an entity merely by its
function.
This last argument is a fairly standard objection to the causal-role approach, but I think
it is unfair—even if it is sometimes embraced by proponents of the approach. I believe
that the dispositional view, in spite of itself, has no more diffi culty in coping with malfunc-
tion than does the etiological view or the intentional view—because the normativity
involved in the ascription of malfunction is not necessarily introduced by (and is thus not
explainable by) intentionality or selection history.
But why should facts about intentions, about selection history, or about causal roles
explain, justify, or even motivate assertions about what ought to be the case? They can’t,
I presume; but this means that the normativity, if such there is, is coming from someplace
else. The basic problem is how to apply the fact-norm distinction to function descriptions
so as to be able to ascertain the extent to which norms and normativity are introduced in
seemingly purely descriptive propositions.