96 Peter McLaughlin
If God’s will, the history of selection, and the contribution to a system’s performance
do not ground the normativity of function ascriptions, what does? I want to consider three
possible alternative sources of normativity, three aspects of function descriptions, where
it is possible that apparently factual statements about functions are actually normative. I
examine three possible places where normativity may have been implicitly presupposed
and thus have been introduced without argument: 1) means-end relations, 2) part-whole
relations, and 3) type-token relations.
Let me just assert dogmatically that propositions about all three of these relations are
per se normative—and see how far I get. Somewhat more circumspectly: let’s investigate
how far apparently factual propositions about these three kinds of relations are actually
normative—or at least involve normativity in some way, shape, or form that we did not
suspect before. I start with means and ends and differentiate the problem of parts and
wholes before turning to types and tokens.
6.1 Means and Ends
Any part of a material system that has some effect within the system can be viewed as a
means to that effect—if the effect is in turn viewed as an end. Any link in a causal chain
can be viewed as a means to the next link; any part of a complex system that contributes
to some performance of that system can be viewed as a means to performing that end.
When we ascribe functions to things, we view them as means to ends. Functions are in a
sense nothing more than effects considered from a means-ends perspective. In artifi cial
systems this is often viewed as relatively unproblematic: a can opener has the purpose or
function of opening cans; a pressure valve has the function of regulating pressure. The
functions of artifacts are also normally the effects intended by some agent, but the same
kind of means-ends relations can also be seen in nature. We can also view natural things
not merely in terms of cause and effect but also as means to ends in an attempt to under-
stand the workings of a complex system. We analyze the causal roles of particular things
in a system or a process. Things with functions are thus conceptualized not just as causes
of certain effects but also as means to certain ends. Viewing something as a cause makes
no particular presuppositions as to why the thing or event viewed as a cause is there, but
viewing something as a means to an end makes very defi nite presuppositions about why
it is there. Although causes are not necessarily there for the sake of their effects, means
are in fact there for the sake of their ends; they are subordinated to the ends; that’s what
we mean by calling them “means.” Sometimes we may consider talk about means and
ends simply to be a fi gure of speech that does not commit us to any such intentionalistic
consequences. If we say that the function of the valves in the veins is to constrain the
blood to fl ow only toward the heart, we may mean that this is what they do, that this is
their effect in the system of blood circulation. But in doing so, we are viewing them from