The Inherent Normativity of Functions in Biology and Technology 111
what they say?^13 The diffi culty seems least severe for “is supposed to” statements. Again,
only human beings suppose things, and what an item is literally supposed to do is therefore
what it is supposed to do by someone. In the case of artifacts, there are candidates avail-
able to do the supposing. It is clear that either the designer or manufacturer of an artifact
or the user or both have suppositions, in the sense of expectations, concerning its behavior.
It is also clear that the fact that a particular object is an artifact that has been designed for
a purpose furnishes strong epistemic justifi cation for certain expectations concerning its
behavior.
Indeed, as I see it, the normative statements associated with function that are phrased
in terms of “supposed to” or “ought to” or “should” are to be understood as expressing
that certain expectations are in order. They are, in a sense, to be taken metaphorically—
they are prescriptive in form but not in content. By saying that an item “is supposed to
show behavior B,” or equivalently, “ought to show behavior B,” the speaker expresses
that he or she is justifi ed in expecting behavior B to occur. Two different ways in which
such expectations may be justifi ed can be distinguished: the speaker can be rationally
justifi ed, more in particular epistemically justifi ed, or can be morally justifi ed in holding
these expectations. Epistemic justifi cation involves the standard considerations of empiri-
cal evidence and logical implication. Being morally justifi ed in expecting something to
occur can be expressed alternatively by saying that one has a right to it that it occurs, for
example, because this occurrence was part of the content of a promise.^14 Epistemic “ought
to” statements, as I call them, apply to all items with functions that have been considered
so far: biological items, social items, and technical artifacts. Moral “ought to” statements,
in contrast, can apply only to artifacts, because moral rights are grounded in the intentional
relations among people, not in anything “merely natural.”
That the epistemic “ought to” statements apply to all items with functions that have
been considered so far does not imply that such statements apply only to functional items.
In fact they extend to any object concerning the behavior of which we form expectations.
When we say that object o “ought to” show behavior B, we generally mean that, although
we cannot be certain that behavior B will occur, we are justifi ed in our expectation that it
will occur. Take the following example: “When I let go of this stone (holding it under
water), it ought to sink.” But some unexpected stream could prevent it from sinking, or it
could turn out to be pumice, which can fl oat on water. Depending on the context, it may
not even seem out of order to say: “When I let go of this stone (holding it up in the air),
it ought to fall.” Perhaps I vaguely suspect that some trick is going to be played on me.
Or it may be that I myself intend to demonstrate to a class of students that one should
never hold an empirical statement to be true with absolute confi dence.
Artifacts are therefore not special in that we have special expectations concerning their
behavior; only the grounds on which we form our expectations are different for artifacts
as compared to natural objects. Only in the case of artifacts can we have morally justifi ed
expectations, apart from the epistemically justifi ed expectations we generally have