562 Karen Neander
come into play. In that case we can say that, while the frog’s perceptual system
was adapted for producingRs in the presence of small, dark, moving things, it
was also adapted for producingRs in the presence of flies or frog food. So, it has
been argued, an appeal to the function of the frogs’ detection mechanism does not
give us determinate content forRs. Depending on how we describe its function,
we get different content ascriptions. Described one way, the content is small, dark,
moving things, and described another way it is flies or frog food or something
nutritious.
Again there are different ways of tackling this problem, and how one goes about
it depends in part on what one thinks the correct content is in this case, and in
similar cases. The main point to note, however, is that it is wrong to assume that,
if functions are indeterminate in this way, then content derived from teleological
theories must be indeterminate in this way too. This doesn’t follow, simply because
teleological theories can do more than mention functions. The content one derives
from a teleological theory depends on how functions are (or selection is) put to
use in the theory.
Millikan’s [1991] position on this is the one that is probably best known. She
claims that the content is frog food. Her teleosemantic theory is quite complex,
but the gist of it, regarding this case, is that the content of the frog’sRsisfrog
food because it wasR’s mapping on to frog food, as such, rather thanR’s mapping
on to small, dark and moving things, as such, that explains why there was selection
for the use ofRs. Talk of explanation can probably be eliminated here, because
what is at issue is what was essential for a contribution to the proliferation of
the frog’s genes. Millikan seems to be thinking something like this. Had the frog
swallowed something that was frog food, the fitness value of itsR-induced snapping
would have been the same, regardless of whether it was small, dark and moving.
Whereas had the frog swallowed something that was small, dark and moving, the
fitness value of it’sR-induced snapping would not have been the same, regardless
of whether it was frog food.
Others argue that the correct content is the sign stimuli: small, dark, mov-
ing things. Neander [forthcoming] argues that this content is better suited to
mainstream (information processing) neuroethological explanations of the frog’s
perceptual capacities, and Paul Pietrosky [1992] argues for analogous content in
an analogous case for more intuitive reasons. Some of Fred Dretske’s writings
also support this result (e.g., see [Dretske, 1994]) although to see this we need to
note that frog food did not strictly co-vary with small, dark, moving things in the
environment in which frogs evolved: the flies sometimes stood still; bits of bark or
small shadows sometimes moved. In these places Dretske speaks of the content of
RsaswhatRs were recruited for indicating (or sometimes, he says, “maximally
indicating”) whereRs indicateCsiffifthereisanRthen there is aCin the
relevant environment. (This manner of addressing the problem speaks only to the
first part of it, not to the second.)
The simple teleological theory with which we began also gives the sign stim-
uli as the content, with only minor tweaking. Remember that it says thatR-