quantifying overthe entities they concern. We will not pursue this further
here.
Fodor has argued, however, that functional-role semantics must give
rise to a much stronger form of holism than this, and claims that scientiWc
psychology can have no use for such a strongly holistic notion of content
(1987, ch.3; Fodor and Lepore, 1992). In which case, since attempts to
naturalise content are ultimately driven by a desire to defend the reality of
the propositional attitudes by showing that they can be incorporated into
science, we had better not pursue that goal by adopting functional-role
semantics. Fodor’s claim is that functional-role semantics must make the
identity of any given belief-content dependent uponallof the beliefs with
which it is inferentially connected. In which case content will turn out to be
radically idiosyncratic, and rarely, if ever, will two people share beliefs
with the very same content. For there will always be enough diVerences in
their belief systems to ensure that there aresomediVerences in the inferen-
tial connections of the candidate belief. But if content is idiosyncratic, then
it must, it seems, be useless for purposes of scientiWc psychology. For what
such a psychology seeks are laws which will subsume all of the individuals
in a population.
There are various ways for adherents of functional-role semantics to
respond to the alleged problem of strong holism. One has already been
mentioned, which is that functional role is really functionalpotential– and
the same set of conditionals can be true of people who diVer in their actual
beliefs. There are also a number of further proposals to draw limits to
those of the inferential connections of any given belief which count to-
wards the identity of its content (see Carruthers, 1996c, ch.4; Devitt, 1996).
While the problem is diYcult, there are no reasons to think that it must be
insuperable, in our view. Here, however, we will concentrate on showing
that even if our folk-psychological notion of content proves to be strongly
holistic, and so idiosyncratic, this need not mean that scientiWc psychology
should abandon the notion of content altogether.
TheWrst point to be made, is that much of psychology could operate
perfectly well even with an idiosyncratic notion of content. For many
psychological generalisationsquantify overcontent, in a way which does
not require it to be true that any two thinkers ever entertain thesame
content. Thus the practical reasoning syllogism tells us that anyone who
desiresQ, and who believesif P then Q, and who believesP is within my
power, will try to bring it about that P,ceteris paribus. This generalisation
can remain true, and can retain its predictive and explanatory power, even
if no two thinkers ever instantiate thesameinstance of it (with particular
contents substituted forPand forQ). But clearly this point by itself cannot
draw all of the teeth from the charge of idiosyncrasy. For it is also true that
182 Content naturalised