underlying behaviour; and it is quite another thing to say that those
commitments are correct. Note that, as we are understanding these posi-
tions, realism of fact entails realism of intention – the folk could not be
correctin their characterisations of the causal processes underlying behav-
iour unless they alsobelieved inthe existence of such processes. But one
could endorse realism of intention about folk psychology while rejecting
realism of fact, hence becomingeliminativistabout folk-psychological
categories.
Over the next two sections of this chapter we shall be concerned to argue
the case in favour of folk-psychological realismof intention. Then in
section 4 we shall turn to the issue of eliminativism, taking a preliminary
look at the strength of the case in support of folk-psychological realismof
fact– ‘preliminary’, because the question of the likely truth of our folk-
psychological commitments will turn ultimately on the prospects for suc-
cess of an intentionalist scientiWc psychology, and on the extent to which
that psychology will endorse and validate the commitments of the folk.
(These will be topics to which we shall return throughout this book.)
Finally, in section 5, we return to the issue of realism of intention once
again, considering the extent to which folk and scientiWc psychologies are
engaged in the samekindof enterprise.
We should also declare that we adopt a realist (of intention) position
about scientiWc theorising in general, on the grounds (a) that realism is
the natural ontological attitude, (b) that it is methodologically more
progressive because it sharpens the competitive conXict between theories,
and (c) because it makes better sense of the role of experimental interven-
tion in science (Hacking, 1983). But obviously we cannot just borrow
arguments for scientiWc realism and apply them to folk psychology. May-
be as scientiWc investigators we ought to be as realist as we can be. But we
cannot assume thatfolkpsychology will conform to desirable scientiWc
methodology.
In fact, it is important to appreciate that there is both anormativeand a
descriptiveissue about realism. In adopting scientiWc realism, we stress that
realistically interpretable scientiWc theories are possible, and where pos-
sible are methodologically preferable to other theories, particularly in
relation to the progress of scientiWc knowledge. But that of course allows
that other types of theory may also exist, and what those theories are like is
a descriptive issue. Rather than telling us about underlying causal mechan-
isms or microstructural constitution (occupations favoured by the realist),
such theories may just be devices which enable us to work something out or
to solve a speciWc kind of problem.
For example, Ptolemy’s astronomical theory, which attempted to
predict and retrodict the movements of the heavens from a complex
Realisms and anti-realisms 25