1991): the states which are nomically related may be role states, rather than
realiser states. That is to say, the states which are important for the salient
causal laws may be states which are distinguished from other states in
terms of their functional characterisation, instead of their physico-
chemical microstructure. Consider, for example,massandtemperature.
Mass and temperature are nomically real all right, even though it is clear
enough that both objects with the same mass and things at the same
temperaturedo not have to share a common microstructure.
4 Realism and eliminativism
Realism (of intention) about folk psychology involves greater commit-
ments and thereby also greater risks of getting things wrong than inter-
pretationalism or instrumentalism. The methodological advantage of
scientiWc realism is particularly connected with theincompatibilityof the-
ories which posit diVerent underlying structures, causal processes, and
generating mechanisms. By contrast, instrumentalism is more tolerant of
the co-existence of diVerent ways of problem-solving. Maybe my cal-
culator enables me to work out the results more quickly than you can on
your abacus. But that does not oblige you to abandon the abacus, if it
serves your purposes well enough; and there may well be some circumstan-
ces (such as battery failure) when I will be quite happy to borrow a
diVerent device. But if, as we have argued, a strong form of realism about
folk psychology is correct, and if – as we are going to argue in chapter 4 –
folk psychology constitutes a sort oftheory, then we have to acknowledge
the possibility that folk psychology may be amistakentheory.
Some people have arguedeitherthat folk psychology can already be seen
to be an inadequate theory (Churchland, 1979, 1981)orthat it is a good bet
that folk psychology will be shown to be wrong by future developments in
cognitive science and/or neuroscience (Stich, 1983, 1988; Ramseyet al.,
1990). Both of these claims are commonly referred to as ‘eliminativism’,
but it is important to distinguish Churchland’s ‘elimination now’ from
Stich’s less dogmatic ‘elimination in prospect’. Either form of elimin-
ativism is disturbing because it suggests that folk psychology isradically
mistaken. What the eliminativists mean by this is not that folk psychology
often makes us get things badly wrong, wildly misinterpreting each other,
expecting people to do one thing andWnding they do something quite
other, and so on. It is not plausible that folk psychology is radically
erroneous as a guide in practical aVairs. Instead what is meant is that folk
psychology is wrong about the sorts of internal states which lie behind our
behaviour. SpeciWcally, it is wrong in supposing that we have thoughts,
wants, desires, beliefs, hopes, fears, and other intentional states. Note that
40 Folk-psychological commitments