theories, for it is re-applied countless times, with endless individual vari-
ation in each new generation.
(3) Churchland’s (1979) claim that folk psychology stands in ‘splendid
isolation’ and is not reducible to any scientiWc theory deserves careful
consideration. It is substantially true, although the isolation is not so
marked as it used to be. But does that count against folk psychology? We
do not think so. As far as the isolation is concerned, there is a problem
which needs to be solved – the problem of explaining how the sort of
intrinsic contentwhich intentional states have can be realised in naturally
occurring systems. We will be grappling with this problem in chapter 7. It
is a diYcult problem, but there is no reason to abandon it as insoluble.
As to reducibility: why would anyone want to have it? We suppose that
reduction requires the sort of type-identity which holds, for example,
between the temperature of a volume of gas and the mean kinetic energy of
its molecules, thus allowing an observational law at the macro-level (such
as Boyle’s law) to be derived from those at a lower level (in this case,
statistical mechanics). But it should be realised that this sort of reduction is
by no means the standard case in science. We can only obtain that sort of
reduction where type-classiWcations at one level map tolerably smoothly
onto theoretically signiWcant type-classiWcations at a lower, micro-level. It
just so happens that there is only one physical realisation of the diVerence
between gases which diVer in temperature, namely, a diVerence in molecu-
lar motion. But even in the case of temperature this neat micro-reduction
only applies to a speciWc range of cases – not to the temperatures of solids
and plasmas, in which diVerences in temperature are realised in diVerent
ways (see Blackburn, 1991). Usually we do notWnd a neat reduction, and it
is not any threat to the unity of science or the ultimate sovereignty of
physics that we do not. (We return to this important point in section 4 of
chapter 7, in the context of our discussion of so-called ‘naturalised seman-
tics’.)
As Fodor (1974) points out, the normal situation in the special sciences
is that weWnd autonomous law-like relations which holdceteris paribus,
and which are not simply reducible to (that is,deducible from) laws at more
fundamental levels. The reason why this is so is that the special sciences
deal with things grouped together as kinds which from other perspectives –
and in particular in terms of their microstructural realisation – are hetero-
geneous. It is just not necessary to the viability of a theory that the kinds
which it theorises about should correspond to classiWcations at a more
general theoretical level. Thus there probably is not anything of neur-
ophysiological signiWcance which unites all and only thosewho love ba-
nanas, or all and only thosewho cannot stand Wagner’s music, or all and
only thosewho have just realised that their current account is overdrawn. But
Realism and eliminativism 43