developmental disorders, such as Williams’ syndrome. As we noted in
chapter 3, children with Williams’ syndrome have poor practical intel-
ligence, poor visuo-spatial skills, and poor theorising abilities. But they
have good (in some ways precocious) linguistic abilities and are close to the
normal developmental proWle in mind-reading (KarmiloV-Smithet al.,
1995). The dissociation between theorising and mind-reading in Williams
syndrome seems to support the hypothesis of a mind-reading module – and
acquisition by theorising is surely ruled out in their case. Williams syn-
drome also poses a problem for simulationism. Simulationists say that
autism is basically a deWcit inimagining,orinsupposing, which they say is
crucially involved in pretend-play, in practical reasoning and problem-
solving, and in mind-reading. But Williams syndrome children are normal
at mind-reading, but have severe diYculties with practical reasoning and
problem-solving. This strongly suggests that mind-reading and imagining/
supposing dissociate.
5 Accounting for autistic impairments
Thus far in this chapter we have done two things. First, we have argued for
the superiority of a modularist/nativist, maturational, form of theory-
theory as an account of mind-reading development, as against the ac-
quisition-by-theorising view. This has been on the grounds: (a) of develop-
mental rigidity, (b) of the diYculties the child-as-scientist view has in
explaining theory-change, and (c) that weWnd exactly the kind of geneti-
cally related deWcit – namely, autism – which we would expect if the
modularist account were correct. Then, second, we have argued for the
superiority of theory-theory over simulationism as an account of the core,
or basis, of our mind-reading abilities. This was on the grounds that
simulationism: (a) has problems in explaining the parity of explanation
with prediction, (b) that it can give no adequate account of self-knowledge,
and (c) that it has problems in accounting for various forms of mutual
cognition. Taken together, these arguments amount to a powerful case in
support of nativistic theory-theory, it seems to us.
It has been claimed, however (notably by Currie, 1996), that the theory-
theory view of mind-reading cannot oVer such a good account of certain
autistic impairments as simulationism can – particularly the absence of
pretend play and deWcits in executive function. Here our strategy will be to
argue that theory-theory can oVer explanations of these impairments
which areat least as good asthose oVered by simulationism; leaving our
earlier arguments to determine an overall victory for theory-theory.
Accounting for autistic impairments 99