but that theory-of-mind mustalwaysbe involved, at least in a controlling
or monitoring role.
To see theWrst point, consider someone wondering whether they can get
a piece of furniture through a doorway or up a narrow staircase. Surely
that will engage simulation. They are going to think, ‘Suppose we have it
this way round and then turn it like that’ – and then imagine what would
happen either from the perspective of an agent or of a spectator. But there
are countless other cases of counterfactual reasoning in which processing is
not going to be so closely tied to particular action-sequences. Very often
the question, ‘What would happen (or would have happened)if...?’ is
going to be answered by combining the supposition in the if-clause with a
rich set of background beliefs, some of which will have to be modiWed in
order to secure coherence with the supposition being entertained.
To see the second point (that theory-of-mind must always be involved in
complex counterfactual thinking), notice that as you think through, coun-
terfactually, some complex problem, you will need to keep track of your
own inferential moves – you will need to represent that atthisstage you
were supposingthis, whereas atthatstage you were supposingthat, and so
on. This requires a capacity to represent your own sequence of thoughts,
thus implicating the mind-reading capacity. If autism is a deWcit in mind-
reading, as modularist theory-theorists suppose, then problems of execu-
tive function are precisely what we should expect.
6 Conclusion
We conclude that a modularist form of theory-theory stands as the most
plausible account of human mind-reading abilities. While the spur pro-
vided by the contrast with the rival simulationist theory has been invalu-
able, we think it is clear that theory-theory is the more progressive and
promising research programme – and that within this programme, it is
the modularist/nativist version which should be preferred. There is, how-
ever, a role for simulation, in working out the consequences of, and
relationships amongst, contents. In assessing the implications of develop-
mental evidence, it is most important to remember that knowledge of
minds is not only knowledge ofotherminds. Self-knowledge should not
be taken for granted. This is an issue to which we shall return at length in
chapter 9.
selected references
Three useful collections on the topics of this chapter are: Davies and Stone, 1995a
(This reprints the papers from aMind and Languagespecial double-issue, plus
Conclusion 103