5.1 Pretend playDuring their second year normal children start to engage in play which
involves some form of pretence. For example, a doll or teddy is treatedas if
it were a living companion. The increasing complexity of pretend-scenarios
and the way in which children can become engrossed in their make-believe
situations is a striking feature of early childhood, and a regular source of
wonder to parents. Lack of interest in pretend play is an autistic impair-
ment which may seem particularly problematic for the theory-theorist, and
yet readily explicable for the simulationist. Normally developing children
have usually started to engage in pretend play by about 18 months. As
noted above, early absence of pretend play – along with absence of
proto-declarative pointing and of gaze monitoring – is used as part of a
checklist for autism in toddlers which has proved highly accurate in
identifying children subsequently diagnosed as autistic. Autistic children
can be inducedto engage in pretend play if instructed to do so by others
(Lewis and Boucher, 1988; Jarrold,et al., 1994b), but hardly ever seem to
do so spontaneously.
SuperWcially, it seems as if this supports simulationism: not going in for
pretend play seems more like a deWcit in simulation than something
explicable by a lack of knowledge about minds. But we should not be too
quick to assume that the deWcit in pretend play is better explained by
simulationism. To do so would be to fall victim to a sort of pun, or play on
words. In so far as pretending actually is describable as engaging in
simulation (so that ‘pretending to be an airline pilot’ is just another way of
saying the same thing as ‘simulating an airline pilot’), then there is a sense
in which it is clearly true that children who do not go in for pretend play
have ‘a deWcit in simulation’. But then that is just a redescription of the
phenomenon. It is not an explanation of why autistic children should have
this deWcit in terms of the lack of some underlying capacity, on which
pretend-play is causally dependent in normal children. If we think in terms
of a simulationist mechanism at the sub-personal level such as the oV-line
running of psychological systems for practical reasoning and decision-
making (seeWgure 4.1 above), then it is no longer so apparent that an
inability to dothatwould provide a good explanation for failure to engage
in pretend-play.
The question remains: how can theory-theory explain the autistic deWcit
in pretend play? One of us (Carruthers, 1996b) has suggested that autistic
children lack any motivation to engage in pretend play because, being
unable to detect and represent their own mental states (including the
mental state of pretending), they cannot appreciate theXuctuating changes
in mental state which provide the reward for pretend play. But this
100 Mind-reading