drome, who in other respects have more severe cognitive impairments.
Autistic children’s incomprehension of false belief is strikingly exposed
by their inability to use deception. Anecdotal evidence concerning decep-
tion appears to conWrm the normal developmental proWle; and this placing
of the ability to deceive within the child’s mind-reading capacity has also
been reinforced by tests such as ‘the fairy and the stickers’ (Peskin, 1992).
A fairy-character (a puppet in a white dress) asks the child which is her (the
child’s) favourite sticker, and then takes the sticker that the child indicates.
While four year olds quickly learn that the way to keep the sticker they
want is to say that a diVerent one is their favourite, three year olds seem
unable to master this simple trick and keep on truthfully indicatingand
therefore losingtheir favoured stickers. In a similar experimental ar-
rangement, even older (teenage) autistic children lose out in the same way
as pre-four year olds. With Box B empty and Box A containing a goodie
they want, even after repeated trials they are unable to work out the
deceptive strategy of misdirecting to the empty box someone else who
wants to take the sweet and asks them where it is (Russellet al., 1991;
Sodian, 1991; Sodian and Frith, 1992, 1993).
Autism is a developmental disorder. Long before four years of age there
are diVerences between autistic and normally developing children. The
staging of autistic deWcits is summarised in the table below. (Proto-
declarativepointing is pointing intended to direct another’s attention to
something of interest – as contrasted withproto-imperativepointing, which
is pointing intended to obtain something.) Some of these diVerences have
been used in screening tests, at 18 months, and have proved very reliable in
identifying children subsequently diagnosed as autistic. The main features
on the checklist (CHAT, orChecklist for Autism in Toddlers) are proto-
declarative pointing, gaze monitoring, and pretend play (Baron-Cohenet
al., 1996; see also Baron-Cohen and Cross, 1992).
The autistic deWciencies are, of course, deWciencies in comparison with
abilities of normally developing children. In adopting the nativist view
that mind-reading is innate in humans we are not maintaining that mind-
reading, or the hypothesised module through which it operates, is itself
present from birth. Rather, we claim that there are genetic instructions
which have been selected for precisely because in normal human environ-
ments those instructions generate processing systems with mind-reading
abilities. Mind-reading is the long-range eVect of those instructions and
their biological function – it is the reason, or part of the reason, why they
are in the human genome. If we are right about this, it still leaves many
questions to be answered about what the shorter-range eVects of those
instructions are. Mapping of the normal developmental proWle indicates
that gaze-monitoring, and capacities for shared attention and pretend
96 Mind-reading