(such as a broken aeroplane wing, or a collapsed bridge) in an attempt to
Wnd out why it incurred damage. In order to cope with explanation,
simulation would probably have to resort to repeated iterations, trying
Wrst one ‘pretend’ input, and then another, until a match with the behav-
iour to be explained is achieved. This seems – unless it can somehow be
constrained or accelerated – a slow, laborious and uncertain process. (You
mightneverget a close enough match.) Yet interpretation and explanation
do not appear to be more diYcult for folk psychology to handle than
expectation and prediction.
It seems plain, then – if simulation is to work eVectively for purposes of
explanation – that simulators will have to have a body of knowledge about
the likely causes of action, and/or about the beliefs and desires which
subjects are likely to possess in various circumstances. Simulationists
generally appeal to a process of learning through development to explain
our possession of such knowledge. The young child is said to begin with
trial-and-error inputs to simulation, from which it gradually learns which
are more likely to be successful in a variety of circumstances. In this way
simulation is still held to be at the core of our mind-reading abilities. We
think that this developmental claim is implausible. For the evidence is that
young children begin successfully toexplainwhat someone with a false
belief has done at about the same age at which theyWrst become able to
makepredictionsfrom false-belief attributions, for example (Wellman,
1990) – there is no developmental lag, here, with explanation trailing
behind prediction, of the sort to which simulationism is committed.
Since it is plain that simulation will have to be enriched with theoretical
knowledge, we shall mostly focus on the claim that simulation is at thecore
of our mind-reading abilities, rather than on the more extreme claim that
mind-reading isexclusivelya matter of simulation. (In addition to the
point made above, consider someone predicting how another will react
when in a state of extreme fear – since it seems unlikely that the eVects of
extreme fear can be achieved by taking one’s cognitive system oV-line, this,
too, will require theoretical knowledge.)
2.2 Simulation and self-knowledgeOur main argument against simulation-theory is that no version of it can
give a satisfactory account ofself-knowledge of mental states. According
to theory-theory, self-knowledge will usually take the form oftheory-laden
recognition. But this does not mean that we work out what psychological
states we are in by applying the theory to ourselves in a process of
self-interpretation. (This does sometimes happen. In fact, there is evidence
that it can happen so quickly and eVortlessly that we are not consciously
84 Mind-reading