aware that this is what we have done; see Gazzaniga, 1994, for some
remarkable examples of this in commissurotomised, ‘split-brain’, patients;
and see chapter 8 below for some more discussion.)
On the theory-theory view, our concept of a certain kind of psychologi-
cal state (belief,desire,hope,fear,pain, or whatever) is a concept of a state
occupying a certain sort of causal role. But we can alsorecognisethe
occurrence of that sort of state in ourselves, recognising itasthe state
occupying that causal role. This kind of recognitional capacity is quite
commonly found with theoretical concepts. For example, a diagnostician
may recognise a cancer in the blur of an X-ray photograph; or when
looking at the moon through a telescope we, like Galileo before us, may
recognisecraters. The possibility of the recognitional application of a
theoretical concept is part of what philosophers of science have had in
mind in talking of thetheory-ladenness of observation.
As noted above, versions of simulationism diVer over the role they
assign to the recognition of mental states by means of introspection.
According to Goldman and Harris, attributions of mental states to other
people are grounded inWrst-personacquaintancewith our own mental
states. In order to predict what you will do in some situation, Isuppose
myself to be in that situation, making adjustments in my belief and desire
set of which I am introspectively aware. I then let my practical reasoning
system run ‘oV-line’, concluding with a pretend intention of which I am
also aware. I then attribute the corresponding action to the other person.
But when I am aware of these states in myself (pretend beliefs, desires,
and intentions)what am I aware of them as? The simulationist cannot reply
‘as states occupying a certain causal role’, or this will become at best a
version of simulation–theory mix (simulation used to enrich theory) –
which would be to accept our own position, in fact (see section 3 below).
So, presumably, we are supposed to be aware of them as states with a
certain kind offeel, or introspectible phenomenology (as Goldman, 1993,
has suggested).
We maintain that assigning this sort of role to phenomenological intro-
spection is an alarmingly retrograde step in relation to the progress made
in twentieth-century philosophy of mind. One thing which most philos-
ophers of mind are agreed upon by now is that the idea that we learn what
a mental state kind is from acquaintance in our own case is a complete
dead-end. Quite apart from other objections, such an account imposes an
overwhelming burden on learning: we must learnwhichfeelings to induce
in ourselves when simulating another; and we must learn which feelings to
correlate with action-descriptions. Compare positivist accounts in the
theory of perception, which say that we start from awareness of unstruc-
tured sense-data, and then have to learn how to construct on that basis
Problems for simulationism 85