4.1 In support of functional-role semanticsOne argument in support of functional-role semantics is an argument from
functionalism about mental statesin general. As we saw in chapter 1, the
majority of philosophers now think that the way to avoid dualism about
the mental, and to understand the relationship between mind and brain, is
by accepting that mental states are individuated by their causal role,
conceptualised at some level of abstraction from the physical mechanisms
in the brain which instantiate those roles. Indeed, we have accepted (in
chapter 2) that folk psychology embodies an implicittheoryof the causal
structure and functioning of the mind, in such a way that diVerent mental
state types can be individuated by their position within the theory. The
argument is then, that when we extend this approach to states likethe belief
that Pwe get functional-role semantics. But preciselywhichof a state’s
normal causes and eVects are to be used to individuate it?Allof them? And
only theactualcauses and eVects within a particular thinker? We shall
argue that the answer to both of these questions should be ‘No’.
To see the latter point (that it is not just actual causes and eVects which
count), notice that functionalism about the mind claims to individuate
mental states in terms of theirpotentialcausal interactions with bodily
stimuli, with other mental states, and with behaviour. Causal relations
with otheractualstates of the subject do not always play a deWning role.
For example, no functionalist would claim that my pain must be a distinct
kind of mental state from yours, merely because I happen to have a desire
to appear brave whereas you do not. On the contrary, functionalists will
insist that our states are the same, provided that theywouldhave the same
eVectsifall our other mental states were similar. In the same way, although
functionalists should accept thatsomecausal connections with other ac-
tual mental states (namely, the unmediated connections) play a deWning
role in individuating some types of mental state, they should deny that
mental state identity is transitive across chains of such connections. Surely
no functionalist would want to deny that blind people have desires, for
example (that is, action-determining states of the same type that sighted
people have), merely because of the diVerences in the remote causal
connections of those states! (Standing-state desires tend to be activated by
the belief that the thing desired is now available, and such beliefs are often
caused, in the normal case, by visual experience of the desired object.)
When the argument is extended tothe belief that P, then, it is obvious
that it will not be all of the causal connections between that state and the
agent’s other actual beliefs and desires which individuate it, but rather the
potential connections. It will not be theactualcausal liaisons of a belief
with other beliefs and desires which constitute its content, but rather the
Functional-role semantics 177