556 Karen Neander
4 BRIDGING THE DIVIDE: NEUROSCIENCEThis view of neuroscience needs to be discarded. Neuroscience is an intentional
science. It counts as an “intentional science” if by this we mean a science that
studies representational states as representational states (and do not, say, equate
intentional states with folk psychological states).
It’s probably right to say that neuroscience began with a bottom-up approach,
first focusing on molecular and cellular phenomena, and only later considering the
representational functions of neural components. However, this focus was never
exclusive, and as neuroscience matures and becomes more thoroughly integrated
with the rest of cognitive science, it is replete with talk of the representational
functions of neural components.^3 As they are practiced today, there is no sharp
distinction to be drawn between neuroscience and cognitive science more generally.
It is a difference in emphasis only. Neuroscience seeks to describe and explain our
neurological processes and understand how they relate to our cognitive capacities,
whereas cognitive science seeks to describe and explain our cognitive capacities
and how they relate to our neurological processes.
To understand why neuroscience is representational we need to consider the
nature of physiology in general and neurophysiology in particular. In brief, phys-
iology is the study of the functional organization of living systems. So we cannot
complete our study of brain physiology until we have described its functional orga-
nization. Since neural components have representational functions, it follows that
we cannot complete neurophysiology until we have described the representational
functions or the information-processing functions of neural components.
But how does this amount to motivation for teleosemantics? The answer is
simple. Central to physiology in general and to neurophysiology in particular is a
notion of function that is normative. If the SE analysis of it is the best analysis
of it, neurosphysiology is already steeped in talk of SE-functions, and indeed of
information-processing SE-functions. This is not the place for a general defense
of an SE analysis of normative functions. Such defenses have been given else-
where (see [Neander, 1991], and [Neander, 2005] for a response to [Lewens, 2004]).
However, I need to say something about the role of SE-functions in physiology in
general and in neurophysiology in particular, since a proper understanding of this
role is needed to see why we should take teleosemantics seriously.
It has become a near commonplace in philosophy of biology to think that there
is a division of explanatory labor between SE-functions, on the one hand, and
CR-functions, or causal-role functions, on the other. SE-functions, as I said in
the opening section, are what traits of a type were selected for. CR- functions,
in contrast, are dispositions, or causal roles that items are disposed to play. SE-
functions are normative in the sense that an item can have an SE-function that
(^3) Now, by the way, the Churchlands would agree with this. Paul Churchand [1995] offers a
similarity-based vector analysis of how neural nets represent, and Pat Churchland [2002, 273–
318] explains why, in her view, neuroscience needs a notion of representation and why, at least
so far, no viable alternative is in sight.