out to map onto brain areas to the extent that they do. For we were long
ago convinced of the falsity of type-identity theories by multiple-realisabil-
ity arguments; and these arguments suggest thatwherein the brain a given
function gets realised may be highly variable.
4 Fodorian modularity
The theoretical case for modularity has been presented with particular
e ́lan and energy by Fodor (1983, 1985a, 1989). According to Fodor,
modular cognitive systems are domain-speciWc, innately speciWed input
(and output) systems. They are mandatory in their operation, swift in their
processing, isolated from and inaccessible to the rest of cognition, as-
sociated with particular neural architectures, liable to speciWc and charac-
teristic patterns of breakdown, and they develop according to a paced
sequence of growth.
Roughly, one can say that the domain of a module is the range of
questions for which that processing system has been designed to supply
answers. Putting this in terms of a computational/representational theory
of cognition, Fodor suggests that modules are ‘highly specialised com-
putational mechanisms in the business of generating hypotheses about the
distal sources of proximal stimulations’ (1983, p.47). The important point
to note is that if we take seriously both the idea of a cognitive architecture
and the surprising evidence from dissociations, we will not in general be
able to lay downa prioriwhat the domains of modules are. It would
deWnitely be an oversimpliWcation to suppose that the traditional sensory
modes of sight, hearing, touch, taste and smell are each separate input
modules with single domains. The evidence of selective impairment sug-
gests instead that hearing, for example, subdivides into the hearing of
environmental sound, the perception of speech, and the hearing of music.
The domain of a module is really its cognitive function. It is an entirely
empirical matter how many modules there are and what their cognitive
functions may be. We can get some insight into the domains of modular
processing systems by asking top-down design questions about what pro-
cessing tasks need to be discharged in order to get from the information in
the proximal stimulus to the ultimate cognitive changes which we know to
occur. But we cannot lay down the boundaries of domains before outlining
the actual modular architecture of cognition.
The point which Fodor insists on, above all, is that modules areinfor-
mationally isolatedfrom the rest of the mental system. This isolation is a
two-way aVair, involving bothlimited accessfor the rest of the system and
encapsulationfrom it. There is limited access in so far as the processing
which goes on inside a module is not available to the rest of the mind; and
62 Modularity and nativism