Figure 3.2 The Mu ̈ller-Lyer illusionInformational encapsulation is a key feature of modules in Fodor’s
account of the modularity of mind. It means that modular systems are
blinkered, unable to make use of anything except their proprietary source
of information. For some purposes this is exactly what is needed. It helps
to explain how modules can process their input so quickly (speech percep-
tion is an impressive example of this). From the evolutionary perspective
there is clearly value in having cognitive systems which are not ‘dogmatic’,
which do not neglect environmental signals in cases where they conXict
with previously stored information. To put it crudely, an animal had better
be set up to react to movements or sounds or scents which may be
associated with the presence of a predator, and it had better react sharpish.
Sharpish counts for more than foolproof. Double-checking on whether it
is really so gets you eaten, and so is a sort of behaviour which would not get
repeated too often. In terms of survival costs and beneWts, reacting to false
alarms scores better than failing to be alarmed when you should be.
So modular input-systems have a sort of useful dumbness about them.
As illustrated by perceptual illusions, they can quite easily be fooled by
appearances, like a VenusXy-trap closing on a child’s pokingWnger
because it has been triggered to respond to the pressure of an alighting
insect. Predation is not, of course, the only force shaping the evolutionary
development of modular systems. Evolution has operated to recruit phe-
nomenological features for their survival value: hence our fondness for
sweet-tasting things, once a reliable guide to something it would beneWt
our ancestors to eat, and our revulsion at the taste of mouldiness, still an
indicator of something liable to make you ill.
There is much in Fodor’s account of modularity which we want to
accept. His position has evolutionary plausibility on its side, and helps to
explain what we know both about dissociations and the rigidity of normal
development. However, we do not believe that all input modules are fully
encapsulated. For example, it is known that visual imagination relies upon
the resources of (shares mechanisms with) vision. Visual imagination
recruits the top-down neural pathways in the visual system – which al-
ready exist in order to control visual search and to enhance object recog-
nition – in order to produce secondary input, in terms of quasi-visual
64 Modularity and nativism