For one thing, it is quite possible for people to fail to appreciate the
interconnections between their beliefs. So you may acquire a belief which
conXicts with some of your other beliefs without being aware of the
conXict, and you may fail to grasp that something you are inclined to
believe is very strongly supported by something else you have believed all
along. Those relevant interconnections will aVect your beliefs, and the
strength of your beliefs,ifyou notice them; but there is no guarantee that
they will get noticed. We must not let the Network Argument fool us into
thinking that they must inevitably get noticed by concentrating on the case
of scientiWc practice, where there are always other people waiting to point
out what we have missed.
ScientiWc enquiries are mostly conducted by people who have nothing
else to think about, for most of the time. Professional scientists are
employed to work back and forth between theory and data, checking
against background information and evaluating alternative theories.
And of course these activities will be conducted by a great many people
simultaneously, often working in groups, normally with a great deal of
overt discussion and mutual criticism. It is plainly a fallacy to think that
the principles which are operative in such collaborative practices must
transpose directly into the cognitive processes of individuals, as Fodor
appears to do in his Network Argument. (See also Putnam, 1988, for a
surprising convergence in views with Fodor here.)
Another point is that the Network Argument takes as its focus modes
of belief-Wxation which are paradigmatically conscious. Scientists formu-
late their theories explicitly and consciously, and explicitly consider their
respective strengths and weaknesses, their relations with other theoretical
commitments, and with other competing theories. And scientists also
reXect consciously on the methodologies employed in their enquiries, and
modify and try to improve on these as they seeWt. But for all that the
Network Argument shows, there may be a plethora of encapsulated
belief-forming modules which operatenon-consciously. Even if Fodor is
right that conscious theoretical inference is radically unencapsulated, it
may be the case that there also exist implicit, non-conscious, inferential
systems which are modular, and fully encapsulated in their processing.
We shall return to this idea in chapter 5.
We do think that it is one of the great challenges facing cognitive science
to explain how science itself is possible– that is, to provide an account of the
various cognitive systems involved in scientiWc enquiry, and to describe
how they underpin the kinds of activities and inferences which we observe
in scientiWc communities. But we are conWdent that progress with this
question will only be made once it is accepted that central cognition
comprises a variety of modular systems. It may even turn out that there is a
72 Modularity and nativism