I
1
l
\
I
)
l
(
1
!
PARTISAN REVIEW
341
of a speaker of a natural language cannot be captured by any auto–
maton known to us. But while such a conclusion would be welcome
by the humanist, let us not forget that what was for the humanist
speculation within an imprecise terminology is in Chomsky a question
of rigorous demonstration. (To look at the formal properties of a gram–
mar and then to consider the structure of an automaton that can ac–
cept that language may be a new way of formulating hypotheses about
the mind, though the general pattern is as old as Plato or at least Kant.
The rationalist philosophers argued from their characterization of cer–
tain competences - e.g., logic or mathematics - to what the human
mind must be like. Lyons errs when he claims that Chomsky's earlier
writings - which include most of the work on mathematical linguistics
- are still within the empiricist tradition: the way in which the ques–
tions for mathematical linguistics are formulated is the rationalist way
of formulating inferences about the nature of the mind.) As a further
consequence of Chomsky's work we should note that attempts to sim–
ulate the interpretation of various fragments of a natural language like
English by computers or other devices will have no theoretical signif–
icance; Chomsky's specifications of creativity and infiniteness decisively
refute the theory - so popular with social scientists and with certain
philosophers like Wittgenstein - that linguistic competence is merely
the sum of simple skills involved in the interpretation of small fragments.
It
is easy to see that Chomsky's pattern of explanation can be
applied to other aspects of mental activity,
to
parts of psychology and
to other social sciences. As I suggested earlier, his way of looking at
linguistic competence involves by extension a new way of looking at
the social sciences: their task as now construed is not merely to cor–
relate observations and to attempt to devise mechanisms of control; it
is rather to give a careful formal analysis of various competences–
those whose possession involves the mastery of a set of rules - and to
argue from such analyses to the internal structure of the human mind.
Such explanations will be far more concerned with understanding than
with control or predictability; in this respect they will come closer to
contemporary work in the pure physical sciences than to social en–
gineering. Inventing controls over behavior is in itself neither good nor
bad; its value depends on how control is used, by whom and for what
purpose. But in any case, following Chomsky, such work should be
the job of the applied social sciences, and not that of theory within
the fields of social science.
Chomsky assumes that we have internalized various systems of
rules, and further that manipulations of them are crucial in under–
standing. These assumptions, and some well-known facts, lead to the claim
