Chomsky's linguistics and military funding: a non-issue

Noam Chomsky mural, Fairmount, Philadelphia, PA., June, 2011. Wikicommons/ Robert Moran. Some rights reserved.Chris
Knight’s thought-provoking
piece makes two independent assertions, though Knight sees them as crucially
linked. The first is that the Pentagon had high hopes of applying the results
of Chomsky’s theorizing for military purposes. The second is that Chomsky’s horror
at the idea of serving the military in any way led him to develop a theory of
language ‘so utterly abstract and
other-worldly – so completely removed from any practical application – that no
matter what insights he came up with, nothing could possibly be used to kill
anyone’.

My opinion is that Knight
is right on the mark in his first assertion and completely wrong in the second.
Contemporary accounts and subsequent history bear out the deep interest of the
military in language-related research. However, there is no evidence that
Chomsky’s research program has been driven by a desire to devise a theory that is
devoid of any potential military applications.

Military interest in
linguistics goes back at least as far as the beginning of the Second World War.
American linguists at the time were convinced, and were successful in
convincing others, that their methods of analysis were directly applicable to
the preparation of the language instruction manuals that the American forces
would need.

As early as 1942, it was
reported that the director of the program that forged a link between the field
of linguistics and the war effort was ‘called upon for advice on language
problems by practically every agency of government which has these problems: Office
of Strategic Services, Board of Economic Welfare, Department of Justice, as
well as the numerous departments of the Army, Navy, and Marine Corps’. 1.

Knight documents very well
the reasons for military interest in grammatical theory, as carried out at
MITRE and elsewhere in the 1960s. The fact that the linguists involved saw
themselves as developing their own personal research programs is quite
irrelevant, in my opinion.

By the 1980s less abstract
versions of generative grammar were being developed, which indeed seemed more
amenable to practical applications (including military ones) than Chomsky’s
research at the time. Much of this work has taken place at Stanford and
satellite organizations around that university. The most important of the
resultant projects, the LinGO system, has never received any military funding primarily
because the manager of that system, a devout Mennonite, refuses to accept it.

Today, the military is a
huge backer of work in natural language processing. However, most work in that
area is based on ‘deep learning’ (that is, neural networks trained on very
large corpora), making use of virtually no insights from linguistics. It might
be pointed out that the neural network approach to cognition is anathema to
Chomsky.

Abstract
approaches?

The idea that Chomsky has
proposed highly abstract approaches to grammar to fend off potential military
applications is implausible, to put it mildly. Knight is aware that ‘from the
beginning of his career at MIT’, Chomsky preferred a formal abstract approach
to language.

But here we are talking
about the mid 1950s, a decade before military interest in his work. So how
could his initial formal abstract treatment of language have been driven by a
desire to develop a theory that could not be used to kill anyone?

Knight goes on to write
that his students at MITRE ‘would tinker with his latest theory to make it more
realistic’ and that ‘Chomsky went along with this for a while, but then
resolved to retreat into pure abstraction’. I have no idea what Knight is
talking about here. I am unaware of any MITRE-employed students tinkering with
the theory to make it more ‘realistic’ or with Chomsky going along with their
tinkering.

In the late 1960s (after
the MITRE period), Chomsky did indeed propose a variant to his theory that was
somewhat less abstract in a certain way than his earlier variants (the
so-called ‘lexicalist hypothesis’), but he motivated it on purely
linguistic-internal grounds. The relationship between this theoretical variant
and the potential for military funding is obscure, to say the least. Chomsky’s
theories have evolved considerably over the past 60 years, but I see little
overall change in terms of their abstractness.

Needless
controversy

According to Knight,
Chomsky insists ‘that human language is purely individual, not a system of
social communication’. That is a gross oversimplification of Chomsky’s
position. It is true that Chomsky has always focused on language as a cognitive
faculty — he has as much right as anybody to follow his own interests. But he
has never denigrated other orientations to the study of language, as he
stressed in the following quote:

Internalist biolinguistic inquiry does not,
of course, question the legitimacy of other approaches to language, any more
than internalist inquiry into bee communication invalidates the study of how
the relevant internal organization of bees enters into their social structure.
The investigations do not conflict; they are mutually supportive. In the case
of humans, though not other organisms, the issues are subject to controversy,
often impassioned, and needless.

In fact, Chomsky has devoted
dozens, if not hundreds, of pages to exposing the manipulative use of language,
in particular by the leaders of the American political establishment and their
apologists, and to showing that the term ‘freedom of speech’ is often used as a
protective cover by those who would wish to deny it to others. In one book for example, he exposes the use of
terms like ‘aggression’, ‘doves’, ‘hawks’, ‘peace process’, and ‘terrorism’ in
American political discourse.

Surely, the fact that Chomsky
sees no inconsistency between such work and his grammatical theorizing (to the
point where both occur within the covers of the same volume) is prima facie
evidence that he believes that the study of formal grammars and their
properties complements, rather than challenges, the study of language in its
communicative setting.

Reference

1.M. Graves and J. Cowan quote from the journal Hispania
in 1942.