How Scientific is Chomsky's theory of linguistics?

by Carolyn Burke
Copyright 1993, Carolyn Lesley Burke. All rights reserved.


In this study, I have endeavoured to appraise and criticise Noam Chomsky's work in linguistics. My intention was to distinguish those parts of Chomsky's work that were non-scientific from those that were. I have concluded that, in large part, Chomsky's work is not comprised of scientific theories. Rather I claim that he has establishd a rationalist progressive research programme in linguistics. In support of this conclusion, I show that Chomsky's primary contribution to linguistic research has been his expositions of the history and philosophy of linguistics. Further, I show that this exposition is standardly presented to defend Chomsky's tacit claim: there is (and has been) a positive (metaphysical) heuristic guiding successful linguistic research since at least the time of the influence of the Port-Royal grammar. This heuristic is better known as the innateness hypothesis---there is an essence of language and it is inborn.

In this study I have presented: the relevant history of the study of language especially with regard to the use of the innateness hypothesis; the relevant history of method in science; and, the basics of Chomsky's work in linguistics. Finally, to contrast with my claim that Chomsky's work is primarily that of a Lakatosian research programme, I examine the claim that Quine's neo-empiricist work on language makes him the leading current proponent of a behaviourist linguistic research programme.


Download entire document (LaTeX format in zip file)
LaTeX files
Download entire document (LaTeX format in PDF file -- coming soon)
PDF files


Introduction for Oral Examination

I started out in this study with a concern for a problem: How scientific is Chomsky's theory of linguistics? ; where, by `scientific,' I mean a method of study used to increase our understanding of the nature of reality.

Chomsky's work itself is in a sense scientific, in that, it is clear that he is the mentor of an entire research programme in lingusitics; although it is as clear that he does not publish many scientific reports himself. His work usually takes the form of a defense for this programme (or an attack on a competing programme, such as Quine's). In his defenses, he continuously explains that the innateness hypothesis accounts for a number of facts of language (such as the fact of its productivity; and the amazing fact that children acquire language quickly and effortlessly).

He then often explains how the use in linguistics of a Generative linguistic theory can account for the first of these facts--- namely through the use of transformation rules. These are rules that transform Deep Structure representations of a sentence into a Surface Structure representation of that sentence. This allows the combination of a finite number of rules and a finite number of lexical items to produce an infinite set of well formed strings, namely the sentences of a human language. Clearly, the solution to the problem of productivity has been revolutionary in linguistics.

This does not yet account for the second fact---namely, there is a paucity of language evidence in the surrounding speech community, so much so that Chomsky claims that the child could not generalize from these samples to become fully linguistically competent. However, if most of the knowledge that is required to formulate this competence is innate, the child will require only a minimal amount of data from the environment---that is, enough data to select one innately generated grammar from amongst several. The presence of innate knowledge of language, of a Universal Grammar, would explain how children acquire language so easily and quickly. Chomsky makes the claim that the Deep Structure, proposed for independent reasons, is innate---that is, it is the essence of language.

As a conjecture, the innateness hypothesis ought to be tested. But, Chomsky does not stress this. This raises a question: What is the role of the innateness hypothesis? I have claimed that its role is as the negative heuristic of a lingusitic research programme---that is, it is metaphysical. The manner in which Chomsky has used the notion of the Generative Transformational grammer has served to establish this notion as a positive heuristic. Together, these two heuristics have safe-guarded a series of ever more complex theories of language; theories which describe our innate knowledge of language. It seems to be the case that the innateness hypothesis will be tested only insofar as its use will result in either a degenerate research programme, one in which novel facts are no longer uncovered, or a progressive research programme, as has been the case to date for Chomskyan linguistics.

It is clear that Chomsky's programme has been, and is being today, carried out by a great many linguists worldwide. These researchers have generated many variants of the theory of Generative Grammar, only some of which accord with Chomsky's variants. These have formed a series of ever more complex theories about language, much like Lakatos has claimed makes up a progressive research programme of science. There is a surrounding belt of ad hoc hypotheses, and a large selection of anomalous facts waiting to be accounted for with the next theory.

However, clearly each of these new theories is constructed upon the presupposition that there exists an innate Universal Grammar, and each uses transformation rules, or replaces them with another device for generating infinite sentences from finite resources. Finally, it is clear that throughout Chomsky's work, from his 1957 publication of _Syntactic Structures_ to his present day publications, all have as _their_ essence, the common the assumption that there is an essence of language---the Universal Grammar.


Carolyn's Diary
[index]|[mail me]|[finale]