Chomskyan versus Formalist Linguistics

Abstract

By formalist approaches in the narrower sense I refer to Montague semantics, Generalized Phrase Structure Grammar (GPSG) and similar theories. Newmeyer (1986) treats Chomsky's Government and Binding Theory, GPSG, Lexical-Functional Grammar (LFG), and Relational Grammar as alternative theories within generative grammar. However, treating the opposition between GPSG and Chomskyan linguistics as internal to generative grammar (or to formalist approaches to linguistics) is misleading. As shown in Ten Hacken (1997), the discussion between representatives of Chomskyan linguistics, LFG, and GPSG shows so-called incommensurability effects. This means that the same data which constitute a convincing argument for one side of the discussion may be unconvincing or even utterly irrelevant for proponents of the other side. An example is the treatment of X-bar theory in Chomsky (1981) and in Pullum (1985). The contrast between Chomskyan linguistics and GPSG is of an even more fundamental nature, transcending the simple opposition between two competing research programmes. The foundations of Chomskyan linguistics depend on the explanatory possibility arising from the essential tension between logical learnability by a sufficiently powerful language faculty and descriptive potential covering the variety of languages as incarnated in native speakersí competence. By contrast, Gazdar et al. (1985) do away with any reference to psychology. For them linguistics does not and cannot aim at explanation. Ten Hacken (2000) shows that it is impossible in principle to construct a model of the research programme underlying GPSG and Montague semantics in a way parallel to the one for Chomskyan linguistics. Linguistics can only be scientific in this approach if it is taken to be a formal science on a par with formal logic, rather than an empirical science. The aim of a formal science is deriving theorems from a set of axioms rather than explaining empirical observations. In linguistics this means that languages are abstract objects and sentences axioms of a formal system.

From the proposed analysis it follows that we should at least assume a tripartition in the approaches to linguistics. Functionalist approaches aim at explanations of the form in terms of the function of language. Formalist approaches in the narrow sense study the form of language with the goal of deriving theorems from sentential axioms. Chomskyan linguistics shares with functionalist approaches the concern for explanation and with formalist approaches the emphasis on the form of language. Whether explanation vs. deriving theorems or emphasis on the function vs. on the form of language should be taken as the principal dividing line is not immediately obvious. Much harm has been done, however, by entirely ignoring the former opposition. Only in this way can it be explained that in many critical presentations of allegedly Chomskyan linguistics the basic assumptions of the narrowly formalist approach are attributed to Chomsky.

References

Gazdar, Gerald; Klein, Ewan; Pullum, Geoffrey & Sag, Ivan (1985), Generalized Phrase Structure Grammar, Oxford: Blackwell.

ten Hacken, Pius (1997), 'Progress and Incommensurability in Linguistics', Beiträge zur Geschichte der Sprachwissenschaft 7:287-310.

ten Hacken, Pius (2000), Reusability and Research Programmes in Computational Linguistics, Habilitationsschrift, Universität Basel (Switzerland).

Newmeyer, Frederick J. (1986), Linguistic Theory in America, second edition, New York: Academic Press.

Newmeyer, Frederick J. (1998), Language Form and Language Function, Cambridge (Mass.): MIT Press.

Pullum, Geoffrey K. (1985), 'Assuming Some Version of X-Bar Theory', in Eilfort, William H.; Kroeber, Paul D. & Peterson, Karen L. (eds.), Papers from the General Session at the Twenty-First Regional Meeting, Chicago: Chicago Linguistic Society, p. 323-353.