This is the first of a projected two volume set of essays, the second of which is entitled Semantics, subjectivity and localism. Three of the nine essays in this volume are previously unpublished. Of the remaining six, four have been extensively revised; the two that have not been revised have epilogues which provide historical context and L's own assessment of their content. In addition, L's 1965 inaugural lecture at the University of Edinburgh is included as an appendix.
One of L's aims in putting this collection together is to substantiate the following contention which appears in the Preface:
[T]he expressions 'natural language' and 'universal grammar' are often employed nowadays loosely and uncritically (if not equivocally) in the case of the former and tendentiously in the case of the latter.' (xi)
L rejects the identification of linguistic theory with the study of universal grammar, defined as the innate linguistic endowment of human beings. L testifies that his own views of linguistic theory have not changed since he incorporated Chomsky's original theory of generative grammar into his conception of traditional grammar and 'a rather eclectic blend of European and American structuralism' (xii). He testifies that he did not become aware that he held a minority view within the field until the late 1960s or early 1970s 'when Chomskyan generative grammar had come to be much more definitely associated with cognitive psychology and a particular kind of universalism than it had been previously' (xii). He remains as committed as ever to generative grammar as it was originally presented; about its current psychological and universalist interpretation (which he terms generativism) he remains agnostic. L is, however, able to encompass generativism in his overall view of linguistic theory; it is a theory of psycholinguistics, a branch of what he calls macrolinguistics (see below). Classical generative grammar, on the other hand, is a theory of microlinguistics, the autonomous study of languages generally. He concludes:
[L]inguistic theory should not be constrained by generativist assumptions: it should be concerned with a broader range of languages (natural and non-natural) and it should cover a much wider range of data. (xii)
The term natural language and its constituents natural and language are painstakingly analyzed in Ch. 3 (27-45) and Ch. 4 (46-72), two of the previously unpublished essays. In his analysis of the term language, L distinguishes between the senses of langue (which L observes is itself ambiguous between 'language-system' and 'language pretheoretically identifiable as human language', for which L uses the term N-language) and langage ('language generally, including nonhuman language'), but also between langue (i.e., 'language-system') and both performance and competence. We return to these distinctions below. L's detailed analysis of natural is very much in the spirit of ordinary-language philosophy lexical analysis. He distinguishes among four senses of the term, which makes it possible for him to describe certain languages, such as the artificial languages of logicians, as natural in one sense, but unnatural in another. Next, he introduces the term non-natural as follows:
More precisely, I will say of any language-system which is (in whole or in part) the product of human construction that it is a non-natural language. (68)
L then concludes that the class of N-languages, as analyzed within theoretical linguistics (71), are non-natural languages, which are no different, ontologically, from the artificial languages of mathematicians, logicians and computer scientists. He bases this conclusion on the following premise:
[T]heoretical linguistics deals with N-languages as rule-governed systems ... [which] are, inevitably, the product of abstraction. ... And, being at least partly the product of human construction, they ... satisfy the ... definition of 'non-natural'. (71)
L uses the term abstractions also to refer to the language systems which are the products of abstraction, contrasting it with the term abstract objects which appears in the title of Katz (1981) so as:
to emphasize that the language-systems postulated by theoretical linguists are the products of the linguist's more or less conscious and deliberate process of abstraction, or idealization, and have no independent or prior real-world existence. (47)
L's ontology of language systems is like Ernest Nagel's ontology of mathematical systems (Nagel 1979): both consider the systems to be abstractions that are justified by the discipline of scientific inquiry, and both elude Katz' classification of ontologies into nominalism, conceptualism and (Platonic) realism.
A second aim of this book, carried out chiefly in Ch. 2 (12-26, also previously unpublished), is to defend the autonomy of linguistics narrowly construed as the synchronic description of N-languages (microlinguistics). L points out that centuries of descriptive work on N-languages carried out without reference to psychology, sociology or anthropology is proof of the possibility of autonomous linguistics (esse valet posse), and he provides three arguments that such autonomy is desirable. The most important of these is that the autonomous study of language-systems has revealed 'interesting distinctive properties that are unique to them as semiotic systems' (20). L notes that these properties have been uncovered primarily as a result of the work of generative grammarians who espouse the view that linguistics is a branch of cognitive psychology. However, the actual investigations have been carried out with little or no regard to psychological theory, so that they have been as autonomous in fact as previous (e.g. post-Saussurean structuralist) studies have been.
According to L, linguistics broadly conceived as the general study of language (in the sense of langage) is not autonomous, and breaks down into the several (and proliferating) branches of what he calls macrolinguistics, including psycholinguistics, sociolinguistics, ethnographic linguistics, neurolinguistics, phonetics, stylistics and the philosophy of language. Each of these disciplines requires its own idealization about language, and crucially brings to bear theories from another discipline outside of linguistics. Hence in addition to theoretical microlinguistics, there is also theoretical psycholinguistics, theoretical sociolinguistics, theoretical ethnolinguistics, etc. In particular, theoretical microlinguistics (the autonomous formal study of language systems) must be distinguished from theoretical psycholinguistics (which involves among other things the nonautonomous study of human knowledge of language systems, or linguistic competence). On L's view, the object of psycholinguistic theorizing differs from the object of purely linguistic theorizing; in other words, linguistic competence (viewed, say, as a grammar represented in the human mind) is different from a grammar of a language studied as a purely microlinguistic abstraction.
An example may help here. Suppose psycholinguistic research were to reveal that linguistic competence is best modeled as a finite-state transducer. That result would not show that N-languages are finite-state. The classical formal arguments that show that English, for example, is not a finite-state language, would still be valid. It would only show that the language that people acquire knowledge of in the normal process of language acquisition is not English. While L draws a distinction something akin to Chomsky's E-language and I-language (the former the product of microlinguistic investigation, the latter that of psycholinguistic investigation), L does not consider the former an epiphenomenon and the latter real. Both are abstractions, and have essentially the same ontological status. The 'real' language systems, whatever they are, are not amenable to direct scientific study. L believes that they are not fully determinate, though of course the abstractions based on them (including the abstractions needed for sociolinguistic theory) must be.
A third aim of this book is to provide a setting for two of L's early and still widely cited papers: 'Phonemic and non-phonemic phonology: some typological reflections', which first appeared in 1962; and 'Towards a "notional" theory of the "parts of speech"', which first appeared in 1966. As I noted above, L has provided extensive notes and epilogues to these two articles which will be of great use to historians of linguistics and others interested in either the London school of linguistics or the early history of generative grammar.
A recurring theme in these essays is the independence of speech and language. L specifically argues that spoken language is only one type of N-language, and that N-languages may have arisen in humans out of gestural, rather than vocal, systems of communication (Ch. 5: 73-95). Two papers on deixis complete the volume.
Katz, Jerrold J. (1981). Language and other abstract objects. Oxford: Blackwell.
Nagel, Ernest (1979). Teleology revisited and other essays in the philosophy and history of science. New York: Columbia University Press.