Defining the undefinable

A quarter of a century ago a leading American sociolinguist observed that the term "sociolinguistics" is one that "means many things to many people". That situation has not changed a great deal, to judge by the present volume. The field continues to flourish, in spite of, or perhaps because of, that diversity. However, as the same sociolinguist acidly remarked: "energetic activity and prolific publication need not warrant confidence in the scientific worth of what is done".

Sociolinguists, as Florian Coulmas notes in his introduction, still tend to be "very defensive" about their subject. One reason is certainly the suspicion that sociolinguistics - as the hybrid term almost seems to imply - is neither fish nor fowl: neither sociology nor linguistics. Or at least not "linguistics proper". By continuing to use the latter expression, sociolinguists themselves invite marginalisation, as their own studies of linguistic usage should by now have taught them.

If sociolinguistics is marginal, the wealth of what lies in these margins is amply illustrated by this vade-mecum. It has 28 chapters by a total of 31 contributors. Between them they cover the history of sociolinguistic research, the demography of language, language varieties, language change, sociophonology, dialects, sex- and age-related variation, relations between spoken and written language, language in the media, diglossia, code switching, language contact, pidgins and creoles, language death, multilingualism, language and identity, language and ethnicity, "global" sociolinguistics, language and culture, linguistic etiquette, language in education, language and the law, language planning and language reform.

As Coulmas notes, it is a very curious situation when all this can be regarded by other linguists as peripheral to the study of language, and even dismissed with some contempt. He cites one of the contemptuous, who observes condescendingly that "any social parameter whatsoever may be the locus of some linguistic difference" and then adds: "Unfortunately nothing of interest to linguistic theory follows from this." A more justifiable conclusion would be that "linguistic theory", if defined from this blinkered perspective, has nothing of interest to offer to those concerned with the study of language(s).

While rejecting the term "linguistics proper", Coulmas surprisingly endorses the no less questionable expression "sociolinguistics proper". This he contrasts with "sociology of language", while at the same time claiming that there is no sharp division between the two. In an alternative terminology he calls the former "micro-sociolinguistics" and the latter "macro-sociolinguistics". This is unhelpful and even misleading since it does not parallel the (now dated) distinction between "microlinguistics" and "macrolinguistics". However, sharp or blunt, the division is deemed important enough to structure the volume, with 13 chapters devoted to one ("sociolinguistics proper" or "micro-sociolinguistics") and only eight to the other. It is difficult to see any useful purpose such a division serves, particularly in a handbook of this kind.

There are two further sections: one, "Foundations", contains the two papers on the history of sociolinguistics and linguistic demography. The fourth section is devoted to "applied" sociolinguistics - another curious classification, since the matters it deals with do not appear to be concerned with the application of findings by sociolinguists. It comprises four chapters, two devoted to language in education, one to language and the law, and one to language planning and language reform.

Wide as the overall coverage of the volume is, there are striking omissions and imbalances. Although the bibliography lists works by Austin, Searle and Grice, there is no account of modern speech-act theory, which focuses on conditions of appropriateness that are manifestly social in nature. This is presumably because speech-act theory is counted as falling in some other domain of inquiry - philosophy of language, or pragmatics. But if this is the case, it implicitly calls in to question the central claim of sociolinguistics to deal with language in its social setting. For inquiries, replies, offers, acceptances, refusals, promises, etc, are acts which seem essential to that "communicative competence" which society relies on in its members.

Although publications by Sacks, Schegloff, Jefferson, Atkinson and Heritage are listed in the bibliography, there is no adequate account of work in the ethnomethodological framework. It is referred to obliquely in the chapters on "Spoken and written language" and "Varieties and variation". The index lists a single reference to "conversational analysis", but none to "turn-taking", "adjacency pairs", "repair" or any of the related terms now commonplace in the description of communicational exchanges between interlocutors.

The index, it has to be said, is a slipshod piece of work. It does not even give a complete list of the languages referred to in the main text, which one would have thought essential for any reference work in sociolinguistics.

Leaving that aside, the other problems mentioned above are to be laid not so much at the editor's door as at the door of sociolinguistics itself. As Coulmas puts it, the subject has produced "theories but no theory", and this is just where the sociolinguist feels at a disadvantage when confronting the lofty superiority of those who speak for orthodox "autonomous" linguistics.

Some sociolinguists have doubted whether any unified theoretical framework for the subject is possible. A more radical view would question whether, even if possible, it is desirable. It is interesting to note that one of the best books to appear recently on a major topic of sociolinguistic interest, Deborah Cameron's Verbal Hygiene, has an author who does not conceal her scepticism about sociolinguistic methodology and speaks of the subject being "stranded in an explanatory void".

This is contested by the author of the chapter on the history of sociolinguistic thought, R.B. Le Page, but not altogether convincingly. He seems not to see that there are three interrelated problems which are endemic in most contemporary sociolinguistic research. One is that of the hypostatisation of "data" and their statistical manipulation. Another is the failure of those who insist that language must be studied "in context" to provide any non-question-begging account of "context" (one more key term that does not appear in the index). The third is talking about "communication" in the absence of any serious theory of communication. But if these problems were recognised and tackled, then sociolinguistics would become a quite different enterprise from the one documented in this book.

Roy Harris is editor, Language and Communication.