What we say when

This is the third collection of Randolph Quirk’s occasional pieces to appear in ten years. Like its predecessors, it’s a good read: each essay is short and its argument easy to follow, each contains a good joke and an amusingly bad one, and each makes a definite point about language. The point is often one which could not very easily be made in a more serious context, not because the point is irresponsible or unscholarly but because it is conjectural. He puts his guesses over so lightly one hardly sees that one is being asked to think about something difficult and indefinite. For this reason the writing is sometimes rather coy: you feel a conclusion approaching, the exposition leads up to it, there is a quick hint, and then ... a joke, a nice example, an elegant phrase distract you back to the conversational mode.

That is the opportunity a book review or an amusing radio talk provides: one can be stimulating without being conclusive. The reviewer of the reviews can take the same opportunity, and suggest what the hidden point is, what the conclusion would have been if the author had paused to bore. And if he gets it wrong, well, it was only a review. I am especially tempted to put words in Quirk’s mouth when he is talking about the normative function of linguistics.

Amateur interest in language focuses on issues of correctness, on questions of what reasons there might be for speaking this way rather than that. Most professionals will have nothing to do with this. Linguistics is a science and is concerned with what is, rather than what ought to be. Quirk hints at a more flexible attitude, one that preserves scientific objectivity while holding out hope of satisfying the popular demand for advice about language.

This attitude is revealed almost comically in a previous book, in which a chapter by a collaborator (helper, not Quisling) explaining that all forms of English are equally OK is subtly tripped up by an earlier chapter by Quirk arguing that Standard English in some ways deserves its standardness. And in this collection there is an essay, ‘International Communication and the Concept of Nuclear English’, which can be read as an argument for the possibility of linguistic engineering, another, ‘Language and Nationhood’, which can be read as an attack on the ‘every dialect a language’ idea, and another, ‘Grammatical and Pragmatic Aspects of Countability’, which tries to find the true linguistic facts which make people say, and purists wince at, ‘there are less islanders than sheep.’

Quirk is cautious and indirect about all these things, though. He cushions these essays with others arguing, for example, that the fundamental character of the language has not changed so very much in 60 years of radio or 200 years of graffiti. Even here, however, when one reads a little closer one finds something quite unorthodox, more in the direction of a theory of the effective use of language. He argues that there is an illusion of greater change in the vocabulary of English than is actually the case. One might thus be comfortably reassured that the language cannot be going to the dogs too terribly fast. But the argument really is that the illusion of great change in vocabulary is produced by great changes of a different sort. The language employs a variety of different registers, of combinations of vocabulary and constructions used for particular communicative purposes. (Quirk does not speak of registers: I am foisting words as well as opinions on him.) These change: at different times and in different parts of the society different combinations of words and constructions are used to do different jobs. An essentially constant grammar underlies an evolution in what we say when.

This may seem quite obvious, and in a way it is. But Quirk is using it to make a point about present linguistics and a suggestion about possible future linguistics. I will return to the suggestion at the end of the review. The point about present-day linguistics is that one can tune one’s ear to just one of the language’s registers and as a result be deaf to much that is really part of the same language.

The deafness can take a number of forms. Selective tuning can make one think that our stock of words has changed in character rather than in the addition of a few items. Three light essays press, under the guise of admiration, the grammarian’s case against the lexicographer: just knowing what words are (or were) used, even in what linguistic context, will not tell one what the speakers of those words were up to in using them. For that one needs something more theoretical – at the very least a grammar. One quite basic example of this is the order in which a dictionary lists the definitions it gives for a word. Some (non-superficial) knowledge of a language is needed to decide whether ‘parking summons’ or ‘theatrical admission’ is the most common meaning of ‘ticket’. Deeper knowledge is required to tell whether, for example, ‘substance made from cellulose fibres’, whether or not it is a more common meaning for ‘paper’ than ‘newspaper’, is a more basic meaning, in terms of its position and function in the word’s range of uses. And extremely difficult questions about language must be faced if one wants to know whether a definition of ‘motorcycle’ ought to say that motorcycles have stronger frames than bicycles. Quirk refers to this last as the problem of distinguishing linguistic from encyclopedic meanings. He says, to suggest its difficulty, that ‘semanticists have long grappled with such problems.’ He might have added that philosophers of language have been quarrelling for thirty years about whether there is any such distinction to be made – whether the lexicographer’s (as opposed to the encyclopedist’s) meanings may not be a myth.

Languages, if we take them as the fixed structures underlying varying registers, are pretty broad things. Illusions which minimise their breadth are produced by selective tuning. Linguistic nationalism produces such illusions, like the lexicographers’ but in space rather than in time. Quirk urges us to see the differences between, say, British and American English, or the varieties of Scandinavian, as quite superficial, not enough to make different languages out of them. The delicate thing about the essay in which he provides much of the argument for this is that it was written to be delivered in Scotland and appeared in a book on Scottish nationhood. So it begins and ends with gestures towards the difference of Scottish from English English! Even this, though, is delivered with a quirk. He pleads for respect for every man’s mother tongue, never saying that tongues are languages, and combines the plea with another, for the provision of a ‘language of wider communication’ bearing no relation to national boundaries. Underneath what I have called ‘registers’ there are tongues, and beneath the tongues there are languages; these may come in various breadths, and underneath them all there may be a single human facility of Language.

In stressing the breadth of language, Quirk is setting up an invisible two-stage trick. Linguists nowadays are supposed to look for unobvious underlying grammatical structure, and are not supposed to worry much about what counts as a language. ‘Eighteen-year-old South Bristolian’ is a rather better object of study than ‘English’. Quirk trips this up by pointing out, in his nice harmless-sounding way, that to look for underlying structure is to look for something likely to be common to variants of a language (over both time and space). So one part of the orthodoxy (‘languages are the regularities behind the utterances of particular people’) is at odds with another part (‘look for deep structure’). This can be phrased so that it is generally acceptable. But if we don’t neutralise it we set things up for Quirk’s second twist: if grammar is not tied to a particular version of a language, then it ought to allow one to compare versions, to say something useful about their relative merits for various purposes. This is in fact the claim of the essay on ‘Nuclear English’ and, in a different way, of an essay, ‘Public Words’, on popular semi-academic writers catering to the public’s desire to be told what to say.

We can most easily see what Quirk is driving at here in terms of the suggestion about the development of linguistics that I referred to earlier. At various places in the book he makes the point that the time is ripe for the development of an account of the communicative use of language, which would make use of the grammatical analyses developed in the past generation, to say with some rigour which communicative intentions in which situations may be realised as which utterances. He owes his realisation of this possibility, he says, in part to the writings of the philosopher J.L. Austin. He does not mention the many more recent writers in linguistics (Lakoff and Karttunen in America, Gazadar and Wilson in Britain, for example) who have developed very similar ideas, sometimes under the label ‘pragmatics’. I take Quirk to be claiming something that none of these writers have suggested, though: that if we had such a theory to explain what people actually say in various situations, we would then rely less on grammatical theory to do this, and, as a result, we could relate grammar more closely to languages and language more widely construed. So one would not take a person to be speaking in a language as described by a grammar and a pragmatic theory: rather, one would take someone to be speaking in a register (on this occasion) of the tongue (which he habitually speaks) which is part of a language, as described by a grammar. Quirk does not say any of this, I must stress, but it draws together a lot of what he does say.

Whatever the nature of a theory of language-in-use according to Quirk, the last two essays in the book must be taken as contributions to it, as examples of the results it might offer. The last essay is an analysis of the ways in which poems, stories and conversations, begin, how the first few words set up the context for what follows. The factors which make a suitable literary opening are, according to Quirk, the same as those which make suitable conversational beginnings, whose grammatical features can be described in some detail. The point is to show how the presence of various grammatical constructions opens up expressive possibilities, which may or may not be taken advantage of on any occasion or conventionally incorporated into a mode of speech.

This essay is preceded by the essay on ‘Countability’. Here his main aim is to show that the simple dichotomy of nouns into count-nouns and mass terms (‘count/noncount’, he says, a little confusingly, since his point is that not all that is not count is noncount) will not do as a description of the way in which any version of the language works. Many nouns are neither like ‘cow’ (‘some cow’, ‘fewer cows’) nor like ‘milk’ (‘some milk’, ‘less milk’): examples are ‘family’. ‘earnings’, ‘data’, ‘government’. He is completely convincing on this point, and goes some way to showing what the real distinctions, which we may misconstrue as a dichotomy, are. It follows that some non-pedantic speech may blur the distinction in ways that better reflect the basic character of the language. And so, as in the passages in the ‘Nuclear English’ essay which praised some local features of American English as showing the way language might adapt to be an easily learned international medium, Quirk turns a defence of the prescriptive power of grammar into a subversion of standard English. The language that is prescribed may not be what the purist clings to, but something clearer and richer.