Comments Welcome
George W. Grace
University of Hawaii

Ethnolinguistic Notes

Series 4, Number 2

LINGUISTIC CHANGE

1. THE "ÉTAT DE LANGUE"

Quite early in my work on Austronesian reconstruction, I began to feel that at least some Austronesian languages must have developed in unexpected ways. I imagined that the evidence of these languages--if we could ever determine what had happened to them--might significantly change some of the fundamental assumptions of historical linguistics.

I said that they must have developed "in unexpected ways", but, in fact, what were the expected ways? As I began trying to pin down what the expectations were that had seemingly been disappointed and what assumptions might have led to them, I soon found myself asking what the fundamental assumptions of historical linguistics actually were. What were the assumptions that had traditionally been made by historical linguists about how languages change and differentiate--the assumptions upon which historical reconstructions rested? As I thought about these questions I came to realize that in my own work I had been taking a lot of things for granted, but I'd taken them so much for granted that I had no idea what they were. So began a continuing effort to figure out where all of us were coming from. Here I'll attempt a kind of progress report--in the form of a hypothetical dialog with myself--on (one thread followed by) my groping efforts to figure it out.

QUESTION: What is a linguistic change? That is, what kind of observation demonstrates that a linguistic change has occurred?

ANSWER: The observation that some language is in a state (état de langue) that is different from its previous state.

QUESTION: Yes, but what kind of observation is it that establishes what its state at any given point in time is in the first place? (The reason that this question is necessary is that we, of course, will never have descriptions of the state of every language at every point in time. Historical linguists therefore need to understand as much as possible about the rules of evidence for determining a language state so that they can evaluate the evidence concerning hypothesized linguistic changes [that is, changes of language state].)

ANSWER: I should begin by citing the fundamental principle that modern linguistics is a descriptive rather than a prescriptive discipline. Accordingly, our synchronic account of a language (that is, our description of a language state) is arrived at by observing what the speakers of the language actually say--i.e., by observing the utterances that are actually produced by them.

QUESTION: Yes, but when I think of the linguistic descriptions that I've seen, they don't seem like descriptions of utterances. Why is that?

ANSWER: I need to be a little more specific. Linguistics has traditionally been concerned, not with all of the properties of utterances, but only with their linguistic properties: specifically, the linguistic expressions (1) that are employed in them. Accordingly, linguistic description doesn't attempt to account for any aspect of the utterances except the linguistic expressions. Therefore, it would be more accurate to say that our synchronic account of a language is arrived at by observing what linguistic expressions are actually uttered by the speakers.

QUESTION: Yes, but the descriptions I've seen don't really seem to be descriptions of linguistic expressions either. Why is that?

ANSWER: A linguistic description is not a simple description of linguistic expressions. The linguistic expressions themselves are not the language. The language consists in the principles, often expressed as rules, that govern what linguistic expressions may be uttered (2). The rules of the language are called its "grammar"; the grammar is said to "generate" the linguistic expressions.

QUESTION: That seems clear. The linguistic description generates exactly the set of linguistic expressions that speakers have been observed to utter, and in this sense it may be said to "describe" (or maybe "explain") them. Have I understood correctly?

ANSWER: That's approximately correct. However, I need to be a little more specific on one point. Any collection of actually uttered linguistic expressions will be an imperfect reflection of the actual language, because any collection can be expected to suffer from both errors of omission (of perfectly well-formed linguistic expressions that have by chance failed to appear in the utterances recorded) and of commission (since speakers misspeak from time to time). Therefore, the linguistic description will be designed to generate the correct set of linguistic expressions rather than the set that any observer might actually observe.

QUESTION: Let me see if I understand correctly.

1. The language really consists of a grammar that generates a set of linguistic expressions.

2. A description of a language is in fact a description of that grammar.

3. The evidence on which this description is based consists in a set of all of the linguistic expressions that speakers of the language have been observed to utter (or profess that it would be possible to utter in the case where an appropriate situation existed).

4. The description doesn't conform perfectly to the evidence, however, because no amount of observation would ever turn up all of the linguistic expressions that properly belong to the language, and also because any sizable set of actual utterances will inevitably include some slips-of-the-tongue or other kinds of errors that shouldn't properly be thought of as belonging to the language. It's up to the linguist to determine what would be included in a truly representative set of utterances, and to design the description so that it generates the revised set of linguistic expressions.

Although I feel somewhat uncomfortable with the idea that the descriptive linguist isn't bound more strictly to the actual evidence, I believe I understand the reasoning involved. What I would ask then is what methods do the linguists follow while collecting the initial data to be sure to get as nearly a representative sample of all linguistic expressions that speakers would accept as possible?

ANSWER: For one thing the linguist, in interviews with native speakers, takes care to obtain translations of sentences illustrating all of the grammatical features that are most likely to be of interest to general linguistic theory.

QUESTION: What method is used for obtaining a representative sample of different speakers?

ANSWER: There has been a tendency to assume that all native speakers of a language (since they've all acquired the same language) are intersubstitutable as informants. In cases where the linguist is a native speaker of the language, he/she can therefore rely heavily on his/her own intuition. However, there is awareness that different dialects exist, and linguists generally do check with different speakers to the extent that they're available. In languages like English which have been studied by many native-speaker linguists, canonical judgments often develop to settle points upon which there had been troublesome disagreement among the linguists (acting in their native-speaker capacities).

QUESTION: I'm still not satisfied. Let's take English as the example, and consider the following linguistic expressions recorded from native speakers in England: (all, in fact, taken from Trudgill 1983):
"You need your hair cutting". (Trudgill 1983: 18)
"He's a man at likes his beer". (ibid: 41)
"He's a man as likes his beer". (ibid: 41)
"He's a man what likes his beer". (ibid: 41)
"He's a man he likes his beer". (ibid: 41)
"He's a man likes his beer". (ibid: 41)

Do theoretical linguists design their grammars of English so that they will generate these sentences, or do they exclude them as misspeakings or incorrect in some other way?

ANSWER: There isn't yet any grammar of English that's complete enough for us to be able to answer that with any certainty.

QUESTION: Then do the statements that linguists make about rules of English, for example in discussions of universal grammar, imply the inclusion or exclusion of such sentences?

ANSWER: I would imagine that where they imply either, it's more likely to be exclusion. Note that most of these constructions are probably not known to most speakers of English (3).

QUESTION: I might add that most of them are quite unfamiliar to me, too, and that they don't seem like what I naturally recognize as "English". However, there are other expressions that I wouldn't normally use that are nevertheless more familiar. For example, Trudgill cites (ibid: 43) "He don't know a lot, do he?", and (ibid. 46) "I can't eat nothing". Or consider the following sentence taken from a fall 1994 issue of the Honolulu Advertiser: "Union County fell in love with them babies". This seems even more familiar, no doubt partly because at one time in my life I'd probably have said the same thing. There are very many (a hundred million or more?) native speakers of English who regularly use such sentences and find them natural. Would the statements linguists make about rules of English imply the inclusion or exclusion of such sentences?

ANSWER: I'd guess they would most often imply exclusion. However, note that it would probably be possible to elicit from many--probably most--of these speakers the judgment that such sentences--although they do regularly use them--are not in fact correct English.

QUESTION: But isn't this "correct" English something that was artificially created by a few members of the intellectual elite of England (notable among them, Samuel Johnson and a few 18th century grammarians)?

ANSWER: Well, some individuals did play very influential roles in prescribing standards to be followed and taught. However, the standard language that has emerged was rooted in existing usage.

QUESTION: Here it seems appropriate to bring in the distinction made by Heinz Kloss between "Ausbau" and "Abstand" languages. An Abstand language qualifies as a separate language by virtue of its linguistic distance (Abstand means "distance", of course) from other languages. As for "Ausbau" languages, here are Kloss's (1967: 29) own words:

"The term Ausbausprache may be defined as 'language by development'. Languages belonging in this category are recognized as such because of having been shaped or reshaped, molded or remolded--as the case may be--in order to become a standardized tool of literary expression."

He says that English is both an Abstand and an Ausbau language (as are French, German, etc.). My question is: Is it English the Abstand language or English the Ausbau language that linguists are describing under the name "English"? Isn't it Ausbau English that linguists are talking about when they say, for example, that "English" has such and such a rule?

ANSWER: Yes, I suppose that statements that purport to be about "English" with no further qualification--statements that implicitly exclude some of the "incorrect" usages that we've been discussing--are essentially about Ausbau English.

QUESTION: How does this square with the claim that linguistics is concerned with description rather than prescription?

ANSWER: Maybe we could say that--at least as far as English is concerned--what linguists do is describe the prescriptions; or more accurately, describe the speech of those speakers who best represent the prescriptions.

QUESTION: Or describe a corrected version of the speech of those speakers who best represent the prescriptions?

ANSWER: Yes.

QUESTION: But Kloss says very explicitly that Ausbau languages do not result from natural processes of linguistic change. I'll quote his explanation of why he chooses to use words such as "reshaping" rather than "development" in speaking of the how Ausbau languages came to be as they are:

"Terms such as reshaping or remolding or elaboration, by focusing on deliberate language planning, help us to avoid a misunderstanding that the term development might lead to, namely that 'Ausbau' might come about by that slow, almost imperceptible and quite uncontrolled growth which we are wont to call natural." (Kloss 1967: 29).

I seem to be getting nowhere in my attempt to find out what is known (or assumed) about how languages change and differentiate. First we established that a "linguistic change" is a change between an earlier and later state of the same language. Then, when we turned to examine a specific example of a language state (the present state of English), we found that the standard descriptions are concerned with an artificial system that (according to Kloss) explicitly does not illustrate the normal processes of linguistic change.

Can any linguistic description of any Ausbau language be of any relevance to historical linguistics?(4)

ANSWER: It's hard to see how it can.

Conclusion:

It seems obvious that if the successive états de langue of English are seen as successive states of the standard language (the English "Ausbausprache"), then they can teach us nothing about the nature of linguistic change in natural languages. In another Note I want to continue by considering the question of language states from the perspective of Abstand languages.

NOTES

1. Although this point isn't especially relevant to this discussion, the discussion nevertheless seems incomplete if it doesn't mention that contemporary theoretical linguistics is very particularly concerned with one type of linguistic expression--the sentence--which it assumes can be identified in all languages. Back up

2. Where did such an assumption come from? I'm no expert on the history of Western thought, but I have the impression that one factor was the tradition of teaching Latin by rules in the Middle Ages and even before. Another was the creation of standard languages in Europe. One explicit aim in designing standard languages was to create rule-governed systems--in marked contrast to what was thought to be the nature of the vernaculars on which they were based. Thus, rules were associated with the imposition of discipline on what were naturally undisciplined systems--in short they were associated with prescription. Back up

3. It might be mentioned that Charles Hockett (1955: 18-22, and 1958: 332-37) proposed two diametrically-opposed conceptions of the relation between a language description and the speech of the speakers of that language. In the first conception, what would be described would be the "common core"--that is, just the total set of features that are shared by all speakers. In the second conception, the descriptum would be the "overall pattern"--that is, everything that is in the repertoire of as much as a single speaker. An attempt to describe the common core of English would presumably exclude the sentences in question, but I'm not aware of any such explanation ever being offered. However, such matters are not discussed much. Back up

4. For that matter, what relevance do statements about Ausbau languages have to any questions about so-called "natural" language? Back up