Comments Welcome
George W. Grace
University of Hawaii

Ethnolinguistic Notes

Series 4, Number 4

LINGUISTIC CHANGE

3. ARE "VARIETIES" THE LOCUS OF THE "ÉTAT DE LANGUE"?

The two preceding Notes in this series have been concerned with the question: What is a linguistic change? We saw that a linguistic change has been defined as a difference between successive states (états de langue) of the same language. But in the last Note we were forced to the conclusion that all of our so-called "languages" are fictions--that there is no general agreement as to how, on the basis of what is (or can be made to be) observable, we are to determine how many languages there are or to draw (non-arbitrary) boundaries around them.

More damaging still to our quest is the fact that even in a case like English, where there is pre-existing agreement on what varieties are to be counted as belonging to the language, there is no agreement at all about how, on the basis of observable behavior of the speakers of those varieties, we can arrive at an accurate specification of the language state.

However, it's noteworthy that our discussion has turned on the question: which varieties belong to such and such language? It has thereby assumed the existence of another entity--the "variety"--essentially a more homogeneous version of the language. This might be taken to suggest that, unlike languages, varieties are real, and that it's perhaps in them that we should seek the locus of linguistic change.

[Here I'll return once again to the hypothetical dialog with myself].

QUESTION: Is it possible that these "varieties" provide the solution to our problems; that we should look to them as the locus of linguistic change? What actually are varieties?

ANSWER: It's hard to answer that question without referring to the fiction of individual "languages". It's been obvious that there was a lot of variation in the ways in which many languages (i.e., what were acknowledged by consensus to be languages: e.g., English) were spoken. This variation was assumed to be explained by the existence of different alternatives in various parts of the grammar.

The idea arose that the grammar could be thought of as being a composite of a number of slightly differing--and more homogeneous--grammars, each of which could be thought of as a variant realization of the general grammar. The way of speaking associated with each variant grammar would be called a variety of the general language.

QUESTION: And that's one of the things we've been looking for, of course: a system that's sufficiently homogeneous that it would be possible to specify its state (its état de langue) at each point in time. Might the variety be the system we've been looking for?

ANSWER: It's not surprising that that possibility should suggest itself because that's really the main idea behind concepts such as variety (and dialect, idiolect, etc.).

QUESTION: Explain.

ANSWER: Concepts such as variety (and dialect and idiolect) can be traced back to the fiction about how language works. According to that fiction(1) what would be required for the most effective communication would be a system in which the set of possible expressions is precisely defined and the meaning of each can be specified unambiguously. The failure of real languages (i.e., Abstand languages) to conform to that ideal has been explained away to some extent by the idea that what individual speakers effectively use are systems that are more homogeneous than the language--that are more nearly the supposedly ideal well-defined systems.

QUESTION: Are you saying that that idea is wrong?

ANSWER: Well, there are some obvious problems with it. For example, we've seen in the two preceding Notes that there's no satisfactory way of defining "a language". Well, the problem of defining "a variety" is, if anything, even greater.

QUESTION: But we said above that there was agreement on what varieties are to be counted as belonging to English. That seems to suggest that we can identify the varieties. If we can identify them, why do we need to define them (that is, if we know the extension of the term, there seems no need to worry about its intension)?

ANSWER: Interpreted that way, what we said above was misleading. It wouldn't make any sense to attempt to list all of the varieties of English unless "variety" were given a special ad hoc definition for the purpose, because the things to which the word has been applied are extremely diverse. For example, David Crystal, in his dictionary (Crystal 1992), makes the following statement in the entry for "variety": "Varieties of English include scientific, religious, legal, formal, conversational, American, Welsh, and Cockney".

In fact, it seems to me that the prominence of the word "variety" (and the decline of "idiolect´and perhaps "dialect") in recent linguistic discourse reflects an abandonment of the objective of classifying the variation within a particular language into the kind of homogeneous systems that were imagined.

QUESTION: Explain.

ANSWER: The existence of what have been called "dialect" differences has long been recognized. "Dialects" were thought of (from the "Abstand"(2) perspective) as very like languages except that they were too close linguistically to another dialect or dialects for them to count as separate languages. However, it also proved impossible to identify dialects that were clearly demarcated linguistically or that didn't themselves have significant internal variation.

At one point, linguists considered the possibility that the homogeneous system that they sought might finally be found in the individual speaker. The term "idiolect" had considerable currency at one time. However, it has turned out that individual speakers don't have homogeneous systems. They are highly sensitive to, and make use of, different varieties spoken around them. They strive to, and are able to, accommodate their speech to their interlocutors. A speaker's choice of variety can itself carry meaning.

Linguists now often use locutions such as "my dialect", but without any implication about who else, if anyone, might have the same dialect. It seems to mean no more than, "English (or whatever) as I know and use it".

QUESTION: And where does the word "variety" come in?

ANSWER: It's my impression that it's come into prominence more recently, and that it's connected to the discussion of "variation". It seems usually to refer--most often with a deliberate lack of specificity--to variation of the sort that we were accustomed to calling "dialectal"(3).

QUESTION: Then the conclusion is that varieties, dialects, etc. are as much fictions--as much a product of as-if thinking--as languages?

ANSWER: Yes.

QUESTION: And therefore of little or no more help in our quest for a linguistic system with a specifiable état de langue?

ANSWER: Actually, they're probably of even less help, because they're much less likely than the ("Abstand") language to maintain their integrity over time. That means that even if you could isolate a well-defined variety at some particular moment and obtain a good description of its linguistic state (état de langue), you might find that it had no clearly isolable successor at any later time (or antecedent at any earlier time).

QUESTION: It seems, then, that the conclusion must be that if we're ever fully to understand linguistic change, we first free ourselves from the idea that there are long-continuing entities that undergo the change.

ANSWER: Yes, but then if we're ever fully to understand any aspect of language (or of anything else, for that matter), we'll eventually have to find a way past all of our fictions and oversimplified working assumptions. As we make progress, we'll gradually find it necessary to dispense with successive fictions and to try to formulate the real questions (or at least, increasingly close approximations to them) and to try to find answers to them.

NOTES

1. That fiction may be roughly described as follows:
Language takes the form of a (large) number of individual systems, each of which consists of an inventory of lexical items that specify bits of (actual or hypothetical) reality, and a system of grammatical rules that specifies relations among (the referents of) lexical items. It's such a system (the synchronic "state" of such a system) that the kind of linguistic description that we know has been designed to describe.

According to this fiction speakers use these individual systems to encode propositions by putting selected lexical items into grammatical constructions, thereby asserting the existence of particular relations among particular elements of (putative) reality. The conventional name for these individual systems is "languages".

However, one problem that arises with this conception is that what we call "languages" in most contexts (the languages that we give names to such as "the English language") turn out not to be the kind of homogeneous systems that the conception calls for. Nor are they the kind of system that linguistic description is designed to describe. On the contrary, they often exhibit extensive internal variation.

This has led to the idea that the "variation" found in these languages is really a matter of a number of (more nearly) homogeneous systems (sometimes called "varieties"), and that it is one of these homogeneous systems that a speaker is actually using when s/he encodes a proposition. Back up

2. The word "dialect" is often used in "Ausbau" contexts to refer to a variety that is subordinate culturally or politically to some other variety, with only the superordinate variety receiving the designation "language". Back up

3. As opposed to stylistic or "register" differences. There has been a tendency in linguistics to distinguish between the differences between speakers from different geographical areas or social classes (geographical or social dialects), on the one hand, and the differences that exist between the ways anyone speaks in different social situations (e.g., formal vs. informal styles or registers) on the other. It has seemed to me that "variety" is used most often for dialect type differences, but note that Crystal in the quotation cited above includes "formal" and "conversational" as varieties alongside "American" and "Cockney". Back up

REFERENCE

Crystal, David. 1992. An encyclopedic dictionary of language and linguistics. Oxford: Blackwell. Back up

COMMENT (26 August 1996):

(From Michael L. Forman, Department of Linguistics, University of Hawai'i at Manoa)

In Ethnolinguistics Notes, Series 4, Number 4, you say that in English there is pre-existing agreement on what varieties are to be counted as belonging to the language. What are the grounds for saying this?

I would think it not to be the case. I wonder, myself, if many patterned usages as seen in the work of some of our students, from, say, Japan, are to be counted as varieties of the language. If we say that they are not to be counted, then what do they count as varieties of? Aren't they varieties of Japanese English? And isn't Japanese English "itself" at least a set of varieties of English, then?

I know that there is a lot of disagreement about whether pidgins and creoles with what is now called "English lexifier" are in fact varieties of English, or are not varieties of English but are "different languages".

So I was surprised to find you saying that there was pre-existing agreement about English. I guess I'd agree that there is pre-existing agreement, or expectation, that if X says so, it must count as English -- but not if Y says so?

So, what are the grounds? What does it mean to say that there is pre-existing agreement about what counts as a variety of English? I don't know, really.

REPLY (26 August 1996):

You're right, of course.

It's not easy to come up with a revision of my statement that really seems satisfactory. For the present let me just say that I think there's pretty good agreement (reaching back a long way in time) about where the boundary between English and its closest Germanic relatives falls. Maybe that's all I should claim.

The point I was trying to make was that there should be pretty good agreement on what utterances (or the utterances of what speakers) should be accounted for in a description of English, but that we nevertheless don't really try to account for all of them, and we have no strategy for doing so.

I'd agree that "Japanese English" would have to be considered a variety (or an indeterminate number of varieties) of English. Of course, linguists would argue that these varieties don't need to be accounted for in a description of English because only the speech of native speakers of the language counts, and these people speak English as a second (or third, etc.) language.

If the distinction between pidgins and creoles is that creoles have native speakers, then the pidgins-but not the creoles--could presumably be eliminated from the description of English on the same grounds as Japanese English. In any case, disagreement about whether these pidgins and creoles count as English directly contradicts my statement about pre-existing agreement as to what varieties belong to the language, and my explanation that what I really meant was what's to be accounted for in a description of English doesn't solve the problem for creoles. Therefore, I don't see any way to salvage the statement.

But I don't feel too bad; you really reinforce my main point. What you say not only demonstrates that there is uncertainty about what some particular varieties are varieties of (e.g., is a certain pidgin or creole a variety of English?) but also that there is uncertainty about what counts as a variety at all (e.g., how many varieties are subsumed under the heading "Japanese English" and what are they?). It's clear, therefore, that what you say actually reinforces the point that the concept "variety" is no help toward finding a locus for the kind of état de langue that our definition of linguistic change requires.

First put on the Web on 19 August 1996
Corrected 28 November 2000