Ethnolinguistic Note 28

Links to pages: 434, 435, 436, 437, 439, 440, 442, 443, 444

Comments Welcome

George W. Grace

University of Hawaii

June 8, 1987

ETHNOLINGUISTIC NOTES

Series 3, Number 28

THE TRANSLATION OF CASUAL SPEECH

I want to call attention here to the translation of casual speech because I think it is more likely than some other forms of language use to be subject to (and therefore to provide illustrations of) the relatively informal kind of translation that I have been calling "perlocutionary". Although I speak of casual speech, I do not intend to exclude writing. In fact, the kinds of language use that I want to exclude as non-casual are particularly likely to be written. These non-casual kinds seem to fall into two general categories in both of which the form of the linguistic expression is more important than it is in casual speech.

The first category involves wording or "language" which may be thought of as "formulations". I will call these formulated linguistic expressions. One kind of example would be a legal document such as a statute or a contract. Those who draw up such a document normally anticipate that its precise wording--its "language"--may subsequently be subjected to intense scrutiny, and they are normally correspondingly careful in its formulation. The category of formulated linguistic expressions is by no means restricted to the language of legal documents; it surely should be thought of as including, inter alia, much philosophical, scientific, and bureaucratic prose.

The second category involves forms especially designed for their aesthetic or other effects. These effects would include allusions and plays on words as well as poetic devices such as rhyme, assonance, and alliteration. For convenience, I will refer to these as literary expressions.

For the discussion to follow I will need some terminology which I have used elsewhere, but which is not generally familiar. Therefore, I will review a few key terms before going on.(1)

To begin with, there is what I will call a speech act. I am using that term in an extended sense to include the uttering or writing of anything of any length from a single word to a speech or lecture or a book. For the moment, we may say that translation (the term here being used broadly so as not to exclude what is usually referred to as "interpretation") involves producing a speech act in a target language, some aspect of which is equivalent in some way to some aspect of the original source-language speech act. Clearly, that characterization is very vague. One of the most obvious challenges for a theory of translation is to clarify it further.

I will use the term linguistic expression to refer to the artifact of the speech act, i.e., that which is uttered or written. (From another perspective the linguistic expression can be thought of as the vehicle of the speech act). In the case of oral speech, the artifact is what a tape recorder would have recorded, or in the case of writing, it is what actually appears on the page (again, regardless of length).

There is a conventional idea that translation is a matter of designing a linguistic expression in the target language which is equivalent in meaning to a given source language expression; that is (1) that translation involves a relation between linguistic expressions and (2) that that relation is identity (or at least, equivalence) of meaning. Translation according to that conception is what I have been calling locutionary translation.

The alternative to locutionary translation is what I call perlocutionary translation (cf. Grace 1986). Rather than equivalence in meaning between linguistic expressions, perlocutionary translation aims for equivalence in effect between speech acts. What is sought in perlocutionary translation is an expression to serve as the vehicle of a target-language speech act which will serve the same purposes that the original speech act was intended to serve.

Now, the translation of casual speech is of interest because it seems less likely than the translation of either formulated or literary expressions to focus sharply upon the source-language linguistic expression as such. It is in the translation of formulated expressions that the locutionary conception of the translation objective seems most appropriate. The translation of casual speech can typically be fairly unself-consciously perlocutionary.

The principal point that I want to make in this Note is that our way of talking about translation gives undue prominence to the source-language linguistic expression. For example, one often hears questions of the form, "How do you say X in Y?" where Y is the name of a language and X is a linguistic expression in some other language, and the addressee therefore is offered no information except that contained in X. Such a question may possibly present no problems if the aim is to elicit dictionary equivalents. However, if the questioner intends for X to be thought of as a (real or potential) speech-act vehicle--i.e., if he/she envisages some situation of use--such a question is likely to be very difficult to answer without further information. The kind of information the addressee is likely to ask for will depend on circumstances. Perhaps, for example, the questioner has encountered X in reading and is attempting to interpret the text, or perhaps he/she is contemplating performing a speech act in Y and is trying to devise a linguistic expression to serve as its vehicle.

Of course, it is generally recognized that this problem exists. However, the interpretation which it is usually given corresponds closely to the sequence of events which I described. I.e., the starting point is the linguistic expression. However, the linguistic expression in isolation is found to be ambiguous--it may have an unmanageably large number of meanings. If translation (or, for that matter, any communicative transaction in which it is involved) is to be successful, it is necessary to disambiguate it--i.e., to determine which of its meanings is intended. Disambiguation is then supposed to be accomplished by considering the linguistic expression in relation to the context and situation in which it was used. As it is sometimes expressed, the meaning we are looking for is the situated meaning of the linguistic expression, and the situated meaning is a selection from the set (or a narrowing of the range, for whoever finds that metaphor preferable) of its meanings in isolation.

Much of the literature on translation (especially translation where isomorphism is not possible) is concerned with how the translator (whether human or electronic) can figure out what the original speech act meant. In the case of locutionary translation that means determining the meaning of the source-language linguistic expression (which in practice does seem primarily to mean resolving ambiguities).

However, our concern here is with perlocutionary translation. Of course most authors do not explicitly distinguish perlocutionary from locutionary translation, but it seems clear that in the case of perlocutionary translation, figuring out what the speech act meant means figuring out what the speaker meant--that is, what the speaker's intentions or purposes were.

Clearly, if a translation--or at any rate a perlocutionary translation--is to be satisfactory, the translator must have an adequate understanding of the speaker's (at least ostensible) purposes. How a translator may obtain that understanding is of course a matter of considerable interest. (For that matter, how anyone obtains understanding of anything is of interest). But that is not the question with which I am concerned here. Rather, I am concerned with the question of what is required for a translation to be judged satisfactory. (We may note that it is possible for the translator to understand the speaker's intentions perfectly well and still produce an unsatisfactory translation. Moreover, this can be either unintentional or intentional--there is nothing in the process which makes it impossible for a translator to lie).

In order to separate the problem of whether or not a translation is satisfactory from the problem of understanding what was originally meant--what the original speaker's purposes were--I propose that we imagine the situation where the original speaker is also the translator. In such a case we may suppose that he/she already understands the original purposes as well as they can be understood.

We may assume that the objective of perlocutionary translation is to satisfy the purposes of the speaker.(2) We may think of these collectively as the effect that the speaker is seeking to achieve. And since, in the case that we are imagining, the translator is the speaker, he/she is the ultimate authority on what the speaker's purposes are and whether or not they have been satisfied by a particular speech act.

Now, we want to consider a hypothetical situation where a bilingual speaker is asked to translate something which he/she has said. First of all, it is possible that the request will be interpreted as calling for locutionary translation. We could imagine the scenario in this way. First, the speaker speaks. Then, having the artifact of the speech act (i.e., the source language linguistic expression) before us, we ask him/her to translate that expression. In such circumstances, and particularly to the extent that we appear to be interested in the linguistic expression as such, the speaker may feel called upon to attempt a locutionary translation (and therefore may undertake to cope with the problems that that may entail).

However, such a constraint seems an unusual one for real life situations. It seems that in the typical case of casual language use it would be more realistic simply to ask the speaker to express his/her "message "(3) in the target language (i.e., leaving him/her free to re-phrase it). Therefore, in the situation which we are imagining, the speaker (being fully bilingual) repeats his/her "message" in the second language without the constraint that the linguistic expression used in the repetition has to match the original linguistic expression in any particular way. The question now becomes: why is latitude permitted the speaker to make the repetition deviate from the original speech act?

Before considering this question, however, let me call attention to the fact that even in a completely monolingual setting, we are not uncommonly asked to repeat something (because for whatever reason it wasn't heard, or understood, or remembered properly), and that in such circumstances it is very likely that we will not repeat the original expression verbatim.

There may be several reasons why the original is not repeated verbatim. One would be that, upon reflection, the speaker has slightly revised his/her purposes. However, in this case it would not be accurate to consider the second version to be equivalent to the first. But another reason would be that the original was underdetermined by the speaker's purposes--that there are one or more alternative expressions which are at least as apt to serve his/her purposes as the original one. And perhaps he/she thinks there might have been something in the original expression that was intrinsically hard for the particular audience to understand, and so chooses an alternative which otherwise seems equally suitable for his/her purposes.(4)

We have been considering the situation where a speaker is asked to repeat something he/she has said. But this is quite parallel to the case where the speaker is a bilingual and we ask him/her to repeat in the other language (perhaps for members of the audience who did not understand the language first used) something that he/she has said. If an isomorphic translation is possible, it is one option (an option roughly analogous to a verbatim repetition in the same language of what one has just said). Or, again, one could improve upon the original, as a result of having revised one's purposes, however slightly. And in that case, again, it would not be accurate to consider the target-language version to be equivalent to the original.

But one could also improve upon the original, making it a more suitable vehicle for satisfying one's purposes, by taking advantage of superior resources made available by the target language. In this case, I would maintain, it would be quite misleading to accuse the speaker/translator of misrepresenting the original.

Our question was why latitude is permitted the speaker to make the repetition of "the same message" deviate from the original. The answer, in short, is that the choice of linguistic expression is often substantially underdetermined by the speaker's purposes. And it seems that we should be as willing to accept the latitude made available by the underdetermination of expressions by purposes in the bilingual case as we were in the monolingual one.

The main point that I want to make here, therefore, is that translation (except locutionary translation) cannot be seen as an operation performed on a linguistic expression. It is the execution of a strategy designed to satisfy someone's (normally the original speaker's) purposes. The original source-language linguistic expression is of interest only in that it serves as a guide to those purposes for a translator who is otherwise not privy to them.

It should be emphasized, furthermore, that this statement applies to a "situated" linguistic expression (i.e., a linguistic expression taken together with the context and situation of its use) with the same force as to a linguistic expression treated as an abstract object. Translation is not an operation performed on a linguistic expression, taken either in isolation or in situ. The contextual information helps the translator in his/her attempts to guess the speaker's purposes, but it has nothing to say about what constitutes a translational equivalent.

NOTES

1. To provide more context, I should perhaps point out the following argument that I have made in Grace (forthcoming) and elsewhere:
1. We know of no way to specify the meaning of a linguistic expression apart from its form.
2. One consequence of this fact is that there is no basis for saying that two linguistic expressions which are different in form mean the same thing (or for that matter that they do not mean the same thing--the question itself is meaningless).
3. In some cases it is possible to design a target-language expression that is isomorphic with the source-language expression to be translated. In many (most?) of such cases, the possibility of isomorphism results from a shared reality-construction, and then it is probably fair to regard the two expressions as equivalent in meaning. However, where isomorphism is not possible, the goal of what I have been calling locutionary translation--i.e., the goal of producing a target-language expression which is equivalent in meaning to the source-language expression--is illusory. (However, nothing said here should be interpreted as denying that the possibilities for satisfactory translation between languages are unlimited as long as enough cultural Gleichschaltung has occurred).
4. The alternative to locutionary translation is perlocutionary translation. (Back up)

2. That the purposes which the translation aims at satisfying are those of the speaker should at least be the most general case, although in particular cases the translator might be working for--and most concerned to satisfy the purposes of--some other party to the transaction. (Back up)

3. "Message" is a "conduit-metaphor" concept (cf. Reddy 1979) which suggests the existence of some intended "content" which is separable from any "container". There is no place for such a concept in the view of the matter which I am trying to develop here. But as Reddy noted, it is impossible to talk or write about language use in a non-convoluted way without falling into the conduit-metaphor way of talking. Consequently, using the word "message" seemed the best way to convey to the speaker in simple, straightforward (although presupposition-laden) English what he/she was being asked to do. (Back up)

4. In fact, it seems probable that our only basis in experience for the idea that different expressions can mean, at least approximately, the same thing is precisely that the choice of linguistic expression is so frequently underdetermined by the speaker's purposes. That is, the differences in the meanings of several expressions that may come to mind are not relevant to his/her purposes, so that for those purposes it doesn't make any difference which one is used. This is surely the basis of what we call paraphrase, and it is surely for this reason that translation so often seems perfectly satisfactory. (Back up)