Ethnolinguistic Note 32

Links to pages: 475, 476, 480

Comments Welcome

George W. Grace

University of Hawaii

September 15, 1988

ETHNOLINGUISTIC NOTES

Series 3, Number 32

THE IDEA OF A THEORY OF TRANSLATION:
SOME GENERAL OBSERVATIONS

In the literature on translation one sees frequent references (cf. for example, Amos 1973, Catford 1965, Frawley 1984, Graham 1981, Grähs et al 1978, Kelly 1979, Longacre 1958, Nida 1976, 1977, Richards 1953, Rose 1981, Song 1975, Wilss and Thorne 1984) to "theory". These references, as represented by the titles of the works, seem to be classifiable very roughly into four groups. Those of the first group appear to suggest that a number of different theories of translation already exist. A second group suggests that there is a single theory of translation which either already exists or is being proposed in the work bearing the title. A third may be said to assume only that such a theory is a desideratum toward which we are working, while a final group probably intends to claim nothing more than that translation practice has some kind (or kinds) of theoretical underpinnings.

However, whatever the differences among these authors, the appearance of the word "theory" in so many titles surely reflects some kind of concern within the field of translation (if indeed translation may be thought of as constituting a field) with the theoretical underpinnings of the field.

Of course, I have been arguing for some time that translation is of very central significance for an understanding of the nature of language. Therefore, any theory that helped to characterize and explain translation would be likely to be of considerable interest to linguistics ("linguistics" being understood here as the name for a general science of language). The more formal and the more general such a theory, the better. However, we are far from a formal, general theory at present. This Note can therefore attempt nothing more than some introductory remarks about the requirements which such a theory would need to meet.

A general theory of translation would presumably have to provide a single overall characterization of all of the things which go by the name "translation" and also provide a basis for the analysis of all kinds of "translation" practice. But what form might such a theory of translation take, and what range of phenomena would it need to explain?

Although it is apparent that the word "translation" is applied to many different kinds of operations with different kinds of objectives, the recent literature seems to be concerned particularly with what might be called "information processing" uses of language (which is to say with what is frequently called "referential meaning"). That is, there seems to be a tendency to consider the prototypical act of translation to be concerned with a text the principal purpose of which is to state facts (or to be somewhat more precise, to affirm, question, deny, etc. what might be described as "fact-like content" [I gave what I hope is a much clearer description of this in Chapter 3 of Grace 1987a and will not devote more space to it here]). Although I said that the literature treats this as its principal purpose, it often in effect treats it as the only purpose.

However, a general theory of translation cannot assume that translation is always, or even typically, of this kind. In fact, there are a number of different ways in which actual translations may depart from this ideal.

To begin with, there is surely never a case where a speaker or author has only a single purpose in producing a speech act (cf. Grace 1986b), although some cases do approximate somewhat more closely to this ideal than do others.

Secondly, sometimes none of the speaker/author's purposes have to do at all with Information processing uses of language. That would largely be the case with poetry. Furthermore, I presume that even what are called "translations" of Lewis Carroll's "Jabberwocky" are part of what a general theory of translation would have to account for.

Thirdly, translations do not necessarily have to be designed to serve the purposes of the author of the original speech act--they can be designed to serve anyone's purposes, and the basis upon which the success of a translation is judged surely will vary according to the purposes for which it was made.

Fourthly, the supposed prototypical act of translation was described above as being concerned with a text, but that presupposes writing, and writing is a relatively recent development in the history of human language. It is surely a relatively recent development in the history of translation between languages as well. And without writing, the availability of the source-language text (in the sense of an exact wording) to the translator cannot be assumed (except for a very brief interval immediately following its utterance).

And finally, something akin to translation (and often referred to as "translation", although often with the suggestion that the word is being used loosely) sometimes is necessary between people who are native speakers of the same language, as when one is using a "jargon" which the other doesn't know (although in such cases it seems to be assumed- -at least by some--that there is some kind of "plain talk" which could be resorted to which would obviate the need for translation). (There is also much evidence that sometimes people who are native speakers of the same language, but who are of different ethnic or social backgrounds, also would greatly benefit from the services of an interpreter or translator).

I have been considering some of these problems during the last few years, and I have been attempting gradually to develop a framework for discussing translation--essentially a way of talking about it (which is, of course, a very much less ambitious goal than a formal theory). The framework which I have developed so far (cf. especially Grace 1986a-b and Grace 1987a-e) may be represented roughly as follows:

1. Translation always involves what I call a speech act (a term which I use in an extended sense to refer to the uttering or writing of anything of any length from a single word to a speech or lecture or a book--cf. Grace 1987b: 435). A speech act requires a linguistic expression as its vehicle.

2. We may think of a linguistic expression by itself as having a meaning. This meaning is what I have called its meaning-by-computation or computed meaning (cf. Grace 1987e: 455). It is--theoretically--computable from the meanings of the constituent lexical items of the expression and the meaning(s) of the construction(s) in which they appear. [Of course, very little has been done toward developing a way of actually computing meanings. Is it possible, for example, for structurally very unlike expressions in different languages to have identical or even similar computed meanings?]

3. I call translation whose objective is likeness of meaning (which I believe has to mean computed meaning) between linguistic expressions (i.e., between texts) locutionary translation.

4. Opposed to locutionary translation is what I have been calling perlocutionary translation. This is defined as translation whose aim is to produce some kind of understanding (where "understanding" is to be interpreted broadly enough to include effects that might not ordinarily be thought of as understanding) in an audience rather than to produce a text (Grace 1986b: 404-5).

5. It is assumed that all individuals have (at all times) their own purposes. The participants in any communicative transaction have at that time their own motivating purposes--those purposes which are actually in play in that transaction (Grace 1986b: 408).

6. A perlocutionary translation can be seen as attempting to satisfy someone's motivating purposes. They can be those of the original speaker, but they can also be those of someone else. Presumably the basis on which the quality of a perlocutionary translation is to be judged is a matter of perspective--most often a matter of its success in satisfying the purposes of whoever commissioned it.

7. A locutionary translation may be said to take the point of view of the original speaker/author. That is to say that, to the extent that the original linguistic expression may be considered as representing that speaker's motivating purposes, the (locutionary) translator may be said to be undertaking to satisfy the purposes of that speaker. In any case, in the absence of provisions to the contrary, it seems to be most usual to expect translation to take the point of view of the original speaker. (Should that be regarded as some kind of default position for a theory of translation?).

8. There seem to be some purposes for which certain kinds of texts--which I have called formulated linguistic expressions [cf. Grace 1987b: 434]--are required. Except in a particular kind of favorable circumstance, formulated linguistic expressions seem to require locutionary translation (i.e., translation which matches the computed meanings of the texts). But the computed meaning nonetheless yields precedence to what I have called the what-they-would-say-in-the-same-situation (WTWS) principle (cf. Grace 1987e) when that principle is applicable.

That is, when something qualifying as "the same situation" can be satisfactorily identified in the life of the target-language community, and when the linguistic expression that speakers of the target language would use in that situation can be fairly exactly specified, then the latter linguistic expression is the preferred translation rather than an expression with the same meaning-by computation.

9. Employing the WTWS principle is complicated by the difficulty of identifying equivalent situations in different communities. The most serious difficulty derives from the fact that the situation must be defined in such a way as to ensure (1) that the speaker will elect actually to speak at all and (2) that what linguistic expression he/she will in fact utter when he/she does speak can be accurately predicted. That is, the speaker's purposes must be recognized as part of the situation. Indeed, the situation in question must ultimately be a subjective one.

However, the WTWS principle is of more use than it might at first appear because (1) the common characteristics of the human species and of the natural environment which we inhabit, and (particularly) (2) shared cultural artifacts and institutions, result in recurrent types of situation which may be spoken of as objective situations. It is possible to define an idealized subjective situation on the basis of (1) such an objective situation and (2) a hypothetical single purpose on the part of the speaker in relation to that objective situation--for example the purpose of reporting the objective situation. (cf. Grace 1987e: 462ff.)

On this basis, the WTWS principle--which apparently must in any event be acknowledged for certain cases by a general theory of translation--can probably be developed as a more widely applicable tool.

It seems clear that the WTWS principle is very widely used in actual translation practice. Therefore, a general theory of translation would have to recognize it and, in fact, to give it some sort of characterization. I am hoping that connecting (idealized) subjective situations to objective situations which themselves can be connected (mainly) with shared cultural experiences will constitute a sufficient basis for such a characterization. (However, this whole section of the theory would mainly be valid only for cases where such shared cultural experiences exist and are available to be invoked. It is my impression that there is almost no appreciation, among linguists and philosophers at least, of the extent to which most of the kinds of translation that commonly get talked about depend on shared cultural presuppositions--cf. Grace 1987c,d.)

I believe that the foregoing considerations make it clear that any attempt at a general theory of translation would have to go far beyond a simple consideration of texts and their meanings. It seems very clear that the cases in which the only problems to be solved concern the identification and translation of the individual lexical items of a source-language text and the placing of those items in a equivalent target-language syntactic construction are very special cases indeed. Possibly a theory of translation dealing only with such matters could be proposed as a theory of a special case of translation. But it would be a very special case. Certainly no theory which was concerned only with such problems could ever be seriously considered as a general theory of translation.

And there is a further concern: to approach the problem in this way might have an unfortunate tendency to perpetuate false conceptions of what language is and how it is used.