EQUIVALENCE DEFINES TRANSLATION
Equivalence could be all things to all theorists
Although descriptions of the relation between the input and output of translational work often refer to notions of equivalence, the term would appear to be the great empty sign of such exercises. Equivalence has been extensively used to define translation, but few writers have been prepared to define equivalence itself. Indeed, it is quite possible that the term in question means all things to all theorists: since it is usually taken to be the result of successful translating, its content as a theoretical term is probably nothing more or less than the theory in which successful translating is defined. Equivalence thus perhaps means achieving whatever the ideal translator should set out to achieve. Yet this is a mere tautology: equivalence is supposed to define translation, but translation would then appear to define equivalence. One senses that something more substantial needs to be said about equivalence itself.
Historical research is of little avail here. The brief survey offered by Wilss (1982, 134-135) simply presents guesses suggesting that the English term "equivalence" entered translation studies from mathematics, that it was originally associated with research into machine translation, and that it has or should have a properly technical sense. But Snell-Hornby has used comparative historical analysis to argue against the possibility of any such technical sense, claiming to have located some 58 different types of equivalence referred to in German translation studies (1986, 15). Moreover, even if one could locate substantial common factors underlying all these variants, there is surely no guarantee that history or etymology alone will lead to the most fruitful future definition. A slightly more creative approach is required.
In what follows, I want to suggest that equivalence-based definitions of translation are fundamentally correct; but I also want to show that they say rather more than the sterile tautologies they ride on. Despite all the problems with historical usages of the term, despite recently fashionable attempts to ignore it altogether, I believe that equivalence in its most unqualified form - definitionally ideal equivalence - does indeed define translation. But to reach this conclusion, to discover what is being said but not heard, it is necessary to discard several false or inadequate notions of equivalence. We must disregard the way structuralist linguistics once used the term to suggest a symmetry of "equal values" between discrete systems; we must turn to the economics of exchange in order to distinguish equivalence from assumptions of natural use values or functions; we must see how equivalence can actually operate within a dynamic translational series based on the primacy of exchange value; and finally, we must appreciate that equivalence is not a predetermined relation that translators passively seek, but instead works as a transitory fiction that translators produce in order to have receivers somehow believe that translations have not really been translated. In all, if equivalence is ideally to define translation, we must take steps to redefine ideal equivalence.
I should stress that my subject in this chapter is no
more - and no less - than equivalence as an ideal. We shall later find
reasons for challenging its limits and for qualifying its lesser modes.
But for the moment, what interests me is the silence of the great empty
sign itself.
Equivalence is directional and
subjectless
The following are fairly representative equivalence-based definitions of translation:
"Interlingual translation can be defined as the replacement of elements of one language, the domain of translation, by equivalent elements of another language, the range [of translation]." (A. G. Oettinger 1960, 110)
"Translation may be defined as follows: the replacement of textual material in one language (SL) by equivalent material in another language (TL)." (Catford 1965, 20)
"Translating consists in reproducing in the receptor language the closest natural equivalent of the source-language message." (Nida and Taber 1969, 12; cf. Nida 1959, 33)
"[Translation] leads from a source-language text to a target-language text which is as close an equivalent as possible and presupposes an understanding of the content and style of the original." (Wilss 1982, 62)
Many further definitions could be added in this vein (cf. Koller 1979, 186 ff.). But the main variants in any longer listing would tend to concern more the nature of what is supposed to be equivalent ("elements", "textual material", "functions", "communicative effect", etc.) than the nature of equivalence itself, which, within this decidedly twentieth-century tradition, is simply assumed to exist. Indeed, in some circles, the assumption is so amorphously present that one hesitates to question its grounding. Even Quine's definition of indeterminacy, despite all its efforts explicitly to question contemporary presuppositions, feigns to be upset about the same text leading to different translations "which stand to each other in no plausible sort of equivalence relation however loose" (1960, 27). But who told Quine that wholly determined translation should depend on equivalence? Is it not strange that equivalence thus appears in the definition of both what we know about translation (determinacy) and what we suspect we do not know (indeterminacy)? But what then is equivalence itself, however loose?
It might of course be assumed that the term means exactly what it says: a relation of equal value. But such a reading would contradict the similarly widespread although perhaps less obvious features I have put in italics in the above definitions:
- In all these definitions, the term "equivalent" is used to describe only TTs, the products resulting from the translating process. It is not used to describe the ST, the abstractly initial material, nor the Y text, the textual material as it arrives in the place of the translator. This one-sided use implies an asymmetry that must be considered at least odd if associated with a relationship of presumed equality.
- The verbs employed or implied ("replace", "reproduce", "lead to", etc.) not only refer to processes, but are decidedly unidirectional in nature. Translating goes from Y to TT, and if the process is reversed it is called "back-translation", as a kind of underhand reversal of the correct way of the world.
- The described processes are also peculiarly subjectless: it is obvious that somebody or something must be doing the "replacing" or "reproducing", but this person or thing appears to have no expressed place in the translational process. Although there must be at least some notion of location implied in terms like "replacement" and "lead to", the subjectless nature of this place suggests that no one particularly cares who or what is doing the work.
Taking all of this together, we find that the term equivalence
is commonly associated with the end result of translating as a one-way
process occurring in an apparently subjectless place. Equivalence is directional
and subjectless. I believe that these distinctive features are highly useful
for the definition of translation. Moreover, their implicit asymmetry presents
significant problems for certain less definite ideals like equivalence
as an affair of "equal values". The first of these problems is the nature
of value itself.
Equivalence is asymmetrical
Although "value" is generally not a technical term in contemporary translation studies, it does make frequent and prolonged appearances in Saussure's Cours de linguistique générale, widely held to be one of the foundational texts of modern linguistics and often cited in arguments against translatability. Saussure describes linguistic elements as having values corresponding to their mutual oppositions:
"Modern French mouton can have the same signification as English sheep but not the same value, and this for several reasons, particularly because in speaking of a piece of meat ready to be served at the table, English uses mutton and not sheep. The difference in value between sheep and mouton is due to the fact that sheep has beside it a second term while the French does not." (1916,115)
Saussurean value is thus positional and relative within a fixed tongue, since "in language there are only differences without positive terms" (120). It is important to stress Saussure's distinction between, on the one hand, "value" as the entire semantic potential left to an element by the presence or absence of neighbouring terms, and on the other, "signification" as the particular use made of that element in a given situation. This distinction is clear in the example of the chess game, where the value of the knight is described as its capacity to carry out any number of moves within the limits of certain rules, its signification then being the import of each individual move. So far, so good.
Since Saussurean value refers to the relative positions of elements within an entire tongue, the fact that different tongues divide semantic space in different ways theoretically denies the very possibility of different elements being of equal value. Vendryes even considered equivalence to be contrary to the nature of the tongue as system, arguing that as soon as two elements become "equivalent" within the same system, one of them is forced to disappear (1923, 381). It is then not surprising that Saussure's synchronic linguistics excludes not only questions of equivalence but also all reference to one-way processes and to places of lesser dimensions than tongues. Saussure does not talk about translation. For example, he chooses not to tell us that the difference in value between "sheep" and "mutton" is due to the historical situation in which Anglo-Saxon servants presented what they called "sceap" to their Norman masters, who called the same object "moton". The positional values of the terms were changed - were exchanged - as soon as the meat approached the master's table and intercultural communication was established. It is only through asymmetric situations like this - which clearly involve translation and quite massive material transfer (of meat, of armies), as indeed does Saussure's description of the example - that the linguist has access to the comparable terms enabling him paradoxically to demonstrate that equal values (and thus "translation" itself) are strictly impossible. But the pertinent translation had taken place centuries before!
If linguistic notions of value should thus suggest that there is no such thing as equivalence, it is because they are logically posterior to beliefs in precisely this possibility. Just as Saussure received his example from the asymmetrical social relations of the Norman conquest, so all comparative or contrastive linguistics necessarily receive their data from situations of transfer and translation. A theory of translational equivalence has very little to learn from the Apollonian lines drawn by linguists or their structuralist acolytes. Equivalence is not symmetrical.
Marianne Lederer goes shopping:
"A word out of context, on the level of the tongue and thus not yet in a message, is like a 50-franc note which has yet to be materialised as something bought. As long as the 50-franc note is not spent, it is potentially groceries, books, a train-ticket or whatever. But its actual materialisation can only be one of these virtualities." (Lederer, in Seleskovitch and Lederer 1984, 24)
Although the analogy basically concords with Saussure's example of the chesspiece, it has several added peculiarities. First, it interestingly refers to a fixed quantity of signifying material: Is it really so important to specify that this is a 50-franc note, or that a TT has two or ten pages? Second, the value of the note is not its potential use as a material object, but what it can be exchanged for. Linguistic value is thus seen as a kind of economic exchange value, although Lederer does not say where one should look for a market where words can be exchanged for their referents, nor exactly how to think about materialisation when, in the place of translation, words are exchanged for words, and francs are exchanged for pounds, dollars or pesetas. The analogy is on the right track, but more thought is needed.
Value is an economic term
Scant attention has been paid to the fact that Saussure's uses of the term "value" - and indeed his fundamental distinction between synchronic and diachronic linguistics - were developed from analogies with economics, or more precisely from comparisons with the most prestigious social sciences of his day, political economy and economic history:
"Here [in linguistics] as in political economy we are confronted with the notion of value; both sciences are concerned with a system for equating things of different orders - labour and wages in one, and a signified and a signifier in the other." (79)
According to Saussure, labour is to wages what the signified is to the signifier. But are these things of different orders really being "equated"? An economist who equated the value of wages with the value of labour would not get very far when trying to explain profits or capitalism. Does linguistics get very far if signifiers are just equated with signifieds? I suspect not, at least not with respect to phenomena of variability and dynamic change. Nor can translation studies make many advances while TTs are simply equated with STs. But Saussure takes up the problem in a second comparison:
"A value - so long as it is somehow rooted in things and in their natural relations, as happens with economics (the value of a plot of land, for instance, is related to its productivity) - can to some extent be traced in time [...]. Its link with things gives it, perforce, a natural basis, and the judgements that are based on such values are therefore never completely arbitrary; their variability is limited. But we have just seen that natural data have no place in linguistics." (80; italics mine)
This is a strange commentary. Here we see that the main point of comparison is the peculiar way nature appears to guarantee the equations of economics. Saussure seems to believe that economic value is determined by a commodity's "natural" embodiment of uses. Then he correctly does all he can to reject this naturalist basis from his linguistics. However, the commentary is strange because no economist who had read Adam Smith would have confused value with this natural basis. In fact, most economists would have agreed with Saussure's basic arguments against natural data.
Since the problem of natural value continues to haunt translation studies (Nida's definition, for instance, refers to "the closest natural equivalent"), it is worth considering what economics really has to say about the matter. After all, economists have been discussing these problems for centuries. Perhaps they can help us avoid a few elementary confusions. Here is David Ricardo giving textbook examples in 1812:
"Water and air are abundantly useful; they are indeed indispensable to existence, yet, under ordinary circumstances, nothing can be obtained in exchange for them. Gold, on the contrary, though of little use compared with air or water, will exchange for a great quantity of other goods. Utility then is not the measure of exchangeable value, although it is absolutely essential to it." (1812, 1-2).
The role of natural data here is clearly limited to use values. To say that an object is useful is to say that it can be exchanged. But natural use value itself has no strongly quantitative relation to actual exchange value. It is a function belonging to a space prior to properly economic activity. In fact, after the moment of necessary recognition, utility is of little interest to economics. The real value to be explained is that pertaining to exchange.
How can this distinction be applied to linguistics and translation? Surely use value is limited to recognising, for example, that "mouton" has utility in French and "sheep" has utility in English, or that a 50-franc note can be used in France but not in Britain. That is, there are certain separate spaces within which each term has utility. But mere use can tell us nothing about the actual value these terms might have when they enter a mutual space, when they exist at a point of contact between the two domains concerned, as when Anglo-Saxon meat is served to the Norman master or Ms Lederer tries to go shopping in London.
What does it mean to say, as do communicative and contextual semantics, that meaning is use, or even that meaning is use within certain frames and scenes? Surely all that is being said is that the space pertinent to some kinds of use is smaller than that of entire tongues. But it is still no more than a space. It is not a point of contact or exchange. Within the natural spaces concerned, only water can be used as water, only mutton as mutton, only francs as francs. Use-value theories of meaning thus do not really rise above the identity equations underlying Saussurean mutual exclusion. All they do is accord a term a domain or series of possible domains (as with the chesspiece). Such theories can only say that a T1 is of value because its use does not correspond to T2, T3...Tn. Which is simply to say that it exists naturally and cannot be equivalent to any other term. Obviously, an equivalence-based definition of translation studies can have no place for such theories.
In order to talk about value as something more than a sterile identity equation of space with use, we must find another way in which T1 can be related to the series T2, T3...Tn. It is not enough to rely on simple comparison and mutual exclusion. The relation must become more dynamic.
An instance of how this can be done might be to look up the dictionary definition of a term, then the definitions of the defining terms, and so on until, according to certain theories, the exercise will exhaust the entire dictionary and take so long that the tongue itself will have changed, the dictionary will have to be rewritten and the process should begin again. The series of terms generated by such semiosis will concern not passive comparison between areas of utility, but active interrelations of exchange.
Interestingly enough, Jakobson described this same process in the following terms: "The meaning of any linguistic sign is its translation into some further, alternative sign" (1959). Which is to say that translation generates series of exchange values.
It is then in the analysis of exchange, not use, that economics provides us with a model of equivalence able to avoid making translation impossible. Rather than consider Saussure's positional problems of natural mutton and natural sheep happily separated by the Channel, translation studies should enter the active situational problems of Anglo-Saxon servants and Norman masters (or competing British and French farmers) who have to negotiate and exchange mutually recognised values before getting down to the undeniably useful business of eating sheep, mutton, or the non-equated leftovers.
Equivalence is an economic term
There is undoubtedly a certain ideological underpinning to approaches that see translation as a mode of relation between social systems, especially when they stress twentieth-century use-value theories of "equivalent effects". Our century has seen sociology overtake speculation; the law of the market has undone philological illusion; and at least for Newmark, translational functionalism has accompanied the triumph of base consumerism (1981, 38). It would seem that formal nineteenth-century exchanges have sunk to the level of economic expediency. Against this background, there is a certain perverse pleasure to be gained from citing a nineteenth-century economist in order to explain why equivalence does not really concern use values.
An incisive analogy:
Marx's analysis of commodities is simple enough: "quantity x of commodity A" = "quantity y of commodity B"; or, in terms appropriate to the first International as a meeting of mostly Jewish tailors, "20 yards of linen = 1 coat". Here is the commentary:
"The linen expresses its value in the coat; the coat serves as the material in which that value is expressed. The former plays an active role, the latter a passive role. The value of the first commodity is represented as relative value, or appears in relative form. The second commodity functions as equivalent, or appears in equivalent form." (1867, I, 63; italics mine)
Marx stresses that no one commodity can assume both the relative and equivalent forms of value at the same time, since the value of the linen is only recognised "when it comes into a communicative situation with the coat" ("sobald sie in Umgang mit andrer Ware, dem Rock, tritt"). The coat is thus that which "brings value" ("der Träger von Wert"). There is no question of this being an identity relationship. Nor is there any question of this kind of economic value being, as Saussure had supposed, "somehow rooted in in things and in their natural relations". The nature of the materials involved is unimportant to their expressions of value. Value is here purely a result of the relationship between the commodities and the communicative but subjectless place ("Umgang") in which this relationship is possible. The relationship is moreover explicitly situational and may be repeated with respect to numerous other commodities. The coat may be equivalent to 20 yards of linen this week and 15 yards next week; to five umbrellas, three pairs of shoes or two pairs of trousers, or any combination of these quantities and qualities, within the spatial and temporal limits of the markets that the coat can reach. The coat can thus potentially enter into a translational series with any number of other items, each time occupying the equivalent position but never having its nature reduced to that of a definitive and obligatory equivalent of any other item. In this sense, equivalence depends only on what is offered, negotiated and accepted in the exchange situation; it is decided each time by what the seller and the buyer situationally believe to be of value and worth exchanging. It is never an exclusive relationship between the natural qualities of linen and coats; there is no suggestion of cause and effect such that whenever a seller offers twenty yards of linen a buyer should exchange a coat.
In short, according to this model, each relation of equivalence is a transitory convention, a momentary link in process of potentially endless exchange. More critically, it is a fiction, a lie, a belief-structure necessary for the workings of economies and the survival of societies.
If we now write "transferred text" (Y) and "translated text" (TT) in the place of "linen" and "coat" - not entirely metaphorically, since some texts are indeed bought and sold, and weaving can be as textual as it is textile - , certain clear correspondences appear between the model of exchange and the definitions of translation cited at the beginning of this chapter. The relationships are in both cases one-way and non-reciprocal equivalence is in both cases expressed in only the latter of the two positions available; and the very possibility of this relationship - the very possibility of an equivalent form - depends on an apparently subjectless locus in which both sides of the relation can be at once mutually present and mutually distinct.
That is, equivalence can be defined in terms of exchange value, expressed as a relationship between texts (TT:Y) and determined in the specific locus of the translator as a silent trader. This is what was being said but not heard.
Equivalence is not a natural relation between systems
The suggestion that equivalence-based definitions of translation unwittingly define their object in terms of simple exchange could justify common usages of the word "equivalence", but it by no means justifies all that is said by the contemporary theories incorporating these definitions.
Most notions of dynamic or functional equivalence are based on a correspondence between use values which are rumoured to exist in distinct languages, societies or cultures, understood as independent systems. Translation is seen as a matching of one use or function with another, rather than as a productive function in itself. The economic definition of equivalence, on the other hand, enables us to focus on value as something manifested through the translation of texts in situations of contact between interrelated cultures. Equivalence is to be understood as emerging from active interrelations, determined by what translators actually do, and not by abstract comparisons between falsely discrete and passive systems. The methodological importance of this point is rarely appreciated.
Since translation is an interrelational activity, it is slightly contradictory to suppose that it can be analysed in terms of non-relational categories. And yet this is precisely the kind of contradiction found in overtly Marxist approaches to translation. When Otto Kade states that untranslatability is the result of the non-correspondence between "two historically developed societies" (in Koller 1979, 156), the fact that no two societies have developed in the same way would logically suggest that translatability and thus equivalence are impossible. Dialectic acrobatics apart, there is clearly something fundamentally wrong with this supposedly Marxist but in fact eminently Saussurean mode of argument.
Marx's critique of use value is perhaps more interesting than the twentieth-century abstractions that have followed him. He saw exchange not as a capitalist plot, but as a result of concrete intercultural communication:
"Just as a Manchester family of factory workers, where the children stand in the exchange relation to their parents and pay them room and board, does not represent the traditional economic organisation of the family, so is the system of modern private exchange not the spontaneous economy of societies. Exchange begins not between individuals within a community, but rather at the point where communities end - at their boundary, at the point of contact between different communities." (1857-58, 882)
The importance of the frontier curiously reappeared when Marx was searching for analogies between money and language, considered difficult because "ideas do not exist independently of language". However:
"Ideas which first have to be translated out of their mother tongue into a foreign language in order to circulate, in order to become exchangeable, offer a somewhat better analogy; but the analogy then lies not in language, but in the foreignness of language." (1857-58, 163)
Exchange value is thus opposed to the idealism of "naturally
arisen communal property", just as translation can be opposed to "naturally
arisen common languages". As cultures become increasingly interrelated,
the foreignness that appears on the frontier tends to overtake relationships
based on the false homogeneity of traditionally discrete systems. Exchange
overtakes use.
Equivalence has become unfashionable
One of the paradoxical effects of the historical increase in intercultural communications is that, through the rise of non-linguistic cultural and historical studies, there is nowadays declining interest in translational equivalence. As it becomes more and more obvious that equivalence is not a natural relation between systems, writers on translation are becoming increasingly inclined to act as if there were no such thing as equivalence at all, throwing out the proverbial baby with the bath-water. The result is that the theorists usefully brought together through the many citations in Koller (1979) can be identified and historically distanced as believers in what is now a fairly reactionary notion of equivalence.
There are at least two good reasons why restrictive ideas of equivalence should have become unpopular.
First, historians of translation are showing that many equivalence-based theories unnecessarily exclude much of the richness of the past. Jeanette Beer correctly points out that, for medieval translators,
"...structural equivalence between source and translation was not of prime importance. By the criterion of appropriateness to target audience a treatise properly could become poetry, epic became romance, and sermons drama - or vice versa! Such dramatic changes in form serve as irritants to those modern theorists who, for the sake of anachronistic criteria, categorise a millennium of translative vitality as one thousand years of non-translation." (1989, 2)
Second, the sheer quantities of weakly authored material nowadays to be translated have brought about significant changes in the professional tasks of many trained translators, who are writing summaries, providing linguistic consultation services, producing new texts for new readers, or processing computer-generated translations. Strict quantitative equivalence to Y is often no longer considered as important as the efficiency of TT, to be assessed as a new text designed to serve a new purpose.
Both these arguments are justified and fruitful. But do they mean that equivalence itself has disappeared? Is not the modern or medieval priority accorded to "appropriateness to target audience" - amongst many other operative criteria - in itself a valid basis for equivalence on a particular level? Translations written as edited reports for a specific reader - as suggested by Mossop (1983) and detailed by Gouadec (1989, 22-29) - do not necessarily break with equivalence, given that what is exchanged, what the specific reader ideally wants and receives, is ultimately a representation of that part of Y which is considered to be of value in the particular exchange situation concerned. Equivalence has thus by no means disappeared. It is still what happens, on one level or another, whenever a translated text is received as if it were a merely transferred text; it is still there whenever translation is distinguished from non-translation; it is still implicit in the way a TT signifies its antecedents, even in cases of pseudotranslations, where no antecedents exist.
What has changed is not so much the way TT represents Y, but the way this representational relation was once considered a natural fact. As the model of exchange makes clear, equivalence is artificial, fictive, something that has to be produced on the level of translation itself. But it must be produced. Whether one likes it or not, just as the exchange value of Ms Lederer's banknote is a question of communal belief necessary to keep economic relations alive (economists know that most of the note represents non-existent wealth), so translational fictions of equivalence remain essential for the maintenance of countless acts of intercultural communication. Their negation or denial should not be thought a simple task.
Equivalence thus neither descends from above nor blossoms from the soil. It is a fiction without natural correlative beyond the communication situation. Yet naturalist assumptions continue to obfuscate its role as an active mode of interrelation. As José Lambert (1978) has remarked, the kinds of equivalence presented in formal theory ("functional", "communicative", "semantic", etc.) tend not to correspond to the notions of equivalence implicit in non-normative descriptive studies. The large-scale paradigms based on precarious abstractions from use habitually fail to perceive that they themselves depend on the equivalence which can only be found through exchange, through translation as a communicative process. They thus falsely convert translators into "equivalence seekers" (Mossop 1983), ignoring that, as Koller puts it, translating itself is the production ("Herstellen") of equivalence and translators are ultimately the people who say what should or should not be proposed to the receiver as an equivalent (1979, 186-193).
But if definitions of translation can consistently omit
all reference to the person or thing determining equivalence, exactly who
or what is the translator?