Translation users need to be able to rely on translation. They need to be able to use the translation as a reliable basis for action, in the sense that if they take action on the belief that the translation gives them the kind of information they need about the original, that action will not fail because of the translation. And they need to be able to trust the translator to act in reliable ways, delivering reliable translations by deadlines, getting whatever help is needed to meet those deadlines, and being flexible and versatile in serving the userâs needs. Letâs look at these two aspects of translation reliability separately.
Textual reliability
A textâs reliability consists in the trust a user can place in it, or encourage others to place in it, as a representation or reproduction of the original. To put that differently, a textâs reliability consists in the userâs willingness to base future actions on an assumed relation between the original and the translation.
For example, if the translation is of a tender, the user is most likely the company to which the tender has been made. âReliabilityâ in this case would mean that the translation accurately represents the exact nature of the tender; what the company needs from the translation is a reliable basis for action, i.e., a rendition that meticulously details every aspect of the tender that is relevant to deciding whether to accept it. If the translation is done in-house, or if the client gives an agency or freelancer specific instructions, the translator may be in a position to summarize certain paragraphs of lesser importance, while doing painstakingly close readings of certain other paragraphs of key importance.
Or again, if the translation is of a literary classic, the user may be a teacher or student in a class that is reading and discussing the text. If the class is taught in a mother-tongue or world-literature department, âreliabilityâ may mean that the users agree to act as if the translation really were the original text. For this purpose a translation that reads as if it had originally been written in the target language will probably suffice. If the class is an upper-division or graduate course taught in a modern-language, comparative literature, or classics department, âreliabilityâ may mean that the translation follows the exact syntactic contours of the original, and thus helps students to read a difficult text in a foreign language. For this purpose, various âcribsâ or âinterlinearsâ are best â like those New Testament translations published for the benefit of seminary students of Greek who want to follow the original Greek text word for word, with the translation of each word printed directly under the word it renders.
Or if the translation is of advertising copy, the user may be the marketing department in the mother company or a local dealer, both of whom will presumably expect the translation âreliablyâ to sell products or services without making impossible or implausible or illegal claims; or it may be prospective customers, who may expect the translation to represent the product or service advertised reliably, in the sense that, if they should purchase one, they would not feel that the translation had misrepresented the actual service or product obtained.
As we saw earlier, this discussion of a textâs reliability is venturing into the territory traditionally called âaccuracyâ or âequivalenceâ or âfidelity.â These terms are in fact shorthand for a wide variety of reliabilities that govern the userâs external perspectives on translation. There are many different types of textual reliability; there is no single touchstone for a reliable translation, certainly no single simple formula for abstract semantic (let alone syntactic) âequivalenceâ that can be applied easily and unproblematically in every case. All that matters to the nontranslating user is that the translation be reliable in more or less the way s/he expects (sometimes unconsciously): accurate or effective or some combination of the two; painfully literal or easily readable in the target language or somewhere in the middle; reliable for her or his specific purposes.
A text that meets those demands will be called a âgoodâ or âsuccessfulâ translation, period, even if another user, with different expectations, might consider it bad or unsuccessful; a text considered a failure by some users, because it doesnât meet their reliability needs, might well be hailed as brilliant, innovative, sensitive, or highly accurate by others.
It is perhaps unfortunate, but probably inevitable, that the norms and standards appropriate for one group of users or use situations should be generalized to apply to all. Because some users demand literal translations, for example, the idea spreads that a translation that is not literal is no translation at all; and because some users demand semantic (sense-for-sense) equivalence, the idea spreads that a translation that charts its own semantic path is no translation at all.
Thus a free retelling of a childrenâs classic may be classified as an âadaptationâ rather than a translation; and an advertising translation that deviates strikingly from the original in order to have the desired impact on target readers or viewers (i.e., selling products or services) may be thought of as a ânew textâ rather than as an advertising translation.
Each translation user, limited to the perspective of her or his own situational needs, may quite casually fall into the belief that those needs arenât situational at all, indeed arenât her or his needs at all, but simply the nature of translation itself. All translation is thus-and-such â because this translation needs to be, and how different can different translations be? The fact that they can be very different indeed is often lost on users who believe their own expectations to be the same as everyone elseâs.
This mistaken belief is almost certainly the source of the quite widespread notion that âfidelity,â in the sense of an exact one-to-one correspondence between original and translation, is the only goal of translation. The notion arises when translation is thought of exclusively as a product or commodity (rather than as an activity or process), and when the reliability of that product is thought of narrowly in terms of exact correspondence between texts (rather than situated pragmatic reliability).
Reliably translated texts cover a wide range from the lightly edited to the substantially rewritten, with the âaccurateâ or âfaithfulâ translation somewhere in the middle; there is no room in the world of professional translation for the theoretical stance that only straight sense-for-sense translation is translation, therefore as a translator I should never be expected to edit, summarize, annotate, or re-create a text.
While some effort at user education is probably worthwhile, it is usually easier for translators simply to shift gears, find out (or figure out) what the user wants or needs or expects, and provide that â without attempting to enlighten the user about the variability and volatility of such expectations. Many times clientsâ demands are unreasonable, unrealistic, even impossible â as when the ...