Wednesday, November 08, 2006
Untranslating
A trawl around the web reveals that the practice of back-translation, far from being the aberration that might be supposed, is widely regarded as a serious tool by what seem like serious people. Here for example:
If the comparison reveals a discrepancy, how do you know it was the original translator who erred and not the back-translator?
You could follow through of course and have the back-translation forward-translated once more by a third translator and then have that version compared with the first translation. But what if you get another mismatch - you now have three translators at loggerheads, do you call in a fourth, a fifth...?
More insidious is the opposite scenario, where the back-translation does match the original, as this may well be the result of both translators fouling up.
This is because the most common translator ineptidute is attempting to translate words individually rather than collectively and thus failing to see the wood for the trees. If the error-checking process then consists of a back-translator who has been enjoined to back-translate 'literally' - as seems to be SOP - such errors will be folded straight back in and the client left none the wiser but with the illusion that everything is fine.
For a glimpse of the back-translation technique in action see this letter from a client (consisting of a five-strong research team) to an English-to-Italian translator detailing their observations on his back-translated translation. The reviewers actually come across as rather thoughtful, having figured out for example that the problems they identified may have arisen only at the back-translation stage. But the inanity of the whole approach still throws up the likes of this:
There seems to be a fundamental misconception at play, the same one that underlies the whole daft notion of back-translation. The translator is conceived of as a sort of carrier-pigeon, a dumb creature endowed with a wondrous facility of which it has no understanding. Its function is to transport not to process the message.
The truth of course is that the translator is as much a participant in the communication process as author and reader. In order for the message to pass it first has to be decoded and re-encoded, it cannot be simply strapped to his leg.
There's plenty more in the same vein:
You have to wonder though why they didn't simply give him the dictionary in the first place, thereby avoiding the whole laborious exercise. Perhaps they didn't think he'd know how to use it.
Earlier post on the lowly station of translators.
Requirements for the Translation of Consent Documents (02/03)Or this:
It is now a requirement of the IRB that all consent document translations be back-translated. A back translation allows the investigator to examine the content, accuracy and clarity of the translated document.
Does an adequate translation into your local language already exist? If not, you will need to have it translated from English into your language, and then, independently, back-translated into English again. Comparison of the original and back-translated English versions will tell you if your translation is accurate.It will?
If the comparison reveals a discrepancy, how do you know it was the original translator who erred and not the back-translator?
You could follow through of course and have the back-translation forward-translated once more by a third translator and then have that version compared with the first translation. But what if you get another mismatch - you now have three translators at loggerheads, do you call in a fourth, a fifth...?
More insidious is the opposite scenario, where the back-translation does match the original, as this may well be the result of both translators fouling up.
This is because the most common translator ineptidute is attempting to translate words individually rather than collectively and thus failing to see the wood for the trees. If the error-checking process then consists of a back-translator who has been enjoined to back-translate 'literally' - as seems to be SOP - such errors will be folded straight back in and the client left none the wiser but with the illusion that everything is fine.
For a glimpse of the back-translation technique in action see this letter from a client (consisting of a five-strong research team) to an English-to-Italian translator detailing their observations on his back-translated translation. The reviewers actually come across as rather thoughtful, having figured out for example that the problems they identified may have arisen only at the back-translation stage. But the inanity of the whole approach still throws up the likes of this:
The use of “ragazzi”: The translation back into English renders this as “boys” where the original English was “kids”. The word “kids” and its singular form “kid” applies equally to males and females... The use of a gender specific reference here significantly distorts the meanings of the items in which this occurs. English grammar has lost most of its gender distinctions. This, I think, is not true of Italian. There may be no single word in Italian corresponding to “kids”. If “ragazzi” really does mean “boys”, then some further work on the Italian translation is necessary for items 39, 47, and 52.Ragazzi, boys, kids - words don't come any more simple and everyday than these. The reviewers apparently assume the translator to be so ignorant of basic vocabulary and grammar (not knowing English had 'lost most of its gender distinctions'?!) as to require their non-linguists' advice on ordinary meaning and usage in both languages. Yet they go on to compliment him on his good work and want to order more. How can this be?
There seems to be a fundamental misconception at play, the same one that underlies the whole daft notion of back-translation. The translator is conceived of as a sort of carrier-pigeon, a dumb creature endowed with a wondrous facility of which it has no understanding. Its function is to transport not to process the message.
The truth of course is that the translator is as much a participant in the communication process as author and reader. In order for the message to pass it first has to be decoded and re-encoded, it cannot be simply strapped to his leg.
There's plenty more in the same vein:
The use of “sconvolto”: The translation back into English rendered this as “dumbfounded (shocked)” where the original English was “upset”. In America “upset” has the connotation of emotional turmoil that may endure over a period of days or longer. In contrast, “dumbfounded” or “shocked” have the connotation of a sharp surprise lasting a brief time. Moreover, “dumbfounded” has the connotation of surprise, but with little or no emotional turmoil. Literally, “dumbfounded” means to be speechless with surprise. A look at our Italian-English dictionary suggests that the translation is likely correct ....Of course... an Italian-English dictionary - why didn't I think of that!", you can almost hear the translator exclaim.
You have to wonder though why they didn't simply give him the dictionary in the first place, thereby avoiding the whole laborious exercise. Perhaps they didn't think he'd know how to use it.
Earlier post on the lowly station of translators.
