Dear Sophie Lechauguette,

You have certainly broached up a very interesting issue. Indubitably, backtranslation is a legitimate procedure for making a rough estimate of the quality of a translated work. In Julian House's model of translation on quality assessment , it has been used as a preliminary estimation of translation quality. However, as you have rightly observed, backtranslation does not provide a clear picture showing the success of translator in conveying the intended message by the author of the original text. In fact, after applying backtranslation, we need to take a further step by using textual analysis considering the context of situation variables (e.g., field, tenor, and mode) and perform lexical, textual, syntactic , thematic, and iconic analyses and by performing a comparative analysis of the translated work  we have to determine  any possible mismatches between the source and target texts. Then, we have to repeat the procedure at the generic level as well. On this basis, you are absolutely right to express professional concern over the under -representativeness of backtranslation as an assessment procedure even though it is quite a common practice to use it for translation evaluation nowadays.

Best regards,

R. Biria 

Back translation is the best means of self evaluation by the translator. In fact, it is through back translation that a translator will be able to know if he/she has done a good job or not. Even though it is generally accepted that back translation can hardly reproduce the original, it is a means of evaluating how close the translation has succeeded in coming close to the original message of the author. The further the translation is from the original text, the more inappropriate it is. 

Dear Sophie, Dear Awa,

I doubt that the same translator is able to perform/use a back translation as a verifying tool for the accuracy of his/her initial translation. Of course, in a way, she does that, in the process of producing the initial translation, by evaluating each option she has concerning each unity she must render in the target-language. But upon larger portions of texts, the back translation would be vitiated by the subjectivity of the translator who remains the same.

Hmmm?

Adina

Dear Adina, Awa and Reza

Thank you for your comments.

Adina, you are of course right but in medical and other technical fields, the backtranslation is carried out by other translators, not the author of the target text.

I still find it strange to rely on a backtranslation to assess a translation. I am told the idea is not to obtain the original sentence or text but to highlight discrepancies in meaning by comparing the original and the backtranslation.

However to me, this means the thinking is mostly done on the source language. I am of course very ignorant of those fields but I would think that it would be better to do the critical thinking on the target text, by testing it and the effects it produces on the intended recipient. Especially in the medical field, I would think you need to listen to patients reactions and answers. Current protocol do have such a test phase, but it happens after the backtranslation stage of the validation protocole.

I remember once reading an article (I was not researching then and do not have the reference, sorry) about the difficulties of translating a campaign against AIDS for countries where it is culturally unthinkable to discuss sexuality the way it is done in the West. So the accepted translation was very different from the orignal phrasing, but conveyed the same idea in culturally acceptable terms.

What of backtranslation then ?

I think the idea of back translation is very okay as a means of vetting the target text. It can never reproduce the exact original text but it will surely depict how close or far away the work is from the original text. And as rightly noted, this is best for very technical oriented translations like Medicine, Engineering..etc where terminologies are almost stagnant in usages and meanings, but not in literary works.

As for the way such works are vetted and accepted or rejected in the West and America, I do not agree to their ways. It is mostly based on Nida's Dynamic Theory of translation. This theory have it that the target language and audience is of more importance compared to the original text. Therefore, the cultural implication of the target audience is considered while that of the Original text is not. Finally, the reaction of the receptor of the target text is the far goal of the whole translation. This sounds sweet and cool, but the fact remains that the translator, in this context, is just like an author and no more a translator. This is the current leading trend in the world of translation in the America and West. However, Peter Newmark disagrees with this and I think his opinion is very weighty and most times more accurate here.

Since translation is taking a very different turn in the contemporary world, I am beginning to think that one needs to understand the environment he or she is translating for and how they perceive translation. This will guide on the best approach in handling the translation work.

I only use (machine) translation and back translation as a technique to gauge how good my original English is!

I examine the places where my back translation failed to convey the original sense, and see how I can improve my original English. After all, if a machine stumbles over my original text, then an English reader may do so also.

Google translate--> Je n'utilise que la traduction (machine) et la traduction en arrière comme technique pour évaluer la qualité de mon anglais original!

J'examine les endroits où ma traduction arrière n'a pas réussi à transmettre le sens original et à voir comment améliorer mon anglais original. Après tout, si une machine trébuche sur mon texte d'origine, un lecteur d'anglais peut aussi le faire.

Google back translate--> I only use translation (machine) and translation back as a technique to evaluate the quality of my original English!

I look at the places where my back translation failed to convey the original meaning and see how to improve my original English. After all, if a machine stumbles on my original text, an English reader can also do so. 

I can live with the original text. 

In the 80s and the 90s, backtranslation was used and known (but not meant, Brislin 1970 as example) to measure (with quantities) the translation quality and by extansion to guarantee it. For several years now, as you may read in the ITC guidelines and other recent test adaptation guidelines, this strategy used alone is not acknowledged in the literature anymore. Since it takes time to institutions to change and to knowledge to be transferred to practitioners, the strategy is still recommanded, but as a complementary way to discuss and to think about the translated material during its review. Ironically, this backtranslation "second life" represents a qualitative process in the heart of a still very quantitative inquiry which I would resumed in simple terms : How to (scientifically) validate the equivalence between a psychometric test, considered objective, and its translated version. Or, even more simply : How to validate a subjective process, i.e. the translation.

Dear all,

I somehow agree with Kennedy. The first time I came across the so-called 'backtranslation' was through machine translation, the purpose of which is to measure the translation quality of machine translators. Human translators vary in everything, which might explain why different high-quality translations of one original (particularly literary ones) are produced. Backtranslation as an evaluative technique, to me, depends on certain determinants (such as domain, genre, context, etc.) and is genre-specific. As a case in point, will it really work for poetry translation? I guess not, because target poetic translations of one source poem will certainly differ in style, phraseology, rhetoricity and so on, keeping, however, the spirit of the original. There are many Arabic poetic translations (brilliant ones) of Shakespeare's 'Shall I compare thee', the backtranslations of which will not bring forth the exact original. That is why a hot debate on the (un)translatability of poetry still exists.

Backtranslation, like machine translation is only a tool in the toolbox of a translator.

I have used backtranslation as a tool to improve the quality of my poems translated from Serbian and Afrikaans. Of course the translator of poetry has to decide which are the essential aspects of the original, and ensure their fidelity, in REWRITING the poem as a translator.