In the translation industry, there are usually unequivocal sentiments surrounding the efficacy of different processes and strategies: Machine Translation is helpful but it’s not perfect; you should always use professional, certified linguists; accurate translations require linguistic validation.
But when it comes to back translations, there doesn’t seem to be a strong consensus.
Back translation refers to translating a document from source to target language, and then taking the target translation and translating it back to the original source text. The idea is that when the target phrase is translated “back” to its source form, it should match the original text 100 percent to ensure accuracy.
Clients often want back translations as a way of knowing that everything was localized correctly and not translated literally. The big issue is determining that there was no loss or change of meaning during the process.
Although back translations are sometimes required, there are many who believe the method is ineffective. Some argue that just because a back translation doesn’t match the original doesn’t necessarily mean the translation was done incorrectly.
Below we’ll explore both sides of the issue, examining the reasoning for and against back translations.
Back Translation and Reconciliation
Back translations are usually used in life sciences or other highly-regulated fields. Clinical trials, for example, are one area in which back translations are used to increase the accuracy of translations. Back translations are employed when the information being translated is critical and accuracy is extremely important.
A separate translator is recruited to complete back translations, and during what’s called the “reconciliation process,” will compare the back translation to the original text to see if the two match up.
Back translation acts as another quality step in the translation process, but not everyone agrees on the method’s efficacy.
Issues With Back Translation
Some compare the back translation process with doing a math problem backwards. The problem with this analogy is that math problems don’t contain the nuance and ambiguity that translations do.
It’s unlikely a back translation will be identical to the source text because different translators use different styles. Each linguist will have his or her own way of choosing which words to use in a certain context. And if a back translation contains different words, but has the same meaning, clients will often think it’s incorrect even though that’s not always the case.
A 2008 edition of the American Translators Association Chronicle touches on the subject of varying word choice and identical meaning. The Chronicle’s article refers to “differences that matter” in back translations and those that don’t.
Another downside to back translations is that the method requires extra resources and cost. The extra step requires finding a third-party translator to complete the back translation and compensating him or her for doing so. On top of this, a back translator could make mistakes, and muddle up the process even further.
Explore Your Options
Unless a back translation is required due to regulatory issues, clients can likely get the same validation results by going through the normal quality assurance processes. This means employing language validation and in-country review after an initial translation is completed.
In addition, a Language Quality Assessment could also be used. During the LQA process, a separate translator scores each paragraph of text, citing minor, major and critical errors within a translation.
However, this doesn’t mean that back translations aren’t valuable. Any language validation is good, and the more options you have for checking a translation, the better.
With that said, it’s important to be mindful of the shortcomings and limitations of any validation method, and to employ multiple review stages to create a successful target translation. No matter what method(s) you use for language validation, strong translators and project management teams are at the core of accurate localization.