We emphasise to our clients that quality translations require a thorough review process. Every document we translate undergoes close scrutiny – not once, but a minimum of three times and as many as nine or ten times. We begin the review process with a draft translation performed by native tongue translators, then the documents are reviewed by editors and proofreaders (linguistic reviewers).
The track changes allow the team members and the project manager to see every change made to the text – additions, omissions or replacements. Each track change is marked with a time stamp and a unique color that is automatically assigned to each linguist with their initials or computer name.
Constructive criticism, citing references and resources is used, by inserting comments, the team of linguists and the project manager can discuss the implications and reasoning behind each change made. This method, along with the tools in MS Word, allows the team and project manager to observe and compare documents throughout their evolution.
Over the years, our project managers have compiled a bunch of questions typically asked by clients about quality control, ‘back translation,’ our translation process, certification and reliability. Here I take a look at the issues raised by ‘Back translations’, one element of quality control, and provide the answers we typically give.
Q: Is back translation an illogical waste of money and a time-consuming method of quality control that does not produce what you are intending it to accomplish?
A: Almost any translation has many equally correct ways that it can be translated into another language; similarly, a back translation also has many equally correct ways that it can be translated back into the source language. This makes it hard to believe that any back translation would be a reliable method of verifying the accuracy of a translation.
Q: What does back translation provide qualitatively?
A: Traditionally, it has been considered as a way to validate, approve, assure, guarantee, or prove that the translation is accurate, follows the original source text, and reflects their ideas accurately. We believe that back translation provides absolutely nothing syntactically or semantically, about the translation and is unreliable as an effective quality control procedure in translation.
Q: If quality control of the document translation process is good, is there a need for back translation?
A: We don’t believe there is if you use a translation method that is thorough. Our work must be able to stand up to the most fastidious reviews.
Q: Who would support the use back translation if it takes more time and costs more money?
A: For certain organisations, there is no better way to validate a translation because they don’t understand the language (linguistics) and have no other viable resources to determine the accuracy of the work. Certain organizations dealing with pharmaceutical case studies, mental health evaluations, or instructions for medical equipment use back translation because they have no reason to believe it does not work.
Q: What are the inherent problems with back translation?
A: There are several translation companies and independent linguists that don't support the concept of back translation as it is usually just a way to get into an argument about syntax and style choice (and a number of other linguistic factors) based upon varied cultural and educational backgrounds, or one of a million other variables another linguist will have to say about how you can improve a translation via the back translation process. Keep in mind, “Translation is an art, not a science.” Unlike science, where a certain scenario, with the same elements and same factors will produce consistent results – translation involves a human factor that adds a unique variable into the scenario. Another typical problem with back translations is that it sometimes involves the use of an in-house bi-lingual staff of the client rather than professional linguists. A client may consider the opinion of their bi-lingual employee as a professional translator and add even more variables into the quality control method supported by back translation.
Q: Consider this translation issue: if the change doesn't improve the sentence or phrase then how necessary is it?
A: If the end result is the same meaning to the reader, the only other reasons to change the translation would be so something sounds nicer than another version, works appropriately given the context with other translations/source text or has a better idiomatic fit. All of these types of linguistic decisions are personally derived at by phonetic, aesthetic or cultural perspectives and do not provide a quantifiable measure of improvement. We all understand that these personal characteristics differ from one person to the next. It is quite typical of a client, using a bi-lingual employee to perform the back translation, to be persuaded by their employee’s unprofessional, personalised opinions about a translation that are unsubstantiated. This not only increases the amount of time spent on the translation, but also the cost for production.
Q: At what point is a translation complete when using back translation for quality control?
A: After the translation from the source language into the target language is completed, then you perform the back translation. But, the process is not over yet… You now have a translation into the target language and document translated back in to the source that allows you to identify differences from the original source file with the new back translation into the source. If you compare documents, you will likely find many differences. At this point, you have to wonder when the translation will be complete. You have two versions of the source and one translation into the target that you don’t know what to do with due to unknown differences. In your attempts to refine the quality of your translations by using back translations, you have been forced to fix your translation with a verbatim translation of your source text. This not only defeats the purpose of using editors and proofreaders in the first place, but also leaves you with a dry and sometime meaningless transliteration of your source text.
Q: What types of text seem suitable for back translation?
A: Technical documents, like MSDS or scientific formulas seem to be suitable for back translation because the source text is usually written by Engineers or Scientist and is less likely to include humour, colloquial expressions or complex literary statements. If the content seems like it is written by a computer, then it is easier to obtain a verbatim translation in the target language and a back translation would be helpful to verify the content. In MSDS or scientific formulas, back translation won’t provide any verification for syntax or semantics and at best provides synonyms for words taken out of context.
By Scott M. Crystal,
American Translation Partners, Inc.,
Raynham, MA, U.S.A.