Evaluating Translation

Evaluating Translation | One Hour Translation

An objective and useful evaluation of a translation project is surprisingly difficult to accomplish for a variety of reasons.

If you think about it for a moment, it becomes very clear that translation is very difficult to evaluate objectively. This goes beyond the sour grapes any translator of any experience might have when their work is challenged, especially by someone who isn’t trained in the work and thus doesn’t understand the true scope of translation, and centres on the very nature of the work, because the main difficulty in evaluating language translation is the fact that one side of the equation is always a second language.

That means for anyone who is not bilingual (at least) in the specific language set being used, evaluation becomes impossible. How can you decide if a translation is done well if you can’t even read half the text in question? And even if you are fluent in both languages, you must also have the experience and training of a professional translator to truly evaluate the work that’s been done. And even then, you must be familiar with the specific requirements of the project in order to be able to judge whether the translation is a success or not.

Back to Basics

Let’s get back to basics, though, for a moment. We need to establish a baseline. In an ideal world where everyone evaluating a translation project does have the experience and training necessary, what are the aspects of a project that should be judged in an evaluation? I’d argue there are just two:

  • Accuracy. The target text must faithfully convey the concepts, facts, and asides of the original, without distortion or unnecessary flourish, as closely as possible. A simple example to use is when the source uses the numeral 1, and a translator decides to use a word like “single” or “solitary.” These are not incorrect (depending on context) but they are also not precisely what was written, are they?
  • Appropriateness. The translation must also convey the correct tone and style. Meaning isn’t just facts and figures, nouns and grammar. Meaning can also be conveyed effectively via the tone and style used in the written work, and this must be preserved for the translation to be judged a success.

Objective Vs. Subjective

Here you begin to see the true problem: There’s a lot of subjectivity there.

Many folks would look at my numeral Vs. “solitary” example and shrug – it would seem perfectly fine to them. Or they might argue that only if the context made the use of the numeral essential would this be a concern. Valid points, but they also speak to the underlying problem: There is rarely a clear consensus concerning what’s correct and appropriate in a translation. People bring their own ideas, and a translation reviewed by ten people – even professional translators – would like get 10 different “grades.”

There isn’t much to be done here. All awe can do is our best to evaluate each other’s work, and there’s value in evaluation even if it is difficult to standardise it, I think.