When a "literal" translation is not accurate (from Jennings in Goma)

In Scripture translation, we're aiming at four goals, which are sometimes in tension. We want the translation to be: 1) accurate, 2) clear and easy to understand, 3) natural-sounding rather than stilted or "foreign" and 4) acceptable to users (especially the local church leaders).

Of these, my first responsibility as a consultant is to try to ensure accuracy. If the translation don't reflect the meaning of the original text, then it doesn't matter how clear or natural it is, it isn't a good translation. So I need to be sure that nothing has been added, omitted or changed.

But accurate is not the same as literal. Sometimes a translation can be very literal -- even word-for-word -- and yet be completely inaccurate. That's because meaning doesn't just come from the text, from the words. It comes from the text plus the context: the linguistic, cultural and historical context of the original language and how it fits -- or doesn't fit -- with the context of the target language.


We've run across two examples of this this week:
(A) "O foolish Galatians! Who has bewitched you?" (Gal. 3.1 ESV) - It is quite easy to translate this literally into the language we're working with. As in many other African cultures, they have a strong history of African Traditional Religion, including witchcraft, even among those who also call themselves Christians. So there is a verb that means "to bewitch" or "to cast a spell over." No problem.

The problem is the context. In the Greek text, "Who has bewitched you?" is a rhetorical question. It isn't a "real" question, it isn't seeking an answer. In this case, it's a way to rebuke the Galatian readers. In order for that rhetorical question to "work", it needs to be understood in the way it was intended. That means that 1) the target language must also use rhetorical questions to rebuke, and 2) it should be clear in the target language that the speaker isn't looking for an answer.

In the context of this language and culture, a question such as "Who has bewitched so-and-so?" would be quite common, but it would be a real question. When something bad happens -- death, illness, financial setback -- it is often blamed on witchcraft, and it is very important to identify the person who did it. So this question could easily be taken as a real question, as Paul trying to find out who "bewitched" the Galatians. Translating it literally could give a completely false meaning.

(B) "For it is written that Abraham had two sons, one by a slave woman and one by a free woman. (Galatians 4.23, ESV) This, too, was very easy to translate literally. The problem was the word "free". A "free woman" in this language would hint at a woman who was "free" to leave her husband or to sleep with whatever man she chose. So the literal translation gave quite an inaccurate meaning. (We went with "woman who was not a slave".)

There are certainly plenty of other places where potential misunderstanding could take place which could distort the meaning. This is why, in addition to translation checking with a consultant, teams also do translation testing -- seeing how people (not involved in the translation) understand what's been translated.