Translations should primarily be accurate. Advocating for a word (or advocating for a thought) when the resultant text is agreed by all translators to be a less accurate translation is inappropriate.
But if we agree on accuracy, why prefer word level translations over sentence level translations: formal equivalence over dynamic equivalence?
One of the reasons is that I think formal equivalence ends up with an accurate translation at the level of meaning, ie. the clause or sentence and potentially delivers less interpretation. The dynamic approach asks what the passage means, an eminently important question—that all translators should ask of each passage—but one that allows more leeway in interpretation. Therefore the formal translator when asking the same question (which he should) is more constrained in how he can translate. While a dynamic translation may be accurate in many places, it is probable it is inaccurate in more places than a formal translation.
There are several other reasons to prefer a formal translation. One is that it better allows for allusions in reading Scripture. Phrases that use identical words in the original are more likely to do so in a formal translation. The reader recalls other passages more easily, or sees them as more obviously related.
Similar to this is plays on words. Repetitions of words, including the use of related nouns and verbs can modify the readers interpretation.
Some passages are difficult to interpret, reproducing this difficulty in a translation means that an interpretation does not need to be settled on, the ambiguity can remain for the reader who is forced to think though the options. Ambiguity in the original may be intentional, it may be that the passage is supposed to have a double meaning. A word for word translation I think is more likely to retain multiple meanings, especially in the situation where the translator is unaware of more than one meaning.
While I advocate this translational philosophy, I think there is unnecessary, or inappropriate adherence to elements of this by translators who favour formal translation.
The structure of the destination language needs to be considered and respected. I find some phrases in formal translations tortuous. This is not to say every passage should be simple to understand. If it is difficult in the Hebrew or Greek, then the English may be appropriately difficult. But there is no need to make simple phraseology complex. The 2 improvements that many formal translations could make are
- not try and preserve word order; and
- use extra words.
The context of a word in the destination language also needs to be considered. Comrade was an useful English term for a friend or helper in a common cause but now the word has communist overtones and these overtones may not exist in the source language.
A word for word translation suggests using a single word in one language for a single word in another. But it is really a word-unit. At times one needs to use several words in the destination language for a single word in the source language. This is not a compromise toward dynamic translation: word → word-unit → clause. A word-unit (in the sense I am using it) is a single concept requiring more than one word. For example ram is a single word for a male adult sheep; ewe for female. There is no single English word for female camel but she-camel is a word-unit that could be used to translate a single source word for a female camel.