- 1. Knowing the text and its context. This is the paramount requisite. No one will be able to translate a text she does not fully grasp.
2. Knowing the context in the target language. One cannot expect to be able to translate the English "rule" as nyāya without taking into account its inner-Sanskrit hues of meaning. Vice versa, one cannot translate Sanskrit philosophical terms without taking into account the complex meaning of the terms they choose to translate them in English (or French or German) philosophy.
This being said, many methodological questions are yet to be faced, e.g.,
- A. Suppose your source text is obscure. Should you reproduce its obscurity?
- B. Suppose your source text, though philosophical, uses also a lively metaphorical language. Should you reproduce its style?
These problems are much less urgent in the case of translations within the same cultural area, i.e., from French into English or vice versa. Thanks to many centuries of mutual influences, English and French writers share a common reservoir of metaphors and poetical devices, understandable to both readers. Furthermore, readers share a common background of references and will easily understand short hints, so that it is rarely the case that a text will be really obscure and that this marginal obscurity can be reproduced without emparing the global understanding.
By contrast, in the case of very different cultural areas, i.e., while translating Sanskrit into English or German into Sanskrit (my last endeavours), the residual obscurity might be massive. And this obscurity might have not meant to be such by the author, who relied on the background he shared with his readers. Thus, the translator needs to solve the problem and to convey the text the author wanted to convey. A common example: Suppose the author refers to a well-known verse of the Gītā by just quoting its beginning. His target reader would have automatically supplemented the rest. But this is not the case for contemporary readers, who need the additional help of the whole verse being reproduced in the translation (either in the main text or in a footnote).
The problem is even more complicated in the case of style. I am personally convinced that dead metaphors (e.g., when we speak of a computer's mouse, we are not thinking of the animal the word originally denotes) should not be translated. As for live metaphors, I would try to translate them only if they are relevant (e.g., when they convey a shade of meaning which might have been meant by the author). An English author speaking about loving surrender to God, e.g., might choose to use the metaphor of falling in love on purpose. And a translator might consider reproducing it (kāmabhāvaṃ nipatati?).
What do you think? How do you translate?
You can read a text of mine about the importance of point no. 2 on my Academia page, here.
Samira Nekooeeyan recently dedicated an interesting talk to the translation of metaphors in poetical texts. Her practical example was the translation of Shakespear into Persian, but she dealt in general with the problem of what to do with metaphors (one of the viable options being "ignore them"). You can read the abstract of the talk if you download the program of the conference, here.
I decided to write this post after an intriguing discussion (on "flos"/puṣpa) with Sudipta Munsi while we were translating Kant's Critique of Pure Reason. You can read the first step of this enterprise here.