| Comment | John,
I agree with AGB. Translation is inherently an inexact process, and there is simply no way to map the words of one language precisely into those of another, for many reasons.
As AGB has pointed out, even figuring out what an original speaker "means" by something is a complicated question, and whole subsections of linguistics are devoted to this question, which is one that interests computer scientists involved in artificial intelligence a great deal as well. Without even getting into the issue of ambiguous words, you have surely been in a situation where different (monolingual) listeners could interpret the same sentence quite differently.
Another way to state the inexactness of translation, is that there is to "prove" that one translation is better than another, other than by simply taking a poll of translators you trust and asking them to vote on it. That's another way of saying, there is no right answer.
Your comment about using "frequency" is one attribute that can indeed be useful. For example, the French say "patrimoine" and "deonotologie" all the time and everybody knows what they are, but if you look up the translation and use "patrimony" and "deontology" in English, nobody will know what you are talking about.
But frequency is by no means the only factor. There is also formality/informality, slang/normal, offensive speech/normal, word used for centuries/neologism, trendiness/normal, and so on.
There's virtually no end to it, and in the end you can only do the best you can.
Sometimes a translator manages to do something magical and hit precisely the right sound and comes up with something that ought to get them awarded a "brilliancy prize" as they do in chess, and you can only wonder and applaud, because there is no way to teach or learn that kind of thing in my opinion, but I'll leave it at this point to the professional translators (which I am not) to put in their two cents. |
|---|
