A Discussion on the Intricacies of Crafting a Delicious Translation

[Asceai]

Originally Posted by Agilis

One thing I don't think most people who start translating plan for is the fact that editing can easily take 2-5x longer than it does to go through a script the first time. Unless you're really experienced and careful at translating, you really can't wrap up any length of text in just 2 passes, the mental stamina required is pretty high.

I'm hoping to be able to do exactly that. Does it help that I've played the game twice so far, and some sections multiple times, and thus have a pretty clear idea in my head of how everything flows?

That said, I'm going to 'edit' each chapter as I finish it, and use the betas as a way to get other people to fix up my shoddy work =P

For reference, my use of the terms 'translation' and 'editing' hopefully doesn't mean
translation: japanese -> barely comprehensible engrish
editing: barely comprehensible engrish -> english

Hopefully I won't have to change too much in the editing stage, since I'm trying to write something as close as possible to the final output, but we'll see. Like I said before, my Japanese is terrible (not that I'm saying my English is that great, but I don't really have any excuse for the latter =p) and I probably shouldn't even be doing this translation, but I don't think I'm doing too badly so far.

[zalas]

Originally Posted by Asceai

I'm hoping to be able to do exactly that. Does it help that I've played the game twice so far, and some sections multiple times, and thus have a pretty clear idea in my head of how everything flows?

Sure, it's a necessary, but not sufficient condition. It solves one of the problems below. The problems I find tend to occur with people starting out in translation are:

1. Not seeing the big picture
A lot of people end up using their translation as the first pass through a game. This means that you will end up with suboptimal wording now and funny inconsistencies later on. Furthermore, if you cannot play the game without sitting down and translating all the lines, it means you have enough problems with the original language that a translation of a full visual novel will really bog you down.

tl;dr solution: Play the game through first and fully understand it.

2. Not seeing the big picture Part II
People often try to be too literal when they first start out. This may be the result of the way their language instruction was structured, or it may be the result knowing 100% what the author wrote, but not 100% what the author. Translation is often a conversion of meaning first, structure second. Not seeing the big picture of the intended meaning will make for translations that don't really sound that great.

tl;dr solution: Make sure you understand the intent and not just the literal meaning of each line.

3. Not able to write in the target language
People often get the misconception that as long as they know two languages, they can translate. This is not so. In addition to being able to understand the source language, the translator must be an excellent writer in the target language. Otherwise, no matter how well one understood the source material, if one cannot convey it correctly in the target language, the translation will not be faithful.

tl;dr solution: Get better at writing.

All of these problems make for long editing sessions. Problem 1 often leads to massive amounts of search and replace at the end when things aren't consistent. Problem 2 and 3 lead to long rewording sessions. Problem 1 generally has a simple solution while the latter two require more experience in translation.

I've seen problem 3 often being handled by having a "translation" and "editing" pass by two different people. This is suboptimal, as without a brain meld, you either don't get enough information across from the "translator" to the "editor" or you spend way too much time in conversation. If it's the only thing you have, then fine, but don't treat it like it's some golden rule that must be followed.

Now, some people might say that it's impossible to find someone who's supremely fluent in one language and a great writer in another. In those cases, find a translator who knows the source language well enough to know when he doesn't understand something. For those situations, the translator can then consult a native speaker or someone more knowledgeable. This is much better than the "translator"->"editor" relationship, as less information is needed to flow from one person to another.

[Asceai]

Originally Posted by zalas

Now, some people might say that it's impossible to find someone who's supremely fluent in one language and a great writer in another. In those cases, find a translator who knows the source language well enough to know when he doesn't understand something. For those situations, the translator can then consult a native speaker or someone more knowledgeable. This is much better than the "translator"->"editor" relationship, as less information is needed to flow from one person to another.

I couldn't see how a translator->editor relationship would work either, so I'm not even going to try it. I don't know any native speakers of Japanese, but there's a couple of people I know who know a lot more Japanese than I do, so I consult occasionally.

[Agilis]

I'm planning on writing a much longer and detailed piece about translation at some point, so I'll just keep things short here...

On top of what zal already wrote, which I agree with 100%, I'll add one practical test.

Take what you've translated, let it sit untouched for a week or so while you work on other things, it could even be more scripts from later in the game. Then, at the start of your day, come back to the piece and read through it carefully, comparing it against the original script and retranslating it in your head.

Unless you've been doing this for years and years and know both languages inside out, you're going to have a number of "WTF was I smoking!" moments before you get past 300 lines, most likely you'll have plenty by 100. It could be from horrific English writing, it could be utter failure to even read Japanese, either way, it's not going to be something that an English-only editor can save without lots of guessing.

New translators have a very peculiar form of arrogance about their language skills. they can say how horrible they are at reading a given language hour after hour, but they automatically seem to assume that they managed to translate a line right (functionally, not even stylistically) on the first try. So they translate to English on the first pass, then happily pass that English script onto their English-only editor, as if all that needed fixing was merely fixing up bad writing. Or they just 'fix the English' themselves, still assuming they got it right J->E on the first try.

When you add in stylistic factors, and you've got much more to worry about.

I can find lurking errors after reading a passage 5 times over the course of months, and I've spent hours mulling over a single line. I highly doubt that anyone without years of experience can do literary translation in 2 passes for over 10k lines and expect it to be of consistent quality the whole way through. Hell, I doubt many pros can do it, 10k lines is a lot of mental strain.

[nachtrabe]

I should add that most (all?) good writers can't even write 10k lines in their native tongue and have it turn out "right" after just 2-3 passes.

As the saying goes, "There are no good writers, just good re-writers."

[drmchsr0]

tl;dr: have epic skills in the original language, and the target language.

Otherwise, drama happens and you lose the game.

History has proven this point well.

Not to mention problems such as limited character numbers in a text box with speech dialogue, or small errors such as spelling and punctuation.

[GreatSaintLouis]

I'll throw my hat into the ring as well and say it's not simply a function of your fluency in the original language, your status as a native speaker in the target language, or the number of editing passes you make; you also need to be a semi-competent writer in your native tongue as well. Let's not forget you're essentially rewriting, in your native tongue, a novella-length work, and as such having that literary background will prove essential. You'll be able to pick up on when the script uses various literary techniques and better translate them in your own work, and of course you'll be able to produce a piece that is not only a faithful translation of its Japanese original, but a work of prose that stands on its own merits as well.