Why does translation look so easy?

After writing a number of articles discussing the management and team building of visual novel translations, I’ve been wanting to write about the actual translation process, and translation in general, for a while now. The problem was that the topic was too huge, even if you know you’ll be writing a series of articles about the topic, where do you start? For the general audience of Neechin.net, who are most likely not translators, nor project managers, nor anything of that sort, going into the details isn’t particularly helpful. Instead, I need a general topic to start things off and introduce ideas as I go.

One thing that soon came to mind that bridged what us translators of visual novels experience, and what the more general public experiences: Why is it that people seem to think translation is easy?

Before we get started, let me explain where I get that rather outrageous statement from. Observing various discussions about visual novels in the English community over the years, it becomes apparent that at least one section of the audience community feels that way. There are of course the people we ignore out of hand, the ones that will post a link to a game that just came out less than a week ago, and ask “is there an English patch out?” Those people we’re more or less used to categorizing as “selfish twits that aren’t even worth getting worked up over.”

But besides that, we always come across questions such as “How come it’s taking so long (a few months)” or more often “why aren’t there any updates? What’s the status?” There are many variations on the theme, but the overall impression is that the translation phase of the project should go much faster than it does in reality. For the moment, I’m not even going to get into the amount of complaints that come up during the editing phase.

Now, I can’t say that we translators are totally blameless. We do take a very long time usually, though often it’s not our deliberate choice. Still, slow is slow in the world and that’s what we are. However, one thing that most translators I speak to seem to agree on is that the general public believes we can be much faster than we ever could be even in ideal conditions. “They think that translation is easy, and fast! When it’s really neither,” more or less sums it up.

Now, what can possibly cause this disagreement to happen? First, let’s look at it from the side of the waiting audience. To be blunt about it, the demand side of the population is spoiled. They are used to speed-subs that can churn out an episode of anime with decent subtitles within 24 hours of the first broadcast in Japan. Sure the quality varies group to group, but the impatient get their weekly fix with some semblance of regularity, the people who care about higher quality can wait a few days after to get their fix.

Next up, are the manga scanslation groups. Again, it’s the same thing. I’ve heard from a few scanslators I know that there’s a very strong pressure to be release regularly in step with the monthly release schedule for running manga, or whatever release cycle they set for themselves if it’s a bound tankokuban. If they get even a few days behind, the insistent questions about “what’s happened? why’s it late?” start building quickly.

So, perhaps because there are groups who push themselves to the limits in the competitive fields of speed-subbing and scanslating, the consumers go “Ah, well obviously if these people can do all that work in a day, then translating must be easy!” It’s like watching an interpreter quickly and seemingly effortlessly translate between two people, well if a person can do it in practically real time, then how hard can this translation business be, right?

Sadly, this only seems reasonable because they haven’t thought about things in more detail. Visual novels are fundamentally different creatures compared to everything I mentioned above, up to and including the professional language interpreter. So here is the version of reality that we see as VN translators.

First, there’s the length of the work we’re taking on. Most people are aware that we work “with a great deal of text” on at least some level. However, I doubt they know exactly just how great that difference is. I have once read somewhere out in the Internet that the amount of text of a ‘typical’ visual novel can be the equivalent of 1 full season of anime’s script or more. For some of the longer games, I’m positive that it’s a few times that value.

To put things in perspective, allow me to draw from my own projects. Narcissu Side 2nd, the project currently occupying most of my project time, and late for a number of reasons that I won’t get into here, is roughly 4267 + 2249 = 6516 lines of text (Narci2 and Narci1 respectively, there’s probably a few hundred lines on top of that to cover miscellaneous items). That’s considered really small in the VN world, on par with perhaps one main branch route for a heroine in a normal game, which could have 15-40k+ lines depending on the game.

However, if we go by word counts, 6500 lines of script text is easily above the 40,000 word line that writers suggest is a good size for a novella. In fact it probably pushes against the 60k barrier, well on its way to being a full fledged novel. Already, 6500 lines would amount to a good 200-300 page book in your hands if it were published. In fact, one of the novelization projects I’m working on easily stretches past the 200 page mark of carefully typeset LaTeX, and it’s not even complete yet.

Besides simply the volume, the nature of the work differs in significant ways to subbing and interpreting. Both of those forms of translation are very similar in the sense that they’re translating speech. And certainly for interpreting, there’s a very strict time limit. So the important part is to get the proper meaning across as quickly and clearly as possible, with the exact words and other nit-picky details taking a back seat to the need to be prompt. It’s not that those sorts of things aren’t important, but in the grand scheme of things, if you have to prioritize, something has to give way to speed and clarity of meaning.

Meanwhile, what visual novel translation resembles something more like literary translation, such as books, plays, poems, and so on. There’s a mix of prose and dialogue, and in most cases the original author actually expended time and effort thinking about what words to use, unlike what often happens in typical speech. If someone spent effort choosing their words to write down, it’s only fair and proper that the translation should expend effort choosing words to carry the meaning over in the best way.

But just how difficult is this whole choosing words thing? After all, that’s why we have dictionaries. Hell, there are computer systems that are supposed to be able to translate text, right? Even if they get things wrong here and there that’s what good editors are there for, spotting where things went wrong and getting those spots fixed.

As you can see from my stream of rhetoric, this can be a rather hot button topic for us translators. I’ll devote a whole article (or more) to this topic at a later date, but the short of it is that there appears to be this strange conception that languages map onto other languages in a clean fashion. The word for head is atama and if I point at something else, there’s supposedly a corresponding word for it too. Translation is then just taking those mappings, then arranging the words in the right sequence so it all makes sense!

Of course, this isn’t true. The languages of the world probably do have a tendency to have matching extremely basic nouns like“a person” or “a tree”, unless they’ve never seen a person or tree before. But where does our “afternoon” end and our “evening” begin? Do other languages have a name for either of those terms? Does that term, whatever it is, cover the same period of time?

To make things even more complex, it’s not simply a case of “does this word exist in the other language.” Most often, there’s more than one word that will express what a given word means. The problem is that each of those different words covers a different shade of meaning.

For example, take the word “Love” in dictionaries, you could find that it could be “ai” 愛, or it could be “koi” 恋. Now, which is it? “Ai” is closer to the love of families, the deep bond between mates. Meanwhile, “Koi” is closer to the love that exists between lovers, the fiery romantic love that makes you walk on air and brings fresh color to the world; it certainly isn’t the word that is used to describe the love between mother and child.

Many such cases exist going in either direction. The trick is to understand the differences, and choose the proper word to use in the context. I can’t count the number of occasions I had to stop work because I wound up debating back and forth over picking just the right phrasing to render a given line. And we haven’t even considered those nasty things known as “idioms” those set phrases that have totally different, and often very difficult to translate, meanings.

Finally, the translator has to bring all this together into a coherent whole. It’s exactly like writing an original work. All the words need to be carefully considered, the work has to be clear and read without too much strain, and so on. An editor can only do so much if the original is a disaster, no amount of editing genius can save a script that had originally be mistranslated to begin with.

Looking at what’s involved in translation, it’s something of a minor miracle that it’s possible at all. But we do it, we take a long time because it’s a difficult skill to master, and for most of us who work in this tiny community, this is our first real attempt at trying it in our entire lives.

Next time, I’ll write another article about translation, perhaps to elaborate on the mapping (or non-mapping) of words from language to language…


  1. MuzikalNotes
    09/27/2008 11:01 AM | #

    It is true, many do believe that translation of anything is easy. The problem is, when someone is pressured to go faster then they can, the result is either ignorance, or a low-quality translation. The latter in a VN will not do well with both the observant readers and the translating team themselves, and the former is just going to annoy both at the same time. If only people could think a bit more…

Commenting is closed for this article.