Localization: Why We Do What We Do

So, one has to wonder: if a blog is posted to on the Internet infrequently, does anyone still check it? Well, hopefully they do, or nobody’s going to read this!

First of all, of course, news. And to lead off… well, shit, the biggest piece of news is right here (or right below this blogpost, if you aren’t scroll-lazy). I’d actually intended at first to get this blogpost out before we announced Fortune Summoners, but, well, that’s not how I roll. (Which is to say, I “roll” by occasionally leaving blogposts to moulder for months while we tackle multiple projects at once.) There’s not too much else to talk about with Fortune Summoners just yet beyond it being A Thing (and yes, we do hope it’ll be coming to all our current digidistro partners including Steam), but the August-September timeframe should see us slinging some words around about it, and we are aiming for a November-early December release timeframe, for reals. I can’t absolutely guarantee we’ll make it, but there shouldn’t be as many hangups with FS as there have been with other projects and we should be able to get this one out the door quite a bit faster (provided I don’t start erasing a week’s worth of progress in one go in a bout of herpaderp-fueled inanity again as I did recently, but, well painful story).

Development on Territoire in Japan continues apace; there really isn’t a “date” for it in Japan yet, so we can’t actually guarantee it’ll see the light of day from us this year at this point (in fact, with it for sure not coming out at Comiket 80, meaning a gold master for us to work on in September at the earliest, and with our plates pretty full for the rest of the year at this point, it’s all but certain to not land from us this year). We’re still super excited for it, though, and assuming EGS wants to hand off their most recent baby to us we’re eager to sink our teeth into it. EGS has even begun to update their devblog again, and if people want, we could begin posting translations of these up to the CF forums.

There’s a few other things going on in the background… that, sadly, I still can’t talk about other than them being A Thing That Exists. Snipers wait outside my window to take my head off if I so much as type a letter of the plans otherwise. While the possibilities behind the Background Stuff are really exciting, the lack of ability to talk about them basically means we have to run entirely silently and look almost totally inactive for the time being. On some level, it’s maddening. We can say, at least, that beyond our recently-announced Project Three (that being Fortune Summoners), there is a Project Four in some stage of “in the works”, and we can hopefully talk more about that in the future. I may even be allowed to drop a few hints over on The Twittersâ„¢ about Project Four in the future. I look forward to the mad speculation, as I always do, since I am fundamentally a very, very evil man.

Finally, in the interest of actually getting some questions answered, I’ve now got a Formspring. I want to use this primarily to answer “smaller” questions from the community that aren’t big enough for a blog post (meaning I can also answer them in a fashion that hopefully doesn’t take months). So feel free to direct questions over there, though Really Big Questions might get reserved for the outer reaches a blog post.

With the news out of the way, there is actually something else I wish to blog about this time around, something I’ve been meaning to talk about for months, even ahead of our own dissection of how Recettear’s release went from our point of view. While Recettear’s English version has been met with near-universal praise, there have been some questions raised in certain corners about how and why we handled the English version of Recettear the way we did. I’d like to use this blog post to explain why some elements of Carpe Fulgur’s version of Recettear are what they are, address the (tiny) number of flat-out changes we made, and more generally talk about our philosophy to this thing called “localization” and why we consider it so desperately important to both the success of Recettear and to imported games in general. This will be split into two parts: the first being an explanation of our general philosophy to localization, and the second addressing loc choices made in Recettear proper.

Yayifications – To Every Character, A Voice

To begin with, I should define what we mean by “localization” and why we use this word instead of “translation”. I mean, that’s all we’re doing, right? Just taking a Japanese videogame and turning the words therein into English. Well, no; by a word-for-word translation, many of the lines in Recettear are wildly different from the original script. Some have been expanded, some new phraseology slipped in there, puns and whatnot are familiar to American audiences, oh god it’s like the 1980s and chainsaws-to-anime all over again send help.

And yet our version of Recettear is also exactly the same: same characters, same voices, same story. It’s different, and it’s the same. That, ideally, is what we’re talking about with localization: the words may be different and “easier” to read for a native English speaker, but the experience should be exactly the same as the original.

To head off certain arguments (from both sides) at the pass, it’s worth pointing out that “straight” translation certainly has its place. In face-to-face business deals or diplomatic negotiations, for example, you definitely want to know exactly what the other person said. There’s no time or inclination to get flowery with the prose. Direct translation being misunderstood can still lead to, well, international incidents and whatnot, but it’s better to be misunderstood at first and then sort  out what was misunderstood later than have to blame some limp-wristed tramp of a translator for getting creative with that letter from the President. Instructions, comments from game developers, many other sorts of non-fiction writing and speech should be directly translated as well. (This sometimes leads to hilarity, but, well, that’s just bad translation and doesn’t have much to do with localization.)

In fiction, however, the circumstances are different. Fictional works, above all, rely on the suspension of disbelief on the part of the reader to retain their potency, their ability to entertain and even educate. And while bad translation has certainly marred a number of Japanese video games in particular over the years, even perfectly accurate straight translation often falls flat. It isn’t exciting, it isn’t engaging, it doesn’t capture players. It doesn’t seem like the translation is “good”, despite being perfectly accurate.

This is because English and Japanese are different languages.

You may think “Oh, SpaceDrake is exploring new frontiers of blindingly obvious herp-a-derp” with that statement. The fact of the matter is, though, that statement is true; more true than its simplicity might imply. Japanese and English are two of the least related languages on the planet; modern Anglic is a frankly bizzare Frankenstein’s Monster of Germanic and Norman-Gallic, and Japanese is derived in parts from both Sino-Tibetan and (possibly) Altaic roots. Massive swaths of the way they treat grammar, diction, case, and even gender and address are all completely different and non-compatible in a 1:1 sense. In many cases, it’s difficult if not impossible to straight-translate Japanese to English and not have it sound at least a little strange, and the opposite holds true as well. Quite simply, they are different languages. To paraphrase a certain famous wise man: “They aren’t just in the same ballpark, they aren’t even the same effin’ sport.” And for fiction, this poses problems.

When you translate something straight, you often lose much of the “voice” of the language or character or whatnot involved. To somewhat paraphrase another wise man who is somewhat less famous than the first (sadly):

“[Japanese] translation further requires an ear for voice. In any decent [piece of fiction], each character has a distinctive style of speech. In some cases it is more subtle than in others… [you must look] at the speech as a whole, and consider the personality, background, and mindset of the speaker.”

This is why Carpe Fulgur does what it does the way it does. Were we to simply translate everything 1:1, the characters would sound flat and uninspired – and almost nothing rips a reader/player/what-have-you out of their suspension of disbelief quite like flat, unnatural-sounding characters do. Getting to the essence of the characters requires reading deeply and extensively into what is said in-game, and opening a dialogue directly with the original authors to ensure we understand everything (the latter of which is mercifully easy to do, now that we live in the 21st century and can rely on things like email). That is the essence of what we do – in dialogue, incidental text, everything, find the spirit of the work and carry it forward into English – even if that requires fairly significant diction changes compared to “straight” translation.

The Carpe Fulgur Method – Finding the Voice

So what, exactly, is involved with making sure we get the voice of a piece right? As mentioned, the first part is sitting down with the creator(s) of a work proper and going over the characters and world with them. What kind of common terminology is used, what inspired the setting, what kind of outlook do not just the various characters but the world around them have, are there any themes in particular we should focus on, etc. In the age of the Internet, there’s really no excuse for not doing this sort of thing even if we can’t travel to Japan directly to hand-deliver questions; this is not the 20th century, and it’s trivial to discuss works with their creators (and, in our experience so far, creators love to discuss what they’ve worked on). It’s also our policy to bring things like name changes to the developers directly, for approval or veto. If we make any truly major changes to a game, you can be sure that the original creators approve of the move.

The next step involves Robin having a go at Japanese language elements directly. This is where everything is first converted into English – and, ironically, it’s in a format many fansubbers would recognize, full of annotations about cultural mores, explanations of Japan-specific jokes, and the like. It is extensively documented, and still left in its most literal-sounding form (although Robin sometimes mentions what he thinks would be the best way to handle a certain bit of text, and I give these opinions a lot of weight).

This is where I then step in. I take the raw Japanese translation and, consulting both the provided notes as well as talking to Robin and/or the developers directly, I proceed to render the dialogue/interface element/voice bit/what-have-you into something that sounds natural in English. Sometimes, this requires very little “rewriting” at all, outside of perhaps a bit of grammar shifting to ensure it sounds like English a person would actually speak. Other times, however, achieving a character’s natural voice – something that really matches what the Japanese sounded like to a Japanese speaker – requires somewhat more substantial changes to diction. Where an idiom might have worked in Japanese, it needs to be completely re-rendered in English to be sensible; where a character might sound “young”, they end up sounding flat in translation and this needs to be rectified. Sometimes the end result ends up deviating wildly from the original Japanese on a word-for-word basis… but assuming I have done my job even remotely right, the meaning should be preserved.

And then – and this is a crucial part of our process – after I’ve ridden roughshod over the glorious fruits of Nippon with my gaijin-fueled wheels of madness, Robin comes back and takes a look at what I’ve written, and tells me if I’ve completely missed the mark on anything. You guys who are worried about the original translation and meaning being preserved? Robin is basically your advocate: not only does he provide the “initial” translation, but a large part of his purpose in this process is to make sure that I don’t completely over-localize something in a fit of herpaderp-fueled over-exuberance and change part of a game into something not even recognizable.

This is actually a fairly crucial part of our philosophy and process – the concept that the script should never have only a single pair of eyes on it. There have been successful translators in the past who have also been their own editors – the aforementioned Matt Thorn, Alexander O. Smith, Ted Woolsey, several others – but to link back to an analogy above: being a master of two different sports at once is hard. Really, really hard. You are very likely to be better at one than the other, and our system relies on this. I studied for a specialized degree in English and wrote creatively extensively before Carpe Fulgur was a Thing; Robin has a degree, and a decade’s worth of near-constant experience, in Japanese. We are both far better at separate things, and it’s to a project’s overall benefit if we both contribute and bounce ideas off of one another rather than trying to do it on our own, where we’d only be relying on our own life experience to make the work sound as natural and “good” as possible. (It also helps prevent a single person’s ego from getting in the way of making changes – or not making changes – that are necessary.)

That, then, is how our scripts are produced. No one person has sole say in how it comes out – Robin contributes, I contribute, and we make sure the original developers contribute, too. The end result, unless we are fantastically poor at our jobs, should be a game script that doesn’t read exactly like the Japanese version – but one that tells the exact same story, hits all the same emotional responses, and has as near as we can come to the same voice as the original. One of our internal goals is that, even with the changes in wording, a fan from Japan and a fan from America should be able to discuss the plot of game with perfect accuracy. Even if what Griff says in Recettear, for example, is different on a word-for-word basis, both fans should recognize the exact same drives and motivations from the character. If we’ve done our jobs right, you should notice no changes to the story at all – the characters will remain intact, and hopefully make even more sense to the new audience.

The Recettear Notes: What Got Changed And Why

With this established, it’s time to address something: many people have noticed that we didn’t include “translation notes” in the game. Fansub-style translation notes are something we’ll probably never include; they’re massive immersion-breakers and at worst are simply a crutch for lazy translators who don’t actually know the meaning of a given word. So we’ll never do that because we’re professionals (we think) and we aren’t going to wreck someone’s immersion to  drive a reference in their face. We didn’t include one in the manual, either, since we wanted to make sure the manual was light and easy to navigate; including translation notes may simply have confused some users who were just looking for the controls. So in one place, they’re unprofessional; in another place, they’re inadvisable.

This blog labors under neither restriction. One of the purposes of this blog is to catalog any “major” changes we make to a game, or note anything that we feel requires special explanation about our particular version of a game. And so we’re going to begin with our first project, and the major notes surrounding its localization: what is “different” from the original script, and why we did this in the first place. Note that this will, by necessity, contain spoilers, so if you haven’t beaten Recettear yet and wish to avoid spoilers, stop reading now.

Character Voices  – Yayifications and other things

The first thing to address is probably some of the “differences” in character diction between the original script and ours; the largest changes center around Recette, Tear, and Griff.

A fair bit of hay has been made about Recette’s “creative” ways of saying “yay”, which sprang partially from her penchant for saying “yatta” as basically a catchphrase. The actual catalyst for making her “catchphrase” a variable phrase instead of a set one, however, was the content of the game itself.

Recette, as she is presented in the game, is not nearly as “dumb” as many people might initially suspect. While she’s clearly not very worldly or experienced, she’s also a very bright, creative young girl – as evidenced early on in her ability to make the sign for Recettear completely of her own volition. Later scenes, like the infamous “tree scene” (and the sequences with Caillou leading up to it), only reinforce this; she’s capable of quite a lot of creative thinking. It therefore made a lot of sense, to me, for her to display similar, nearly-subconscious creativity in her language. She’d practically invent affirmative responses on the spot; she couldn’t not just stick to one answer. So from there we get yayifications, yepperoni, yayness, and a host of other replies. When coming out of Recette’s mouth, they seemed so natural to me that I couldn’t not include them.

(For the record, however: “Capitalism, ho!” is also purely an invention and one of the very, very few I can offer no explanation for beyond “it felt right”. Given the memetic status it’s reached, however, I don’t feel particularly bad about that one bit of liberty taken with the script.)

With Griff, the one large “liberty” taken with his text was actually its volume. While Griff did have a tendency to make somewhat long speeches in the original script, in a few places in the English script, particularly atop the Obsidian Tower, this was played up even further for effect. The concept around Griff’s character is that he’s trying to be a supervillain, more or less, and is frankly not doing a particularly good job of it for various reasons (not the least of which is the fact that his heart doesn’t truly seem to be in it, despite what he may say). If we produced an English voice for him, he’d pretty much sound like a handsome twentysomething doing a Cobra Commander impersonation. That’s exactly how his voicework, and his tone in general, come across in Japanese.

The way we developed our localization methodology behind the scenes, however, meant that we’d become very adept at handling script insertion so as to reduce the load on our Japanese partners… which meant that, in some places, we could add textboxes to the game. With no character did I abuse this ability quite like I did with Griff. To properly come across as a “would-be villain”, especially one who was trying to hard, he really HAD to try too hard – which meant expanding out his speech in places. His speech atop the Obsidian Tower is the most obvious – in terms of textboxes, it is nearly three times as long as the original script. Nothing was omitted from his speech… but the imagery and, ah, violence of the language was enhanced. A lot of “hammy” villains in English tend to make overly-long speeches (hell, the very word hammy is derived from the famously long monologues in Hamlet), so Griff felt the most “right” when he was given leave and encouragement to start gorging himself on the scenery. As far as I’m concerned, it worked brilliantly, and the scene at the top of the Tower is my favorite scene in the entire game, both for what Griff says and for Recette’s equally epic rebuttal. So mission accomplished on this front, I think.

For the curious, the character who second-most-abused the “we can add textboxes!” function was Tear. Which segues nicely into…

Tear’s “voice” both was and wasn’t changed in the English script. Tear actually went through a complete script revision around halfway through the project, in fact (that is, everything she said in-game was thrown out and redone from scratch) as I realized she wasn’t coming across precisely as I wanted her to.  The problem was twofold: her level of formality, and linguistically foreshadowing certain events later in the plot.

Tear, naturally, is very formal in Japanese, being basically a businesswoman. In the original script, just as in our version, she is very much the straight man to Recette’s silly comedian. She also has a tendency to launch into long, overly-technical explanations of things, which I tried to keep intact as much as I possibly could. The main thrust I wanted to convey, however, was not only that she was almost overly-stuffy, but that “human speech” is practically a second language for her. She was raised in fairy society to be, essentially, something between and envoy and a slave to powerful humans. Thus, her grasp of the language that Recette and the other characters speak is actually not perfect; this is represented largely by the fact that she never uses contractions except in a tiny number of circumstances. Part of the early patching to the game was correcting one or two instances where contraction use had survived from the “old” script.

And in the interest of full disclosure, this is the one “experiment” with the script that I’m not actually entirely sure worked as I meant it to. Tear can sound a bit stilted or overly-formal a lot of times, yes – but I’ve also seen comments in various places that Tear led people to believe early on that the game was translated “poorly” (see above). Because she’s one of the two omnipresent characters and the very first one we hear speak, readers have to deal with her contractionless, hyperformalized, slightly stilted speech constantly, and I think it may have grated on some readers. I’m ultimately still pleased with how she came out – some of the scenes later in the game are very poignant with the approach we took – but if I had to do it all again, I might take yet another look at how I’d approach a character like Tear.

References – Clinton jokes are clearly timeless

Some questions have been raised about the amount of “meta” humor we may have inserted into the game. To be frank, while we did slip some references in, in my view we didn’t go overboard with it compared to the tone of the original script. A good bit of Recettear’s humor operated on a “meta” level from the very start; the item list in particular was filled with references that didn’t even carry over at all (such as the Totally-Not-Druaga gear, which nobody outside of Japan would recognize but would be instantly familiar to any Japanese game fan of a certain age) or with references that even we missed (Japanese fans pointed out a few to us post-release that had totally gone over our heads!) Recettear is a game that knows its audience, and knows what they’d find humorous, regardless of the language. While we did slip in some additional references, especially in the item list (in particular I tried to get some western game references in there, to balance out all the obscure Japanese references a little) a good deal of that was already present in the original script and the game as a whole had a self-aware spirit about it. We simply tried to stay true to that.

Names – Third time is the (c)sharme

On the topic of name changes, generally we didn’t engage in any at all. A few of the names were slightly naturalized in spelling – “Louie” vs. “Lui” and “Griff” vs. “Griffe”, but by and large we left the names alone, and even with a few of the “altered” names, the theme naming and the puns therein (as virtually all the characters have French words for names) remained perfectly obvious and intact.

The one name we did change to a significant degree was the “third” adventurer you might get.

Tielle, in the original Japanese katakana, is named as “ティエール”, which one might think would actually be a cut-and-dry situation. The original site for the game, however, indicates a bit of a different spelling than you’d expect at first: “Tiers”. That is, the French word for “third”. When we spoke with EGS, we found out that this referred to her place in her family; that is, she was literally the third of three sisters.

This actually ended up being a bit of a dilemma and a debate for us. On the one hand, it matched the rest of the somewhat silly theme naming well enough (“Griff(e)” literally means “claw” in French, for example). At the same time, however, it ultimately struck us as both a bit too silly and also perhaps a little too cruel on a cosmic level. (“Oh, Third was eaten by a monster? Well, missus, time to get to making Fourth, then!”) This was compounded a little further by the fact that this character and her sisters may well make an appearance in a future EGS game (or three). Further not helping our decision was the fact that, like many of the characters, she wasn’t consistently labeled in the data files themselves – sometimes she was references as “Tiers”, other times as “Tiel”.

The one thing that finally pushed us over the edge to render her name as “Tielle” was, in fact, the original katakana. “Tiers” is actually pronounced with an effectively silent s (much like “Paris”), but I was fairly certain that most Americans would try to pronounce it with a hard S, meaning it’d seem like her name no longer matched up with the voiceover still in the game. This ultimately drove us to decide to render it as “Tielle”, and after giving EGS a chance to veto the spelling change, the elven archer became Tielle in our script. It’s a policy of ours to still try and stick as close as we possibly can to the original katakana if we have to “change” a name, and given the voicework and given If her sisters show up in the future, we’ll deal with them as they come.

The one other, more minor, name change worth noting here is the name of the town itself. Originally the town was called “Heartsease”. It was altered largely as a result of one item: the “town magazine”. This was originally the “Hartzworker”, which operated on a pronunciation pun that works to some extent in Japanese (T and D sounds are separated only by dakuten marks in the Japanese alphabet, as are S and Z sounds) but didn’t really quite transfer over right to English. There was the added issue of the name not quite gelling with the very French feel of the rest of the town. We therefore asked EGS if they were alright with us changing the name to “Pensee”, and thus allowing us to call the magazine “Le Penseur”. If you’re wondering why we would go with “Pensee” of all things, look at what kind of flower a heartsease is.

Food – The kimono that changed a town

The other change to the script that made a bit of hay in the media was the alterations we made to some food references; basically every reference to foodstuffs in Pensee was made “western”. This is a fairly controversial action to take (some previous anime and game projects have been slammed a bit for making similar decisions), but our reasons for doing this were twofold. The first was simply the extreme “French-ness” of the setting otherwise; we’d talked with EGS and found out that Pensee had a lot of French influences, and so some of the casual references to tofu being a diet staple and whatnot felt a little disjointed in that light. This wasn’t helped by the fact that many other food references were “western” to start with, making it all feel a little disjointed in terms of setting.

Even that, however, might not have been enough to justify such a (relatively) radical change of content had there not been one character in particular who shined a spotlight on the issue.

The entire concept around Nagi as a character is that she’s a bit of a stranger in a strange land; a naginata-wielding female samurai who is trying to make the best of a life among a whole bunch of foreigners. This worked well enough on its own, and she’s one of my favorite characters hands-down, but she created a problem: there were scenes wherein her “foreign-ness” was played up by the rest of the cast to make it clear she’s Not From Around Here… and this included her diet. She’d occasionally describe making some kind of elaborate Japanese dish… and then, occasionally in that same scene, Recette and Tear would talk about the paucity of their meals of things like tofu and the like. To a Japanese reader, this might not seem that odd, but to us this read extremely strangely; it created a sense that this “foreign” character both was and wasn’t foreign to the setting at the same time. So, in order to make sure that Nagi’s “foreign-ness” came through as it should, food in Pensee was made to, well, be foreign to her. With the setting consistent, she properly sticks out like she should.

Questions?

This more or less covers the large changes made to Recettear’s script, and this post is already a monster, so we’ll stop here. If you have questions about other parts of the script, however, or are just curious, feel free to drop a question in the discussion thread on the forums or in my Formspring. I’ll be happy to tackle any questions that might arise, and I do take some pride in the fact that any “change” made to the game was made for a reason.

With that, we’ll see you next time!

Click here to discuss this blog post

Wait, what’s that?

You say the news section of the blog isn’t complete? You want to know what we’re doing with Chantelise?

Well, we do have an update to make on that.

It’s a little something. I doubt anyone will find it of all that much interest.

But I suppose we can share it.

You know, if people will find it interesting.