Q&A: Translating the humor & tone of Yakuza games for the West
Each of Sega's Yakuza games contains multitudes.
For example, Yakuza 0 (the series prequel released outside of Japan early last year) sets itself up as a playable crime serial set in '80s Japan.
But it can also be a cabaret management game, a blind date sim, a 3D beat-'em-up, an emulator of old Sega arcade classics, a rhythm game, a light-hearted RPG about solving stranger's problems, a real estate management game, and a surprisingly good place to learn the basics of tabletop games like shogi and mahjong.
From a developer's perspective, the scope of such a project seems daunting. From a player's perspective, it can be overwhelming -- my partner and I recently completed Yakuza 0, and saw that our 87 hours of combined play merited a "completion" score of roughly 33 percent.
The hook that holds together all these disparate game design elements, that pulled us and scores of Western players like us through the game, is the writing. For all its focus on cold-hearted criminals and petty evil, Yakuza 0 is a remarkably funny game; it affords players the freedom to quickly jump from incredibly serious, macabre scenes (a criminal is tortured in a warehouse) to (teach a shy punk band to act like Cool Rude Dudes in public).
That writing is translated and localized for the West by Atlus (alongside longtime series translator Inbound) starting with Yakuza 0 and continuing on through Yakuza Kiwami (a remake of the original 2005 game) and Yakuza 6. The localization of all three games has been overseen by Atlus' Scott Strichart, who recently sat down to chat about the ins and outs of adapting these games' humor and gravity for Western players.
It was an interesting conversation that went beyond the localization process (Atlus uses a translator/editor tag-team approach, rather than relying on translators alone) to touch on how, exactly, you translate humor, and how players in different regions can view a game or its characters completely differently. What follows is a version of that conversation we've edited for clarity.
The Yakuza games have a winning sense of humor. How do you translate that for a Western audience?
Strichart: The humor of a Yakuza game is a fine line to walk. It's very clear when the developers want to be funny. It's very clear to us when the writing that exists in the game is supposed to be like, "ha ha here's a joke." So we just want to make sure that if they intended for it to be funny, it also has to be funny to our audience. Whether or not that means changing the dialogue a little bit, changing the style in which it's delivered, making dialogue options a little bit more punchy, that kind of thing.
One of the perfect examples is, [in Yakuza 0] Majima encounters this...did you play the substory where he goes to infiltrate a cult? In that substory, one of the options he has is to crack a pun, in order to get this girl to snap out of her cult tendencies. So in Japanese, that pun is "futon ga futon da", which means "a futon is a futon" or, "a futon flies." It's a pun on words. It's basically a "why did the chicken cross the road" kind of joke.
If we literally translated that, it wouldn't work. If we put in "why did the chicken cross the road?" it's not much of a pun; it doesn't feel in line with Majima's character.
So that's where we had to come up with this pun that we ultimately went with, which was "how do you avoid dangerous cults? Practice safe sects!"
Okay, let's drill into that localization process -- how, exactly, did you go from a flying futon joke to a safe sex joke?
Well it wasn't just me -- we have a team of translators and editors who approach these games. These games are massive; if they were left to just me, i'd be buried under each of these games for years! Most of these Yakuza games actually are on par or greater than your average JRPG, in terms of volume of text.
How many lines of text were you working with in Yakuza 0?
Yakuza 0 is 1.8 million JPC (Japanese characters), and the average JRPG is, I think, 1 million to 1.2. So we were well above the average there.
So anyway, how our process works is, the scripts come in from Japan, and we divvy them up to certain translators and editors to make sure there's consistency among the sections that they're doing.
"We just want to make sure that if [the devs] intended for it to be funny, it also has to be funny to our audience. Whether or not that means changing the dialogue a little bit, changing the style in which it's delivered, making dialogue options a little bit more punchy, that kind of thing."
So that particular pun was in a substory, for instance, so the translation team who was on that was one of our outsourced translators, who was doing most of the Majima substories. Because they were familiar with Majima as a character, they were familiar with the substory content, and all that kind of stuff.
These kinds of localization issues, we don't know about them until we hit them. We're doing it line by line, and suddenly we're like "oh shit, here's a pun." And when you hit that, you have to take a step back and say okay, this isn't going to be a direct translation. We have to deal with it -- sometimes that comes down to a discussion amongst editors and translators, or sometimes a translator will flag it for the editor to say "I didn't know what to do with this, man. Let's talk about this."
So when that gets to the editor, the editor's job might be to come up with a way to make that pun palatable to the Western audience. That's generally the Atlus approach, is we use editors to refine the translated English text to make sure it makes sense to Western players.
A lot of companies don't do it this way. And it's not right or wrong, but a lot of companies dedicate a single translator who has the ability to translate Japanese to English, and make it good English. Whereas we use a method where the translators give an editor, not a literal translation, but translated words off the page that don't necessarily scream "this is great English." That allows an editor to come in and refine that text to make it palatable to a Western audience, whether or not that editor even speaks Japanese.
I really enjoy that style, actually. Back when I was the first localization employee at Level-5, it was kind of up to me to establish that style. I could have just thrown the work at a translation agency and let them go, but I thought it was a good idea to give it a more personal touch.
So we brought in a translator and, you know, when an editor meets a translator and they learn their style through working with them on a game or two, and you can almost feel their style through the text, you develop this like, symbiotic relationship with that translator. And that's kind of how I felt about that translator who was working on Attack of the Friday Monsters with me. It worked out really well.
So why do you think it's a good idea to use an editor/translator team, rather than just a single translator?
Well, it's not right or wrong. I think that both ways of doing localization have advantages and disadvantages. The advantages of this method are that it's a little more collaborative.
A translator gets to talk to an editor, and if you hit a line you're gonna stumble on, you can actually get a lot of insight into what that character is about, what they're saying, and kind of hash it out between you. If the editor kind of strays too far, the translator can kind of rein them back in and say "well, it's a little closer to the Japanese this way", and you end up refining a line to the point where it's a bit more accurate, or more true to what a Japanese developer was trying to convey. And that's something you can't get if you're just a single translator, a single mind trying to parse a game's text as best you can. There's no dialogue.
The second advantage is that, when editors are working on something, and they have less or no Japanese fluency -- I don't claim to have Japanese fluency -- we bring a creative writing aspect to these texts that might otherwise be just a direct translation. And a direct translation, that's not a localization. Not if you just translate something directly, that's often not enjoyable to read. It leaves humor behind, it ends up dry, it ends up boring to read.
What do you lose by using this tag-team method? Besides the extra costs, of course.
Here's the disadvantages: we're slower. When a game is translated, that's work that could be finalized, but then it then goes to an editor, who is then tasked with finalizing it. We try to mitigate that by making sure translators are going when editors roll onto a project, so you end up with this kind of lock-step process. But at the end of the day, it is a bit slower than if a single mind just translated the text and delivered it.
Is that why the Yakuza games seem to take a year or more, on average, to come to the West?
No, that is not why! [Laughs] We are closing the gap. That is my directive from management: "close the gap." Yakuza 5 was two years in the process, and that was because Sega literally closed their office up here [in San Francisco] and the project sort of stumbled out the door. 0 came back into our hands, once we'd established operation out of Atlus, and that took us, I want to say, a year and a half?
Kiwami was a little bit less, a year and two months. And Yakuza 6 was....a year? A year and three months? If you look at the past 6 months, where we've released essentially 3 titles, a Yakuza game every 6 months, no one else is accomplishing this. This is practically impossible. The fact that we're getting it done, with the quality level tht we're held to for this series, is a marvel for my team. I'm nothing but impressed with everyone's hard work on this.
Fair enough! What challenges have you faced along the way towards closing that gap?
To go back to that other disadvantage, on the translator/editor route. It's consistency. The more people you bring onto a team, the more people who touch it, the more likely you are to create inconsistencies. Amongst terms, amongst the way characters talk, amongst spellings, all of that stuff has to be mitigated. You have to have a strategy for that.
"The more people you bring onto a team, the more people who touch it, the more likely you are to create inconsistencies...You have to have a strategy for that."
What I'm talking about is like, this person is capitalizing the word patriarch. Or this person is using "jeez" with j instead of "geez" with a g. And you end up with a character that seems to speak in two different ways, or worse, you end up with literally different interpretations of a character.
When we did Radiant Historia, I decided that a character was going to have a little bit of a British lean in his voice. And I talked to the other editor about it, and even with that understanding...we ended up with completely different British leans! Because one of us understood cockney and one of us didn't.
And that's something that's now actually getting fixed. And you can mitigate some of that in QA, but those little miscommunications, if you don't have very strong communication amongst your group, especially as it grows beyond a couple editors and translators, these consistency problems balloon, and it has to be mitigated by a producer and/or a consistency pass on the text.
What advice would you give others trying to do something similar with their own games?
Focus on consistency and clarity. As a producer, I am the consistency pass. I am the voice on how characters should sound. I'm very upfront with my teams; we have daily standing meetings about our challenges that we're hitting, and we discuss these things upfront in terms-less meetings.
We try to frontload the localization as much as possible, with these meetings, with regular discussions to hammer out the fine details. And at the end, you know text goes through a translator, it goes through an editor, and then I act as a consistency pass.
With Yakuza 6, for instance, we rolled a new editor into writing the main story, and he didn't have a strong grasp of Kiryu's character. Until probably, halfway through the game. So in going back through his text, I end up tweaking probably 80 percent of the text. Because I'm a perfectionist. I shouldn't be doing it that much, but I do. And I told him, you know, your Kiryu's way stronger in the latter half than in the beginning. So I'm going to go back and kind of rewrite your beginning. And he was totally onboard with that.
Do you have any sense of how Japanese players perceive the Yakuza protagonists?
In Japanese, I feel like Kiryu is a little bit more of an avatar for the player. He uses a lot more ellipses than we do in the English version, because we actually want our audience to identify with Kiryu as a character. Whereas in Japanese, you might want to be like, I can put myself in Kiryu's shoes. I can be this Japanese badass.
It's a bigger leap to expect a Western audience to be like "ah, I can be this Japanese badass." So we give Kiryu a little bit more of his own characterization, that is very much in line with the Japanese when they do characterize him. So there's no gap there; it's just a matter of trying to bridge that gap Western audiences might face in trying to fully identify with a Japanese character.
How do you decide when to write in an original response, one that's not in the Japanese version?
When a Japanese person would know what he's thinking, because they're able to put himself in his shoes, or they might have a better idea of what's going through Kiryu's head, that's when we bridge the gap. When we decide okay, the player needs to understand a little bit more of what's going through Kiryu's head, because this isn't an immediately known quantity to them.
And Japanese storytelling can be very subtle, at times. And we don't want to be blunt and hammer players over the head with "hey this is what they're trying to say", but sometimes we do have to lead them a bit. We have to look at the text and be like, this is clearly trying to convey this subtle aspect of masculinity, or honor, because this is what these cultural signifiers mean. And then we have to present something to the player that makes it clear hey, this is a Japanese concept of honor, so please understand that through this thought process that Kiryu is having in the middle of it all.
What's a good example of that?
So, Majima's whole transformation from Yakuza 0 to Yakuza Kiwami. Longtime Yakuza fans know Majima is this class clown, kind of joker-esque personality. And when 0 [a prequel] came along, they decided to completely flip the script on this character [and present him very differently]. So you might get the idea that in playing through 0, Majima is going to be going kind of insane.
But that's not really the Japanese take on that. The whole point of 0 is not to show that he goes insane, but to show that he makes a conscious choice to...let loose. And it is subtle in the game, even despite our best efforts to be a bit more blunt about it, especially in the scenes with Nishitani [a loud, brash character] where it's clear Majima is gaining insight into how someone like that would live.
We brought that out a little bit more, specifically asking the devs "did you mean to communicate this?" Because even we were like, this is really, really subtle, guys. Were you trying to communicate a direct kind of correlation between Nishitani and Majima? And they were like, "yes. If you feel like you can bring that a little more obviously, without adjusting the text to the point that it doesn't make sense or it's not in line with the original, then go for it." And we did! And it's still very subtle. People, at the end of the game, they say well, Majima went insane. And it's like well, heh, let me show you the scenes where he didn't go insane, where it's supposed to be clear he made a deliberate choice.