Original title: “R&D Lab A”

This is an edited version of a blog post that originally appeared on 01/27/2007.


The slap of the junklight on the cove;
The freeze-frame of ducks
Below us; your shots; the wounded flop
And skid of one bird to the thick brush;
The moon of your face in the fire’s glow;
The cold; the darkness. Young,
Wanting approval, what else could I do?
And did, for two hours, waist-deep in the lake,
The thicket as black as death,
Without success or reprieve, try.

Charles Wright, “Blackwater Mountain”


SegaGaga localization continues according to schedule. (We have no schedule.)

We have just finished the translation of the in-game text for the first RPG segment, R&D Lab A. I want to focus on that, as well as describe the long companion PDF file that I want to make available with the patch.

I’m dealing with a breadth of side characters for the first time in SegaGaga. My goal with these characters is to make each of them distinct insofar as befits them to be distinct. By contrast, I want to make certain main characters less distinct from each other for narrative purposes.

I think it this approach as an attempt to simulate an un-self-conscious observation of a gallery of freaks. Essentially, the main character–Taro Sega–needs to have as unobtrusive of a voice as possible. He needs to sound like an all-American Dudley Do-Right Jr. because he exists as a shell for the player’s ego to inhabit. (Yes, I know that Dudley Do-Right was Canadian. Taro is the result of what Dudley calls his “Yank infidelity.”)

Japanese games (and especially RPGs) tend to characterize their leads so that they assert themselves minimally upon the player, who is likely using the game as an emotional escape from reality. I think that the fact that characters like Link and Chrono are regularly voted among “the greatest game characters of all time” suggests a widespread appreciation for the freedom to role-play a pre-determined identity.

At any rate, Taro needs to be as translucent as a character. Alis and Yayoi (the Heroine) need to be likewise translucent, but they need to vary subtly from each other and Taro. The translucence works to make the main characters more accessible to the player.

By analogy, consider them the glass on your television screen. If it wasn’t translucent, you couldn’t see the displayed images. Going along with the terms of the analogy, of course, the displayed images are the forces external to the main cast–the forces that help and hinder them.

By writing the voices of the core characters as distinct from the rest of the cast, we can keep the story’s main narrative lens clean while also giving the core characters the illusion of personality. Comparative differences create personality, and the slight differences between the core characters’ voices should be enough to make them seem as though they have more than perfunctory roles in the story. (In the end, of course, they are 80% narrative tools.)

The next category of characters are the Supporting Characters. They are the figures in the RPG segments who have more than three lines of dialogue and with whom the player must interact in order to advance the game.

These characters are trickier and more fun than the main characters. The original Japanese text can be pretty bland, and the supporting characters are often distinguished according to codified caricatures, another narrative trope of many Japanese games. It’s also common that these caricatures–taken exactly as they appear in the text, without comparing them to other cultures’ caricatures–are extremely self-referential within contemporary Japanese culture. They are recognized types of people. It might help a localizer to know what it means when a character is supposed to sound like “an Edokko,” but that stock type’s grammar, vocabulary, and tone (recognizable to most Japanese from television, if nothing else) won’t literally carry over into English.

An Edokko, by the way, is a person from Tokyo. He usually has the hip, jive ways of a city-slicker–specifically a Tokyo city-slicker, whose defining idiosyncrasies mostly work only when seen beside the manners of the Kansai ruralite, another Japanese stock narrative type.

This is a pretty looming problem for translation. Language, of course, is one of a game’s main signifiers for the caricatures. This might suggest that it makes the work easy, but the opposite is true.

One part of the problem is the fact that Japan has more or less definite limits on its cultural landscape, so they have much less subjectivity when it comes to perceiving a ruralite. By contrast, the West has many different sets of provincial stereotypes.

For example: a Southern American accent might strike a Midwestern American as indicating rural ignorance. It might also strike a Southerner as indicating someone familiar, salt-of-the-earth. A Scotch accent might strike an American as… what? Maybe comic and intensely virile, as in Mike Myers’ use of the type? Yet it would have different associations for an English gamer.

The overall problem resembles a game of telephone. The Japanese text uses linguistic cues to assign caricature types. The translators try to use stereotyping linguistic cues to translate those stereotypes. Finally, on the audience’s end, the dialects and accents imply different qualities than the original text carried. A rural, Kansai accent will not immediately imply ignorance to a Japanese listener in the same way that a Southern accent will to most American listeners.

Let me clarify that I do not categorically disparage accent-based translations. Final Fantasy XII stands as a great example of how accent-based translations can create a memorable localization. (Square-Enix should offer Alex Smith’s team many fattened calves to keep them on staff.) I think, though, that Final Fantasy XII‘s particular circumstances make the various accents more justifiable. The characters traverse a whole physical landscape, so they need to also traverse a believably varied linguistic landscape.

So, to use the dilemma’s metaphor, this is our rock: the composition techniques of some Japanese text relies upon fixed (or, at least, agreed-upon) social patterns that are not exactly replicated on a large scale over the rest of the world. The easy way to handle this problem–“mimicking” the accents–often dulls the writing and burdens the final work with social associations and biases that distract players from the contents of the text.

And this is the rock’s complementary hard place: inventive expression can solve the problem as long as the invention doesn’t alter essential content. If I write with voices invented off the general types evoked in the initial Japanese text, I think we’ll come out with an interesting translation. I have tried to establish the RPG areas’ supporting characters’ voices with idiosyncratic rhythms, vocabularies, and grammar patterns.

One set of characters is potentially tricky since they use the same sprite model, despite being different characters. I worked to distinguish them by their expressions. Drawing from the general tone of the characters, I decided to pursue the first character as a 1950s superhero (think Powdered Toast Man) and the second character like a vaguely Hopkinsian Yoda. Again, these voices are lightly invented from the tone and content of their Japanese origins.

I think it’s important to distinguish between this sort of expressive characterization and the regional stereotyping often used in localization. I would like to think of briefly seen, definitely caricatured characters as “grotesques.” (I use the word “grotesque” in the sense that it is used with respect to the fiction writings of Flannery O’Connor and Sherwood Anderson.)

Here’s an example of the differences between the two grotesques that use the same sprite design. Here’s the design they share:

Here is how we’ve got this character, Flagman, introducing himself to Taro.

FLAG: Heyo! The name’s Flagman! I’m in charge of the flags in this Development Studio.

TARO: Oh, hey. Nice to meet you.

FLAG: So you wanna know about flags, do you?

TARO: Huh? No, I didn’t–

FLAG: Can’t say I blame you! Now listen: flags control everything that happens in a game, whether it’s an epic quest or a love story. Life and death themselves hinge upon whether countless little flags are up or down.

Taro meets the second Flagman at the entrance of a puzzle. The second Flagman explains the rules of the puzzle. Since Taro has the Golden Flag, smaller flags will sprout at his feet as he walks through the sterile Flag Garden. I versified most of his speech, playing with rhythm, rhymes, and occasionally drafting his lines in hyperbolic, self-consciously poetic quatrains.

There’s no one left to raise a new flag,
and all our flags are withered and dead.
Young man with Golden Flag in your hand,
go forth and the Flag Garden tread.

While these lines are enjambed here, they’re not meant to appear as a poem on the screen. The dialogue is meant as speech rather than lineated verse, but drafting his lines in this format helps to establish a sing-songy quality to his words that the masculine, staccato assertions of a 1950s superhero would foil.

A thrush, because I’d been wrong,
Burst rightly into song
In a world not vague, not lonely,
Not governed by me only.

Richard Wilbur, “On Having Mis-identified a Wild Flower”


This entry was posted in Original Blog Posts. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *