Author: DubstepKazoo

Howdy, friends. DubstepKazoo is back with the second leg of our journey through the interesting parts of the Wakabayashi. Let’s jump right back into it!

 

The next big thing to discuss is English present in a Japanese text. Unsurprisingly, Wakabayashi has many ideas, and the one you go with largely depends on the purpose of the English within the text. Here’s the one we went with in Senmomo:

 

Mark the non-Japanese expression by italics, underlining, parentheses, quotation marks (‘scare quotes’), angle brackets, uppercase (including all caps), a different typeface, a coloured font, or phonetic or non-standard spelling. (p. 183, emphasis mine)

Katakana English in Japanese can convey anything from trendiness to austerity, but in Senmomo, it is literally English. It’s “the language of the Republic.” It’s in katakana, of course, but characters who use English words are explicitly using words from a different language in their speech. Oftentimes, it’s a word Soujin doesn’t know, so we have to set it apart somehow. We chose the tried-and-true method of angle brackets. There’s actually something else we did, but I think I’ll keep that a secret and see if you can spot it.

 

I will say that we considered replacing the English with another language—such as French—but we quickly abandoned that idea. One, the Republic is very obviously based on the US. It goes on imperialistic conquests, it has a hard-on for freedom, and its most famous restaurant is WcDonald’s. It would be absurd for its language to be anything but English. Two, the reader needs to be able to understand the English involved. I can’t think of a second language that the majority of English readers will be able to recognize. And three, there’s just no language with the same relationship to English as English has to Japanese. It’s that simple.

 

Funnily enough, it’s not just English that this happens to in VNs. As I recently discovered, Tsuriotsu 2 has a handful of lines written in French with the Japanese translation provided as ruby text. My French-speaking compatriots in Operation Bellflower assure me it isn’t shit, though it is missing accent marks and makes a few odd word choices. Truly, I wonder how a hypothetical translation effort might handle that.

 

Wakabayashi also talks about metalinguistic references, such as discourse regarding kanji, and provides several approaches you can probably guess by now (rewriting to English metalinguistic discourse, dodging the issue, keeping the Japanese and adding explanations, and so on). In Senmomo, we do have some proper nouns whose kanji’s meanings are significant. We have our cake and eat it too by including said meanings as ruby text the first time the terms appear, such as sappanwood for Empress Suou or Floral Splendor for Hana Akari. This approach obviously wouldn’t work in a novel, but hey. You better believe we’ll take advantage of the medium.

 

She also has a very long section about wordplay—that thing every translator dreads—and dedicates most of it to categorizing the various types of it. If you want to know what she suggests you do about it, look no further than any translated VN that contains it, as she is of the same mind as us. Interesting to note is that she also proposes just axing the pun if you can get away with it and can’t think of anything else to do with it, perhaps adding your own pun in somewhere else to compensate.

 

The next worthwhile area of discussion is onomatopoeia. In English, onomatopoeia is rather juvenile, and exclusively represents sounds. Not so in Japanese. No, Japanese onomatopoeia can also represent information from other senses (sight and touch, for instance) or even states of mind. They’re more accurately referred to as mimetics, and they can be the bane of my existence at times.

 

If you’re lucky, you can translate them with an English onomatopoeia, a well-chosen verb, or a creative simile or metaphor. If you’re not, things get a lot harder, and Wakabayashi throws a lot of things at the wall in hopes that something will stick. Here’s a very interesting piece of advice she has to offer, though:

 

An acquaintance with English sound symbolism might likewise be helpful—e.g. knowing that fl at the start of a word expresses movement (flee, flutter), gl at the start of a word is often associated with light (glow, glitter), sn at the start of a word often relates to the nose or mouth (snout; snicker) and can have somewhat unpleasant connotations (sneer, snap), and sl at the start of a word relates to frictionless motion (slide; slithery) or unpleasantness (slug; slovenly). (p. 202)

And here’s a very funny one:

 

Omission. This is acceptable if the propositional content is not affected and the sound effect lacks significance. Studies show that omission is in fact one of the most common techniques in Japanese–English translation of mimetic expressions. It should not, however, be used as an excuse for not attempting to find a solution. (p. 205)

One final note on this section:

 

Ultimately, the best way to translate mimetics is to understand the context, deverbalise the meaning and express it without consciously resorting to the above techniques. (p. 205)

That’s my approach with everything, really.

 

The next section she talks about is a relatively minor thing: slogans and catchphrases. Japanese advertising copy isn’t worded even a little similar to English advertising copy. Here’s a short quote:

 

They are often characterised by elements such as brevity, wordplay, and repetition or parallelism for effect. Unless the translation is purely for informational purposes, it is vital that the English rendition also sounds catchy. This might involve modifying the Japanese or substituting a creation that has impact in English. (p. 206)

This was only relevant for one line in Senmomo. Hotori has a side job as the idol (“famous singer” within the text, as “idol” isn’t a word in the Empire’s language) Natsumi, produced by her friend Shino. At one point in Chapter 2, Shino muses over a slogan to advertise a new product with. Predictably, it reads badly when directly translated, so Lonesome threw it out and came up with his own. I don’t have access to our script at the moment, but if I remember, I’ll come back and copy/paste it here later.

 

That’s it for this one. Just one more post in the Wakabayashi train, and it’s gonna be an exciting one: it’ll start with a discussion on honorifics and end with an analysis of how I translated Senmomo’s title. Let me know what you think about this one while you wait warmly for the next. See you next time!

It’s been one good omelet rice lunch and an hour and a half since I finished writing the last post, but I’ve figured out what to write about next. It’s a topic that was staring me in the face all this time: all the stereotypical bullshit that comes with translating dialogue in Japanese media. Honorifics, speech register, Engrish, the works. It’s gonna take several posts (unless you want to read many thousands of words in one sitting), so buckle up. I’m gonna be referencing the Wakabayashi a lot in these ones, since it dedicates a good four chapters to this stuff.

 

Wakabayashi starts by talking about how English and Japanese use direct and indirect speech differently, even saying “[I]t is not always appropriate to reproduce material inside kagi kakko as direct speech in English. Doing so can result in a more dramatic effect than in the original” (p. 146). This is only relevant to VNs when speech is indicated during a line of narration—a relatively rare occurrence—but I don’t think Wakabayashi’s advice here really applies. It’s perfectly fine to render such instances as direct speech with quotation marks in English; the impact will be sufficiently lessened by the lack of voice acting and the presence of narration surrounding it.

 

But if you’ll humor me, I’d like to talk about it a little anyway, even if it’s not related to VNs. Here’s a quote from the Wakabayashi about kagi kakko specifically:

 

Despite the apparent similarity to English quotation marks, dictionary definitions of kagi kakko state that their primary function is to highlight something and distinguish it from the surrounding text. They therefore have a wider range of use than English quotation marks… Kagi kakko do not necessarily indicate that the enclosed utterance is a verbatim quotation. Conversely, their absence does not necessarily indicate indirect speech. (pp. 73-4)

Indeed, sometimes a character’s dialogue looks just like any ordinary narration, particularly if said character is a first-person narrator. This creates a bit of a blurring of lines between the character as a participant in the story and the character as a narrator, as if presenting us with the character’s inner monologue and merely implying that they said something to that effect, to which other characters react.

 

I’m not just making that up, mind you. Have you ever read the Haruhi series in the original Japanese? If you have, you’ll know that very little of Kyon’s dialogue is actually presented inside kagi kakko. That effect I just talked about distances him a bit from the situation at hand and brings him closer to the reader.

 

But as it turns out, the latest installment of the series, The Intuition of Haruhi Suzumiya, makes explicit reference to this. The main thrust of the book is about Tsuruya-san sending the SOS Brigade a series of mystery stories regarding vignettes from her life and asking them to solve them. In them, not a single line of her own dialogue is enclosed in kagi kakko.

 

However, toward the end of the second story, the brigade has cause to suspect that some lines inside kagi kakko, that look like they belong to another character, were actually said by Tsuruya-san. Here’s some of the brigade’s discussion about the affair, translated by yours truly:

 

“The question is why Tsuruya-san was talking like the attendant. What do you think, Koizumi-kun?”

 

“According to the rest of the text of Episodes 1 and 2, Tsuruya-san seems to talk to her close friends in a very direct and candid way. But for all we know, that could just be her inner voice; when she actually speaks, perhaps she does so in the sort of polite language we see in the kagi kakko. Nobody but her can guarantee that her speech matches perfectly with her inner monologue, after all.” (p. 249)

And of course, the other characters in Tsuruya-san’s story reacted to her “inner monologue” as if she’d actually spoken. So you can see this blurring of the lines between narration and dialogue. It’s actually a really cool effect, and it lets you get away with cheap gimmicks like the one Tsuruya-san pulled. I’d like to see that tried more often in English fiction, if possible. Heck, maybe one day I’ll translate a light novel (already have my sights set on one) and render kagi kakko-less dialogue without quotation marks, see how that goes.

 

Just a bit of an aside before we really get into these issues. What we really care about is the content of the dialogue. For several pages, Wakabayashi details many, many techniques for writing colloquial dialogue in English, but I think I’m just gonna skip over that. I mean, come on, that’s how we all write on the internet! I get why a translator of literature or more formal documentation might need a refresher, but all we have to do to make dialogue sound casual is imagine how we’d say it out loud. (Or more precisely, how someone like the character in question would say it out loud.)

 

Things get interesting when Wakabayashi starts talking about keigo. In case you don’t know, keigo refers to the more formal registers of Japanese speech. It involves completely different verb endings from the ones you’re used to, so it’s baked into the very grammar of the language. Unsurprisingly, whether a character speaks in keigo—and if so, what level of it—can say a lot about that character and the interaction they’re in. And just like politeness in English, it can be used sarcastically. Since English has a hard time distinguishing between casualness and politeness based on word choice alone—at the very least, it’s not as clear as Japanese—one must be very careful, or the distinction may be lost.

 

Case in point: Haruhi again. Mild spoilers, so skip these next couple paragraphs if you care, but it turns out that Tsuruya-san’s kagi kakko dialogue was her speaking in English. Oh, and before you ask, T is a new character. She’s introduced as a foreign exchange student, a member of the Mystery Club, and Tsuruya-san’s friend. I like her a lot.

 

Koizumi turned to T and asked, “Does Tsuruya-san’s English sound polite to you, like we saw in Episode 2?”

 

“Indeed,” she declared, brushing aside her bangs. Apparently they’d been enjoying their newfound freedom a bit too much with the hairpin gone. “She utilizes precise idioms and stiff grammar. The single blemish upon the pearl is that her pronunciation has just slightly too many vowel sounds.”

 

“Ahaha, guess I ain’t quite native yet,” laughed Tsuruya-san. “I be tryin’.”

 

Thinking of it another way, I bet if we translated T’s English to Japanese, it’d come out sounding decent instead of the awkward and rough Japanese she actually speaks. Obviously, since it’s her native language. The difference in nuance would probably be something like her first-person pronoun being translated as watashi instead of atashi, which is what she actually uses. (p. 341)

That’s such a small difference! But such changes in speech register do make clear distinctions in how a character’s dialogue is perceived.

 

In many works, the loss of that distinction isn’t terribly detrimental, but it would be catastrophic in Senmomo with its clear hierarchies. Here’s a rather salient quote Wakabayashi has:

 

When honorifics are translated, there is often a loss of nuance and subtleties, and some translators suggest that honorifics can simply be ignored altogether. Yet summarily dismissing the keigo in dialogue passages fails to convey how “They add authenticity to the relationships between characters and situate them within an appropriate social milieu” (McAuley 2001: 65). These relations can be expressed in English by other means. (p. 154)

This is what I did with Inui Takahito, the manager of Koujiya Flowers. He always speaks in teineigo, the standard desu/masu forms everyone learns first, whether he’s talking to the leader of the warriors or the man he’s housing. You may think this somewhat distant of him, but based on his words and actions, I disagree. He’s using it to build up an easygoing, approachable identity, playing up the mild-mannered and wispy image that has all the women of the neighborhood enthralled. The more casual speech register—tameguchi, it’s called—can feel overly blunt and rough at times. As such, I translated his dialogue as suave and colloquial English in order to create the same air of friendliness and reliability that his politeness in Japanese did. I achieved the same effect by almost the opposite method.

 

Next, we have this:

 

Compensation is another way of handling keigo, as in the following sentence where the formality of my dear conveys the politeness of ご滞在 (p. 154)

This is part of what we’re doing with Kotone. Despite being the grand poobah of the Empire’s religion, she speaks in very humble kenjougo to absolutely everyone and invariably exalts their actions with honorific sonkeigo. In particular, she enjoys the sentence-final phrase de gozaimasu. It’s practically her catchphrase. Therefore, we’re making her use sir and ma’am a lot to retain both her politeness and her idiosyncratic way of speaking.

 

We’re doing more than just that, of course. We’re giving her more sophisticated diction, for instance, and we’re also abolishing most contractions from her dialogue and hedging questions with such phrases as might you.

 

One of the things Lonesome and I did before we got started on editing was sit down together and nail down how we’d render each character’s speech, with a particular focus on politeness or lack thereof. Obviously I already had my own ideas when I did the original TL, but I wanted to see what he had in the kitchen. One idea he had for Mutsumi, the elegant and refined proprietor of a popular warrior watering hole, was to have her use one instead of you to make her questions more indirect. That was left on the cutting room floor, but it just goes to show the level of creativity you need to deal with many different speech registers in a setting where they’re important.

 

Wakabayashi has another warning, though:

 

It is particularly important to indicate when speakers shift between formal and informal language. If not conveyed directly through the mode of speech, this can be done indirectly by a phrase such as slipping into informal conversation/language/speech. (p. 156)

This is often very hard to represent. Thankfully, the extreme politeness with which many characters speak in Senmomo makes our job easy on this front. How do we portray Kotone speaking in tameguchi with Isuzu? Drop the fancy bullshit we usually give her. How do we render Soujin’s deferential speech toward his liege at the end of Chapter 1? Pepper his dialogue with qualifying phrases like I would be honored to or I humbly request. The real tricky ones are the smaller shifts, like how Hotori takes a blunt yet friendly tone with Soujin and a gruff, almost militaristic tone with the rest of the Sworn Blades. But we managed, as you’ll see when the patch comes out.

 

Incidentally, that last part of the Wakabayashi quote about hijacking narration to make up for the lack of change in dialogue is something she suggests for a lot of the things I’m going to discuss in these posts. I consider that approach a copout, and thus we avoid taking it as much as possible.

 

Wakabayashi starts the next chapter out with a discussion on dialects, a problem we don’t have in Senmomo, but she makes all the suggestions you’d expect, such as replacing it with an English dialect that creates the same effect, mentioning it offhand in narration or a footnote, or even not bothering with it at all. Hasegawa even has this to say:

 

The protagonists’ social prestige is amply depicted throughout the story; therefore, Seidensticker does not attempt to express it through their dialect. However, the story also includes speakers of the Tokyo dialect. He strives to preserve the dialectal contrast by using the difference in speech tempo: the Kansai dialects usually sound slower and employ longer sentences than the Tokyo dialect, which is generally perceived as crisper (Seidensticker and Nasu 1962: 207–8)… Lines (a), (c), and (d) are in the Tokyo dialect, while (b) and (e) are in the Senba dialect. Seidensticker employs contractions only with the former. He hopes that the non-contracted lines convey the nuance of unhurried speech, rather than formality. (p. 62)

Personally, having actually read the lines she’s referencing here, I don’t think Seidensticker went far enough, and I don’t get the feeling from his Senba lines that he wishes I would. The difference between the two characters’ speech certainly doesn’t feel as drastic as a dialectical difference. Just goes to show that even the bigshots have trouble with this stuff.

 

She also says this:

 

Birnbaum (2006: 207–8) considers that in Japanese-to-English translation, the most conspicuous place to demonstrate one’s creativity is in conversations. In his translation of Ikezawa Natsuki’s 池澤夏樹 Mashiasu Giri no shikkyaku マシアス・ギリの失脚, a story about a corrupt president of a fictional South Pacific island nation, Birnbaum decided to make the protagonist’s mistress (who speaks ordinary Japanese in the ST) speak non-standard English. (p. 63)

Lonesome cited this principle when he made that proposal about Mutsumi’s dialogue, but I axed it anyway because of how unduly conspicuous it would be for a side character.

 

Wakabayashi even makes a nod to fansubbers when discussing dialects, suggesting experimenting with typeface, font size, and color to convey what the text itself cannot, but she ultimately admits that her various ideas have their flaws and a translator needs to experiment to figure out what works. In other words, the pros are just as clueless as we are.

 

She also gives brief consideration to differentiating male and female dialogue. In Japanese, the differences between the sexes’ speech patterns are much clearer than they are in English. It’s often impossible to denote a clear difference, but it’s also often unnecessary. Unless you’re translating a trapge, that is, and I suspect that’s a big reason why gems like Tsuriotsu have yet to see the light of day in English—differentiating the protagonist’s speech register between his male and female personas is often prohibitively difficult, and may require what academia calls compensation in kind, i.e. showing the distinction in some way other than speech register. Funnily enough, she has an aside about gyaru jargon and other modern slang, advising using online slang dictionaries to make sense of them.

 

Hoowee, that’s a lot, and there’s even more left to go. I’m gonna call it for this post. Next one’s gonna start with the problem of Engrish in the source material. Exciting! As always, leave your feedback in the usual places—Reddit, my email inbox, Discord, et cetera—and I’ll see you next time.

I’m writing this post on March 29^th. It’s been a hot second since the English release of Cyanotype Daydream, hasn’t it? As was pointed out last night in a polite and helpful manner by a highly valued and respected member of the Reddit VN community, Cyanotype Daydream contains a smattering of typos and areas where one line didn’t flow into the next. I haven’t played it, so I don’t know firsthand, but I do remember the controversy surrounding the game at launch. Even the more recent Café Stella supposedly had an untranslated line at first. What’s going on? Do the localizers just not care about the quality of the work they put out?

 

Of course they care. They just let some dumb mistakes slip through the cracks, that’s all.

 

It happens to the best of us, really, even the people who wrote these games in the first place. Many times, I’ve seen incorrect kanji, accidentally duplicated hiragana, accidentally omitted hiragana, and every other kind of typo you can imagine. Even Senmomo has a few typos in the original script. One time, they even got the Empire and the Republic backwards!

 

I’ve been known to make dumb brain farts in my Senmomo translation, too. For example, I evidently didn’t hit the 0 key enough times, reducing the ranks of Okonogi’s Forbidden Guard from 200,000 to a mere 20,000. I also make the odd typo here and there, though Lonesome maintains the stance that the relative scarcity of such dumb mistakes in my script is praiseworthy.

 

Lonesome, too, makes these oopsies, despite how good an editor he is—more than I do, in fact. As was likely the case in Cyanotype Daydream, he’ll change one line and forget to alter the next to match it, or he’ll forget to change one of my verb endings to match the new subject he gave it. Maybe he’ll forget to delete one of my words, or maybe he’ll forget to add one of his. None of this means he’s a bad editor; it’s all a normal part of the process.

 

Of course, the problem is when these mistakes survive long enough to reach the consumer’s eyes. I’m obviously not defending the blunders rampant in Cyanotype Daydream at launch; I was just trying to explain where they came from. No, a good QC job should be eradicating them before the master copy hits the servers of Steam, JAST, Mangagamer, Denpasoft, or wherever.

 

Granted, that’s easier said than done. These games are long. If you look at JPDB’s page for Senmomo, you’ll find that this game contains nearly fifty thousand lines. That’s a lot of text, and plenty big-name games are even longer. These companies are trying to meet budgets and deadlines. They might not have the time or money for multiple people to read through several-dozen-hour-long games multiple times in search of a forgotten comma. That’s just the unfortunate reality; all we can do is hope the people working on the script are more careful in the first place.

 

That’s somewhere fan translations have an advantage over official ones: we’re not constrained by time or money. If we wanted to, we could sit on Senmomo for years as we make minute adjustments to the script, even after editing is finished (a maneuver I will henceforth call “pulling a Eustia”). We won’t, of course, but we could. The obvious tradeoff is that fan translation groups are just a collection of randos, while employees of companies went through a hiring process during which they were held to some sort of standard (the actual value of said standard notwithstanding), so fan TL quality is a lot more volatile than official TL quality. Sometimes we’re a lot better, and sometimes we’re a lot worse. Personally, I’d say the Senmomo script gives a lot of official translators a run for their money, but even if you’re skeptical, at the very least, you can expect it to be polished.

 

Believe me, I get just as annoyed at typos in commercial works as you do. It’s sloppy and unprofessional, and it shouldn’t happen. But at the same time, they’re a fact of life, so while you’re in the right to criticize them, don’t be a dick about it. A lot of these game engines are held together by bubblegum and duct tape, making them hard enough to work with as it is. These translators and editors and what have you are trying their best, so be respectful in your dealings with them.

 

Also—not that I’m saying any of you would do this—don’t assert that an obvious QC goof is the product of secondary MTL and proceed to unduly attack a translator/editor without evidence. That’s a dick thing to do.

 

That’s all for this time. I kept a post relatively short for once! Are you proud of me? Let me know in all the usual places. See you next time, when I’ll talk about… I don’t know, actually! Stay tuned.

In case you didn’t know, whenever Lonesome finishes editing a script file of Senmomo, we get on a Discord call together and go over it line by line, looking for odd turns of phrase, typos, grammatical errors, and edits that stray too far from the original Japanese. (He’s usually very good about keeping to the original tone, but sometimes he gets too excited, and I’ll make him redo a line.)

 

But wait. He proclaimed to the world when he first joined me that my translation is already good as is and could be released in its current state after a simple sweep for typos. Why am I bothering to have him go over the script?

 

Or, alternatively, he’s the editor. He’s supposed to touch the script now, so why am I suddenly coming back in and interfering with his job?

 

There’s a lot of good answers to these questions, ones that I’m sure are fairly obvious to most of the people reading this, so I won’t bother enumerating those. In this post, I’d like to talk about one that often goes overlooked: bias inherent in a translator’s perception of English.

 

Lonesome and I come from fairly different linguistic backgrounds. I grew up my entire life in Florida, that strange pocket of the American southeast that seems to defy the influence of all the states surrounding it. I did spend my second year of college in Japan, but I mostly spoke Japanese there, so outside a few short periods of my life, I was mostly exposed to people of backgrounds similar to mine.

 

Lonesome, meanwhile, finds his English origins most strongly in Canada. He also went to college on the east coast of the US, and now he lives somewhere in Asia. With more of an international background than mine, he tends to prefer the conventions of British English in writing, even if his accent when speaking is essentially no different from mine.

 

With all this in mind, it’s actually remarkable how similarly we do speak. But the slight differences between our English sensibilities do make themselves apparent in the script. And just think of the untold hundreds (maybe even thousands, hopefully) of people who will eventually read Senmomo. Each and every one of them comes from a unique background, and they will read the game through that lens. What sounds normal to one person may sound strange to another. (To give an example, I was once criticized for using “sense of equilibrium” over “sense of balance” in a screenshot I posted, but all of us on the team agreed “equilibrium” worked better there.) The final script needs to reflect a neutral standard of English that will sound natural to as wide a range of readers as possible.

 

As such, we get on our calls together to make sure the script reads in a way we can both be happy with. For example, I tend to prefer “too” where he would more readily go for “as well.” I’m an “okay” person, but he likes “alright.” He’ll go for “towards” where I’d sooner use “to” or “toward.” Sometimes, one of us will point at a line the other wrote and say, “This sounds kinda off.” Translators and editors are only human; we can only write based on what makes sense to us. When we make a word choice, we do it based on the nuances we personally consider that word to have. Your personal biases are going to creep into your translations no matter how hard you try. You can obviously counter this with some effort, but it always helps to have another pair of eyes. See this quote from the Wakabayashi:

 

Like writers, translators have their own idiolect. A predilection for certain turns of phrase can result in a ‘translator’s tic’. Ways of overcoming this include paying careful attention to one’s writing, soliciting critical feedback, and consciously looking for alternative expressions when reading other people’s work. (p. 169)

All of which I do regularly.

 

This can easily be a slippery slope, of course. So what if both of us supervise the script at once? Won’t it just become the average of our sensibilities and potentially still not mesh with other readers? Considering that you don’t really hear about this in discussions of translations, I think you know the answer to that question. I’ve gone on about differing linguistic backgrounds, but those are ultimately backgrounds of the same language. Said differences are going to be minute—mostly on the level of “oh, that sounds fine, I guess, but I might’ve gone for something else first.” Indeed, when we went over Chapter 1 on a call with Garudyne and Silverlight, neither of them said a word about our English. I take this to mean that we’ve reached a comfortable medium that most people can enjoy.

 

Indeed, I actually advise against having too many people touch the script, lest its voice get too schizophrenic. One editor and maybe one TLC is enough, with perhaps a QC person dedicated solely to catching typos (and doing other things that don’t involve editing the script). I have a relatively long history with fansubbing, and one time, I subbed a show I was really looking forward to with a group I’d just joined. I’d already read all the source material. I knew the story inside and out. I had a plan of attack for everything. I remember watching the first episode on the TV in my dorm’s common room (this was during my study abroad), then racing into my room to write my TL. Once I’d finished and uploaded it, I went to bed, leaving the rest in the hands of the other members of the group.

 

I woke up to an incomprehensible mess that I couldn’t even recognize as my own script. It was all over the place; I couldn’t tell what sort of voice it was intended to have. Some lines just made no sense at all, and even some subtle foreshadowing had vanished without a trace. I quickly discovered the reason for this: the project had one editor, two TLC people, two QC people, and the group leader. All six of them had editorial authority over the script (which begs the question of what “TLC” and “QC” mean). And more than one of these people spoke neither English nor Japanese fluently.

 

Needless to say, I was less than thrilled at this arrangement. And I, uh, let them know it. I’m still not proud of how I conducted myself that day, but that conversation resulted in the arrangement of letting me go over scripts for future episodes again with the TLC people in a desperate attempt to salvage what little I could. I’m only still in contact with one of the individuals involved in that project, and I distinctly remember him being the one I enjoyed working with; he had a good head on his shoulders, and I’ve gladly worked with him on other projects since then.

 

So as you can see, there is such a thing as too many cooks in the kitchen; once you start getting to three or so people with decision-making authority on editorial work, a script starts to lose its identity. Whenever possible, I’ve liked to work directly with my editors. For instance, when I subbed Attack on Titan: The Final Season with DameDesuYo, myself, the editor, and the group leader would all sit down and watch each episode together before release to make sure we were all satisfied with it. The workflow on that project was much better than the nightmare I described above, so I rarely had anything to say, but I was glad to have the opportunity.

 

When I first recruited staff for Operation Bellflower, I was deliberately vague about what I wanted from an editor. Basically all I said was “DM me your application,” upon which I would send applicants the test I had prepared. I wanted to see what people would bring to the table of their own accord. Most people just gave me very short messages along the lines of “Hey, I’m interested, so send me the test.” A couple even mentioned existing projects they’d worked on.

 

Lonesome, though, was different. I’d seen him responding to my WAYR posts during my translation effort, and I knew he was interested in editing, but I certainly wasn’t expecting multiple paragraphs going into detail about his background and editing philosophy. Even over the course of the next week, while I was waiting for people to submit their tests, he filled my Discord DMs with questions about such things as the workflow, my style, and even my competence. He even asked to have the opportunity to look at the script before fully committing to the project, a request I happily granted. If it was the unreadable MTL mess people feared it was, he would not have considered it worth his time. I was testing him, but at the same time, he was testing me. I was impressed, to say the least.

 

This alone didn’t earn him the position, of course. Even if people hadn’t brought these topics up, I intended to have these discussions with the submitters of the best tests from the very beginning. But even then, Lonesome’s test submission was so much better than the rest that I couldn’t bring myself to care about furthering anyone else. And in case you were unaware, one of the editor applicants does this for a living. That’s how good Lonesome is at this, despite having no prior experience. On Reddit, our camaraderie is a bit of a spectacle at this point, what with how much we talk each other up, but personally, I’m very glad to have him. This is by far the best work relationship I’ve ever had with an editor. More often than not, once we finish going over a script file on a call, we’ll stay on the line for another hour or two shooting the shit about all kinds of things, ranging from what we’re reading to our daily lives, or—and this is by far the most common subject—translation theory. Several of these blog posts I’m writing originated from conversations I had with him.

 

Perhaps you disagree with my statement above regarding how many people ought to have editorial license over a translation. I’d certainly be interested in hearing your perspective. I’m still DubstepKazoo in all the usual places, so by all means, feel free to give me a shout. Next time, I want to talk about something everybody’s guilty of: dumb mistakes.

When I formally announced my Senmomo translation to the world, people said to me in disbelief, “DubstepKazoo, you sexy beast, how could you possibly have translated this entire game in a matter of weeks? Think of all the elaborate, flowery narration it has!”

 

Nah, man. That shit was easy. English has plenty of twenty-dollar words custom-made to provide any specific effect you could ever dream of, so all I had to do was pick the right one at the right time and throw in some appropriate grammar structures when necessary. (Obviously I’m being facetious here; I’ll go into my methods more seriously in later posts.) You wanna know what the real challenge was? The H scenes.

 

This being my first VN translation, I had no prior experience with translating H scenes, and it, uh, showed. My prose in them is pretty shit, and my team loves to make fun of me for it. In this post, I’m gonna go over what makes H scenes so difficult to translate and how I’m going to deal with them moving forward.

 

First of all, there’s the mental exhaustion that comes from doing it. I feel like this should be obvious, but translation takes much more time than just reading. A line that takes two seconds to read could take ten to type, and that’s assuming the time it takes to devise the translation is short enough to be negligible. If you translate at five times your reading speed, you’re blazing fast.

 

If we assume the average H scene takes thirty minutes to read—a bit of an underestimate, admittedly, assuming you let all the voice acting play out—then multiplying that by five gives you two and a half hours minimum of staring at an anime vagina while desperately trying not to make your descriptions of it sound cringey as fuck. You can see how that might wear on a person, no?

 

And chief among the tangible, textual considerations you have to keep in mind is the gap between what Japanese people consider hot and what English speakers consider hot. Oftentimes—and especially in Senmomo—you’ll find that the narration contains long, impersonal, downright scientific descriptions of the various bodily fluids involved, or even of the precise texture of the inside of the vagina the protagonist is railing. This shit ain’t sexy! It’s just weird. Even the terminology for body parts frequently turns out to be weirdly clinical in Senmomo, likely in reference to Soujin’s straitlaced personality. I swear, the only other time I’ve seen inkei used so consistently is in the doctor roleplay scene of Study Steady. And on the opposite end of the spectrum, you have body parts described with words that, as far as I can tell, do not exist, as if the writer just played mix-and-match with various kanji (such as drawing the first one from a pool of secret, shadow, or shame and the second from, say, rift, bud, hole, and so on).

 

If you translate this prose even slightly directly, you get—well, you get my original script. Something that can be described as “awkward” at the most generous. While the original prose may be all well and good to the Japanese reader, it’s downright unacceptable in English, even if that’s what most translations end up going with. Translators need to get creative. After all, English is far less tolerant of repetition than Japanese is, so at the very least, a translator needs to have a toolbox full of synonyms for such common things as body parts and fluids.

 

This especially holds true with Senmomo, where Soujin goes out of his way to avoid the coarser language you’d readily find in other works. (For instance, in the fan disc, he uses a circumlocution to get around having to say iku, even though he’d just be quoting Elsa.) Even with most of the heroines, the lewdest their dialogue gets is iku, which isn’t even particularly profane. Only Elsa refers to the deed as ecchi instead of something more poetic, and only Shino dares refer to her breasts as oppai.

 

I naturally recognized this problem when I was translating the game, and I tackled it by embracing the clinical terminology as a way of conveying Soujin’s autism, for lack of a better word, but it came out jank and basically the opposite of arousing, which is obviously not the intended effect. Lonesome’s solution to the problem was to introduce a lot of metaphor and other figurative language to better fit the game’s poetic tone. Sometimes, this involved throwing lines out entirely and rewriting them from the ground up, as I mentioned in another post. It’s a bold move, one I didn’t have the courage to make last year, but having seen Lonesome’s work and read those textbooks, it’s something I’d very much like to try my hand at in my next project should it become necessary.

 

Then there’s the dialogue. Now, I’m still a virgin, but even I know people don’t talk during sex, not nearly as much as they do in visual novels. But that doesn’t stop these scenes from having dialogue, and frustratingly enough, it’s ridiculously hard to translate well for the most mundane of reasons.

 

For example, consider the word ureshii. Even someone in their first semester of Japanese education can tell you that this is an adjective that means happy. However, imagine now a girl in the middle of passionate sex saying, “I’m so happy…”

 

Sounds stupid, right? The word happy bears childish, silly connotations that ureshii does not, and any even tangentially related circumlocutions don’t fit either. This is just one of many mundane words in H scene dialogue that are deceptively difficult to translate without breaking the tone. Add in the words that are annoying to translate at the best of times (setsunai, for example, is a popular one), and you can see why I struggled. This sort of thing isn’t much of a problem in ordinary dialogue because ordinary dialogue isn’t a tightrope walk. In H scenes, you have to constantly uphold the same tone, and the slightest slip-up can break it completely. (This is, of course, assuming you’re not dealing with the rare breed of H scene that’s supposed to be funny or disturbing, rather than sexy.) This problem is further compounded by the fact that in H scenes, the heroines often narrate precisely what’s happening to them in any given moment. I swear, just once I want to see a protagonist shoot back with “Yeah, I know. I’m the one doing that to you.”

 

Let us also not forget the fact that anime girls behave very differently from real ones, at least in Western culture. They’re generally very meek and submissive when it comes to sex, they’re always virgins (because that matters to some people, apparently), and, well, they fit a lot of the stereotypes Westerners have about Japanese women. Are Japanese women actually like that? Again, I have no experience myself, but anecdotal evidence says at least some are. But you’ll note this as another big difference with Western erotica, in which the women tend to be more confident and assertive, and more hot than cute. To rewrite character dialogue to better reflect Western sensibilities would be a step too far, in my opinion, but you still have to figure out how to render these Japanese sensibilities in a way that’s palatable to an English-speaking audience.

 

In case you couldn’t guess, I threw in the towel and stayed relatively faithful to the Japanese, but once again, Lonesome displayed the courage to take some liberties with the letter of the text to stay true to the spirit, and I take my hat off to his dexterity.

 

One more banal but very big consideration is moans. There’s a lot of them. How do you render them in English? English-language erotica doesn’t contain them to anywhere near this degree. How many A’s should you type? What about H? Or N? Maybe sometimes you’ll need a few F’s or O’s, but how many?

 

Obviously there’s no good answer to this. You just have to play it by ear. Though it gets really bad when the voice acting differs significantly from the text; Kanami’s voice actress, Nekomura Yuki, is frequently guilty of this. You’re doing a good job, mate, but you’re making mine way harder than it needs to be.

 

A related issue is that of sound effects. Kissing, sucking, slurping, squelching, you name it. Which of these do you try to sound out? Which of them do you render as just a verb between asterisks? Do you put the sounded-out ones inside asterisks, too? It’s a very arbitrary decision to make, like the moans, and there’s no real guideline you can follow beyond what looks fine to your eyes.

 

Then there’s the more awkward predicament of words like iya, dame, or muri. They’re all variations on “no,” which in any real-life sexual encounter would be a cue to, you know, stop. This obviously doesn’t happen in these games, and neither character involved seems to think much of it. (In fact, I’ve even seen protagonists try to stop because of these words, only to promptly get yelled at.) Should they be translated directly? Should they be replaced with expressions like “yes” to more comfortably fit a Westerner’s conception of intercourse? Or would that be too treacherous a change?

 

I hope you can see by now how H scenes present their own, very unique difficulties to a translator that regular scenes do not, and they’re likely to trip anyone up the first time they encounter them, no matter how much experience they have with other media. Now that I’ve dealt with Senmomo, though, I think I can do much better on my next endeavor.

 

Then again, a lot of the time, translators bumble through the things just as awkwardly as I did. One of the first things I read in English after finishing Senmomo was Study Steady, and its H scene translations exhibited all the problems I just complained about. The rest of the game’s TL wasn’t good by any stretch either, but the H scenes reminded me of my own work. Considering how much of the game’s content is sex, you can see how that would be a problem.

 

Except I’ve never seen anyone but me complain about it. You don’t even have to know Japanese to notice this stuff, and yet not a single person takes issue with the way that game’s translator handled H scenes. It could easily be that I’m just blowing things out of proportion or overanalyzing things. Indeed, I’ve been told before that the various cultural differences I’ve mentioned above aren’t a problem, that people come to VN H scenes specifically for those cultural differences, so even if dialogue or prose sounds weird and stilted in translation, people see it as a sort of cultural exchange experience, rather than getting put off. That’s certainly a valid take. The obvious best solution is to retain those cultural differences while presenting them elegantly in English, which is what I believe we’re doing in Senmomo, and quite well at that with Lonesome’s magnificent editing.

 

With all this in mind, I now feel a lot more confident about dealing with H scenes in the future: I’ve recognized my mistakes, analyzed them, and learned from them. There’s this one nukige developer whose works intrigue me, and since they’re so short, I think I’m gonna use them to hone my H scene translation skills. Before anyone gets excited (assuming anyone even does get excited about an announcement like this), I’m going to take them very slowly, and I’m gonna sit on them so I can edit them myself. I have a full-time job now, unlike when I did Senmomo, and even if I did have well over a dozen hours a day to translate, I wouldn’t do it what with all the other cool stuff around me I can do. Heck, I’m not even gonna start translating these games until I reach a good point in my VN backlog, so it could be a while before you hear anything about this again.

 

That’s all for now. From now on, I think I’m gonna write about more generalized subjects regarding translation. It’s gonna be a bit more casual—like when Lonesome and I just shoot the shit about these topics during our calls. Next time, I’ll tackle a popular topic of discussion between us: how a translator’s perception of English influences the translated script.

 

As always, I’d love to hear your feedback, be it on Reddit, in an email, or in our Discord server. See you next time, everybody!

Hello again, friends. It’s your friendly neighborhood DubstepKazoo, here to talk about more of the translation caveats innate to visual novels. Today’s topic of discussion? Visuals and audio.

 

If you’ve read a decent number of VNs in English, there’s a good chance you’ve come across at least one voiced line where the translation seems to contradict what you’re seeing and hearing, where the given English isn’t technically an incorrect rendition of the Japanese text alone, but was very clearly written in a boring old text file without the visuals and audio of the game to provide more context. (No, I’m not talking about stuff like “Noa opens her eyes wide” when she clearly has them shut or the Higurashi translation mixing up the kanji for “plate” and “blood.” Those are just objective mistakes, with or without visuals.)

 

Unlike traditional novels, VNs provide extra information to the translator. We can actually see the face of the character who’s currently speaking, and we can hear the tone of voice they use. Even if it’s an unvoiced, faceless protagonist, we at least have other characters’ reactions to aid us. These data points can clue us in on how intense a statement is meant to be, for example, which might influence the words we use in English. It can even clear up ambiguity—for instance, if the textual context is somehow making it hard to tell if you’re looking at the standard janai or the rhetorical janai (assuming a noun predicate, since janai is grammatically unambiguous in all other cases), the delivery of the voice acting can help you. (Though it’s usually obvious, so if you really can’t tell, you should probably be paying more attention to the text.) My point is that the visuals and audio of a VN are important resources to account for, and every translator should be translating with the game open.

 

But this is a double-edged sword. On one hand, voice acting and visuals provide extra stipulations; on the other hand, they provide extra stipulations. If a character is shouting in one line and calm in the next, she’d better do the same in your translation, even if the English flows much better the other way around. If a character does that thing where they make the sentence look like it’s going one way only to take it somewhere completely different at the last minute, the usual solution—a moderate rewrite—won’t line up with the tone of voice on display and may stand out to some readers. If you’d like to shuffle some narration around, but the BGM and sound effects make that awkward, you might be in a pickle. You can imagine how visuals (facial expressions, screen effects, and so on) can cause similar problems. Sometimes, the most elegant translation has to be replaced with a compromise to accommodate these added restrictions (especially since you usually can’t, say, move a line of narration to the other side of a line of dialogue).

 

This even affects line length! Let’s be honest here: sometimes, a long-winded Japanese sentence can be summed up in just a couple English words, and sometimes something succinct needs clauses upon clauses of prose in English to achieve the desired effect. (Stay tuned for an upcoming “Top Ten Words Translators Hate” article.) But when the translated text differs too greatly in length from the voice clip that accompanies it, no matter how good a translation it may be, it’s going to sound strange to the reader.

 

This is an issue I’m quite used to dealing with due to my history in anime fansubbing. In fansubbing, there’s this metric called CPS—characters of text on screen per second. Once your CPS gets too big, you’re moving too fast for the average native English speaker. This obviously isn’t directly applicable to VNs, but having to be mindful of CPS for line length in anime has gotten me pretty used to keeping translations at a reasonable length, though you have a lot of latitude there; as long as you’re not constantly translating one word as twenty or vice versa, you can get away with considerable discrepancies. I also know that even if you don’t match all the pauses and emotional coloring perfectly, you won’t be looked at funny as long as you’re in the general ballpark.

 

There’s another facet to this problem that I’d like to at least mention: some readers with a little Japanese knowledge might pick out individual words in a voice clip and be surprised to see them excluded or replaced in the translation. For instance, a character’s name being translated as a second-person pronoun. However, I don’t tend to give much consideration to this. If I’ve changed or removed a word in the translation process, it’s for a good reason, and if this hypothetical reader is able to understand that word but not the ones surrounding it, there’s a good chance they won’t have the context they need to derive that reason themselves. I’m not going to dumb my translation down just for the benefit of this weird, in-between class of reader. Translations are written primarily for those who understand nothing. Our hypothetical beginning Japanese learner here is just in the unfortunate position of having enough language knowledge to recognize they’re not listening to complete gibberish, but not enough to understand the considerations that go into translation.

 

Then there’s the ones who think they know it all and get mad at you because you didn’t make the word choice they consider the objectively correct one. Sometimes they act like each word only has a single valid translation, which is such a laughable notion that I’m not even going to bother acknowledging it beyond the end of this paragraph. Yes, I’ve dealt with people like this. Translation does lie in a very awkward space between subjective art and objective science, but when people argue that ojama shimasu can’t be translated as “thank you for having me over” because it doesn’t contain arigatou? Yeah, ignore ‘em and move on with your life.

 

That should do it for now. As always, you can contact me easily on Reddit or Gmail, and the Operation Bellflower Discord server is always a great place to make your feedback visible.

In this post, I covered problems introduced by the characteristics of the visual novel medium. Next time, I’ll write about a predicament largely unique to visual novels, but not because of the medium itself: H scenes. Ugh.