don't click here

Sega's Inconsistent Romanization (Feat. A Brief History On Japanese)

Discussion in 'General Sega Discussion' started by Chimes, Jul 19, 2024.

Tags:
  1. Chimes

    Chimes

    The One SSG-EG Maniac Member
    886
    617
    93
    Okay, let's cut the shit.



    upload_2024-7-19_13-12-55.png

    Sega's Inconsistent Romanization
    Or, How I Learned to Stop Yapping and Look at Words



    This is something I've always talked about on my social media circles for ages now, but after I surveyed just how obscure this is to people (outside of the L and R thing in that thread, if you know you know), I've decided to make a thread that jots down all the instances that this has happened. But before I make any other posts to this thread, we need all of the backstory. All of it.

    So if you know anything about Japanese, you would know that the way we spell Japanese in the English language is officially called Hepburn. Let's go on a crash course. Hepburn focuses on how we read and pronounce words in written English. Japanese has five vowels, and those can be linked with a consonant. These sounds when written are called kana, much like how we call ours in English letters. A I U E O, Ma Mi Mu Me Mo, Ka Ki Ku Ke Ko, you get the idea.

    There are some sounds in Japanese that do not always fit this, however. Chi and Tsu in the t- kana row, Fu in the h- kana row, and Shi in the s- kana row. To accommodate for these extra sounds, Hepburn changes them to how we would pronounce them.

    You got that? Now let's talk about the elephant in the room. That's the sound that we use in Canada and the United States, as well as people who learn Japanese as a second language. But what do the Japanese use? You might've noticed that I have been staging this in a distant manner. Well, when Hepburn was introduced in the 19th century, Japan also created their own English system to accommodate and/or supplant Hepburn. That system is called the Kunrei system, itself a offshoot of the Nihon system, and this is what Japanese people mainly use to spell their speech in English.

    I should add that there's a whole fascinating story on how these two systems were sitting in a strange cultural space due to the whole... horrifying mess that was turn-of-the-century Japan, but let's not get lost in the weeds here. Just know that they were going to use Hepburn, but Kunrei stuck around on the books.

    Right off the bat, unless you're actually from Japan (and I know most of this forum aren't), this is where things start to feel "wrong". You start to really appreciate linguistics a lot more once you get over the initial shock. In this system, everything sticks to the consonant and vowel rule. Shis become sis, tsus become tus, and when dakuten get included in the mix, jis are now zis. Things get even messier when you deal with you'on, where normally you append kana with smaller characters to create new sounds. While Hepburn accounted for this, the small ya/yu/yos get included with the consonants in Kunrei, leading to stuff like syo, tya, zyu, et cetera... It's like talking without idioms. If you're not used to it, begins to feel strange after a while.

    Up until quite recently (it was earlier this year, if you're curious about that sort of thing), Kunrei was what was taught to Japanese kids for generations. Typically, when romaji are introduced to kids in the classroom, Kunrei and Hepburn are both included but with little context as to why there's two systems. While to us they're two different spellings, to Japanese speakers they simply just represent the same sounds and are interchangeable when writing.

    After the mid-1940s following World War II, the United States temporarily watched over Japan like a sheepdog. It was there that a wider world of (reluctant) cultural exchange opened up. For the United States, the memory of the war left a bittersweet taste on the public's mouths for decades, with many reluctant to accept goods and culture from the country. In popular culture, the image of Japan was often the same image two to three generations ago. Importing goods from Japan came with this mild hesitancy, with toys often thought to be "cheap" and some filmmakers feeling uncomfortable when admitting that they took inspiration from Japanese cinema.

    But for Japan, there was less hesitancy when it came to the invitation of new ideas. There was still some reluctance, but the most apparent effect of this exchange was the wider proliferation of English as a language. I have a hard time explaining it, but basically for the younger generation it became "cool" to use English words and sentences in any way possible, much like how some of us use French phrases to describe something i.e. je ne sais quoi or déjà vu. Whether it'd be using English loanwords to describe something that doesn't quite exist in Japanese, or making new English words and wearing clothes that often have English text on it.

    (As a tangent, while Japan found itself in a economic bubble in the late 20th century, this was actually why Nintendo had to be careful when exporting their games to the United States, such as the Nintendo Entertainment System. Even in the 1980s, Japanese products still had an unspoken aura of caution: at best, it was still a guilty pleasure, and at worst you had people smashing up cars because it came from a Japanese manufacturer and the American cars weren't selling well. It got really ugly, and I mean really ugly.)

    Nearly all mass media in Japan starting c. 1950s have used English and Japanese together. It's basically the norm; heck, even Sonic the Hedgehog uses both at times. While the English lexicon have become a part of Japanese culture (to the point where it's weird for a game to have only Japanese), Romaji is still shaky in a literary sense. For most people, it's very common to use the Kunrei system. At times, both Kunrei and Hepburn were used interchangeably either in the same sentence or the same word. This isn't really out of the ordinary, heck, I often flipflop between the Canadian way of writing my words and the American way. Most Japanese people don't know that there is a meaningful difference between the two systems, other than it being a different interchangeable way to spell Japanese in English.

    But to an average English reader, some of the ways Japanese gets spelled might seem off, missing letters and all, and without further context you might be tempted to scullydad it off as "just being a typo". But, well, no, that's actually how the system works. We don't say bologna, we say baloney. Sometimes, a tsu is a tu.

    You might actually be familiar with a instance of this already! If you know anything about kaizo games, you might already be familiar with the Japanese Windows game Shobon no Action. It's this really hard parody of Super Mario Bros. that's specifically designed to piss you off. But before you pull up Google, let me assuage you ahead of time by saying no, you're not being gaslit, you did know it as Syobon Action. Nearly everything on YouTube before 2023 calls it that, so why the name change?

    Long story short, that's how Chiki (or Tiki, if you prefer) spelled it in his blog. Like most Japanese citizens, he was taught Kunrei as his primary system for romaji and that's how he types it with his IME. Indeed, it was "Syobon no Action", but in his blog's URLs he shortened it as "syobon_action_description", and anyone not fully familiar with Japanese but blindly poking his website in 2009 to download the game would assume the game's English title was Syobon Action and that it's pronounced "So yo bonn action". It's... technically correct, but no, it's just "Shobon no Action" and this curious instance of Kunrei being used is probably the more popular instance in the current Internet consciousness.

    You know what's also a part of Japanese mass media? Sega. How do they handle English? Yes.

    While not unique to the company, Sega's credits have a particular predilection of carsmashing Hepburn and Kunrei together whenever a piece of text or the credits have to use Romaji. Sometimes it can be mistaken as a typo as I mentioned earlier, but other times it might give you pause if you speak English primarily as your main language and know about Japanese.

    This isn't just your regular mistransliteration i.e. Harry and Hurry, Bas and Bus, Mirror and Miller, this is when names have the given name written in Kunrei, while the family name's written in Hepburn. Stuff like that. This might be the choice of the developers and how they personally write their names, but from experience I've seen given names written with you'on use Kunrei while family names that use differing consonants (i.e. Chi or Tsu) are written in Hepburn.

    Sometimes they just use Kunrei, other times it's entirely Hepburn, and sometimes, they use both at once which might make you think there's some new combination of kana you're not familiar of in your studies. But no, they really just mishmashed two systems together. A part of this comes from force of habit; the Wapuro system for keyboard IMEs allows you to use Kunrei and Hepburn interchangeably, and although I haven't done research on this front yet, it seems Wapuro was a staple among early Japanese computers. Many of these games were developed using these kinds of computers, so it's easy to see that they would overlap. (It might also explain why "key" was often a synonym of "button", i.e. Push Start Key).

    While not strictly related to Sega, there's some oldschool games that write out their Japanese in romaji and while I can't think of a specific example, they also flipflop between both Kunrei and Hepburn. I think it's both interesting and admittedly dizzying at the same time, because you have to store a mental dictionary of which sound they're trying to write out and you have to read between the lines on what this actually says or not.

    For fun, this thread aims to list any instance of Sega's inconsistent romanization, or when they use Kunrei. As a facet of Japanese itself, other developers also use Kunrei, and I aim to make note of those too, but otherwise it's just Sega and their habit of mozidouri syukudai o konashite iru. Sorry, force of habit.

    Normally this is where I end my post, and I wish I could provide some further reading to you, but... uh, I have a confession to make; there isn't really any? At least not in the format us zoomers like to read in. This is the kind of stuff where unless you're REALLY studying Japanese or if you're actually in Japan, you wouldn't really know about it and it wouldn't be useful for casual stuff.

    In fact, I only came across this by sheer accident while going off the beaten path. I was looking up something about Cardcaptor Sakura (it's this oldschool anime about a girl and capturing magical things in cards while wearing costumes) and one of the guys in it had his name spelled weirdly. I checked Wikipedia and next that I know I'm reading this old fan website's blurb about how this character has the same pronunciation but differing spellings. And now I got bit by the bug. Once I noticed it I could not unsee it... it's like getting bit by a hydrophobic dog and next that you know you start seeing special things and you must spread the goods everywhere.

    Edit: Some of the replies here provided very good insight in regards to how Romaji is used, so I recommend checking those out too. I also changed some of the paragraphs here to reflect the new information that I learnt after I initially wrote this.

    Unlike my other threads, it's hard to say if people can provide examples. Developing a eye for this kind of stuff is actually a bit hard, considering the relative obscurity of Kunrei in the West. Some things such as QuackShot's Japanese title using "Guruzia" can hide in plain sight unless you cross reference a dictionary and your romaji system of choice. If you can send other examples of these curiosities, be my guest!
     
    Last edited: Jul 20, 2024
    • Informative Informative x 7
    • List
  2. Chimes

    Chimes

    The One SSG-EG Maniac Member
    886
    617
    93
    [​IMG]

    First example that comes to mind is Altered Beast... this is the Genesis version, but the Arcade version has a secret message that also makes use of Kunrei-Hepburn mixing (specifically, Shinobi and tukutta)

    [​IMG]
     
  3. Black Squirrel

    Black Squirrel

    no reverse gear Wiki Sysop
    9,060
    2,896
    93
    Northumberland, UK
    steamboat wiki
    I always assume Japanese to English translations are wrong. Unfortunately the translation tools can be wrong too - Google Translate sometimes does as its told, but sometimes it goes for literal translations, and sometimes it struggles with kanji shared with Chinese. Yes there are "better" translation tools but they still run into similar issues.

    Christ there was one just last week - "Madou Monogatari: Fia and the Mysterious School" made the Western news, which I had to re-translate as Madou Monogatari Fia to Fushigina Gakkou. I never know if I'm 100% right, but I'm certainly less wrong.


    Sega's all over the place within games. There has never been a period in its history where it hasn't had at least one eye on America, so much of its output has a very Western "feel" to it, but they're always flip flopping between styles in their credits. The same game can have two different translations of someone's name depending on where they worked.



    The absolute worst thing that I hate is punctuation. Is "Fia to Fushigina Gakkou" meant to be a subtitle? They've not written it as such. Except for when they do. VGMdb's policy is to name their pages by exactly how things are printed on the covers so you get all sorts of inconsistencies - "Persona5", "PERSONA5", "Persona 5", "PERSONA 5", "P5". Argh.
     
    • Informative Informative x 4
    • List
  4. Gryson

    Gryson

    Member
    389
    349
    63
    I don't know. I think you might be thinking about this (at least some points) from the wrong angle.

    I'm fairly certain that developers were able to choose how their name was shown in the credits, which is why you get a mix of Hepburn and Kunrei. Some people prefer to write their names using Hepburn, others Kunrei. It's just personal preference. Take for example Naoto "Ohshima" - that "Oh" is not Hepburn or Kunrei, as far as I know (they both use lines over the long vowel). It's just how Ohshima prefers to romanize his name.

    In elementary school, students are taught Kunrei, but the Hepburn will be listed in parentheses when they differ. In general, I think most Japanese people are not very conscious of the difference between the two. They know there is a difference, they may have been told to use one over the other, and their keyboard typing habits may have a bigger influence than anything. But the average person probably isn't super sensitive to the difference between "Tasi" and "Tashi" - they are just spelling variations of the same word. So people develop personal preferences for their own names.

    I agree, though, that it would be pretty strange for someone to mix different styles within their own name. And it would definitely be inconsistent if a company used different styles for the same word in their official materials or something. But individual names seems more like personal preference to me.
     
    • Like Like x 4
    • Informative Informative x 1
    • List
  5. Chimes

    Chimes

    The One SSG-EG Maniac Member
    886
    617
    93
    Oh, do tell! This is something that is often documented rather passively on Wikipedia, and I fear there could be errors as for some points (particularly the 1950s point, I knew this because some 50s medical films that I knew of were bilingual, but it could go earlier or later). I had to go off observations instead of documentation while I was writing this (which can be a big no-no in some contexts), and while performing sanity checks I had trouble trying to find information on a specific concept that I read somewhere, but got buried in one page or another. It's very possible that I'm missing some vital context that's present in other documents that I couldn't find yet.

    Admittedly, I did write that under the assumption that one or two people would be responsible for transcribing names onto a game's credits. I believe that could be a Hepburn thing? It's buried inside a sentence in another page, but from what I've found, Kunrei traditionally uses the circumflex for that purpose i.e. syônen. That's one gap in my knowledge that I've been able to infer thanks to words such as "wolf" (ohkami/ookami), but it appears that Oh comes from the choh-onpu in his family name's first kanji and how they type it on a keyboard (see next point). From what I can gather, it's on and off if in English someone decides to discard the choh-onpu's macron in their name or if they want to keep it. This Wikipedia disambiguation article demonstrates as such.

    That's a really good point too, and that's also one thing I've been able to infer from how I've been able to use my Japanese IME. While I believe some developers might have used thumb-shift keyboards (I have no real way of proving this and no one that I know of really uses thumb-shift), a common way to type has been the Wapuro method (a cursory search shows that this was introduced on IBM PC and NEC PC-88 machines). On my Windows IME, it does accept both Kunrei and Hepburn. Kunrei is faster to type on using Wapuro, while Hepburn can be used in the same sentence. The underlying kana still is the same, so that might play a factor on why some messages and credits use both systems.

    While it isn't Sega adjacent, Bubble Bobble for the Game Boy might be of important reference. Buried inside the game's ROM are several messages which both use Kunrei and Hepburn. This might be a indication that whoever wrote the messages was using Wapuro, because it doesn't use Kunrei exclusively and seems to use it to save on typing time.
     
    • Like Like x 1
    • Informative Informative x 1
    • List
  6. Brainulator

    Brainulator

    Regular garden-variety member Member
    Yeah, I wouldn't say this is necessarily a Sega thing: KonoSuba's full name, according to its logo, is written in romaji as "KONO SUBARASHII SEKAI NI SYUKUFUKU WO!" - changing systems in the same word. Ultimately, they're the same sound as far as the Japanese are concerned, and the meaning is understood. This strikes me as being not too far off from how medieval scribes writing Old English would interchangeably use the letters thorn (uppercase Þ, lowercase þ) and eth (uppercase Ð, lowercase ð) for <th> sound(s).

    I say all of this not to put you down, but with a smile on my face. :-)
     
    • Agree Agree x 1
    • Informative Informative x 1
    • List
  7. Chimes

    Chimes

    The One SSG-EG Maniac Member
    886
    617
    93
    Oh, no, that's exactly the kind of thing I'm looking for! That is super interesting.
    On that note, I'd like to know if there were any errors on my introduction; there's certainly some things that I may have missed.
     
  8. Gryson

    Gryson

    Member
    389
    349
    63
    I guess my point is just that the differences between these two romaji systems are much more apparent to English speakers like us given our language background. For Japanese speakers, it's more of a trivial difference that doesn't have any impact.

    It's true that if you have a list of "Yoshimoto"s and then there's one "Yosimoto" in the middle, it will stick out to anyone. However, I think most Japanese people would not notice or care if the list had a mixture of styles for different names, like "Yosimoto" and "Takahashi".

    To give more background in how these are learned: students are taught romaji only in the 3rd grade of elementary school and are exposed to both systems. Here's an image of a textbook that shows how both "si" and "shi" are included. The teacher may or may not explain why there are two ways to write it, and the students probably don't care at that age. Students have to learn to write all of the letters, and then they are tested on it, and then they're done and never learn it again. One of the main points during that time, however, is learning to write their own name. The teacher will most likely help each student with how to write their name, and the student will adopt whatever the teacher came up with as the romaji form of their name. So if the teacher came up with "Tushima", then that's what the student will go with. (Here's someone writing on their blog about how their romaji name is Tushima.)

    Nowadays, I'm pretty sure most teachers will go with Hepburn, because that's required on Japanese passports. In the past it was more of a mess, with people using one style or the other, or mixing them, or using something different entirely. But it didn't really matter because everyone could read it and most people probably didn't even notice the differences. Well, some people would notice, like the person who made this blog post of examples of mixed-up romaji, but that's an exception.

    Beyond writing their own name, there is basically no opportunity for the average person to learn to produce romaji correctly according to one system or the other, so everyone is going off of the little they remember from 3rd grade. I guarantee you the vast majority of Japanese people do not know when to use "m" or "n" at the end of a syllable in Hepburn, for instance.

    I think it might be going a bit too far to make a generalization like "Sega was inconsistent with romanization". It's probably better to say "The Japanese language is not strict about romanization consistency." You will find this in every informal listing of romaji names in Japan, and game credits put together by young developers is no exception.

    I'm not trying to devalue your post, by the way. I think what you're saying is valuable info, since people were questioning why the S&K credits listing had different spellings. It's just that this isn't really a Sega thing, it's a fact about the language.
     
    • Informative Informative x 2
    • Agree Agree x 1
    • List
  9. Chimes

    Chimes

    The One SSG-EG Maniac Member
    886
    617
    93
    Whoa, that textbook photo was exactly what I needed in terms of that missing context I was talking about. Thank you!
    Yeah, with that in mind I'll go ahead and tweak parts of that introduction. I didn't know those blog posts existed, but that does explain a lot about the back-and-forth that I've observed with mixed-up romaji
     
  10. Cooljerk

    Cooljerk

    Professional Electromancer Oldbie
    4,770
    432
    63
    One thing to understand is that all these romaji transliterations are attempts at mapping japanese phonemes onto english phonemes, and they don't all exist. In truth, the real pronunciation of these japanese phonemes is sort of an in between of Hepburn and Kunrei. It's not Zhou or Jyou, it's a sound in between that kinda sounds like both at the same time.

    This is something understood in english as well when comparing to spanish: the rolling R, or the ll consonent. They are attempted to find english equivalents using double r's and a y, but they are not perfect matches for the same sound in spanish. As we acquire our first language, there's actually a period where children become "deaf" to non-existent phonemes in their language, which accounts for why it's particularly hard for someone who speaks one language that lacks a particular phoneme to go to another which doesn't. The L/R thing in english to japanese is the opposite: neither of those phonemes really exist in japanese, instead a phoneme which is sort of like a midway point for both exists, and thus all the translation problems you get.

    Sometimes these different spellings were attempts to codify the intonation of, say, chinese, a characteristic the english language completely lacks. Zhou jyou etc were meant to reflect the on'yomi readings, which came over to japanese given the natural cross over. It's the old beijing/peking situation.

    Also worth noting that actual spoken japanese doesn't map correctly onto kana either. There are more hard rules with kana than actual spoken japanese regarding vowels and consonents. In actual spoken japanese, for example, there are multiple words that don't end in a vowel or n, which when written in romaji from kana, do. Basically every trailing "su" in japanese is actually pronounced "s" for example. It might be spelled, for example, wakarimasu, but it's pronounced "wakarimas" 99% of the time unless you're being ultra formal and punctual.
     
  11. Chimes

    Chimes

    The One SSG-EG Maniac Member
    886
    617
    93
    Yeah I noticed that the phonology of Japanese is both simple yet hard to communicate
    When I learned how some of the phonemes were pronounced, I attempted to explain it to others to the best of my ability (although I feel there's a better way to do it somewhere)
    On that note, on and off I sometimes attempt to try and transcribe Japanese by ear. There are some words that still give me trouble (for example, じゃない). It can be hard and sometimes, when I figure out the actual sentence using a dictionary, it's far off from what I initially heard! But I could hear that phrase from the original audio once I knew what that phrase was
     
  12. Gryson

    Gryson

    Member
    389
    349
    63
    Romaji isn't trying to map Japanese sounds onto English sounds, though. It's just mapping Japanese sounds onto an alphabet (an arbitrary series of symbols). That's how all languages with phonetic alphabets work. Yes, the general sound categories tend to align across languages that use the Roman alphabet since they're all derived from a common source, but "r" in Japanese is not meant to be "r" in English or "r" in Spanish: they are all different sounds that just happen to use the same symbol. Romaji is a Japanese alphabet and it operates independently of any other language. It is its own thing. It just happens to use the same symbols that the Romans used (well, similar ones). The romaji "Zyo" and "Jo" are the same sound in Japanese and map 100% to Japanese phonemes. Don't be tricked into thinking they're meant to sound like English just because they share the same symbols. As with any language that uses the Roman alphabet, you won't get very far trying to pronounce the letters as if they were English. You first have to learn that language's sound-symbol correspondence, by either listening to it or using the International Phonetic Alphabet. I hope that makes sense :)

    You're talking about the phenomenon of vowel devoicing. Most Japanese phonologists will say that the "u" is still there, it's just being devoiced. There's quite a bit of evidence in favor of this interpretation (getting technical here, but the devoiced vowel can sometimes be detected via coarticulation of the preceding consonant). It still maps onto the kana, though, it's just that the kana has some variation. す (su) can be either "su" with a 'u' or "s" without a clear 'u'. The kana doesn't necessitate that there be a vowel attached. It's just that every textbook (and romaji system) says す = su, and that confuses some Japanese learners when they don't hear the "u".

    Sorry to nerd out on you, but I'm slightly into linguistics.
     
    • Like Like x 2
    • Agree Agree x 1
    • List
  13. Cooljerk

    Cooljerk

    Professional Electromancer Oldbie
    4,770
    432
    63
    I used "english" as a shorthand for the understood "roman" in my post because I didn't think I needed to be this specific, but the word "romaji" gives away the meaning of the character set: it is a linguistic set mapped to the total phonemes of the roman/latin derived languages. This is known as latin orthography, and there is a limit to the mapped phonology of latin orthography. While individual romantic languages may have more or less distinctive characters, they all map to the same core phonology set which derived latin orthography. þ is a latin orthographic character which we do not use anymore, but the phoneme it represented is still in the latin phonology set that latin orthography is derived from. There is a total number of sounds represented in this set. The basic jist of my post is that the character set was built for a group of phonemes which inherently do not exist in japanese and vice versa. The written languages were derived from spoken languages, not the other way around. No matter what sort of abstraction you want to put on the roman character set, it was never built to handle certain sounds or linguistic characteristics of asian languages, and as such will always be an approximation using one orthographic set that wasn't derived from an equivalent phonology.
     
  14. Chimes

    Chimes

    The One SSG-EG Maniac Member
    886
    617
    93
    IMG_1750.jpg
    Here's an interesting example from the hidden credits stuffed inside the Ferie line of personal organizers. These two entries in these credits are Romaji that use a mix of the two systems.

    Normally, you can't see this in normal operation; the device checks for the word "TEST" in RAM, and since none of the minigames ever write that string in RAM you have to take apart your Ferie, rig it with a setup that allows custom ROMs and use a patched version of the Ferie's ROM modified to write that string.
     
    • Informative Informative x 2
    • List