Stuff I Say

Romanization or Why I hate X, Ys and Zs

Posted in Uncategorized by 51future on July 13, 2009

I’m going to go ahead and start this off with two hypotheses: Blue is blue and green is green, and also, Japanese romanization is a linguistic orphan, whose only real strength is conciseness, and (to a lesser extent) interoperability.

So let’s talk about romanization first.

The Jesuits back in the 1600s, more specifically a Japanese Catholic priest or group of priests, were responsible for the first iteration of romanization, which was created for interoperability with their Western printing presses. They spread the word of God and his romanization of the Japanese language until Japan eventually closed itself off from foreign intervention. It wasn’t until a man named Hepburn popularized another romanization system in the late 1800s, which eventually came to be known as Hepburn romanization, that it came back into use. Now it’s unclear to me, just how the next series of events transpired, as I’m not a scholar studying romanization, but according to my digging, Hepburn popularized Japanese romanization proposed by the Society for the Romanization of the Japanese Alphabet (社団法人日本ローマ字会). Shortly thereafter, a Japanese physicist, Aikitsu Tanakadate, created his own variation of romanized script– Nihon-shiki (日本式ローマ字). His intention was to completely replace kana and kanji with the roman alphabet, in order to make Japan more competitive with the Western world. Had he succeeded, he would have at the very least heralded great change, for better or worse, to the Japanese language. After that, in the 1900s, the Japanese cabinet declared Kunrei-shiki (訓令式ローマ字) as the unified standard, which was based on Nihon-shiki, probably because it is more “regular” and thus more logical to the Japanese who are used to the logical progression of kana along phonetic lines and columns. While this is not something I’ve ever really understood very well, learning kana in phonetic groups can aid one when it comes time to conjugate verbs, apparently. Kunrei-shiki was abolished by the Allied occupational forces, and then later reinstated by the Japanese cabinet. A modernized version of the Hepburn system is used for the majority of official documents and public works (road signs, place names, etc.) in Japan today.

Now, on Twitter, @Tomoakiyama notes that Japanese psychologically assign the さしすせそ line of hiragana to sa-si-su-se-so and za-zi-zu-ze-zo respectively in romaji. This is an unfortunate consequence of bureaucracy. Japanese children learn Kunrei-shiki romanization in elementary school in order to write their names and practice the alphabet. However, when they graduate and go on to junior high school, they spend their first month in English classes trying to unlearn Kunrei-shiki so that they can write their names according to the Hepburn romanization that’s outlined in their junior high textbooks.

Kunrei-shiki “makes sense” to Japanese people in that it remains regular down the phoneme lines. The sa-si-su-se-so line is easy to write because all you have to do is change the vowel, however, unfortunately it isn’t very accurate. In my opinion, the ultimate goal of any system of romanization should be to map the roman alphabet to corresponding Japanese sounds in the way that allows a reader to render the sounds accurately. Thus, it makes sense that sa represents the sound さ in Japanese– it is the phonetic equivalent. Su, se, and so furthermore adhere to this rule. This happy go lucky world falls apart at the “siems” so to speak, as “si” doesn’t correspond to the Japanese sound for し, which is most certainly “shi”.

Now here’s where I’m going to lay a few things out.

First of all, the only reason “si” psychologically appeals to Japanese people is because they learn it first. If you taught your children at home than an apple was an orange, and vice versa, they’d be confused and upset to find out the opposite at school. Similarly, if you spent most of your early life memorizing kana from a chart that had the first kana in each line redacted or omitted, you’d find yourself crippled and confused when the time came to correct your mistake.

Second, teachers who teach romanization, and, I’m going to go out on a limb here and speculate too, that this includes most Japanese people in general, don’t understand the concept of phonics (in relation to say, Hangul or the Roman alphabet). In Japanese, kana represent one sound and one sound alone. In the cases where the spoken sound changes or diverges, so does the kana. In English, “ki” can be “ki as in kit” or “ki as in kite.” Letters are like tinker-toys. You put them together and you create a larger whole (say a robot) that dictates the way the different parts of the whole sound (suddenly, those connected bits become arms and legs, and are labeled and perceived differently than they were before the whole came into view), whereas Japanese kana are like predefined computer variables, no matter where they appear, they have the same value, always. This leads to the situation where teachers teach romanization from a chart. They tell their students, who are tiny and have absolutely no exposure to anything comparable, that “si” corresponds to し and that’s that. I think that most Japanese children internalize “si” as a unit, much like they do し in Japanese, and that if you were to ask them whether, say, two consonants sitting together, say, “sk” were a valid construction, they’d almost surely respond in the negative. (And they’d be correct, as far as the romanization of their own language is concerned.)

Really, the blame here lies upon the Japanese government’s wishy-washy policies regarding official romanization and in fact, highlights how apathetic Japanese people are about creating ONE MORE writing system for a language that already has two more than most.

Now that I’ve reached the end of this diatribe, I have one other thing to say, and that is that Japanese culture fosters a lot of memes that gain noteriety specifically because they are Japanese and supposedly indicative of Japanese philosophy, even though when you really sit down and think about it, they’re really more stupid than anything. Failure to come to terms with one, single, worthwhile way of romanizing the Japanese language is just one example of this fervent desire to do it the Japanese way, irrespective of practicality. Two other examples I can think of off the top of my head involve color: namely, the color of the sun and traffic lights.

Scientifically, the sun is closer to white than yellow, but when viewed through the atmosphere, the scattered wavelengths of light appear to be yellow– most of the time. As the sun sets, this scattering becomes more pronounced and the sun can take on a red, orange, or even purple hue as it sinks beneath the horizon and colors the surrounding sky. I’m not sure how I was first introduced to the color of the sun, but I surmise that I just looked up in the sky one day and make that deduction on my own. In Japan, however, children are told, over and over again, until it permeates their very being, that the sun is, in fact, red. Ask any Japanese person what color the sun is, and aside from a few scientists who know better, they will tell you that it’s red. This comes from the flag and traditional Japanese tradition. The Rising Sun, the symbol of the Japanese empire, is red, and thus, culturally, the sun is red. The same red they paint their postboxes with. In the same vein, green lights in Japan are blue. Japanese people will tell you that blue (青) can mean either green or blue, but while a green traffic light can be called blue, a blue sky can’t be referred to as being green (緑). There are a number of traffic lights, some blue and others green, that tell you to “GO.” There are some green and some blue lights in both Japan and America. I’ll even go as far as to say that colloquially, Americans might refer to the “go-light” as a “green light,” when discussing lights that tell you it is legal (and safe) to go. However, the color green does not and will not ever equal the color blue. If you were to show an American a Japanese traffic signal that had a blue (rather than green) signal in it and then ask them what color the light was, the American would tell you it was blue. Whereas, if you were to show a Japanese person an American traffic signal that had a typical green bulb in it and ask them what color the corresponding light was, the Japanese person would invariably tell you it was blue.

This allegiance to specific cultural memes in spite of the ultimate truth to the contrary being both easy to grasp and widely understood, is too responsible for the notion that [psychologically “si” is closer to し than “shi” is], even though the truth is simply that the government mandates that “si” is taught to kids before “shi” is taught to them and that their teachers, teachers who have been teaching “si” before “shi” for who knows how many years, don’t really understand the difference anyway, because, aside from romanization’s function in the modern QWERTY keyboard, its usefulness on road signs, it’s role in helping to compact J-E and E-J dictionaries, romaji as a replacement for kanji or kana in written documents is all but useless to both Japanese people and Japanese language learners alike.

Note: Tomo notes that you can type “zi” on a computer keyboard and get じ and while this is absolutely true, I think this is less about linguistics and more about the QWERTY keyboard and corresponding IME software being flexible enough to accommodate everyone, Kunrei and Hepburn enthusiasts alike.


5 Responses

Subscribe to comments with RSS.

  1. Tomo said, on July 14, 2009 at 12:55 am

    Very interesting post, and this makes me want to go back to linguistics! It’s one of my favorite topics and you got me started 51Future… and thanks, it’s been a while since I was this intellectually stimulated!

    I have to say some of your claims seem to be misguided though. Let me add some explanation. Bear with me!

    ([ ] represents phonetic sound, / / represents phonological representation and | | represents how the word is spelled in the given orthography)

    1. Phonetics, phonology and orthography are entirely different.

    You seem to be mixed up a little but most people are, and yes, they are confusing (I struggled in graduate school!). In short, the acoustic sound, or output of a linguistic form (phonetics – ex. whether a syllable sounds like “she” or “sea”), its subjective representation (phonology – ex. combination of /s/ and /i/), and how it’s spelled (orthography – ex. |si|, |shi|, |ci|, |see|) should be clearly distinguished. If you are American, you may think the /t/ sound – not the letter |t|, the sound you think is /t/ – in “tea” and “water” are the same (i.e. phonologically identical), but you end up pronouncing them quite differently (i.e. phonetically different) and it’s unintentional – the former is an air blockage and the latter is a flap.

    A similar thing happens in Japanese. We think the /s/ sound – not the letter |s|, the sound we think is /s/ – in さ and し are the same, but we end up pronouncing the /s/’s in each environment quite differently and it’s unintentional – the former sounds like the consonant of the English word “sea” and the latter like “she”.

    2. し is indeed a combination of consonant /s/ and vowel /i/. Proof: Japanese does not have [si] sound

    My post that placing し as |si| is “psychological” was a little misleading and I’m sorry about that. It may have given you an impression that it’s affected by extra-linguistic factors, like language education and the hiragana table. But it’s not. Also, you can’t really make an analogy with color, which is visual, not symbolic (think about color blind people).

    For example, how do you explain the lack of [si] sound (sound like English “sea”) in Japanese? Just think about it, why?

    The sound like English “see” (i.e. [si] in International Phonetic Alphabet, or alveolar fricative followed by high-front vowel) simply does not exist in Japanese.

    Actually there is a commonly known allophonic alternation rule in the Japanese phonology as follows:

    RULE1: /s/ becomes [sh] in front of /i/
    RULE2: /t/ becomes [ch] in front of /i/
    RULE3: /t/ becomes [ts] in front of /u/

    They are in just about every phonology textbook.

    For further information, read from page 32 of N. Tsujimura, Introduction to Japanese Linguistics:

    It’s very interesting stuff! You seem highly analytical so I think you’ll enjoy it.

    3. The “accuracy” of Japanese Roma-ji spelling can never be free from second-language bias.

    You said spelling し as |si| is “inaccurate” but if we follow the same logic:

    – |Fuji| (as in the mountain) is inaccurate and should be spelled |Fouji| (In French, |fu| would be pronounced with high-round vowel almost like “few”)
    – |Fuji| is inaccurate and should be spelled |Foudschi| (In German,
    |j| is pronounced like an English [y])
    – “Fuji” is inaccurate anyway because you are not supposed to bite your lower lip to pronounce the Japanese ふ sound.

    My point is, we can never make the Japanese romanization totally neutral and fair to all languages. The “Nippon no Roma-zi Sya” is trying to make their system as free as possible from other languages and they applied the sa-si-su-se-so/ta-ti-tu-te-to system because it is actually the most conforming to the Japanese phonology.

    To give you another example, a common Japanese family name “Takeuchi” looks like “ta-ko-ee-hee” to Germans. In your post’s logic, this is inaccurate and thus should be spelled “Takeoutschi”. And さしすせそ would be sa-schi-su-se-so.

    But if you consider that English is the second language taught in Japanese schools now, yes, the Hepburn system, which is intuitive to native English speakers, should definetely be used.

  2. 51future said, on July 14, 2009 at 1:24 pm

    I’m not a linguist, so I’ll just say that my views are colored by experience and are made up mostly of opinion. I took a linguistic course in college and found it really boring. The aim of the course was to deconstruct Japanese grammar analytically so that we could determine why any given sentence was grammatically correct. However, I’ve always considered language and grammar to be somewhat subjective, so I had difficulty coming to terms with analyzing it objectively like that and by the time the course was over I had determined that linguistics wasn’t quite my forte.

    I champion Hepburn in the Japanese school system for two reasons. First, Hepburn is the way Japan romanizes its official documents and signs, so teaching Hepburn seems to correspond to the interests of policymakers in the government. Given that every Japanese person is required to study English for a good chunk of their time in school, focusing on Hepburn romanization serves a dual-purpose: it aids in the teaching of English phonics and allows Japanese people to romanize content with English speakers in mind, something that I think is important– or at least, should be, otherwise, why aren’t Japanese people studying German, for instance?

    However, generally, I think that romanization on the mother-tongue side of things is a waste of time. I’ve already talked about what I believe the point of romanization is, and in that same vein, I believe that romanization should be a function of the learner’s language. German students should use romanization that corresponds to German pronunciation, English speakers Hepburn, etc. However, with that said, I also believe that romanizing Japanese is a detriment to studying anyway and someone who studies from romaji is just prolonging the amount of time it takes to become functional in the (Japanese) language.

    Romanizing Japanese with Japanese people in mind is a waste of time. Japanese people can read Japanese, type it into a computer, and print it out. The only conceivable use I could see would be in programming languages and Japanese people seem to be getting along fine without being able to compile kanji and hiragana and whatnot in those fields.

    • Tomo said, on July 14, 2009 at 11:27 pm

      Thanks for your response. After all it seems like we agree about the conlusions.

      I’m for using Hepburn in the Japanese school system. Idealists and neutrality nazis like the Roma-zi Sya people and Esperantists may criticize it as “linguistic iimperialism” but I think we should go for the de facto standard. I understand ALTs like you are required to read out numerous Japanese names and quickly use them in classes but if they are spelled like “Tuyosi”, “Tika”, or “Syoitiro” it would be very inefficient. (I still think learning Nihon-siki is easier than learning English spelling though! Too many exceptions and inconsistency. Imagine what I had to go through in Japanese schools! Also, kanji/kana system must be really hard for non-Japanese native speakers.)

      I also agree with you in that Roma-ji should be used as an auxiliary system for Japanese-as-a-foreign-language learners and tailored to their first languages. I don’t think forcing a “neutral” romanized orthography just for the Japanese will help Japanese become any more globalized. I would have to repeat my last lengthy reply everytime people ask me why “Mishima” is now spelled “Misima”! Besides I like the kanji/kana system and I don’t think there’s anything wrong about it.

      Oh, and I didn’t go for Ph.D after I got my Master’s in linguistics because of what I think you find boring about linguistics. I do think it’s interesting because it answers many questions I have about languages but I found speaking different languages more fun than speaking about them.

  3. Ady said, on September 16, 2009 at 1:08 pm

    Very Very Interesting post.

    I really wish Katakana did not exist. I believe its the sole culprit behind the lack of English proficiency in this country.

    Also I would like to add that, being an Indian, I feel privileged that both Japanese and English syllables are a subset of what I have learned back home in India. We have four different sounds for T. e.g. त थ ट ठ. I’ll have to add audio to explain the difference.

  4. Hiko said, on January 13, 2010 at 10:56 pm

    Interesting article – I’m not a linguist or a teacher and so don’t have much of a stake in this, but was drawn here from finding @Tomo’s tweets on the subject interesting.

    Tomo’s made a few points about the criticisms of Nihonshiki that are fair – we might get annoyed at seeing “si” because English speakers are capable of distinguishing “shi” and “si” and “fu” and “hu”, but Japanese aren’t and so it is natural that they would prefer a structured approach that uses fewer characters.

    Taking “si” for example – In Romanian, they differentiate “si” and “shi” by always writing “si” and changing the inflection over the “i”. In this way, they are doing something pretty similar to what the nihonshiki standardization does. The lesson is that with most any language, people have to learn to adapt to the way that they write their language. English is of course the worst offender (maybe second to French) at spelling words completely inconsistently with the sound of words.

    Above said, I agree with the general thrust both of you are getting at – if you are building a standard romanization system from the ground up, you might as well pick the system that makes the most sense to both foreign speakers and Japanese, and doesn’t overburden typing. Fortunately, you can get away with typing pretty much using either romanization system and get away with it – and I might add, that I quite often type “si” instead of “shi”, and “tsu” as “tu” just because it saves pointless keystrokes. All that said, I think the Hepburn system still offers the best balance of both worlds.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: