Romanization in Taiwan

I just spotted this article on romanization in Taiwan at a good and brand-new Taipei related blog with the unfortunately bland name of Taipei, Taipei. This article he(she?) links to, as well as the blog post on Taipei2 do a good job of introducing the problem of completely un-standardized, incompetent romanization of place names in Taiwan.

At least the situation seems to be improving in a way. Here in Taipei, all official signs now use standard Hanyu pinyin.

Let’s look at the way 古亭 has been romanized. The MRT stop is labelled “Kuting.” But because the apostrophes are routinely omitted in Taiwan, it is completely impossible — even for the relatively few people who are familiar with Wade-Giles — to know if the name is really Ku-ting (Guding ㄍㄨ ㄉㄧㄥ), K’u-ting (Kuding ㄎㄨ ㄉㄧㄥ), K’u-t’ing (Kuting ㄎㄨ ㄊㄧㄥ), or Ku-t’ing (Guting ㄍㄨ ㄊㄧㄥ). (Note that hanyu pinyin, Guting, has no such ambiguity and works well to show the correct pronunciation.)

That’s four equally likely possibilities — and that’s without considering tones, which are an essential component of Chinese. If tones are included in the computations, there are 64 different possible pronunciations of the two syllable “Kuting” — hardly a useful representation of 古亭.

As it so happens, I live right by 古亭 MRT station, and it’s official romanized name is, of all things, Guting! Exactly what it should be. The Taipei city/county government has, sometime in the past few years, rewritten all of the signs in proper Hanyu pinyin.

Another example that those of us currently living in Taipei luckily do not have to deal with.

But Tamshui is the historical Taiwanese name for the city.

No. Tamsui (no h) is the correct historical spelling, reflecting the Taiwanese name for the city. Tan-shui would be correct Wade-Giles, and Danshui correct hanyu pinyin. Of course, the “Tam-shoo-ee” pronunciation formerly used on the MRT is quite beneath contempt.

Like Guting, 淡水 is now rendered in the correct Hanyu pinyin of ‘Danshui.’ It may not match 17th century Dutch maps, but it sounds closer to the Chinese pronounciation, and it’s consistent with, for a start, the way people write Chinese words in the Roman alphabet in the other 99% of the planet that isn’t Taiwan.

Unfortunately the problem persists in other areas of Taiwan. For example, I have seen the character 中 romanized as, zhong, chong, chung, jhong, and now thanks to Taipeitaipei, the inexplicable ‘jhorg.’

How on Earth is this inconsistency helpful to anyone?

12 thoughts on “Romanization in Taiwan”

  1. Stories like this remind me how annoyed I get when some American conservatives (not that I write this as a righty) put down pinyin as a standardization system solely because “the Commies made it.” A single romanization system can work wonders.

    Now if only Koreans could get around to creating standard romanization for their language…

  2. Pinyin does have problems, though. There are quite a few character combinations that are pronounced absolutely nothing like how people unfamiliar with the system would expect, since they don’t match usage in English or any Romance language. “x,” “q,” “c,” and “z” are particularly bad. There are also a couple of different sounds that Hanyu Pinyin fails to distinguish while other systems do.

    Another thing is that Pinyin only works for Mandarin– at least, you can’t use it to romanize certain Wu dialects (or languages) like the Shanghai dialect that actually do have voiced consonants. Same goes for Taiwanese as well, which has voiced consonants in addition to the aspirated and unaspirated voiceless consonants. If you use Wade-Giles, you have the voiced consonants left over and then can use them for Taiwanese and Shanghaiese, grafting onto Wade-Giles. Of course, you can also just use a whole different romanization system.

    Of course, since Shanghaiese is practically banned in the PRC, and everybody official wants to stamp out other Chinese languages and use Mandarin, who cares.

    The biggest problem with Wade-Giles is that people ignore the non-alphabetic apostrophes, even though they shouldn’t. This is a problem with most Japanese romanization systems, too, even the official version of Hepburn. Most people seem to use the modified Hepburn that doesn’t use diacriticals though.

    Wade-Giles is a little bit easier for a non speaker of the language to produce something closer to correct, but harder for a non-speaker to distinguish between certain sounds which sound similar in Engish. Of course, if you don’t know the language, you’re going to pronounce things horribly wrong anyway– what’s the use of distinguishing between two sounds that sound really similar to Westerners if that’s just going to make them pronounce one of them horribly wrongly.

    Perhaps different romanization systems are appropriate for those learning the language and for those who don’t but need to pronounce the occasional place name and proper name.

  3. And of course, since no one regularly writes the tone information, it’s still never going to be right.

  4. Pinyin definitely has a number of problems, but the benefits of ANY even half-way decent standardized system absolutely outweigh the drawbacks of the complete lack of standardization that exists in Taiwan.

    As an accurate phonetic description of Mandarin I find Pinyin, and every other Romanization system, a bit confusing. I used Pinyin for both semesters of elementary Chinese in the States before I came to Taiwan, but the local Bopomofo system (which I just taught myself recently) gets rid of all of those problems. By not using Roman letters at all, it takes away all the ambiguity.

  5. Yeah, Bopomofo works extremely well as a syllabary, since it was actually designed for the language.

    Mandarin and English just have way too many phonetic differences for any good romanization system to really exist. Japanese is at least much closer, and you don’t have the problem in Japanese like you do in Chinese where multiple Chinese sounds each sound most similar to the same Engish sound– Japanese maps one to one pretty well.

    The needs of someone who just needs a decent approximation at place names and proper names, are completely different from those of the language learner, who needs to easiy distinguish sounds which are different in Mandarin but sound identical to most English speakers. Pinyin in way too many ways fails the “is this the spelling which, when pronounced by an English speaker who doesn’t study Chinese, sounds closest to the correct Chinese of any possible spellings?” Renders it insanely annoying for the occasional users.

  6. Totally agree with you on the annoyance of the multiple romanization schemes. It’s pretty easy to handle any Japanese romanization scheme, and there are only a handful of places where you’ll make significant errors based on faulty or incomplete romanization. The situation in Taiwan is so much worse.

  7. John, I’ve got to disagree with you on a few points. First off you said that standard pinyin’s “x,” “q,” “c,” and “z” don’t “match usage in English or any Romance language. Not true. Many Portugese words that use “x” are pronounced very similarly to the sound represented by the same letter in Pinyin (an alveolo-palatal fricative). Furthermore, several Spanish words also follow this spelling. One example woule be the second largest city in Guatemala, Xela. Qs can also be found in Romance langauges. The Chinese sounds represented by “c” and “z” DON’T EXIST in any Romance langauges. And not they are NOT the same as the English “ts” and “tz”.

    Also, Mutant Frog, it simply isn’t true that zhuyin is less ambiguous than pinyin. Both are ambiguous at the single character level. For example, in pinyin “i” changes depending on whether it follows “j,x,q” or “zh,sh,ch”. In zhuyin the word 重 is spelled ㄓㄨㄥ. However, if each zhuyin character were read separately and then put together it would make the sound “zh+u+eng” => “zhueng” (sounds roughly like an English like jwung).

    However, both pinyin and zhuyin are 100% accurate at the syllable level, and Chinese characters can be converted to either with no ambiguity about what sound should be made. Pinyin and zhuyin can also be converted back and forth between each other completely accurately with no loss. Of course what isn’t possible is to take pinyin or zhuyin like bo2 ㄅㄛˊ and make a 1-1 mapping into a Chinese character. 柏,勃,博,薄, and 駁 all have that sound.

    I totally agree with John about the tones. It’s great progress that at least Taibei is finally using standard pinyin on their signs, but with the tones it would be a lot nicer for the businessmen and tourists who just study a few weeks of Chinese before they come. It’s not that hard to learn pinyin well enough to be understood when reading it plus a few dozen characters. But, learning how to just read the characters for EVERYTHING is thousands of hours of work.

  8. As the author of the Web page that got this discussion started, perhaps I should explain a little. That page is about five years old and thus considerably predates Taipei’s change from bastardized Wade-Giles to Hanyu Pinyin. So it’s out of date, and I really should get around to doing something about it. (My thinking on the issue has probably changed some as well, so I’m not looking forward to reviewing that particular blast from my past.) Most of my Web-related energies have since been focused on a different but related site:

    I have a post on my new site related to the issue of Taipei street names. It may be of some interest to your readers, as may also be this large chart comparing various romanization systems. (It is more comprehensive and also, I believe, more accurate than that of the GIO.)

  9. My G-d are foreigners such asses when it comes to our issues. The fact of the matter is, the disparate (now actually dichotomous) romanization standards are but a reflection of the Blue-Green divide in our political arena. Think red-state/blue-state, but a lot more serious. (Well, romanization is arguably NOT one of those serious issues). The fact is, as the South becomes more Green and the North becomes more blue, you’re going to see romanizations based on their local population’s stances on China. To adopt a singular standard would not only have bad implications for the in-country divide, but also cross-strait relations.

    P.S. “Tamsui” comes from the Dutch rendering of “Tam-tzui”, as they had no way to render “tz”.

  10. Yes, I realize that the pinyin division is based on political leanings, but that doesn’t make it any less stupid. Don’t forget that foreigners are the only people who even know pinyin – I doubt one Taiwanese out of a hundred does. Do you think that politicizing an issue that is basically invisible and incomprehensible the voters is really worth it?

Comments are closed.