Q and A about romanization in Taiwan

OK, so Wade-Giles isn't official. But it's still used more than any other system in Taiwan.

No, not really. What's usually seen in Taiwan is a bastardized form of Wade-Giles. The proper use of the apostrophe (') is essential to Wade-Giles. Because the apostrophe is almost always omitted, the spellings that result are impossible to read with certainty. In other words, it's close to useless.

Certainly. 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 ㄍㄨㄉㄧㄥ), K'u-ting (Kuding ㄎㄨㄉㄧㄥ), K'u-t'ing (Kuting ㄎㄨㄊㄧㄥ), or Ku-t'ing (Guting ㄍㄨㄊㄧㄥ). (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. Here is a chart of the possibilities, with tones indicated by numbers and the correct pronunciation marked in green.

	ting1	ting2	ting3	ting4	ding1	ding2	ding3	ding4
ku1	ku1ting1	ku1ting2	ku1ting3	ku1ting4	ku1ding1	ku1ding2	ku1ding3	ku1ding4
ku2	ku2ting1	ku2ting2	ku2ting3	ku2ting4	ku2ding1	ku2ding2	ku2ding3	ku2ding4
ku3	ku3ting1	ku3ting2	ku3ting3	ku3ting4	ku3ding1	ku3ding2	ku3ding3	ku3ding4
ku4	ku4ting1	ku4ting2	ku4ting3	ku4ting4	ku4ding1	ku4ding2	ku4ding3	ku4ding4
gu1	gu1ting1	gu1ting2	gu1ting3	gu1ting4	gu1ding1	gu1ding2	gu1ding3	gu1ding4
gu2	gu2ting1	gu2ting2	gu2ting3	gu2ting4	gu2ding1	gu2ding2	gu2ding3	gu2ding4
gu3	gu3ting1	gu3ting2	gu3ting3	gu3ting4	gu3ding1	gu3ding2	gu3ding3	gu3ding4
gu4	gu4ting1	gu4ting2	gu4ting3	gu4ting4	gu4ding1	gu4ding2	gu4ding3	gu4ding4

Those who know nothing of tones or of Wade-Giles, however, will simply say Kuting (ㄎㄨㄊㄧㄥ), which is of course wrong and does not aid communication. Why, oh why does the government keep giving foreigners the wrong information? Or is it just that no one in the government knows how to pronounce Chinese?

I cannot stress this too much: Other than perhaps Taipei ㄊㄞㄆㄟ for 台北 (hanyu pinyin Taibei ㄊㄞㄅㄟ) the only proper "English" pronunciation of a Chinese name is same as the Chinese pronunciation. There is no such thing as a proper "English" pronunciation of Chinese. Everything else brings unnecessary complications and confusion. The tones needn't always be rendered to ensure communication, but the vowels and consonants must.

the only proper "English" pronunciation of a Chinese name is same as the Chinese pronunciation.

The whole reason for having a romanization system is to allow those who cannot read Chinese characters to know the pronunciation of characters. The purpose of romanization is communication, plain and simple, not sticking a few letters from the roman alphabet up on signs because someone thinks it looks cool, not face (which in any case is not received if the system is poor or its implementation is botched), and most certainly not petty nationalism.

To force locals to speak pidgin English and foreigners to speak mock Chinese to attempt to communicate with each other is simply ridiculous, esp. when the use of hanyu pinyin can eliminate most misunderstandings. Actually, Taiwan's approach to this has been far more than ridiculous; it is patronizing of and offensive to foreigners.

The apostrophe is the only thing that allows users to dintinguish between p from b, k from g, ch from j and several others.

For example, the sound most people associate with the letter b is written p in Wade-Giles. The sound of the letter p, on the other hand, is written p'.

There are sound linguistic reasons behind the Wade-Giles approach. Unfortunately, however, its complete failure as a popular system is beyond doubt. The approach of t=d, p=b,k=g, etc., is unnecessarily counter-intuitive and misleading; and it has led to far too much "English pronunciation."

What about the so-called nick-numbering system, in which some of Taipei's major roads were given numbers, like in New York City? Surely that helps foreigners.

I'm willing to believe that whoever came up with the idea meant well. But the fact remains that it was a bad idea that should never have been implemented.

Instead of giving foreigners a proper romanization system that would allow them to know how the street names are pronounced in Mandarin, thus facilitating communication with locals, the numbers create yet another layer of confusion. It's a useless system that neither locals nor foreigners know, much less use.

Anyone who asks directions to, for example, "Seventh Boulevard and Twelfth Avenue" is almost certainly going to be met with incomprehension. (You're an educated person, but do you know where that is? I doubt it.) Similarly, "Twelve A and Seven B." Who knows this stuff? No one. Nor is anyone ever going to bother learning such useless nonsense. What a joke!

But the nick-numbering system is great for people who take taxis, because now all a foreigner has to do is point on the nicknumbering map that taxi drivers have.

First, no matter what the Taipei City Government may say, most taxi drivers do not have this certain map. But even for those who do, the numbers don't do much.

Because the communication involves pointing at a map. A map that labels 南京東路 "Seventh Blvd." and 敦化北路 "12th Ave." does not magically become far superior to a map that labels the same streets "Nanjing E. Rd." and "Dunhua N. Rd." It is no harder to point at "Nanjing E. Rd." and "Dunhua N. Rd." than to point at "12A" and "7B," and it certainly makes a lot more sense.

If a map has a proper romanization system, then the nick-numbers are redundant. Worse still, they are another source of possible confusion -- not to mention a big waste of taxpayers' money. (And don't forget: resident foreigners have to pay taxes, too!)

Besides, even if people knew the system, it still works only for major intersections. Anyone who wants to go anywhere else is out of luck. The nick-numbers have no sections (段), and therefore cannot be used for addresses. They are, quite simply, useless.

I've often heard that. But it's still nonsense. For example, I could come up with a system for representing the sounds of Mandarin by using only icons of the faces of Chen Shui-bian, Ma Ying-jeou, James Soong, Lien Chan, Kofi Annan, Michael Jackson, Hello Kitty, Karen Carpenter, the members of Boyzone, etc. It would thus be both local and international, hooray! It would also have the virtue of being cute (ke-ai). Surely this would be Taiwan's salvation! I believe I will call it Ke-Ai Pinyin (可愛拼音). Once we apply it consistently, the problem will be solved, right? Perhaps I should go plan my marketing campaign now.

Come on, don't be ridiculous. Besides, it's not a romanization system if it doesn't use Roman letters (a, b, c . . . ).

Aw, don't you like my idea? Maybe if I were given a hefty government grant to study the idea for several years, and then the opportunity to lobby politicians I could perfect it.

My point is that consistency alone is not an answer. Even if Taiwan consistently used Ke-Ai Pinyin, if the system doesn't make sense to those who need it most and it isn't supported by readily available reference materials, it doesn't do any good. It doesn't do much good, either, if the foreigners who are supposed to benefit from it think it's stupid. The situation is the same for tongyong pinyin.

But if it's romanization you want, how about using Gwoyeu Romatzyh? It's linguistically sound, and was Taiwan's official system for many years, so there's a precedent.

Few people have. The chief recommendation of Gwoyeu Romatzyh is that it incorporates the tones into the words without using extra marks. The rules for doing this, however, are more than a little complicated. Here's how the GIO describes it:

But that's just the beginning. It's actually even more complicated. Anyway, as the GIO notes, "The difficulties in popularizing this system are obvious."

If you want consistency, then the MRT largely has this in its bastardized Wade-Giles. I say "largely" because a few stops are different: Nanking E. Rd. (using the old Chinese postal system, instead of the Wade-Giles Nan-ching) and Tamshui (which is simply wrong).

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.

OK, let's get back to why you think consistency -- any consistency -- isn't enough to solve the problem.

Gladly. As I was saying, the MRT is basically consistent in using bastardized Wade-Giles. But that's not particularly useful, because, as I explained earlier with the example of Kuting station, even those few who know the rules of Wade-Giles will find it impossible in most cases to know the real pronunciation of the Mandarin.

OK, so some systems are better than others. Consistency is necessary but isn't enough by itself. But what's wrong with using tongyong for Mandarin?

Tongyong can be used for all the languages and dialects of Taiwan: Mandarin, Taiwanese, Hakka, and the languages of the aborigine tribes.

What "people say" isn't always reliable. As I've been showing, Tongyong supporters have made many unsubstantiated, exagerrated, and misleading claims. Why should this one be any different? Of all the many people who repeat this universality claim, how many of them have seen the evidence with their own eyes? Does it even exist? Where is it? Where is the Tongyong scheme for Atayal? For Sedeg? For Saisiyat? For Bunun? For Tsou? For Kanakanabu? For Saaroa? For Paiwan? For Puyuma? For Ami? For Yami? And for all the other languages and dialects of Taiwan?

Let's see them -- in English as well as Chinese -- and with clear, scientific, hype-free explanations and examples. Show how these schemes -- assuming they really exist -- compare to other systems that have been used for these languages.

all methods of romanizing Chinese are systems that must be learned to be used properly

Given the exagerrated claims made by Tongyong supporters, I'm inclined to be skeptical of this, especially regarding the languages of Taiwan's tribes, which have sounds not found in any dialect of Chinese.

It cannot be stressed too much that all methods of romanizing Chinese are systems that must be learned to be used properly. For a system to succeed on the popular level (and not just on an academic one) it is desirable that it should match standard English usage, where possible and practical. But there is no system, none, that is such a perfect fit with English that a native speaker of English who does not know Mandarin can with certainty pronounce properly without study. Not even the Yale system, which was devised with American English speakers in mind, is up to the task. This does not mean that Yale or any other system is therefore "defective" or "deficient." English and Chinese are completely different languages. Anyone who expects or demands a perfect, unambiguous fit between English and pinyin has either not thought about the matter much or is a fool.

Looked at from a scientific standpoint, the English alphabet isn't even a good match for English itself. The letters and combinations of letters are made to carry many, many different sounds. Take the letter a, for instance. Compare the a's in ago, gas, ate, any, all, homage, was, and father. The sound of the a is different in each one.

The consonants also change their pronunciations depending on context. There's the well-known example of how ghoti can represent fish. (Gh as in cough, o as in women, and ti as in friction.)

input methods

What about input methods? Tongyong supporters have said that their computer input method is the strongest point in favor of Tonyong pinyin, because q, x, and z don't need to be typed as often.

This only serves to show how insubstantial the arguments in favor of Tonyong are, because this point is so weak as to be laughable.

There are already several methods people can use to key Chinese characters into a computer. These methods can and do coexist on individual computers with no trouble or confusion. To switch a computer from one to another is a simple matter that takes but a moment. Most people in Taiwan know this for themselves through personal experience. If you are reading this on a Chinese Windows system, you probably have one of the following images in the bottom right corner of your screen.

In order for input methods to be roughly comparable to the romanization mess, you'd have to imagine something like this. As you are typing Chinese into your computer, your system randomly switches input methods every few seconds. To make matters worse, you would not be warned or even have the new system identified; you'd be left with little more than trial and error. And as soon as you finished one character the whole farce would begin again.

To paraphrase the well-known feminist simile: Taiwan's screwed-up romanization system, street signs, company names, place names, MRT stations, personal names, etc., etc., ad nauseum, need a new computer input method like a fish needs a bicycle. (Or should that be a "ghoti"?)

Tongyong vs. hanyu pinyin is most certainly not a battle of equals. hanyu pinyin is the one and only international standard for romanizing Mandarin. It is used by the United Nations, the U.S. Library of Congress, and countless other institutions worldwide.

It is now the standard for translation dictionaries printed everywhere but Taiwan (and even a few here, too). It is the standard for students of Chinese throughout the world. All international sinologists know hanyu pinyin. It is supported by a large number of websites in a variety of languages. There are dozens of hanyu pinyin dictionaries. It has been used successfully for nearly half a century. Before its initial adoption it had to demonstrate its superiority to a huge number of other proposed systems. It is respected. It is linguistically sound. The foreign community in Taiwan wants it.

Tongyong, on the other hand, has basically nothing behind it. Nothing. A few sites give a little information. Of the few English sites that give the most information about this supposed godsend of internationalization, not one is in favor of it -- quite the contrary.

Taiwan culture

I've heard it claimed that anything other than tongyong will cause Taiwanese culture to be lost.

Baituo! This is an astonishing assertion, and one I think is demeaning of local culture. What Taiwanese culture has been preserved in tongyong? What of anything has been done in tongyong, other than a few batches of error-filled, confusing street signs written using a now abandoned version of tongyong.

If Taiwan does not use tongyong to romanize Mandarin, then Taiwanese culture will be destroyed? What nerve!

learning multiple systems

But if Taiwan uses hanyu pinyin, won't students have to learn more than one system?

If students need to be able to read romanized Mandarin, Taiwanese, Hakka, Ami, etc., then, yes, hanyu pinyin isn't enough and they will need to learn other systems. If tongyong can really take care of all the other languages accurately and efficiently (which I very much doubt), good for it.

But since when have students had to learn even one romanization system? Most locals do not know any system at all. They usually engage in guesswork. This is the main reason the street signs, company names, etc., have remained in such a mess.

But why should elementary school students have to switch from BoPoMoFo to hanyu pinyin?

This is an unrelated issue. If BoPoMoFo is working well for Taiwan's students (as it seems to have done), there's no compelling reason for change. But there can be no doubt that the mishmosh of romanization systems and additional layer of misspellings are a complete, embarrassing failure that must be corrected.

Some people will never accept spelling their names using a method associated with China.

People should be allowed to romanize their names in whatever language/dialect they prefer. The supporters of hanyu pinyin aren't trying to force people to spell their names as someone in China would. But I do suggest that people who would like their Mandarin names pronounced correctly by foreigners should use hanyu pinyin, and not any other system.

Anyway, this is again a rather silly argument, because it is inevitable that some -- indeed many -- names will be romanized in exactly the same way in hanyu pinyin and tongyong or whatever other system.

Let's look at the most common family names in Taiwan: 陳,林,黃,李,張,王,吳,劉,蔡,and 楊. Half of the population of Tawain has one of these 10 names. All but two of them are identical in their hanyu and tongyong spellings.

The opponents of PY often like to say, "Oh, hanyu pinyin and tongyong are very similar. So it won't be a big problem to use tongyong, because so much will be the same." But in the next breath they'll say, "People will never permit their names to have a hanyu pinyin spelling." Quite simply: You can't have it both ways.

But maybe that wouldn't happen as much because in tongyong, words are written differently.

CapITalIzIng The InITial LetTer Of EvErY SylLaBle Makes ReadIng HardEr, Not EaSiEr.

CapITalIzIng The InITial LetTer Of EvErY SylLaBle Makes ReadIng HardEr, Not EaSiEr. And It Is ObNoxIous And QuickLy BeComes TireSome To Read. But If SomeOne InSists On A ComProMise, Then MayBe This Could Be It: Keep HanYu SpellIng But Use This StuPid RollEr CoasTer Text ForMat. At Least The Sounds WouldN't Be As ConFusIng As OthErWise. I've Got To Stop This Now, BeCause I'm FeelIng SeaSick.

Let's get back to what you said a moment ago. Tongyong and hanyu pinyin are 85 percent similar. So why the big objection to tongyong?

Even if they were 95 percent similar, there would still be a problem. When you change a standard, even just a little, confusion is the result. And I don't mean confusion just 5 percent of the time, but confusion most of the time. Also, in this computerized world, any change whatsoever can render improper results.

But I should take this opportunity to point out something crucial: Tongyong and hanyu pinyin are not 85 percent similar! This frequently cited figure is, quite simply, bullshit. At the syllable level, the differences are 19.5 percent, giving similarities of 80.5 percent, not 85 percent.

And consider that some syllables appear more frequently than others. If we look at syllables from a real-world standpoint, the similarity drops to 72.5 percent.

Moreover, most Chinese words are not monosyllabic, which means they combine syllables and therefore are even less likely to match in different romanization systems. Nearly half (48.84 percent) of Chinese words do not match in hanyu pinyin and tongyong pinyin (source: "Similarities Between Tongyong Pinyin and Hanyu Pinyin"). As the author of the most scientific studies on the topic concludes, "Obviously, there is no reason to say that hanyu pinyin and tongyong pinyin are compatible."

You think so? Remember my Ke-ai Pinyin? I've just decided to rename it. The new term for it is "The Taiwan Freedom, Democracy, Preservation of Taiwanese Culture, Super-Sophisticated, Internationalized, U.N.-Membership-Gaining, Defeat the Evil Communists, Really Easy to Learn, Makes Your Children Smarter, Increases Your Sex Appeal Pinyin."

Using tongyong pinyin would be no more useful an expression of national identity than requiring people to wear their underwear on the outside of their clothing.

The "differentness" Taiwan would achieve by adopting tongyong pinyin would be no more beneficial than, for example, deciding that stoplights on the island will no longer be red, yellow, and green like the system used in China (and almost everywhere else), but blue, orange, and pink. No matter how many blue-orange-pink supporters cry "See how wonderfully unique we are!" the rest of the world will still look on this as a ridiculous change from the standard.

Using tongyong pinyin would be no more useful an expression of national identity than requiring people to wear their underwear on the outside of their clothing. Difference for its own sake would make Taiwan look petty and ridiculous. Taiwan does not benefit from looking ridiculous.

That's a lovely cliche. But it doesn't do anything to address the issue: Taiwan, whether or not most locals realize this, has an enormous problem rooted in its haphazard, slipshod, inconsistent, confusing and even patronizing approach to romanization.

Variety? My God! Taiwan's screwed up romanization has nothing if not variety. Pa-te Rd., Bade Rd., Pateh Rd., Patch(!) Rd., Bader Rd., etc. Besides, the anti-hanyu crowd are the ones who oppose variety. They tout a one-size-fits-all system. (Even though their claim is a lie.)

In short, tongyong pinyin, despite claims to the contrary by its supporters, is:

Questions and answers about Taiwan's romanization situation, and information about why tongyong pinyin is not what it claims to be

input methods

Taiwan culture

learning multiple systems