MOE approves Taiwanese romanization; Tongyongists protest

Years of valuable time has been lost in the squabbling over romanization systems for Taiwanese. And that squabbling will no doubt continue, as the links below make clear. But an important step was taken on Thursday. Finally, finally, Taiwan’s Ministry of Education has approved a romanization system for Taiwanese: Tái-luó-bǎn Pīnyīn (台羅版拼音), to give its Mandarin name.

I’m already on the record as having called Tongyong Pinyin, in its various incarnations, a national embarrassment for Taiwan, so I won’t bother to disguise the fact that I got a real kick out of the fact that the Tongyong Pinyin scheme for the Taiwanese language was roundly rejected. I know that more than a few readers of Pinyin News will be cheering this news. For many, this has as much or more to do with the methods used to push through the much-despised Tongyong Pinyin system for Mandarin than any defects, real or imagined, in the Tongyong Pinyin system for Taiwanese.

Predictably, Yu Bo-quan (余伯泉, I’ve given up bothering to figure out which of the various spellings for his name he’s using now), the main person behind the Tongyong romanization systems, is unhappy. Reportedly, after it was clear things were not going his way he stormed out of the meeting. After he left the new system was approved unanimously.

Yu’s remarks make clear the political nature of his approach.

Tái-luó-bǎn pīnyīn xìtǒng zuó chuǎngguān chénggōng hòu, Yú Bó-quán qìfèn de shuō, Tái-luó xìtǒng de qǐyuán shì Táiwān Mǐnnányǔ yīnbiāo xìtǒng (TLPA), shì Guómíndǎng shídài de chǎnwù, ér 2002 Tōngyòng Pīnyīn shì Mínjìndǎng zhízhèng nèi tōngguò de, zhìyí wèihé Jiàoyùbù wúfǎ hànwèi zhízhèngdǎng de Mǐnnányǔ pīnyīn xìtǒng zhǔzhāng, Jiàoyùbù duànrán tōngguò Tái-luó-bǎn, Táiwānyǔ Tōngyòng Liánméng hòuxù jiāng zhǔnbèi kàngzhēng. (台羅版拼音系統昨闖關成功後,余伯泉氣憤地說,台羅系統的起源是台灣閩南語音標系統(TLPA),是國民黨時代的產物,而二○○二通用拼音是民進黨執政內通過的,質疑為何教育部無法捍衛執政黨的閩南語拼音系統主張,教育部斷然通過台羅版,台灣語通用聯盟後續將準備抗爭。)

That doesn’t sound all that far from calling those on the committee dupes of the KMT, which isn’t likely to win him any friends with those in power. But it may well be that by this point he has so alienated others he thinks he has nothing to lose.

Apparently Tongyong for Taiwanese will retain something of a foothold in southern Taiwan. (See source no. 8 below.)

Later, I’ll try to put up more about just what system was approved and under what circumstances it will (and will not) be used — unless the ever-knowledgeable a-giâu beats me to it.

Because there’s a lot of confusion about Tongyong, a few notes are in order:

  • Tongyong is not one romanization system for all the languages of Taiwan but rather a group of related systems, some of which could be said to work better (or worse) than others.
  • When Tongyong (for Mandarin) was officially approved in Taiwan in 2002, the Tongyong system for Hakka also received approval but not the Tongyong Pinyin system for Taiwanese.
  • As the vote should make clear, plenty of strong supporters of romanization (and other scripts) for Taiwanese have never much cared for Tongyong.


  1. Tái-luó-bǎn pīnyīn míngnián shànglù; Jiàoyùbù duànrán dìng’àn; Tōngyòng liánméng jiāng kàngzhēng (台羅版拼音明年上路 教育部斷然定案 通用聯盟將抗爭), Píngguǒ Rìbào (Apple Daily), September 29, 2006
  2. Guóxiǎo lǎoshī: xiāngtǔ yǔyán zuìhǎo zìrán xuéxí (國小老師:鄉土語言最好自然學習), Liánhé Xīnwén Wǎng, September 29, 2006
  3. Zuóyè zuìxīn: Mǐnnányǔ xiāngtǔ jiàoxué quèdìng cǎi Táiwān Mǐnnányǔ Luómǎzì pīnyīn (昨夜最新:閩南語鄉土教學確定採台灣閩南語羅馬字拼音), CNA, September 29, 2006
  4. Táiyǔ Tōngyòng liánméng kàngyì Jiàoyùbù cǎi Mǐnnányǔ Luómǎ pīnyīn (台語通用聯盟抗議教育部採閩南語羅馬拼音), CNA, September 29, 2006
  5. Mǐnnányǔ xiāngtǔ jiàoxué quèdìng cǎi Táiwān Mǐnnányǔ Luómǎzì pīnyīn (閩南語鄉土教學確定採台灣閩南語羅馬字拼音), CNA, September 29, 2006
  6. Mǐnnányǔ pīnyīnfǎ quèlì: Luómǎ pīnyīn shèng chū (閩南語拼音法確立:羅馬拼音勝出), Zhōngguǎng Xīnwén Wǎng, September 29, 2006
  7. Pāibǎn dìng’àn! Jiàoyùbù tōngguò Mǐnnányǔ jiàoxué; cǎiyòng Tái-luó pīnyīn (拍板定案!教育部通過閩南語教學 採用台羅拼音), Dōngsēn Xīnwénbào, September 29, 2006
  8. Nánbù sì xiàn-shì dǐzhì; Tái-luó pīnyīn jīn chuǎngguān (南部四縣市抵制 台羅拼音今闖關), Zhōngshí Diànzǐ Bào, September 29, 2006

20 thoughts on “MOE approves Taiwanese romanization; Tongyongists protest

  1. Well, hopefully this is a step in the right direction. I’m very much looking forward to hearing more about this system, how it works and where it will be used.

  2. Yeah, I bought the same ????. It was talking about the argument of “P as ?” vs. “P as ?”.

    BTW, Mark, your old site had a lengthy page about various incarnations Tongyong had taken for Mandarin. Have you moved that onto this site yet?

  3. Mark, I don’t remember having had much detail about that. I recall that Tongyong for Mandarin originally had zh- instead of jh-. Yu Bo-quan changed to the latter against the advice of some of the linguists working with him. Dan Jacobson has some remarks on the changes in Actually, it was I who influenced many of Tongyong Pinyin’s critical changes. (Scroll down that page to find that headline.)

    I tend to avoid looking at my old site, for fear of all the mistakes I’ll find. But I know I need to read through it again soon, if for no other reason than the anti-Hanyu screed mentioned in a recent post quotes from it several times.

  4. Hopefully this is a sign of a backlash against Tg Pinyin in general. Maybe when the KMT regain power, Tg (cursed be its name) will be stamped out once and for all, replaced with Hanyu Pinyin.

  5. I ran into Yu Bo-quan, the inventor of Tongyong Pinyin, earlier today. We had a very pleasant conversation about romanization, though that may well be because he didn’t seem to know who I am or what I write. Regardless, in the spirit of niceness I’m going to try to be less cranky about Tongyong for at least a little while.

  6. I haven’t read any of the articles you’ve so thoughtfully linked to yet, but from what I understand through the grapevine, there was basically a rebellion on the part of Taiyu activists who wanted to stick with what they know: the so-called “church romanization” system. This avoids the problems of a generation gap in literacy, and allows them to help their children with their homework. The new system isn’t pure “church romanization” but seems to be close enough so as to appease this constituency.

    Can anyone confirm that this is, indeed, what happened?

    It has always made sense to have two systems, since there can never really be “one system” due to the different phonetic values of the same symbols in each of the languages.

  7. If this report is to be trusted, the southern DPP officials have now (grudgingly) given their approval of the MOE decision, which may sound a bit odd given that education is quite centralized on the island (but then local DPP administrations were pushing mother tongue issues at a time when the KMT-held central government was vehemently against). My suspicion is that nearly all pols have absolutely no idea how romanization works, and therefore are not competent to give their approval or disapproval. I’d be happier if they’d just say, “It’d cost money and we prefer the money go to this or that”.



    Indeed, the new proposal is a hybrid of two fairly similar romanization systems. For the sake of some readers, I’ll run the risk of over-explaining the background below, so please bear with me. One is, of course, the Peh-oe-ji (POJ) system traditionally used by the Presbyterian Church (PCT; but now largely irrelevant within the church except as a historical/cultural emblem). The constituency is now, by informal estimates, mostly not affiliated with the PCT or Christianity for that matter. The other system is the TLPA, which is based on the POJ, but with a few consonants and vowel clusters modified. The most significant difference is the replacement of tone diacritics by numerals. This makes it suitable only as a phonetic system, not a full-fledged orthography. A theory among POJ enthusiasts goes that the KMT promulgated the TLPA in order to undermine the PCT (which has historically been anti-KMT). I personally think the KMT was merely trying to appropriate the issue, just as Lee Teng-hui hijacked the DPP’s “Join UN” issue.

    Anyway, the Tailuo (TL) system pretty much sticks to TLPA’s changes but retains POJ’s diacritics. Thus, whereas POJ would write “goá” (I/me) and TLPA “gua2”, TL now writes “guá”.

    It’s clear to me that without Tongyong’s “insurgency”, neither POJ nor TLPA would likely have gotten together to come up with this consensus proposal, given the prior history of interpersonal animosity between leading proponents, the tension between POJ’s perceived “Christian-ness/foreign-ness” and the TLPA’s perceived “pro-KMT” and “academic” root, and so forth.

  8. I’ve made tentative arrangements to speak with one of the members of the committee involved. He’s quite knowledgeable about romanization and Taiwanese. It should be an interesting talk. But we probably won’t have a chance to meet until a week or two after the holidays.

  9. Hi Y’all,

    Maybe you will wonder who is this dumb daring to post and ask a question soooo stupid…

    Anyway, let’s do it. I am trying to learn Taiwanese and I can’t find two books with the same romanization. Besides, as I can’t speak Taiwanese, I have no idea about how to find an “accurate” book.

    You can quite easily find Mandarin teachers but it’s quite hard to find a Taiwanese ( I mean Holo of course) teacher. And when you start talking about learning how to write and read, they run away.

    Here is my question: Which system should I use? Do you know any association teaching Taiwanese (in Ping Tung) with this sytem? (Oops, I said one question)

    My problem is actually more related to these fights about romanization than the language by itself.

    Cyril (Hsia)

  10. I am going to be somewhat contrary and recommend another system. I have studied several Taiwanese romanization schemes for over a decade, and in recent years have fallen into the MLT camp:

    After several years of trying to teach English-speaking youths in the US Church Romanization (POJ), and failing each time, a group of Taiwanese Americans decided to try out Modern Literal Taiwanese, and have reported much greater success.

    It’s not easy to explain the advantages of MLT over POJ, so I would simply ask you to check it out yourself. At first glance, there is not much to recommend, because MLT looks alien and awkward. However, people begin developing proficiency in MLT very rapidly—much more so than POJ. If young children and teenagers are any guide, they demonstrate much better reading comprehension and recall using MLT than POJ.

    MLT was invented in the 1930s by a group of Taiwanese students and scholars, who wanted to improve upon the deficiencies in POJ. Prof Liim Keahioong, the main advocate for MLT, was one of those students.

  11. Hsia,

    Pe?h-?e-j? is the most widely used romanization system and the one I would recommend you use. The majority of Taiwanese/Hokkien-language books you will encounter will be written in Pe?h-?e-j?.

    The romanization system mentioned in the article above is Tâi-ûan Lô-má-j? (Tâi-lô), the standard sanctioned by the government for use in schools. The differences between it and POJ are mostly minor and it is relatively easy to read Tâi-lô if you
    already know POJ and vice versa.

    Wikipedia article on POJ:

    Here is a useful conversion table of different romanizations:

    S. C. Goh,

    MLT seems to indicate tones with different spellings instead of the diacritical marks used by Pe?h-?e-j? and Tâi-lô-j?. It reminds me of Y. R. Chao’s Gwoyeu Romatzyh system for romanizing Mandarin Chinese which also uses “tonal spelling.”

    The main advantage of tonal spelling is that it may be easier to memorize the difference between si, sy, sii, six, and sie than s?, si, sî, sì, and sí (yes, poem, time, four, die). It is also easier to input text into a computer with tonal spelling than deal with all those annoying diacritical marks.

    The drawback is that such tonal spelling are complex and require the reader to identify which letters are used for tones. The letters used to represent tones also change with different vowels.

    However, regardless of its benefits or drawbacks, the problem with learning an obscure system such as MLT is that it is rarely used and few (if any) books and textbooks are written in it.

    Wikipedia article on MLT

    More info on Gwoyeu Romatzyh:

  12. Paul,

    “The drawback is that such tonal spelling are complex and require the reader to identify which letters are used for tones. ”

    This is an often-cited criticism, and which is probably accurate The start-up cost of learning MLT is higher than that of POJ in some ways. I know when I first saw MLT, I was taken back at how strange and unaesthetic it appeared. Once, one overcomes that initial hurdle, though, MLT continues to confer far greater benefits than POJ.

    Here is an experiment that I have proposed to many POJ practitioners: Have teams of fully-trained POJ adherents compete against an MLT team in several measured trials.

    Trial A (Transcription)

    A sample piece of text of is read allowed by a neutral party. A member of each team then transcribes the spoken speech using the team’s orthography: POJ or MLT. The writer than passes his text to a second member of his own team, who then proceeds to read it out loud. The trial is judged on both speed and accuracy.

    Trial B (Translation)

    A newspaper article written Chinese text (Mandarin) is presented to a member of each team, who then proceeds to translate the text into Taiwanese . The resulting translation is presented to a second member of the team, who then reads it out loud in Taiwanese. This trial is also judged on speed and accuracy.

    I’ve only tried this out informally, but to me it is clear that there is a clear quantitive difference in reading and writing ability among these orthographies. The great difficulty is in actually getting POJ practitioners to agree to participate in this experiment becaue they are often very unconfident about their own abilities in POJ, despite their years (sometimes decades) of practice.

    “However, regardless of its benefits or drawbacks, the problem with learning an obscure system such as MLT is that it is rarely used and few (if any) books and textbooks are written in it.”

    The same may be said of POJ. There is such a lack of literature, and so few proficient practitioners in POJ, that it is almost as starting from a blank slate in either case. Note, though, I am not against POJ. I think that some Taiwanese writing is better than no Taiwanese writing; furthermore, writing systems often change with time. Indeed, I suspect that even if POJ should become accepted, people will start to gravitate towards a more MLT-like system, as the demand for proficiency increases.

  13. Pingback: Pinyin news » Taiwan premier calls for support for romanization of Taiwanese

  14. So glad I found this in 2017!

    I just started to learn Taiwanese romanization and wanted to know how MOE came up with ?? (tâi-lô), when there was a perfectly good and well known romanization in pe?h-?e-j?.

    I should have suspected it was politically-based decision as always.

    MLT seems very interesting. It must be faster to write and read because tone sandhi is incorporated into the script, so on-the-fly conversions wouldn’t have to be made by the writer and the reader.

    The official romanization should have been something totally traditional and established to keep backward compatibility, such as pe?h-?e-j?, or something easier to type and display on modern computers with diacritic-free script, such as MLT. Yet they chose neither.

Leave a Reply

Your email address will not be published. Required fields are marked *