Neither Fish Nor Fowl: Radical Romanization

tan.jpgThere are a multitude of romanization systems out there for Taiwanese which do the job they are designed for, so you might think that there is no reason to go around inventing new systems when others accomplish the work perfectly well. Well, some people would disagree with you, specifically in this case one Mr Tân (the gentleman pictured on the right).

Recently I was given a series of Mr Tân‘s books by Mark of which outline a new approach in to the “problem” of finding an effective written system for Taiwanese (I say new, but I think the books were published in the late nineties, so we’re talking relatively here). The system is interesting in that it combines two different approaches from the tradition of Taiwanese writing, although it has to be said that the attempt leaves a lot to be desired.

One of the issues always mentioned in connection with writing any Chinese language in romanization is that of information loss – the pro-character types assert that characters contain more information, more succinctly expressed, than any romanized system can. They would say that Chinese languages have such a high level of homogenity that no alphabetic system can convey the layers of meaning necessary. As an example, the Chinese character input system on my computer brings up a total of 247 possibilities for the Mandarin syllable “shi“, and even if we narrow it down by tone to “shì” there are still 36 possibilities for that one sound.

Cover of Crkunl - a manual to Tan's romanizationTo combat this perceived defect in romanizations of Chinese languages, the inventor of this system has combined romanized writing with a system of semantic signifiers which indicate the category to which the sound belongs. To this end he has created a total of 40 categories into which words can fall, such as the “woman” category, the “vital” category, the “electricity” category, and so on. The category of the syllable is indicated by a letter or symbol after the sound. It’s as if the English word “boy” was written “boy♂” and “lightning” was rendered “lightning↯”.

Examples of words given in his books include “bòΛ” (cloth; written as pò͘ in the standard POJ romanization), “dwā%” (big; toā), “cuib” (open; khui), “cỳ→” (go; khì) and “kàᚑ” (to teach; kà).

The main problem with the whole system, besides the sometimes arbitrary assignation of words to categories, is the assumption that the greater number of homophones at the character level renders Taiwanese incomprehensible if written in romanization. This would be true if Taiwanese were a monosyllabic language, but in fact it is far away from being so, with the majority of both nouns and verbs in the language being either di- or trisyllabic. The system therefore does not address a need, or a lack in the existing romanization systems – meanwhile it does introduce another layer of complexity in to an already complex system. For a comprehensive dismantling of the “monosyllabic myth”, see the chapter of the same name in John DeFrancis’ book “The Chinese Language: Fact and Fantasy“.

For those who might doubt that written Taiwanese can be understood without the need for characters or the symbols which Mr Tân employs, it suffices not only to see the relatively large amount of material printed in the Pe̍h-ōe-jÄ« (POJ) romanization over the past 100 years, but also to note that Taiwanese speakers have no problems with verbal communication – so why would they struggle with a system (POJ) which represents the spoken language very accurately?

4 Responses to “Neither Fish Nor Fowl: Radical Romanization”

  1. […] Tailingua on romanising Taiwanese. […]

  2. Mr. Tan’s effort at least represents a new way of thinking and creativity, which most Hoklo speakers lack. POJ also requires special software and font. Want to see a way of writing Hoklo with just plain English keyboard? Take a look at

    See, what I fail to understand is, all Hoklo speakers today have a keyboard in front of their computers. Yet so few even dare to try to write down their own daily speech. As if writing Hoklo is a difficult task.

    Another thing I fail to understand is with all the Hoklo teaching in Taiwan, the Hoklo speaker still fail to appreciate the single most important feature of their language: the tonal phrase structure. When will they ever take notice that this is very feature that makes their language so unique and special?

  3. Akira says:

    Re: “Taiwanese speakers have no problems with verbal communication”

    Of course, when speaking, one can use gestures and emphasis to clarify confused meanings. One can also see the faces of the person being spoken to, and immediately sense any lack of comprehension.

    + + +

    Thanks for this great site.

  4. Pinyin Info says:

    Akira, the principles are the same, regardless if we’re talking about
    Taiwanese, English, Russian, or Swahili: a natural language is a natural
    language is a natural language.

    Surely you’re not under the impression that Taiwanese never speak with
    each other over the telephone, that there are no Taiwanese radio stations,
    or that blind Taiwanese cannot speak.

    Of course native speakers can become confused about what someone else is
    saying. But that happens in all languages.

Leave a Reply