Battle for Taiwanese

« previous post | next post »

From karts deffle:

Tái wén zhànxiàn


"Battlefront for Taiwanese writing"

Though that's what it must feel like to those who are trying to preserve their mother tongue, the English translation of the title of this magazine is not incorrect, but oh what a metaphor, both visual and verbal!

Selected readings


  1. Chips Mackinolty said,

    January 24, 2023 @ 3:31 am

    A curious cross reference to Iwo Jima!

  2. Qere said,

    January 24, 2023 @ 5:07 am

    I believe it would be more appropriate to romanize the title as "Tâi-Bûn Tsiàn-Suànn", since the journal is apparently written in Taiwanese Hokkien, even though the title is "language neutral" in a sense.

  3. Jonathan Smith said,

    January 24, 2023 @ 7:34 am

    @Qere one could hardly ask for a better example of why Chinese characters are the (ok, a) hill Taiwanese, etc., will die on — they are psychically speaking Mandarin-occupied

  4. Victor Mair said,

    January 24, 2023 @ 9:29 am

    @Jonathan Smith


    See the posts and comments by Lañitri Kirinputra and other Hoklo speakers (e.g., "Confessions of an Ex-Hokkien Creationist" [9/20/16]).

  5. Qere said,

    January 24, 2023 @ 10:35 am

    @Jonathan Smith
    Chinese characters are not exclusive to Mandarin or even Chinese languages in general. A Japanese speaker or a hanja-proficient Korean speaker would read this title easily, pronouncing it as "Taibun sensen" and "Daemun jeonseon" respectively. It may take a bit more to parse 台文 as 台灣文學 though.

    For Hokkien, however, it is sadly more true, mostly due to its orthographical tradition being interrupted, re-invented, and filled with mandarinisms (such as 的 for possessive ê, written as 個/个 in earlier works) along the way. Yet, it is more viable for Hokkien to reclaim the characters, rather than retreat to Pe̍h-ōe-jī, which is still strongly associated with the Church.

  6. Jonathan Smith said,

    January 24, 2023 @ 11:48 am

    Right; my comments pertain to the Chinese sphere.

    The situation is indeed annoying because — contrary to the intuition of many native speakers — there is obviously nothing about standard Mandarin which renders it inherently more suited to representation in Chinese characters than are Taiwanese, Cantonese, etc.

    However, the status of Mandarin within the Chinese sphere is such that Classical Chinese becomes pseudo-Mandarin (>> zhī hū zhě yě…), Japanese becomes pseudo-Mandarin (>> Ānbèi Jìnsān, Běntián, oh and Rìběn…), Korean in the same way becomes pseudo-Mandarin, etc. How could Taiwanese represented in Chinese characters expect to fare any better?

    This is an inevitable result of the socioeconomic position of Mandarin across the "Sinosphere," including of course in Taiwan — where practically *everybody's* Mandarin is at this point stronger than their Taiwanese, to say nothing of the writing they are exposed to (99.9%+ Mandarin) — indeed, it is only via Romanization of Taiwanese that people are truly forced to confront the fact that the language they are looking at on the printed page is not, in fact, pseudo-Mandarin.

    Funny comment from someone looking over my shoulder at a Taiwanese text — "台語?好像好難哦,好多生僻字!" :D

    So Taiwanese can either live, however marginally, in some alphabetic form, or die like Classical Chinese before it, as a pile of "生僻字".

  7. Jerry Packard said,

    January 25, 2023 @ 6:55 am

    “… there is obviously nothing about standard Mandarin which renders it inherently more suited to representation in Chinese characters than are Taiwanese, Cantonese, etc. …”

    I don’t see how you can say that. Just about any utterance in SM is easily transcribable in characters, something that can’t be said for Taiwanese, Cantonese, etc.

  8. Jonathan Smith said,

    January 25, 2023 @ 8:37 am

    @Jerry Packard
    key word *inherently*

    The standard happens to be easily represented in characters *because* it is the (arbitrary) standard, not because of some inherent quality. We might say the standard language, by virtue of its status, has tremendous character-sucking power; i.e., as novel morphemes enter the language, they instantly require written representation, making us feel as if the gap never existed and the connection to the written form is "natural".

    Notice how many CHIN 101 characters are relatively recent coinages or reapplications to fill such gaps: 你 他 嗎 也 這 那 的…. If the standard language of China were say a Southern Min variety, these would be nothing more than random crap, "偏僻子" made up to write some low-grade northern patois :D

    Probably I could have made my point clearer above by simply saying that while I understand it *seems* that using Chinese characters makes Taiwanese more accessible to a wider group, the truth is just the opposite — it makes Taiwanese much *less* accessible to a wider public.

    Give Taiwanese texts written in (A) characters and (B) Romanization to a crowd of learners, shaky half-speakers, etc., and observe what they produce:
    A: zombie mixture of Mandarin and Taiwanese, of course, with the subject focused on how to "read" particular characters "in Taiwanese"
    B: slow/mispronounced/otherwise shitty Taiwanese

    which is better?

  9. Jonathan Smith said,

    January 25, 2023 @ 8:38 am


  10. Victor Mair said,

    January 25, 2023 @ 8:54 am

    @Jonathan Smith:

    "So Taiwanese can either live, however marginally, in some alphabetic form, or die like Classical Chinese before it, as a pile of '生僻字'".

    Truer words were never spoken about the relationship between Hànzì 漢字 ("Sinoglyphs; Chinese characters") and Taiwanese language.

    For those who are unfamiliar with the term, "shēngpì zì 生僻字" means "obscure glyphs; uncommon characters; esoteric Chinese characters; rarely used Chinese characters".

  11. Jerry Packard said,

    January 25, 2023 @ 11:06 am

    @Jonathan Smith “… which is better?…”

    Is there a correct/intuitive answer? They seem equally good/bad to me.

  12. Qere said,

    January 25, 2023 @ 12:06 pm

    @Jerry Packard
    >Just about any utterance in SM is easily transcribable in characters, something that can’t be said for Taiwanese, Cantonese, etc.
    Most of the time Taiwanese (and especially Cantonese) can also be easily transcribable. Often, however, there are multiple ways to transcribe an utterance, but that is another problem.

    The orthography was not uniform even for Classical Chinese. That is especially the case with ideophones. Take for instance the word wui4 wong4* "to hesitate". Below is an incomplete list of its attested spellings found in 漢語大詞典 and 大漢和辭典:


    Even modern Mandarin is affected by it. Not too long ago I have seen a discussion about the word 眦毛 brought up in a forum. A person found it in a modern novel (obviously written in MSC) and asked about its meaning. Although it was not found in any dictionary, it was soon identified as Mandarin cīmáo[r]/zīmáo[r] "to make trouble; to be untameable". Turns out it was actually present in many general and dialectal dictionaries under various spellings: 呲毛,刺毛,齜毛,髭毛,滋毛 (often with final 兒s). Although the 現代漢語大詞典 labels the word as 方言, it seems to be common in North-Eastern and Beijing Mandarin.

    *read as huê-hông in Hokkien, huí huáng in Mandarin, kaikou in Japanese, 匣灰^平 匣唐^平 in Middle Chinese, etc.

  13. KIRINPUTRA said,

    January 30, 2023 @ 1:02 am

    @Jonathan Smith
    Your points are deep, but they intuitively assume the real battle is elsewhere. The common intuition is that the “battle for the sinographs” is a trivial can’t-win battle with little strategic upside and infinite downside.
    It’s probably true is that sinographs have winner-take-all characteristics in the context of the modern state. That makes them momentarily pro-Mandarin in the Taioanese-speaking context. But that’s just a symptom of a disease that we’re already fighting (or not).
    Macro-delinking Taioanese from sinographs a là Vietnamese COULD be a good idea. But it can’t be implemented without the kind of organization & resources that only a state (if not a nation-state) would have. (Not to say individuals shouldn’t go ahead and micro-delink….)
    And for a tribe with that kind of organization & resources, ending the psychic Mandarin occupation of sinographs would be child’s play in the long run (20+ years). Even the short run effects would be mind-boggling.
    We should also keep in mind that neither sinographs nor Mandarin (!) is (fundamentally) secular in the context of people psychically descended from Qing subjects, living in a psychically non-foreign land. There are profound confounding factors….
    Staying off that tangent, the intuitive idea is that Taioanese could fight better & survive longer “w/o sinographs”. What I’m pointing out is that — counterintuitively — the “lost” battle for the sinographs is just another section of the struggle, and not a particularly frivolous or unwinnable section.
    As an unsavory (and they all are) exhibit, consider how most fluently Mandarin-literate Taioanese speakers use romanization. If a guy named 旺 has “ONG” on his shop sign, he’s probably from Indonesia. A literate Formosan 旺 would feel invisibly compelled to “anglicize” 旺 as “WANG”, via Mandarin.
    ㊟ BTW, people whose Taioanese is stronger than their Mandarin are still abundant! There’s just a strong correlation between where they hang out vs where the Informed People don’t.
    ㊟ The Roman script is psychically English-occupied for most Taioanese speakers.
    Simply put, the interface between Taioanese & other languages is “Mandarin-occupied” — even when sinographs are taken out of the picture, such as when a Formosan writer writing in English writes that the borough of Bángkah “used to be called Bangkah”. It’s obviously still called that, but something complex & interesting tends to happen on the way to stating that simple fact in a psychically foreign language.
    At this stage of the war, all the hills are Mandarin-occupied, not just sinographs. Sinographs are just one of many hills that Taioanese is dying on.
    If the idea is to curb the all-pervading Neo-Chinese presence on Formosa (as was done by with the former Japanese omni-presence, although not so much by Formosans themselves), there is nothing to gain from “giving up” sinographs. And if the Neo-Chinese presence on Formosa is not curbed, there will continue to be no margin wide enough for community transmission of Taioanese & Hakka — full-blown languages even without sinographs.

    The point is counter-intuitive and I need to be concise, to let it breathe. A few tangents to acknowledge in closing, though….
    On reflexes: It’s possible for a Mandarin-educated adult to curb or even gut their Mandarin reflex to sinographs. Any serious Formosan learner of Japanese goes through such a process, rarely w/o success. It’s equally possible to curb the Mandarin reflex in favor of Taioanese or Hakka, but a LOT of anguish swirls around this idea, tellingly. (Bíphang Hôa 米芳華 vlogged about doing it, though.) The Mandarin reflex instilled at the age of 5 or 6 is a deep violation of the colonized mind; for some, the unspeakable beauty of romanized Taioanese is that it bypasses this problem in the context of Taioanese, which is the only context where autopilot doesn’t somehow seem to take care of things. (The fabric is warped, though, as most easily seen when young or middle-aged people want to talk or write about a Japanese person, place or thing in Taioanese.)
    (So while the idea of delinking sinographs from Taioanese is reasonable and seriously worth considering in itself, a lot of its emotional support seems to come from people wanting to avoid the anguish of having to confront the Mandarin reflex in an alley at night, w/o an army & a navy for backup.)
    On “cold sinographs”: There’s a stark difference between traditional sinographic Taioanese and what has emerged from erudite modern (post-1920 in the Formosan context) efforts to “write Taioanese in Zhongwen”. The latter is a dialect extension of 中文, packed with 生僻字 in accordance with Neo-Chinese notions of orthographic hygiene. Traditional sinographic Taioanese tends to recycle sinographs instead, and when graphs are coined, they’re coined w/o any effort to avoid shapes that may have been used elsewhere once upon a time. From a (say) 20th century Korean POV, traditional sinographic Taioanese should be much less esoteric-looking than even Mandarin or Cantonese, let alone the “Taioanese Zhongwen” show orthography.
    On sinographs vs romanization: There’s an assumption that the now-150-year coexistence of these two orthographies is unnatural & bad. I want to peel that assumption away for the purpose of this discussion (& beyond). Let nothing I’ve said here be wrongly taken to mean I’ve taken a side in the imaginary death match. And if there’s such a thing as a reasonable advocate of script unification, may they voice their reasons, for they never have. Nor do I mean each person should have to learn two scripts. Much respect to all who communicate & create in either script!
    P.S. Well said! “The standard happens to be easily represented in characters *because* it is the (arbitrary) standard, not because of some inherent quality.”

  14. Jonathan Smith said,

    January 30, 2023 @ 11:06 pm

    aha, the stank of my hyperbole attracts the elusive KIRINPUTRA :D

    I'm afraid that at heart my objection is selfish: I can't learn Taiwanese and other languages of interest to me via Chinese-character texts; accidental exposures to such cause the immediate, irreversible and literally palpable decay of what is at my age very hard-earned Taiwanese (etc.) brain-wiring.

    I thus conclude via confirmation bias that Chinese-character texts are terrible Taiwanese *learning* tools for the Mandarin-literate (and maybe more generally), whatever purposes they may serve / significances they may carry *within* the communities that produce and share them. As I see it, such texts simply can't go forth efficiently to interested non-initiates, and can't go down to a Mandarin-acculturated next generation.

    But it is indeed stupid of me to (appear to) advocate some particular direction or solution in this respect given my literal and sociocultural juwairen status, or to "miss the forest" of which you speak so eloquently "for the tree" that is the script question in particular.

    Random —

    re: Roman script — I think the situation is different. IMO it is not so much that alphabets are associated with English/foreignness per se as it is that "native" users of Chinese script intuit that alphabetic representations consist of mere indicators, pointers to words as opposed to "the things themselves" which (one imagines) Chinese characters to numinously manifest. That is, the attitude seems to be, ah, yes, I see that by using this script, you are able to recover the words of the underling message via indirect means :D To apply this to Taiwanese, etc., thus seems rather clumsy.
    Incidentally here, advocates of Romanization for Taiwanese, etc., may fail to appreciate the extent of the neurological task involved — reading such script fluently is hardly simply a matter of "learn the sound-symbol relationships in 20 minutes and be on your way." Ideal approach of course is "be 3"

    Re: "traditional sinographic Taioanese" I recall and appreciate your views but admit I remain skeptical that the preference here is more than sentimental or a function of the sense that the one is (relatively?) imposed, the other a (relatively?) organic product of the community itself (both of which are of course themselves meaningful considerations.)

    Darker —

    What I sometimes wanted to tell students from e.g. Zhuang communities in Guangxi but didn't have the heart to is that just as serving the still-living (language community) is a sacred task so is serving the dying — one day young kin will hear tales, get curious, and be deeply gratified to see that decedent's papers are in order. Romanized or equivalent has to the the right choice here…

  15. KIRINPUTRA said,

    February 5, 2023 @ 1:12 am

    Right again on pretty much every count, @Jonathan Smith.

    True, sinographic Taioanese “can't go down to a Mandarin-acculturated next generation”. But can the spoken language?

    There’s no macro-equilibrium where acculturation to Mandarin & fluent Taioanese can overlap. The Mandarin-acculturated, Taioanese-fluent individual is a transitional phenomenon. There are rarely more than two such generations in any family (or “lineage”), and often less. As a rule, if the parents & grandparents speak passable Mandarin, the kid doesn’t inherit passable Taioanese; nor do they learn it by osmosis in the community as immigrant kids often did during the first three decades of Neo-Chinese rule. National languages are jealous gods, esp. in E/SE Asia.

    As for documentation, isn’t there enough out there — recordings, romanized Taioanese writings, etc., a good deal of it stored offshore — for the language to be brought back from the dead a là Hebrew, if things came to that?


    Honest of you to admit that your objection (if that’s what it is) is selfish! Much of the debate on these themes is (evidently) driven by individuals trying to avoid rewiring at all costs.

    ㊟ Your inability to “learn Taiwanese and other languages of interest … via Chinese-character texts” is unremarkable, but the exaggerated effect of “accidental exposures” MIGHT be tied to a rigid, monolithic — but understandable — internal characterization of kanji as “Chinese characters”.


    Sinographic Taioanese writing doesn’t help people learn Taioanese. No doubt about that.

    (But transcription facilities for learners are not the end-all & be-all of writing. The two may conflate for a dying language, but the difference becomes non-trivial the moment the language “gets out of deathbed”. Transcription facilities for learners are a niche concern when a language is well.

    Transcription facilities for learners matter! But Taioanese writing doesn’t exist to help learners learn the language, functional overlap or not. That idea lingers because most people still can’t imagine Taioanese writing ever not being superfluous or niche.)


    The generally unquestioned assumption is that if you’re for rom. Taioanese, then you’re against sinographic Taioanese, and v.v. What I question is the logic that they can’t both exist indefinitely.

    Rom. Taioanese is obviously much more helpful to learners.


    My subpoint there was regarding just 偏僻字: they’re abundant in the Neo-Chinese show script, probably by design; but relatively scarce in the native script, even compared to non-show Chinese scripts.

    The native sinographic script is obviously nowhere as helpful to learners as the rom. script. For Mandophones, even the “dialect Zhōngwén” show script might be more helpful since it’s essentially a reconceptualization of … Hokkien through a Mandarin lens, with built-in mnemonics, etc.


    “[T]he attitude seems to be, ah, yes, I see that by using this script, you are able to recover the words of the underling message via indirect means. To apply this to Taiwanese, etc., thus seems rather clumsy.”

    (Although the prejudice that’s described may historically be Neo-Chinese rather than inherent to sinography.)

RSS feed for comments on this post