Victor Mair has a Log post about bopomofo, a phonetic system for transcribing Chinese. Mair says, “Chinese characters are supposed to be able to represent speech, so it seems odd that users need to resort to a separate writing system (bopomofo) to clarify how a word should be pronounced,” but goes on “As a learner of Chinese in Taiwan four decades ago, I was deeply grateful for the existence of extensive reading materials at all levels that were phonetically annotated with bopomofo. That saved me endless hours of dictionary drudgery and frustration,” and describes the bookstores in Taiwan that “are stocked with hundreds of premodern texts, both classical and literary, that are not only annotated with bopomofo, but accompanied by translations into Mandarin and extensive commentaries and notes to assist the reader…. Time and again, I have urged the educational and cultural authorities in China to use Pinyin for the same purposes, although they have been very reluctant to do so, partially because of technical difficulties of setting the ruby symbols in an orthographically correct and esthetically pleasing manner.” It’s a good read, and I expect an interesting discussion—and a lively one, especially now that Trimegistus has left a comment asking “if the Chinese themselves can’t even read their writing system, why not quietly abandon it and switch to a phonetic alphabet already in use?”
It will be fascinating to see what effect the arrival of (almost) universal cheap Chinese script-reading software on mobile phones or similar devices will have: “script amnesia” will no longer be a problem as far as reading Chinese script is concerned, which will surely solve the enormous problem of billions of “legacy texts” that might otherwise be unreadable to future generations if China switched to an easier-to-learn writing method.
I still have some of those bopomofo versions of classics. A Chinese language learner spends a lot of time just looking up the exact pronunciations of words, often rare words used only in names, and sometimes very rare words whose pronunciation is uncertain and which are used only in names. The value added of this effort is slight; you improve your skill at flipping through dictionaries, and that’s it.
One thing: the bopomofo in Taiwan is used to teach Hokkien and Hakka speakers Mandarin. I don’t think that for native Mandarin speakers learning the sound of the grafts is especially difficult, though they still usually have to learn correct educated Mandarin pronunciation. In other words, the problem is less a problem with non-phonetic script than it is with standardizing pronunciation nationally.
Just as I feel about most well meaning linguistic reforms – my answer would be “why bother?” Do we English or French speakers really suffer compared to Spanish or Turkish speakers because our writing system is far more illogical? Recent economic and cultural history might suggest otherwise. Look at all the waste and nonsense around the German spelling reforms – whose life has improved because of it? China seems to be doing pretty well right now, so what’s the issue? And if they do switch to pinyin, it won’t be a tragedy. Most people in China won’t suffer too much, and won’t care they now can’t read old texts they probably would never have read anyway. The simple fact is for most of Chinese history the vast majority of Chinese never knew characters anway.
Wish I could share hat’s optimism about an “interesting” discussion — I’ve seen this debate so many times (on the Language Log itself, even) that I feel like I know the lines of argument before they come up. vanya’s covered the “they seem to be doing okay with characters” argument (no offense) and soon we’ll see somebody point out how many Chinese “words” are pronounced “ji” (or whatever) and how these will all become completely, totally indistinguishable without characters, written Chinese will become incomprehensible, dogs and cats living together, etc.
Graphs. Dyslexia is really hitting me hard.
One quality I do envy in the CJK characters is their flexibility with ordering. When reading about any of these character sets, you always see, either diachronically or synchronically, a sort of laissez-faire attitude about whether to write right to left, top to bottom, left to right. You name it. It makes my lefty heart envious. I tried to see if I could adopt a top-down RTL script for writing English, but I didn’t find any good candidates.
Z. D. Smith,
I too envy the CJK characters’ multidirectionality, which I would attribute to their notional squareness. Conversely, in my opinion at least, the Latin alphabet can’t be used left-to-right without reversing the letterforms. And this introduces all kinds of problems. Not to mention that top-to-bottom, for more than about 8 letters, is totally illegible.
Thanks Bob!
I should say, as a mea culpa, it’s worth it if the discussion spins off into unexpected tangents, which is already happening. So god knows nobody should take my carping seriously.
I still think boustrophedon was a great idea.
Until Rivers4 came along with the predictable “Yet another rant from a foreigner who wants China to abandon its writing system because he finds it too difficult” line of attack. These rote exchanges are indeed tedious.
(That was to Bob, not Boiko.)
[Chinese] characters are supposed to be able to represent speech
As wall-eyed as it is, in order to evade possible accusations of rote and tedium, I suppose I should just let this supposition slide.
“As wall-eyed as it is, in order to evade possible accusations of rote and tedium, I suppose I should just let this supposition slide.”
By all means, let us know what you think.
From what I can tell, the consensus of linguists who examine how Chinese people actually use characters is that characters serve as a very redundant syllabary (that is on a day to day basis characters are treated more as carriers of phonetic rather than semantic information).
Oh yes: higher numbers of illiterates and dyslexics where the economy is comparable. And grown people like present company needing to use a fucking spellchecker on the fucking Internet.
Those of everyone who’s younger than me? 🙂
Actually, Chinese has already been through script reform. So has Japanese. A hundred years ago, everyone could read everyone else’s Chinese characters. Now they’re all different.
This is not a trivial issue. Mainlanders find it an effort to read things written in traditional script, and Taiwanese/Hong Kongese feel the same about simplified characters. Not impossible, just tiring and unfamiliar. In other words, they’d prefer not to be reading it. And if you’re one of those foreigners who do happen to need to learn more than one variety, there’s a whole lot more stuff to learn! It’s great if you’re really into it, not so much if you just want to read and write.
Oh, yes. The breakup of the Sinosphere is, I should think, a direct result of the Western impact. They’d all been using the same character-set for well over a thousand years and suddenly, pouf, ‘modernisation’!
By all means, let us know what you think.
If we suppose a strong form of the claim:
Writing is an (i) historically conservative (and therefore inadequately symmetrical), (ii) causally subordinate (and therefore functionally dependent), (iii) surrogate for speech.
Then a few of the more obvious counterclaims might include:
The supposition is (a) empirically incomplete (unidirectional causality?), (b) analytically reductive (independently characterizable explanatory facts?), (c) simultaneously overspecified (writing is functionally transitive/relational with respect to speech?), and (d) undertheorized (semantic constraints?).
In a weak form, ‘[I find this situation to be extremely thought-provoking. After all, Chinese] characters are supposed to be able to represent speech…’ is just that – a supposition, one which, insofar as it concerns the interdependence, complementarity, or relative autonomy of writing and speech, doesn’t strike me as having provoked much thought at all.
(I don’t know – is that rote and tedious?)
A small point: the “bopomofo” are also known as “zhùyīn fúhào” 注音符号, ㄓㄨˋ ㄧㄣ ㄈㄨˊ ㄏㄠˋ. When I went to Taiwan to study spoken Chinese and Tang poetry for a summer at the Mandarin Center in 1968 as part of my graduate work in Korean, I found these symbols the hardest thing to master since struggling with the four corners system to look things up at the Gest Library at Princeton.
I notice that most of the commenters on this site have moseyed (mosied?) over to languagelog to comment over there.
At any rate, while much that Victor Mair says makes good sense, he does rather come across as an old warhorse. Since he has campaigned about these issues for many years, it is a little incongruous when he uses turns of phrase like “I find this situation to be extremely thought-provoking”, as though he has only just noticed what is going on.
I do notice that comments seem to have become more sophisticated (leaving aside knee-jerk responses from know-nothings like Rivers4). Perhaps that is only because Hat regulars were commenting, but there was evidence of people actually thinking and questioning what a writing system is. I’ve seen past comments on similar threads sink to pretty low levels of chauvinism and ignorance.
Actually, more alarming to me than 旻, which I’d encountered in the past but forgotten (again) how to read was the sign below it — in particular 贗復重建, the first character of which I did not know at all. After considerable searching I found the pronunciation and the meaning of the character (贗 / 赝 yàn, ‘fake, counterfeit’) and, I think, the phrase (贗復重建 / 赝复重建 yànfù chóngjiàn ‘prosthetic reconstruction’).
I was somewhat relieved to find out that 赝复 is largely confined to Taiwan. But it reinforces Victor Mair’s point about the difficulties that can be thrown up by the Chinese writing system.
I’m just tickled by the fact that in Taiwan the word for “dentist” is what would be understood in Japanese (if you wrote it in the same characters) as something like “tusk chirurgeon”.
Yes, that’s because the Chinese word for ‘tooth’ is 牙/牙 or 牙齒/牙齿. I suspect that it is Chinese that has changed, not Japanese.
@Bathrobe: The first character occurs in the normal word for “fake, ersatz”, 赝品. I have never heard the word 赝复, though, being a die-hard Mainlander.
Yes, 赝品 is a normal word, but my Chinese is still stuck at the 假货 stage 🙂
People like Rivers4 need to read and understand John DeFrancis’s book Chinese Language: Fact and Fantasy. If afterwards they still think that the Chinese writing system is well suited to its purpose then I’ll be surprised.
I won’t. People like that have their ideas set in concrete.
I don’t think people like that would even care enough to read such a book. They are all attitude and no content. The most you could expect (but are unlikely to get) is a grudging ‘Whatever’. But I’d be surprised if he’s still following the discussion. Having lobbed his cheap shot, he’s long gone.
Characters may not be easy to use, but I think a good case could be made for their beauty; it’s simply more pleasant to read characters than pinyin. The form of a character also adds a bit of complexity to a word’s flavor in a way that would be lost with romanization. Simplifying written language may only cause small problems in comprehension (of the Война и Мiръ variety), but it makes things blander. If we ignored etymology and spelled phonetically, it would still be possible to read Australian and Canadian novels witha bit of mental adjustment, but the associations we form with words of different origin would probably start to fade. Ultimately, a language should be judged not based on how well it is used by most, but by how well it can be used, by the pleasing complexity of connotations and denotations at its command.
Besides, without characters, books would have to be translated into several dialects.
Hope this isn’t too tedious. I love your blog and was a bit reluctant to sully it with banalities you’ve undoubtable heard before, but I really feel strongly about this.
Besides, without characters, books would have to be translated into several dialects.
Hope this isn’t too tedious. I love your blog and was a bit reluctant to sully it with banalities you’ve undoubtable heard before, but I really feel strongly about this.
I have to disagree with the “breakup of the Sinosphere” claim, at least with respect to highly educated Sinophones. Before 1949, everyone had to learn both standard and “grass” (informal) character styles, and so they still do: it’s just that on the Mainland and Singapore nowadays you learn the grass style first.
As I wrote at the Log, the serious technical barrier to full romanization (as distinct from the many social ones) is the wènyàn (Classical Chinese) embedded in high-register writings, which cannot be understood when spoken in Mandarin (it’s not redundant enough) and therefore cannot be romanized by a system based on speech. I wonder whether a Middle Chinese romanization scheme like David Prager Banner’s Neutral Transcription System, which reflects traditional Chinese philology rather than making an attempt at reconstruction, might be fit for the purpose. I asked him by email, but haven’t yet heard back from him on that — he’s a busy man.
Alternatively, the General Chinese of Yuen Ren Chao (or Dhyao Qiuan Remm, if you like), if adapted a bit for Hanyu Pinyin compatibility, would likely serve. Chao estimated that 90-95% of the characters in running text would be uniquely represented, and the result can (as a bonus) be read equally well in Mandarin, Yue, Hakka, Min, Hakka, and Sino-{Japanese, Korean, Vietnamese}.
David Branner’s own comment at the LL post was:
I’m sorry to say that the major dialect groups are not representable by a single metasystem, unless you’re writing a purely literary form of Chinese in them.
ㄨУ
ㄓЖ
ㄈF
ㄐЧ
ㄇМ
Characters may not be easy to use, but I think a good case could be made for their beauty
Nobody denies their beauty; some of us just don’t think the tradeoff is worth it. English would doubtless be more beautiful if written with Chinese characters, but would that be a good idea? (As for the “tedious” bit, I didn’t mean to discourage anyone from commenting—well-written and impassioned comments like yours are always worth having. I just didn’t want the thread to degenerate into the standard “Beauty!” “Utility!” “BEAUTY!” “UTILITY!” exchange that so often distinguishes the Log threads on the subject, complete with insults. As long as partisans of each view can acknowledge the points of the other and discuss things like tradeoffs and what the facts of the matter are, it can be a productive discussion.)
One of the tradeoffs could be that it would make the memorization and comprehension of Classical poetry harder. Even for a populist 8th century poet like Bai Juyi, some of his lines have turns of phrase that would be difficult to explain without recourse to the characters.
In the second line of the Song of Everlasting Sorrow (“御宇多年求不得”) for example, to explain “御宇” (He ruled over the land , you need to tell “御” (yù, to control, to rule, but not the usual morpheme for it) and “宇” (yǔ, vast space > land, again not the usual morpheme for it) from the learnèd words formed on them. Considering the number of the homophones, it’s quasi-impossible to explain them if you don’t mention the written shapes.
Of course the case could be make that it doesn’t concern the written shapes per se, but an abstract notion standing for Old Chinese morphemes. This is doubtlessly the mechanism that permits even the blind people to know whether a certain meaning of a syllable belongs to the same character with another meaning or not. But without a daily use of the characters, this knowledge of repartition of spoken syllables into abstract categories will gradually die out. In my humble and unjustified opinion, that is a bad thing.
The Neutral Transcription System doesn’t look too much like a real writing system (I mean, who would want little subscript 3a, 3b’s after each syllable?) This normalcy is a significant virtue of every Romanization of Chao, for all my rants for their unlearnability.
Would the “General Chinese” or “Neutral Transcription” approach be more or less like agreeing that every Romance language should be written with the Latin spellings of the Latin predecessors of the words of common origin (with language-specific knowledge of how to transform those ancestral spellings into the daughter-language pronunciations)?
@minus273: Branner’s article goes at some length into the subjective willingness (at least among Western Siniticists) and perhaps even objective desirability to give up some theoretical precision in order to use a system that requires no characters not found on whatever typewriter the relevant academic already had on his desk. (You have to go beyond that — Wade-Giles apparently doesn’t work well in practice in Taiwan because people often don’t bother with the apostrophes, and it’s not because they didn’t have access to typewriters, qwerty keyboards etc. that featured apostrophes. Understanding the actual basis for user resistance to the apostrophes would be helpful in avoiding similar blunders in future proposed systems.)
Bathrobe: DPB is right, of course, but my concern isn’t using a Middle Chinese system as a diasystemic spelling: the need for that is over. It’s using it to represent Classical Chinese embedded in Mandarin that’s represented in HYPY.
Minus273: The trouble with things like Gwoyeu Romatzyh isn’t the tonal spelling, it’s how complicated the tonal spelling rules are. I have no trouble remembering how to spell Gwoyeu Romatzyh ever (if you have mastered English, spellings like that are trivial), unlike the diacritics on Hanyu Pinyin, of which I remember three on a good day. But as to the actual tones of the former expression, I have no clue because the system is so, as you say, unlearnable. From what little I’ve discovered about General Chinese, though, it’s much less so.
One of the tradeoffs could be that it would make the memorization and comprehension of Classical poetry harder.
Oh, there’s no argument about that either. Any move toward romanization would devastate classical literature; it would become even more the province of a few than it is now. It took me a long time to overcome my instinctive aversion to that idea, but I finally realized true mass literacy was worth it. After all, quite a few Westerners take the time and trouble to learn Classical Chinese starting from zero simply because the literature is so impressive; surely even more Chinese, starting from someplace higher than zero, would do so. Something that has become a minority taste has not vanished from the earth.
@John Cowan: Just relooked at GC on Wikipedia. I think it would be much improved by two simple changes:
– giving ng- back to MC [ŋ]
– uniformizing the tone marking to doubling for 上 and -h for 去.
(As the series 平, 上, 去 for a simple BA is ba baa bah, for a coda’d JANG, jang jaang jangh would be much more straightforward than jang jag jaq.)
If I’m allowed to do a little more change, the stop/affricate onsets should be rationalized: if [p ph b] is b p bh, [k kh g] should be g k gh not c k g. (c could then be rescued for a Pinyin-like [tsʰ]) Chao just cared too much about brevity over doing something usable by everyone.
There is something comical about lecturing the Chinese on phoneticity, and doing it in English. I move that any further such lecturing be done in Finnish or Hawai’ian.
@caffiend: But still, in English you won’t write “coc” and pronounce “lock”.
I dislike the idea that “true mass literacy” requires a switch to a phonetic system, especially when there are plenty of countries with phonetic writing systems that have problems with literacy. Spanish-speaking countries, for example. Literacy is fundamentally an economic and educational problem, as evidenced by high rates of literacy in Japan and Taiwan despite their use of characters. PRC will be the same in a few generations when they have less poverty.
It is obvious that there is not a one-to-one correlation between a writing system that is easy to learn and mass literacy. It should be equally obvious that there is a causal correlation, even if other factors are at work.
Of course there is a causal correlation, but not one so large that it justifies demanding wide-scale reform of the writing system rather than demanding that the government make education a bigger priority. Literacy problems are fixed almost automatically as soon as a country is able to lift its citizens out of poverty and educate them for free. You’re never going to see 99% literacy in a poor, overpopulated country with little or no free public education, no matter what writing system they use.
None of this matters because China will soon be able to improve its literacy rate as it becomes stronger economically. That will take much less time than changing the way everyone writes. When this happens, I wonder if the endless discussions about whether or not hanzi should/could be done away with will finally end. (Probably not, since there will always be frustrated students of Chinese or Japanese who want to complain about how hard it is to learn.)
Quick to learn is not the same as efficient for experienced users.
Mongolian traditional script is quite a nightmare, and attempts to reinstate the script usually founder on the fact that it is much harder to learn than Cyrillic. Still, from what I’ve heard, for people familiar with it, it is a faster and more efficient script for note taking than Cyrillic.
Before 1949, everyone had to learn both standard and “grass” (informal) character styles, and so they still do: it’s just that on the Mainland and Singapore nowadays you learn the grass style first
Sounds good as a theory, but that’s not how it works. It’s true that they all have their roots in the same tradition, that does not mean that they are the same.
There are now essentially two printed standards in Chinese-language areas; three if you include Japanese. Hand-writing may be slightly different, but for most people the printed and written standards are basically the same. People who learn the Simplified Characters (many of which are based on grass script) learn only the Simplified Characters. They don’t learn Simplified first and then go on to Traditional, although they may pick up a nodding acquaintance with the Traditional character set. And people who Traditional may learn to use grass script, but that is not the same as learning the Simplified character set.
Even though the ultimate roots may be the same, the Sinosphere has truly been split up into three different character sets.
Well, I was being a bit ironic when I implied that Simplified actually was grass script. And I did confine my remark to the highly educated, who do still in practice need to learn to read, if not write, both Simplified and Traditional styles.
This is already the case, avoided only by the fact that everyone learns Mandarin. Written non-classical Chinese is written Standard Mandarin.
Written Cantonese is pretty well developed, and it’s quite obviously a different language. They even had to invent characters for morphemes that happen not to exist in Mandarin.
Very good points, but how about:
ㄓЏ
That’s much closer phonetically, and the shape still isn’t far off.
Oh FFS. The problem of memorizing and comprehending this kind of stuff is about the same that I face with aurea prima sata est aetas, quae vindice nullo, and so it should be. 8th century! For crying out loud!
Pretty much.
How high are they really? (Ignoring kana in Japan.) There are book excerpts at http://www.pinyin.info which suggest the figures are all overblown, and that’s before we get to Qingdao University students of Chinese who forget how to write “sneeze”.
He cared about making things easier for the fully fluent writer, not for the reader or learner.
So, BTW, does the Mongolian script.
Qingdao University students of Chinese who forget how to write “sneeze”
Qingdao University? The one in Shandong Province?
The second character in 喷嚏 is a seriously difficult character. (The first one is a shoo-in. I don’t know why people insist on saying Chinese people can’t write ‘sneeze’. The problem is 嚏, not 喷). Yes, it’s a failure of the Chinese writing system that such a commonly used word (although arguably more commonly spoken than written) should be so hard to write. But it’s a bit unfair to magnify this single example right out of proportion. When one character is constantly used to prove the extraordinary difficulty of the whole writing system, you know there is something wrong.
I beg to differ in this case. It’s more comparable to Singinge he was, or floytinge, al the day; He was as fresh as is the month of May, save for those poetic coinages according to the literary morphology scattered everywhere. (aurea prima sata would be the Book of Odes, quite another matter, which exists in popular minds only as a few selected verses and stock phrases.) Saying this make me feel like a premodern Chinese particularist bigot, but I do feel that the equivalent to an Eugene Onegin to a modern Chinese mind is exactly the semi-canonical since late-Ming selection of Tang-dynasty shī poems and Song-dynasty lyrics. It follows that ditching them along with the characters should only be considered as a very last resort, in a deep national crisis maybe, but certainly not when the economy is doing quite good for the education to suck some blood from.
Bathrobe, I think you see the “sneeze” example everywhere because we ignorant barbarians all copy our examples from 莫大伟, the famous 相声 performer.
*shame* It’s Peking University. I had misremembered, and I had forgotten that I have Why Chinese Is So Damn Hard in my bookmarks.
“Sneeze” is just the example I remembered because it’s presented in the most detail. Here’s the relevant excerpt:
That’s what I mean.
The stuff about literacy rates must be hidden somewhere here.
Uh, sorry. I wasn’t clear on what the meaning of “is” is. :-]
What I mean is that I don’t see why anyone should expect to breeze through such old poetry. It’s in a different language, even if the writing system partially hides that fact.
I don’t actually know anything about that other than the name.
Thanks for that wonderful Moser excerpt; I’ve added the page to my bookmarks now.
Has anyone figured out whether native learners of Chinese script have similar rates of dyslexia to ours?
Chao just cared too much about brevity over doing something usable by everyone.
Yes indeed. Barnett-Chao romanization, which adapts GR principles to Cantonese, looks like it might work when you see it explained syllable by syllable. But then you see that 廣州話 ‘Guangzhou dialect’, which is guǎngzhōuhuà in Mandarin and gwóng jàu wá in sensible Yale Cantonese romanization, comes out Gworngzhawwraav. Worse yet, 粵語 ‘Cantonese language’, yuèyǔ in Mandarin and yuht yúh in Yale, is written jrytjryr. Now that’s just scary. Even if you wrote the Yale spellings as single words Pinyin-style, Gwóngjàuwá and yuhtyúh are perfectly readable. (These examples come from WP, which is constrained from pointing out just how awful they are. For people not familiar with Yale, it uses h as a diacritic for tones that begin low, as opposed to those that begin mid or high.)
Actually none of the letters I’ve used has any meaning except for the letters A & I in English.
O, I think you forgot one.
guǎngzhōuhuà […] Gworngzhawwraav
Then there is accent-free Pinyin, where it is guangvvvzhouvhuavvvv. This is also apostrophe-free, since there is no ambiguity about syllable boundaries.