Character Amnesia

« previous post | next post »

Pessimists and alarmists have long been lamenting the negative impact of computers upon the ability of Chinese to write characters by hand.  See, for example, Jennifer 8. Lee's article entitled "In China, Computer Use Erodes Traditional Handwriting, Stirring a Cultural Debate" in the Technology section of the New York Times for February 1, 2001.

If the situation was bad already a decade ago, it is far more grave now that short text messaging is so wildly popular.  In "China worries about losing its character(s)," Los Angeles Times (July 12, 2010), Barbara Demick provides graphic evidence of the starkly diminishing powers of supposedly literate Chinese to produce many characters that are essential for daily usage.


I've seen people stumped by even the simplified form of the character for "shrimp," xiā 虾, never mind the traditional form, 蝦, and Demick tells of "literate" people who cannot write zàijiàn 再見 ("goodbye") or "shampoo" (there are several possibilities).  Even before computers, exceedingly few people could write both characters for "sneeze" (pēntì 噴嚏, simplified 噴嚏 [same]); though I've asked scores, I personally have never met any Chinese, including individuals with master's and doctor's degrees, who could do so, and David Moser — much to his astonishment — had similar results (see his classic piece entitled "Why Chinese is So Damn Hard," written nearly twenty years ago; further extensive and very recent research by Moser has only strengthened and confirmed his original findings).  Given that such a common word as "sneeze" cannot readily be written out by the vast majority of allegedly literate Chinese, one can only imagine what trouble they would have with a word like zhā 皻 or 齇 ("red flecks on the nose of a drunk person").  Well, they wouldn't have any trouble with the word zhā or the word pēntì (I hear people say the latter all the time; I've only heard the former spoken a couple of times); it's the characters for these words that flummox people.

Because of their complexity and multiplicity, writing Chinese characters correctly is a highly neuromuscular task.  One simply has to practice them hundreds and hundreds of times to master them.  And, as with playing a musical instrument like a violin or a piano, one must practice writing them regularly or one's control over them will simply evaporate.

Demick's article ends thus:

"It will take a lot of effort to preserve our Chinese characters. It is the same way they try to preserve these old hutongs," said Zhu Linfei, 24, a Beijing graduate student, referring to the traditional Beijing alleys, now rapidly succumbing to the wrecking ball.

Zhu, who was touring the old bookstores of Liulichang with her classmates to buy calligraphy books, estimated that she had already forgotten about 20% of the characters she knew in high school.

"But it's not such a big problem," she said. "If I don't know a character, I take out my cellphone to check."

Zhu Linfei is mistaken.  It is a big problem that she cannot write 20% of the characters she knew just 5 or 6 years earlier.  By relying on her cellphone to check those characters she can't recall, that percentage will increase with each passing year.  Furthermore, every time Zhu Linfei has to stop to take out her cellphone crutch to remind her how to write a character, she is wasting time, and that in itself is a problem.

Unlike aphasia, a type of language disorder that usually occurs suddenly because of physical injury, the impairment brought about by frequent cellphone checking is gradual.  Nonetheless, the attrition that results is just as real as that brought about by dysphasia (limited aphasia).

Last year, I surveyed nearly two hundred individuals who are literate in Chinese, asking them what their preferred IME (Input Method Editor) was.  About half of them were professional teachers of Chinese, and they hailed from around the world. Around 98% of the respondents used Pinyin (romanization) to input Chinese characters, with the remaining tiny handful using a shape-based system such as Cangjie or a stylus to write them on a pad or window.  Both of those who used Cangjie were from Hong Kong and Taiwan, and a couple of respondents from Taiwan said that they used bopomofu (Zhuyin Fuhao), a phonetic inputting system somewhat reminiscent of Japanese kana, though it is not strictly syllabic.

Computers, cellphones, smartphones, and all other such electronic gadgets are wonderful tools for communication, but they all exacerbate the predicament of declining ability to write characters among the Chinese population, and they are hastening reliance on alphabetical access to literacy, instead of a direct approach through the 11 or so basic strokes, the 200 or so radicals, and the 850 or so phonetic components.  Are these worrisome trends?  Can anything be done to stanch the hemorrhaging of active character proficiency at the hands of cellphones and computers?  Finally, is romanization inevitable?  That is a question to which I shall return in a future post.

[A tip of the hat to Stefan Krasowski.]



140 Comments

  1. le_sacre said,

    July 22, 2010 @ 6:44 am

    You wrote: "Zhu Linfei is mistaken. It is a big problem that she cannot write 20% of the characters she knew just 5 or 6 years earlier."

    That's a rather confident statement of opinion as fact! I hope in your future post you'll address the rationale for this position. If the ability to write characters from unaided memory were crucial in any way to day-to-day living, it would not be dropping off so precipitously. Of course it would be a tragedy for the characters to vanish altogether, but it seems extremely unlikely that literate Chinese will forget how to read them, even if they can no longer write them by hand (without an electronic aid).

    It's simply extraordinary how many hours are devoted to this rote skill acquisition in Chinese childhood. Especially if the years of painstaking mastery appear doomed to irrelevance or obsolescence, one simply can't avoid questioning whether there isn't anything better kids could be spending those hours on.

  2. Randy Alexander said,

    July 22, 2010 @ 6:56 am

    I'm surprised none of your respondents use wubi.

  3. Yuval said,

    July 22, 2010 @ 6:57 am

    I'm sorry. Jennifer 8. Lee?
    Huh.

  4. Victor Mair said,

    July 22, 2010 @ 7:22 am

    Randy: one started out using Wubi (a shaped-based system) because everybody in her high school was forced to use it, but she hated it viscerally because of the heavy, arbitrary memorization load it imposes and difficulty in touch-typing. Later, she gladly gave it up in favor of Pinyin. At least that's what she told me when I asked her about it.

  5. Ginger Yellow said,

    July 22, 2010 @ 7:25 am

    So, when these people want to write "sneeze", say, what do they do, other than checking their cellphone? Do they look it up? Do they write the wrong character? Do they do an inaccurate version of the correct character?

  6. Ginger Yellow said,

    July 22, 2010 @ 7:27 am

    Or do they write it in pinyin?

  7. Bill Walderman said,

    July 22, 2010 @ 8:00 am

    How often do literate English-speakers have occasion to write the word "sneeze" in the course of their daily activities? Isn't there a large number of common words in any language that are primarily spoken and rarely written, given the functions that writing serves in daily life? And if that's the case, is it possible that the inability of literate Chinese speakers to write words such as "sneeze" or "red flecks on the nose of a drunken person" is a not recent phenomenon indicative of a decline in the ability of Chinese speakers to fluently manipulate the Chinese writing system, but rather something that has been true throughout the history of written Chinese?

  8. John Cowan said,

    July 22, 2010 @ 8:06 am

    Randy: In order to use wubi, you have to know how the character looks to begin with, not to mention what the stroke order for writing it is. Wubi is well-suited to professionals transcribing written text that's in front of them (because they do not need to know how to pronounce a character in order to type it), but not to the composition of novel text by non-professionals.

    (I know, IMHBT.)

  9. Matt Jadlocki said,

    July 22, 2010 @ 8:10 am

    Is there any evidence, other than anecdotal, showing increase in 'character amnesia'? I mean is the percentage of people who can't write the characters for, say, 'sneeze' within a given age group greater now than in the past based on some sort of series of longitudinal studies on character writing ability? Perhaps there are studies showing that people with less access to computers/cell phones are better able to recall characters for writing. Something more than just personal stories and sentiments.

    Any argument seeming to invoke the ‘good old days’, tends to arouse my suspicion. And one based mostly, it seems, upon anecdotal evidence adds to that suspicion. Lee’s article mentions a saying that translates as “forgetting characters upon lifting the pen”. I’m curious to know if this saying has a recent origin related to the advent of technology or a long history perhaps indicating an unrecognized national pastime of forgetting how to write characters. And have the writers Lee mentions really been able to notice a shift in how often they forget characters that correlates with the increasing prevalence of computers?

    I’d also be interested to see if these same people who are unable to produce a character on command would have been able to produce the character given the opportunity to write a paragraph or so with a topic prompt(that would result in the writing of the desired character). Emphasis not being on character recall but on the topic in question.

    I have many more questions based on character amnesia being a real and growing problem, but I’m not quite sure I want to make that jump.

    Regardless, one upside to the advent of alphabetic input, learners of Chinese (and other languages employing Chinese characters) have an easier time of becoming writers in the new language since there is little need for handwriting anymore. (Though I guess this applies more to self-taught learners than those stuck in a classroom requiring endless repetitions.)

  10. John Cowan said,

    July 22, 2010 @ 8:21 am

    Bill W: Here's the context of the "sneeze" example from Moser's article:

    In fact, one of the most gratifying experiences a foreign student of Chinese can have is to see a native speaker come up a complete blank when called upon to write the characters for some relatively common word. You feel an enormous sense of vindication and relief to see a native speaker experience the exact same difficulty you experience every day.

    This is such a gratifying experience, in fact, that I have actually kept a list of characters that I have observed Chinese people forget how to write. (A sick, obsessive activity, I know.) I have seen highly literate Chinese people forget how to write certain characters in common words like "tin can", "knee", "screwdriver", "snap" (as in "to snap one's fingers"), "elbow", "ginger", "cushion", "firecracker", and so on. And when I say "forget", I mean that they often cannot even put the first stroke down on the paper. Can you imagine a well-educated native English speaker totally forgetting how to write a word like "knee" or "tin can"? Or even a rarely-seen word like "scabbard" or "ragamuffin"?

    I was once at a luncheon with three Ph.D. students in the Chinese Department at Peking University, all native Chinese (one from Hong Kong). I happened to have a cold that day, and was trying to write a brief note to a friend canceling an appointment that day. I found that I couldn't remember how to write the character 嚔, as in da penti 打喷嚔 "to sneeze". I asked my three friends how to write the character, and to my surprise, all three of them simply shrugged in sheepish embarrassment. Not one of them could correctly produce the character. Now, Peking University is usually considered the "Harvard of China". Can you imagine three Ph.D. students in English at Harvard forgetting how to write the English word "sneeze"?? Yet this state of affairs is by no means uncommon in China.

  11. Ginger Yellow said,

    July 22, 2010 @ 8:22 am

    How often do literate English-speakers have occasion to write the word "sneeze" in the course of their daily activities? Isn't there a large number of common words in any language that are primarily spoken and rarely written, given the functions that writing serves in daily life?

    The general point is fair enough, but "sneeze" shows up fairly often in written English, for instance as part of the phrase "nothing to sneeze at" (14.9m Google results, for what little it's worth).

  12. Ginger Yellow said,

    July 22, 2010 @ 8:24 am

    767 results in the NYT archive.

  13. JimG said,

    July 22, 2010 @ 8:36 am

    Yes, Jennifer 8. Lee.

    She's a journalist and writer, not given a middle name at birth. She chose "8." as a teenager. It's a lucky number associated with prosperity and good fortune.

    Imagine the possibilities: Some folks go through phases of anarchism, nihilism and antisocial attitudes, and might choose less "desirable" numbers. Hmmmm, maybe I'll change MY middle name.

  14. John Cowan said,

    July 22, 2010 @ 9:19 am

    My given names are from my grandfathers, and to change them would be tantamount (in the ways of verbal magic) to changing them.

  15. Doctor Science said,

    July 22, 2010 @ 9:27 am

    Meanwhile, James Fallows recently observed that written Chinese is ideally suited to Twitter, because a 140-character limit is, in Chinese, a 140-*word* limit, which isn't much of a limit. I hadn't realized, though, that pinyin is being used as the bridge between the keyboard and the display.

    Are people seeing "character amnesia" for reading as well as writing? Are literate Chinese just as able to *read* "sneeze" or "elbow" as they were before electronics, or is reading following writing?

  16. Janne said,

    July 22, 2010 @ 9:34 am

    Hm, I have to agree with Matt above: anecdotes don't prove anything. Newspapers and pundits regularly use the exact same gambit of taking a few examples, then weaving a elaborate, flimsy tapestry of a story around it, either to give support for a predetermined conclusion or simply to score a selling headline. Now, they tend to do this on purpose, and I think you are sincere. But sincere or not, while "data is not plural of anecdote" is a trite old expression, it is still very true. Find data. Good, solid comparable data on this over time, and show a real statistical effect.

    People are panicking about the exact same thing here in Japan. But according to my teacher people have been panicking about declining kanji knowledge throughout modern history, at least beginning with the rise of pens over ink brushes, so there's little new.

    Besides, even if true, is that so bad? After all, a major reason people would be forgetting how to write them by hand (if they really are) is because they no longer do. And if you no longer write by hand, then there's not much lost as far as practical literacy is concerned, is there? If you can read and write the characters effortlessly then where's the problem? I know my handwriting is barely legible today, and I can't write much more than a paragraph before my hand starts to cramp up. But as I just about never need to write more than a single word or two by hand it is not a problem either.

  17. Jarek Weckwerth said,

    July 22, 2010 @ 9:55 am

    Forgive the naivety of the questions — I know far too little about Chinese writing. Would the people who don't know the character for 'sneeze' be able to read the character? Also, why is stroke order so important? What if somebody can write a character but uses the wrong stroke order? That wouldn't change the end result, would it?

  18. Jen said,

    July 22, 2010 @ 9:57 am

    Hello, I am Chinese, though I do not speak Mandarin or Cantonese or any other Chinese language. 8 and 9 are considered very lucky numbers in China and among Chinese. For example, there is a Chinese supermarket on the west coast called "99 Ranch" and one in Boston formerly called "Super 88". My aunts also have a US license plates which read "88888" and then their initials.

    I have tried to learn Mandarin, but it is very frustrating for the following reasons:

    1. The tone system is a pain to learn for native Englsih speakers.
    2. There is not a one-to-one correspondence between sounds and orthographic characters like there is (for the most part) in English, necessitating a phonetic alphabet called "pinyin".
    3. There are 4 or 5 different pinyins in use, depending on how old you are.
    4. It's a pain to learn and write each single character.
    5. My mom is Chinese, born in Taiwan, and she has no idea of what pinyin is. I have no idea why.
    6. My mom does not like the idea of my learning Mandarin since she is ashamed to be Chinese. That is why she never taught me Taiwanese.
    7. As a linguist, Mandarin textbooks are not advanced enough for me. Too much of the material is "dumbed down".

  19. language hat said,

    July 22, 2010 @ 9:59 am

    Amazing what hoops people will jump through to avoid seeing the problem with Chinese characters.

    Finally, is romanization inevitable? That is a question to which I shall return in a future post.

    Be sure to put on your hazmat suit first.

  20. Daniel von Brighoff said,

    July 22, 2010 @ 10:23 am

    I'm not a native or even fluent speaker of Chinese, but I am (barely) literate in the language. Yes, just as users of alphabetic scripts recognise far more words than are in their active vocabulary, Chinese-speakers recognise far more written characters than they can actively produce. I couldn't write 噴嚏 either if called upon to do so, but I've never had trouble recognising it when I see it in print.

  21. J.Marshall Unger said,

    July 22, 2010 @ 10:39 am

    The interesting historical point here is that, back in the 1980s and 1990s, many authors were confidently predicting that ubiquitous, cheap computers and "intelligent" software would guarantee the robust use of Chinese characters in China and Japan for the indefinite future. Some of us wrote books and articles explaining why such predictions were wrong. Of course, even back then, Chinese characters on computers were not totally intractable. The problem has always been with the capacities, desires, and priorities of end-users. Until computers can read minds, Chinese character input will always be a problem.

    Back in the 1980s, Yamada Hisao at the University of Tokyo demonstrated empirically how to design a large character set input system as efficient as English touch-typing. Typists require special training and the keying system must be be non-mnemonic with respect to the "sounds" and shapes of the characters, but it makes (high) keying speeds and (low) error rates as good as those for English attainable. People missed the implication: Yamada's research proved that this is the only kind of input system that can reach this standard.

    Manufacturers ignored these results (though they were well-known in the R&D community) because they wanted products that could be used immediately by consumers upon purchase. Now Chinese and Japanese are paying the price.

    As for romanization, the point is that they are already using it for everything except text display.

  22. chris said,

    July 22, 2010 @ 11:14 am

    If the 8 is the entire middle name, isn't the period improper, as it is for Harry S Truman?

    Also, do you translate it depending on which language you are speaking at the moment?

    Finally, I concur with the other posts skeptical of the idea that there was a time when everyone could remember all the characters in Chinese's ultracomplex writing system. Aside from professional scholars (and if you go back far enough in history, scribes and scriveners), how many people ever had more than a limited repertoire of characters?

    Isn't the smaller number of glyphs to memorize commonly accepted as one of the advantages of alphabetic and syllabic writing systems? (Although contra Moser, I think educated English speakers misspell words all the time, especially uncommon ones. This may be limited to languages whose spelling is as unsystematic as English's, though.)

  23. digory said,

    July 22, 2010 @ 11:27 am

    I was inspired by Moser's article, oddly enough, to start dabbling a bit in Mandarin, and I find it fascinating. Early on I decided to focus on reading the characters, but not bothering to try to write them. This discussion makes it clear that that was the right decision, for the casual learner of Mandarin, and that in fact many Chinese are making the same choice!

  24. vanya said,

    July 22, 2010 @ 11:29 am

    Amazing what hoops people will jump through to avoid seeing the problem with Chinese characters.

    I don't think that's fair. If you were starting from scratch of course you would choose an alphabet or syllabary to write your language. But Chinese aren't starting from scratch. Is it worth throwing out 3,000 years of knowledge and literature for some amount of greater efficiency? If the Chinese do move to romanization, then probably within a generation almost no one will be able to read characters anymore. Every thing, every inscription, every sign, every family heirloom, every restaurant menu, every blog post, ever written in Mandarin, Cantonese or other Chinese language before 2010 will be inaccessible to future generations – it will be a massive massive break in the Chinese cultural tradition. I can certainly understand why a lot of Chinese would be opposed to that. Obviously Chinese characters are a very inefficient way to write, but there's still no evidence that China, Hong Kong or Taiwan are suffering much of a drop in quality in life from using them. The move to romanization may well be inevitable, but lots of inevitable developments aren't necessarily worth celebrating.

  25. Jen said,

    July 22, 2010 @ 11:38 am

    It would be tragic if the Chinese were no longer able to read their characters, but at the same time, English speakers can no longer read (or speak) Gothic or Proto-Germanic (if they were to have writing) or Old Norse. Languages change naturally, for the better, just as species evolve to become stronger. Therefore, in my opinion, changing to a Romanized script for Chinese would be ultimately beneficial.

  26. Mr. Shiny & New said,

    July 22, 2010 @ 11:41 am

    As an English-speaking student of Mandarin, I can attest to the concept of character amnesia. I can recognize far more characters than I can write, but thanks to Microsoft's pinyin IME I can even type characters which I don't recognize, with a fairly high certainty.

    For example, if I an writing a word or phrase that the IME recognizes, which has multiple characters, the IME is very good at suggesting the right characters. Library is tushuguan. In Simplified Chinese I easily recognize 图书馆 but the Traditional 圖書館 I only partly understand. But the IME knows more Chinese than I do, and so even though I don't recognize the first Traditional character, and only vaguely recognize the middle one, I can be pretty confident that it is the choice I want for this common word.

    I am not sure how much evidence is really required to prove that the degradation of memory actually occurs. I have seen the exact same thing happen for math: students in school who used calculators were much worse at simple math than students who didn't use them. This didn't have much consequence on final grades, since calculators were available and reliable, and the math being tested was not multiplication, but the result was entire classes of students who have extreme difficulty calculating sales-tax or tips. This is a purely mental exercise analogous to memorizing characters. The more you use it the easier it is to use. But whether the rate has increased, or whether there is an actual problem, is open to debate I would say.

  27. Daniel von Brighoff said,

    July 22, 2010 @ 11:47 am

    @chris: Moser's point isn't that Chinese-speakers occasionally get a stroke or a radical wrong (the Sinographic equivalent of a misspelling), it's that sometimes they "cannot even put the first stroke down on the paper". (My emphasis.) Can you really conceive of a literate, educated, native speaker of English who not only can't spell "causeway" or "recipe" correctly but can't even take a stab at it because they literally can't remember a single one of the letters that make it up?

  28. Greg said,

    July 22, 2010 @ 11:55 am

    @Jarek: Stroke order is pretty important when writing complex characters with subtle differences. Even two- or three-stroke characters might be confused if your stroke order/direction is not consistent (人 vs. 入, for example). Also, in regular hand-writing (not even ultra-cursive), strokes tend to flow one into the next in a patterned and recognizable way; drastically changing order/direction might produce something legible but not pretty if written carefully, but entirely illegible if carried out too quickly. All that said, I know that at least in Japan there are a fair number of people whose character writing repertoires include some "non-standard" components wrt stroke order (usually limited to a few characters).

    @language hat: Not sure about the situation with Chinese, but my impression is that romanization (or pure kanafication) in Japanese would create a highly inefficient system, largely due to the extraordinary number and nature of homophones that are differentiated with kanji.

  29. vanya said,

    July 22, 2010 @ 12:03 pm

    "romanization (or pure kanafication) in Japanese would create a highly inefficient system, largely due to the extraordinary number and nature of homophones that are differentiated with kanji."

    Of course a Japanese person can always write down any word using kana which makes the burden of characters far less severe. It may well be that Chinese will eventually move to a similar "mixed" system of pinyin and characters. Having 500-1000 basic characters available is pretty handy shorthand for homophones, and the discontinuity with the past would be far less severe.

  30. Bill Walderman said,

    July 22, 2010 @ 12:46 pm

    'Now, Peking University is usually considered the "Harvard of China". Can you imagine three Ph.D. students in English at Harvard forgetting how to write the English word "sneeze"??'

    But I'll bet the Ph.D. students at Peking University would have no problem with the specialized vocabulary of the areas in which they work and abstract vocabulary that they would need in the course of their studies–words like "relationship" or "introduction." And, if my experience learning the vocabulary of foreign languages is any guide, those Ph.D. students' passive vocabulary–the ability to recognize written words like "sneeze"–is probably much larger than their active vocabulary–the ability to write the same words.

    That's not to suggest that Chinese characters don't create problems for Chinese literacy. I don't know very much about Chinese characters (what little I do know comes mainly from McCawley's "Eater's Guide to Chinese Characters," which allows me to identify my favorite dishes without knowing a word of any Chinese language and without being able to write any characters at all from memory–evidence of how passive abilities outstrip active abilities). But I suspect that what I've read elsewhere is true: the complexities of Chinese characters do impose a heavier burden on education and other aspects of life than alphabetic or syllabic writing systems.

    With regard to the subject of the current post, however, it strikes me that it's an open question whether the inability of otherwise literate Chinese speakers to write common words, especially those that they don't have to write on a daily basis, is a recent phenomenon due to computers or a long-standing situation prevalent throughout Chinese cultural history.

  31. Daniel von Brighoff said,

    July 22, 2010 @ 1:29 pm

    @Greg: In Asia's orthographic dilemma, William Hannas makes the point that the vast number of homophones is a bug resulting from the nature of these writing systems. In other languages, their proliferation would be kept in check by the necessity of differentiating the neologisms in speech. Only character-based systems make it trivially easily to coin new written words without any consideration of the problem of differentiating them when spoken.

    (It strikes me that a possible counterargument to Hannas is the rampant proliferation of acronyms and initialisms in English. But relatively few of these become actual "words" in their own right.)

  32. Rodger C said,

    July 22, 2010 @ 1:37 pm

    @Jen:

    "English speakers can no longer read (or speak) Gothic or Proto-Germanic (if they were to have writing) or Old Norse." What in Gothic is as worth reading as the Shi Jing, the Daodejing, or anything else written in Chinese before the last few centuries?

    "Languages change naturally, for the better, just as species evolve to become stronger." A complete misunderstanding of both biological and linguistic evolution.

  33. Kai Samuelsen said,

    July 22, 2010 @ 2:06 pm

    "Furthermore, every time Zhu Linfei has to stop to take out her cellphone crutch to remind her how to write a character, she is wasting time, and that in itself is a problem."

    So wasting time is a problem when you're checking your cellphone for the correct character, but not for the years and thousands of hours you've spent memorizing characters that you can input through the much more efficient pinyin or romanization set-ups.

    We revere traditional chinese characters for no reason other than that they are old, while ignoring the fact that their abandonment would bring literacy to millions of the poorest peasants. You lament the loss of "literacy" for Zhu, but she's not losing literacy, just one kind of literacy. Her ability to communicate is undiminished, as her cellphone usage shows. So what is lost if young Chinese let their calligraphy fall into disuse?

  34. J-P said,

    July 22, 2010 @ 2:15 pm

    Some people are (rightly) failing to see an actual problem here. It's orthographical, a slight curiosity, but neither a crisis, nor a linguistic matter. It's borderline-irresponsible to even compare simple *forgetting* to aphasia/dysphasia. And Dr. Mair may not be aware of dysgraphia – which this also does not resemble.

    Get back to me when people start forgetting how to say '噴嚏'. That's a problem.

  35. Eric said,

    July 22, 2010 @ 2:37 pm

    Re: Fallows' blog

    That recently occurred to me when I responded to a Chinese person's tweet. My first thought: "Hey, no fair!
    "That's cheating!"

  36. J.R. Omahen said,

    July 22, 2010 @ 2:41 pm

    Once again, I come out to comment, but hopefully with less ignorance. :)

    What I find interesting, is why do the only options have to be "retain the characters as-is" or "use romanisation"? Why ignore the obviously relevant solution of Bopomofo? Political reasons aside, this system has proven quite effective at representing the spoken language phonologically, and as an aide in learning how to read and write the characters growing up. I should think that having a phonologically-based syllabary similar to Japanese kana is far more efficient for Chinese (especially with the tones) than a specialised subset of Latin characters.

    Lest it not be considered, while Kanji has helped Japanese with its homophonic ambiguity in writing, the same could be said for Chinese characters for Chinese. Switching to a purely syllabic writing system could cause initial shock with homophones. Certainly, humans having robust minds, and language requiring context to function, it wouldn't be an issue in the end, as it would likely be abundantly clear what word is really meant in a phrase or sentence.

    Interesting discussion, as always. :)

  37. Don Sample said,

    July 22, 2010 @ 2:49 pm

    I don't know about Chinese, but using a keyboard for any serious writing that I've done in the last 30 years, has pretty much destroyed my ability to write English cursive script. Just about everything that I have written in the last 30 years that I wanted anyone else to read, has been typed. When I write by hand, its usually quick notes that are for myself only, and its made up of a mix-mash of printed and cursive characters.

    I just tried an experiment, writing the first sentence of this post out, in cursive. I was immediately struck by the fact that I was uncertain how to write a capital "I". The rest of it went along, sort of okay, but it looks awful. Sort of like what you'd expect to see from a grade four student (but at least my spelling is better than it was back then.)

  38. Lazar said,

    July 22, 2010 @ 2:55 pm

    @Greg: I've never understood that point regarding homophones. If they're not distinguished in speech, why do they need to be distinguished in writing?

  39. vanya said,

    July 22, 2010 @ 3:10 pm

    their abandonment would bring literacy to millions of the poorest peasants.

    This is the argument that drives me crazy. There's really no evidence for this. Millions of poor peasants in Latin America cannot read despite living in countries that use writing systems that are far more logical than English, never mind Chinese. Millions of poor Taiwanese (and mainland) peasants have learned to read with characters. China historically has been one of the most literate societies in the world despite it's cumbersome system. Abandoning characters will probably make day-to-day life easier for most Chinese. But it won't be a magic bullet, and any improvement in literacy or economic efficiency will probably be marginal. There is this assumption that all the hours "wasted" learning characters will magically be channeled into some more productive activity. Some people may take advantage of that time to do great things, but more people will probably end up just having more time in their childhood available to text friends, play sports, watch TV and play video games. Certainly not a bad thing if you're concerned with human happiness, but abolishing characters probably won't do much to improve China's competitive position in the world.

  40. Sili said,

    July 22, 2010 @ 3:10 pm

    Not strictly relevant, but I was a good speller while in school – the only subject I liked it Danish. Actually composing texts of my own has always been loathsome to me.

    Over the last ten years or so, I have increasingly been reading, writing and speaking English, to the point that I know think in English to a considerable degree.

    As a result I often find it hard to spell even common Danish words, when I have to, and I joke that I'm sesquilingual because learning the English word for a specific concept often leads me to forget the corresponding Danish one.

    (As for my handwriting it has always been poor, and I doubt that I, myself, can decode text I have no memory of having written.)

  41. Greg said,

    July 22, 2010 @ 3:14 pm

    @Daniel von Brighoff: Sounds like an interesting read. I'm curious, though, as to the scope of the book. The writing systems of "Asia" are quite varied — the "dilemma" seems to be with Chinese characters specifically. I often thought the Japanese situation was quite unique, though, in that the majority of homophones are phonologically Japanized Chinese readings of two or more characters; this in combination with Japanese's relatively simple sound system (five vowels and a handful of consonants; no complex vowels; no distinguishing tones) makes for a whole slew of true homophones that are actually brought about BY the writing system.

    My impression of the various languages of China is that their sound systems are robust enough so that an entirely phonetic writing system would not cause as much of a problem as it would in Japan. If they're anything at all like Vietnamese (I can't even count how many diphthongs and triphthongs there are, plus so many consonants and tones), they might be much better off with an entirely phonetic system — but then there's the problem of a widespread writing system understood by all despite no/low mutual intelligibility in speaking (or is this a gross exaggeration I have taken as a fact about "Chinese"?)…

  42. language hat said,

    July 22, 2010 @ 3:18 pm

    @vanya: I can certainly understand why a lot of Chinese would be opposed to that.

    So can I, and I've never said otherwise. I was talking about a refusal to admit that there's anything to discuss. J-P provides a classic example: "Some people are (rightly) failing to see an actual problem here." Nothing to see here, move along…

    The move to romanization may well be inevitable, but lots of inevitable developments aren't necessarily worth celebrating.

    I'm not celebrating it. I love the characters and the whole cultural complex that goes with them; I'm your basic laudator temporis acti. But I'm not so much in love with them that I fail to see the basic injustice in preferring cultural literacy for a few to true literacy for the masses. As Kai says, "We revere traditional chinese characters for no reason other than that they are old, while ignoring the fact that their abandonment would bring literacy to millions of the poorest peasants."

  43. language hat said,

    July 22, 2010 @ 3:22 pm

    @Greg: my impression is that romanization (or pure kanafication) in Japanese would create a highly inefficient system, largely due to the extraordinary number and nature of homophones that are differentiated with kanji.

    I ask the same question I ask when people say the same thing about Chinese: how do you suppose Japanese manage to communicate orally when they don't have the kanji to differentiate those pesky homophones? Sure, if they eliminated the characters they'd have to change their writing style accordingly, but that doesn't seem a huge barrier.

  44. Terry Collmann said,

    July 22, 2010 @ 3:26 pm

    Vanya: "Every thing, every inscription, every sign, every family heirloom, every restaurant menu, every blog post, ever written in Mandarin, Cantonese or other Chinese language before 2010 will be inaccessible to future generations"

    Nah, you're not thinking it out. You'll point your cellphone at whatever you want to read in Chinese characters, (or similar device) and it will either read it out to you aloud or display it in whatever sort of character you choose. C'mon – Zhu Linfei is already doing this in reverse to generate characters she can't remember: it won't take much of an upgrade to do it the other way round.

  45. Greg said,

    July 22, 2010 @ 3:30 pm

    @Lazar: There are actually cases in spoken Japanese when people refer to the writing of a kanji (or use other sorts of circumlocution) in order to distinguish homophones. The only example I can think of off the top of my head is shiritsu, meaning (roughly) either "private" (私立) or "public" (市立). To distinguish, people will often refer to the former by using a different reading of the first character than the one actually called for. Like I alluded to in my comment to Daniel, I think these homophones come about from taking morphemes that are phonologically distinguished in the source language, and using them in words where they are not distinguished in the target language. The book he recommends seems to have another thesis, but it could be thecae that these both apply.

  46. J-P said,

    July 22, 2010 @ 4:14 pm

    language hat said,

    I was talking about a refusal to admit that there's anything to discuss. J-P provides a classic example: "Some people are (rightly) failing to see an actual problem here." Nothing to see here, move along…

    You're selectively reading me. There's nothing linguistic here to discuss. Certainly nothing neurolinguistic. That's an effortless conclusion; no hoop-jumping required.

    Philology is great. Orthograpahy is fun to ponder. Sociology is… something. But this is Language Log – what sort of response do you expect? Keep your chocolate out of my peanut butter.

  47. Katy said,

    July 22, 2010 @ 4:15 pm

    "Not sure about the situation with Chinese, but my impression is that romanization (or pure kanafication) in Japanese would create a highly inefficient system, largely due to the extraordinary number and nature of homophones that are differentiated with kanji."

    I'm a native English speaker currently studying Japanese. I don't know about homophones being a big issue, but it certainly is easier to read Japanese when kanji is interspersed. I think this is because each word is better differentiated from the next when kanji is used. Since there aren't spaces between words in Japanese (and, I believe, Chinese?) it's a bit more difficult to read phonetically. ImagineifEnglishwereallwrittenlikethis, itwouldbeatadmoredifficulttoread. Kanji helps to solve this problem.

  48. Xmun said,

    July 22, 2010 @ 4:55 pm

    itmightbeatadmoredifficulttoread
    nodahpertsuoboslaserehtfiyllaicepse

  49. Nathan Myers said,

    July 22, 2010 @ 5:15 pm

    I don't think I'm alone in seeing this as a wholly positive development. The writing system is a yoke on the Chinese neck. As noted, machines can remember the traditional characters at need. I predict the simplified forms will disappear as superfluous and insufficiently traditional. Only idle classicists will bother learning to write traditionally. Regular people will use modern alphabets, and the whole nation will benefit enough that a small fraction of the new surplus could support a corp of classicists, along with a wholesale translation of old literature to the new representation.

    Educated people in western societies have mostly abandoned Latin and Greek, and suffered little for it. The Chinese will suffer even less.

  50. Kalirren said,

    July 22, 2010 @ 5:19 pm

    "What I find interesting, is why do the only options have to be "retain the characters as-is" or "use romanisation"? Why ignore the obviously relevant solution of Bopomofo? Political reasons aside, this system has proven quite effective at representing the spoken language phonologically, and as an aide in learning how to read and write the characters growing up. "

    Seconded! I'm surprised Putonghua education in the western world doesn't use this natural "Chinese alphabet". It would solve a lot of accent problems.

    I think what may eventually be lost with the advent of digital Chinese is the idea that "writing" a character in its entirety is important. At present, I don't think any sane Chinese person would call someone uneducated for forgetting how to -write- as obscure a character as the "ti" of "penti", but would definitely call them uneducated for not knowing what it meant if it were presented to them and they were asked to -read- it. If at some point in the future, you can reliably input a character with a combination of pinyin/Bopomofo and principal radical, what's to lose?

    "then there's the problem of a widespread writing system understood by all despite no/low mutual intelligibility in speaking"

    This is true. This might actually be something to lose. Local dialects can be very different from Putonghua. You can't expect to get a service job in Shanghai if you don't speak Shanghainese.

  51. dw said,

    July 22, 2010 @ 5:30 pm

    Literacy in both Korea and Vietnam improved dramatically after those countries switched away from Chinese-based orthographies to more phonetic writing systems. Is there any reason to think that things would be different for China?

  52. Kalirren said,

    July 22, 2010 @ 6:04 pm

    The difference, I would venture, is that Chinese speak Chinese, whereas Koreans speak Korean and Vietnamese speak Vietnamese.

    If we started teaching English in Tamil script I'm sure our literacy would drop.

  53. Rodger C said,

    July 22, 2010 @ 7:05 pm

    First of all, let me say that my knowledge of Chinese is rudimentary and I have no competence to tell the Chinese people what would be good for them. However, I'm struck by what seem to me the obvious flaws in some of the arguments for romanization here.

    What I always envied the Chinese for was the fact that their whole literate history was open to them without translation and all its pitfalls. "[A] wholesale translation of old literature to the new representation"–does that mean there'll be a single Putonghua translation of each Wenyan text, foreclosing all other possible meanings, to say nothing of the inevitable mistakes and tendentious distortions? I'm sure Beijing's masters will love that prospect; it'll finally give them the power of the Qin emperor to abolish the past except for what they want to be known about it, and without a single bookburning or execution. What experience does Nathan Myers have with translation theory and practice? Or with the economics of the publishing industry?

    When I teach Latin and Greek literature, I often recite passages aloud in the original just to work up my own enthusiasm for the text in the face of the prosaic or sputtering version on my students' pages (a different one with every new edition). When I teach classical Chinese literature, I give them several versions of, say a Shi Jing poem (Legge, Waley, Pound, etc) to show them how drastically versions can differ. I don't want to see the Chinese people put in my students' position of disadvantage, at the mercy of intermediaries.

    "Educated people in western societies have mostly abandoned Latin and Greek, and suffered little for it." To the contrary, students today can't learn scientific and medical vocabulary etc. without a great deal of memorization that would have been easier had they learned the words in the context of actual reading at an earlier age. My colleagues in the sciences are always complaining about this. Oh, but they're idle classicists. My own love of Latin in school came first and foremost from the fact that it *explained* so much. And at the time I expected to be a scientist.

  54. Janne said,

    July 22, 2010 @ 7:18 pm

    "I ask the same question I ask when people say the same thing about Chinese: how do you suppose Japanese manage to communicate orally when they don't have the kanji to differentiate those pesky homophones?"

    As in most languages, you use a smaller vocabulary in speech than in writing. Many of the troublesome homophones only occur among words used in writing (the "私立・市立" example above being an exception; "科学・化学" (science/chemistry) being another). That is most likely not a coincidence of course; people will naturally gravitate towards unambiguous expressions in speech.

    But Unger above seems to be right if I understand trends here: computers and cellphones have indeed caused an increase in the use of kanji here. In Japanese, you always have an option to use kana instead of kanji, and there's frequently more than one possible character you can use for a given meaning. People today tend to use more characters and more rare and complex characters than they used to.

    This most likely a benefit of offloading part of your memory. When you don't have to remember how to write a rare character it becomes much easier to make use of it. When your software helpfully pops up all the possible options you have, you only need to vaguely recognize the character to make use of it. The list of standard kanji to learn in school was recently increased with about 200 characters, in part because of this.

  55. Leonardo Boiko said,

    July 22, 2010 @ 7:37 pm

    @languagehat
    > how do you suppose Japanese manage to communicate orally when they don't have the kanji to differentiate those pesky homophones?

    Well, for one thing they have accents, which for some reason neither kana nor hepburn or kunrei romanization register (perhaps because they vary so much between dialects). For example, “bridge” in kanji is 橋, and “chopsticks” is 箸. In spoken language, chopsticks is ha↓shi, and bridge his hashi↓ (with the arrow denoting a drop in tone in the next syllable). In kana and romanization these different words become homographs: simply「はし」 or “hashi”.

    In my very humble opinion a romanization proposal for Japanese could perhaps start by coming up with a way to write pitch, thus reducing the number of homographs. But yeah, you’re right—the existence of spoken Japanese should be enough to prove that kanji, no matter how much we love them, are not _necessary_ for communication. William Hannas, I believe, also has a point that the use of kanji probably encouraged the proliferation of homophones (I haven’t studied older forms of Japanese yet, but it’s clear that the number of (native) kun-yomi homophones is much lower than the number of on-yomi homophones, which came with the characters).

    [(myl) You wouldn't have to work hard to "come up with a way to write pitch", since in the Tokyo dialect anyhow, the only options are to mark one (or none) of a word's syllables as accented, which could trivially be done (for example) with an acute accent.

    Kansai dialects also have a lexical distinction in "melodies", which basically amounts to the choice of a word starting low or not. Again, it's not hard to find a way to write this — say with an optional grave accent.

    But this points to a more significant difficulty — accent varies a lot by region. There are even dialects with no lexical accent. And even within (say) Tokyo dialect, there is a perhaps-surprising amount of variation in which (if any syllables) individuals treat as accented in a given word.

    So my impression from all of this is that the functional load of accent in Japanese is probably not terribly high. It would be easy to test this — look at speech comprehension in the dialects that have no lexical accent, or in whispered Tokyo speech.]

  56. dw said,

    July 22, 2010 @ 7:49 pm

    @Kalirren:

    But isn't pinyin widely used already in the Chinese educational system, in some cases before Chinese characters are taught? According to Wikipedia, pinyin "was introduced to primary schools as a way to teach Standard Mandarin pronunciation and used to improve the literacy rate among adults".

  57. Nathan Myers said,

    July 22, 2010 @ 8:01 pm

    Kalirren seems to labor under the common misapprehension that the numerous Chinese "dialects" are all written the same way, and only differ in vocal pronunciation. In fact, as I understand it, they differ from one another as much as do European languages, and are mostly unwritten. When speakers of those languages are called upon to write, they translate to and write Mandarin. Curiously, the overwhelmingly most numerous group who so misapprehend appears to be exclusive Mandarin speakers. On reflection this shouldn't be surprising.

    Rodger C seems to assume that a transcription necessarily loses the essence of the original, but there need be little more loss than in switching from uncial script to Times Roman. Prof. Mair frequently presents romanized names of characters that map unambiguously to a corresponding traditional character, lacking only the actual calligraphic strokes. Puns resulting from accidental resemblance between various traditional characters would become unintelligible, but that seems rather a small tradeoff. (No doubt there will be compilations of traditional puns, for reference.) In any case, the outcome does not depend on how any of us feel about it.

  58. Kalirren said,

    July 22, 2010 @ 8:05 pm

    I haven't been through the educational system in China, I don't actually know. I wouldn't be surprised if that were true, though. The CCP has put a lot of effort into promoting Putonghua.

    The comment that pinyin is used to improve the literacy rate among -adults- is very interesting to me. The only adult educational use for pinyin I could see would be to promote Putonghua over local dialects, since recognizing the characters is what makes for literacy in Chinese. No sane Chinese teacher would use pinyin alone to teach the language itself.

    On that thought, I wouldn't be surprised if literacy figures are skewed in China because people know how to read the characters in their local dialect can't read them in Putonghua, which is Standard with a capital S.

  59. Kalirren said,

    July 22, 2010 @ 8:13 pm

    "Kalirren seems to labor under the common misapprehension that the numerous Chinese "dialects" are all written the same way, and only differ in vocal pronunciation. In fact, as I understand it, they differ from one another as much as do European languages, and are mostly unwritten. When speakers of those languages are called upon to write, they translate to and write Mandarin. Curiously, the overwhelmingly most numerous group who so misapprehend appears to be exclusive Mandarin speakers. On reflection this shouldn't be surprising."

    I think the best way to say it is that there's no -other- way to write the dialects. The phrases are different, the compounds are sometimes different. Sometimes when writing Cantonese you do even see the use of defunct characters that no longer exist in Putonghua. And there's even a special tag on Cantonese wikipedia that says, "This article was copied from the Putonghua Wikipedia and needs to be un-Putonghua-ified," which gives you a good idea about exactly how far apart/close together the dialects are.

    But the dialects are generally similar enough that any string of characters written in Putonghua is legible in any Chinese dialect – a native dialectical speaker might think the wording awkward, but it's legible.

  60. Rodger C said,

    July 22, 2010 @ 8:19 pm

    "Rodger C seems to assume that a transcription necessarily loses the essence of the original, but there need be little more loss than in switching from uncial script to Times Roman."

    I wasn't talking about "transcriptions" at all, but about translating Wenyan into Putonghua, which involves (at the very least) greatly increasing the number of syllables to avoid things like the example discussed elsewhere, "Shi shi shi, Shi shi, shi shi, shi shi shi shi," etc. That's a made-up example, but Wenyan simply can't be written in Pinyin without massive ambiguity. Ancient Chinese syllables were much more phonetically complex and thus could carry a correspondingly greater semantic load. Rendering from Wenyan into Putonghua isn't "transcription" but translation into a different language, with all the pitfalls of translating from Latin to French.

  61. Kragen Javier Sitaker said,

    July 22, 2010 @ 8:45 pm

    John Cowan writes: "My given names are from my grandfathers, and to change them would be tantamount (in the ways of verbal magic) to changing them."

    Most women's surnames are also from their grandfathers, and yet in the US (although not in China) they commonly change them upon marriage. So did I.

  62. Victor Mair said,

    July 22, 2010 @ 9:13 pm

    On the charge of anecdotalism: If every single Chinese person one knows (and one knows hundreds upon hundreds of Chinese individuals), who regularly uses computers and cell phones to write characters and does not regularly practice their handwriting, complains about the marked decrease in their ability to produce characters by hand, is one not entering the realm of statistics?

    One of my very best friends in China, Xu Wenkan, a senior editor of the Hanyu Da Cidian (China's equivalent of the Oxford English Dictionary) knows how to write Chinese texts on a computer or a cell phone, but he almost never does so. Instead, he writes everything out by hand. If he needs to submit a manuscript somewhere, he can always hire a WUBI drone who will do the typing cheaply and thoughtlessly. When he sends me messages, he always writes them out longhand (and a very beautiful, exacting hand it is) on a piece of paper. In the past he would fax the messages to me, but now he scans them as a pdf and attaches them to an e-mail that has just these words:

    ====

    Please to read the attachment.
    Wenkan

    ====

    It was a conscious decision on Mr. Xu's part NOT to type things with a computer so that he would retain his wonderful ability to write by hand. It is sort of like the samurai Giving up the Gun [: Japan's Reversion to the Sword, 1543-1879 — Noel Perrin's great book] or Middle Easterners rejecting the wheel in favor of their tried and true ship of the desert, the camel (see Richard Bulliet's seminal social history on that topic).

  63. Kragen Javier Sitaker said,

    July 22, 2010 @ 10:39 pm

    Kalirren: We do actually teach English in Latin script, and the phonology of English is probably as close to the phonology of Tamil as it is to the phonology of Latin. So I'm not sure switching to Tamil would create a greater problem, aside from the practical difficulty of transliterating everything written so far in English into Tamil so that New English readers could read it, and the somewhat worse complexity of the Tamil writing system.

    Nathan Myers: Now that I more or less speak Spanish, I can read Portuguese, Galician, and Catalan without much difficulty, although I frequently make hilarious mistakes, and I can read quite a bit of Italian and Romanian (and French, although I studied French before, so the comparison is not quite apt). German, Danish, Finnish, Basque, Hungarian, Icelandic, and Turkish are quite another matter, though. (Dutch isn't too bad, really; almost at the level of Italian for me as an English-speaker.)

    So I suspect that the proper European reference point for the Sinitic language family is "Romance languages", not "European languages".

  64. Qov said,

    July 22, 2010 @ 10:55 pm

    I only know how to write one Chinese character (it's the ma that looks vaguely like a horse) but I don't think the "technology erodes traditional language skills" issue is confined to characters. In school I was a really good speller in English, but I think my ability to recognize and correct misspellings in my own language has declined considerably due to the ubiquity of the squiggly red line and right click suggestions. Most spellcheckers use either American or British spelling rules and I'm afraid that the unique Canadian spelling system will disappear.

  65. Nathan Myers said,

    July 22, 2010 @ 10:56 pm

    Kragen: Are all Chinese "dialects" really in the Sinitic family?

  66. Nick said,

    July 22, 2010 @ 11:01 pm

    One more vote for the utility of the cellphone — when I was taking courses in Chinese in 2001, the professor (of a kind of Dao/Ru/Fo 'basics' class) would often forget how to write characters he wanted to put on the board, and would ask the Chinese college students in the class to walk him through some characters (there is a set of fairly quick ways to describe Chinese characters using spoken Chinese). They, in turn, almost always went to class with a small electronic dictionary — not because they needed it more because they were students (people right after the college entrance exams know the most characters) but because anyone who's serious about keeping their writing up needs to carry a dictionary of some kind.

    Cellphones may have decreased the number of notes that people write to each other — but my subjective experience is that they greatly increase the amount of writing people do to one another, and plus they've expanded the number of people who carry a character dictionary around, so they make writing possible in some situations (with the character lookup).

    I think in general, without the support of IME pinyin and other digital facilitators, the Chinese character would be dying a much swifter death.

  67. Veiyi said,

    July 22, 2010 @ 11:41 pm

    Just curious, is 噴嚏 ever pronounced pen qi?

  68. Ellen Marker said,

    July 22, 2010 @ 11:53 pm

    @ JimG re: "Yes, Jennifer 8. Lee. She's a journalist and writer, not given a middle name at birth. She chose "8." as a teenager. It's a lucky number associated with prosperity and good fortune. "

    THE TRUTH: When she went to Harvard as an underclassmen, first year, there were like 17other Jennifer Lees in the freshman class, and her mother suggested to her that she take a special middle name in order to stand out, and mom suggested the number 8 since it has magical properties in Chinese culture, where her mom came from, so Jenny adopted 8. with a period as her middle name and it began appearing on her byline at the Harvard Crimson during her undergrad years there. Then when she joined the NY Times, the editors graciously allowed her to retain the 8. with a period as part of her byline, since she had already used it at the Crimson and the Boston Globe, BUT the Times rarely allows these things to happen.

    Which reminds me: there is a news photog in Florida whose byline carries a "tilde" in her name for photo credits is: google to find it

  69. John said,

    July 23, 2010 @ 1:23 am

    Its very interesting. I have been trying to learn mandarin for 20 years- pretty much unsuccessfully. But I'm going to Japan for a few weeks at the end of the year so tried to learn some Japanese. Within three months I can already make sense of some headlines in newspapers; the katakana and hiragana are incredibly helpful in giving the kanji context, even if remembering the pronunciation in context is difficult. I have 'on the road' in Japanese and can work out one or two sentences; I still can't achieve that minimal level in Chinese after all this time. And the main problem for me is working out where words start and finish, as the article mentions.

  70. PKA said,

    July 23, 2010 @ 2:40 am

    I" think in general, without the support of IME pinyin and other digital facilitators, the Chinese character would be dying a much swifter death."
    No, I can live WITHOUT PINYIN.
    PINYIN IS EVIL.
    First, pinyin forces you to think in Mandarin pronunciation and that helps wiping out other Chinese languages, which many of them are unintelligible with Mandarin.
    Second, pinyin input method does not require you to remember HOW TO WRITE it.
    "Cangjie", which requires you to remember the stroke order and shape of a word, is a better input method for not to forget how to write Chinese character totally. (even for Chinese people.) There are so many cases that Chinese (or Taiwanese) people who uses pinyin (or bopomofo) forget how to write Chinese character.

    How do you input through Cangjie?
    蝦 -中戈口尸水
    再見 – (一土月) (月山竹山)
    噴嚏 – (口十廿金) (口十月人)

  71. pka said,

    July 23, 2010 @ 3:12 am

    Here's a video of me using cangjie in here.

    http://www.youtube.com/watch?v=qWnLMP5zyeg
    Most of the time, you don't need to choose anything from table.

    I apologized that I left so many comments but I feel I have to tell the world that there is a stroke-order input method and it is quite fast.

  72. C. Custer said,

    July 23, 2010 @ 3:57 am

    Can't say I think this is a huge problem. When was the last time anyone who commented here wrote "sneeze" by hand in English?

    Most educated Chinese know how to write by hand the characters they often have cause to write by hand. That's common sense and also, I assume, rather natural. If, on occasion, they come across something they don't know and have to look it up on their phone, that wastes a little time (generally about 5 seconds).

    But how often does that really happen? Once a day? Once a week? And how much time would it take of consistent review each week to ensure they actually remembered how to write these characters?

    Obviously this is purely speculative, but I suspect that checking characters by phone every now and then actually saves a little time in the overall scheme of things. And while I understand the romantic attachment to the idea of a civilization full of people who can write characters fluidly and cherishes China's writing traditions, there's a reason everyone uses pinyin input and email rather than memorizing stroke orders and writing letters by hand. It's quicker and easier.

    Not that quicker and easier always means better. But I can understand why these "educated Chinese" would rather spend their time on other things than break out their character workbooks and practice the characters for "sneeze", a word I suspect they have cause to write by hand only once every few years.

  73. outeast said,

    July 23, 2010 @ 6:51 am

    @ Victor Mair,

    I don't want to be a arse, and I know it's probably annoying to have total ignorami like myself challenging the view of an expert like yourself; but your colleagues over there at Language Log Plaza have spent years now relentlessly drumming it into us that we should be seriously sceptical of hell-in-a-handbasket claims about language – and from here in the bleachers this does sound remarkably reminiscent of the TXTNG W/ DSTRY ENG mania and suchlike language claims.

    Prefaced thus – and with apologies if this is both annoying and obtuse…

    I would have thought your last claims about the large numbers of people anecdotally reporting this effect is unlikely to be prophylactic against the recency illusion (or at least the possibility of it). Just because 'everyone' now attributes the erosion of their character recall to computers does not mean they are right; it may be that many people have always 'lost' their character recall over time but either denied it, downplayed it, or attributed it to some other cause (or variety of causes). Now there is an appealing and plausible cause, and one that 'everyone' is affected by.

    Your counter example of Mr Xu may not be a very useful data point; after all, if he chose not to use computers specifically in order that he should retain his ability to write characters then obviously he already placed a high premium on character recall. Plausibly that factor (and related values, such as a certain cultural conservatism?) might have been at least as important in his skill retention as his rejection of computers.

    Again, sorry if this comes across as thoughtless or rude.

  74. Steffan Fennander said,

    July 23, 2010 @ 8:17 am

    @ Ellen Marker wrote.." there is a news photog in Florida whose byline carries a "tilde" in her name for photo credits is: google to find it."

    This one? Mari Darr~Welch?

  75. minus273 said,

    July 23, 2010 @ 8:23 am

    @ Nathan Myers,
    Yes, those non-Sinitic ones are called "ethnic languages", and their speakers classified as non-Han. There is only one exception for the ethnic classification, the Kradai Lingao speakers in Hainan are classified as Han, their language still refered to as a "language" not a "dialect".

  76. Frans said,

    July 23, 2010 @ 8:23 am

    @Eric:

    That recently occurred to me when I responded to a Chinese person's tweet. My first thought: "Hey, no fair!
    "That's cheating!"

    Using ligatures could probably boost your output a bit. You could automate it through some UserJS.

  77. Kai Samuelsen said,

    July 23, 2010 @ 9:25 am

    "On the charge of anecdotalism: …"

    It's more like flirting with the realm of statistics. If you go by the people you know, even if you know a large number of people, they probably aren't a representative sample of the population as a whole, because they are a self-selected population, rather than a random sample. If I were to take all the people I know, I'd find an abnormally high percentage who were extremely knowledgeable about the physical properties of asbestos – much higher than in the population as a whole, because I used to work at an asbestos laboratory. Not to discount anecdotal evidence, because with hundreds and hundreds of people it can be a good indicator, but it's still not quite statistics.

  78. vanya said,

    July 23, 2010 @ 10:00 am

    But I'm not so much in love with them that I fail to see the basic injustice in preferring cultural literacy for a few to true literacy for the masses.

    Again – there is no evidence I've ever seen that characters are really the root cause of cultural illiteracy in China. Please show me a country where "true literacy for the masses" exists. It certainly does not exist in the US, it doesn't even exist in Russia, one of the most culturally literate countries I know, nor does it exist in Turkey, despite Ataturk's reforms. Pretending that characters are the problem is just a convenient way to ignore the myriad of other social, cultural and genetic factors that influence literacy.

  79. David J. Littleboy said,

    July 23, 2010 @ 12:05 pm

    "computers and cellphones have indeed caused an increase in the use of kanji here."

    I think it's settled down to a sensible state here. I showed up here in '86 when word processors and peecees were first starting to do kanji seriously (and mfrs were making it a point of pride to handle more kanji than the competitor's current model), and people went crazy overboard tossing in strange kanji by the boatload. It was quite fun. Since my Japanese was fairly good, I could tell if a character was rare enough that the bloke I wanted to embarass was unlikely to know it, and I would innocently ask him or her how to read it. (This was at a large electronics companies that was one of the guilty parties in overdone kanji; and the internal memos that came by were pretty wild with kanji overuse.)

    But nowadays, everything I see at work is very civilized in terms of not using off the wall kanji. Of course the literary magazines are quite dense of unusual characters, but those blokes are using them for good reasons.

    By the way, Japanese without kanji isn't going to happen. Ever. So there's not much point telling the Japanese how wrongheaded they are. And I don't think that they are being wrongheaded, at all.

  80. Eric said,

    July 23, 2010 @ 1:20 pm

    @Frans:

    Yeah, but I don't like to tweet in non-ASCII characters (or at least non-Latin-1) because it's ostensibly destined for people's phones. Good lookin' out, though

  81. Tom said,

    July 23, 2010 @ 1:26 pm

    @Chris: Harry S. Truman is generally written *with* the period. The President used to joke about the lack of a period, and some people took the joke too seriously. He himself usually wrote it with a period.

    @Jen: If you find most college Chinese textbooks to be too dumbed-down for a linguist, then you might try the Pinyin Primer from Princeton. The books are designed for college freshmen, but the notes use a whole bunch of linguistic terms. On occasion, they even refer to French and German when it's easier to explain by analogy. Note: there are *four* books in the first-year syllabus, but only *three* of them are sold in a bundle. Be sure to buy the Character Text separately.

    @vanya: There is already a massive break in the Chinese cultural tradition. That's the switch from Classical Chinese to vernacular Chinese. An educated English-speaker can read Shakespeare without much trouble, and Chaucer with some puzzlement. But an educated Chinese speaker would have difficulty reading anything earlier than about 1920, more-or-less. Especially if they know only simplified characters.

    Don't forget that alphabetization was one of the Utopian goals espoused by the Communists back in the 1940s. Simplification was intended to slowly move the country in that direction, without a single big-bang break. Due to the backlash from the Cultural Revolution, even the latest round of simplification was rolled back, as it was associated with leftists and became discredited. Now, with Malaysia and Singapore also having standardized on the rolled-back simplifications, it would be difficult to proceed any further.

  82. Rodger C said,

    July 23, 2010 @ 1:42 pm

    I don't think Minus273 directly answered what I took to be Nathan Myers' question. Yes, the Sinitic family is pretty much defined by the Chinese "dialects"–how could a family be *smaller* than a language group with one dachsprache?–plus, I understand, one "ethnic" language that seems to be descended from Archaic Chinese, or a sister language, but whose speakers didn't form part of Chinese civilization.

  83. Helmut said,

    July 23, 2010 @ 4:22 pm

    Zhu Linfei is mistaken. It is a big problem that she cannot write 20% of the characters she knew just 5 or 6 years earlier.

    Is it though? I'm 24, just like Zhu, and I'd guess that my knowledge of state capitals, organic chemistry, SAT vocabulary, and landmark Supreme Court decision details has similarly declined. Not because I now use a computer or cell phone rather than memory when I need that information, but I so rarely need that kind of information. It'd be a grand waste of time to maintain the kind of ready access to information that I had in high school.

  84. Jen said,

    July 23, 2010 @ 7:41 pm

    @Rodger:

    What in Gothic is as worth reading as the Shi Jing, the Daodejing, or anything else written in Chinese before the last few centuries?

    I do not purport to judge and rank the quality of the writings of a language. Each language has its own intrinsic value. That you would think or do otherwise shows a complete lack of respect for the subject matter at hand and makes me question your understanding of the field.

    "Languages change naturally, for the better, just as species evolve to become stronger." A complete misunderstanding of both biological and linguistic evolution.

    Actually, I'm dead on. Prove me wrong.

  85. Eric said,

    July 23, 2010 @ 9:33 pm

    Re: Mr. Xu

    Can I just use this opportunity to mention Dr. X's legendarily exquisite handwriting in The Diamond Age? OK, thank you.

  86. Rodger C said,

    July 23, 2010 @ 10:40 pm

    @Jen: Evolution doesn't imply improvement. Ask any biologist OR linguist. And what field do you accuse me of not understanding when I say I'd rather have the Shi Jing than the Skeireins? Have you read both of them?

  87. Uly said,

    July 23, 2010 @ 11:11 pm

    "English speakers can no longer read (or speak) Gothic or Proto-Germanic (if they were to have writing) or Old Norse." What in Gothic is as worth reading as the Shi Jing, the Daodejing, or anything else written in Chinese before the last few centuries?

    I don't know. I can't read Gothic.

    But I can, if I want, read any number of Chinese classics, despite not being able to read or speak Chinese either. How? In translation.

    If they were to switch wholesale to Romanization, it wouldn't take very long for people to republish all those worthy books.

  88. Victor Mair said,

    July 23, 2010 @ 11:41 pm

    To those who imagine that computers and cell phones have not had any deleterious impact upon the ability of literate Chinese to write characters by hand, I can only say that you simply have not been observing what has been happening over the course of the past thirty years and more. When people from all walks of life (top translators at the United Nations, intellectuals with and without advanced degrees, scientists, business people, military personnel, teachers at all levels from kindergarten to university, and so forth) constantly use electronic devices to write Chinese for years consistently and uniformly admit to (and frequently complain about) pronounced attrition in their ability to produce characters by hand, you cannot tell them that it hasn't taken place. I have lived and worked closely with countless such individuals for decades, and watched them evolve from people who used to confidently write out letters, papers, and notes longhand into insecure pen and pencil wielders. In contrast, this loss of character-writing ability did not occur over the course of the lifetimes of the previous generations.

    To those who insist that there's no difference between what's happening with Chinese and what's happening with alphabetically written languages as a result of computers, I can speak from personal experience. After thirty years of using personal computers to type my papers and books and twenty years writing tens of thousands of e-mail messages, I still have no hesitation whatsoever to pick up a pen or pencil to write out a note, letter, or manuscript. The same is true of most of my family, friends, colleagues, and students. In fact, I still prefer to write out my serious academic and literary work longhand. Naturally, we recognize the convenience of composing texts with the assistance of electronic devices, but this has not resulted in any shyness about writing things out by hand when necessary or convenient. The natures of alphabet-based writing systems and character-based writing systems are dissimilar, so they are bound to make different demands and have different impacts upon their respective users.

    To those who would assert that I am attempting to be a soothsayer about the future of Chinese characters or that I am monocausally attributing character amnesia to electronic devices, I would reply that I am merely aiming to identify and track observable trends, trends that have been repeatedly remarked upon by literate Chinese themselves.

    To those who keep trying to shift the focus of discussion from the impact of electronic devices upon handwriting ability to the influence of literacy upon society, please let's save that for a more appropriate time.

    BTW, Lane Green was inspired by this LL thread to write not just one, not just two, but three consecutive posts on the subject of "Character deficit" on his new Language blog at The Economist: http://www.economist.com/blogs/johnson

  89. Rodger C said,

    July 23, 2010 @ 11:48 pm

    I repeat, romanizing Wenyan with Putonghua sound values produces unintelligible results. If Putonghua is romanized, the Chinese will be forced to read their own classics in translation, the way most of us read the Icelandic sagas (to pick something much more worth preserving than the Skeireins).

    I also repeat, my knowledge of Chinese is very general–far inferior to my knowledge of Icelandic, for that matter–but does anyone on this thread who's so enthusiastic about romanization even know what Wenyan is? To my recollection, none of them has used the word so far. Is it too much to ask that the actual question be addressed?

    As for the frequent assurance on the part of romanization enthusiasts, English spelling reformers, etc., that everything worthwhile would be quickly republished, that just isn't the way the publishing industry works anywhere, let alone in a totalitarian state whose rulers know that to control the past is to control the future.

  90. Rodger C said,

    July 23, 2010 @ 11:51 pm

    The above was of course meant as a reply to Uly.

  91. Jer said,

    July 23, 2010 @ 11:58 pm

    Watching my Chinese friends chat online using the keyboard, it's clear they have solved the pinyin homophone problem. They have (unconsciously) memorised the position of each character in the homophone list that the computer presents to them. So horse is m + a +right arrow + right arrow and mother is m + a + three right arrows. They can type without looking at the screen. This might be the worst solution, but it is the solution people are using.

    And speaking of worst solutions, I remember how painful writing things by hand was. Trying to do three pages in an hour during an exam left me with cramps for most of the day. Is it possible that people are embracing pinyin because that their hands don't hurt so much?

    The newspaper stories have no comparisons in them, so we can't tell if people without electronic help lose… maybe 50% of their learning after school?

    BTW you can buy an electronic stylus and board that lets you draw characters for the computer to turn into regular chinese text on the screen. Available at most of the computer stores that I've seen.

  92. Rodger C said,

    July 24, 2010 @ 12:00 am

    One more word. It's entirely possible, as far as I'm concerned, that romanization of Putonghua is worth the drawbacks. I'd just like to see some acknowledgment of what the drawbacks are.

  93. Jer said,

    July 24, 2010 @ 12:16 am

    @RogerC everybody understood it the first time. The response to your question wasn't posted because the answer is clearly the same as to the concerns that Greeks can't read the Iliad today, Italians can't read, say.. Virgil, and the English can't read Beowulf.

    * I'm out by 800ish years, but what's 800 years in a forum discussion?

  94. Keith M Ellis said,

    July 24, 2010 @ 12:55 am

    Jen, you wrote:

    "Actually, I'm dead on. Prove me wrong."

    Just to be clear, this is what you had claimed:

    "Languages change naturally, for the better, just as species evolve to become stronger."

    I'd like to be charitable and assume that in both cases you were using the word "better" as "adapted to its environment", but then you write "species evolve to become stronger" and make that generous reading impossible. You clearly mean "better" in some absolute sense.

    Which is utter nonsense. As Rodger wrote, it's nonsense both with regard to biology and linguistics. This is a common misconception about evolution, however, and as you're not a biologist one is inclined to be a bit forgiving. However, you have claimed to be a linguist—or at least a linguistics student—and you're commenting in a linguistics-oriented forum, so this is really unforgivable. I suspect the only reason that a horde of linguists, including our hosts, aren't eviscerating you is because your error is so egregious, and so arrogant, that most people are embarrassed for you. It brings to mind the exceptionally gentle treatment a commenter received here a week or so ago when she made all sort of odd assertions about Latin and Greek.

    I'm not a linguist, certainly, and only know what I know about linguistics from reading what linguists write popularly, such as this blog, but I'm pretty darn sure that an elementary foundational assumption of modern linguistics is that no language is in any objective sense "better" than any other. That would necessarily include a language at one point in its evolution in comparison to any other. But I'll leave that argument to the linguists here.

    With regard to evolutionary biology, I suppose I might start by mentioning that your statement is literally false—of course there are numerous examples of species becoming less muscularly "strong". But you didn't intend that meaning. You meant "better".

    Perhaps you've confused "more fit" with "better" in an absolute, objective sense because in common language, "more fit" does have this connotation. But with regard to natural selection, it certainly does not. And in modern evolutionary theory, "more fit" is more precisely formulated as "more reproductively fit within its environment"—a formulation that allows numerous deviations from the common conception of something being "better" than something else in the objective sense.

    Related to the misconception that evolution proceeds toward some objectively overall "superior" form, is the misconception that evolution proceeds toward greater complexity. This only decades ago was widely believed among biologists and is still widely commonly believed, including among the scientifically literate.

    One authoritative source that comes to mind who directly debunks these claims is George Williams in the first chapter of his seminal book, "Adaptation and Natural Selection". I suggest you start there.

    Or, you know, Google.

  95. Potomacker said,

    July 24, 2010 @ 3:31 am

    I think that somewhere in this discussion (wow, who knew such conversations took place?) there deserves some mention of the fact that the Chinese writing system is what holds the PRC together. If the Chinese language had developed (been allowed to be represented by) a syllabic or alphabetic system, mainland China would highly likely represent a patchwork of linguistically defined nations like Europe. The once highly regarded and utilized lingua franca of Asia has been in retreat for some time. Linguistically all Asian nations bordering China have borrowed from it, yet only Japanese and South Korean maintain the use of zhongwen. (please correct me if Vietnamese or others still use characters in any capacity.)
    The Japanese or Korean speaker who hesitates when writing a word normally requiring a Chinese character has Hiragana or Hanggeul respectively to allow him to continue writing.I think we can agree that both writing systems benefit from this hybridization.
    The difference in China is that such writing can/will be regarded as a political act. I don't have all the resources nor scholarly abilities to adequately comment, yet I detect a waxing sense of anxiety relating to Han Chinese identity, so heavily invested in Chinese orthography, as the Chinese language and nation attempt to expand its cultural influence around the world. (I cannot find the link but there was an oddly heated debate on the mainland, posted online at least a year back, about a Taiwanese brand name that was a compound of letters and characters!)
    As a further example of this anxiety, we can debate separately the nature of the effects of the English language upon other colonized regions, but it goes without mention that as English speakers spread their language globally, the language itself was modified as a result of the contacts. I suspect that the deeply conservative Han Chinese wonder what consequences are in store as they begin to watch their language studied and spoken outside their control.

  96. ahkow said,

    July 24, 2010 @ 4:06 am

    I tend to side with Rodger C here — romanizing would really have an impact on the ability to comprehend literary Chinese. For example, in literary Chinese 首 head and 手 hand are both pronounced shou (third tone), which could lead to ambiguity. Similarly, the adjectives 不肖 good-for-nothing and 不孝 unfilial are pronounced the same way in modern Mandarin (thus providing a visual minimal pair in the forms 不肖子孙 (mediocre descendants) and 不孝子孙 (unfilial descendants)).

    Also, there exist visual homophonous minimal pairs in surnames: 张, 章 (Zhang) 古, 谷 (Gu) 严, 颜 (Yan), 江, 姜 (Jiang), and at least a triplet 于, 余, 虞 (Yu) — something tells me that folks with these surnames might be upset if the powers-that-be decreed that everything be romanized…

    Third, there is an implicit assumption made by pro-Romanization-ers that everyone speaks something close to standard Mandarin, being able to distinguish between the phonemes represented in pinyin as x, sh, and s (and their affricate counterparts q, ch, c; j, zh, z), or the final nasals n and ng. In many southern Chinese languages speakers do not distinguish these sounds well, leading to, eg., the pun "我愛彰化, 不說髒話" (in pinyin Wo ai Zhanghua (N.B. Changhua, in Taiwan), bu shuo zang-hua, which I will gloss as "I love Zhanghua, NEG say dirty-speech" and translate as "I love Changhua and will not use vulgarities"). The pun works in Mandarin dialects that lack retroflexes.

    In another case, Shanghainese has only one single nasal, the palatal nasal, occurring in syllable-final position. Speakers of Shanghainese are known to have difficulties distinguishing between the alveolar nasal (n) and the velar nasal (ng) found in standard Mandarin (cf. http://202.121.166.17/tpzs/shrzy.htm — in Chinese, written for Shanghainese speakers).

    The reason why I brought these up is to argue that romanization does not necessarily make communication easier for speakers of non-standard varieties of Mandarin, or speakers for whom Mandarin is a second language. Imagine you were from Shanghai, and you are trying to write something (in Mandarin, since that's the official language) that ends with a nasal — but it's confusing! because you don't know whether to type "n" or "ng," because you don't distinguish between these two sounds. Or imagine you are from Hong Kong (Hong Kong Cantonese does not distinguish between sh, s): "Yikes, does it start with 'sh' or 's'?" Isn't this just as arbitrary orthographically as characters are? Arguably, the current character-based system allows speakers to readily share and understand a set of morphemes, without having phonological differences interfere with the orthography.

    Lastly, an issue of switching costs: although standard written (Mandarin) Chinese is rooted in the spoken vernacular, the usage of classical/literary allusions is still prestigious (since it connotes familiarity with literature, culture, etc). So romanization would almost definitely entail getting rid of existing styles of writing, in favor of a style that more closely resembles the spoken language. Not a bad thing in itself, but I suspect the costs might be too high to compel enough people to do so.

    While we are at it, why doesn't someone advocate reforming English orthography? That one's pretty arbitrary, too, the last time I looked…

  97. J.R. Omahen said,

    July 24, 2010 @ 4:38 am

    At the risk of sounding dumb, I would like to mention something small, tiny and inconsequential.

    "@RogerC everybody understood it the first time. The response to your question wasn't posted because the answer is clearly the same as to the concerns that Greeks can't read the Iliad today, Italians can't read, say.. Virgil, and the English can't read Beowulf."

    "What in Gothic is as worth reading as the Shi Jing, the Daodejing, or anything else written in Chinese before the last few centuries?"

    It sounds to me as though direct analogies are being drawn between Modern / Ancient Chinese and other Modern / Ancient civilisations in relation to writing and orthography. To which I would like to have one small point: Old English, Gothic, et cetera, are completely different languages to Modern English and other modern languages in particular Western civilisations. While there certainly is a distinction with Ancient Chinese (to grossly over-simplify the history :p), and it is a different language, the relative continuity of the writing system has smoothed over some of the linguistic change and variation that has happened since then. Thus, it can be (not always, but occasionally) far easier for a modern Chinese-speaker to read the Daodejing than, say, teaching any one of you to read Dream of the Rood in its original MS.

    And yes, languages are not objectively superior or inferior to any other language. They are all unique, and fully capable of expressing ideas, beliefs, moods and thoughts to their speakers as any other language is.

  98. Rodger C said,

    July 24, 2010 @ 10:35 am

    Actually modern Greeks *can* read the Iliad, with some training (not full foreign-language instruction), because the spelling of both Homeric Greek and Modern Greek is assimilated as closely as possible to that of 5th century BC Attic. They pronounce it as modern, like a Chinese reciting Confucius. Plus Greeks who go to church are used to hearing the liturgy in ancient Greek with modern pronunciation.

  99. Steven said,

    July 24, 2010 @ 10:46 am

    I came to this discussion via the Economist language blog. I speak Spanish and have studied Italian and French, but never Chinese and this is fascinating.

    But I have a question, about this passage:

    "If I don't know a character, I take out my cellphone to check."

    Can someone explain, or point me to an explanation, of how one uses a cell phone to look up a Chinese character when one doesn't know it?

    It reminds me of a classmate in fifth grade, who asked a teacher, "But how can I look up a word in the dictionary if I don't know how to spell it?"

    Thanks

    [(myl) At least for the standard Putonghua language, the relationship between pronunciation and pinyin spelling is transparent and not subject to significant ambiguity. Therefore, if you speak Chinese and know how to say a word, you know how to write it in pinyin. Cell-phone texting programs (like input systems for Chinese characters in word-processing programs and so on) usually work by taking in a pinyin spelling and putting out a list of possible Hanzi characters. If you can passively recognize the correct character, then such a cell-phone system will solve the problem.]

  100. stephen said,

    July 24, 2010 @ 11:12 am

    Have you seen this? [(myl) Yes, we have…]

    It seems a lot of people getting Chinese characters for tatoos, don't know what the Chinese characters *really* mean. The tatoo artists either don't know Chinese or are pulling pranks on their customers.

    I wonder, are Chinese people getting English tatoos which convey strange messages? [(myl) Yes, they are….]

    http://blogs.static.mentalfloss.com/blogs/archives/22251.html

  101. Jer said,

    July 24, 2010 @ 11:21 am

    @Steven
    If the speaker knew how to say the word, but not write it, she could look in a phonetic dictionary (http://en.wikipedia.org/wiki/Chinese_dictionary#Phonetically_organized_dictionaries). Have a browse through the rest of the wikipedia page to find out how to look up a character you are reading. It's fairly complicated.

    @RodgerC
    Thanks for the tip, I haven't tackled ancient greek yet. I thought they were more different. And since my point wasn't clear, I was saying that if Chinese is romanised then China will join the list of countries where people can't read their classic texts without much study. Personally I'll feel sad if the whole word switches to a small alphabet, more so if it's roman/english. It's nice to have a bit of variation.

  102. Rodger C said,

    July 24, 2010 @ 1:24 pm

    @Jer: Your point was the same as mine then. Ancient Greek is much more highly inflected than modern Greek, but my impression is that educated Greeks have enough passive familiarity with it to blunder their way through it.

    I studied translation practice, and worked a bit with Sappho, under Willis Barnstone, who had (has) the disconcerting (to me) habit of pronouncing even Sappho's Aeolic as modern Greek. His wife is Greek, and I gather he learned modern Greek first.

  103. Rodger C said,

    July 24, 2010 @ 1:37 pm

    "It brings to mind the exceptionally gentle treatment a commenter received here a week or so ago when she made all sort of odd assertions about Latin and Greek."

    I believe she was from eastern Europe, no? Can anyone here confirm or deny my impression that a lot of very fringey historical notions come from there? I think that people in the former Soviet bloc realized with some shock after 1989-91 that a lot of the history they'd been taught in school had been systematically falsified, creating a fertile field for notions to the effect that our whole received idea of history is bogus.

  104. Rodger C said,

    July 24, 2010 @ 1:50 pm

    "Rendering from Wenyan into Putonghua isn't "transcription" but translation into a different language, with all the pitfalls of translating from Latin to French."

    Finally, let me say that I couldn't write that without thinking of the legendary student who rendered "Timeo Danaos et dona ferentes" as "J'estime les Danois et leurs dents de fer."

  105. Eric said,

    July 25, 2010 @ 11:34 am

    @ahkow:
    FWEIW, many English speakers have to learn the difference between "pen" and "pin" when written, too. They manage.

  106. Nat Hillard said,

    July 25, 2010 @ 5:41 pm

    It is surprising to me that no one has yet mentioned the fact of literacy rates.

    [(myl) That's partly because there's very little good information about what literacy rates really are. The cited numbers in many countries, China and Japan included, seem to be very much politically influenced, and probably very much too high as a measure of the level of ability of the population to read (e.g. newspapers) and to write, especially a few years after various exams have been passed. See here for some discussion of the historical situation in Japan.]

    According to http://en.wikipedia.org/wiki/People's_Republic_of_China#education and http://en.wikipedia.org/wiki/List_of_countries_by_literacy_rate , China's literacy rate in 2007 was (at least reported as, anyhow) 93.3%, putting it at 83rd in the world. China's youth (age 15 to 24) literacy rate was 98.9% (99.2% for males and 98.5% for females) in 2000 — in fact, at least nominally higher than that of the United States ( http://en.wikipedia.org/wiki/Literacy_in_the_United_States )

    Notably, literacy itself is notoriously hard to judge, and in fact there are numerous methodological issues that may significantly lower figures such as the US's 99% or China's 93.3%. A 1992 Department of Education study (The National Assessment of Adult Literacy) noted that in fact only 15% of US adults were in the highest category of literacy (cited above), while nearly 40% were at basic or below basic levels. To quote again, "If this bottom quantile of the study is equated with the functionally illiterate, and these are then removed from those classified as literate, then the resultant literacy rate for the United States would be at most 65-85% depending on where in the basic, minimal competence quantile one sets the cutoff." I would wonder what this "actual" rate would be in China, and perhaps this is harder to tell.

    The difficulties that computer-aided character writing presents are to the already most literate subset of the population, which likely coincides with those with computer access. Claiming, therefore, that this is a problem that affects the overall literacy of the nation is, it would seem, alarmist–for the vast majority of the population, this is not an issue at all, in fact–just as it has not been an issue throughout most of history, where, at times, literacy was as low as 5%. Importantly, with the advent of baihua in the first part of the 20th century, this figure improved, but it is believed that the rate was still only 20% as of 1949. It again improved again following the advent of simplified characters.

    Taking a finer-grained look at what "literacy" itself means ultimately leads to a more modest appraisal of western literacy (perhaps here taken as an implicit counterargument) and, one would hope, a greater appreciation of the tremendous progress that has been made due to infrastructural improvements in Chinese rural education.

    Seeing phonetic alphabets as the populist contender in this rather bizarre race to obtain the greatest "literacy" for the greatest number is therefore a necessarily flawed representation of a deeper reality: a field where "functional" literacy is ultimately of greatest importance–as we have seen, even in a fully industrialized nation such as the united states, as much as 40% of the population is functionally illiterate, if not simply because it is not necessary in the course of daily life to be able to maintain written correspondence on matters of complex intellectual import. In a nation such as China, which is still overwhelmingly rural, it is likewise unnecessary for the majority of the population to be able to write the majority of the language's characters.

    Finally, who is to say that "literacy" itself is not to be defined differently for every culture? Chinese has a long and storied history, and characters still in use today were developed almost 4000 years ago. This is not the case with English, even if one takes particular liberty with the specifics of historical linguistics. The classical Chinese register is of a completely different sort than modern highly-literate English and has a vastly different historical upbringing. Being unable to write the word for 'sneeze' (which to a Western European audience is unconscionable–evidence of a lack of efficiency and expediency) is a feature that exists right alongside the ability to comprehend the geometric significance of a phrase as simple as 人仁也 : benevolence is necessarily a human characteristic.

  107. Victor Mair said,

    July 25, 2010 @ 10:14 pm

    @Nat Hillard

    In reply to your final sentence: Yes, and rén rén rén rén rén rén rén 人人人人人人人, which also has a certain "geometric significance."

    Since I'm hearing and seeing figures of nearly 500,000,000 people using cell phones in the PRC, that's hardly an insignificant subset of the literate population — probably more like the whole of (or greater than) the literate population.

    Finally, in addition to Prof. Liberman's good reason why no one (properly) has raised the issue of literacy, it's because the topic of the thread is about the attrition of the ability to write characters by hand as a result of the use of computers, cell phones, and so forth. There are no reliable literacy figures available that speak to the matter at hand. Hence literacy in the context of the present discussion is a canard.

  108. Jacob said,

    July 25, 2010 @ 11:58 pm

    This post prompted 3 blog posts at the Economist and has now prompted me to actually start a blog.

    I have put up a picture of the keyboard template that I'm using to help me learn Wubi.

    Thinking about Wubi makes me wonder if there is any correlation between the ability to write characters longhand and input method used; are people who use Wubi and cellphones with handwriting recognition better able to write characters than those who use pinyin?

  109. Nat Hillard said,

    July 26, 2010 @ 3:28 pm

    @myl
    Thank you for the link to this other posting, which is interesting, and I believe relevant, in the case of the discussion at hand. It, too, is concerned with cellphone input methods, and their impact on general literacy, especially in historical context.

    @myl + Victor Mair
    As you rightfully point out, these figures are highly politicized. Perhaps I misspoke with reference to "literacy rates" by identifying them as institutionally reported entities, rather than "literacy" in general, which I believe is very much at issue here, and which the post on Kanji discusses as well. Professor Mair, in your original post, you mention the "supposedly literate Chinese" and " 'literate' people". You speak of " hastening reliance on alphabetical access to literacy". Some general insight into how literacy as a figure is calculated in China would certainly be relevant here toward an understanding of the changing face of this very concept. As you again mention, there are no reliable relevant figures–as we all know, the Chinese government is prone to exaggerating numbers.

    But along these lines, can one still be considered literate if they need the aid of a device to reproduce a character but cannot do this by hand? Does reliance on the IME pinyin input method speak to reliance on pinyin itself? People must still recognize and choose from the characters they wish to input. The characters are in many ways a necessary intermediary. Being unable to reproduce them by hand is a different matter from being unable to pick them from a line up. Additionally, the number of characters that are auto-completed based on context-sensitive pinyin data varies by cellphone and IME method–certain systems are "smarter" than others. Even with the most advanced systems, simply typing pinyin and letting the system do the work assuredly results in gobbledygook: 这个橘子没有仁棒他写 ("this sentence has no one to help it be written" is rendered "this orange has no benevolence stick he write"–admittedly based on characters i have recently input). Is one less "literate" if he/she is able to understand and reproduce a character but not physically write it? Taking this one step further, and making the former question obsolete, If one were to take automated methods out of this entirely, would it really be possible to communicate in only pinyin? This gets into issues, such as the vast differences between spoken and written Chinese–especially at formal registers.

    I am curious to see your followup post, which, if I read your tone correctly, will perhaps advocate the eradication of characters entirely, and their replacement with pinyin or another romanization method. I should hope this is not your position, but I am curious to see how this can be argued.

  110. Victor Mair said,

    July 26, 2010 @ 6:37 pm

    @Nat Hillard: "I am curious to see your followup post, which, if I read your tone correctly, will perhaps advocate the eradication of characters entirely, and their replacement with pinyin or another romanization method."

    No such thing! I'll just ask the question and present some realities, then let people decide for themselves what they think.

    To satisfy your curiosity about how Chinese script reformers have been advocating for more than a century in favor of romanization or some other phonetic script, you'll have to read the relevant historical studies. The best place to start would be with John DeFrancis' Nationalism and Language Reform in China (1951; probably still in print somewhere), though there are many other works to which I could refer you if you want to pursue this topic beyond that.

    Meanwhile, I want to point out to a few commenters who mentioned that Pinyin doesn't work with Literary Sinitic (Classical Chinese) that it was never intended for that purpose. That is the realm of the characters and always will be; HANZI and WENYAN go hand in glove. Pinyin is only suitable for what the great linguist Y. R. Chao called "Sayable Chinese."

  111. Victor Mair said,

    July 26, 2010 @ 7:12 pm

    @Jacob

    Interesting project you've got started. If you tell us the URL for your blog, I'll follow it eagerly, especially if you keep a record of how much time you spend on acquiring competence in Wubi, whether it interferes with your thinking when you try to touch-type with it, and so forth.

    The vast majority of people who use Wubi are secretarial professionals. Wubi is not for the casual, avocational typist, nor for someone who wants to compose literary or scholarly texts on a keyboard. I only know one Chinese professor who is (sort of) able to use Wubi, but he often has to resort to Pinyin when he can't get the Wubi to produce the characters he's thinking of.

    I am not aware of any studies, formal or otherwise, on the impact of shape-based systems such as Wubi and Cangjie upon the ability of individuals to produce characters by hand. One thing about most of these systems that makes them so difficult to learn, and why they are not meant for casual nonprofessionals, is that they decompose characters (or, conversely, build them up) differently than one does when writing characters by hand. In other words, there is a lot of arbitrariness about the way the characters are constructed in most shape-based systems, and this puts a heavy burden on their users to remember the special sequences of key strokes that are required. There are some systems that do attempt to mimic, one-for-one, the actual sequence of strokes that one uses when writing characters by hand, but these have not caught on. I suspect that this is because: 1. the user would have to be very exacting in following just one possible, "correct" way of writing a given character, whereas there is a lot of wiggle room among different individuals about how to write the characters, 2. it would be very slow and inefficient to enter 10 or more strokes just to call up one character / syllable, which would happen fairly often in the course of composing a text, etc.

    Pinyin entry is more popular than shape-based entry systems because: 1. it is a lot easier, 2. it is tied to speech, 3. doesn't require much of an initial learning period, and so forth (these are just a few of the reasons people have given to me when I ask them why they use Pinyin).

  112. Richard said,

    July 28, 2010 @ 3:22 pm

    Tom:

    Wenyan is much, much, much more comprehendable to a Chinese college graduate than Chauncer is to an American college graduate.

    This is because phrases in wenyan are still commonly used in the written language (especially in Taiwan & HK, but also southern parts of the mainland) as well as the non-Mandarin spoken languages in the south. In fact, you could make the case that Cantonese is more similiar to wenyan than it is to Mandarin.

    This is a point that all you non-Chinese who are against characters forget: if you replace characters with pinyin, you're essentially imposing a different language on southerners. The best example is if France, Spain, Italy, & Portugal were still one country/empire, and they'd been using characters for writing for thousands of years (though early in the 20th century, they started writing French using characters rather than Latin, but the grammar is similiar enough–imagine in this example that all the Romance languages had grammar that was as similiar as the Chinese ones). What you are proposing is that everyone stop using their characters (which the Sicilians would still consider Sicilian, the Catalans Catalan, etc.) and start using Parisian French in the Roman alphabet as the written language.

    You may understand why some people may be against that (and how it may actually hurt literacy rates).

  113. Cooper said,

    July 29, 2010 @ 12:41 am

    Victor, in a (much) earlier post about the Guoyu Ribao you mentioned how "slowly, with practice, I also became capable of writing in characters", after much reading. I'm curious, how do you maintain your ability to write in Chinese now that (I'm guessing) you spend much more time reading and typing than handwriting? Or do you face the same predicament as Zhu Linfei etc?

  114. Victor Mair said,

    July 29, 2010 @ 6:14 am

    @Richard: "Wenyan is much, much, much more comprehendable [sic] to a Chinese college graduate than Chauncer [sic] is to an American college graduate."

    As someone who wrote an undergraduate thesis on Chaucer and who has taught Classical Chinese / Literary Sinitic to college and graduate students from around the world (including from the PRC, Taiwan, Hong Kong, Singapore, Korea, and Japan) for more than thirty years, I can assure you that this extremely difficult, dead language (with many very different styles of its own) is definitely not "much, much, much more" comprehensible / comprehendible for Chinese college graduates than Chaucer is for an American college graduate.

    Maybe if you had said Beowulf, the comparison would have been more apt.

    Before I advise students who claim they can already read Classical Chinese whether to take my introduction to Classical Chinese course or to go into second-year Classical, I ask them to read a couple of very simple, easy passages from different periods. Often they can read most (but by no means all!) of the characters, just as American students — even from junior high or high school — can read out Chaucer or Beowulf. But then, when I start to ask them what the passages mean, about the grammar and the syntax, it turns out that they don't really understand at all.

  115. Victor Mair said,

    July 29, 2010 @ 7:52 am

    @Cooper

    I still read materials from the Guoyu Ribao publishing house, and also assiduously seek out intellectually stimulating texts with Pinyin annotation.

    It's only very recently that I have learned to produce characters on a computer (mainly for the purpose of writing Language Log posts). Just like my friend Xu Wenkan, I have resisted writing characters with a computer, and I shall probably never acquire a cell phone, much less use one to write STMs, whether in English or in Chinese. I still write Chinese letters and exams, etc. by hand. When I need to send a message in Chinese to friends quickly, I often use Pinyin, although lately I've also begun to type characters via Pinyin. Nonetheless, I make sure to spend time every week writing something in characters by hand. If I didn't, I would very soon lose the active ability to produce characters by hand and would end up, at best, being able to recognize them passively.

  116. Richard said,

    July 29, 2010 @ 6:45 pm

    Victor:

    Chaucer's pretty damn difficult.

    I know that I can't really follow what's going on in the Canterbury Tales (even though I can identify certain words), and actually find modern Dutch to be easier to understand (maybe the 2 years of German I took in HS help?) BTW, English is the only language I'm literate in, having been in the US since 2 (though Mandarin & Ningbo-ehwo were my first spoken languages).

    On the other hand, even while recognizing only a few hundred characters, I can make out passages in the Shijing. For that matter, Tang poems (which I believe Chinese college grads, at least of my parents generation, can understand) seem closer to current Chinese than the "Tales" to modern English.

    Granted, I know the Shijing isn't "dense" wenyanwen (being essentially the common speech of that time, unlike the wenyanwen written by later officials who wrote in a denser style). Still, my parents can read wenyanwen. Maybe I'm biased because its so easy for them. Maybe its because they were history majors. Maybe its because, last generation, wenyanwen was taught more in both HS and college classes (in Taiwan). I know that, even though my dad first went to school after the Communists had already taken over, he was learning/reciting wenyan texts (in Ningbo dialect) in grade school.

    BTW, Sir Gawain & the Green Knight is just a shade more comprehensible to me than Beowulf (but not enough for me to make out what the hell is going on). This points out a strong point about characters: If "Sir Gawain" was written using characters, I'd probably be able to follow the story with the aid of a dictionary (since most of the words would be ones I recognized, even if a lot of them are antiquated or regional now), but as it is, it's too heavy a slog as every other word I encounter isn't spelt like something in Modern English. Imagine if the Tang poems were written by pronunciation rather than by characters. No one would read them (unless they're translated in to each modern Chinese regionalect) because not even the Cantonese & Hakka would be able to make out what the heck was written down. Even if translated, though, I imagine few people would find them beautiful (just as few people these days find translated versions of the "Canterbury Tales" or "Beowulf" beautiful).

  117. Jonathan Tang said,

    July 31, 2010 @ 12:57 am

    Since Professor Mair brought up Y. R. Chao's name, but not his poem "Lion-Eating Poet in the Stone Den," I figured I might as well.

    The Wikipedia article (http://en.wikipedia.org/wiki/Lion-Eating_Poet_in_the_Stone_Den) contains a detailed summary. For those lacking the time to read it, Chao wrote the poem in classical Chinese, using the same sound *shi* in Mandarin Chinese, but with different tones. In other words, were the whole of it romanized, per the urgings of many of the commenters, then the text would be completely incomprehensible.

    Put another way, the poem is Chao's argument that characters could never be entirely divorced from the Chinese language. Chao being the creator of the Gwoyeu Romatzyh system of romanizing Mandarin, I find his claim to be quite a convincing one.

    Using characters, the text is as follows (all taken from the above linked wikipedia page):

    《施氏食獅史》

    石室詩士施氏,嗜獅,誓食十獅。
    氏時時適市視獅。
    十時,適十獅適市。
    是時,適施氏適市。
    氏視是十獅,恃矢勢,使是十獅逝世。
    氏拾是十獅屍,適石室。
    石室濕,氏使侍拭石室。
    石室拭,氏始試食是十獅。
    食時,始識是十獅,實十石獅屍。
    試釋是事。

    In English:
    « Lion-Eating Poet in the Stone Den »

    In a stone den was a poet Shi, who was a lion addict, and had resolved to eat ten.
    He often went to the market to look for lions.
    At ten o'clock, ten lions had just arrived at the market.
    At that time, Shi had just arrived at the market.
    He saw those ten lions, and using his trusty arrows, caused the ten lions to die.
    He brought the corpses of the ten lions to the stone den.
    The stone den was damp. He asked his servants to wipe it.
    After the stone den was wiped, he tried to eat those ten lions.
    When he ate, he realized that these ten lions were in fact ten stone lion corpses.
    Try to explain this matter.

    Using pinyin romanization:

    « Shī Shì shí shī shǐ »

    Shíshì shīshì Shī Shì, shì shī, shì shí shí shī.
    Shì shíshí shì shì shì shī.
    Shí shí, shì shí shī shì shì.
    Shì shí, shì Shī Shì shì shì.
    Shì shì shì shí shī, shì shǐ shì, shǐ shì shí shī shìshì.
    Shì shí shì shí shī shī, shì shíshì.
    Shíshì shī, Shì shǐ shì shì shíshì.
    Shíshì shì, Shì shǐ shì shí shì shí shī.
    Shí shí, shǐ shí shì shí shī, shí shí shí shī shī.
    Shì shì shì shì.

  118. Viktor Brech said,

    July 31, 2010 @ 10:50 am

    Dear Viktor Mair,

    is there any evidence that these people affected by "amnesia" also lose their ability to recognize characters?

    As long as passive knowledge persists, this "amnesia" seems to me to be unimportant, given that handwriting is increasingly irrelevant and most people are familiar with Mandarin pronunciation (and hence able to use pinyin input even when producing a text meant to be read in a different dialect).

    I am very much looking forward to your follow-up post. Thank you for your work.

    Best regards,
    Viktor Brech

  119. Viktor Brech said,

    July 31, 2010 @ 10:51 am

    Dear Mr. Mair, sorry I misspelled your name in my comment above. VB

  120. Victor Mair said,

    August 2, 2010 @ 5:52 pm

    @Richard

    If those texts you mention (Shi jing, Tang poetry, etc.) were so easy, especially for people like you who only know a few hundred characters, how come they're all being *translated* into Mandarin? All of the Chinese classics and major histories are being translated into Mandarin (many of them repeatedly) for people who are literate in Mandarin but who are not specialists in Classical Chinese. It's like translating Classical Greek or Latin texts into Modern Greek, Italian, or English.

    @Johnathan Tang

    Y R Chao's (in)famous "shi shi shi…" essay has been brought up many times before on Language Log and elsewhere, but all too often by people who misunderstand why he wrote it. He was not trying to prove that romanization is impossible. Quite the opposite, the point he was making is that romanization only works for the vernacular, i.e., for texts that are "sayable" (his term). I will discuss this a bit in my forthcoming blog on romanization.

    @Viktor Brech

    For those who constantly resort to Pinyin to compose texts, the ability to recognize characters also diminishes, but at a much lower rate than the ability to produce characters by hand. I do not know of any formal studies comparing the two rates of attrition, so my comments here are based on my own observation of individuals who complain that they are not only forgetting how to write the characters, but — to a lesser extent — also how to read the characters. This is why some scholars are predicting that "Pinyin in, characters out" will gradually morph into "Pinyin in, Pinyin out." I will also touch upon this briefly in my forthcoming blog on romanization.

  121. Richard said,

    August 3, 2010 @ 12:12 pm

    The more I think about it, the more a mixed script (like the Japanese use) makes sense. In all of the Japonic & Sinitic languages that I know of, there are a lot of homonyms, and a pure alphabetic script will lead to many difficulties:
    1. People will have to severely change the way they write & read; the simple fact that people can speak to each other in a language doesn't mean writing the way people speak is ideal; as many other people have pointed out, when conversing, communication is a 2-way street, and homonyms/words based on context can be cleared up easily; when reading, there's no way to ask the author what exactly they meant to say if there is confusion.
    2. Even then, you introduce many possible new ways to misinterpret the writing (deliberately or not), which makes it likely that characters will still been needed in the legal sphere/writing laws or anywhere else where clarity (or conciseness) is required.
    3. People who currently don't know written Mandarin will have to learn written Mandarin. You could argue that they read written Mandarin now, but with characters & the fact that the Sinitic languages don't differ a lot in grammar, written Mandarin in characters is readable even to Cantonese speakers who don't know Mandarin. Written Mandarin with an alphabet definitely would not be. In fact, the difference within registers of Mandarin (between, say, colloquial Mandarin spoken on the streets of Taiwan and written Mandarin on a newspaper in Taiwan) can be greater than the differences between different Sinitic languages (between spoken Mandarin & spoken Wu, for instance) when it's all expressed in characters. As an example of the beauty of characters, even though the Japanese don't fully use kanji to express everything and the grammar is wildly different from any form of Chinese, my parents can still make out the gist of 90% of a Japanese newspaper.

    BTW, I'd recommend a modified hangul rather than Roman characters in the mixed script. It'd fit in better, and allows writing to be vertical as well as horizontal. Modify the hangul enough, and you could fit in virtually all regionalects as well.

  122. Ellie said,

    August 19, 2010 @ 10:52 pm

    Prof. Mair: I think the suspicion of anecdotal evidence has to do with whether cell phones and computers actually cause the loss of characters, and whether the loss has increased since before computer input. It seems pretty clear that post-schooling character attrition does occur among educated Chinese, and in fact the news articles refer to studies which quantize that to some extent.
    Unfortunately, the articles don't provide full citations. Do you know where one could find those studies, in any language? Is there any other formal work on this topic at all? In particular, historical studies of, say, how many characters college graduates in the 1950's or 1970's still knew a decade out of school would be very useful for comparison.
    Thanks!

  123. Victor Mair said,

    August 26, 2010 @ 10:59 am

    @Ellie

    J. Marshall Unger, in some of his works (perhaps in The Fifth Generation Fallacy) refers to the sort of studies that you are seeking. I remember one in particular that followed character loss in Japanese army recruits. William Hannas, either in Asia's Orthographic Dilemma or in The Writing on the Wall, cites similar studies for China. However, the radical type of character amnesia that results from reliance on electronic devices is of an entirely different order than the gradual attrition you describe. It is because the amount of characters being forgotten is now so large and it happens so quickly that people are bemoaning their dwindling ability to produce characters far more than they did 15 or 20 years ago.

  124. Alex said,

    August 27, 2010 @ 1:25 am

    A few readers are making the following argument (taken from one of Daniel von Brighoff's comments above):

    Can you really conceive of a literate, educated, native speaker of English who not only can't spell "causeway" or "recipe" correctly but can't even take a stab at it because they literally can't remember a single one of the letters that make it up?

    However, I think the intended point is moot. Chinese and English orthographies are not comparable in the way Brighoff, for example, is making them out to be. For example, the word strength is written in English the way it sounds (barring dialectal differences)–the letter "s" directly signifies its phoneme, as does "t," and so on. However, in Chinese, the character for strength–"力"–does not directly signify the 3 phonemes used to pronounce the character (I only know Cantonese, so the pronunciation would be "lik6" in the Wade-Giles romanization), for if it did, then one might find it somewhere in the character for experience–"歷"–, which has the exact same pronunciation (in Cantonese). To complicate matters, when two characters are combined into a new one, the composite character may not have a pronunciation that obviously follows from the combination. For example, when the characters 重 (jung6/chung4/chung5 — meaning: serious or to repeat) and 力 (lik6) are combined to make 動 (meaning: to act or to move), the pronunciation of the latter is "dung6." There may well be some order of logic that states that 動 necessarily has the pronunciation it does, but that logic is not as transparent as the logic that states that the letter "s" is necessarily pronounced with, well, an s-sound.

    Furthermore, although there are, as Victor Mair states, around 850 phonetic components to the characters, that number is a far cry from the 26 phonetic symbols used in English orthography. As a result, even if one were to try to rely on the 850 or so phonetic components of the characters to remember pronunciation or to remember how to write the characters from their pronunciation (assuming such a reliance is even possible), one would still have more difficulty remembering the pronunciation of those components in comparison to the 26 letters of the English alphabet.

  125. dhd said,

    September 3, 2010 @ 12:50 am

    This adorable cartoon always comes to mind when I think of this:

    http://www.wretch.cc/blog/nellydyu/7629539

    Maybe somebody in the comments talked about this and I missed it, but have there been any studies about the regions of the brain involved in reading versus writing characters, and in writing characters versus writing in alphabetic/syllabic scripts?

    The asymmetry between perception and production for characters is pretty striking. I know that after a year without Mandarin classes I've forgotten how to write all sorts of ridiculously commonplace ones, while at the same time the number of characters I can read has continued to increase just through reading Chinese text and browsing through the dictionary on my iPhone, whose handwriting recognition makes it really easy to look things up that I don't recognize (I was really disappointed to find that Android phones don't have this feature, although I think maybe the HTC-branded ones do)

  126. Wentao said,

    April 29, 2011 @ 1:34 pm

    @ Victor Mair: "If those texts you mention (Shi jing, Tang poetry, etc.) were so easy, especially for people like you who only know a few hundred characters, how come they're all being *translated* into Mandarin?"

    I'm a native Chinese, and I believe myself to be fairly literate, at least literate enough to be able to write 喷嚏 correctly by hand; I do admit that 诗经 is extremely difficult, full of archaic and obscure characters, and that good translations, or good footnotes, are always very helpful; however, I've never seen any Tang poetry completely translated into modern Mandarin, except in textbooks for primary schools. Most of them only contains a couple of notes explaining cultural background and difficult phrases.

  127. bryan said,

    May 7, 2011 @ 7:19 pm

    ""sneeze" (pēntì 噴嚏, simplified 噴嚏 [same])"
    NOT! It is not the same in Simplified & Traditional Chinese:

    喷嚏 [Simplified] is not the same as 噴嚏[Traditional]!
    喷 has 贝
    while
    噴 has 貝

  128. bryan said,

    May 7, 2011 @ 7:56 pm

    @ Victor:

    As someone who wrote an undergraduate thesis on Chaucer and who has taught Classical Chinese / Literary Sinitic to college and graduate students from around the world (including from the PRC, Taiwan, Hong Kong, Singapore, Korea, and Japan) for more than thirty years, I can assure you that this extremely difficult, dead language (with many very different styles of its own) is definitely not "much, much, much more" comprehensible / comprehendible for Chinese college graduates than Chaucer is for an American college graduate.

    Literary Sinitic? [Do you not speak English? SInitic is an adjective. An adjective describe things, but what literary things is there to describe? ] Classical Chinese? You people pull a whole lot of stuff out of thin air. There's no such thing as "Literary Sinitic"! Sino- is Latin for "Chinese" supposedly, but who knows, it's a dead language. -itic is of Greek origin and I doubt it has anything to do with the Chinese language. "Sinitic" is synthetic. "Classical Chinese" is the Sinologists' way of saying "the spoken form of Ancient Chinese" which is what Wenyenwan is. Literary Chinese = Wenyanwen = Ancient Chinese.

    Foreigners most often than NOT would look at a more comprehensible version of a translation [in English, etc…] form of work of the Chinese classics, then translate that into their home language, whichever that may be. Then they get the English translation, then they decipher each character by meaning, composition, etc… via the dictionaries available. Then they try to transcribe Wenyanwen back into Mandarin, due to most foreigners learning Mandarin first. But having more than one translation might steer away from the original, and create something new but untrue.

    BUT… BUT most of the poems in Chinese were written using Cantonese, Shanghainese, Fujianese, or other southern Chinese dialects, & NOT Mandarin, meaning that learners of Mandarin would NOT be able to understand what's written UNLESS the meaning is in the footnotes.

    And sometimes, foreigners would argue with Chinese people about such and such and Chinese people who have learned written Chinese [Traditional form… Simplified is just simplifying the strokes according to variant or slang or alternative forms, which might not have a cursive form and turned into a cursive form in Simplified], spoken Chinese, in any dialect except Mandarin. Mandarin, as some of you don't even know, after learning Mandarin for so long, is that it is a reduction of the sounds of Classical Chinese, whether it's through Middle Chinese or Wenyanwen. Actually what most foreigners assume is incorrect. Middle Chinese didn't start from the Tang dynasty, but rather the Sui dynasty. I won't argue, but I dislike those who call themselves "Sinologists" whom tell tales worst than a tall tale.

  129. bryan said,

    May 7, 2011 @ 8:04 pm

    "850 phonetic components to the characters"?

    What an assumption! How many Chinese characters are there in current use? They don't even know the answer and tell you that there's 850 phonetic components? Mandarin DOES NOT have that many!!! Are you including Korean Hanja, which actually uses the same Traditional script as Chinese people with NO simplifications & Japanese Kanji, with both Japanified simplifications and Traditional Chinese script forms, to this list also? There's about 2-5 ways to pronounce a Hanja, while a Kanji could have up to ten pronunciations in Japanese! How about Vietnamese, which used to but no longer use the Chinese characters? Are you including Chu Nom & Chu Nho plus other possible yet unknown variants that's not even talked about?

  130. bryan said,

    May 7, 2011 @ 8:22 pm

    "the vast majority of allegedly literate Chinese":

    "allegedly", due to them NOT being literate in the first place. Simplification of Chinese is for illiterates, at least that's how Mao looked at it. Mao certainly didn't care about the people: While he read and wrote in Traditional Chinese BEFORE demanding that Simplification be used, should be ashamed. Simplification makes no sense and has no link to the original meaning of the characters, due to absurd changes. People, as they age, tend to forget and return to that of a baby: an old person forgets, and keeps asking for food, while they have already eatendue to some condition such as Alzheimer's, etc… or they forget simply due to old age. So, they write in simpler steps with fewer strokes, like a young child learning to write the characters, without learning what it means yet! That's why people don't know how to write in Chinese anymore. And pinyin makes it even worse!!!

  131. bryan said,

    May 7, 2011 @ 8:41 pm

    My grandfather is in his 80's. He knows both scripts and can still write in Traditional Chinese! Mao was heading in the wrong direction with the Civil war, Communism, and the absurd changes to the writing system.

    Example:
    [Simplified Chinese] – 皇后后面有个姓黄的后人在吃面 [say this in Mandarin and try to reproduce the writing, and try to remember which character's which, and try to explain to someone who's not Chinese in English what it means, Does the translation take more or less time? : 後 has been substituted by 后, 麪 has been substituted by 面.]

    [Traditional Chinese] – 皇后後面有個姓黃的後人在食麪 [say this in Cantonese: In Cantonese, the more difficult situation: 皇 & 黄 are the same pronunciation. Cantonese uses 食 for "to eat". Now try to do the same exercise as above. Which takes more time for translation? ]

  132. Bathrobe said,

    May 19, 2011 @ 12:17 am

    After all that, I'd just like to remind people that the problem is not 噴, it's 嚏. 噴 is a common enough character that means 'spurt, spout, gush, or jet'. 嚏, on the other hand, is hard to write, even if you can find a model to work from, unless you look really closely at the character.

  133. Michael Robinson said,

    August 17, 2011 @ 9:21 pm

    Fourpoints:
    1) Do Taiwanese students using traditional characters perform worse in active knowledge of characters? This seems likely, if complexity and number of strokes is a problem.
    2) Why isn't there more use of voice recognition systems? I realize homophones and tones would create problems, but surely there are solutions.
    3) Are there benefits to the difficulties of written Chinese? More of the brain is devoted to visual matters when reading Chinese than in reading European languages, according to some research. Does this pay off in increased visual sophistication/memory for Chinese learners? The uses of adversity are sweet.
    4) Scientific papers are all written in English. Status of a language determines use, growth, strength of a language.
    5) Chinese literacy rates are increasing. Sales of books, papers is increasing – at least that's my understanding.

  134. Hill Gates said,

    December 13, 2012 @ 5:37 pm

    Nothing better could happen to Chinese society than a switch to an alphabetic or even a syllabic writing system. If I want to read Aeschylus in the original, I can study Attic Greek; if a Chinese wants to read Discourses on Salt and Iron, s/he can delve into characters. Ditto for beautiful calligraphy. No one should be forced to waste childhood years on a Bronze Age writing system. Nor should ordinary adults be half-blind to travel directions, newspapers, and information on food packets or medicine bottles. Keeping the "broad masses" essentially illiterate is the oldest trick in the long, repetitious story of Chinese hierarchy.

  135. Victor Mair said,

    December 19, 2013 @ 8:53 pm

    @bryan

    "Sinitic" is both an adjective and a noun:

    http://www.thefreedictionary.com/Sinitic

  136. Penti said,

    October 6, 2014 @ 1:35 am

    I don't expect anyone to read this comment 4 years after the fact, but I found this all very interesting and thought I'd add some random and uninformed thoughts.

    The choice of 喷嚏 as a litmus test is perhaps a bit arbitrary. Though sneezing is indeed a common activity and "sneeze" is easy to write in English, there isn't always a correlation between ubiquity and Hanzi-ease-of-writing, or between English-ease-of-writing and Hanzi-ease-of-writing. It's rarer to speak of uranium than of oranges, for instance, but I think 铀 is easier to remember than 橙. And from the other side, a literate Chinese person might say, "I see English speakers misspell 'necessary' all the time, whereas I can always remember '必须,'" or "English speakers get confused by 'plenary session,' but can you imagine not being able to write 全会?" This is not really a substantial criticism of your main points, and I myself look forward to a time when Pinyin/BPMF/some other system more systematically supplements Hanzi; it's just 嚏 is such a one-off character.

    Also, on a tangential note, perhaps a small part of the problem is the standardization of Hanzi in the modern age. Many of the entries in the Kangxi dictionary are those odd, handwritten variants that people probably improvised on the spot when they couldn't remember the standard character. So, in ye olde days, I figure that if I couldn't remember the "嚏" in "喷嚏" when I was writing a handwritten letter to someone, I could probably get away with writing "噴 '口提 [as one character]'" or "噴'口替'" or something and assume that my reader could at least guess what I meant (similarly to how if I write "nessessery" you can guess what I mean). Now that characters are encoded and it is more difficult to make up new ones, if you can't write 嚏, then you just can't write 嚏.

  137. Victor Mair said,

    October 6, 2014 @ 5:02 pm

    Seriously, Penti, what prompted you to write your comment at this particular time, more than "4 years after the fact", as you say? I'm really curious about the synchronicity of your comment, since I myself mentioned this post to a colleague just a few hours before you posted your comment, and I hadn't mentioned that post in an e-mail to anyone else for a long time.

    I'm also amazed that you chose the name "penti" as your web moniker, as though this word pēntì 噴嚏 were a preoccupation of yours.

    Anyway, there's a difference between not being able to write pēntì 噴嚏 and not being able to write "necessary", so it's not simply "arbitrary" or "one-off" to mention pēntì 噴嚏, nor is it a "litmus test". People misspell "necessary", "accommodate", "misspell", etc. yet still get their message across. But if you can't follow through with all of the strokes of pēntì 噴嚏, you're stuck. You'll either simply not tell your correspondent that you've been sneezing a lot lately or use some awkward, convoluted expression to get the idea across in a watered-down fashion.

    The reason pēntì 噴嚏 has become such a cause célèbre in writing reform circles is the irony / disparity between the quotidian nature of the semi-autonomous reflex and the arcane characters chosen to write the simple word pēntì.

    As for "necessary", I have seen people hesitate over whether to write bìxū 必须 or bìxū 必需 and get stymied by one or the other second character of both locutions, to the point that they end up writing bìyào 必要, which is not what they wanted to write in the first place.

    Moreover, in our numerous posts and comments on Language Log and elsewhere, David Moser and I have cited many other surprisingly common terms in Sinitic languages, including MSM, that people have a hard time writing, e.g., the word for "shrimp" mentioned in the original post above and also in this classic article by John DeFrancis, "The Prospects for Chinese Writing Reform" (see Exhibit 2 near the end), which also reveals that a PRC researcher from the Chinese Academy of Social Sciences with a Ph.D. had to give up when she couldn't write the second syllable of the word for "egg" and the first syllable of the word for "chives". In all three instances she scribbled over the botched attempts to write these characters and replaced them with pinyin! The researcher could not write three out of the eight items on this shopping list, and these are all very common ingredients for making jiǎozi 饺子 ("dumplings")

    http://www.pinyin.info/readings/defrancis/chinese_writing_reform.html

    You might also be interested in these other LLog posts on character amnesia.

    "Character amnesia revisited" (12/13/12)

    http://languagelog.ldc.upenn.edu/nll/?p=4367

    "Spelling bees and character amnesia" (8/7/13)

    http://languagelog.ldc.upenn.edu/nll/?p=5669

    "Character amnesia and the emergence of digraphia" (9/25/13)

    http://languagelog.ldc.upenn.edu/nll/?p=7142

    BTW, I think that there's more than a grain of truth to what you say (in the last paragraph of your comment) about "odd" characters in "ye olde days" and the effects of standardization in modern times. After all, in Shakespeare's time and in succeeding centuries, there were many different ways to write his name, but now there's only one.

    http://en.wikipedia.org/wiki/Spelling_of_Shakespeare%27s_name

    http://shakespeareauthorship.com/name1.html

  138. David Moser said,

    October 6, 2014 @ 8:39 pm

    I admire Victor's tireless exposition of these issues. Character amnesia is NOT the same as forgetting how to spell a common word. Misspelling a word is not the same as forgetting how to write it entirely, which is one of the biggest problems with character amnesia. Having said that, I might add that very often Chinese people are able to come up with a rough approximation or a mis-remembering of a character, which might SEEM to be similar to a misspelling in English, but the results are still different from the English phenomenon. The other day my Chinese wife was making a list of duties for me and wanted to write hui4kuan3 汇款, "wire transfer". For the kuan 款 character she got as far as the component 士 and then got a little confused. Rather than flounder around for the rest of it, she simply wrote "bank wire" in English. For reasons that should be obvious, Chinese script is not really conducive to "making a stab at it." If you forget, you forget. Misspellings never seriously hamper the communication process. Character amnesia usually puts a halt to it. Which is why my wife just gave up and switched to English. (In a more ideal world, she might just switch to pinyin, which would be a step in the right direction.)

  139. Conal Boyce said,

    October 7, 2014 @ 12:32 pm

    In connection with a pair of banners seen recently in the news about the demonstrations in Hong Kong, there is a kind of thought experiment we might try, regarding the following question covered only indirectly (?) in the thread so far:
    Are there situations where a different semantic payload, so to speak, is delivered by characters versus pinyin?

    The pair of 4-character banners looked like this:
    公民抗命 佔領中環
    Somewhere in the news, I saw their text translated as follows:
    Civil Disobedience. Occupy Central!

    Two points of reference:
    Google Translate has the following for 'civil disobedience': 公民不服從
    A 1968 dictionary I happen to have at hand shows this for 'civil disobedience': 不合作主義
    I.e., a reasonable translation of the word 'disobedient' or 'disobedience' is 不服從 , certainly not 抗命. And yet, in context, the journalist's translation doesn't look completely off-the-wall; so what really is going on with that first banner?

    To me, all other things being equal, 抗命 feels much closer to 'defy orders [from the capital]' than to 'be disobedient'. But of course, "all other things" are NOT equal in this case, which has a rich context: I have a hunch that many of those HK protesters are well informed about the history of democracy and the ENGLISH term 'civil disobedience', so their banner very likely has a western subtext. How much of a western subtext? That's the question. And finally, coming to my point: If the banners were written using pinyin, would they have exactly the same kind and amount of subtext, the same 'semantic payload'? It seems clear to me that they would not. If the first banner were written in pinyin, I believe the reader would be nudged sooner and with no complication in the direction of the western subtext idea; as it is written, using characters, I find myself aware of the western subtext, yes, but also unable to ignore the many centuries of Chinese-historical reverberation in 抗命 which is triggered visually, in ways with which 'kangming' could not possibly hope to compete. Not that one is necessarily 'good' and the other 'bad' — I'm just pointing out that there exists a real difference. It is not a simple case of 'the same thing using different orthography'. (Note re simplified vs fantizi: The four characters of the first banner are ambiguous on this point, but the 1st, 2nd and 4th characters of the second banner are clearly fantizi. This too adds a nuance, causing the impact of the first banner, when read together with the second banner, to be located a tad further toward the Chinese-historical end of the spectrum.)

  140. Victor Mair said,

    October 7, 2014 @ 9:17 pm

    @Conal Boyce

    You've taken us off in a different direction from the original intent of the post (a number of the other commenters have also done so). In the original post, I did not want to talk about pinyin replacing characters or the differential semantic weight of characters vs. pinyin (of course, characters, being morphosyllabic, have a much higher semantic weight than pinyin, whereas pinyin has a much greater redundancy than characters — speaking in cybernetic terms). I really just wanted to talk about how people are forgetting how to write characters at an alarming rate, and I wanted to point out how pinyin input systems are a major cause of this phenomenon.

    If writing in pinyin ever becomes widespread (I don't think anyone will try to bring it about all of a sudden, as Ataturk did with the Roman alphabet in Turkey (although Mao — as John DeFrancis has documented — was close to doing so); rather, it will come about gradually through digraphia [to a certain extent, this is already happening, as we have demonstrated in various Language Log posts]), it will certainly not be a mere transcription of writing in characters. Instead, it will develop styles and modes of expression that are suitable to the script, just as characters have done so.

RSS feed for comments on this post