The future of Chinese language learning is now

« previous post | next post »

When I began learning Mandarin nearly half a century ago, I knew exactly how I wanted to acquire proficiency in the language.  Nobody had to tell me how to do this; I knew it instinctively.  The main features of my desired regimen would be to:

1. pay little or no attention to memorizing characters (I would have been content with actively mastering 25 or so very high frequency characters and passively recognizing at most a hundred or so high frequency characters during the first year)

2. focus on pronunciation, vocabulary, grammar, particles, morphology, syntax, idioms, patterns, constructions, sentence structure, rhythm, prosody, and so forth — real language, not the script

3. read massive amounts of texts in Romanization and, if possible later on (after about half a year when I had the basics of the language nailed down), in character texts that would be phonetically annotated

The problem was that all of my language teachers insisted that I memorize hundreds of characters right from the very start.  I rebelled against their demands as much as I could, but had to give in to the extent necessary to get a good grade in their courses.  The first year of learning Mandarin was pure torture in the classroom, though I managed to preserve my sanity by reading old missionary manuals for learning Mandarin that provided romanization for all sentences.  In particular, I was enormously grateful for the splendid work of the Belgian priest, Jozef Mullie, whose magnificent The structural principles of the Chinese language: an introduction to the spoken language, Northern Pekingese dialect was like a lifeboat for me.  Without Mullie's great work, I don't know how I would have survived that first year of intensive Mandarin.

Fortunately, I went to Taiwan not long thereafter and encountered another savior, Guóyǔ rìbào 國語日報 (Mandarin Daily News).  This was a daily newspaper that, at that time (1970-72), had a variety of materials for individuals of all ages (news, literature, essays, recipes, social criticism, philosophy, and so forth), and everything was annotated with phonetic symbols, so I didn't have to struggle with radicals and residual strokes.  I read every word of that paper each day for two years, and by the end of that time I had painlessly acquired the ability to recognize three or four thousand characters and probably could write more than a thousand characters.

The same folks who published the newspaper also put out a series called Shū hàn rén 書和人 (Books and People) and another wonderful series entitled Gǔjīn wénxuǎn 古今文選 (Anthology of Ancient and Modern Literature), in both of which I read voluminously, thus greatly enriching my exposure to texts of all sorts from every period of Chinese history.

Thus I struggled to find my own path to a rational, workable method for learning Chinese without suffering unduly from rote memorization (sǐbèi/jì 死背/記 [lit., "deadly memorization"]).  I have summarized these methods in "How to learn to read Chinese" and "How to learn Chinese and Japanese".

My wife, Zhang Liqing, told me that, in her long career as a professional language teacher at Harvard, Bryn Mawr / Haverford, Swarthmore, Penn, Oberlin, and Middlebury, a few of her best students instinctively had the same antipathy to the characters and zest for the spoken language that I did during the initial stages of learning Mandarin.  Of course, virtually all foreign students of Cantonese, Taiwanese, Shanghainese, etc. do so through Romanization.

But now I want to jump to the present and then take a look at the future.

On Tuesday, April 1, David Moser (of Beijing Capital Normal University and Academic Director of Chinese Studies at CET Beijing) delivered the following lecture at Penn:  "Is Character Writing Still a Basic Skill?  The New Digital Chinese Tools and their Implications for Chinese Learning".


In the past decade an explosion of digital devices and technological innovations have virtually revolutionized the study of Chinese language and Chinese script. These include tools such as comprehensive digital dictionaries (such as Pleco), hand-held multilingual dictionaries and translation devices, a plethora of online translation tools, hundreds of Hanzi-to-pinyin conversion tools, various optical character recognition (OCR) programs, text-to-speech converters, and a dazzling selection of pedagogical software (such as Wenlin), databases, and digital tools that can translate, annotate, gloss, format, read, write, speak and explain Chinese texts in various fashions and to varying degrees of sophistication. These tools, easily installed on smart phones and pads, allow beginning and intermediate students to easily navigate a vast range of texts that would have been nearly impossible a decade ago. Thus these digital tools clearly have the potential to utterly revolutionize the teaching of Chinese at all levels. Yet many students fail to make full use of such aids, and a great many Chinese teachers have not been able to incorporate these tools into the standard curriculum. These new aids important implications for the learning of Chinese characters in particular, and have prompted a reevaluation of the importance of written character memorization (e.g. daily tingxie exercises). Since even native Chinese speakers are losing the ability to produce characters by hand, to what extent is writing still a basic skill? What are the cognitive and psycholinguistic implications of these new methods of processing Chinese script? How can we teach students to make better use of these tools? How many of these digital aids are truly useful, and how many are merely distracting and counterproductive?

It was a brilliant lecture.  David demonstrated many of the new tools to the audience.  For example, he used his iPad to take a photograph of a character that was written on the blackboard.  He then ran that character through the dictionaries in the iPad, retrieved the meanings of the characters, and played the pronunciation for the audience.  What is really fantastic is that, utilizing the advanced OCR capabilities of the various new programs he introduced, David showed how you can scan texts, phonetically annotate them, and obtain detailed explanations for each character and translations of whole texts.

Basically, one could say that the miracles that David presented to us relied on two fundamental techniques:  digitization and phoneticization.  Watching David perform all of this magic was like a dream come true for me.  These are exactly the things that I have been wanting for over forty years, and now they are a reality.

How shall we use these marvelous tools, and what are their implications for Chinese language teaching / learning?  I could expatiate on the virtues of these new applications for days, but here I shall merely outline them under several points:

1. they will eliminate the need for the dreaded, boring, antiquated, stifling tīngxiě 听 写 / 聽寫 ("dictation") exercises

2. they will banish the fear of character amnesia (electronic devices are already writing our characters for us)

3. they will enable students to read massive amounts of quality texts on the widest possible variety of subjects without having to endure the agony and drudgery of looking up characters by radicals, stroke order, shape, etc. — I wasted years of my life on exactly those tasks — and it is precisely the reading of large quantities of real Chinese that facilitates the acquisition of a confident Sprachgefühl for the language in diverse contexts

In "Sneeze, hiccup, cough", I summarized David's research on the inability of Chinese to write common characters.  If native speakers of Chinese are increasingly unable to reproduce the characters for simple words ("egg", "shrimp", "toad", and so on), why should we force non-native speakers to do so?

Chinese characters are hard.  Do not believe anyone who tells you that they are easy:  whoever does so is a quack.

ShaoLan Hsueh, who doesn't even bother to tell us how to pronounce Chinese characters, thus further divorcing them from language than they already are, claims that "English-speakers can start learning to read Chinese in less than 10 minutes."  10 months for a start, maybe, and 10 years for proficiency.

"Chineasy? Not"

The road to Chinese literacy is hard, hard, hard — just like that fabled road to remote and inaccessible Shu.

Let the machines serve us so that we no longer have to be slaves of the characters.


  1. Mara K said,

    April 5, 2014 @ 3:57 pm

    Hear hear.

  2. Rubrick said,

    April 5, 2014 @ 4:09 pm

    All you say seems eminently sensible — and yet I can't help but feel there is something ominous about tying "literate" to "wealthy enough to afford a smartphone".

  3. Milan said,

    April 5, 2014 @ 4:30 pm

    Rubrick, yes: this will probably revolutionize the teaching of Mandarin to non-Chinese students, but it won't help most of the millions of speaker of Chinese topolects other than Mandarin who at the moment effectively have to learn a second language to achieve any literacy.

  4. adbge said,

    April 5, 2014 @ 4:44 pm

    For the unfamiliar, you might want to check out this review of spaced repetition for facilitating language learning.

  5. Matt Kosko said,

    April 5, 2014 @ 4:48 pm

    A minor point, but do you mean to write "Shū hé rén" rather than "Shū hàn rén"?

  6. Ned Danison said,

    April 5, 2014 @ 5:48 pm

    Matt: 和 is pronounced hàn in Taiwan.

  7. Jeremy said,

    April 5, 2014 @ 6:18 pm


    Anyone wealthy enough to want to learn Chinese as a foreign language is wealthy enough to afford a smartphone.

  8. Daniel said,

    April 5, 2014 @ 8:22 pm

    You don't need to be wealthy enough for a smartphone. You could pick up a inexpensive local brand phone with these character recognition abilities if you're in China. Or, just install (free) youdao or the new bing dictionary/translator (also free) on your computer – both are better than Pleco, in my experience. Or, install qq and talk to real Chinese people in whatever level of Chinese you have… It's still a very hard language, but learning Chinese has never been easier

  9. Victor Mair said,

    April 5, 2014 @ 8:24 pm

    A major point concerning language acquisition that David stressed in his talk but that I forgot to mention in the original post is that the basic skills of speaking, listening, reading, and writing reinforce each other in a close-knit cycle when one is learning a language that uses a strongly phonetic writing system, but that the cycle is broken when one is learning a language with a weakly phonetic writing system such as Chinese characters. The good news is that the new tools which David introduced to us serve to erase the gap in the learning cycle by phoneticizing the characters, thus joining the cycle of speech and script as they are in languages that use phonetic scripts.

  10. Ken Brown said,

    April 5, 2014 @ 8:36 pm

    "Anyone wealthy enough to want to learn Chinese as a foreign language…"? Really? North Korean refugees? Migrant workers from ethnic minorities? Even non-Mandarin-speaking Chinese who did badly at school and never quite managed to become fluent or fully literate? There must be millions of them.

  11. Matt said,

    April 5, 2014 @ 9:52 pm

    @Ken Brown:
    Now you're just nitpicking. He said "foreign language". Out of the examples you gave, North Korean refugees are the only ones for whom Mandarin would really be considered a fully foreign language.

    Regardless, smartphones aren't as rare as you seem to think. Not every smartphone is a luxurious iPhone or Samsung Galaxy:

  12. CThornett said,

    April 6, 2014 @ 12:47 am

    Shinichi Suzuki followed similar principles in developing his 'mother tongue' method of teaching music–listening, more listening, and guided imitation of the music listened to and the teacher's stance, bow hold, violin hold and so forth. Reading music notation and explicit music theory can wait until listening and reproducing skills are in place. The Suzuki Method is used mainly with children now, but was initially developed for adults. (And of course much traditional music is learned without any notation at all.)

    The contrast is with those schools of music pedagogy which, like your early Mandarin lessons, tend to begin with notation and theory. Early-stage learning by ear doesn't necessarily impede learning to sight read music, either.

    Attending my youngest's Suzuki violin lessons was a real help to me as a language teacher. Listening first, then speaking.

  13. Brendan said,

    April 6, 2014 @ 12:54 am

    David was also talking about the future of language teaching, and smartphones, computers, and tablets are only going to get cheaper and more widely available. Pleco runs great on a three- or four year-old iPhone, and presumably runs pretty well on low-cost, low-spec Android devices as well; Wenlin, the desktop computer app that started all this, is lightning-fast on any computer made in the last 15 years or so, and it runs on every platform. (There's no official Linux support, but it works through WINE without a problem.) Browser addons like Peraperakun are available free of charge, and would work even for people who had to use shared library or internet cafe computers.
    I haven't sat down and added up the costs for all the software I have, but it would be possible for a student using a hand-me-down phone to download Pleco (base cost: $0) and get pretty far using nothing but the built-in free dictionary. Add-ons cost money, of course, but so do physical textbooks and dictionaries.

  14. Jeff W said,

    April 6, 2014 @ 4:21 am

    If native speakers of Chinese are increasingly unable to reproduce the characters for simple words ("egg", "shrimp", "toad", and so on), why should we force non-native speakers to do so?

    Isn’t the answer that native speakers of Chinese can, presumably, already recognize the characters for simple words but that non-native speakers—which, I assume, means “learners” here—are learning the characters and that reproduction, being harder than recognition, more fully ensures that the characters are learned?

    If I, as a learner, have to reproduce the character, I have to attend to features that I might overlook in merely recognizing it. I might know—or think I know—the character if I can recognize it but if I can reproduce it, I really know it—and I can recognize it more easily or more accurately (or both) later. Reproduction for the non-native learner acts as a kind of proxy—maybe not an ideal one—for the enriched environment (in terms of exposure to the characters on a constant basis for years and discriminating between, say, similar-looking characters) that native speakers experience. Am I missing something here?

  15. Jacques LEGER said,

    April 6, 2014 @ 6:01 am

    For all these reasons and many more, I have developed a new Chinese Word Processor for Non-Chinese.
    You may say "one more !”.
    But if you visit you will understand our approach : "Use the computer for remembering and use your brain for thinking".
    I am available to share the product with anyone interested by the development of such techniques through the community of Western people learning Chinese.

  16. Zhaoning Wang said,

    April 6, 2014 @ 6:30 am

    We have prepared a set of Pinyin-Chinese.

    A input method named "DD Shurufa Pingtai" is provided by NGOs. With this 输入法 most of "Euro-Letters" can be included in our Pinyin-Chinese or Pinyinese creation.

    We are now trying to teach Pinyinese in 速记.

    I can mail to you the work in process in excel format if you are interested.


  17. Victor Mair said,

    April 6, 2014 @ 6:58 am


    Thanks for drawing the excellent analogy with the Suzuki method for learning violin.

    @Jeff W

    Yes, you are missing something. The tools that David introduced are liberating language learners from the mindless labor that you advocate. By letting machines do the labor for them in the early stages, while reading massive amounts of interesting and authentic materials — something your old-fashioned method will not permit — the learning process is jump-started, and the learning process far more quickly and with less suffering reaches the stage that native readers operate at.

    I don't think that you've read my original post and all of the comments carefully and with an open mind — at least you haven't taken them to heart.

    How many years have you spent learning Chinese the old way? What level are you at? Can you read a newspaper? As David observed in his lecture, after 4-5 years using the old methods (like the one you advocate), students still can't read newspapers with a high degree of comprehension, but with the new methods (machine-assisted reading), they start to read them effectively within a couple of years. Before your turtle students have gotten out of the gate, our rabbit runners are miles ahead, and enjoying the thrill of reading interesting texts, and using what they read in conversation and work. Your turtles will never catch up to the rabbits, and more often than not they feel defeated and fatigued. Moreover, our sprinting rabbits more quickly are able to read and write Chinese the way most Chinese are now doing.

    But if you truly enjoy the masochistic drudgery, you certainly have to right to stick to the old way of learning Chinese.

  18. Lugubert said,

    April 6, 2014 @ 7:19 am

    After retirement in 2001 following a stroke, I moved back to my Alma Mater city and wanted to go on from 1969 first semester Sanskrit. They were dismantling the department. Still a part time translator, I wanted to know how I learned languages, and how my mind worked when translating. I had to choose a language I knew enough of to be fascinated by it, but not enough to have a head start. I went for Chinese.

    My goal at the time apart from the two abovementioned aspects was to learn enough to read newspapers and really easy texts. Traveling to China and using the language unaided felt unrealistic on many levels. I soon became too fascinated by the language as such to remember my first goals… I still have absolutely no idea how my learning and translating work. I’ll have to accept that the processes do work, and forget about taking up psychology and neurolinguistics and whatever at the age of 71.

    To reconnect to the thread, professor Mair’s prescription would probably have worked wonders if I’d had the idea and finances to apply it half a century ago. Now, I’m grateful that my Swedish teacher taught how to recognize character parts. Real etymology, not Heisig and clones mnemonics, work for me. University funds unfortunately didn’t cover indigenous conversation, and the proverbial haystack needle is/was easier to find than a local teacher.

    I’ve made a couple of China journeys together with a former classmate for a trip organizer and heptalingual interpreter. Shrewdly, I was left on my own a day each in a couple of cities, to prove that I somehow managed, and I survived! But on the few occasions when I managed to have locals seemingly understand what I said, if more than complicated than to point to a menu item featuring an ox or pig radical or asking for a bottle of beer, I mostly was totally out of my depth when they answered.

  19. Gene Callahan said,

    April 6, 2014 @ 9:19 am

    @Rubrick: "and yet I can't help but feel there is something ominous about tying "literate" to "wealthy enough to afford a smartphone"."

    Literacy has always been expensive and always will be. If the State can subsidize years of learning to write characters by hand to raise literacy rates, it can surely much more easily subsidize the much cheaper distribution of smart phones to poor school children, right?

  20. Elessorn said,

    April 6, 2014 @ 9:35 am

    I think mindless drudgery is a little harsh. At the least I think it's worthwhile getting students to understand the possibility of non-utilitarian standards of evaluation, to keep them from looking at the numerous activities cognate with character memorization, in greater and lesser ways, that you find all over life in East Asia, with something other than bemused contempt.

    BUT, I completely agree it is absolutely pointless, perhaps cruel, to adhere to more traditional ways of teaching in the West when the tools and methods Prof. Mair mentions have the potential to make learning Chinese or Japanese so much faster and more fun. Taking that as a granted baseline, then, since after all free apps are going to be discovered by students anyway, how do we get students to take the next step which these methods have made easier to take? To learn characters at all? If technology has made it possible to deal with characters without learning them, what answer do we give to the question, "why bother?"

  21. Victor Mair said,

    April 6, 2014 @ 9:55 am

    From Perry Link:

    Yes, interesting. At the Princeton pedagogy conference last year, I asked the question publicly about whether we should simply stop teaching stroke order, since everybody writes on a keyboard now anyway. There was push-back, even though I had not made a statement, but had only asked a question. The sanctity of Chinese characters somehow undergirds not only cultural pride but people's senses of personal security.

  22. Milan said,

    April 6, 2014 @ 10:16 am

    Victor Mair: Naturally, a lot of people balk at accepting the notion that something that has cost so much labour shall now be useless. Adhesion to obsolete technologies — or rather the attempts to keep them relevant — has been a major motor of cultural and artistic innovation; think for instance of the development of abstract painting after the invention of photography, so maybe we shouldn't be too quick too deride such technological conservatism.

  23. Victor Mair said,

    April 6, 2014 @ 10:17 am

    Ping Xu (from China) and Theresa Jen Kuo (from Taiwan) are both highly experienced professional Chinese language teachers, and Theresa is an outstanding calligrapher (I don't know about Ping in that regard). Together, around fifteen years ago, they developed "penless Chinese", for which see their "Penless Chinese Character Reproduction", Sino-Platonic Papers, 102 (March, 2000), 1-15 (free pdf) and "'Penless' Chinese Language Learning: A Computer-Assisted Approach," Journal of the Chinese Language Teachers Association, 40.2 (2005), 25-42.

    Ping sent in this comment:


    Our "penless" website was discontinued after we completed the federally funded project. But in case you don't have it, I will send you the article we published on the Journal of Chinese Language Teachers Association. You can also find discussions and applications of "penless Chinese" on the Web.

    At the time when we started with the "penless" project, the machines we were considering were mostly PCs. But we did actually anticipate what hand-held devices can further accelerate the adoption of the "penless" idea. We even discussed the possibility of voice recognition that will allow us to speak to the device and have it display Chinese characters, therefore even bypassing the need to hand input pinyin. That would be the last logical step if the "penless" idea is carried to its fullest! The tech world has shown us that they have been trying to do this with products such as Apple's Siri and Microsoft's Cortana. They are not quite there yet, but they will be, and soon. I think in ten years we will see our students dictate to their smartphone that will spit out sentences and articles in Chinese characters. That will realize the direct transfer between speech and text without relying on the intermediary of hand — a truly "penless" world, indeed!


  24. Victor Mair said,

    April 6, 2014 @ 10:19 am


    Young people are voting with their fingers and voices. They are the future.

  25. Stephan Stiller said,

    April 6, 2014 @ 11:15 am


    I agree with your facts, but I disagree with your conclusion. I'd say that the right to literacy is a modern human right (all the while acknowledging that human rights are an ever-evolving idea). The notion of "cultural innovation" that you mention must be subordinate to more essential rights.

    In fact, in order to participate in a society's culture, literacy is often a prerequisite.

  26. Milan said,

    April 6, 2014 @ 12:19 pm

    @Stephan Stiller: Definitely. I don't think that new technologies shouldn't be used in language teaching and I regret if I made that impression. I just don't think the teaching of obsolete techniques should be abandoned altogether — limited yes, and we shouldn't make mastery of a language dependent on mastering an obsolete technique, but we should encourage students to also learn about more traditional aspects of the language they are studying. Especially if they already have already achieved bare literacy in one or another language.

  27. Lane said,

    April 6, 2014 @ 12:33 pm

    I have a theory as to why change and reform of this kind of thing (and this thing in particular) is so hard:

    1) everyone would swear that they learn a hard language system like Chinese writing to take part in a wide and deep language community

    2) but really we know that people are (understandably) proud of their difficult achievements, and those who have learned to read and write Chinese have learned something very difficult. It is also valuable, and so

    3) to make it really easy on the next generation is to both destroy a powerful club membership (I survived this and have the scars to prove it!) and, practically speaking, to give lots more people a rare and valuable skill that, previously, few possessed.

    In my book I compare it to fraternity hazing: the club (and the bond) are worth little if they are open to everyone. And everyone in the club today (and ergo everyone in position to make tomorrow's rules) survived yesterday's hazing ritual. It takes a very grown-up person to say "I did this the hard way, but child, I want you to do it the easy way, for the greater good."

    So bravo to Victor and David and co. for doing just that.

  28. Stephan Stiller said,

    April 6, 2014 @ 1:03 pm

    I very much agree with your comment. (Though I have no particular opinion on whether your explanation is exhaustive. I'd need to give this some thought.) What you've written is a deep truth about social dynamics. I recall people describing a difficult class at my undergraduate institution using phrases such as, "But it's a good pain!" The truth is: The class was too much work for everyone, sapping students of energy needed for other things. The class should long have been split into two. Yet those that finished the class seemed to flaunt a sort of sadistic pride.

  29. Conal Boyce said,

    April 6, 2014 @ 1:40 pm

    The top of Volume 4 of my tattered old Gwoyeu Tsyrdean makes the perfect "cradle" for my smart phone. Before seeing this post, it hadn't occurred to me how symbolic that arrangement might be.
    First, a digression, though: What Victor describes regarding Chinese writing can be subsumed under a huge pattern that I call "dumplexity": In chemistry instruction, there are more and more virtual experiments, less and less contact with "dangerous" things like glass and stoppers that fly off and colored vapors and — well, c-h-e-m-i-s-t-r-y. In that field, I think the plasticky, sanitizing trend is a travesty (even though, as mentioned below, I think it makes perfect sense with Chinese writing to distance oneself and let computers do the heavy lifting). Speaking of computers, the area where the concept/term 'dumplexity' first occurred to me a few years ago was computer software education. I think there is an over-emphasis on "How many layers of complexity can we pile on?" versus "How does a simple 'Hello World' program even WORK in the first place?" (Eventually, only two or three people on the planet will know the answer to the latter question.) Two more examples: Airbus fly-by-wire and Toyota accelerator problems: software so 'advanced' and so deeply layered that no one knows what it does anymore. (Here, the analogue to stroke-count in Chinese [mentioned by Perry Link] turns out to be a detail that actually matters!) "Mysterious" transportation mishaps ensue: Air France for sure, possibly MH 370?
    So, IN GENERAL, I take a dim view of dumplexity — hence the pejorative name. For Chinese specifically, though, I think it is a good thing. Anyone who has actually "been there" would surely agree with Victor's take on the subject.

  30. Bruce Humes said,

    April 6, 2014 @ 10:41 pm

    I am a native speaker of English who learned Chinese at university, and went on to do many things with that knowledge, including train 8,000 Chinese managers in export management, launch 10 or so Chinese magazines/web sites, and translate a few novels.

    My road to fluency was similar to Professor Mair's. I loved reading Hànzì, but I found early on that intensive speaking and reading was the path — for me — to greater fluency, both spoken and oral. Unlike him, early on I preferred characters to romanized Chinese, but once I'd graduated I made no attempt to learn how to write what I could say or read. I too found the Guóyǔ rìbào approach very useful when I first got to Taiwan as a student, and even 2 years ago in Kunming when I read Sima Qian's “Shǐjì," I made a trip to HK to buy the Taiwan edition complete with bo po mo fo. Very useful!

    I have never learned to write more than a thousand characters or so by hand, and when I do so without a dictionary, I mix Japanese, traditional and simplified.

    About ten years ago I learned how to write Chinese on a computer, and now regularly write posts and even full-length essays. So it seems to me that:

    1) To attain reading and writing fluency in Chinese characters, a good foundation in the basics of writing is useful, but this need not require rote memorization of thousands of them;

    2) It is markedly easier to learn how to write a character once you know what it means, and can use it in conversation;

    3) Home-grown Chinese methods for teaching the language have long been misguided, and students of the language are better off doing what seems to work best for them.

  31. Guy said,

    April 7, 2014 @ 4:40 am

    Has anyone ever studied the differences in the rate of attrition for character amnesia in writing vs reading?

    As a native Mandarin speaker, I was a terrible student at Chinese, constantly failed my "ting xie" (spelling tests) and would confidently suffer from character amensia now if I ever bothered to write by hand.

    However when it comes to reading, my character amnesia is almost non-existent. I don't think I've ever forgotten how to read a character. I think the only exception might be when I come across personal names with rare characters no-one ever uses. In that sense, I much prefer reading characters over zhuyin/pinyin (which is just a pain), but definitely prefer to type using the latter methods.

    I wonder if other native or foreign speakers have this experience too?

  32. Dave Cragin said,

    April 7, 2014 @ 5:43 pm

    @Guy, As a native English speaker who is learning Mandarin as an adult, I have an experience that is different because I almost never handwrite characters, but it might be of interest: I can write an e-mail easier than I can read one.

    Why? Context. When I’m writing, I know what I want to say and Windows is good about offering the most probable characters I need when I enter the pinyin. I can often recognize the correct character in context.

    In contrast, when I try to read an e-mail, I don’t know the context so I’m much less likely to recognize the characters.

    This is opposite of what usually happens when learning another European language, that is, it's generally easier to read than write. As a simple example, in looking at German text, I can readily understand the various forms of "the" (der, die, das…), but using them myself correctly would be much much harder.

    (Caveat: I’ve been learning Chinese via a path recommended by Victor, i.e., 1st spoken, then pinyin and finally characters.)

  33. Dave Cragin said,

    April 7, 2014 @ 6:03 pm

    For those non-Chinese speakers, an example might help show how essential technology has become in learning Chinese. While a dictionary is normally the best friend of anyone learning a language, before the new technology, using a dictionary in Chinese was typically an exercise in frustration: Consider, with a character-based language, there is no alphabetical order.

    To look up a “word” in Langenscheidt or Berlitz dictionary, first you had to identify the character’s radical (not an easy prospect), count the strokes (sometimes difficult), and look it up in a “Radical Chart.” The chart then gives a number to look up in “Radical index” and this gives one the pinyin to look up in the text of the dictionary.

    However, it’s actually much harder than this because when you are looking up the character, you don’t know if the character is part of a 1, 2, 3 or more syllable word (no spaces are used with characters) so trying to determine meaning is extremely difficult.

    Finally, the above presupposes you actually find the character you are looking for. It’s quite possible you picked the “wrong” radical to look up the word. On top of this, Wikipedia notes that dictionaries don’t necessarily index characters in the same way. Technology eliminates this dreadful process.

    See also David Moser’s classic article: who noted "Imagine a language where simply looking a word up in the dictionary is considered a skill like debate or volleyball."

  34. Matt said,

    April 7, 2014 @ 8:31 pm

    I wonder if other native or foreign speakers have this experience too?

    For what it's worth, my experience with Japanese has been very similar, and I followed Bruce's approach — once I was no longer being formally assessed on it, I just stopped practicing pen-and-paper writing altogether. So now the number of characters I can write is a tiny fraction of the number I can recognize.

    It's kind of embarrassing that my handwriting is so childish when I have to write a note to the doctor or whatever, but the proper solution to that looks to me more like calligraphy classes than character memorization in and of itself. (The problem isn't lack of technical knowledge of the characters, it's an inability to write them in a balanced and visually appealing way, even when copying direct from a smartphone screen.)

  35. dainichi said,

    April 8, 2014 @ 12:36 am

    @Jeff W:

    "but if I can reproduce it, I really know it"

    If you define "really know" as "being able to reproduce", then sure. But I don't see how that's necessarily a meaningful definition. For example, I'd claim that I really know my wife's face, but I wouldn't be able to reproduce it if my life depended on it.

    I'd expect the same or similar parts of the brain to be used for facial recognition and character recognition. I'd be interested if someone knowledgeable could provide more details.

    Slightly off topic, and then maybe not: In my mind, I draw parallels between character writing and NP-problems. Solutions to NP -problems can be verified in polynomial time, but might take exponential time to compute (unless P=NP). Just another example that creating something and recognizing something take a different order of magnitude in effort.

  36. Stephan Stiller said,

    April 8, 2014 @ 2:44 am

    @¶1: In many cases the ability to reproduce entails the ability to recognize. This might simply be a side effect of the relatively higher difficulty of attaining the ability to reproduce, or it could have to do with the fact that active knowledge requires full knowledge, while passive knowledge is simply less knowledge or context-conditional knowledge (and I'm leaving the distinction between procedural and declarative knowledge aside here). So Jeff W's point was that direct acquisition of active knowledge (likely) includes acquisition of passive knowledge. That said [and starting here I'm also addressing others], Victor's (justified) reply to that was that direct acquisition of this particular kind of active knowledge is extremely tedious and less effective than learning through lots and lots of (passive) exposure. Let me also add that retrieval of a character from a pronunciation-meaning pair as an active skill does in fact not include knowledge of contextual choices of pronunciations of polyphonetic [← I just made up that word] characters or knowledge of word meanings (or in fact knowledge of any words at all), but extensive reading (of material with both characters and a phonetic transcription) will reinforce those aspects of character knowledge. Let me finally repeat something I wrote earlier, namely that I am fairly sure that characters are attached to words stored as phoneme/syllable sequences as opposed to words being built from graphically stored characters.
    @¶2: Facial recognition is a highly specialized skill, and I would expect Chinese characters to be different from letters, but I'd check which brain regions are associated with prosapagnosia and dyslexia. It turns out that the fusiform gyrus is involved in both facial recognition and "processing of Chinese characters" ("Brain Activation in the Processing of Chinese Characters and Words: A Functional MRI Study", Human Brain Mapping 10, 2000; pp. 16-27), though this provides only a partial answer. There must be papers on dyslexia for the Chinese language.
    @¶3: Well, construction of a solution is often more difficult than verification, and this is what your analogy relies on. In the case of character knowledge, we are talking about retrieval and recognition of fixed (static) knowledge. Also, the learning process (for active or passive command of this subject matter) is fairly simple and transparent, albeit lengthy. As for the paragraphs above, it is hard to come up with meaningful or useful analogies unless the underlying processes are known.

  37. David Moser said,

    April 8, 2014 @ 4:43 am

    Lane wrote:
    It takes a very grown-up person to say "I did this the hard way, but child, I want you to do it the easy way, for the greater good." So bravo to Victor and David and co. for doing just that.

    That's it in a nutshell. This makes me feel very good.

  38. Jeff W said,

    April 9, 2014 @ 5:42 pm

    @Stephan Stiller

    Jeff W's point was that direct acquisition of active knowledge (likely) includes acquisition of passive knowledge

    Thank you. Yes, that was exactly my point.

    I agree with you on everything you said about my point and about Victor Mair’s response, except that it was justified. You can see my response to him below.

    If you define "really know" as "being able to reproduce", then sure.
    I appreciate your comment but I wouldn’t define it that way—that would be tautological. By “really know,” I had in mind something like that someone would be able to discriminate between very similar characters or notice if a stroke was missing or added, to about the same degree as a native speaker of Chinese. In addition, Stephan Stiller got my point. See above.

    @Victor Mair
    You asked, in essence, why force non-native speakers to do x when native speakers themselves cannot.

    My response was that the non-native speakers are in a different position with respect to the Chinese language than native speakers and that x as a method for learning Chinese might make up for that difference in position. (I was not claiming that that x was a good method or that the rationale I proposed for using x was an especially good rationale—I was just proposing it as as one that seemed to be at least reasonable.)

    Your response to me is that y as a method for learning Chinese is better than x for non-native speakers.

    I don’t understand your response to me.

    The entire post, as far as I understood it, was about assessing one method [machine-assisted learning] over another [reproducing characters] for non-native speakers who are learning Chinese. Looking at what native Chinese speakers can or cannot do in assessing either method for non-native speakers didn’t make a whole lot of sense to me—maybe it had to do with the way they typically learn the characters—but, as long as you were asking, I suggested what seemed to be a least a reasonable possible rationale as a way of trying to understand the question.

    I was trying to make sense of your question and offering an answer to it. Your response to me seemed unduly combative, challenging my level of reading Chinese, referring to my “turtle students” as opposed to your “rabbit runners” and assuming that I enjoy “masochistic drudgery” or that I “advocate” “mindless labor.” I am not at any level of Chinese, I don’t have any students, turtle, rabbit, or otherwise. I was not “advocating” anything. I was not picking a fight with you but it seems like you are fighting with me. If you are going to respond to me, I would like you to respond appropriately to what I posted, in the spirit that I posted it—for clarification and understanding—and not as a challenge to your position.

  39. Victor Mair said,

    April 11, 2014 @ 3:47 pm

    @Jeff W

    You asked some questions and I strove to answer them from the vantage of someone who has been through more than four decades of learning Mandarin. The reference to rabbits and turtles was an attempt to make the exchange light and humorous, and "your" was not referring to Jeff W but to whoever advocates such methods. I never dreamed that someone would take offense at my reply, especially since I have made it clear throughout this post and the comments to it that I think it's understandable that some people will prefer to continue the traditional ways of learning Mandarin. I do hope that you will allow me to suggest for beginners that there is an easier, more effective way — if that's what they want. There's no single way to learn Chinese languages (or any languages, for that matter), and there never will be.

  40. arthur waldron said,

    April 13, 2014 @ 2:01 pm

    Why is he han in Taiwan? It is so pronounced, I believe, in official guoyu which the Nationalists had invented or begun to impose earlier.But in PRC such pronunciation is ridiculed. I don't think it was assigned after the NATS got to Taiwan though.

    What I see here is developing linguistic chaos. No romanization really works, I think, but characters pose the problems mentioned. Japanese they could just romanize with no problem. But Chinese is not so easy (can anyone comment on Vietnamese? It is romanized)

    I'm thinking of an article on the end of Chinese culture as we and they know/knew it, over the last sixty years Linguistic chaos is certainly part of this. As always so many ways out are proposed that no one will ever be adopted uniformly, and the best one, as with GR, will certainly not be adopted.

    You know that just getting spoken commands into a form intelligible to the troops is still a problem for the Chinese army?

    In any case. It is over. The 5,000 years no have a terminus–round about 1956 or so or a little later. That was two generations ago. When the President of China buys the 論語 and declares "I should read this" then we have clear evidence that the culture is gone. I do not underestimate how it will perdure tenaciously. It is a great culture. But as a shared national asset, no longer.

    So what takes its place? The linguistic equivalent of Beijing architecture? When visible through the toxic smog?
    world copyright–ANW

RSS feed for comments on this post