Spelling mistakes in English and miswritten characters in Chinese

« previous post | next post »

In the comments to "Character amnesia revisited", Joanne Salton remarked that "It doesn't have a tremendous effect on the ability to communicate because the odd mistake doesn't matter all that much."  I started to dash off a brief reply, but my answer soon grew to such inordinate length that it seemed to merit separate posting under the above title.

If one is talking about written communication, which I assume Joanne is since she's dealing with characters, then unfortunately occasional mistakes have a huge impact on the ability to communicate.  Joanne herself makes the comparison to "having every word in the English language being something that might cause a spelling mistake", and that is actually a very good analogy, though perhaps a bit extreme.  But there's a great difference between mistakes in writing with Chinese characters and mistakes in writing with the Roman alphabet.

If you mess up the spelling of English words, the person who reads what you wrote can almost always make perfect sense of it.  For example, you can write "it's" or "its'" when you mean "its", you can write "neccesary" or "necissery" or any of a dozen other permutations and combinations when you mean "necessary", you can write "beleive" when you mean "believe", and so on and so forth.  I have correspondents who constantly make exactly the same sorts of mistakes, but it doesn't hamper our communication in the slightest, though for someone who cares about spelling, grammar, and punctuation, it can be a little bit jarring and occasionally even a little bit humorous to encounter this sort of fractured English writing.  And we've all seen those experiments where the internal letters of English words are jumbled around or other sorts of transformations are worked on English words in a passage, and yet you can usually read them off without much difficulty.

The situation is very different with Chinese characters.  If you forget one stroke or misplace one line or dot, or if you write one stroke in slightly the wrong direction or if you cross strokes that are just supposed to touch or that are supposed to get close but not quite touch, or if one of your strokes curves in the wrong way, etc., then your attempt to write that character will usually FAIL altogether.  What frequently happens is that the writer who cannot remember exactly how to write ALL of the strokes of a given character in their proper orientation (the correct sequence is also important) will simply give up.  They will either smudge, cross out, or otherwise erase the character they were trying to write, and in handwritten texts I've often seen Pinyin inserted to replace the failed character or following what remains of the crossed out attempt to write a character.

I suppose that the cold, hollow, empty, sinking feeling one experiences when one realizes that one can't remember how to write a particular character is best encapsulated in the expression "tí bǐ wàng zì" 提笔忘字, which means that you pick up your brush, pen, pencil, stylus, fingertip, or other writing instrument, but the character just won't come out.  You simply have to confront the fact that you have forgotten how to write the character as a whole, though you may know one or more of its components.

Contrast that with the situation in English.  You pick up your writing instrument and have the gnawing feeling that you can't remember exactly how it is spelled.  If you're a perfectionist, it will grate upon you, but you can still write it out in an approximate form and be confident that your reader will understand what you meant — so long as you get the sounds roughly write (whoops! I meant "rite", no, "right"!).  And if you're typing, well nowadays the spell checker will tell you you're wrong, and you can keep retyping till you get it right.  And I've been astonished at how incredibly good Google is at suggesting correct spellings for words that you've mangled.  Think of all the different ways you can misspell "innocuous", for example.  Google seems to be able to catch all of them instantaneously.  It's like a miracle to me that Google can do that.  Right while I am mistyping, Google is correcting me!

Now, there's nothing like that in Chinese.  If you are trying to write a character by hand or enter its shape into a computer via components or stylus / fingertip, if you don't really know for absolutely certain how to write it, the computer will not be able to help you.  However, if you type in the sound of the character, the computer / cell phone / gizmo / gadget will write the character correctly for you.  No more "tí bǐ wàng zì" 提笔忘字 ("lift brush forget character").

Here are a few relevant Language Log posts:

"New radicals in an old writing system"

"The unpredictability of Chinese character formation and pronunciation"

"The esthetics of East Asian writing"

"The cost of illiteracy in China"

Share:



35 Comments »

  1. J.W. Brewer said,

    December 18, 2012 @ 12:45 pm

    How do the gadgets deal with Mandarin's ubiquitous homophones? Just looking at e.g. wiktionary, it's apparently not unusual for a single pinyin monosyllable (with tone specified to narrow the field) to correspond to a dozen or so different hanzi. I assume the gadget can give you a pulldown menu from which to choose, but that assumes that (even if you couldn't produce it with a brush or pen) you can recognize by eye which alternative is the one you want, right?

  2. Matt Whyndham said,

    December 18, 2012 @ 1:03 pm

    … basic frequency (most common presented to the user first), and frequency given the other Sounds/Characters preceding the character of interest. You do have to know enough characters to pick the right one when presented. If you know the tone, you can narrow it down that way, eg. bi3 –> 笔 . The menus usually fly out rather than drop down. Try playing with Windows/Mac Chinese input methods if you have a spare moment.

  3. Eric Vinyl said,

    December 18, 2012 @ 1:41 pm

    If you are trying to write a character by hand or enter its shape into a computer via components or stylus / fingertip, if you don't really know for absolutely certain how to write it, the computer will not be able to help you.

    I don’t know about all that. Between mandarintools.com’s character lookup by radical + strokes, and the couple webpages that let you draw (badly) a character with your pointing device and give you a list of potential matches, those are some invaluable tools that have helped me find the character I’m looking for when I can almost-but-not-quite see it in my mind. They’re usually not even that picky about stroke order.

  4. Victor Mair said,

    December 18, 2012 @ 1:44 pm

    @J.W. Brewer

    Moreover (following up on what Matt Whyndham said), if you type in a whole sentence with Pinyin Romanization, the software will select the right characters with a very high degree of accuracy. You usually only have to choose from a selection in a small number of cases. I myself normally type word by word (a Mandarin word averages out to about two syllables / characters in length), and the software almost always gives me the ones I want for the word in question (homophony is not a big problem when you treat Chinese languages as consisting of polysyllabic words rather than syllable-long written characters). If not, the automatic drop down menu offers the alternatives right before your eyes. You don't have to go looking for them. Some sophisticated Pinyin inputting systems are syntax sensitive as well.

  5. Eric Vinyl said,

    December 18, 2012 @ 1:45 pm

    @J.W. Brewer

    Reading is a different process than writing. Oftentimes you can recognize by sight which character is the one you want, even if you can’t for the bloody life of you think of how to go about drawing it.

  6. Victor Mair said,

    December 18, 2012 @ 3:48 pm

    @Eric Vinyl

    "…mandarintools.com’s character lookup by radical + strokes…."

    Well, first you have to know the radical and then count the strokes, and that is often both very hard (maddeningly frustrating) and time-consuming to determine.

    "…couple webpages that let you draw (badly) a character with your pointing device and give you a list of potential matches…."

    I've seen them. They're slow and messy and frequently hit or miss, not very useful for those for whom time is at a premium.

  7. Adrian said,

    December 18, 2012 @ 4:49 pm

    Hmm. Although you're the expert, Victor, I'm a tad sceptical. Is the situation really so different between the two languages? If you "misspell" a character can't the reader often work out what it should be, from the context? For example, if the sentence is "He's a nice person" but instead of writing ren (person) you write ru (enter), can't a reader figure this out?

  8. wren ng thornton said,

    December 18, 2012 @ 5:08 pm

    If you are trying to write a character by hand or enter its shape into a computer via components or stylus / fingertip, if you don't really know for absolutely certain how to write it, the computer will not be able to help you.

    I don't know how it is for Chinese, but that's certainly not the case for Japanese. When using the handwriting recognition on my iPhone it always gives me a list of closest matches to pick from— including autocompletion for the character (sometimes) and autocompletion for the word (often). And that's not including pseudo-manual input systems like SKIP or searching by combination of components (no, I don't mean radical+stroke, though that's an option too). It's not just an Apple thing either; my Zaurus did the same thing over half a decade ago— and frankly, was better at it than the iPhone is. I'd be shocked if Chinese character support doesn't have these same facilities on any modern system.

  9. John Meissner said,

    December 18, 2012 @ 8:05 pm

    Now, there's nothing like that in Chinese. If you are trying to write a character by hand or enter its shape into a computer via components or stylus / fingertip, if you don't really know for absolutely certain how to write it, the computer will not be able to help you.

    At least on my Android phone with QQ 输入法, this is quite false. If there's a part of the character you're unsure about, you just can put in a suitable amount of random squiggles instead. The program will make its best guess, and it will also give a list of other alternatives. I just tried it for 赢. I wrote the top part, and doodled around in the bottom, and it worked fine. Other alternatives given were 蠃, 癔, and 羸.

  10. The Ridger said,

    December 18, 2012 @ 9:09 pm

    I constantly read that characters must be drawn in the correct order, but why is that so? Particularly with machine/digital characters, can anyone really tell that you drew it, say left to right instead of in proper order? Does it distinguish homophones? What I can find about it says it makes it easier, or more logical, to write, but why is it such a shibboleth, that if you forget the stroke order and yet manage to get the character to look the same, you have written it "wrongly"?

  11. dainichi said,

    December 18, 2012 @ 9:41 pm

    > What frequently happens is that the writer who cannot remember exactly how to write ALL of the strokes of a given character in their proper orientation (the correct sequence is also important) will simply give up.

    If China is anything like Japan, you are overstating the situation by far. Japanese people make mistakes when writing kanji all the time, and the meaning still comes across.

    Sorry, not trying to put words into your mouth, but it sounds a little bit like you can just take any character, change any one stroke, and you necessarily get another character which means something completely different. Not so.

    Yes, it is true that some pairs of characters differ by very little, so one misplaced/missing stroke can alter the meaning, but firstly, writers tend to be more aware of these pairs thereby knowing to be careful with them, and secondly, readers are usually also alert that some characters are more likely to be mis-written than others and therefore mentally correct them than other.

    Humans excel at that kind of "fuzzy" logic.

  12. dainichi said,

    December 18, 2012 @ 9:43 pm

    That should be:

    therefore mentally correct them quickly.

  13. Victor Mair said,

    December 18, 2012 @ 9:44 pm

    @John Meissner

    "I just tried it for 赢. I wrote the top part, and doodled around in the bottom, and it worked fine. Other alternatives given were 蠃, 癔, and 羸."

    That's really a big help! (tongue in cheek) Now how are you going to decide among these if you're not sure how to write the one you were looking for? Do you know the meanings and pronunciations of all these characters? Three of them look very similar and are very hard to distinguish, even for advanced scholars; the other is quite distinct in sound, meaning, and even shape. How long did it take you to do the doodling? How long did it take you to verify that you had chosen the right form of the four that your doodling produced? HOW did you verify that you had the right form?

  14. John Meissner said,

    December 18, 2012 @ 10:33 pm

    @Victor Mair

    Now how are you going to decide among these if you're not sure how to write the one you were looking for? Do you know the meanings and pronunciations of all these characters? Three of them look very similar and are very hard to distinguish, even for advanced scholars; the other is quite distinct in sound, meaning, and even shape.

    I have no trouble recognizing characters. It's actively recalling the strokes that is hard. I was under the impression that "提笔忘字" in native speakers (who are otherwise literate) works in much the same way, and that they would have no problem picking out 赢 among the line-up above, even if they don't know how to write 赢, and even if they (like me) have no clue what the three other characters mean.

    I mean, the exact same problem exists in pinyin input systems, doesn't it? You still have to choose among options which may potentially look very similar, but people seem to manage that mostly fine.

    As for how long it took to make the doodle: I literally just swipe my fingertip around a bit, so half a second, maybe? All in all, I think the time taken is comparable to using Google or a spellchecker to correct misspellings in English.

  15. Victor Mair said,

    December 18, 2012 @ 10:48 pm

    @John Meissner

    So you know the sound and meaning of 赢, 蠃, and 羸 right off and can write and distinguish among them in half a second? If so, you're much better than my best graduate students in China.

  16. John Meissner said,

    December 18, 2012 @ 11:11 pm

    @Victor Mair

    No. As I said, I had no clue what 蠃 or 羸 meant, nor how they were pronounced. I'm only saying that I recognized 赢 as the character that I wanted, and so that's the one I picked. What I did know about 蠃 and 羸 was that they looked like I didn't know anything about them.

    Surely, this is the same process that users of pinyin input systems go through, at least in the cases when they do have to input single characters. If I type "ying" into my pinyin IME, it gives me lots of options. Some are characters I don't recognize, but that's fine, because I still know that they are not the character I want.

    Note also that the input system did show 赢 as the first option, presumably because it takes frequency of use into account, just like pinyin input systems. That helps too.

  17. John Q said,

    December 18, 2012 @ 11:22 pm

    Victor, it seems as if you're being unnecessarily obtuse. If I intend to write 贏 (without the aid of a computer) and I had forgotten the exact components of the bottom half (which I expect is quite common if one does not write this character frequently), I would write the 亡口月 and then just scribble something in the bottom right corner. I believe most readers would not fail to understand what I wrote.

    @The Ridger
    The order matters in calligraphy, and also a bit for character recognition. Most handwriting is scribbled. For example, check this out:
    http://graphicinsight.co.za/stalinwriting.htm
    A Russian speaker would have no problem reading that, but I can't even figure out what it says though I do know the Cyrillic alphabet. Same goes for Chinese.

  18. dainichi said,

    December 18, 2012 @ 11:42 pm

    @Victor Mair

    If you and I have a common friend, would you be able quickly draw him in a way that I could recognize him? Unless you have good drawing skills, my guess is: no.

    If you have 3 photos, one of which is of a friend of yours, can you immediately recognize which one? My guess would be: yes.

    Characters are pretty much the same way. Remembering how to write them requires substantially more memory than being able to recognize them. Is this memory ratio bigger than for native English speakers writing and reading English words? I have no sources, but I would guess that the answer is yes, since English orthography is a lot more predictable from pronunciation than Chinese characters are, in spite of the former's messed-up-ness.

  19. rgove said,

    December 18, 2012 @ 11:52 pm

    I have to speak up in support for John Meissner. I'm an intermediate learner who often forgets exactly how to write 赢, but I would have no problem picking it out from a list of those four options as the only one I recognized. Once one has learned to analyze a character into its components, recognition is MUCH easier than production, and even visually-similar pairs like 衣/农 or 乐/东 start to look distinct.

  20. michael farris said,

    December 19, 2012 @ 1:51 am

    What I remember from my brief foray into kanji before deciding I wasn't cut out for them in any way:

    - active writing was much harder than passive recognition,
    - analysing characters into parts (not necessarily the 'official' way) made recognition easier
    - learning characters in combinations (as used in real words) was far easier than learning them one by one.

    I wonder what the longterm consequences are for a language community to have almost entirely passive knowledge of their writing system (all these doodads are great but reinforce that writing in Chinese is more and more a one way street)

    And, I'll mention again that Mandarin users (at least in the PR) seem to be training themselves en masse to actively write Mandarin in pinyin (and associate pinyin with meaning).
    If there's a tipping point for pinyin it will come with those who've spent far more time actually writing in pinyin than in characters – at what point are some of them going to start perceiving characters as an extra, redundant step in the writing (and reading) process?

  21. Victor Mair said,

    December 19, 2012 @ 1:46 pm

    @John Meissner

    You have not responded to my skepticism that you did all that in "half a second". I've watched lots of Chinese friends, students, colleagues, and strangers using systems like the one you described, and it takes them much longer to achieve the sort of feat you claim for yourself.

    Now for some more detailed questions.

    1. If you are going to look up 赢 with your system, you're probably going to have to enter at least three, and probably four or more, strokes to narrow down the amount of characters from which you have to choose to a workable number, and then you have to take some time choosing, especially since those three characters — 赢, 蠃, and 羸 — look a lot alike and they have many strokes jammed into a small space, so you'll have to squint at them to distinguish them clearly.

    2. You stated that one of the alternatives given was 癔. How did that happen? If your system brought up 癔 after three or four strokes, then potentially it might also have brought up hundreds of other characters written with the "sickness" radical (#104 in the Kangxi system). Moreover, 癔 ("hysteria") is not a particularly high frequency character, so there should have been many other characters in the queue before it.

    3. How many strokes did you have to enter to get 赢 to come up as the first option? John Q mentions 亡口月 — that's a lot of strokes! You could not possibly have done that, or anything near that, in "half a second", much less decide which of the proffered characters was the one you wanted.

    4. If we call your way of inputting the "doodle method", we should admit that doodling takes some time (check standard definitions of the word) and is often aimless to boot, so your blazing "half second" doesn't jibe with your own characterization of your method as "doodling".

    @John Q

    "Obtuse" is a recognized synonym for "stupid".

    I've been studying, observing, and teaching about these issues carefully for nearly four decades, so my approach to them is not that of a dullard.

    @dainichi

    I have no problem with anything you've written, and the analogy of the photos is a good one.

    @Michael Farris

    Thank you very much for your common sense and realism.

    Finally, to all of those who think it's easier to struggle with handwriting input systems to call up the full 17 strokes of 赢 (simplified form) or 20 strokes of 贏 (traditional form), I submit that it would be far easier, faster, and more efficient, and vastly less frustrating, simply to type "daying" (you don't even need tone marks) into your pinyin entry system, and you would effortlessly get the character you're after, which, in any event, usually occurs in the expression dǎyíng 打赢 ("win"). Even if you only want / need the 赢 of 打赢, it's easier just to produce 打赢 through Pinyin entry and get rid of the 打 than to wrestle with handwriting this admittedly refractory character. If you're lazy and don't want to be bothered by typing "daying", then you can type just "ying", and guess what? 赢 will pop up within the first five characters offered to you.

    As I shall demonstrate unmistakably in a long post that I am preparing on the present subject, this (using Pinyin inputting) is exactly what the vast majority of Chinese speakers do. It is ironic that it is a few foreign students who are clinging most tenaciously to handwriting input systems, but there is a reason for that too, and I will explain what it is in my forthcoming post as well.

  22. John C. said,

    December 19, 2012 @ 4:30 pm

    >I submit that it would be far easier, faster, and more efficient, and vastly less frustrating

    I have only anecdotal evidence, N=2, but my Taiwanese-native in-laws would disagree strenuously. We bought them an iPad after they visited us and discovered that its handwriting recognition was *vastly* easier for them than IME-based inputting.

    Perhaps that's because they learned bopomofo rather than pinyin; perhaps it's because they're in their 60s and prefer more traditional mechanisms. It's also possible that input via strokes is objectively slower, but subjectively feels faster and more fluid (there are various human-computer interaction studies that show such a discrepancy, as you are no doubt aware).

    For what it's worth, my wife, who spent her childhood in Taiwan and today does a lot of business with mainland China, finds pinyin very hard to work in. For her, "simply to type 'daying'" is an oxymoron, as there is nothing simple to her mental process of converting a Mandarin word to pinyin orthography, which she never learned. It's far easier for me, who learned pinyin orthography from the start — and in fact, it's not uncommon for her to ask me "How do you spell [in pinyin] X"? Or to type an English phrase into Google Translate to get it to spit out the characters she wants, and then copy and paste them into her document.

  23. Soris said,

    December 19, 2012 @ 4:39 pm

    I know that the reason I have clung to handwriting input systems, as a student of Japanese, is that there are times when I wish to type a word that I can read, but cannot remember the correct pronunciation of. This is because I have learned the written and spoken languages in tandem; I imagine it is less likely to be a problem for someone who spoke the language fluently before they learned to read a single character.

  24. Tong said,

    December 19, 2012 @ 4:53 pm

    if you don't really know for absolutely certain how to write it, the computer will not be able to help you.

    I'm a proficient user of the stroke-based input method WuBi (5 strokes) and obviously wubi keyboards nowadays (e.g., BaiduInput on Android) are smart enough to not only allow errors in certain keystrokes, give suggestions to words that can be typed out by replacing with a neighboring key on the keyboard, but also have language models to help guess and promote the most likely character or phrase given the typed history.
    Of course you still have to remember something to begin with but given that each character/phrase is typed out by a max of 4 keystrokes, even one correct correction is hard and impressive enough to have corrected at least 25% of errors!

  25. joanne salton said,

    December 19, 2012 @ 6:28 pm

    Well, if I may attempt to bring some synthesis to the controversy I have stirred up, the truth is of course somewhere in the middle and forgetting a spelling is somewhat less important that forgetting a character. I do think that it is important to remember that it often may not really be a great problem to forget small things though – the people here are after all speaking from experience – and that recognition is much, much easier than writing.

    Pinyin though, as others have noted, is also a great problem for these feckless young character amnesiacs! Many more mistakes are made with that, in my experience – Westerners are quite often consulted for advice. Watching the Chinese write Pinyin is often reminiscent of watching English beginners struggling with the "phonetic" alphabet used in an English dictionary. No doubt these skills would improve with more usage, but in the end we process words much like characters once proficient with them. Thxt xs why yxx cxn undxrstxnd thxs kxnd xf thxng quxtx wxll.

  26. Victor Mair said,

    December 19, 2012 @ 6:32 pm

    @John C.

    "Perhaps that's because they learned bopomofo rather than pinyin…"

    True!

    "…perhaps it's because they're in their 60s and prefer more traditional mechanisms…."

    True! Naturally, such folks are in a distinct minority in the overall population.

    "It's also possible that input via strokes is objectively slower…."

    True!

    "…then copy and paste them into her document…."

    I've met many people like your wife. It really is a psychological block / bar, as you indicate when you talk about "mental process". I have plenty of dear friends from Taiwan who have lived in America and are fluent (more or less) in English, so they know how the alphabet works and use it to write things everyday. But they can't get their brains to wrap around Mandarin written in Pinyin, although if I sit down with them and very, very patiently show them how it works, they will get it — but still they don't want to get it.

    MY wife was different. She came to America from Taiwan and somehow became an ardent advocate of Hanyu Pinyin. She used to go out on the street and hand texts written in Pinyin only to total strangers, then ask them to read them out. When what they read off could be understood by fluent speakers of Mandarin, she would be triumphant and say, "See, I told you it is easy to learn to read Mandarin!!"

    @Soris

    "I imagine it is less likely to be a problem for someone who spoke the language fluently before they learned to read a single character."

    Indeed! And that is why, when I began learning Mandarin, I instinctively did NOT want to see a single character until I had become fluent in the spoken language. I wanted to learn Chinese the natural way a Chinese does: master the spoken language, then move on to the written — it is much, much easier that way! Unfortunately, Mandarin pedagogy in the West overemphasizes the characters from an early stage, which causes all sorts of problems in learning grammar, syntax, and practically all other aspects of language acquisition, and even adversely affects learning to read and write.

    This is why I consider Eleanor Jorden to have been the best theoretician and practitioner of East Asian language pedagogy, for she wouldn't let her students look at characters until they became really good in Japanese. Some of the very best teachers of Chinese in America — including Ron Walton, Galal Walker, and Neil Kubler — were trained under Eleanor Jorden.

    @Tong

    "I'm a proficient user of the stroke-based input method WuBi (5 strokes)…."

    You belong to a small minority of the Chinese population, especially if you are not a professional typist.

    I will cover all of these things and more in the long post on Chinese inputting methods that I am preparing.

  27. Eric Vinyl said,

    December 19, 2012 @ 6:44 pm

    Victor, it seems as if you're being unnecessarily obtuse.

    Professor Mair is one of my heroes, he can be a little bit nerdview—not exactly in the sense as it’s used here on LL, but something I recognize in a few of my friends and now familiar to much of America via Sheldon on “Big Bang Theory.”
    It’s difficult for me to analyze and articulate but it goes something like: “This situation is inelegant, not ideal, or irrational, so I cannot understand it. Once my rigorous predetermined conditions are not satisfied, this way lies chaos! and who can comprehend such senselessness.”

    Most normal folks who don’t really care about the ideal or most efficient solution (and therefore are arguably responsible for Why Everything Is a Mess These Days) and just want to get along seem to have no such, or much less, emotional attachment to disproving its utility.

  28. Matt said,

    December 19, 2012 @ 7:42 pm

    "Obtuse" is a recognized synonym for "stupid".

    Yes, but "being deliberately obtuse" is a fairly common expression for "pretending not to understand something (for rhetorical effect) even though you really do", and "being unnecessarily obtuse" is quite clearly a variant with the same meaning. It's obvious that John Q was not calling you stupid — he was saying that John Meissner's point is quite obvious, but that "it seem[ed]" as if you were pretending not to understand it because it undermines your argument. And, frankly — and I say this as someone with a great deal of respect for you and your work — I agree.

    I don't have much cause to input Chinese text, but I can easily believe that typing pinyin is the best entry system by far, much better than handwritten entry. (The same is true of Japanese, and even alphabetic languages for most people.) But that's not the issue that John Meissner was responding to. He was responding to your claim that "if you don't really know for absolutely certain how to write it, the computer will not be able to help you". This is demonstrably not true — it is an exaggeration at best — and it remains untrue even though pinyin input is superior (faster, etc. as you point out).

    I think you do your position on the Chinese writing system, pinyin, etc. a grave disservice with this kind of extremism. On the one hand, you are asking people to put their prejudices aside, look at the facts, and reconsider whether the things they have been taught all their life are really true. But on the other hand, you let your own prejudices seduce you into describing hanzi and hanzi-related phenomena in such an exaggeratedly negative way that those particular statements are easily refutable as untruths, even though your overall arguments about hanzi as a writing system are sound.

  29. John Meissner said,

    December 19, 2012 @ 8:06 pm

    @Victor Mair

    The "half a second" was in reply to your question about long it took to make the doodle, i.e. the parts where I didn't actually need to take care to write anything sensible.

    Perhaps "doodle" was not the best choice of word. "Move the fingertip back and forth across the screen without aiming to write anything specific" is more what I was looking for. With this definition, I hope you can see that it really can be achieved in half a second.

    I never claimed that the whole process from starting to write to having selected the right character took half a second, so I feel like you are putting words in my mouth. Nonetheless, It still didn't take more than maybe two or three seconds, and as I have already said, I think this is comparable to the amount of time it takes to type "innocous" in English and having Google correct you.

    As for your specific questions:

    1. I think I entered the first 5 strokes fairly clearly, had probably a hint of the first stroke and second strokes in 月 in there, and the rest was not anything specific. But the point is that the handwriting recognition really is good enough that you don't have to do anything near perfect strokes; even rather sloppy strokes will usually be fine, so you can write quickly.

    As for "squinting" to see the difference between 赢, 蠃, and 羸: I never felt this was an issue. The font used in the IME was better for this purpose than the one used on this blog; presumably the designer of the IME have put some thought into that. I think recognition of the character was more or less instant, just like you would instantly recognize a friend's face.

    2. I don't know how 癔 got into there. But I think this is the point. The recognition system uses "fuzzy" algorithms, so that you can write very sloppily and even incorrectly, and the algorithm will still come up with a decent best guess. A consequence of the fuzziness is however that it will make guesses that when written out in neat and well-proportioned manner look very different from the intended character.

    3. Again, I never said I did the whole thing in half a second. But again, I should point out that with these systems, strokes can be joined together, crooked, or missing altogether, and it doesn't matter too much. I do think it's possibly to write a rough version of 亡口月 in less than a second.

    4. As said, "doodle" was a poor choice of words on my part. I apologize for the misunderstanding.

    Finally, I don't know if the second-to-last paragraph of your post was addressed to me or not, but I would like to point out that I don't think the handwriting thing is faster than pinyin input. At least not for me. I was merely arguing that your initial claim

    If you are trying to write a character by hand or enter its shape into a computer via components or stylus / fingertip, if you don't really know for absolutely certain how to write it, the computer will not be able to help you.

    is false, as anyone with access to a modern smartphone can easily check for themselves.

    But I would also like to point out that one should make a distinction between computers and smartphones. On a computer, or anything with a proper keyboard, I quite agree with you. Pinyin input is going to be much faster and much more convenient. On a phone with a small touchscreen, the difference will be smaller, the problem being that pinyin input on a touchscreen is also kind of awkward.

    handwriting recognition comes close to pinyin input. One needs to bear in mind that inputting pinyin is also quite awkward, when all you have is a smart.

  30. wren ng thornton said,

    December 20, 2012 @ 4:55 am

    It is ironic that it is a few foreign students who are clinging most tenaciously to handwriting input systems, but there is a reason for that too, and I will explain what it is in my forthcoming post as well.

    Again speaking of Japanese rather than Chinese, I fully admit to using wapuro romaji as my preferred entry method (never having learned to use kana keyboards). Handwriting is slow, whether it's on a computer or on paper. However, my point still stands. The claim you're making is that when one cannot remember exactly how to write a kanji, this is a crippling problem that cannot be surmounted. While I do not deny that logographic systems are much harder to master, your portrayal of the situation is so exaggerated that it makes even bystanders question the realism of your claims. Without denying your years of experience on the matter, the fact that decent quality handwriting input systems do in fact exist for Japanese (as well as Chinese, according to this thread) means that your claim, "if you don't really know for absolutely certain how to write it, the computer will not be able to help you" is false.

    In addition to the empirical evidence already provided, I'll add some more. You've questioned how it is that one can choose between three very similar looking characters. On OSX when typing things in and multiple options are presented due to homophony, the popup not only presents the choices available, but also presents quick definitions if you hover over that option. Naturally, if you are fluent in the language, you need not even read the definition per se; a quick glance is usually enough to have a word or two pop-out to know what ballpark you're in. I'll fully admit that input methods for Japanese in the late nineties where atrocious and lacked all these modern niceties; but, again, this modern technology has been on the market since at least 2005. If you were arguing that typing in Chinese is difficult, or cumbersome, or slow, or any number of other complaints, then I'd have no particular reason for disagreeing with you. But your argument that computers cannot and do not assist in typing logographic languages is demonstrably incorrect.

  31. Victor Mair said,

    December 22, 2012 @ 8:56 am

    I am grateful to Eric Vinyl for understanding the point of my post.

    @ John Meissner

    "Perhaps 'doodle' was not the best choice of word. 'Move the fingertip back and forth across the screen without aiming to write anything specific' is more what I was looking for."

    This directly contradicts the following admission you make a few paragraphs below in the same comment:

    "I think I entered the first 5 strokes fairly clearly, had probably a hint of the first stroke and second strokes in 月 in there, and the rest was not anything specific."

    Here you are being specific in the strokes you enter, not aimless. Note that you had to know accurately a considerable amount of information about the construction and configuration of the complicated character for which you were searching. If you had not known the beginning portion of the character that you entered into your device, your search would have come up empty.

    Furthermore, to do all that, plus choose from among the possibilities offered by your device, is going to take you more than the "maybe two or three seconds" you've now expanded to.

    I don't know about you, but I can type "innocuous" — or one of its possible misspellings — within a second or two, and Google corrects me instantaneously — while I am typing. This is the point I was trying to get across. Chinese shape-based inputting systems of the sort you use cannot do that.

    As for Matt (not my graduate student Matt!), wren ng thornton, and John Q who think that I am being obtuse and speaking falsehoods, I shall try to redeem myself by presenting a mass of anecdotal and empirical evidence that demonstrates how native speakers and the best foreign students of Chinese actually use various types of inputting schemes in daily life. In fact, the reason I didn't respond earlier was because I was soliciting this information from hundreds of contacts in China, Taiwan, Hong Kong, Singapore, and elsewhere around the world — plus grading exams and papers. I hope to make a long, detailed post on this subject before December 25.

  32. Jacob said,

    December 24, 2012 @ 1:35 am

    Not entirely related, but the act of 'tricking' pinyin or handwriting IME systems into finding the character you want, but don't know how to write reminds me of the opening of an article I read long ago.

    "Imagine this as a common telephone conversation between two well-educated native speakers of English:
    A: Yes, you can just send the letter along to Mr. John Smith
    B: John? As in John Donne?
    A: Yes, and Smith as in blacksmith, that Smith
    B: OK. And your address?
    A: Number 1, Lane 13, Isle City Street.
    B: [ail]? What [ail]? Church aisle?
    A: No, Isle as in the Isle of Wight.
    B: [ail]? of [wait]? As in White House?
    A: Heavens no, THE ISLE OF WIGHT, the place name, as in England.
    B: Hmmmm, I am not clear about that. . .
    (At this point the first speaker turns to another person in the room and asks how he should contextualize the word 'isle', to which is replied 'Isle of Wight.' The speaker says he's already tried that but the person on the other end of the line doesn't know the term. This is met with raised eyebrows, so he continues)
    A: This 'Isle' has two straight standing lines, one at the beginning and one in the middle, and on each side of the second one there are shorter curvery lines.
    B: Oh! THAT isle, as in island. And [siti] as in New York City?
    C: (in an exasperated voice) NO, as in Buffalo City, Arizona, that city."

  33. Jacob said,

    December 24, 2012 @ 1:51 am

    And – especially on my phone – I will use whatever system to get at the character I need. I start with pinyin, then sometimes need to do something along the lines of Prof Mair's example of 打赢 (though the first pairing that comes up is 答应), and sometimes have to resort to handwriting recognition.

    Though I do understand the arguments against John Meissner, it does bring to mind the phrase '看不懂就看一半'. Could it be that it's possible – in limited circumstances – to extrapolate this to '写不出就写一半'?

  34. Dave Cragin said,

    January 3, 2013 @ 8:48 pm

    Like Victor, I started learning Chinese via speaking only (using Pimsleur) and I agree it's the best way to start.

    His comment on using google for spelling made me want to offer an idea that could be useful to those learning characters: When I first started writing notes to colleagues in China, one asked "I don't mean to be offensive – but are you using google translator to write the notes?"

    I wasn't, but her comment gave me the idea of using google translator as a "quasi-" spell checker to avoid some of the "odd mistakes" Joanne mentioned, that is, after I write an e-mail in Chinese, I have google translate it into English.

    If the translation looks really weird, I may have picked the wrong character, forgotten a de 的 or le 了 or made some other error. It's obviously not perfect and you have to know enough to judge the quality of its translation, but it's the best way to "proofread" my Chinese I know of (and it's actually a good teaching tool because it makes me examine what I wrote – as opposed to just sending a note that was wrong).

  35. Mike said,

    November 20, 2013 @ 6:20 am

    I'm puzzling why Google translator displays miswritten characters when translating from Czech to English. "Z důvodu netěsnosti potrubí je přerušena dodávka tepelné energie do oblasti Prahy 3 a Prahy 10." = "The dà …  ¯ water Neta "â € º leaks potrubàthe PA … â" ¢ eruà …  ¡¡ena dodàvka tepelnà© energies in Prague 3 and Prague 10." That's strange, isn't it? With a translation from Czech to German it's the same.

RSS feed for comments on this post

Leave a Comment