Chinese characters and eyesight

« previous post | next post »

There was an interesting article in the Economist a couple of day ago:  "Why So Many Chinese Children Wear Glasses" (11/9/2014)

Myopia is epidemic in China, and the percentage of those with this affliction is increasing each year.

Here are a few passages from the article that provide food for thought:

The fastest increase is among primary school children, over 40% of whom are short-sighted, double the rate in 2000. That compares with less than 10% of this age group in America or Germany.

The incidence of myopia is high across East Asia, afflicting 80-90% of urban 18-year-olds in Singapore, South Korea and Taiwan. The problem is social rather than genetic. A 2012 study of 15,000 children in the Beijing area found that poor sight was significantly associated with more time spent studying, reading or using electronic devices–along with less time spent outdoors….

The biggest factor in short-sightedness is a lack of time spent outdoors. Exposure to daylight helps the retina to release a chemical that slows down an increase in the eye's axial length, which is what most often causes myopia.

A combination of not being outdoors and doing lots of work focusing up close (like writing characters or reading) worsens the problem. But if a child has enough time in the open, they can study all they like and their eyesight should not suffer, says Ian Morgan of Australian National University

Yet China and many other East Asian countries do not prize time outdoors. At the age of six, children in China and Australia have similar rates of myopia. Once they start school, Chinese children spend about an hour a day outside, compared with three or four hours for Australian ones.

(VHM: emphasis added)

A similar article appeared in the LA Times of 7/5/12:  "China's myopia epidemic comes into focus"

"As many as 90% of urban youth in China are nearsighted. Researchers think they can fight myopia by forcing youngsters to put down their books and expose their eyes to natural light."

Selected passages:

By the time they complete high school, as many as 90% of urban Chinese youth are afflicted by the condition known as myopia, in which close objects can be seen clearly but things just a few feet or inches away start to blur.

That's about three times the rate among U.S. children. Even more troubling is the severity of the Chinese cases. Between 10% and 20% of nearsighted Chinese children are expected to develop "high myopia," which is largely untreatable and may lead to blindness.

Eight years earlier, a Chinese news outlet focused on the same problem:

"30 Mln Chinese Become Nearsighted Per Year" (7/11/06).

This is not an affliction that has been plaguing generations of Chinese students and adults only in recent years.  Very high levels of myopia have been a serious issue in China for a long time.  See, for instance, The Ophthalmic Year Book, Volume 18 (1922), p. 49a:

Li's experience convinces him of the necessity for thoro cycleplegia in the great majority of cases.  Of a large number of Chinese students examined, 53% were myopic, 36% hyperopic, and 11% showed mixed astigmatism.  He ascribes the high percentage of myopia to the peculiar construction of the Chinese characters, writing them being more difficult than reading.  Poor hygiene generally prevails among the students.

Rush lists myopia as one of the four chief causes of blindness in China, and states that among students he found 58% were myopic.

(VHM:  The studies of Li and Rush appeared in the National Medical Journal of China, 6 [1920] and the China Medical Journal, 34 [1920].)

Although myopia rates are lower in rural areas, even there the rate is higher in China than in Australia or the West when the children do have a chance to go to school.  In earlier posts on this subject, I commented on the strained, squinting look on the face of Chinese children who are reading and writing.

See item 4 and the photograph of a little girl accompanying it in this post, "The cost of illiteracy in China" (3/31/12), also the photograph of the little boy in this post, "Copying characters" (2/11/13).

Commenting on the photographs accompanying these two posts, I remarked how Chinese children reading and writing often have a strained look on their face.  This may due to a variety of factors, including density of strokes, dim lighting, poor printing, and so forth.

There is plenty of anecdotal evidence that would seem to indicate a connection between myopia and Chinese characters, e.g.:

"The reason why so many Chinese have bad eyesight?" (discussion started 12/24/11)

"Reading Chinese characters contributes to myopia" (4/21/09)

Of course, such anecdotal evidence is necessarily inconclusive of a link between myopia and Chinese characters.  In searching for scientific studies on the subject, I turned up a few interesting studies.

"The effect of simplified and traditional chinese character on accommodative reponses in myopes and emm[e]tropes" (8/6/12)

by D. Xu and W. Xin both from School of optometry and ophthalmology, Wenzhou medical college (Wenzhou)

also here


Purpose This study investigated the effect of simplified and traditional chinese character on accommodative lag, accommodation microfluctuation, near work-induced transient myopia (NITM) during the near task as well as the subsequent decay time. We aimed to analyze if the traditional Chinese character had the trend of inducing the myopia progress.

Methods Young adults were classified into emmetropes and myopes . Two size (12pt,8pt) and two type (simplified and traditional chinese character) were used to give four different reading targets on computer screen. It was presented in random order at 25 cm and was read for 5min. For each target, accommodation response , accommodation microfluctuation NITM and its decay were measured using the free space Grand-Seiko auto-refractor.

Results For myopic subjects, accommodation microfluctuation was greater for traditional chinese character 0.35±0.17 D than for simplified chinese character: 0.29 ±0.11 D, (p=0.017) ,NITM was larger for traditional chinese traditional character 0.61±0.34 D than for simplified chinese character 0.53±0.33 D,p=0.028, however there was no difference in emmetropic subjects. Myopes had larger accommodation microfluctuation 0.35±0.17 D than emmetrope 0.24±0.11 D p=0.02,myopes had greater NITM0.61±0.34 D than emmetropes 0.42±0.25 D P=0.013. For all targets, The accommodative lag of myopes is 0.74±0.32 D while the emmetropes 0.53±0.35 D P=0.000 ;The decay time for myopes was15.88±19.53 seconds while the emmetropes 9.04±12.62 seconds P=0.012.

Conclusion The study showed that compare to the simplified Chinese character, the traditional chinese character had significant effect on myopic subjects, it may be more susceptible of inducing the myopia progress.

This study suggests that, since the density of strokes of simplified characters is less than that of traditional characters, the rate of NITM is greater for the latter.

On the other hand, the following study found no significant difference in accommodative responses between traditional and simplified characters, though both resulted in NITM (nearwork-induced transient myopia) after 10 minutes of reading.

"Accommodative response and nearwork-induced transient myopia induced by simplified and traditional Chinese characters"

Ye YANG;  Jin-hua BAO;  Jing-jing XU;  Chen-xiao WANG;  Fan L(U)

Chinese Journal of Experimental Ophthalmology 2012; (12) : 1100-1103

Background Epidemiologic studies found that the incidence of myopia is higher in Hong Kong and Taiwan regions of China than that of the mainland. So whether the general reading words with traditional Chinese characters and simplified characters is associated with myopia deserves attention.  Objective This study was to test the accommodative responses and the regression levels of nearwork-induced transient myopia (NITM) induced by traditional complex Chinese characters and modern simplified ones in the same size, and to explore the differences and inherent relationship of the accommodative regulations based on the structures of the two types of Chinese character in causing reading triggered myopia.  Methods Twenty two volunteers aged 24-29 years were included in this study with informed consent.  The corrected vision of both eyes from each subject was ≥ 1.0 with a mean spherical equivalence of (-1.86±2.34)D.  Accommodative response was tested with 4 different reading texts using the rapid sequence visual presenting model with the GRAND SEIKO-WV5500 infrared autorefractor, and this procedure was performed after full correction of refractive error.  An initial test of looking at a certain distance was performed (as baseline), and then the subjects read intensively at the targets for 10 minutes at 33 cm to calculate the accommodative responses.  After a 10 second pause, the ocular refractive status was obtained exactly at 15 seconds, 20 seconds.  The one-way ANOVA method was used to determine the effects of the different font types and sizes on the adjustive responses and the causation of NITM.  Results Accommodative response induced by simplified and traditional Chinese characters showed an accommodative lag of (1.11 ±0.38),(0.95 ±0.43),(1.18 ±0.33) and (1.06±0.28) D,showing a significant difference among the 9 pt and 12 pt simplified and traditional Chinese characters (F =1.62,P =0.19), and significantly different accommodative lag values between 12 pt simplified characters and 9 pt traditional characters was found (t =5.56,P =0.02).  NITM induced by the four different targets were (-0.45 ±0.45), (-0.47 ±0.46), (0.45 ±0.82) and (-0.46±0.78) D in the 4 types of characters, without a significant difference among them (F=0.01,P =0.99).  Conclusions Near-distance reading causes accommodative lag regardless of the type of reading texts.The target demonstrated stimuli spatial frequency and font size play an impact on accommodative responses.  NITM appears when one reads simplified or traditional Chinese for 10 minutes. The accommodative lag and NITM trend might be responsible for the onset or regression of myopia, yet it is not supportive for the hypothesis that reading traditional Chinese causes more strain since there is no difference between the two.

(VHM: emphasis added)

The last two cited studies only indicate that both simplified and traditional Chinese characters cause NITM, but that there is no significant difference between them.

The biggest breakthrough in my research on the subject of the relationship between Chinese characters and eyestrain came from an unexpected source, this compelling article in Mutant Palm:

"Widespread Myopia and the Chinese Language"

Among other insightful observations and valuable data proffered, Dave Lyons (the guy behind Mutant Palm) states:

In one study, rates of myopia in Chinese peasants were found to be around 5%, while scholars had nearly 85%. A study in Taiwan found that older Chinese people and older white people had comparable rates of myopia, but younger Chinese had it far more than younger whites. All of this suggests that with increased literacy and reading in China comes myopia.

As Lyons remarks, written Chinese clearly involves more nearwork than written English.  He cites a recent post by Chinese blogger Hecaitou that provides a compelling example of this.  It comes from an “Alzheimer’s Test” that has three questions like this one:

This is a REAL neurological test. Sit comfortably and feel calm.

Find the C below. Do not use any cursor help.


I found the "C" within one or two seconds, and most of the commenters easily found the C.

Now look for the yǐ 已 ("already") amidst all the jǐ 己 ("self; sixth heavenly stem") in the following sample (I won't trouble you with throwing in an occasional sì 巳 ["sixth earthly branch"]):

己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己 己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己 己己己己己己己己己己己己己己己己己已己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己 己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己 己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己 己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己 己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己己

Not only is this a myopia inducing exercise, it is also apt to bring about cephalalgia.

Commenter EGG suggests a pair of characters with more strokes:  gǎo 搞 ("do; make; care; get; start; get hold of; participate; go in; produce; set up")  and gǎo 稿 ("draft; manuscript; sketch; stalk of grain; straw") — with 13 and 15 strokes respectively, 搞 and 稿 have just above the average number of strokes for a Chinese character, which is around 12.

稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿 稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿 稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿搞稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿 稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿 稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿稿

You will be forgiven if you were not so persistent as to find the 已 and the 搞 among the masses of 己s and 稿s.

I love the way Lyons ends his article:  "Perhaps the incidence of myopia in China would be reduced if text were segmented. I wonder what John DeFrancis would say."  I'm sure that my old friend JDF would have heartily applauded such a move.

Just before the concluding section of his article, however, Lyons introduces us to Gord Hotchkiss's fascinating piece entitled "Chinese Eye Tracking Study: Baidu Vs Google" (6/15/07), which presents striking heat map comparisons.

Judging from eye scans and mouse clicks, Baidu searches are much more scattered, inefficient, and relatively unsuccessful, whereas Google searches are more patterned, efficient, and comparatively successful.

From all of the materials assembled above, it would appear that Chinese reading and writing require more nearwork and eyestrain than are needed for reading and writing with the Roman alphabet.  Chinese teachers are well aware of the burdens placed upon their students and the consequent deterioration of vision.  So they resort to all sorts of measures to alleviate the detrimental effects of reading and writing characters, such as:

"New eye exercises for Chinese students, but just as useless as before" (11/29/08) and desk railings to keep their students from getting too close to their books and papers!

The good doctor Ian Morgan of Australian National University, who is cited in some of the above articles, insists that Chinese students wouldn't have such a high incidence of myopia if only they would go outdoors more often.  That makes a lot of sense.  The questions is, though, why don't Chinese school children spend more time outdoors?  Perhaps it's because they want to master those high maintenance characters, and to do so requires writing each one of them hundreds and hundreds of time so that one can recognize them accurately and reproduce them correctly when called upon to do so in tīngxiě 听写 ("dictation" [lit., "hear-write"]) quizzes.

It's a series of vicious cycles and circles:  the further those railings keep the children from their books and papers, the harder they strain to distinguish the tiny strokes on them; the more time the children spend outside playing, the poorer they do on their character quizzes.  Somehow, a balance needs to be struck between the demands for literacy and the need for good eyesight.


  1. flow said,

    November 12, 2014 @ 8:44 am

    it must be said in all fairness that when you compare O/C vs 己/已, the difference within the latter pair is rather smaller than within the first; I/l/1 vs 己/已/巳 looks fairer, or else O/C vs 己/巳.

    there are more characters in Chinese that resemble each other simply because there are so many of them. because of the potential internal complexity of characters, characters should really be much larger than Latin script. To obtain a fair ratio, one could compare the space taken up by close translations of texts written in, say, Chinese and English, German, or French. in normal printing practice, where the space taken up by a CJK character equals approximately the footprint of a W, Chinese texts are regularly much shorter than their English counterparts, although English is known to have rather shorter words than, say, French (true? i believe it is). one could then enlarge the characters so mean printed areas of comparable texts become more comparable. just as a rough guide.

    as anecdotal evidence, i remember driving down Taiwan highways years ago and trying to read those road signs written in parallel CJK/Latin. i've been blessed with book myopia for many years, so even back then it was *always* the latin i could read earlier. and yes, differences in fluency with the two scripts must have played a role, but then the features of the Latin letters became apparent when the Chinese 方塊字 were still blobs in my perception. there's so much more packed into each single one of, and that detail needs room.

    as another anecdotal evidence, last week i was shown a product battery-disposal advisory blurb originating from IKEA, printed in dozens of languages on a single sheet, including Chinese for the Mainland, Chinese for Taiwan, Japanese, Korean, and German, French etcpp. of course, the East Asian languages were printed in type the same nominal size as the Latin script languages, in other words: definitely too small!

    my general impression is that printed matter in modern times show a tendency to use ever smaller type. i have one 《新华写字字典》 here, a work intended for use by elementary school students. in the main entries, explanatory remarks are scaled down to a type size of 1.95mm!!!; in accompanying 'side stories', the size is 2.15mm, but to make things less readable, they use a sans-serif font and a backdrop with red pixels. the introductory remarks have a size of 3.6mm (around 10pt, which is sort of small even for Latin), which i feel is about the minimum size for comfortable reading (still no big print). maybe someone has more 'traditional' books etc. lying around to comment how type used to be sized?

    in conclusion, it is not so much the script that is too complicated (although it is), but the printers who don't give the script the room it needs. if more materials were printed at a size that is easier for the eyes, probably less people would suffer from this kind of myopia. that said, reading books is always a strain on the eyes, and so is the monitor and the mobile device. when i look up from someone's smartphone, i often feel like i've been ZING! flashed. not good.

  2. Ralph Hickok said,

    November 12, 2014 @ 9:15 am

    I have been told, more than once, by opthalmologists that reading doesn't damage the eyes, and this list of myths and facts seems to bear that out:

    Research led by Kathryn A. Rose seems to establish that near work is negligible as a cause of myopia and that the problem is actually the lack of sunlight:

  3. Victor Mair said,

    November 12, 2014 @ 9:34 am


    Thank you very much for your astute remarks. One important thing I think you've established clearly is that, for comparable ease of reading, Chinese characters need to be printed at a significantly larger size than Roman letters. This also speaks to the oft-asserted claim that Chinese characters convey more information than Roman letters for the same amount of space.

    As for premodern printing, it was traditionally done with woodblocks. The nature of the medium required that characters be much larger than is the case with typical contemporary typesetting.

    I remember an American typesetter of Chinese who was inordinately proud of his technical ability to produce characters so tiny that the strokes looked like mosquito legs or the threads of a spider web. But there were two major problems with such miniscule typesetting: 1. few printers could produce copies that retained the clarity of what the typesetter could create in his computer (I believe it may have been over a thousand dpi); 2. even if the printing could capture most details of the spidery, web-like characters, few human beings could read them with the naked eye.

    One of the many reasons I decided to create the first ABC dictionary of Chinese was because, back in the 80s when I was teaching both 3rd-year and 4th-year Mandarin, I was upset that my students were going blind by having to read the microscopic print of the little Oxford Chinese-English dictionary (the narrow one with the flexible red cover), which was about the only thing that was available in those days. John DeFrancis and I spent a lot of time doing legibility studies with different sizes and fonts, and I can tell you that John was very mathematical and scientific about it. In the first edition of the ABC dictionary, we quite consciously made the Chinese type larger than the surrounding English type so that the reader would not have to strain to differentiate among the characters. Although it necessitated a lot of extra work for us and for our typesetter, we think this innovation was worth the effort, and I suspect that you would approve of it.

    Concerning the differences among 已, 己, and 巳, there are lots of pairs and larger groups of characters that are distinguished by small details, but such is not the case with the letters of the alphabet. I couldn't agree more with your statement that "there are more characters in Chinese that resemble each other simply because there are so many of them."

  4. J. W. Brewer said,

    November 12, 2014 @ 10:37 am

    What conclusions are we to draw from the claim that urban youth in South Korea have equally high rates of myopia as those in HK and Taiwan? Does hangul cause comparable problems? (To the extent there is still *some* semi-vestigial degree of instruction in reading/writing hanja in the ROK schools, it's got to be dramatically less in terms of hours-per-week than the time devoted to hanzi in places where the primary language of instruction is Sinitic.)

  5. DCA said,

    November 12, 2014 @ 11:16 am

    There is an interesting discussion of high East Asian myopia rates in Alan McFarlane's "The Glass Bathyscaphe".–he also references a rate of 80% amongst Orthodox Jewish males.

  6. flow said,

    November 12, 2014 @ 11:23 am

    @VHM "The nature of the medium required that characters be much larger than is the case with typical contemporary typesetting."—certainly true, given that the glyph shapes had to be manually chiseled from a woodblock. then again, there speaks some pride out of the remarks of 沈括 Shen Kuo about 畢昇 Bi Sheng's printing efforts (around 1050): those glyph shapes were 薄如錢唇 "slim as a coin's rim". also, let's keep in mind the Chinese are famous the world over for carving whole little worlds out of a cherry pit, so maybe the really limiting factor was the lack of spectacles.

  7. un malpaso said,

    November 12, 2014 @ 12:21 pm

    I found the C among the O's in less than a second. I gave up on the 己 exercise after about a minute and a half of searching. But weirdly, I found the 搞 in just a couple of seconds! Probably because I was specifically looking at the left side of the characters for the missing stroke.

  8. Curt said,

    November 12, 2014 @ 1:07 pm

    It's interesting to know why China didn't proceed with further simplifications and improvements to the current system.
    I would think that the main drawback of phonetic spelling is that it requires people to know prescribed pronunciation of every word.
    I wonder if it's true or just another myth:
    In those days the main form of contact with standard language for chinese population was through print and ink/paper. So even if people learn correct pronunciation ( and tone ) of all words that they already know and use every day ( just like foreign language ), they may forget it later or not be able to reproduce it correctly. And when the only audio source of correct pronunciation is radio ( no subtitles ) it's not reinforcing. Characters helped. So replacing characters with pinyin wasn't practical and party decided to simplify and eradicate complexity as much as possible.

    But these days with IME you must know correct pronunciation; television and mandarin popculture on the internet instead of radio and print and traditional folk culture. I wonder do chinese children all over China think in Putonghua phonetically close enough to prescribed standard?

  9. Mara K said,

    November 12, 2014 @ 1:21 pm

    @JW: In the case of South Koreans, the balance of factors causing myopia probably leans more toward "time spent inside studying."

  10. David L said,

    November 12, 2014 @ 1:22 pm

    The idea that a lot of close-work or reading by children will lead to nearsightedness is dubious at best. The strongest predictor is heredity (if your parents have myopia you probably will too), but this page and this one will only go as far as saying that close-work may worsen vision problems in children who are already disposed toward them. The NIH says only that myopia tends to develop in childhood and the teenage years — which can make it appear as if it's a result of some behavior.

    In any case, although it may be that kids today are spending all their time peering at their smartphones, it's also true that in many countries, not so long ago, children were put to work at an early age, and girls in particular may have spent a lot of time doing needlework and similar jobs.

  11. Jonathan Badger said,

    November 12, 2014 @ 1:23 pm

    "The questions is, though, why don't Chinese school children spend more time outdoors? Perhaps it's because they want to master those high maintenance characters"

    Perhaps. Or if they live in one of large cities it may be because breathing the polluted air isn't very fun…

  12. David L said,

    November 12, 2014 @ 1:42 pm

    @Ralph Hickock: Sorry, I didn't see your post earlier when I wrote mine

  13. Stephen said,

    November 12, 2014 @ 2:40 pm

    I couldn't find the C among the O's at all. I gave up after about a minute. Ought I to be concerned? I couldn't find the 已 in the 己s either. It was reassuring that I spotted the 搞 in 稿s in less than a second.

  14. Richard W said,

    November 12, 2014 @ 3:18 pm

    When I started learning Japanese, I found myself a language exchange partner — a Japanese woman who worked as a typist in the Japanese department at a university. I'm talking about a Japanese typewriter, one with hundreds of characters, like this:

    (This was in the 80s, before they replaced typewriters with word processors.)

    She eventually left her job acrimoniously after failing to get recognition that her headaches and deteriorating eyesight were caused by eyestrain resulting from the nature of her work.

  15. J. W. Brewer said,

    November 12, 2014 @ 3:44 pm

    I had never previously considered this angle, but when my family temporarily moved from the U.S. to Tokyo (when I was 8 years old) I had not yet been diagnosed with myopia, yet within a year or so after I first started rudimentary study of kanji at the American School I received such a diagnosis and got my first pair of eyeglasses from a Japanese optician. And here I am still wearing glasses many decades later.

  16. hector said,

    November 12, 2014 @ 3:50 pm


    "it is not so much the script that is too complicated (although it is), but the printers who don't give the script the room it needs."

    — don't blame the printers. Printers will give you as much space as you need, if you're willing to pay for the larger size of the book that will be produced as a result of the larger script sizes.

  17. Eidolon said,

    November 12, 2014 @ 7:16 pm

    I think we need to have comparative studies in other East Asian countries – especially Korea, which uses a completely different writing system – to draw a solid conclusion. I reckon @Brewer, in case the data is correct, that the bigger issue is the much larger amount of time East Asian children spend studying indoors, especially in light of what @Ralph wrote. Although, the other argument is that hangeul is itself difficult to read, despite being an alphabet. The way it's organized into blocks, and the fact that a lot of its characters have similar shapes, open us up to the same criticism.

    Consider the following snippet I shamelessly took from a web search:

    나라 = Nara, ナラ(奈良, a city in Japan)(In Korean, 나라 means a "country".)
    나가 = naga (Get out!).

  18. Adrian said,

    November 12, 2014 @ 7:41 pm

    I've had two tries at each of the Chinese character searches without success. I have to give up after about 30 seconds because it soon becomes quite painful.

  19. Lazar said,

    November 12, 2014 @ 9:10 pm

    @flow: The thing about I/l/1, though, is that very little informational load is hinging on their looking different. You can almost always tell which character is intended from a variety of clues – whether it's at the beginning of a word, whether that word is at the beginning of a sentence, whole-word recognition, sentence context, whether it's accompaneid by numbers, etc. Some fonts (like Arial) use the same glyph for the first two, and many people write all three identically.

  20. hwu said,

    November 12, 2014 @ 10:04 pm

    I have to say it’s so sad that even Chinese children already spend so much time on reading/writing Chinese, the character amnesia still seems inevitable. Although Chinese government is taking actions (such as “Chinese Character Dictation Competition” or “Chinese Characters Hero”) to help addressing this issue, those actions seems only focus on middle school students. Forcing 20 million junior school students to participate in CCDC (yes, this is what the government is doing this year) is totally not a solution for character amnesia, because those students only participate in the competition once in their early middle school years, while character amnesia is particularly prevalent among white collars who use computers heavily. Preparing for the competition will only bring more workload and risk of myopia for those students, and in the future when they become university graduates, they may still have character amnesia.

  21. JS said,

    November 12, 2014 @ 10:14 pm

    Same is true of 已己; they could be (and sometimes are) written identically and it would scarcely matter thanks to context clues. There are a million plus google hits for "己经", for instance…

  22. hwu said,

    November 12, 2014 @ 11:18 pm

    By the way, an interesting solution for this “Alzheimer’s Test” is to scroll the browser page up and down quickly & repeatedly. This method works best for IE browser.
    More examples can be found at

  23. Suburbanbanshee said,

    November 12, 2014 @ 11:53 pm

    If kids don't have time to play outside, maybe they have time to practice reading and writing outside. Maybe with big fat pieces of chalk and a nice big parking lot.

    And in the winter, you could have your lunchroom or the hallways painted in blackboard paint, and kids could practice somewhere sunny.

  24. Movenon said,

    November 13, 2014 @ 1:38 am

    What about Vietnamese people? They seem to have a lot of nearsightedness, despite their Roman script.

  25. Stephen Benfey said,

    November 13, 2014 @ 2:35 am

    In some US states (e.g., North Carolina) you can take the driver's license exam "orally," that is you don't have to read the test, the examiner will read the questions to you.

    If the USA took the effort to achieve the literacy levels of China, Japan and Korea, I expect that the incidence of myopia would rise proportionately.

    "32 million adults in the U.S. can't read. That's 14 percent of the population. 21 percent of adults in the U.S. read below a 5th grade level, and 19 percent of high school graduates can't read."

  26. Chas Belov said,

    November 13, 2014 @ 2:54 am

    I got the C virtually instantly, maybe less than a quarter of a second. I still haven't gotten either of the Chinese tests.

  27. maidhc said,

    November 13, 2014 @ 5:12 am

    I developed myopia as a teenager. I don't really think it had anything to do with sunlight though–I got plenty of sunlight. More likely heredity as both my parents wore glasses.

    One advantage is that when you get old you can see things close up without reading glasses. Now that I've moved into that age group I appreciate that. Whenever I want to look at tiny print I just take my glasses off.

  28. Zeppelin said,

    November 13, 2014 @ 7:23 am

    I don't know how Firefox handles chinese characters, but the way they're being displayed for me the 搞 is very easy to find, sticking out much more than the C for example (the characters are quite small and 搞 is much darker than 稿).
    On the other hand I'd definitely have to squint if I wanted to make out any specific features of 搞 – from a normal reading distance the, uh, main bit just appears like a blob.

  29. Victor Mair said,

    November 13, 2014 @ 12:51 pm


    "I wonder do chinese children all over China think in Putonghua phonetically close enough to prescribed standard?"

    Every child in China learns to read and write through Pinyin, which is taught according to prescribed pronunciation rules for Putonghua (Modern Standard Mandarin). The pronunciation of Putonghua throughout China has become remarkably standardized. I don't know exactly what goes on in people's brains, but, as I have pointed our many times in the past (e.g., here), you don't have to speak standard Mandarin to be proficient in Pinyin inputting.

    @Mara K

    Even though you're speculating, I'm glad to hear you speak of "balance of factors causing myopia ".

    @David L

    "The strongest predictor is heredity (if your parents have myopia you probably will too), but this page and this one will only go as far as saying that close-work may worsen vision problems in children who are already disposed toward them."

    You have misinterpreted what those linked pages actually say.

    The relevance of your second paragraph to the question at hand is unclear.

    @Jonathan Badger

    I know that you were attempting to be funny, but what's keeping the kids glued to their desks after hours is not the pollution, bad as it is.

    @Richard W

    Excellent evidence from a Japanese typist.

    For the Chinese typewriter, see here and here.

    @J. W. Brewer

    Thank you for the account of your personal experience learning kanji as a child and its apparently lasting impact.

    @hector, @flow, et al.

    If people know that small kanji / hanzi / hanja put a strain on the eyes that is greater than the strain experienced from alphabetical writing of the same size, why don't they increase the size of Chinese type?


    Excellent points about the nature and shape of hangul. This is something that William C. Hannas has written about in his books on East Asian writing.


    Sheer pain. The headache effect.


    I really appreciate the testimony of someone like yourself who comes from the Mainland about the futility of enforced dictation exams and contests in combatting character amnesia. I'm also grateful to you for the link to other examples of seeking for slightly different characters among masses of lookalikes.


    Lots of people don't know for sure what the difference is among 已, 己, and 巳. That's why I have to teach my students in Classical Chinese, most of whom come from the Mainland, Taiwan, Hong Kong, and Singapore, a mnemonic device for distinguishing them.

    Two other characters that people often confuse, especially in handwriting, and particularly when they form parts of other characters, are tǔ 土 ("earth") and shì 士 ("scholar"). My wife, who was very learned in Chinese literature, both modern and premodern, and who was one of the best Mandarin teachers ever to walk the face of the earth, was often frustrated — both for herself and for her students — about when to use 土 and when to use 士. Mind you, moreover, that these are both very simple and very common characters.

    Here's a pop quiz for everyone (find the one character that is different from all the rest):



    That's actually a good suggestion for getting kids to go outside and practice BIG CHARACTERS. Cf. this post:

    "Writ in water"


    "They seem to have a lot of nearsightedness…."

    Please be more specific.

    @Stephen Benfey

    "I expect that the incidence of myopia would rise proportionately."

    Pure surmise. Anyway, what needs to be measured is the incidence of myopia among literate individuals in East Asia versus the incidence of myopia among literate individuals in the United States.


    Were both of your parents highly literate?

    When I taught at Tunghai University from 1970-72, one of my neighbors was a very highly literate, old-fashioned Chinese scholar who had extremely thick glasses and could barely see a few feet in front of him. Because of the war (WWII), neither of his sons received a proper education, and neither of them wore glasses.

  30. brandon seah said,

    November 13, 2014 @ 5:30 pm

    Instead of desk railings, maybe the schools could throw out their old stocks of writing paper with the tiny 1 cm squares and have kids read and write in a more reasonable size! Saying this as a long-suffering myopic….

  31. maidhc said,

    November 13, 2014 @ 11:12 pm

    Victor Mair: My father was not particularly literate as a child, he was more of an athlete. After serving in the military he took the opportunity to get a university education and made himself very literate.

  32. David J. Littleboy said,

    November 14, 2014 @ 10:13 am

    "What about Vietnamese people? They seem to have a lot of nearsightedness, despite their Roman script."

    And computer nerds. At a meeting of a computer club at MIT in 1972, 25 or so nerds in attendance, someone shouted "Hey, everyone close your eyes." So we did. "Given that 60% of the population wears glasses, what's the probability everyone in the room is wearing glasses" Well, duh, we can all do that: .6^25 is, like, a really small number. But, of course, everyone was wearing glasses.

    (Our next claim to fame was telling Bill Gates to take a walk, since we worked with real computers (Multics mainframes), when he came looking for software engineers. Really. Oops.)

  33. Jim Breen said,

    November 14, 2014 @ 7:56 pm

    My first degree was in statistics, and I seethe when I see badly designed studies dishing up quite unsubstantiated results. The medical profession is a major offender (social scientists are worse.)

    BTW, I had great difficulty finding the C among the Os, but I suffer from mild AMD (Age-related Macular Degeneration), so the rows of Os kept waving around when I looked hard at them…


  34. Nanani said,

    November 15, 2014 @ 6:33 pm

    I am fluent in Japanese but not Chinese, so for the third test, it was quit easy to find the one Not-Japanese character (稿 is used in Japanese, the other is not).

    However, the 己/已 test took me rather longer… Until it occurred to me to try "reading" the characters vertically. The odd-character-out jumped out at me then!

    So, the followup question suggests itself: What are myopia rates like for people using horizontally vs vertically printed texts?

  35. Tom V said,

    November 16, 2014 @ 12:15 pm

    I had to enlarge the font (on a small laptop) just to see the difference between yi and ji in isolation, although I didn't try it with my stronger pair of reading glasses.
    It seems easier for me to read characters written with a brush than typeset ones. In part this is probably due to the larger size, but perhaps the greater variation in stroke width helps as well. This might be related to the fact that I find sans-serif fonts in English more difficult to read.

  36. Akito said,

    November 17, 2014 @ 10:05 am

    "There are a million plus google hits for "己经", for instance…"

    This is interesting. It implies a large number of people don't use a phonetic-based input system (using e.g. pinyin or bopomofo), with a less likelihood of suggesting that particular combination.

  37. Richard W said,

    November 17, 2014 @ 2:30 pm

    @Akito: It seems likely that, with a Wubi-style input, there would be a significant proportion of typos for characters so similar as 己 and 已, but there are 359 million Google hits for the correct form 已经, and only 1.5 million for 己经. I'm not sure that one can draw the conclusion that "a large number of people don't use a phonetic-based input system", especially when it's not clear how many of those errors are scanning errors rather than Wubi typos.

  38. Richard W said,

    November 17, 2014 @ 3:25 pm

    For example, under the assumptions that
    – the Google hit count is more or less correct
    – Wubi users input get 已 right 90% of the time
    – half of the 己经 errors are due to OCR failures,
    I estimate that Wubi users might account for 2% of inputters.

    (1.5 million) / (359 million) = 0.004. This is the proportion of 已经s on Google's Web that are incorrectly written as 己经.

    If half of this proportion is due to OCR failure, that leaves 0.002 as the proportion of errors due to Wubi.

    If Wubi users get 已 wrong 10% of the time, we should multiply that figure by 10 to find the proportion of Wubi users. That would give us 0.02, i.e., 2%.

    Given that Prof. Mair has to "teach [his] students in Classical Chinese, most of whom come from the Mainland, Taiwan, Hong Kong, and Singapore, a mnemonic device for distinguishing [己 and 已]", can we be confident that Wubi users even achieve 90% accuracy with 已? If their accuracy is lower than 90%, the estimated proportion of Wubi users is also lower. For example, an assumption of 80% accuracy would suggest that perhaps 1% of inputters use Wubi.

  39. Richard W said,

    November 17, 2014 @ 5:45 pm

    Just now, I wrote to a contact who is a software engineer working at a company that develops multilingual input methods for mobile devices. I asked him if he knows what proportion of inputters use shape-based methods. He replied:

    As for Wubi, I know only 1 person using it… But since you mentioned "shape-based", I think "stroke" input is quite common for Chinese people to write a character they're not sure how to pronounce or for people who are not accustomed to pinyin (too old to have gone through pinyin education in school).

    The company I work at – and which makes a predictive keyboard – has just released Chinese (in a beta version) and so I hope that soon we can get access to such statistics. We support pinyin (9key and full keyboard) and stroke input for simplified characters, bopomofo and stroke for traditional (somehow the product people, despite being Chinese citizens, believe that nobody would want to type traditional using pinyin…) and jyutpin for Cantonese.

  40. hwu said,

    November 17, 2014 @ 8:03 pm

    @Victor Mair

    Thanks. I commented here because I was surprised to know that here was the source of the english term "character amnesia".

    Maybe this is a little off topic, but I am really excited to reveal that I have developed an innovative software to help combat character amnesia.

    Using computer vision technology, my software allows users to "write" those "amnesia prone" characters using their head movements/rotations. In fact, their head movements/rotations can be considered as daily neck exercises to help reduce neck strain, which is also common among heavy computer users.

    Some smart wristband manufacturers from China are integrating my technology into their mobile apps as activity reminders, and a free PC version will be publicly available very soon.

    To everyone: I can safely say my PC software is quite interactive and enjoyable, so if anyone here want to learn some Chinese in a totally new way, just stay tuned:)

RSS feed for comments on this post