Japanese survey on forgetting how to write kanji

« previous post | next post »

According to a recent survey of more than 2,000 people, 66.5 percent of Japanese think they are losing the ability to correctly handwrite kanji. Moreover, the level is above 50% for every age group except for the youngest (16-19), who are of course still actively studying characters and thus must be prepared for tests; and even there the figure is still very close to the 50% mark. (See graph at the end of this post.)

This phenomenon is no surprise to anyone who regularly reads Language Log — see, for example, "Character Amnesia". Still, it is interesting to have the numbers. Also, the shift from the previous survey 10 years ago is dramatic. Read the article below for more details.

"Computer users say their kanji writing skills are getting rusty"
The Yomiuri Shimbun (Sep. 22, 2012)

A government survey found about 66.5 percent of people think they are losing the ability to correctly handwrite kanji or Chinese characters because they write e-mails with computers and cell phones on a daily basis.

The figure is 25.2 percentage points up from the previous survey 10 years ago, according to results of the Cultural Affairs Agency's survey on Japanese, which were released Thursday.

The survey also showed that 42 percent of respondents feel "writing by hand is a bother," up 10.1 percentage points from the last survey. The agency said, "This trend will certainly become stronger and be a serious problem for children in learning Japanese."

The agency in February and March conducted personal interviews with 3,474 subjects aged 16 or older across the nation. About 60 percent of them, or 2,069 people, gave valid answers.

Asked how they were affected by the diversification of information exchange media, such as e-mail, the number of respondents who replied that their ability to correctly write kanji is falling showed a sharp increase.

Looking at the percentages by age groups, those in their 40s who gave this answer had the highest figure at 79.5 percent, and those in their 20s, 30s and 50s were also over 70 percent. Though the percentages of teenagers and people aged 60 or older were lower than 30 percent in the last survey, the figures rose to around 50 percent this time.

The percentage of those who feel handwriting is a bother increased, while the number of people who find it troublesome to meet people and directly talk to them increased to 18.6 percent, up 7.3 percentage points from 10 years ago.

About 29.5 percent of the respondents said they would write e-mails even about simple things they could communicate orally, which showed 12.3 percentage points up from the last survey.

The agency said, "We must seriously consider how children, who have not yet mastered Japanese to a sufficient level, should be taught to use computers."

[Thanks to Mark Swofford]


  1. Chris Kern said,

    September 24, 2012 @ 9:33 pm

    Since this is only perception, it will do little to convince people — I'm continually surprised at the almost religious fervor people have when asserting that all Japanese people can read and write thousands of kanji in their sleep.

  2. M (was L) said,

    September 24, 2012 @ 10:15 pm

    Oh I disagree; perceptions do far more than facts, when it comes to convincing people.

  3. Stu said,

    September 24, 2012 @ 10:49 pm

    As a learner of Japanese, I can attest to the speed with which I have lost the ability to write kanji (and read them too, for that matter). Of course, as far as most Japanese are concerned, I'm just reverting to a position consistent with their expectations for hakujin.

  4. kamo said,

    September 24, 2012 @ 11:17 pm

    Stu – I've been told more than once that 'kanji are difficult for foreigners', to which my standard response is, 'yes, but then they're difficult for everyone.' Nice to have some numbers to back that up.

    "The agency said, "We must seriously consider how children, who have not yet mastered Japanese to a sufficient level, should be taught to use computers."

    This line of reasoning ('Not until they know Japanese properly') was also used not too long ago by people opposed to introducing English in Japanese Elementary schools. Part of me also wants to dismiss this as knee-jerk blaming of technology for social trends ('Student Death at Facebook Party' and the like), but for once it may be valid.

    Now here's a genuine question, because most people here are far smarter than me. My Japanese is still 'functional', but it seems to me that at present and in the future it would be far easier to work around an absence of kanji (with kana and spaces between words, say), than it would be to work around an absence of computer skills. Thoughts?

  5. Josh McNeill said,

    September 24, 2012 @ 11:40 pm

    I'm confused about why Japanese insists on using kanji anyway. I can come up with motives for Chinese as it's their native writing system and the only one that's really used. But Japanese is different. They're holding onto something they got from China when they have a perfectly good phonetic system that's already widely known and used and created entirely by the Japanese. I've even tried to get an answer from Japanese people and the usual response is something like, "Because there are so many homophones, the kanji make the meaning discernible." A piss poor argument since no one gets confused about what someone is saying in rapid speech where this luxury doesn't exist. Does anyone actually have a good practical answer for this? I'm sincerely curious.

  6. rwmg said,

    September 24, 2012 @ 11:58 pm

    I wonder if anyone has done any research on cross-cultural peevery. Would similar results be obtained from a survey on e.g., the effects of modern technology on spelling skills of English-speaking respondents themselves and the younger generation?

  7. Ethan said,

    September 25, 2012 @ 12:48 am

    @JoshMcNeill: Perhaps you should not be so quick to discount the usual response. With regard to your specific comment that "no one gets confused about what someone is saying in rapid speech" – (1) well yes, it happens and (2) you might enjoy reading the Wikipedia article on "Japanese pitch accent". At least some homophones that are distinguished in writing by differing kanji are distinguished in speech by differing pitch. Omitting or misplacing the pitch changes is perceived as contributing to a strong foreign accent.

    [(myl) I'm skeptical of this explanation, since there are large dialect differences in pitch accent; there are some regional varieties said to lack pitch accent entirely; and there is also quite a bit of idiolectal variation in pitch accent. However, if the information provided by normative Tokyo-dialect pitch accent were crucial, it could of course be indicated in the writing, as [standard Hanoi-dialect] tones are indicated in Vietnamese.]

    My Japanese is only 'functional' to borrow a description from Kamo above, but even as a late learner I agree with the claim that it is easier to read text with kanji than it is to read pure a pure kana verion so long as you recognize the kanji. Of course if you don't recognize the kanji this argument falls apart. Fortunately the number of kanji recognized in context can be far larger than the number you can write on demand. Which brings us back around to the original point of the "character amnesia" posting and the reported poll results. People are not necessarily forgetting how to read/recognize the characters, only how to write them spontaneously.

    [(myl) Fluent reading, in whatever orthographic system, is a highly over-learned skill. If you've ever tried to read English written in IPA, or for that matter in cyrillic, you'll understand that reactions and even experiments about how easy/hard it is to read language X in system Y are not very meaningful.

    There are a few large-scale social experiments we can draw on, for example the change in Vietnamese from a Chinese-character-based system to a latin-alphabet-based system (quốc ngữ.). I believe that the amount of homophony in Vietnamese is at least as great as in Japanese, but no one seems to have trouble reading and writing in the current system. Korean has similarly changed from the use of Chinese characters mixed with special symbols for endings, roughly as in modern Japanese, to the phonologically-based hangeul system, with the Chinese characters dropping out of use gradually. Again, the result seems to work perfectly well.

    I don't think there can be any doubt that Japanese could follow the path of Vietnamese or Korean. The arguments against doing so are social and cultural, not linguistic or psycholinguistic.]

  8. Ruben Polo-Sherk said,

    September 25, 2012 @ 1:20 am

    I can write correctly by hand maybe 15% of the kanji that I know–only the ones that I have had a need to write by hand. I type almost exclusively, and so I've never had a need to write the vast majority of the kanji that I know. (Also, I've never taken a course in Japanese, and I've never 'drilled' hand-writing kanji, so I never 'learned' them by hand in the first place, but still it is true that I don't need to know how to write them by hand if I know how to type them.)

    There a few reasons why kanji are necessary. One is legibility–Japanese written entirely in hiragana is almost impossible to read. You can try to get around this by putting spaces between the words, but then you have to force there to be a clear distinction between words that sometimes doesn't really exist. (Japanese is often written with spaces *between "phrases"*, particularly for providing clarity when representing a spoken style of Japanese, but that's something very different.)

    Another use of kanji is to differentiate between (often very subtle) various meanings of what is fundamentally one word. Context alone does not necessarily make clear which one is intended, and ocassionally, when it is necessary to be very precise, use of kanji is required to specify exactly what is meant.

    Actually there is also a way in which a system using kanji makes things easier to learn. A great many words and technical terms are composed from common characters and are much much easier to understand with kanji. If a student comes across 性的二形 or 弾道 for the first time, they'll know immediately what it means, which would not be true if it were written phonetically.

  9. Chris Kern said,

    September 25, 2012 @ 4:02 am

    Can you give one or two concrete examples of where a kanji is necessary to "be very precise" in a way that can't be picked up through context?

  10. Jan said,

    September 25, 2012 @ 5:30 am

    Does this correlate at all with the number of Americans who can no longer read/write longhand? May be a sign of the times.

  11. richard howland-bolton said,

    September 25, 2012 @ 6:05 am

    only anecdotal, but I find that my own handwriting in English is suffering from years of keyboarding: I note much more inconsistency in letter forms, even to the extent of (occasional, possibly random??) upper case forms intruding in inappropriate places.

  12. Ruben Polo-Sherk said,

    September 25, 2012 @ 6:28 am

    @Chris Kearn,

    Sure; the the smallest significance of the differences start at almost unimportant: その事が分かる・解る. (Perhaps if you want to make a difference, you can say the difference is something like that between "I get it" and "I understand it".)

    Then you can have more important, but still subtle, cases like Aの意味を表す・現す言葉 (Here, the difference is basically in where the "source" of '意味' is; emphasizing the difference, maybe the first is like "the word that expresses A" (where A is more abstract, and more of a platonic form), and the second, "the word that represents A" (where A is rather concrete in its existence at some level), which can be important, for example, if you are discussing a translation.)

    And finally, there are cases with very significant differences, like 返事を伸ばす・延ばす (The first case is like "lengthening the reply", and the second like "delaying the reply").

    All of this does not mean that normal people in everyday writing make these sorts of distinctions, but they are there to be made if you want.

  13. Chris Kern said,

    September 25, 2012 @ 8:03 am

    Hmm, I've seen that claim before but I just don't see it in actual writing — those are potential examples you made but I wanted to know if you could find any examples of someone doing this in an actual text.

    It was sort of a loaded question since I'm strongly in the "kanji are not necessary" camp, but I've just never seen anyone back up that "shades of meaning" claim with hard evidence that it actually occurs.

  14. RMilner said,

    September 25, 2012 @ 9:05 am

    This "forgetting kanji" problem has been around since the introduction of electronic word processors in the 1970s.

    Without time series data there is no way of telling if it has become worse.

  15. J.W. Brewer said,

    September 25, 2012 @ 9:43 am

    My never-more-than-minimal kanji-writing skills (last subject to formal testing in the 1975-76 academic year, when I was in fifth grade) are totally gone, but more to the point, my ability to produce handwritten English is likewise much much rustier than it was in the 1975-76 academic year – cursive is totally gone, and it's probably been 15 years or more since I last generated a multipage document in printed handwriting as opposed to on a computer screen. So this survey is a big unsurprising "duh" unless one actually empirically demonstrates the implied correlation between rusty handwriting skills and the inability to either reliably read or reliably produce via keyboarding the relevant characters. Notably, if I want to generate minimally legible handwriting I now need to slow down to a pace inconsistent with the speed of my thought and, more to the point, to a pace substantially slower than my typing wpm. I would be curious as to how many of those who said their kana-writing is rusty can't produce the characters correctly at all, versus just have to slow down and pay attention in a way they perhaps didn't in earlier decades and which is slow enough (compared to keyboarding) that it interferes with the flow of their thought. Presumably they are non-rusty in kana in the sense that they haven't forgotten how to produce any of the glyphs, but if you asked them to write out a text by hand in kana transliteration I wonder how fluid/fluent they would be at that task compared to a comparable population a few decades earlier who had not become as keyboard-dependent.

    As to implications for schoolchildren – my daughters have had no problems learning handwritten English (and have enjoyed the entirely vestigial skill of cursive) even though they were taught touch-typing starting I believe in 3d grade (I started in 10th grade) although they will I expect lose their fluency in handwriting with time as happened with their parents' generation.

    Are there any major language communities which have changed their writing systems absent either foreign occupation/colonization or a fairly brutal/authoritarian domestic government? Perhaps the period of MacArthur's proconsulship (when there was mild writing-system reform in Japan) was a felix culpa given the general arc of Japanese history and was fairly benign as these things go, but I wouldn't wish it on them again, nor would I wish the political regimes under which e.g. Turkey and Vietnam changed scripts upon the Japanese.

  16. Victor Mair said,

    September 25, 2012 @ 9:55 am


    And now we have that time series data you hoped for. The amnesia is getting worse.

    @J. W. Brewer

    Ataturk was a tyrant?

  17. Victor Mair said,

    September 25, 2012 @ 10:43 am

    To all those who compared character amnesia in China and Japan to forgetting how to write longhand in the West, these are two quite different phenomena. Although I compared the loss of cursive and character amnesia in "Cursive and Characters: Dying Arts", http://languagelog.ldc.upenn.edu/nll/?p=3117 I did not mean to equate the two.

    One may forget how to write cursive or longhand, but one will still be able to print out by hand the words the same way one enters them in the computer: a letter at a time. But when one relies on Pinyin or Rōmaji or kana or bopomofo to enter characters in electronic information processing and transmitting devices, one has become totally cut off from the sequence of strokes required to produce hanzi / kanji / hanja, which is so heavily dependent upon brute memorization and extraordinarily complex neuromuscular coordination.

  18. J.W. Brewer said,

    September 25, 2012 @ 10:51 am

    "Tyrant" might be a more demanding test than "brutal and authoritarian," which were my words. You might ask the descendents of the millenia-old Greek-speaking communities in Asia Minor ethnically cleansed into oblivion by Ataturk's henchmen what they think of his human rights record. To be fair, he probably got the idea that modernization was tied up with stupid nationalism and narrow ethnolinguistic chauvinism (accompanied by bogus claims about historical linguistics!) from 19th century European sources (and/or by the example set by Meiji-era Japan as to how to beat the Westerners at their own game). But even if Ataturk warts-and-all was the least-bad alternative available to the Turks at that particular historical juncture, that doesn't mean the Japanese (at their present historical juncture) shouldn't hope to do better.

  19. J.W. Brewer said,

    September 25, 2012 @ 11:27 am

    I don't know enough about the fading use of characters in favor on hangul in Korean to be certain of the political context, but it sounds from wikipedia that: a) the transition was mandated to occur more or less instaneously in the extremely tyrannical north; and b) the slower and perhaps more organic transition (which is not yet, I take it, 100% complete) had mostly been accomplished in the South during the period in which the ROK's various governments tended to be imperfectly democratic and might at least at times be fairly characterized as brutal/authoritarian (even though orders of magnitudes better than the North). But I don't know how government policy interacted with the transition in writing practice (in terms of whether it was pushing the change versus trying via educational policy etc. to fight an unsuccessful rear-guard defense of the older practice), so perhaps that's an arguable counterexample.

  20. Victor Mair said,

    September 25, 2012 @ 12:00 pm

    @J.W. Brewer

    Are you linking language / script reform with "brutal and authoritarian" (your words) governments?

  21. Ethan said,

    September 25, 2012 @ 12:22 pm

    @myl: I did not intend to imply that pitch accent was crucial to understanding, only that it is an example of information present in spoken Japanese that is not captured in the standard phonetic (hiragana) written representation. It is thus a counter-example to Josh's assumption that a written transcription necessarily captures all the nuances of a spoken phrase. I have encountered several schemes for encoding pitch using diacritical marks written over the kana, but only in foreign language dictionaries or textbooks. With regard to the difficulty of reading all-kana text, I am not supporting the claim that the difficulty is specifically a result of homophony, only offering admittedly anecdotal confirmation that it is hard to read.

  22. J.W. Brewer said,

    September 25, 2012 @ 1:01 pm

    Prof. Mair: Well, that or foreign occupation/colonialism. I don't claim that the correlation is perfect, and counterexamples would be welcome. Of course, arguably most human governments over the course of history might be classified as brutal/authoritarian and/or colonial/imperialist. Also, once two systems co-exist side-by-side and most people who are literate are literate in both, there can be a more incremental/organic drift that needs less top-down brute force to carry out than a switchover of the Ataturk variety (or the CCP introduction of simplified characters). Perhaps to some extent that is what happened in South Korea, and something like that may now be happening in Montenegro, where I am told there is a noticeable drift from cyrillic to latin, which varies by region within the country.

  23. J.W. Brewer said,

    September 25, 2012 @ 1:47 pm

    There could also be cases where the sort of liminal moment in a particular culture's history that can follow independence or the overthrow of a brutal prior regime would be a propitious time to change writing systems. There might under such circumstances be enough popular support for getting rid of a perceived relic of the ancien regime and/or foreign occupier to overcome the transition costs. I believe some languages spoken in the former Soviet Union which had previously been switched back and forth between Latin and Cyrillic according to the whims of Muscovite policy made further switches as a gesture of independence following the demise of Soviet rule, for example. But that's not a scenario for present-day Japan.

  24. Victor Mair said,

    September 25, 2012 @ 3:35 pm

    @J.W. Brewer

    Why can't there just be gradual evolution that happens because people simply prefer to do things a different, new, and perhaps easier way? Why cannot fashion be a powerful factor in language / script change? Why does it have to be top down, heavy, authoritarian, brutal (or a bottom up reaction to such regimes)? I believe that language usage is changing rapidly and radically in China today, and it's certainly not because of the government, but because of spontaneous preferences of masses of people, especially those who are active on the internet. That's why I speak of "emerging digraphia", never of imposed digraphia. Of course, there are individuals who may believe in digraphia and nudge things forward as best they can, but their impact is really miniscule in comparison with the practices of the people as a whole.

  25. Jongseong said,

    September 25, 2012 @ 4:37 pm

    A summary of the Korean situation:

    It should be noted that this was not a simple case of digraphia. Historically, the written language was Classical Chinese. So you had to learn an entirely different language from Korean if you wanted to read and write. Various cumbersome strategies were available to record the vernacular language (e.g. song lyrics) using Chinese characters and signs derived from them, but it wasn't until the invention of the Korean alphabet in the 15th century that the possibility arose for Korean to become a literary language.

    However, Classical Chinese remained the sole literary language in the public sphere, with writing in Korean relegated to private letters and the sort. Classical Chinese monopolized public written discourse all the way through the 19th century, when some Korean slowly began appearing even in public writing. By 1896, even the government's official weekly newspaper Hanseong jubo (한성주보, 漢城周報) had switched to Korean from the Classical Chinese of its decadal (published once every 10 days) predecessor, the Hanseong sunbo (한성순보, 漢城旬報).

    The Hanseong jubo was still full of Chinese characters though, because it used a mode of writing where Sino-Korean vocabulary was represented with Chinese characters, sort of how Japanese is written.

    Chinese characters of course are not necessary to write Korean. The first Korean novel written exclusively in the Korean alphabet came out in 1612, and by the time of the Hanseong sunbo, such publications were increasingly common. So one could say that the Chinese-character-heavy style of Hanseong sunbo represented the hold-out of an era where you were not considered literate unless you could read and write Classical Chinese, even when it was realized that a newspaper actually in Classical Chinese wasn't actually viable.

    The mixing of Chinese characters in Korean texts continued well into the 20th century because Classical Chinese wasn't instantly forgotten. You have to remember that people like Syngman Rhee, the first president of South Korea, began their education with the traditional curriculum of Confucian texts in Classical Chinese (interestingly Rhee was involved in the publication of a Korean-alphabet-only newspaper, and later, as president, set South Korea on the path to Korean-alphabet-only writing). But by the time Korea regained independence from Japanese colonial rule in 1945, whole generations had grown up without learning Classical Chinese.

    In 1948, when the respective governments were established in North and South Korea, both chose to use the Korean alphabet exclusively in official documents (although the South allowed the simultaneous use of Chinese characters "when necessary"). Textbooks were written exclusively in the Korean alphabet and Chinese characters were taught as a separate subject in both Koreas (yes, Chinese characters continue to be taught even in North Korea). There is a world of difference between having to learn Chinese characters merely to read your textbooks and learning Chinese characters as just another subject. Knowledge of Chinese characters would be essential in the former case but not the latter.

    In South Korea, because Chinese characters were never banned from the public sphere, the "mixed script" style persisted for a few decades (due to its more lofty associations), newspapers being the last major holdout. I think it was around the 90s when Chinese characters virtually disappeared from all newspapers.

    Unless the "mixed script" style had been made mandatory from the earliest textbooks, as in Japanese, Chinese characters would have disappeared naturally from Korean writing anyway. You never see Japanese written exclusively in hiragana and katakana, whereas a tradition of writing Korean exclusively with the native alphabet has existed for several centuries already. Mixing Chinese characters made sense for those who were also educated in Classical Chinese, and for the next few generations that grew up familiar with such a style of writing. After that, the utility of Chinese characters in Korean texts rapidly approaches zero.

  26. The Ridger said,

    September 25, 2012 @ 5:16 pm

    Oh, dear. I don't really know Japanese (I'm about halfway through a teach-yourself book) but the first thing I thought of was Sailor Moon's friends identifying the letter from the future as being indisputably written by her adult self because it had no kanji in it.

  27. Josh McNeill said,

    September 25, 2012 @ 7:12 pm

    @Ethan: Because of your reply, I thought maybe homophones are orders of magnitude more profuse in Japanese than English (ie. where English may have as many as 3 words that are homophones, Japanese might have 12 or more). So, I tried to figure out how extreme homophony is in the language and came across this on one site:

    "The distinctions between kun-homophones are often so subtle that even native speakers do not know how to use a particular homophone correctly. Ask a Japanese friend or colleague to explain in which circumstances you should use 硬い, 堅い, and 固い (all are pronounced kataiand mean "hard") . . . "

    It strikes me that the kanji almost seem to be forcing multiple words to exist where there's really just one word that can be used in various ways. For instance, Spanish has the word "llevar" and if you pick up an English/Spanish dictionary you'll find possibly up to 20 translations for this word, almost all of which are very different. Does this mean that there are 20 distinct words here with the same pronunciation or just 1 word with 20 uses? Maybe this is why, as the site I found this on claims, native Japanese speakers would have such a hard time distinguishing the difference between these 3 homophones that all mean "hard"; they could simply be 1 word that seems like 3 words because the writing system implies that, not because that's actually the case.

    Maybe an even better example (that doesn't require translation) would be the word "make" in English, which Merriam-Webster lists as having over 30 meanings. We could write out all of those with different spellings, for instance, and people might get the idea that they're actually many words that are homophonous and would probably have just as much trouble figuring out which one to use. But are there 30 words here or just 1 with 30 uses?

    I am nowhere near an expert, especially when it comes to semantics, so I'm only posting my deductions in the hope that some light can be shed on the argument for me.

  28. Josh McNeill said,

    September 25, 2012 @ 7:14 pm

    Oops, sorry, I should've quoted the whole paragraph from the site:

    "The distinctions between kun-homophones are often so subtle that even native speakers do not know how to use a particular homophone correctly. Ask a Japanese friend or colleague to explain in which circumstances you should use 硬い, 堅い, and 固い (all are pronounced katai and mean "hard") or 柔らかい and 軟らかい (both are pronounced yawarakai and mean "soft")–the reply you receive is likely to include head-scratching and "soo desu ne"-ing."

  29. Ethan said,

    September 25, 2012 @ 10:12 pm

    @Josh: You are conflating two things that I believe deserve separate consideration. The example you give of 堅/固 is roughly analogous to alternative spellings in English. Think jail/gaol or gray/grey. No one would argue that they are separate words, but a writer striving for atmosphere might argue that "gaol" conjures up connotations of quaint antiquity, at least for an American reader, which are entirely missing if "jail" were substituted. If you look up "katai" in an English/Japanese dictionary you will find something like:
    かたい [堅い, 固い] (adj) hard (esp. wood), solid, tough, stiff, tight,… etc through a long list of possible translations.
    I.e. these two kanji are considered alternative ways of writing a single word, although that word itself can have a large cluster of related meanings.

    The issue of ambiguity arising from homophony is something else. Think bare/bear pale/pail coop/coupe etc. It is plausible that Japanese has more such ambiguous pairs because so many longer words are made up of short Chinese-derived roots, but I do not have any real statistics to back up that plausibility. For example, a dictionary search for しんこう (shinkou) returns 17 separate meanings represented by 7 different kanji in the first position (shin) and 15 different kanji in the second position (kou). Now those meanings range from 'close friendship' to 'pickled vegetables' and so are unlikely to cause confusion in practice, but I think that is the sort of thing people have in mind when they argue that homophony requires disambiguation.

  30. Ruben Polo-Sherk said,

    September 25, 2012 @ 10:18 pm

    Chris Kern: I don't really know how to answer your question. There's no way you can have some distinction exist in theory without it ever being manifested in 'reality'. Anytime you see anything written in Japanese (even in hiragana), it was either written with these distinctions in mind, or it wasn't, and the only way to find out for sure is to ask the person who wrote it.

    From the perspective of language as communication, where it makes a difference is when the character written is somewhat surprising by being a little different from the meaning expected by the reader. Then they have to think about whether the written meaning is the one that makes sense in the immediate and larger context. But this effect isn't really that exotic–the same happens in English when you come across a slightly weird usage, and need to figure out what the underlying meaning actually is.

    Basically, it happens all the time (though in a minority of cases).

    About how much the writing system is tied to the language:
    If you were to arbitrarily make Japanese written phonetically, life would go on, and in a hundred years nobody would care. But there are linguistic aspects that will change directly as a result of this. Particularly, three: 1) if you change to a phonetic script, you'll have to introduce spacing. This is a subtle change, but it is more than just orthography. 2) The amount of homophones in Japanese is much larger than in most languages written phonetically. Where this is mostly relevant is in the two-morpheme Chinese loanwords. Take, for example, a word like "telescope"; imagine then that "tele" and "scope" each have 10 unrelated meanings. Perhaps eventually there will be semantic (if the meanings happen to be similar in a particular respect) and/or morphological change in the words (if context doesn't always distinguish them, even if their meanings are very different). 3) there will be no morphological distinctions for the kind of polysemy that I mentioned earlier; depending on the nature of the differences, in some cases the distinction in meaning will be preserved, and in others, it will be lost.

    Josh McNeill:
    You've actually brought up a different point, about one word in language A having many different translations in lanugage B. This happens even if those different uses are not distinguished in A. That's a different effect from the one that kanji distinguish. (Though there is some overlap in the kinds of distinctions provided by kanji and those illuminated by translation)

    With 分かる・解る, for example, in some sense they are just one word, and in some sense they are two different words. You can't really say that the different characters are forcing the existence of different words. It depends on what you want. Most of the time, it doesn't matter, but sometimes it does, and then people will use a particular character, or even write it purposely in hiragana to get the "one word" meaning.

    Things like "make" are, I think, more analogous to thing like する (which is written in hiragana, or if you want to be pompous, with one particular kanji), where there's one fundamental meaning stretched or extended to many different uses, rather than there being different "kinds" of a particular meaning that can be distinguished by orthography.

  31. The other Mark P said,

    September 25, 2012 @ 10:31 pm

    Josh McNeill said,

    I'm confused about why Japanese insists on using kanji anyway

    Does anyone actually have a good practical answer for this? I'm sincerely curious.

    Why do the English resist simplification of their spelling?

    Because elites, having spent a lot of time and energy mastering a system, are reluctant to move to a simpler system. It degrades their point of difference from the oiks. And most other people follow the lead of the elite rather than look like yokels, just as they do with accents, dress etc.

    A good example is the French way of counting. Ninety in Parisian is four-twenties and ten (quatre-vingts dix) for absolutely no good reason at all. The Swill use ninety (nonante), which is quicker and much easier for kids. All the French recognise that nonante is better, but won't use it because of its peasant origins. Basically total snobbery. But good luck trying to persuade the French to change.

  32. Jason said,

    September 26, 2012 @ 7:27 am

    Why do the English resist simplification of their spelling?

    Because elites, having spent a lot of time and energy mastering a system, are reluctant to move to a simpler system.

    Allsoe, Im afrade, bekoz its hedake induceing tu adapt tu an allternativ sistem since the spelling ov vertchualli everi werd wood hav tu be chanjed.

    Even rational Japaneze hu doent bi the kultcher or nesessiti arguement, understandabli, doent wont tu render everi book printed since the 19th sentcheri unreedabel tu a puteativ futecher jennerateion.

  33. J.W. Brewer said,

    September 26, 2012 @ 8:09 am

    For a culture like that of Japan or most Anglophone countries which has already transitioned to mass literacy, I don't think the elite/snob explanation is necessary or even helpful (I think if you ranked the various European languages by the "simplicity" of their orthography you wouldn't get any sort of meaningful correlation to the egalitarian-ness of the relevant societies, because it's mostly just a bunch of contingent historical accidents often having to do with how recently the language's orthography got standardized and how the language's phonology has or has not changed since then). Think of the status quo in either Japanese writing or English writing as more like the QWERTY keyboard. Is it the optimal system to adopt if we were, as it were, writing on a blank slate? No. Are we going to undergo the massive transition costs of wiping the slate clean and starting over from scratch? No as to English, and probably no as to Japanese, although existing digraphia (actually trigraphia given the kana . . .) might make the transition somewhat less onerous.

  34. Ruben Polo-Sherk said,

    September 26, 2012 @ 9:05 am

    I think that "blank slate" is a helpful image. Kanji wasn't introduced into Japanese in order to fit some deep truths about the language, but now that it has been, it's part of the language in a much more than superficial sense. The things that I said earlier kanji is necessary for are essentially so because of the writing system.

  35. Victor Mair said,

    September 26, 2012 @ 9:06 am

    @J. W. Brewer

    "existing digraphia (actually trigraphia given the kana . . .)"

    I tend to think of the Japanese writing system as composed of four main components: kanji, katakana, hiragana, and Rōmaji. Still, I don't think that this constitutes a genuine quadrigraphia, since none of these components functions alone as a means for representing Japanese language as a whole. Perhaps, though, some people will nonetheless want to call the Japanese situation a "sort of quadrigraphia".

    Similarly, as I've shown in many recent posts, the Chinese writing system as currently configured, consists of two main components: characters (in simplified and traditional forms) and Roman letters. Although I believe that a true digraphia (Romanization and characters) is emerging, it has not yet arrived, since Romanization (Hanyu Pinyin) so far is used only as auxiliary phonetic annotation for the characters or mixed in with the characters for special purposes (e.g., zi4mu3ci2 字母詞 ["lettered words"]), but do not yet independently function as a full representation of Mandarin or other Sinitic language (except for Cantonese, Shanghainese, and especially Taiwanese among restricted groups and for pedagogical purposes).

  36. Josh McNeill said,

    September 26, 2012 @ 4:09 pm

    So, if I'm understanding, the homophony argument isn't really about making sure people can understand what the writer is trying to say all together; it's about having options that allow one to conjure up various connotations? As in, there's no rule that says in one context you must use 堅い and in another you must use 固い, you can use either one any in context, but the option allows you to imply more, if you desire, than what could be implied if you simply wrote かたい all the time?

    I can see the utility in that. But this makes it seem strange to me that the argument would be framed as "it clears up ambiguity." It doesn't seem to be doing that to me; it seems like it's simply adding annotations to what is already an easily decipherable text. If that's the case then you don't really NEED kanji for Japanese writing to make sense, it's just kind of a nice tool to have (if you can remember all of them). So the whole issue pivots on whether one puts more value in the usefulness of being able to add annotations through orthography or whether one puts more value on the simplicity of a more streamlined writing system that can be learned in a week as opposed to years.

  37. wren ng thornton said,

    September 26, 2012 @ 4:48 pm


    I dunno. Personally, I find all-kana writing nearly impossible to read, even with spaces. (Though as Ruben Polo-Sherk mentions, the question of where word boundaries belong in Japanese is highly non-trivial, albeit conventionalized in romaji.) Part of it is the high degree of homophony, though presumably this could be overcome with training since it's something we have to deal with in speech. (Adding a notation for pitch would help though.) But part of it is that the rhythm of switching between kanji and kana helps parsing significantly: you can tell at a glance where the phrase boundaries are, and can often get a feel for the style/register even without reading it.

    But, as has been suggested, the main thing is cultural. Even as an outsider, I'd hate to lose the Japanese writing system. The ability to imbue different nuances by using different kanji or by using kana instead (or kanji instead, where kana is the default) gives a wonderfully artistic air to writing. This isn't just exoticizing the other; it fits in with their love of language games, numerous artists' work which focus on language, and the overall way that the Japanese conceive of language generally. This sort of playfulness is all too uncommon in the modern era of regimented spelling and standardized orthography, and it'd be sad to lose the last bits of that.

  38. Victor Mair said,

    September 26, 2012 @ 5:42 pm

    @wren ng thornton

    Ask the South Koreans whether their literature has been impoverished because of the near total absence of hanja in their writing today.

  39. Ruben Polo-Sherk said,

    September 26, 2012 @ 6:42 pm

    Josh McNeill:

    Yes, exactly. (though, actually, the option to write it in hiragana can be considered to be something that gives even more expressive ability (maybe the author wants the "ambiguity" or the unified meaning, or is talking about something where the differences distinguished by the different characters don't matter)).

    I agree: kanji doesn't make understandable an otherwise-undecipherable text. I guess when most people make the "clears up ambiguity" argument, they are talking about the bunch of two-morpheme words (Ethan gave the example of しんこう). But there, context can take care of ambiguity in almost all cases; at most, it makes the text easier to read (for some people, at least–I can read Chinese faster than English).

  40. kamo said,

    September 26, 2012 @ 7:49 pm

    @wren ng thornton

    Thanks for the response. My question was intended to be slightly different to the 'are kanji necessary?' debate that's being played out above, interesting though that is.

    I completely agree that word boundaries are non-trivial, and that losing kanji would be, well, a loss. But working in Japanese education it seems that there's often a tendency for the bureaucracy to seek comfort in tradition at the expense of newer, perhaps more relevant skills (which is hardly unique to Japan, of course).

    The part of the report I quoted seemed to be suggesting that computer skills, and thus computer literacy, shouldn't be taught until sudents' Japanese was 'sufficient'. I'd suggest that for most current Japanese students a lack of computer literacy would be a bigger impediment to their future success than a lack of kanji literacy. But that's just a gut feeling and I'd be happy to hear something more concrete.

    And I'm afraid it does sound a bit like you're exoticising the other. Either 'they' love language games, or the playfulness is uncommon. Can't have it both ways, I'm afraid ;) Believe me, there's nothing playful about the hundreds of hours the sutudents devote to writing and remembering kanji, sadly.

    And finally, an anecdote for everyone's consideration and amusement. My son's Japanese picture books are written entirely in hiragana. When I first read one, it does take me a bit longer to recognise words if I'm used to seeing them written in kanji, but it's obviously still possible. It also means that I'm capable of reading any words that I wouldn't know the kanji for (which, given that these are pre-school picture books with a commensurate level of vocabulary, is pretty shameful. But we'll gloss over that).

  41. M (was L) said,

    September 26, 2012 @ 9:10 pm

    @Victor Mair

    > Ask the South Koreans whether their literature has been
    > impoverished because of the near total absence of hanja
    > in their writing today.

    If they can't read it, they may have no idea what they've lost. In which case, asking them would be rather unfair and misleading, no?

    All change is both gain and loss. Pretending there are pure gains is like pretending there are free lunches.

    The change may well be a good bargain, often they are favorable exchanges. But make no mistake, they are always exchanges. Something is paid, and something is received.

    There is a kind of cowardice that holds onto the past and ignores the alternatives. There is a kind of cowardice that runs like hell from the past and demands alternatives. They are of course exactly the same thing; in the end, there's only one kind of cowardice.

  42. Victor Mair said,

    September 26, 2012 @ 10:08 pm

    @M (was L)

    What you call cowardice may be viewed by others as bravery, especially when one marches forward to embrace the future.

  43. M (was L) said,

    September 26, 2012 @ 11:57 pm

    Yes, they may call it that.

    It's a damn good thing nobody reads Plato or Shakespeare anymore, or the US Constitution, or their last paycheck.

    Because the past is all kinds of icky and stuff and brave people are far too courageous for any of that crap.

  44. Victor Mair said,

    September 27, 2012 @ 5:02 am

    @M (was L)


    Lots of people still read Plato and Shakespeare, and with great gusto! I even named a journal after Plato! Yet I look toward the future with hope and expectation, partly because I am sustained by Platonic ideals.

  45. Rodger C said,

    September 27, 2012 @ 7:46 am

    I believe M (was L) was being snarky.

  46. M (was L) said,

    September 27, 2012 @ 8:59 am

    @Rodger C


  47. Seralt said,

    September 27, 2012 @ 11:37 am

    Re: kanji necessary?
    I don't know if you want to count it as an example of an eggcorn, or a necessity of kanji to distinct homophones, but I had a (Japanese, native speaker) friend who thought the idiom 「便りがないのは元気の証拠」 ("no news is good news"; lit. "no news is the proof of health") was instead 「頼りがないのは元気の証拠」 ("no crutches is the proof of health").
    On a separate occasion, I was telling a friend about how 晦渋(/kaijuu/ – opaque; dense) an author's writing style, but he misinterpreted it as being 怪獣 (/kaijuu/ – monster; beast). In my experience, the ambiguity of speech/homophones doesn't always hinge solely on the words being used. It also depends partially on mutual vocabularies (in this case, 晦渋 seems to be a not-so-common literary word), and the trust that the speaker is employing the words correctly (from a second-language speaker, I'd be guilty of second-guessing all their utterances, even if they were employed correctly.)

    Re: different words for different nuances
    Words like "make" and する are generic catchall terms that exist across all languages. "take a class" doesn't mean the same thing as "take a photo", but in Japanese, it's similarly 「授業をとる」 and 「写真をとる」, even though in both languages, more specific verbs exist than the general "take/とる". The way it looks to me, Japanese has a similar range of polysemous words that acquire different meanings with their complements. But unlike in English, where we specify the meaning by making the complement explicit (e.g. " 'take' as in 'take a bus'; not 'take a nap'"), Japanese allows its words to wear different kanji to explicitly mark its function in a given context (「取る、¬撮る」). As for homophones of sinitic Japanese nouns, I would love to read some research on whether they cognitively reside as separate lexical entries, or whether they similarly acquire meaning only through context.

  48. Victor Mair said,

    September 27, 2012 @ 11:40 am

    @Rodger C, M (was L)

    Of course, he / she was being snarky, but snarkiness proves nothing — except that one is snarky.

  49. M (was L) said,

    September 27, 2012 @ 12:34 pm

    @Victor Mair –

    Not so, as it happens. It should prove nothing else, and at the hour I wrote that I was nothing but snarky. This I freely, and snarklessly, admit to – and I ask no forgiveness because I see no reason why I should expect any to be forthcoming.

    But all the same something was proven in your answer. You speak of the value of Plato, who wrote in a dialect of Greek that modern Greeks tell me they find archaic and difficult; and in an alphabet containing a few letters no longer in use. As I read effectively no more Greek than one gathers in a geometry class, it is both literally and figuratively all Greek to me; but not to you – and right there is the point.

    Whatever is special about period Greek, including whatever is special about period Greek spelling and writing, is all part of the history of the language, of the culture, and in the case of Plato of a great deal more than just this.

    When you see an archaic spelling, do you not see the hints and echoes of etymology? Remove the thorn from Beowulf's paw, and you remove all of that.

    And here I thought this would be obvious to linguists. On this and another thread, I am learning otherwise.

  50. Victor Mair said,

    September 27, 2012 @ 2:23 pm

    M (was L)

    Thus, as you argue, what was proven is to be found in my non-snarky reply to your snarky comment, not in your snarkiness. So snarkiness by itself still proves nothing.

    @Seralt and many others who wonder whether kanji are indispensable

    May I simply point out once again what I've said so many times in the past: no one ever spoke a hanzi / kanji / hanja, except in cartoon balloons and manga cartouches, yet people have been communicating with each other at greater or lesser length and on vital and frivolous topics for thousands of years, even before there was writing at all.

  51. M (was L) said,

    September 27, 2012 @ 7:42 pm

    Nope, you were right the first time. It proves I'm snarky.

  52. Victor Mair said,

    September 27, 2012 @ 8:34 pm

    @M (was L)

    Yuánlái rúcǐ 原来如此

    Sou desu ka そうですか?

  53. M (was L) said,

    September 27, 2012 @ 8:56 pm

    If you say so.

  54. Victor Mair said,

    September 28, 2012 @ 6:17 am

    @M (was L)

    Naru hodo なるほど

    Yappari やっぱり

  55. M (was L) said,

    September 28, 2012 @ 9:57 am

    Ya know, I don't understand either Chinese or Japanese; I gather from the transliteration that you've said something in Chinese once, and three things in Japanese, in your last two posts.

    On that basis, and unreliably assisted by http://translate.google.com – which does the best it can, but that generally means a very crude gloss – I think that you have chosen to fight snark with snark. Please understand that I cannot be absolutely certain; but I'm pretty sure all the same. In any case, that's the conclusion I'm going with.

    Your use of both Kanji and (I think?) hirigana as well as Romanization suggests again that, despite your protestations, you respect the whole history of writing, excluding for obvious reasons my own. What does this prove? Not very much, as each of us is an n of one – the exchange mostly proves things about you and about me.

    But there it is all the same; despite your protestations that whatever is written in Kanji is just as well written in Hirigana, you have chosen to write in both – revealing that you see some utility in using Kanji all the same.

    You beat me with my own stick, and it hurts like any other stick, but I thought you'd like to know which stick you're hurting me with. Don't worry overmuch, it only hurts a little bit and I'll quickly get over it. I have many others, you can keep that one if you like.

    As the things all this snarkiness proves about me are entirely uncomplimentary, I choose at this juncture to staunch the wound as best as I can, apologize for the whole mess to you and to the general reader and to the host of LL, and to stop right here.

    I'm sorry.

  56. Oliver said,

    September 28, 2012 @ 11:42 am

    All the French recognise that nonante is better, but won't use it because of its peasant origins

    Peasant origin? The form looks very much derived from Latin.

  57. Victor Mair said,

    September 28, 2012 @ 6:39 pm

    from C.S.S.:

    This [kanji amnesia] is very, very true. We use computer for writing letters in Japanese; it changes KANA into KANJI, all we have to do is choose the correct one. So we don't have to remember how to write them.

    It's the same as our calculating ability. I remember hearing Dr. Bok, then Harvard President, when he came to Philly and talked about the use of computers and calculators.
    He said the students forgot to do simple arithmetic because they use calculator/computer for everything.
    When I am writing an essay in Japanese, I have to look up words in a Japanese thesaurus. I have forgotten words and how to write them. If I see KANJI on the computer, I know whether it's the right choice or not. But if I have to start from complete blank, I'll have a problem.

  58. tomoda said,

    September 29, 2012 @ 12:51 am

    Actually, the phenomenon that many people here seem to bemoan is the result of a massive improvement in the convenience and ease of writing Japanese. In other words, it is a _good_ thing.

    Of course people's ability to write complicated Chinese characters by hand is declining–because people almost never need to write the characters by hand anymore. Computers and cellphones and other digital devices have made writing Japanese much, much easier than it was a generation ago. Incidentally, this has probably put paid forever to the argument that "Japanese would be much better off without kanji," since the burden and inconvenience of the writing system is now much less.

    The phenomenon discussed in the article has little to do with people losing the ability to write Chinese characters altogether, I think. More likely they are talking about those times when you do decide to write a letter by hand, or have to complete a form, and find that you can't quite remember whether a certain character requires element A or B in its bottom corner, for example. And what does one do in these situations? You whip out your mobile phone or similar device, which will remind you within seconds of which character you want, and on you go.

    The issue with homophones tends, I think, to be overstated. There are large numbers of Chinese-derived compounds that have the same pronunciation in Japanese. (And often the same pitch too, for what it's worth.) As with the example of 晦渋 above, however, many of these are almost never used in normal conversational Japanese. There are large discrepancies between the vocabulary used in learned written discourse and the vocabulary of speech (even learned speech). This is the case in all languages, but I think particularly pronounced in Japanese, where to a certain extent two different lexicons exist, with minimal overlap between them. Stripping Japanese of the kanji _would_, I think, have a deleterious effect on written Japanese in this sense. I don't know much about the situation in Korea, but I understand that at least some people there feel that people's ability to understand difficult technical vocabulary (largely Chinese-derived, as in Japanese) has suffered as the Chinese characters have fallen out of use there.

  59. C.S.S. said,

    September 29, 2012 @ 8:30 am

    [(myl) I've replaced this commenter's name with initials, in keeping with our general policy for communications whose authors have not specifically indicated that they want to be identified more fully.]

    Dear Victor,

    I protest. You told me to look at all the comments you have received for this column, so I did, and I was astonished. The e-mail I sent to you was just to you, and I had no idea that such fine arguments and discussions had taken place. My simple comments were meant for you and not for publishing. Yes, this argument, complex and thought-provoking, has been taken to nth power – not that no more discussion is possible. But it's been discussed enough to make sense – from the length of it – since I can't read them all, I judge it by the length!!!
    I read some, but I stopped because my eyes are so bad I cannot tolerate much concentration, but many people have developed many interesting and logical argument.

    A movement to eliminate the use of Kanji always existed – since the 1930s. I think Shiga Naoya at one time advocated that, too. There was a movement to use Romaji – the alphabetized writing, but it was simply silly and they soon gave up the idea. Tanizaki Junichiro, god of the Japanese language, pointed out the subtlety of the Japanese language and its shaded nuanced use including kanji. So after diverse argument peopled always went back to the use of kanji – the quadrographic writing Victor referred to.
    To me, it's also a matter of economy – maybe someone has pointed this out too. Think of writing out everything by kana, for instance – I can't think of any good example, but 神仏, which can be readしんぶつ、which is not too long, but in vernocular, it is かみほとけ、5 letters instead of 2. In conversation they would be using かみほとけ, and that will never lead to misunderstanding. If everything is written out in kana, you have to stop and think sometimes, but with kanji, you know the meaning at one glance. It is not only economy of space but it's also economy of conception, cognizance.
    I am going to quit now since my eyes hurt very much.

  60. Victor Mair said,

    September 29, 2012 @ 10:49 am

    My apologies to c.s.s., who is a dear friend of mine, and who in the past, whenever I've asked her permission to add her comments to LL posts, has always generously granted it. I'll be sure to ask her again whenever I want to post any of her good comments in the future.

  61. Rodger C said,

    September 29, 2012 @ 11:03 am

    @Oliver: "Peasant associations" would have been more accurate.

  62. JMU said,

    September 29, 2012 @ 11:08 am

    Deterioration of active kanji recall was already happening in the 1980s, only a few years about the first JWP machines were introduced, as I reported in my book The Fifth Generation Fallacy (1987). It is not that computer use is a like a drug that does something to your brain. It is simply that daily practice writing kanji turns out to be necessary for maintaining active recall. This was not fully appreciated before computers because there had never been a natural experimental setting in which many kanji users were able to get along without a lot of daily handwriting.

    I am reminded of newspaper articles about early automobiles. People at the start of the 20th century were well aware of the health risks of the exhaust spewed by internal combustion engines, but they saw them as a small price to pay compared with well-known and diverse costs of keeping and using horses in cities and the many health risks they entailed.

  63. JMU said,

    September 29, 2012 @ 11:10 am

    (Pardon the typos.)

  64. Victor Mair said,

    September 29, 2012 @ 2:52 pm

    from Bob Ramsey:

    I was a bit amused by this comment, which repeats an old saw about kanji in response to a question about how they clarify semantic subtleties:

    "…the smallest significance of the differences start at almost unimportant: その事が分かる・解る. (Perhaps if you want to make a difference, you can say the difference is something like that between "I get it" and "I understand it".)" –and the commenter went on from there.

    What's being said here, as I'm sure you know, is that through the adoption of a distinction in the use of Chinese characters (which often stems from a difference originally found in Chinese), the Japanese have developed some kind of mental awareness of a semantic distinction that does not exist in the sounds of their language, and can only be accessed visually by looking at the characters. The most famous kanji distinction introduced this way is that of 早い versus 速い, both used to transcribe はやい. And dictionaries, including Kenkyusha, list each as a separate word, one translated 'early' and the other 'be fast, quick'. Many linguists have pointed out, though, that there is really only one word here, and that its semantic range includes both 'early' and 'fast'. In other words, context reveals any distinction to be made. But when Japanese learned kanbun and kanji, they started artificially representing this distinction in their own language by transcribing hayai as 早い in contexts where the word was used in reference to a point in time, and as 速い in contexts where it referred to an elapse of time. Now, your commenter assumes that this artificially introduced distinction has now become a natural part of the language, and it may be argued that that's true. (But again, the context already makes the interpretation clear without the graphic distinction.) I suggest only that this hypothesis needs some unbiased testing beyond just asking native speakers to verify through their introspective impressions. You know the drill. I'm sure Bill Hannas has already talked about it.

    We all know (and I suspect most of your bloggers will agree) that this kind of reasoning breaks down most clearly in talking about such things as those sad character distinctions Chinese introduced into their writing system in order to superimpose Western pronoun distinctions ('he', 'she', 'it') onto ta. Does anyone still imagine that that female determinant added to the character makes a distinction that has actually become part of the language? I suspect very few still labor under that illusion. It seems to me that kanji distinctions introduced into Japanese without phonological support represent much the same kind of artificiality. –But, as I said, maybe there exists psycholinguistic experimentation that has already tested these things…

  65. elessorn said,

    September 30, 2012 @ 2:35 am

    I can't see that there's hope of a resolution to this argument. None of the examples here adduced of various distinctions which only kanji usage makes possible will ever convince anyone who argues from a utilitarian standpoint. The utilitarian may or may not agree that such distinctions are a loss worth mourning–either way it will hardly effect his judgment that the utility gained by an objectively simpler writing system outweighs any such losses.

    Nor will the complaint of the very real rupture that giving up kanji would create with the past be likely to sway the utilitarian. The example of Korea (or even of Turkey) demonstrates that such a rupture is not literally fatal for a culture, not necessarily unbearable or practically disastrous–which removes the only hurdle that a utility argument might pause before. There may be utilitarians swayed by such a historical appeal, but I doubt there will ever be many outside of the culture in question.

    It boils down to this: one feels enough respect for human monuments to hesitate before abandoning what is incontrovertibly individual and irreplaceable–a historical tradition; or one does not feel so, or at least not so strongly enough. One feels either so strongly that what has been built up should not be hastily unbuilt, that one is willing to tolerate all manner of small and large inefficiencies just to keep things going, or such reparable inefficiencies come to seem intolerable, and one joins the camp of reform. I personally think Japan's case shows that the various drawbacks to kanji usage are not strong enough to demand or even strongly urge an all-kana policy. And I cannot accept what seems obvious to so many foreigners: that the objective ease of kana is a better argument than the even more objective fact that most speakers of Japanese alive today write, read, and think in kanji without much trouble–since it is they, after all, and not foreign learners, whose system would be eradicated. But then I'm also sure that Victor Mair is right that if kanji ever are thrown out in Japan, the children of the revolution will most certainly sing the bravery of the decision. I don't think I would actively argue against a successful indigenous reform movement–it's not my language, anyway. But I do question the ideology (and it is one) that the principle of maximum utility is obviously persuasive, and skepticism clearly some silly romanticism.

    I have to say that the only thing that confuses me is Victor Mair's position. Not on the article, which provides nothing but another example of how computer mediation atrophies previously robust skills, and nothing more. On other fronts Professor Mair fights a host of campaigns against short-sighted utility, be it political, linguistic, or scholastic. Why when it comes to Chinese characters does the trend shift? Fighting against the sort of government-supported purists who would actually suppress grass-roots English digraphia in China is one thing, but against kanji usage in Japan?

  66. Seralt said,

    September 30, 2012 @ 11:14 am

    I totally agree that verbal communication has continued to be sustainable without the support of kanji for centuries. I also think that your posting of Bob Ramsey's comment made more succinct what I clumsily tried to express — that in many cases, the kanji introduced an artificial split between two different senses of the same Japanese word. But as commenter Tomoda says, there is a more pronounced discrepancy between written discourse vocabulary and normal conversation vocabulary, especially in Japanese — which I can certainly agree with on a personal level. Admittedly, the conversation I had involving 晦渋 was with a 'science-stream' (理系) person, who also didn't know the word 婉曲 — for which I believe there exists no homophone. So I suppose it was more of an issue with vocabulary than of kanji knowledge. (Whether 頼り("support; reliance") and 便り("letter; news") is another example of such an artificial split, is harder for my English-speaking brain to conceptualize.)

    But to follow up with what Tomoda has said about normal conversation and written discourse, I am curious (as a thought experiment) if an elimination of kanji would remap (admittedly less-common) vocab of written discourse into the more common homophones of common speech, or would they simply fall out of use, even in written texts. I also wonder whether such a similar remapping of vocabulary has already occurred in Korean when they switched to Hangul.

  67. Ruben Polo-Sherk said,

    September 30, 2012 @ 11:22 pm

    Bob Ramsey:

    I agree with part of your post. As I said in a previous post, any necessity that kanji has it has essentially created itself.

    But it is not quite true that context reveals "*any* distinction to be made" (unless you extend "context" to dictionary entries): はやく動く can be either 速く or 早く in many contexts; for example, in sports. Which, again, goes back to something I said earlier: If Japanese were not written with kanji, it is likely that there would be a phonological or grammatical change to differentiate things like 早い and 速い when the context is not sufficient and the difference is important.

    I think it's a bit strong Sapir-Whorfian to toy with the idea that the lack of a phonological distinction makes any cognitive distinction "artificial". Now, it may be that that is the case for some people (which may be confirmed by testing), but if it isn't true in general, you are talking more about psychology instead of the linguistic properties of Japanese.

    Also, written language is somewhat independent from spoken language, and is able to inform people's understanding of the language in general. In a sense, it's useful to talk about spoken and written language as separate entities. In English, for example, nobody pronounces commas or quotes or semicolos or colons. But in written English, they are not just orthographical marks included to make understanding the text easier–they convey information. And so it would be quite incorrect to say that because these elements are not present in the spoken language that they aren't really *in* the language itself.

    If there were no distinctions made in either the written or spoken lanugage, then sure, distinctions that can be made in your head but aren't reflected in what the language can express are not part of the language. But if it is a distinction that can be made in your head and can also be expressed in the witten language (even if not in the spoken language), it certainly is part of the language.

  68. Ruben Polo-Sherk said,

    October 1, 2012 @ 3:18 am

    One thing I left out of my previous post (although I mentioned this idea in earlier ones) is that things like 早い・速い certainly *can* be just one word, with both 'fast' and 'early' in its semantic range; it's just that that's not the end of the story.

  69. Yoshi said,

    October 1, 2012 @ 10:27 pm

    I enjoyed the discussions above. As a Japanese who can read Japanese much faster than any other language, it is always surprising to see someone who wonders why Japanese does not abandon kanji. I'm sure Japanese language would not malfunction by getting rid of kanji — there would be workarounds, but of course that cannot be the reason to throw away the cultural assets.

    I'm not saying that the tradition rules over efficiency. Rather, it's about the tradeoff between richer nuances, more visual saliency and distinctiveness, and larger costs of learning and writing. The latter is more likely to surface to consciousness, and thus it is dangerous to assume that the latter outweighs the former. As long as there is no evidence that kanji is hurting Japanese's economy or well-being or whatever, it had better be preserved. If lost, the unique and intricate writing system of Japanese would never be recovered and the world loses its writing system diversity, at the expense of uncertain benefits.

    The computer technology, in my opinion, works in favor of ideograms, because it has an effect of lowering the costs of memorizing and writing. If I can be fanciful here, I expect that in future more colorful and picturesque characters will be invented and added, because a character no longer needs to consist of monotonic strokes. What? The Japanese text messages are already full of emoji?

  70. Victor Mair said,

    October 2, 2012 @ 7:12 am


    I'm not sure that I've espoused a position of "fighting… against kanji usage in Japan". I thought that I have simply been describing trends and reporting conditions.

  71. Chromatix said,

    October 3, 2012 @ 2:33 pm

    English has a pretty horrible orthography, with some common words having little relationship between sound and spelling – for example "does" could be either plural female deer or present tense of do, each with a different sound. But since most of the less common words follow a relatively simple set of rules, you don't need to learn the whole dictionary by rote. That I think makes it much easier to learn than kanji.

    There is at least one country though that has massively simplified it's orthography in recent times – Finland. This came about through some accidents of history which effectively changed the language of the ruling class from Swedish (whose orthography is equally bad as English) via Russian to Finnish – Finnish having until then been the peasants' language. In the mid 19th century the first railways were built with the civil engineering maps annotated in Russian and the first timetables in Swedish. By the turn of the century, both types of document were produced in Finnish. And because Finnish orthography was laid down by an academic within the past 300 years, it is fully regular (and, happily, so is most of the grammar).

    The only problem is that spoken Finnish is not the same as book Finnish. You can happily speak book Finnish and be understood, but it has the same effect as Received Pronunciation does in English.

    Incidentally, I am in the early stages of learning Japanese, and have already decided that I won't bother trying to learn the handwriting part. Why handicap myself when merely reading and understanding will be tricky enough, and the computer I always have to hand allows me to parley those skills into writing ability anyway? If anything it I'd gratifying to learn that native Japanese feel the same way.

  72. Josh McNeill said,

    October 5, 2012 @ 8:55 pm

    I suppose, just to add fuel to the fire, I've recently come across another potential issue with kanji. I asked my native-speaking Japanese teacher whether I should pronounce 開く as ひらく or あく when it's displayed on a button on my computer that I know to mean "open the selected file." She gave a quick answer then immediately withdrew it, thought for a second in a confused manner, then told me she'd have to get back to me. It's somewhat remarkable to me that a native-speaker would have so much trouble figuring out which pronunciation is correct in the given context. I'm sure she's seen this button before, so how did she "read" it? My guess is that she didn't and just registered it the same way one would a picture of a stick figure man or woman over a restroom door. Either that, or the kanji actually ADDS ambiguity between clearly (phonetically) distinct words to the point where a native-speaker might not understand the difference.

    I really appreciate what elessorn and Yoshi are saying, though. They make a very convincing case for the cultural value of a writing system, regardless of how complicated it is. I am curious how Yoshi would determine whether the writing system is hurting Japan in any way. It seems like a good principle but a bit vague. Is it simply about literacy rates? Is it even known whether Japan's literacy rate is supported by things like furigana despite the use of kanji? What about lost time? It takes massively more time to learn kanji than kana; what if all that extra time gained meant more could be learned in other areas? It's not as if Japan is failing in education now, so it's hard to say that kanji is hurting in this way, but that doesn't necessarily mean that the gained time wouldn't improve education even more.

    I'm aware that, as an outsider, my opinion isn't very important, though. I don't think the opinion of any outsider should alone evoke a change unless there is a real, severe, demonstrable problem (I'm thinking of literacy in Vietnam when they used Chinese characters, for instance). But I also think it's important for insiders to get outside perspectives. It's difficult to figure out that you're living in thick dangerous smog if you've never seen a clear sky before. Don't take that to mean that I think fully phonetic systems are the heavenly sky while the kanji system is smog, though, as I really don't know the answer to my questions above and am still pretty convinced by the cultural value argument.

  73. Yoshi said,

    October 19, 2012 @ 10:20 pm

    @Josh McNeil
    Yes, it is difficult to argue against the claim that kanji is taking up the study hours of Japanese children that could otherwise be used for something more useful. American children have to learn the spelling of each word, but probably that does not sound as daunting as learning kanji, as the spelling is much more predictable than kanji even with the notorious orthography of English.

    But my point was that we would miss the whole picture if we only discussed the time taken by the early stage of learning. Once memorized, kanji could facilitate reading. In my opinion it is not crucial whether they disambiguate homonyms. English does have a bunch of ideograms so I can make a point with English examples: I suppose many people would agree that % can be recognized faster than percent, and 1,625 can be more easily recognized than one thousand six hundred and twenty five. Japanese is just doing this in larger portion of lexicon. And if you think that the number of the distinct symbols used in English is just right, while Japanese has too many, that might be just because you are used to English.

    Also, It is sometimes pointed out that the Japanese writing system is suitable for speed reading. Because content words are visually distinct from grammatical words by design, it is easier to pick up keywords while flipping pages. Although, of course, I cannot tell whether that's just because I'm more used to Japanese.

    By the way, as for your example of 開く, it must be ひらく, because あく is an intransitive verb that does not fit the context. But your main point is right — it happens all the time that you know the meaning but have only a vague idea about the pronunciation. But as in your example of restroom, there is no problem as long as you know what they mean, right? And that's also why Japanese and Chinese sometimes successfully communicate with a sheet of paper. Not having phonological representation is sometimes a good thing.

  74. Josh McNeill said,

    November 13, 2012 @ 12:22 pm

    Haha, thank you for finally clearing up which pronunciation to use for that particular version of 開く. My professor never did get back to me on it.

    I completely agree that, once learned, kanji provide a number of advantages that don't exist in alphabetic writing. This post actually led me to do a research paper on the Meiji reform attempts for one of my classes this semester and it's certainly eye-opening. Outside of the hurtle of learning the kanji (a pain I'm currently working through), I can really find no strong reason to abandon the system. This seems to be the only argument the Meiji reformers had against it, too, but even then it was arguably more a problem with the archaic style than the actual script.

    It all sort of makes me wonder why Professor Mair is so strongly in favor of pinyin in China, unless it's just that their education system is simply not universally strong enough to get over that first hurtle of learning characters to begin with, a circumstance which doesn't seem to exist in Japan.

  75. internet history said,

    February 21, 2013 @ 8:24 am

    wow! finally i have found the information i was looking for? thank you for providing this great info..
    thanks again

RSS feed for comments on this post