Columns on View All Columns
Visit ELTBOOKS - all Western ELT Books with 20% discount (Japan only)

Thoughts on Japan

Kingaku kara no omoi - 金額からの思い

Thoughts on Japan from the National Institute of Japanese Studies. University of Sheffield

February 27, 2009

Japanese is a Hard Language

Last week, I talked about all the ways in which Japanese was not that difficult a language to learn. This week, I’m going to start putting the opposite viewpoint, and mention some of the challenges to learning it. So, in my view, there are four main factors making Japanese ‘difficult’ – for native English speakers, that is. These are: lexicon, writing system, social deixis and discourse structure. I’ll deal with the first two this week, and the second pair next.

First, lexicon – the vocabulary of the language, that is – this can be problematic for English speakers, because the vast majority of the word-stock comes from either pure Japanese or Chinese roots, meaning there’s a proportionally greater burden on the memory in learning the words of Japanese than there is a language like French, for example, which has much in common with English. Something like sixty percent of the word stock of modern Japanese is the result of borrowing from Chinese, with the remainder being pure Japanese. It’s true, of course, that in more recent years there has been considerable linguistic borrowing from English, but it is impossible to get by just with katakana 片仮名 gairaigo 外来語 (‘loan words’). Even where borrowings do take place from English, they often change beyond all recognition once they’ve been adopted into Japanese. An example which is frequently given is the element kon コン, which crops up in words like pasokon パソコン (‘personal computer’), eakon エアコン (‘air conditioning’), and mazakon マザコン (‘Oedipus complex’). As you can see, three terms with wildly different senses in English have been conflated into a single one in Japanese, meaning that your English knowledge will do you little good. It just means, though, that you have to put more effort into learning Japanese vocabulary – something that’s tedious and time consuming, but not difficult per se.

I also acknowledge Mr Torley’s point about there being proportionally fewer words with a one-to-one correspondence in meaning between English and Japanese than there are between, say, English and French: this is certainly true, but I am less convinced that this is as significant a problem for human language learners as it is for machine translation. Machines – and programming – like one-to-one matches, but the real world rarely works that way, and the human mind recognises it. Actually, one of my very first lecturers as a student in the mid-1980s was working on Japanese-English machine translation at the time. He argued, quite convincingly, that because of its syntactic regularity, it was relatively easy to ‘teach’ a machine how Japanese ‘worked’ – particularly as the verb plays such an important role in a Japanese sentence, and the various affixes (negative, passive, etc.) always occur in a predictable order. (You can even do this with classical Japanese, which is much more morphologically complex than the modern language.) The difficulty arose in getting the programme to choose accurately between different possible interpretations of the same sentence, given that the language displays a relatively high degree of context-dependence. My lecturer eventually left university life to work full-time on machine-translation for Sharp, and I’ve often wondered how his work progressed, given that, as Mr Torley says, it’s still the case that Japanese-English machine translation works poorly, if at all.

Let’s move on to the second factor: the writing system. This is a problem, as for historical reasons, the Japanese writing system is undoubtedly the most complex in use in the world today. Briefly, the Japanese first encountered writing in the form of Chinese characters, which, for reasons I won’t go into now, were unsuited to writing Japanese. This led to the rapid development of both hiragana 平仮名 and katakana 片仮名, and also to Chinese characters in Japanese being read as Japanese words, and also with a Japanese approximation of their pronunciation in Chinese. These are what are now known as kun 訓 and on 音 readings. As if that didn’t make the situation complicated enough, many characters were borrowed more than once, after their Chinese pronunciations had changed, meaning that they have multiple on readings, and they could also be used for multiple Japanese words, giving a variety of kun readings. To give one example, the character 生, which has a core meaning of ‘life’, has two on readings (sei and shō) and no fewer than fourteen kun readings (ikiru 生きる, ikasu 生かす, ikeru 生ける, umareru 生まれる, umu 生む, ou 生う, haeru 生える, hayasu 生やす, naru 生る, nasu 生す, ki 生, nama 生, o 生 and nari 生). As you can see, in contemporary orthography, the okurigana 送り仮名, the kana characters after the kanji, will generally tell you which reading is meant, but it’s perfectly possible to write both ikasu and hayasu as 生す, for example, meaning you have to work out from the context which of these is meant, and you just have to know the word in order to tell whether 生 is ki or nama in a given context (o and nari are readings only used in names, but again, either could be possible).

So, with its two syllabaries, hiragana and katakana, and kanji, there are a very large number of characters to learn. Hiragana and katakana are not too problematic, though – with a bit of effort you can get a reading knowledge of both of them with about 48 hours of study – the difficulty lies with the kanji. Even though you only need to learn about half the number of characters – about 2,000 – to be literate in Japanese, that you do in Chinese (about 4,000), the fact that almost every single kanji has more than one possible reading, and you need to know a word to be certain of which reading is being used, makes learning to read, and write, Japanese an immensely time-consuming and tedious process. By comparison, written Chinese is much easier, as each character has only one reading, so once you’ve learnt it, you can be confident of how it's read when you come across it in an unfamiliar word, unlike in Japanese. Again, though, there’s nothing really difficult about the learning process – you just have to put up with the tedium of banging your head against the wall of characters and readings, and accept that it’s going to be an incremental one. Obviously, it helps to be in Japan during at least some of this time, because you are surrounded by kanji all the time and can constantly reinforce your knowledge.

It’s difficult to understate the barrier the writing system places in the way of speakers of languages with alphabetic writing systems, if they want to learn Japanese, but on the other hand, you don’t want to make too much of it, either. It’s just something you have to keep working at, and prevent yourself from giving up in frustration. There will always be new kanji to learn, or new readings, and you will always find that you know some readings, but not others for a kanji, or you know the meanings, but not the readings. Part of the problem for us as English readers, I think, although I’ve no objective evidence for this, is that we are so used to associating being able to read with knowing the sound of something, that we tend to panic and think we know less than we really do when we look at a character and find that we can’t say it. What we have to also have to bear in mind is that it’s the same for the Japanese – it’s just that they have more time to learn their characters, and learn them at a time when their minds are more receptive to language learning. They also develop strategies for coping with characters they don’t know if they come across them, and any foreign learner of Japanese needs to do the same, which is something I’ll talk about in a future column.

Next week: social deixis and discourse structure.

« Japanese is Not a Hard Language | Main | Is Japanese a Hard Language? »


I agree about the writing system: it is the most complex in use in the modern world. To be literate in modern Japanese you need to know nearly 100 kana letters and about 2,500 Chinese characters (the kanji prescribed for teaching in schools plus several hundred others in common use). Waiting for a train on Tokyo Station a couple of weeks ago, I occupied my time by counting the number of different kanji you have to know in order to read the names of the stations on various lines. You need over 200 just to get round the Yamanote Line. To learn this number of kanji (and all the various readings they may have in Japanese ­ in this respect Chinese and Korean are easier) in adulthood requires a huge investment of time and effort. It's like learning to become a concert pianist or a professional soccer player. Learning to play well enough for a family sing-song, or reach the standard where you get picked for an amateur team in a Sunday league is enjoyable, life-enhancing and altogether a good thing, but to play at Carnegie Hall or be bought by Manchester United is something else, and requires dedication. These two professions also require talent. Learning to read Japanese doesn't -- or not in the same sense. It demands a certain dedication, which presupposes a certain love of the activity for its own sake, and that's a sort of talent, but it doesn't require any great intellectual gifts. If very few foreigners do learn to read Japanese well, it's for lack of motive or opportunity, not means.

I haven't, of course, asked the obvious question -- hard for whom? It may be hard for foreign learners, but is it hard for the Japanese? Well, in the developed world, which in principle demands that everybody be literate, any writing system that defeated a high proportion of the members of the language community would pretty soon be abandoned or modified. (The Japanese, Chinese and Korean writing systems were all modified in the middle of the last century, for that reason.) Many English-speaking countries anguish over their levels of functional illiteracy, but Japan doesn't seem to have that much of a problem. (No room, I'm afraid, to discuss what's meant by 'literacy', whether the level of literacy in Japan is really as high as claimed, whether or not there is such a thing as dyslexia in Japan, and so on. Let's just say that nearly every Japanese person manages to learn to read.) How do they manage to gain a command of such an immensely complex system when many native-speakers of English find it difficult to cope with a system that only requires them to learn 26 letters? I think the answer, very briefly, is that, while English uses a small number of phonetic signs and a set of spelling rules so complex as to appear to some to be random, the kana sets use a somewhat larger number of phonetic signs with a very simple and perfectly reliable set of three or four spelling rules. The consequence is that by the age of seven or eight -- often much earlier -- all Japanese children have at their command a writing system that enables them to set down accurately anything at all that they hear. You could, in principle, dictate the Equal Employment Opportunities Law (not an easy read) to a group of nine-year-olds and have them take it down perfectly accurately (as long as you told them when wa or o was a particle), even though they would understand next to none of it. From that point, learning the kanji, and substituting them as appropriate for kana, is only a feat of memory. And since the Japanese education system requires students to learn about 2,000 kanji over nine years (four or five a week) it's not really that much of a feat. Foreign learners, if they're not to lose heart altogether, need to learn at a much faster rate, and the pressure tells. If you don't feel that you're starting to get to grips with the written language in a reasonable time, the temptation just to give up is strong.

I said that learning kanji didn't require any particular talent. Is that true? Well, since about 50 per cent of the 126 or so million Japanese must have average or lower-than-average abilities, even if we assume for the sake of argument that the average level of ability of the Japanese is somewhat higher than that of say, native-speakers of English, it must still be true that 80 or 90 per cent of English-speakers have the intellectual equipment to learn to read Japanese. The message is: it can be done, but it demands determination, and, like training for a marathon, it needs to be a regular daily regime. And don't be discouraged by the fact that the kanji you've learned don't all stay in your head. That's the way it is ­ for native-speakers of Japanese, too. I've been studying Chinese and Japanese for forty years. Assuming (very conservatively) that over that period I've tried to learn on average four or five new kanji a week, some 8,000 or 10,000 kanji have passed through my head. How many are in there at the moment, I've really no idea. This week I've been working on a translation of a 19th century book which includes long and highly poetic (in the Chinese style) descriptions of the mountain scenery of Switzerland. Just at the moment I could probably write out from memory thirty or forty characters peak, pinnacle, slope, cliff, crag, lofty, soaring, dizzying, looming, towering, and so on. But in six months most of them will have retreated to the 'read only' part of my brain, and some of them will have gone altogether.

As always, Dr. McAuley has written a fine article indeed. And this time around we are blessed with a sub-article (far more than just a "comment") by Dr. Healey as well! I'd like to chime in (much more briefly) with a few points. One is that learning individual kanji characters is only part of the struggle. Learning the innumerable combinations of said characters is a staggering challenge. I have The Kanji Dictionary (1996: Spahn & Hadamitzky) at hand that boasts 47,000 character combinations. And yet I regularly (yes regularly) find compounds it does not contain.

As for literacy, I've been teaching university students for over 10 years here in Japan. My students display an embarrassing lack of kanji knowledge. They can, of course, read most non-technical books, but when it comes to writing, I'm sorry, they can't do it without the help of a machine. This is no exaggeration. Most 20 year olds in Japan simply cannot write kanji. Take away the computer or cell phone and you have the makings of illiteracy.

As for the struggles of the foreign learner, I find Heisig's creative mnemonic approach incredibly helpful. I'm convinced that this method displays a patent superiority over the rote memorization technique used by the natives (just look at the results!).

((By the way, thanks for teaching us the "kanji game" Dr. Healey. It has been most helpful!))

The devotion and joy in learning required to bash away on a daily basis; a positive attitude to, or a least, acceptance, of the inevitability of forgetting: these things I did not have when I took up Japanese for the first time 22 years ago. That is why I stopped; it was plain and simple - that legion, 'kanji', and all their readings and meanings.

Living in Japan, I have been able to pick it up again. I second Matt White's recommendation of Heisig's mnemonic system. I would add to that a system whereby new vocabulary cards are reviewed for three days daily; then weekly; then bi-weekly and so on up to bi-monthly. This has brought me to the level where I can now read most non-technical writing. I wonder whether Sheffield (where I studied, thanks to the patience of Graham Healey) shouldn't take a more active role in at least introducing a vocabulary reviewing system to students at the scratch level?

Overall, Sheffield was a great experience for me.

It seems to me that one of the major problems for foreigners in learning Japanese is the inevitable`separateness` of the tasks of learning to speak and understand to a reasonable standard (probably much easier than a lot of other languages) and learning to read and write. In the case of languages with purely alphabetic or syllabic scripts, however complex, there comes a point, pretty early on, when the learner can begin to tackle `real` texts, thereby reinforcing and complementing the speaking and understanding elements of the learning process. In Japanese, this is simply not possible unless you make a determined and `separate` effort to learn the kanji.
When I started studying Japanese, I tried to learn the kanji as I went along, at least to be able to recognize the kanji for each new vocabulary item I was picking up. This method, combined with an electronic kanji dictionary (so you can jump straight from the reading of a kanji or compound to its meaning in English) and dogged persistence at reading simple books with entertaining and somewhat predictable (i.e. guessable) storylines (it gets easier once you`ve looked up all the kanji in the first twenty pages ...), worked quite well for me, but there was a negative side effect that I still struggle with now. That is, I learnt to read a lot of kanji and was soon able to tackle newspapers and more complex texts, but I didn`t have a visual image of the sound of the words I was learning. By `visual image of the sound`, I mean that, when learning a European language, or even a language like Arabic and perhaps Korean (I`ve never tried), you can immediately relate a written word to a sound image with the result that you can readily understand spoken words that you have learnt in written form, thereby incrementing the all-round learning process. If I had been able to spend longer periods in Japan, my exposure to the spoken language would probably have compensated, but as I haven't been able to do that, I am now in the peculiar situation of being able to read quite high-level texts in Japanese, but not able to understand very much of even `standard` spoken language such as TV news (actually, very similar to many Japanese, who read English well but don`t understand it when they hear it). This seems to confirm the `separateness` theory.
On a slightly different, but related point, I agree with Dr McAuley when he says "we [English readers]are so used to associating being able to read with knowing the sound of something, that we tend to panic and think we know less than we really do when we look at a character and find that we can’t say it." This maybe explains why the Japanese are able to read so fast (I read statistics on this somewhere once - maybe one of Dr Healey`s Rambles - sorry, I can`t remember where): maybe they are just picking up the semantic information from the kanji when they read without bothering about the sound at all? Which is also why, after ten years of learning Japanese, I can read Chinese newspaper headlines ...

Recent Columns

Recent Comments




World Today