The “science” of Kanji, part 2: readings

This is part 2 of a series on how Kanji work in Japanese, again, based on my experiences so far. You can read part one here. The last post covered radicals, while this one will cover readings of Kanji.

The key to understanding how to read kanji is knowing that there are two different ways to read a given Kanji. To understand why there are two ways, let’s look at history a bit. Kanji are Chinese Characters, in that they are natively used in China. As Japan, Korea, and Vietnam came into contact with China, they absorbed a lot of the culture and philosophy. So Confucian thought has a strong influence on all of the culture above for example, but also they borrowed lots of Chinese words and writing.

Japan already has a language of its own. In fact Korean, Okinawan and Japanese language belong to a language family entirely different from Chinese. There’s no relation whatsoever, but since they didn’t have a writing system of their own, they retrofitted Chinese Characters to describe words in their native languages. At the same time, if they borrowed a word from China, they just borrowed the word and gave it a more native-sounding pronunciation.

So the result is that Chinese characters can be used to describe native words, or they’re used to describe compound words originally imported from China.

Take a classic example, 車, which means cart, vehicle, car, etc. If you read it the “native” Japanese way, called the kun yomi reading, it can be read as kuruma. But the Chinese-style reading, or on yumi reading is sha. So, depending on the context, either you read this as sha or as kuruma.

As a general rule of thumb (but by no means definitive):

  • Stand-alone words use kun yomi
  • Place names in Japan often, though not always, use kun yomi
  • Surnames often use kun yomi
  • Many, many compound words use on yomi
  • Buddhist sutra chanting uses on yomi almost exclusively2
  • Shinto liturgy and prayer uses kun yomi almost exclusively3

So, when I am talking about my car, I would call it kuruma, but if used in a compound word like nihonsha (日本車, Japanese car), or sensha (戦車, tank) you can see the Chinese-style on-yomi is used instead.

That’s a relatively easy example. What about something trickier like 便 which has a native Japanese reading, たより (tayori), but has two Chinese-style readings, ビン (bin), ベン (ben)? 1 This is a common example where a kanji can have multiple Chinese-style and/or Japanese-style readings. So, in a compound word, you could see either 不便 fuben (inconvenient) or 便箋 binsen (stationary). There’s no hard-fast rule which Chinese-style reading would be used for which word. As you come across new words that use this, they could go either way.

Generally though, usually kanji don’t have a lot of Chinese-style readings, so at most you’d have to memorize two, maybe three for a character.

Another dimension to this is that kanji can also be used in verbs, and often are. An extreme example is 下 and 上. These two words have many, many ways to read them, both Chinese-style and Japanese-style.

Let’s take 下 for example and just look at Japanese-style kun-yomi:

  • shita – 下 – down
  • shimo – 下 – down, bottom, lower
  • sagaru – 下がる – to be hung down, lowered
  • sageru – 下げる – to hang something down
  • oriru – 下りる – to descend, go down stairs
  • orosu – 下ろす – to lower something (e.g. hands)

These are all verbs and nouns that come from using this one kanji. It’s an extreme example, but it gives you an idea how many different ways there are to read the Japanese-style reading. You’re not expected to memorize all the different readings of 下. Instead, as you learn new vocabularly, you can see how it gets used in so many different ways.

So how do you remember so many different readings for so many different kanji? I think this is the question Japanese-language students want to ask most and I think there’s two different things you have to do at the same time:

  1. Invest the time to study kanji. Kanji flashcards, especially professional ones, are highly recommended, but they provide accurate information, and require less time to setup. But you can’t just force your way through and memorize them. Instead, you have to develop “flight time” by visually exposing yourself to that kanji over and over again over the long-term. If you see a kanji 10 times, you start to remember it. If you see it 50 times, you get pretty comfortable. If you see it 200 times, it becomes rote.
  2. The above helps, but is not enough by itself, because you can’t construct most words from single kanji. So the other half of this approach is to just learn vocabulary words, and keep learning new words. I believe that much of kanji learning happens by reverse-engineering words you learn, figuring out the kanji they contain. For example, if you learn two totally different words, but they have a common kanji, you can start to figure out that this kanji is read such-and-such way. For example, take words like 消去 (shōkyo to purge, erase) and 消毒 (shōdoku to disinfect). If you never encountered 消 before, you can already figure out that the Chinese-style reading is shō and it has the context of erasing, purging, etc. If you encounter a third word like 消費 and can begin to guess how its read and its meant. The word is shōhi or “consumption” since it “erases” expenditures (the second kanji).

These are just some suggestions. If you do both at the same time, you’ll find one compliments the other, but always try to keep your study of Japanese balanced, not lop-sided, or you will continue to struggle. Good luck!

1 Notice that for native-Japanese readings, it’s spelled with hiragana, while with Chinese-style readings, with katakana since they’re technically “foreign” words.

2 The Chinese Buddhist canon is a very, very extensive set of Buddhist texts used throughout East Asia, and were imported into Japan as the Taisho Tripitaka. Japanese readings of texts maintain the Chinese-style reading as much as possible, hence, the use of on-yomi when doing Buddhist chanting.

3 Shinto is a native Japanese religion, and much of the vocabulary existed in some form before the importation of Chinese culture. Hence the native kun-yomi. In Shinto chanting, if they recite the “sacred numbers”, these are just the numbers 1 through 10, 100, 1000 and 10000 recited in native Japanese pronunciation: hito, futa, mi, yon, i … hyaku, chi, yorozu. This was from the book The Essence of Shinto: Japan’s Spiritual Heart.

About Doug

A Buddhist, father and Japanophile.
This entry was posted in Buddhism, Japanese, Language, Shinto. Bookmark the permalink.

5 Responses to The “science” of Kanji, part 2: readings

  1. ロバート says:

    Did you mean to say that Shinto liturgy uses kun yomi almost exclusively (rather than on yomi like Buddhist liturgy) ?

    My favourite example showing onyomi and kunyomi for the same kanji is illustrated by two words that mean the same thing, but the kanji inverse and the reading changes
    切腹 seppuku
    腹切 harakiri
    both mean slitting one’s belly in ritual suicide. Seppuku is on-on and harakiri is kun-kun.

    Two further points about readings.
    Onyomi can be further broken down into when the word arrived in Japan (in 3 waves I think). I believe Japanese dictionaries also give this information. I doubt it is widely known or studied in school however. And I would have to look up a book (I don’t have here) to give you any examples.

    To return to radicals. A significant proportion of kanji use radicals as a phonetic component rather than a semantic component. Therefore it is possible to make educated guesses about an unknown kanji’s potential reading. You can also group kanji by shared phonetic components and readings to help you remember them.
    形声文字 (85% of kanji) have a semantic component and a phonetic component
    for example 寺 in these kanji gives the on yomi ji (but no semantic reference to a temple, perhaps other than one that is made up ) 時 侍 持
    (but note not in 特 doku or 待 tai)
    Again unfortunately I’ve never really come across a reference that spells this out for all kanji in a systematic manner.

  2. ロバート says:

    oops you corrected it as I wrote my comment!

  3. Doug says:

    Ha ha ha, yeah, I caught it a bit late. I bet Google Reader already sent out the incorrect version to a few people. ;)

    As for the phonetic component, I’ve noticed this too, though I wasn’t sure if it was just a coincidence. In my kanji studies for JLPT3, I noticed a few similar looking kanji all had the same “on yomi”, which now tells me it is not just a coincidence. :)

  4. JonJ says:

    I think a lot of linguistic scholars doubt that Korean and Japanese belong to the same family, despite some resemblances. The last I heard, there was no consensus among the experts about what family Japanese belongs to. It also has some resemblances to Polynesian languages, apparently. In any case, it is certainly not related to Chinese.

    I often say that the Japanese were very unfortunate living off the coast of China/Korea, and not the coast of Europe, as the English did, because then they would have picked up the Latin alphabet the way the English did, and saved a lot of English-speaking students of Japanese a lot of trouble in later centuries!

    And I want to underline your point that learning kanji takes a lot of time and practice. A good way to learn the Buddhist virtue of patience and fortitude. Reading a lot of varied materials and (if you’re in Japan) seeing them on many signs as you walk around the streets is important.

  5. Doug says:

    Hi JonJ,

    You’re right in that the origins of Japanese are still not clearly understood. However, I do recall a native Korean speaker once telling me that when he learned Japanese, the grammar was exactly the same, just that all the worlds and reading were different. Importing words is one thing, but grammar’s not something you can import very easily, and Japan’s colonial rule wasn’t enough of an influence (look at Taiwan ;) ), so I believe they are largely the same. But that doesn’t rule out other influences such as the Polynesian ones. :)

    P.S. From what I’ve heard, those researchers doubt the connection of Japanese/Korean with the Altaic language family, which was an older theory linking the two. I understood it to mean that Korean/Japanese/Ryukyuan were a language family unto themselves. :-/

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <pre> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>