If I am reading a Chinese text to a Japanese audience, how can I know which reading to use for each character? Of the 音読み, there can be 呉音, 漢音, and 唐音 to choose from.

Also for pronunciation, would rhythm be modified to stress individual characters in deference to their monosyllabic origin in Chinese speech?

(I am a beginner and have no background on this topic.)

For example if I was trying to impress a Japanese Buddhist scholar by reciting the 「般若心経」, a classical Buddhist text written in Chinese, how would I pick the readings for the first sentence: 「観自在菩薩行深般若波羅蜜多時、照見五蘊皆空、度一切苦厄。」 As far as I know there is no Japanese translation with words rearranged and inserted. This is probably not a useful skill since I doubt I will ever find another person interested in hearing me recite classical Chinese literature in Japanese. In fact, I could probably just pick readings as I choose for mnemonic purposes only, but I was curious if there is a set of 'correct' readings.

  • By 漢文, are you referring to 漢文訓読? For example, given the Chinese 人不学不知道, this could be rendered into Japanese as 人学ばざれば道を知らず. As a form of written translation, pronunciation and rhythm are typically not taken into much consideration. Also, the generated Japanese is semi-classical so does not make for good spoken reading for a general audience. Rather, you'd likely then need to translate this to modern Japanese. I suggest updating your question with more details with a few pertinent sentences.
    – Dono
    Commented Mar 10, 2013 at 3:30
  • Thank you for the detail. Not 漢文(訓読) as I was expecting. 般若心経 is a very popular Buddhist text in Japan. Many, many people spend years memorizing the text. The start is so famous that even non-religious people are likely to know this. It is 観自在菩薩行深般若波羅蜜多時(かんじざいぼさつぎょうじんはんにゃはらみったじ)照見五蘊皆空(しょうけんごうんかいくう) 度一切苦厄(どいっさいくやく). Though I would consider these to be the normal readings for these characters.
    – Dono
    Commented Mar 10, 2013 at 4:34
  • Sorry, I didn't know what 漢文訓読 was until now. Do typical Japanese people (the general public included) study only 漢文訓読 texts or do they venture into the original as well?
    – HAL
    Commented Mar 10, 2013 at 4:38
  • 3
    漢文訓読 (more commonly just called 漢文) is a subject taught in middle or high school. The text is classical Chinese and we learn how to read it into semi-classical Japanese. It is an old practice going back over a 1000+ years ago. It is a systematic form of automatic translation. Normally, texts are not read in full with the Japanese approximation of Chinese readings. But for religious purposes, some may do this to remain true to the original.
    – Dono
    Commented Mar 10, 2013 at 4:49
  • 2
    It is not as simple as reading each character in isolation. There are words made up multiple characters which typically have a fixed reading found in a dictionary. 般若 is one example, and is always はんにゃ. In this sense, knowledge of this or a dictionary may be necessary. Also note that the various 音読み readings arrived in Japan during specific periods. The age of the text (or when the text arrived in Japan) may also affect which 音読み is used.
    – Dono
    Commented Mar 10, 2013 at 5:29

1 Answer 1


It's a question somehow unanswered for a fairly long time, while it doesn't seem to be too complicated to answer.

If I am reading a Chinese text to a Japanese audience, how can I know which reading to use for each character? Of the 音読み, there can be 呉音, 漢音, and 唐音 to choose from.

There are two major types of situation to read Classical Chinese text in Japanese you might encounter: buddhist sutras, and all others.

In the latter case, it's usually recited through 漢文訓読, where all untranslated words (= remained as 音読み) are read prescriptively in 漢音 (though the rule was established relatively recently that there's no assurance older documents before Meiji conform to it).

兄弟 (brothers)

  • everyday: きょうだい (both in 呉音) 「うちは三人兄弟だ」 My family has three brothers. ("I (who's male) have two brothers / I have three boys.")
  • 漢文訓読: けいてい (both in 漢音) 「人は皆兄弟有り」 Other men all have their brothers.

Anyway, in most cases you don't need to practice the method first-hand, unless you are an enthusiastic Japanology researcher, because almost all of well-known texts already have translations, and what you have to do is just follow the textbook.

The buddhist sutras, on the other hand, are read as is and the pronunciation differs roughly between sects and schools. The largest portion of them use 呉音, but according to here and here, 天台宗{てんだいしゅう} has some 漢音 sutras, and 黄檗宗{おうばくしゅう} (a Zen sect) uses 唐宋音 in every piece.

南無阿弥陀仏 (Namo Amitābhāya)

  • 呉音: なむあみだぶつ
  • 漢音 (天台宗-style): なもあびたふ
  • 唐宋音: なむおみとふ (Compare to Contemp. Mandarin nāmó ēmítuófó)


  • 呉音: かんじざいぼさつぎょうじんはんにゃはらみたじしょうけんごうんかいくうどいっさいくやく
  • 唐宋音: かんつざいぷさへんしんぽぜぽろみとすちゃうけんういんきゃいくんといちぇくえ
    (Again, compare to Contemp. Mandarin guānzìzài-púsa xíng shēn bōrě-bōluómìduō shí zhàojiàn wǔyùn jiē kōng dù yīqiè kǔè)

Note that sutras contain many Sanscrit and Pali transcriptions that often deviate from dictionary reading of kanjis.

Also for pronunciation, would rhythm be modified to stress individual characters in deference to their monosyllabic origin in Chinese speech?

Yes, in buddhist tradition. In this YouTube clip you'll notice that every character (syllable) is pronounced in the same length.

  • Great answer, especially the video. Quick question about うちは三人兄弟だ. Does this mean the speaker has 3 siblings (implying the family has 4 children total) or that the family has 3 children total?
    – HAL
    Commented Jan 8, 2015 at 0:37
  • Wow, I don't know why I've overlooked such a mistake. The "My family has" part is the correct one. I'm going to edit it. Commented Jan 8, 2015 at 4:09
  • By the way: it doesn’t seem the 漢音 norms are THAT compulsory. I mean the cases when there are established 慣用音 pronunciations. Surely, no-one forces to say 不 as フウ (or フツ)? What does the norm say on such exceptions? Commented Nov 26, 2019 at 22:01
  • @AlexanderZ. My answer focuses on 呉音/漢音 contrast and 慣用音 is an orthogonal factor. 慣用音 has many types, for example, those by pure analogy should be avoided, but that the merge of -p and -t is wrong or a natural development is something hard to say. FWIW フ of 不 is a correct 漢音 form: id.nii.ac.jp/1629/00005863 Commented Nov 27, 2019 at 14:23
  • A more explicit example: 増 has very much a voiceless consonant and is supposed to be そう no matter what reading you choose, but in Japan the only reading used is ぞう – does that mean ぞう will be chosed, though strictly not a 漢音? Commented Dec 3, 2019 at 18:29

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .