It is often said that Japanese contains a large number of homonyms because there was no way to render tonal Chinese words into Japanese. This makes complete sense. But I am looking for specific examples of Sino-Japanese words which are homonyms but where their original Chinese counterparts are not homonyms, i.e. the Chinese words have tonal differences, but the Japanese words do not.

  • Have you tried looking up Japanese homophones in a Chinese dictionary such as YellowBridge? yellowbridge.com/chinese/dictionary.php
    – mamster
    Mar 21, 2018 at 13:57
  • @mamster I'm not sure how you would do that. For example, if I want to search for the Japanese word きゅうよう (which is a homonym), how can I search for that in the dictionary you linked?
    – kandyman
    Mar 21, 2018 at 14:20
  • For future reference, you’d do this by searching for the words in kanji, which are usually the same as traditional hanzi.
    – mamster
    Mar 21, 2018 at 15:18
  • @mamster The tones that are relevant to the question are the four tones in Middle Chinese. The Mandarin tones, while having a correspondence to Middle Chinese tones, aren't a 1-to-1 relation, and most dictionaries (including yellowbridge) won't give Middle Chinese reconstructions.
    – dROOOze
    Mar 22, 2018 at 0:44

2 Answers 2


Keywords: MC, Middle Chinese; OC, Old Chinese: MJ: Middle Japanese; OJ, Old Japanese; 呉, Go'on; 漢, Kan'on; 唐, Tō-on; /(absence of superscript)/ or 平, level tone; /X/ or 上, rising tone; /H/ or 去, departing tone; /p̚/, /t̚/, /k̚/, or 入, entering or checked tone*

On'yomi homophones are numerous, but the loss of syllable distinction comes from multiple sources, of which loss of tones upon importation into Old - Middle Japanese from Early / Late MC is just one. There are at least two other major sources:

  • Phonological shifts in the Japanese language itself. If you've dabbled in historical kana orthography, you may recognise patterns when going from historical spellings to modern spellings. In fact, the historical kana spellings for many on'yomi represented a more accurate mora-sequence approximation of MC syllables than modern Japanese spelling, but these written distinctions were later lost in spoken Japanese. E.g.:

    • 「協」(呉音: originally [けふ]{kefu}, MC /ɦep̚/),「京」(呉音: originally [きやう]{kyau}, MC /kˠiæŋ/), now both reduced to きょう
    • 「帳」(漢音: originally [ちやう]{tyau}, MC /ʈɨɐŋH/),「兆」(漢音: originally [てう]{teu}, MC /ɖˠiᴇuX/), now both reduced to ちょう
  • The waves of importation of vocabulary, with readings now classified as 呉音, 漢音, and 唐音. All of these are snapshots of both different eras of Middle Chinese phonology and different Chinese topolects during those eras. Combined with the above phonological shifts in Japanese, you can get even syllables which sounded vastly different in different topolects of Chinese in different eras (not even considering tone) converging in Japanese. E.g.

    • 「清」(唐音: しん, MC /t͡sʰiᴇŋ/), 「森」(呉音: originally [しむ]{simu}, MC /ʃˠiɪm/), now both reduced to しん
    • 「生」(呉音: originally [しやう]{syau}, MC /ʃˠæŋ/ ),「匠」(漢音: originally [しやう]{syau}, MC /d͡zɨɐŋH/), now both reduced to しょう; for reference, the 漢音 of「生」and 呉音 of「匠」are せい and originally [ざう]{zau}, respectively.

The applicable class of words to answer this question are then those which had syllables distinguished in MC by tone only and have merged in OJ purely due to this loss of tone information upon importation. E.g.

  • 京 (/kˠiæŋ/, 平) vs. 敬 (/kˠiæŋH/, 去) both became [きやう]{kyau} -> きょう;
  • 真 (/t͡ɕiɪn/, 平) vs. 震 (/t͡ɕiɪnH/, 去) both became しん
  • 朝 (/ɖˠiᴇu/, 平) vs. 兆 (/ɖˠiᴇuX/, 上) both became [てう]{teu} -> ちょう;
  • 理 (/lɨX/, 上) vs. 吏 (/lɨH/, 去) both became り
  • 相 (/sɨɐŋH/, 去) vs. 想 (/sɨɐŋX/, 上) both became [さう]{sau} -> そう

*Entering/Checked tone is not really a tone.

  • Wow, this is very surprising. It was my understanding that the failure to discern tones was directly responsible for the existence of many Japanese homophones. For example, Richmond (2005) claims that "morphemes which could be distinguished in spoken Chinese from their differences in tonal qualities, were all assimilated as one indistinguishable morpheme in Japanese". He gives the examples 馬, 麻, and 間 as being phonologically distinct in Chinese but homophonous in Japanese. But your point would be that in Early Middle Chinese that wouldnt have been the case, right?
    – kandyman
    Mar 22, 2018 at 14:11
  • 1
    @kandyman sorry, I made a huge mistake when copying the MC reconstructions and misinterpreted them. Please see revised answer. BTW, 馬, 麻, and 間 is a very strange set of examples to demonstrate homophony; only 馬 and 麻 are pronounced with the same syllable in Chinese, with [間]{ま} being a kun'yomi and so is irrelevant; also the common Jōyo pronunciations of 馬 and 麻 are ば and ま, respectively, and so aren't homophones.
    – dROOOze
    Mar 22, 2018 at 15:32
  • @ droooze Thank you again for your thorough answer. I'd like to use those examples in a presentation. Do you know of a good academic source or paper from which I can provide some supporting citations?
    – kandyman
    Mar 22, 2018 at 17:38
  • @ droooze Another comment I have is this: if the number of homophones caused by tonal difference is as small as you say, then the majority of the homophones must be produced by waves of importing and phonological shifts, as you mentioned. But are those factors really enough to explain the large number of homophones in Japanese? It seems a bit of a stretch to me, but I'm certainly no expert in Old Japanese/Middle Chinese.
    – kandyman
    Mar 22, 2018 at 17:49
  • @kandyman The first cause of the majority of the homophones seems to have occurred when OJ tried to capture complex MC syllables (tone is but just one aspect of this). See en.wikipedia.org/wiki/… for the kinds of loss of phonological information that occurred. There were only 3 distinct phonemic tones in MC, but over half of that chart details 4-6 consonants that were collapsed into 1-2 consonants. This collapse occurs on average slightly less than twice in a syllable (occurring almost always both in the onset and final).
    – dROOOze
    Mar 22, 2018 at 18:27

The most homophones that I know of are the various words pronounced きかん (ordered by frequency in Japanese)

  1. 機関 jīguān 機關・机关 engine/institution
  2. 期間 qī​jiān 期间 time interval/period
  3. 器官 qì​guān organ
  4. 基幹 jīgàn mainstay/nucleus
  5. 帰還 guīhuán 歸還・归还 repatriation/return
  6. 気管 ​qìguǎn 氣管・气管 trachea
  7. 旗艦 qí​jiàn 旗舰 flagship
  8. 季刊 jì​kān quarterly (e.g. magazine)
  9. 奇観 ​qíguān 奇觀・奇观 wonderful sight

Formatting: kanji Mandarin pīnyīn [traditional Chinese・]simplified Chinese English.

In (Modern) Chinese all are pronounced differently.

In (Modern) Japanese, they can be only partly differentiated using pitch accent. The pitch accent given in 新明解 日本語アクセント辞典 groups them as follows:

  • きかん【LHH】 ⁴基幹、⁵帰還、⁶気管、⁸季刊、⁹奇観

  • きかん【LHH】、きかん【LHL】 ⁷旗艦

  • きかん【LHL】、きかん【LHH】 ((新は きかん【HLL】)) ¹機関、²期間、³器官

Note that 94% of the occurrences of きかん (numbers from the written language) are 機関 and 器官. For disambiguation, context is always king.

  • Great work, thanks very much. Can I ask for your source on the Chinese translations?
    – kandyman
    Mar 21, 2018 at 17:38
  • Unrelated to the question, but how did you format the words to include red pitch accents with lines, small superscript, etc?
    – kandyman
    Mar 21, 2018 at 17:39
  • The "translations" (more loose associations) are given without warranty — the English words are somewhere between the Chinese (I used mdbg.net/chinese/dictionary and dict.leo.org/chinesisch-deutsch) and the Japanese (jisho.org). The translations are not really related to the question anyway, but I thought it would be better to have imprecise translations than no translations at all.
    – Earthliŋ
    Mar 21, 2018 at 19:31
  • For general formatting, see How should I format my questions on Japanese Language SE? — for pitch accent it's part of the furigana engine, giving the "reading" as, say, LHLLHL. The superscript numbers are part of Unicode (and are part of my keyboard layout), but you can achieve the same using the HTML tags <sup>...</sup>. Finally, you can always peek at the code of an answer by clicking edit under the question.
    – Earthliŋ
    Mar 21, 2018 at 19:37
  • I don't think that it's appropriate to use Modern Mandarin pronunciations to answer the question. Some of these were still distinguished in Middle Japanese and them becoming homophones in Modern Japanese is to do with sound shifts from Middle Japanese to Modern Japanese, not to do with the loss of tones in these words upon importation from Chinese.
    – dROOOze
    Mar 22, 2018 at 0:41

You must log in to answer this question.

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