The character is a Cantonese character meaning "sleep". It is pronounced xùn in mandarin and fan3 in Cantonese. I was wondering how two so different pronunciations (I would expect xun to match syun or similar and fan to match fen or similar) came into being. I did some research. The Wiktionary has pronunciations mén, mèn, shuì for Mandarin and fan3 for Cantonese. mdbg.net has xùn on the main display, but opening the dictionary entry (clicking on the character) reveals shuì and fan3. Baike baidu has fèn, which finally matches the Cantonese. CantoDict, however, has kùn and fan3, and states the origin is 困, which in Cantonese means "sleepy", plus three other very different meanings, according to CantoDict, but is pronounced kwan3. So the Cantonese sound is definitely fan3, but the Mandarin sound? THe shuì pronunciation suggests a pronunciation loan induced by meaning. The other pronunciations seem totally unrelated. fèn matches the Cantonese sound, but the origin of the character is pronounced kùn, curiously matching the CantoDict Mandarin sound of this character. So:

  1. How was it pronounced in older times (i.e. Middle Chinese)?
  2. How did the pronunciations come into being?
  3. Which is the correct Mandarin sound?
  4. If there are more, how did they come into being? Were there semantic loans at any point?
  5. Could xùn be a misspelling of kùn? It is given by baike in several instances, though.
  6. Could mén, mèn be distortions of fèn?

Hope I was clear. My info on this character is very messy :).


Wiktionary has since updated its pronunciations to fèn-fan3. The original mén/mèn/shuì were added by Rukhabot in an edit dated 22:30, 18 August 2012, and removed by Suzukaze-c's 05:30, 22 September 2015 revision. I wonder where that bot got those pronunciations…

2 Answers 2

  1. How was it pronounced in older times (i.e. Middle Chinese)?

I haven't found a record of 瞓 in classical Chinese, but since 瞓 and 训 are both read as fan in Cantonese, I'll take 训 instead. It is read qhuns in reconstructed Old Chinese that is before the 1st century B.C. In Middle Chinese it is pronounced as hyonh.

  1. How did the pronunciations come into being?

If you mean , it is obvious that it's created after as a phono-semantic character or 形声字. the x in Mandarin xun comes from qh, then to hy, then to x. The f in Cantonese fan might have come into being along the qh -> hy -> h -> f route.

  1. Which is the correct Mandarin sound?

瞓 isn't a Mandarin word. The dictionaries say it's fèn and it comes from Hokkien or Cantonese. fèn is an approximation of the sound in the dialects.

  1. If there are more, how did they come into being? Were there semantic loans at any point?

If you mean more of "dialectic words" that don't exist in Mandarin such as 瞓, there are quite a few. They exist because there are no counterparts in Mandarin or rather classical Chinese, so people have to make new characters for these words. I don't think semantic loan happened here, but it does involve borrowing existing characters to make new ones.

  1. Could xùn be a misspelling of kùn?

It's not misspelling. See answers to question #1 and #2. k to x shift is possible. Similar is the case in western languages where the hard c (read k) shifts to a soft c which sounds similar to the Chinese x.

  1. Could mén, mèn be distortions of fèn?

Not impossible: m -> w -> v -> f, For example, in Chinese we have 没(m) 无(w) 否(f) all mean no. But regarding to your question, we need to know who uses the sound men and when it is used. In most online Chinese dictionaries 瞓 is pronounced as fèn, I didn't found any references to men.

  • 1. How was qh actually read in Old Chinese? I guess MC hy was /hj/. Other examples of the route to Cantonese fan from qh? 4. No I meant Mandarin readings. Being into Min and Hakka songs, I know about dialect-specific words with controversial spelling. For example tsia, hia, hia-ni, tsia-ni, khah and so on in Min, and as for Hakka I would wait for my friends at Hakka thian on fb to comment :). Since I've found so many mandarin readings (fen men xun kun shui), I was wondering if there were any more. Now I've come to wonder if there are other possible Cantonese sounds.
    – MickG
    Dec 15, 2014 at 6:55
  • The notation qh according to the Baxter-Sagart system for Old Chinese is a voiceless aspirated uvular stop. But note that it is mainly an abstraction from Middle Chinese.
    – Michaelyus
    Dec 15, 2014 at 14:36
  • Other examples of Old Chinese Baxter-Sagart *qh > modern Cantonese f- include 熏 and 葷. There's also *qwh for 化, 揮; *qhʕ for 呼, *ɢʕ for 乎, *qwhʕ for 賄 etc. The vast majority of these are x + some sort of glide in Middle Chinese. Lots of velar and "guttural/laryngeal" Middle Chinese initial consonants developed to f- in Cantonese before the 合口 glide w.
    – Michaelyus
    Dec 15, 2014 at 15:01

As Wang Dingwei notes in his answer, 瞓 is a phono-semantic character that uses 訓 (also pronounced fan3 in Cantonese) as the phonetic component to represent the word fan3 in its meaning of "sleep". However, 瞓 is a character that was invented in recent times. The phonetic 訓 was chosen because it happens to have the same pronunciation in modern Cantonese, but that does not mean that the historical phonological analysis of 瞓 should be the same as 訓.

Your CantoDict source mentions that the origin of the word is actually 困, and this is indeed the correct origin. 《康熙字典》 mentions that "tired and weary" is among its definitions:


The character 睏 was sometimes used in order differentiate it from the other meanings of 困. The modern Cantonese pronunciation of 困 is kwan3 rather than fan3 though, so how did it end up getting its current pronunciation of fan3? It turns out that it is the result of regular sound change from Middle Chinese, and in fact it's the kwan3 reading that is the exception.

Compare the following words, all of which have an f- initial in modern Cantonese but had a /kʰu-/ initial in Middle Chinese:

  • 苦 MC: /kʰuo/ > C: fu2
  • 堀 MC: /kʰuet/ > C: fat1
  • 快 MC: /kʰuai/ > C: faai3
  • 寬闊 MC: /kʰuan kʰuat/ > C: fun1 fut3

As an aside, Mandarin retained the MC /kʰu-/ initial (pinyin: ku-) essentially unaltered (cf., , , kuài, and kuān kuò, respectively).

困 was pronounced /kʰuən/ in Middle Chinese. 《康熙字典》 mentions that 困 has the fanqie 苦悶切, so by analogy it should have the same initial as 苦 and should be read as fan3 rather than kwan3 in modern Cantonese if the sound changes were consistent. However for some reason its literary pronunciation in modern Cantonese happened to preserve the older /kʰu-/ initial.

Given this, the answers to your first three questions should actually be analyzing 困 rather than 訓:

  1. How was it pronounced in older times (i.e. Middle Chinese)?

As mentioned previously, 困 was pronounced /kʰuən/ in Middle Chinese.

  1. How did the pronunciations come into being?

Cantonese f- came from Middle Chinese /kʰu-/. MC /kʰ-/ had at one point become /h-/ somewhere in the course between MC and Cantonese. /h-/ initial words that did not have a /-u-/ medial remained /h-/, for example:

  • 開 MC: /kʰai/ > C: hoi1
  • 看 MC: /kʰan/ > C: hon3
  • 可 MC: /kʰa/ > C: ho2
  • 客 MC: /kʰɐk/ > C: haak3
  • 肯 MC: /kʰəŋ/ > C: hang2
  • 口 MC: /kʰəu/ > C: hau2

However, those with the medial (i.e., /hu-/) underwent a further change to become f- in modern Cantonese, as previously noted.

As an aside, the /kʰ-/ to /h-/ shift is not an unusual sound change. A similar shift happened in the Germanic languages (e.g., compare Latin canis vs. English hound, cor vs. heart, centum vs. hundred, etc.; see Grimm's Law).

  1. Which is the correct Mandarin sound?

This is simply the pronunciation of 困 in Modern Mandarin, which is kùn.

Your remaining three questions are in regards to the character 瞓 itself rather than the actual word that fan3 was derived from (困). Because it's a character that was invented to represent dialectical pronunciations, it's hard to determine how the other "Mandarin" pronunciations came about without additional references to where they were found, as Wang Dingwei mentioned in his answer.

  • I don't think you can claim that fan3 is predicted given regular sound changes. You seem to propose that /kʰu-/ to /hu-/ to /f-/ in all syllables is a regular sound change but there are too many words in Cantonese with a kh- and khw- initial for that to have been a regular sound change. I think that certain vowels and finals conditioned the sound change so that only a subset of syllables underwent the /kʰu-/ to /f-/ set of changes.
    – HAL
    Jul 3, 2015 at 3:28
  • @無色受想行識 I'd also like to see Claw's response to your question. However, I think you're using an incorrect definition of "regular sound change". "Regular" doesn't mean "unconditional"; your proposal that "certain vowels and finals conditioned the sound change" is an example of regular sound change. Jul 23, 2015 at 14:50
  • I haven't had the chance to look into this yet, but will provide an update once I've done so. You could also pose this in its own question so that others can chime in.
    – Claw
    Jul 24, 2015 at 16:45
  • 1
    Wiktionary added fan3 to the readings for 睏. I think this is correct given 睏 is the normal word for "sleep" in Min and fanqie initials of most khw- to f- words are 苦 (see 塊-苦對切 堀-苦骨切 寬-苦官切 快-苦夬切). The other khw- readings in Cantonese seem to derive from /kˠ/ 愧-俱位切 or /g/ 葵-渠追切 狂-巨王切 except for the 夸 series 苦瓜切. This is possibly why 郑张尚芳 used kʰˠ for MC 夸 series. Mar 16, 2020 at 14:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.