This page on "ChinaKnowledge" as well as Pulleyblank's Middle Chinese (published 1984, p.86) say that such rhymes were placed in grade 3 hekou (合口三等) in the standard rhyme tables. However certain rhyme groups like 缶/否 are kaikou (開口) in the Yunjing, even though Pulleyblank considers them to be an "f" sound (p.89).

Is there any logic to how they are split between kaikou/hekou?

  • Ima keep it 💯 with ya chief, I aint understood a word you just said Commented Sep 18, 2020 at 5:52
  • @小奥利奥 If you're interested you could look through this page on the: Rime table.
    – Mou某
    Commented Sep 18, 2020 at 6:47
  • Baxter’s A Handbook of Old Chinese Phonology, pp. 62-63 probably has answer. He describes that labial-initial syllables (exactly 非敷奉微) have no contrast between finals with -w- and without (there are no pon <> pwon). Hence, the compilers of rhyme tables have freedom to put them to whichever class they want. Commented Sep 18, 2020 at 7:57
  • For anyone else who may be interested in this, I still have not found a completely satisfactory answer. One guess is that because the "light"/"heavy" distinction was current in the Late Middle Chinese pronunciation used by the creators of the Yunjing, it may have been "obvious" which was which. I hightly recommend Shen's "A Phonological History of Chinese" for anyone else interested in the topic.
    – bellkev
    Commented Apr 27, 2021 at 21:48

2 Answers 2


The following answer is a translation and summary from Lin (2013:6-8). This also serves to supplement C.K.'s observation.

  1. Rhymes placed in grade III include eight chongniu rhymes (重紐八韻), ten labiodental rhymes (輕唇十韻), and normal grade III rhymes (普通三等韻).

  2. Openness is determined by the absence of the rounded prenuclear glide /u/ or /w/. The following is a table that summarises the openness (column 2) of each labiodental rhyme (column 1) reconstructed in five references (from left to right: Kalgren, Li Rong, Dong Tonghe, Baxter, Zhu Jianing). The fact that three out of ten labiodental rhymes were unanimously reconstructed as kaikou (東, 陽, and 尤), and that two were either reconstructed as kaikou or hekou (鐘 and 虞), violates the traditional rule that 'grade III hekou is the phonological environment for the labiodentalisation of /p/ to [f]':


    In Lin's words,


    If the traditional saying hekou sandeng indeed refers to openness (i.e. absence or presence of the rounded prenuclear glide /u/ or /w/), then it is bound to differ from Yunjing's classification scheme.


    • /j/ is the necessary prenuclear glide to have the rhymes classified as grade III.
    • The 東-rhyme is not hekou because /u/ is not prenuclear.
  3. Lin further noticed alternative phonological rules were proposed for labiodentalisation in MC:

    (Grade III hekou rule) /p/ → [f] / _ [j] [w] V

    (Baxter 1992:47) /p/ → [f] / _ [j] [V, +back]

    (Yang 2005:147) /p/ → [f] / _ [j] [V, +central, +back]

    which certainly are more lenient. Baxter permits not only [w] but all back vowels (ranging from [u] to [ɑ]), whereas Yang even permits central vowels. Yang's rule is more flexible (Baxter did not use much central vowels except /ɨ/), but not without fault. Lin argued with Yang's rule, words of the 蒸-rhyme would also undergo labiodentalisation (presence of mid-central vowel /ə/, see below, from Wiktionary), although 蒸-rhyme is not any one of the eight known rhymes:

    Zhengzhang Shangfang Pan Wuyun Shao Rongfen Edwin Pulleyblank Li Rong Wang Li Bernard Karlgren
    /t͡ɕɨŋ/ /t͡ɕɨŋ/ /t͡ɕieŋ/ /ciŋ/ /t͡ɕiəŋ/ /t͡ɕĭəŋ/ /t͡ɕi̯əŋ/
  4. Lin subsequently proposed the following phonological rule

    /p/ → [f] / _ [j] [+rounded],

    where [+rounded] means EITHER a rounded prenuclear glide (/u/ or /w/, or the hekou element) OR a rounded nucleus. Obviously the first part corresponds to the hekou sandeng rule. The second part accounts for any remaining inconsistencies. Sadly, there is one exception: the 陽-rhyme (neither /a/ or /ɑ/ are rounded). So to compensate for that,

                        [V, –front]
     /p/ → [f] / _ [j] {           }

    was proposed. I think this is a big improvement from the historical hekou sandeng rule.


  1. 林智凱(2013)〈中古漢語重紐、輕唇與普通三等三者交互關係〉。

  2. Baxter W. A Handbook of Old Chinese Phonology. Berlin: Mouton De Gruyter; 1992.

  3. 楊劍橋(2005)《漢語音韻學講義》,復旦大學出版社。

Appendix: IPA vowels and their backness

  • Great answer! I saw the Baxter rule, but the additional alternatives you mention are very interesting. I’m still curious what e.g. the Yunjing authors were thinking when laying out the structure and mixing “light” and “heavy” rhymes so freely (compared to later books that split them out as separate initials). That might be impossible to know though, and your thorough answer still deserves to be accepted :)
    – bellkev
    Commented May 8, 2021 at 18:01

To be short, the labiodentalization took place in ten rhyme groups called the Ten Labiodental Rhymes (輕唇十韻), which is 東(三), 鐘, 微, 虞, 廢, 文, 元, 陽, 尤, 凡. All of them are hekou (合口) except 尤. All of your examples belong to the rhyme group You (尤韻).

This is because that the labial-initial syllables of the rhyme group You(尤韻) had merged into the rhyme group Yu(虞韻) before the labiodentalization took place.

Your Answer

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

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