How Middle Chinese /h/ was adopted into Japanese, which lacked it

Discussion in 'Etymology, History of languages, and Linguistics (EHL)' started by Mugi, Jul 2, 2007.

  1. Mugi Senior Member

    NZ English
    MC pronunciation of 胡 was probably /ɣuo/, beginning with a voiced velar fricative, which is much closer to modern day Japanese /k/ than what /h/ is. 
  2. Anatoli Senior Member

    Melbourne, Australia
    Native: русский (Russian), home country: English
    Mugi, are you sure about the symbol? You are using the "voiced velar fricative" "ɣ" but I think it's the "voiceless velar fricative" "x" (German 'ch', Spanish 'j', Russian 'x'). English speakers also have often problem with this sound, Russian 'x' is usually pronounced as 'k' by English speakers.

    The first ("ɣ" ) appears in languages such as Arabic and would be quite unpronounceable for a Chinese person, IMHO but the second is quite common and is also used in modern Mandarin. Modern Mandarin 胡 is 'xu', not 'hu' (x is an IPA symbol, not pinyin!).
  3. Mugi Senior Member

    NZ English
    Having not lived in China 10-15 centuries ago, and not being an expert on Middle Chinese, no I'm not completely sure that it was voiced and not voiceless, but the references I have checked so far tend to give it as a voiced velar fricative. In fact none give it as voiceless, although some do give it as /ɦ/.

    I would be very surprised if after 1500 years any sound managed to remain unchanged from MC into any sinitic language. So I don't know that pointing out that a modern day Chinese would find it difficult to pronounce ɣ is at all relevant. You're quite correct to point out that 胡 is /xu/ in MSC and not /hu/ though - my silly oversight!

    For a quick reference on MC pronunciation of specific characters, Google "中古音查询"
  4. Anatoli Senior Member

    Melbourne, Australia
    Native: русский (Russian), home country: English
    Don't get me wrong, Mugi, I am not picking on you. Just surprised. Have you got any English language reference about MC pronunciation? Anyway, this voiced sound can hardly be confused for a 'k' sound by any speaker, IMHO, have you heard this sound?
  5. Mugi Senior Member

    NZ English
    I have two or three English lang references on MC pronunciation. The rest are in Chinese.
    I agree that /ɣ/ doesn't sound like /k/, but to my ears it sounds closer than /x/ does. Unfortunately I know even less about historical Japanese phonology than I do about Chinese, but what is pronounced /k/ in modern Japanese probably wouldn't have been /k/ 1,000 years ago. I was merely pointing out that it's not like Japanese speakers heard /xu/ and the closest they could approximate in Japanese was こ. The Chinese pronunciation at the time and their rendition of it in Japanese at the time were probably much closer than they appear now after centuries of phonological divergence at both ends.

    Ps - Just checked Kalgren and Pulleyblank, and they both give ɣ as well.
  6. Flaminius

    Flaminius coclea mod

    capita Iaponiae
    日本語 / japāniski / יפנית
    Hello Mugi and Anatoli,

    I used this dictionary to check the MC pronunciation of 胡 (sorted by MSC pronunciation in pinyin). If other dictionaries cite /ɣ/ as the MC pronunciation of 胡, my source was not as accurate as I thought.

    In any event, that older Japanese did not have /h/ is pretty well established. I thought I could account for the phonetic change that occurred in Chinese words with [h] or similar phonetics when there were admitted into Japanese.

    This is becoming a full discussion. I will create a new thread later.
  7. Mugi Senior Member

    NZ English
    I often use Baxter too, although I didn't in this case. His reconstructions are usually cited as being more accurate than either Kalgren or Pulleyblank. It's strange that he would differ here from practically everyone else.
  8. Mugi Senior Member

    NZ English
    Anatoli, you may be interested in the following.

    After confirming that 胡 begins with /h/ in both Korean and Vietnamese, I checked my copy the 汉字方音字汇 to see how the character was pronounced in other Chinese dialects besides MSC and I discovered it is pronounced /ɣəu/ in 双峰 (湘语). Most places use an /x/, /f/, /w/, /h/ or /Ø/ initial, although 温州 uses /v/ and 苏州 predictably /ɦ/.

    I looked for other dialect places with the /ɣ/ initial to find that it also occurs in 太原 (classified as 官话 in the 方音字汇, but sometimes as 晋语 in other sources).
  9. Flaminius

    Flaminius coclea mod

    capita Iaponiae
    日本語 / japāniski / יפנית
    In search of other MC morphemes that had initial /h/, I looked up the following characters in Baxter;
    汗 鼾 杭 行 豪 核 河 荷 褐 恒.

    They are all adopted into Japanese with initial velar stop (豪 having voiced stop /g/ in Jp.). Is anyone interested in confirming the MC pronunciation of the 10 characters in other resources to cast more light on my Ch. /h/ —> Jp. /k/ thesis?
  10. Mugi Senior Member

    NZ English
    - 古寒切:kAn胡安切:ÄAn侯旰切:ÄAn
    - 許干切:xAn侯旰切:ÄAn
    - 胡郎切:ÄAN
    - 下更切:ÄŒN / ÄaN下浪切:ÄAN胡郎切:ÄAN;户庚切:ÄŒN / ÄaN
    - 胡刀切:ÄAu
    - 下革切:ÄQk户骨切:Äu«t
    - 胡歌切:ÄA
    - 許箇切:xA胡可切:ÄA胡歌切:ÄA
    - 胡葛切:ÄAt
    - 胡登切:Ä«N

    As you can see from the above, there is no occurrence of /h/, although Pulleyblank gives it for and (許干切). He also sometimes gives /ú/ where others give /Ä/.
  11. Flaminius

    Flaminius coclea mod

    capita Iaponiae
    日本語 / japāniski / יפנית
    I thought I selected diverse enough samples to have at least two or three reconstructed forms with initial /h/, only to be completely betrayed. :D

    Is it possible to conclude from the evidence (I admit it's somewhat scarce) that MC did not allow /h/ at initial positions? :confused:
  12. Mugi Senior Member

    NZ English
    It seems that way, although in addition to Baxter, Pulleyblank and 潘悟云 do offer /h/, especially where others reconstruct /x/.

    However, I doubt that MC /h/, if it existed, would have then developed into a more difficult sound (/x/, /ɣ/, /ɦ/, etc) as it is now represented in many of the sinitic languages. This is the very reason I posted initially; /h/ -> /k/ seemed a highly unlikely development, even given the different phonological constraints MC and Japanese at the time would have been under. I would be interested to discover exactly on what grounds the likes of Baxter, Pulleyblank and 潘 have reconstructed /h/.
  13. Qcumber Senior Member

    UK English
    I am not a sinologist, but I have :

    SCHUESSLER, Axel (2007)
    ABC etymological dictionary of Old Chinese
    Honolulu: University of Hawai’i Press

    First point
    Each entry is the modern item in pinyin representing the Modern Chinese pronunciation (ModC). It is followed by the Chinese logogram and the reconstructed Middle Chinese (MidC) pronunciation in IPA phonetics.

    1) In the pinyin H [x] chapter, one can see that some ModC [x-] items come from MidC [å-] items, while other ModC[x-] items come from MidC [x-] items.
    For instance “bone” has MidC [å] > ModC [x] whereas “sea” has MidC [x] > ModC [x]. (both p. 270)

    2) There are two characters, “dewlap” and “steppe nomads”, both read hú [xu/] in pinyin and both reconstructed as MidC [åuo]. (p. 281)

    3) Chinese hàn [xan\] is read kan in Japanese because it was read [xânC] in MidC.

    Second point

    Now let us examine what happened with Arabic loanwords in Malay. Why Malay? Simply because it has about the same collection of consonants as Japanese.
    Arabic [å] became [g] in Malay, e.g. غسول Arab. [åusu:l] > Mal. gusúl [gu'sul] “ritual cleaning before prayer”.
    Arabic [x] became [k] in Malay, e.g. ختان Arab. [xita:n] > Mal. kitán [ki'tan] “circumcision”.

    So, we may infer [å] in a given language becomes [g] in the other languages that do not have this phone.

    If we apply this observation to Japanese borrowings from Chinese, we can see why MidC [åuo] became go- in Japanese, e.g. in goma 胡麻sesame”.

  14. Flaminius

    Flaminius coclea mod

    capita Iaponiae
    日本語 / japāniski / יפנית
    I could avail myself of a chart of initial consonants of Ancient Chinese (Sui-Tang period) reconstructed by Karlgren (角川新字源 1968: 1185 Table A). According to the chart, four initials traditionally classified as gutturals (喉音) in AC were; 暁 x, 匣 y, 影 ·_, 喩 ʾ_. Nowhere in the chart was found /h/. Is "x" of 暁母 equivalent to Ä in Mugi's 反切 list, to Qcumber's å and to IPA /ɣ/ (voiced velar fricative; commonly known as Ghayn)? As different notations are flying around in the thread, I feel the necessity of striking a commensurable definition.

    Reasonable explanation but 胡 can also be pronounced with a voiceless initial; kojin 胡人 (western nomad). I could extract four characters from the 反切 list above that have the same initial with 胡 and for which no different pronunciation is recorded; 杭 褐 豪 恒. The first two are always pronounced with /k/ in Japanese, whereas the latter two have two initials, /k/ and /g/.
  15. Qcumber Senior Member

    UK English
    I agree with you, Flaminius, that there is more than one Japanese version of Abarbarian@.
    Here are some additional remarks of mine, a non-specialist.
    I suspect it refers to the Turkic tribes of the steppes.
    It is now pronounced [xu/], but was pronounced [γuo] in Middle Chinese (Schoessler 2007:281). In Japanese, it has three main readings: u, go and ko. (Nelson #3753)

    1) I don=t see where the reading u comes from.
    e.g. uron 胡乱Asuspicious [< disturbance caused by a Barbarian ?]@

    2) To me, the reading go reflects MidC [γuo].
    e.g. goma 胡麻Asesame [< barbarian hemp]@

    Its variant is gu.
    e.g. agura 胡座Asitting cross-legged [i.e. like a barbarian]@
    Note. The logograms represent the meaning. They do not correspond to the three components of the term: a-gu-ra.

    3) To me, again, the reading ko does not reflect the MidC pronunciation, but a later one, when the unvoiced velar fricative [x] and the voiced velar fricative [γ] merged into the unvoiced velar fricative [x].
    e.g. koba 胡馬Anorthern barbarian=s horse@
    koshÇ 胡椒Apepper [< barbarian=s mountain ash]@

    Its first variant is ku.
    e.g. kurumi 胡桃Awalnut [< barbarian=s peach tree]@
    Note. The logograms represent the meaning. They do not correspond to the three components of the term: ku-ru-mi.

    Its second variant is ki.
    e.g. wakiga 腋臭Aarmpit stench [< barbarian=s stench]@
    Note. The logograms represent the meaning. They do not correspond to the three components of the term: wa-ki-ga.

    kyãri 胡瓜Acucumber [< barbarian=s melon]@
    Note. The logograms represent the meaning. They do not correspond to the three components of the term: ki-ã-ri.

Share This Page