When does Middle Chinese kh become Cantonese h?

[Hwong_DsiKiem]

I've noticed that, for a number of characters, what was historically kh- (pinyin k-) became h- in Cantonese (like 去), but not all kh- characters.

In what environments does the shift happen? In what doesn't it?

This shift is similar to the one in Greek and Germanic languages, but there all the kh's become h's.

[Hofmann]

I don't know when. If followed by an unrounded vowel, /h/. If rounded, /f/.

[Hwong_DsiKiem]

I was asking when does kh become h.

hw assimilates to f not just in Cantonese, but that isn't the question.

[Michaelyus]

By /kh/, I assume you mean the Middle Chinese 溪 initial, right? Though people usually attribute it to 合口 actually it's way more complicated than that. Specifically 戈一合 [and 桓一合], 灰一合, and 模一开 categories seem to lead to /f/ most readily.  

[Hwong_DsiKiem]

I have been asking about when kh- becomes h- (曉) but it seems that all commentors here have been talking about kh(w)- or h(w)- becoming f-.

But yes by kh I meant the TCC kh, which is the theoretical equivalence of the Middle Chinese 溪 initial.

I don't use the // for it because // is for the phonemic values in the IPA, and it isn't that.

[Hofmann]

You want to read your other questions and my replies more carefully?

[Michaelyus]

Sorry: have been quite seriously archiphoneme-ing!

 

Just to clarify then, just as 合口 is said to lead to fricative f from 溪 kh, 開口 is meant to lead to fricative h from the same initial. That's what's usually assumed.

 

However, there is a much larger number of syllables that have undergone this change. The majority of them are 開口, mostly 一等, one 二等, quite a number of 三等, two 四等; there is also the 合口三等 final 元 and 合口四等 final 先 (though curiously not in 入聲 for either, where they both remain as kh- / Jyutping k-).

 

Conversely, of the 開口一等 finals, only a very small subset have retained kh-, being 咍 and 侯 finals. All the rest have progressed down to h-, f- or null.

Of the 開口二等 finals, only 江、山、刪 in 入聲 have retained kh-. Of 開口四等 finals, 齊 retains kh-. (Some of this analysis might be somewhat incomplete, as many of the rare characters I've yet to find Cantonese pronunciations for; plus 多音字 I've ignored).

[Hwong_DsiKiem]

I was really referring to what causes 去, 謙, 口, 可, etc to get an /h/ initial but not say, 卡, 跨, etc for what used to be kh- in Middle Chinese in Cantonese. Sorry I wasn't being clear. The f isn't what I'm concerned with.

 

EDIT: kh- to h in 入聲: 客, doesn't that contradict with what you said?

 

So for the specific finals, they retain kh, while the others they became h (hw further becoming f)?

[Michaelyus]

客 is 二等.

[Hwong_DsiKiem]

You said "(though curiously not in 入聲 for either, where they both remain as kh- / Jyutping k-)", but 客 has an entering tone

[Michaelyus]

The initial kh does not become h- in 入聲 for either 三等 nor for 四等 finals. But 客 has an 二等 final.

[Hwong_DsiKiem]

I'll try to take my time to comprehend all that. Thanks.

[Hwong_DsiKiem]

I'm investigating on this matter by listing all the finals, and on "Conversely, of the 開口一等 finals, only a very small subset have retained kh-, being 咍 and 侯 finals. All the rest have progressed down to h-, f- or null", I found 口 khêu/ > hau2, but 扣 khêu~ > kau3

http://tcchinese.tumblr.com/post/86774464660/i-tried-to-find-a-pattern-for-the-kh-h-includes

[Michaelyus]

Well then, I suppose 口 was one of those that was affected by kh- > h-, despite being a 侯 final. 㸸 was also affected (that's also 上声); might that suggest a literary/colloquial split? According to an article published in 2013 in 语言科学, entitled 〈广州话溪母字的历史层次及音变〉  says that there is some indication of dialect mixing. The more colloquial and more archaic layers of the language have turned to h-, whereas more recent and literary layers brought down from the 中原 probably by the Hakka migrations in the southern Song. The same paper goes onto suggest that dialect mixing was endemic to the whole area, with Hakka, Pinghua, Gan as well as Cantonese being affected and splitting out the 溪 initial into many more phonemes.

[Hwong_DsiKiem]

So you're suggesting Hakka keeps the kh- initial more than Cantonese does?
 

I noticed that kh has a tendency to become h-, and hi- clusters are sometimes simplified to i, like 綺 and 欣 (han does not exist after all). This might have been a chain shift instead of the kh splitting into many phonemes at once, similar to 快 khwai~ > hwai > fai.