I kind of think the retrolex just as variants of regular s, z, ts, tsh & dz while preceding a /ɣ/, so it doesn't usually bother me that the retroflex never precede vowels directly.
@Phjong But they also precede /j/. It seems strange that a sound would become retroflex before a palatal glide. In fact, you'd normally expect the opposite. (Consider the case of modern Mandarin, where retroflex /ʂ/ or /tʂ/ become /ɕ/ or /tɕ/ before /i/ and /y/)
Mind you, I am not trying to criticize you. I'm just genuinely curious about this and very rarely get to discuss it with someone who knows the material.
In 鄭張's reconstruction, retroflex don't prcede /j/ directly, but can precede /ɣj/. like, '所' is reconstructed /sɣjə/. I ｏｆｔｅｎ ｆｏｒｇｅｔ ｔｏ ｐｒｏｎｏｕｎｃｅ ｔｈｉｓ ɣ．
@Phjong /ɣj/ as a medial is hard for me to believe because it requires two places of articulation which are very dissimilar. Moreover, it means that, since you *never* have pure /ɣj~j/ or /ɣ~j/ minimal pair after retroflexes, you end up with a redundant /ɣ/ in the /ɣj/ combination carrying no functional load whatsoever. Wouldn't it be easier to have a back semivowel like /ɰ/?
indeed /j/ never appears directly after retroflexes. however it's not necessarily something weird: e.g. in some American accents, /k/ are always articulated [qh] when preceding /l/, but normal [kh] otherwise. Could one then argue that, there's no /l/ - ∅minimal pair after [qh]?
@Phjong Sure. In fact, that's true of *my* accent coincidentally. However minor allophony like that isn't something you'd expect native speakers of American English to be aware of without training. Likewise, I find it hard to believe that such an allophonic distinction would have been audible enough, and seemed important enough, to Chinese speakers in the 6th and 7th centuries for them to include it as a distinguishing feature in their 反切 spellings.
The tones are very nicely done, but I'm not convinced of the -ɣ- in div II (per Zhengzhang's work.)
It is problematic because it would mean that retroflex affricates were always followed by a medial -j- or -ɣ- and were never directly followed by a vowel (which is typologically strange,) and that there was a triple opposition –w- (hekou 1), -ɣ- (kaikou 2) and -wɣ- (hekou 2). I find that hard to believe.
Apologies for writing in English. I'm not good at discussing linguistics in Chinese.
And the triple opposition -w- -ɣ- -wɣ- is indeed quite difficult in practice. But when I asked Prof. 潘悟雲 whether the -ɣ- could actually be uvular, he said that there's rarely any language that distinguishes velar and uvular fricatives, and by /ɣ/ he is not trying to exclude the possibility of /ʁ/. In this way, the triple opposition is -w- -ʁ- -wʁ-, which looks okay to me.
Thank you :)
I kind of think the retrolex just as variants of regular s, z, ts, tsh & dz while preceding a /ɣ/, so it doesn't usually bother me that the retroflex never precede vowels directly.
Phjong 1 year ago
@Phjong But they also precede /j/. It seems strange that a sound would become retroflex before a palatal glide. In fact, you'd normally expect the opposite. (Consider the case of modern Mandarin, where retroflex /ʂ/ or /tʂ/ become /ɕ/ or /tɕ/ before /i/ and /y/)
Mind you, I am not trying to criticize you. I'm just genuinely curious about this and very rarely get to discuss it with someone who knows the material.
szfski 1 year ago
@szfski
Don't worry, it's certainly fun discussion.
In 鄭張's reconstruction, retroflex don't prcede /j/ directly, but can precede /ɣj/. like, '所' is reconstructed /sɣjə/. I ｏｆｔｅｎ ｆｏｒｇｅｔ ｔｏ ｐｒｏｎｏｕｎｃｅ ｔｈｉｓ ɣ．
Phjong 1 year ago
@Phjong /ɣj/ as a medial is hard for me to believe because it requires two places of articulation which are very dissimilar. Moreover, it means that, since you *never* have pure /ɣj~j/ or /ɣ~j/ minimal pair after retroflexes, you end up with a redundant /ɣ/ in the /ɣj/ combination carrying no functional load whatsoever. Wouldn't it be easier to have a back semivowel like /ɰ/?
szfski 1 year ago
@szfski
indeed /j/ never appears directly after retroflexes. however it's not necessarily something weird: e.g. in some American accents, /k/ are always articulated [qh] when preceding /l/, but normal [kh] otherwise. Could one then argue that, there's no /l/ - ∅minimal pair after [qh]?
Phjong 1 year ago
@Phjong Sure. In fact, that's true of *my* accent coincidentally. However minor allophony like that isn't something you'd expect native speakers of American English to be aware of without training. Likewise, I find it hard to believe that such an allophonic distinction would have been audible enough, and seemed important enough, to Chinese speakers in the 6th and 7th centuries for them to include it as a distinguishing feature in their 反切 spellings.
szfski 1 year ago
The tones are very nicely done, but I'm not convinced of the -ɣ- in div II (per Zhengzhang's work.)
It is problematic because it would mean that retroflex affricates were always followed by a medial -j- or -ɣ- and were never directly followed by a vowel (which is typologically strange,) and that there was a triple opposition –w- (hekou 1), -ɣ- (kaikou 2) and -wɣ- (hekou 2). I find that hard to believe.
Apologies for writing in English. I'm not good at discussing linguistics in Chinese.
szfski 1 year ago
@szfski
And the triple opposition -w- -ɣ- -wɣ- is indeed quite difficult in practice. But when I asked Prof. 潘悟雲 whether the -ɣ- could actually be uvular, he said that there's rarely any language that distinguishes velar and uvular fricatives, and by /ɣ/ he is not trying to exclude the possibility of /ʁ/. In this way, the triple opposition is -w- -ʁ- -wʁ-, which looks okay to me.
Phjong 1 year ago