Reflexes of Archaic Chinese *g- in the Min Dialects.

The Ancient Chinese initial ɤ -... occurs only before the finals of divisions I, II, and IV, while the initials g-...and j-... both only occur before division III finals. Theoretically speaking, the initials ɤ - and g- or j- are in complementary distribution. It is equally reasonable to reconstruct ɤ - with either g- or j- as one phoneme in Archaic Chinese. In 1971, Dr. Li Fang Kuei discovered that Anc. j- mostly occurred before ho-k'ou finals while g- occurred before k'ai-k'ou finals. He relegated Ancientɤ -, g- and j- to one Archaic phoneme *g-, and set up the following rules of change: Arch. *g + -j- finals > Anc. gj-... Arch. *g + non -j- finals > Anc. ɤ -... Arch. *gw + -j- finals > Anc. jw-... Arch. *gw + j + i > Anc. gjw-... Arch. *gw + non -j- finals > Anc. ɤ w-... For reasons stated elsewhere, I (Ting 1977-78) revised the rules as follows: *g +j > gj-... (...) *gw +j > gju-... (...) *ɤ +j> j-... (... ) *ɤw +j > ju-... (...) *ɤ + non-j finals > ɤ -...- (...) *ɤ w + non-j finals > ɤ u- ... (... ) One problem with this scheme would be why the common initials g- and gw- only occur before -j- finals. In the present paper I will present evidence from the Min dialects to support this proposal. There is a group of Ancient ɤ -words pronounced with initial k- which co-occur with tones in the Yang category in Colloquial Min. (Cf. p.3). This group differs from the other initial k- words which have tones in Yin category. It is natural to reconstruct voiced initials *g and *gw for this group in Colloquial Min as well as in Archaic Chinese. This reconstruction would make Archaic Chinese *g and *gw become well distributed phonemes. Therefore, the following two rules may be added to those I proposed before: *g + non-j finals >ɤ - (...) *gw + non-j finals >ɤ u- (...) I have checked the Hsieh-sheng relations of these *g- and *gw- words, and discovered nothing special. (Cf. p.4). [ABSTRACT FROM AUTHOR]