Проблемы китайского и общего языкознания. К 90-летию С. Е. Яхонтова

 454    Christopher I. Beckwith verses, which have no poetic elements remotely like them 1 . The songs have still other poetic devices, including alliteration and assonance, which need to be analyzed further. It must be emphasized again that the native Pai-lang character of the songs shines through despite the strong Chinese linguistic and cultural influence on them. The appendices give the identified Pai-lang rhymes for all three songs. Note that these are Pai-lang rhymes, not Chinese rhymes. Since it is unlikely that T’ien Kung had such a poor knowledge of Pai-lang that he could not rec- ognize and transcribe the most obvious of the Pai-lang rhymes, it is necessary to conclude that his dialect of Later Han Dynasty Chinese was quite different from the Central dialect. For example, several rhymes make it is fairly clear that T’ien’s dialect did not have a phonemic distinction between final *- m and final *- w , both of which were evidently articulated as a bilabial glide [ w ]. Cor- respondingly, the apparent absence of labial stop codas in the Pai-lang texts suggests that the language did not have any 2 . The same also applies to coronal stop codas in the end-rhymes; however, because there appear to be examples line-internally, especially in the third song, in what appear to be Chinese loan- words, the existence or non-existence of coronal stop codas would need to be established through a more thorough study. Their probable nonexistence in Pai-lang itself would seem to be supported by the fact that the apparent (‘tran- scriptional’) alveolar nasals are not distinct from the transcriptional velar na- sals, indicating that phonemically there was only one nasal coda, probably more velar than alveolar phonetically (cf. Japanese) 3 . By constrast, in view of the consistent transcription and good comparative evidence, there clearly was a liquid coda, *- r , in Pai-lang. The majority of codas, and the only clear, consistent stops in the transcription, are velars; yet even in these cases the final may not have been articulated as a velar stop but rather as a glottal stop, because the vast majority of syllables are open. This distribution is of course typologically normal for the southeastern parts of Eurasia. The establishment of some characteristics of the initials of Old Chinese since Coblin’s article was published allow corrections to be made to his 1 If they were indeed composed in such a way, the Pai-lang poet must have been an even greater artist than I have imagined in order to be able to make the supposed ‘translations’ into art. But the idea that so much energy would have been expended on songs in a language unknown at the court where they were sung, rather than on the Chinese — hich would also have been sung or chanted — does not make much sense. In my opinion the songs must have been composed originally, or primarily, in Pai-lang. 2 On the interchangeability (in the reconstruction) of these two phones in word-initial position in earlier forms of Old Chinese see [Beckwith 2002, 2006]. 3 The nasal phone disappears (probably nasalizing the vowel) after *- Vw .