The language of Chu is lost except for the few pitiful shreds of vocabulary given on our Chu Lexicon page. We here try to track and define what seems to be a phonetic feature of that language, which consists of some degree of mixing of -n and -ng finals in rhyming within a Chinese text. That aberrant feature may be due to influence from a non-Chinese language then spoken in the area of the Chu capital. There seem to be four possible instances of the feature in the present Chu Tsz anthology.
1. In Li Sau 3-4, we have a rhyme between and , which might have been approximately -vng and -vn in the Chinese of the time. Such a rhyme is impossible. The finals -ng and -n are carefully distinguished in the remainder of the Li Sau, and there is no evidence that the Chu author's Chinese pronunciation was generally affected by this merging of dental and velar nasal codas. But in this one case there does seem to be just that sort of confusion. What shall we do with this? Dai Jvn remarks, in connection with this passage, that the people of the Wu area of his day (the Yangdz delta region) still pronounced -vng words as -vn. It seems that he meant to suggest the persistence of an articulatory habit in a region which was once under the domination of Chu. We think that this explanation is essentially correct.
Shvn Zhongwei, in a paper presented at the WSWG 10 Conference in 1998, identified this articulatory trait instead with the Myau/Yau languages, which are still spoken nearer to the original area of Chu, in the middle Yangdz valley. On its face, this would imply cognate relations between old Chu and modern Myau/Yau, which would be in conflict with the Thai vocabulary affinity which has been claimed for the Chu word tiger (the Myau/Yau group are not currently thought to belong to any larger language grouping). Of the Myau/Yau group, it is the Myau (also called Hmong) languages that show nasal mixing. In the words of Ramsey 279, Myau "has only a single nasal ending, which is realized phonetically as -ng after back vowels and as -n after front vowels." It may be best, in the current state of knowledge, to attribute the word "tiger" to the original stock of the Chu language, and to attribute the phonology of nasals in the Chu language to an area phenomenon which might have affected languages which, although in contact, were of different ultimate linguistic affinity.
The question of geographical probability depends on when we date the earliest poems of the Chu Tsz, which are usually attributed to Chyw Ywaen. An often quoted date for Chyw Ywaen's literary activity is c0300, but Gopal Sukhu, in a paper given at the WSWG 13 Conference in 1999, convincingly argued that the Li Sau is a response to the Northern intrusion into Chu culture symbolized by Sywndz. Sywndz did not take office in Chu until after the conquests of 0255, and thus after the shift of the Chu capital to the lower Yangdz, which had occurred in 0278. If this reading is correct, we will have to consider the lower Yangdz as the center of Chu culture as of the composition date of at least this early piece within the Chu Tsz. Dai Jvn's observation thus appears to be relevant.
Other Suggested Cases of nasal mixing in Chu Tsz rhymes are the following (after Wang Li):
2. Tyen Wen 169-170 and
Hawkes, following Jyang You-gau, corrects to . Hawkes notes that this was a taboo substitution for the personal name of the second Latter Han emperor. This change is thus posterior to Lyou Syang's work in codifying the Chu Tsz, and the problematic rhyme vanishes as part of the Tyen Wvn poem as it existed in Han proper or in pre-Han times. It has no phonological importance. The rhyme given by the restored original is unexceptionable. Dai Jvn rejects an earlier suggestion, that Tyen Wvn here invokes the loose rhyming of Shr 305; he concludes that there is "no rhyme" in this line. But this conclusion is unacceptable in view of the pervasive pattern of rhyming in this piece. We note that all instances where Jyang You-gau found "no rhyme" in the Li Sau are capable of resolution into rhyming lines by reasonable emendations of the text. We follow Hawkes in finding that there are no non-rhyming zones in the Chu Tsz..
3. Jyou Jang (A43-44) and
The Jyou Jang poems have been attributed to Chyw Ywaen, but as Hawkes suggests, the group was probably formed from earlier single poems by the late Han editor Lyou Syang. As a single poem, Jyou Jang 1 is not attributed to Chyw Ywaen by any Han authority prior to Lyou Syang. On the evidence, the poem is therefore later than the earliest Chu Tsz pieces. Jyang You-gau groups these lines plus the preceding two (seemingly in -ang), that is, the final four lines of the poem, as rhyming together. It is also possible, and more consistent with other Chu Tsz usage, to take the two couplets as having different rhymes. Hawkes considers that the second couplet, A43-44, does not rhyme, and accepts an emendation of Wvn Yi-dwo which supplies a rhyme. We think that the received text may stand, and that it counts as an instance of (approximately) -vn/-vng rhyme mixing in Chu poetry.
4. Jyou Byen (A4-14):
The Jyou Byen group has never been attributed to Chyw Ywaen. The traditional ascription was instead to the Chu court poet Sung Yw. Thus neither this piece nor the preceding one was regarded, in early Han times, as among the earliest stratum of known Chu poetry. Literarily, both instances occur in the first poem of the group as eventually edited, and both occur at the very end of that poem. They would thus appear to have a signature function, perhaps to give prominent place to a known feature of Chu language, but giving that feature (especially in the Jyou Byen example) more structural scope than did any genuine early Chu poet. They would then be intentional, and not (like LS 3-4) seemingly inadvertent. We conclude that they are labeling gestures, designed to make the poems sound, to a non-Chu audience, like genuine authentic Chu poems.
What exactly was this "known feature of the Chu language?" If we follow Hawkes in eliminating the Tyen Wvn example as arising through the operation of a Han taboo and not as reflecting the phonology of the original work, and consider only the remaining examples, the trait in question would seem to be merging of -ng and -n after one particular vowel (we have suggested -vn/-vng), but not otherwise. (Note that lyen2 "pity," in the Jyou Byen example, is glossed by Jyang You-gau as pronounced "lin2"). Dai Jvn seems to have in mind this situation, which he describes as the mixing of the rhyme groups
but not of a more general mixing of all -n and -ng finals.
We might then have this situation: (1) The old Chu language was non-Sinitic; it had lexical affinities with what are now Thai and Lao. (2) Chinese had begun to replace Chu language at the Chu court in the 03rd century. (3). When writing Chinese, the Chu poets carefully observed the -n/-ng distinctions of Chinese, with only rare lapses into a phonologically limited -n/-ng mixing to which their own language was subject. (4) A similarly limited -n/-ng mixing was reported in the lower Yangdz area by Dai Jvn; this agrees with the date and location of the probable composition of the Li Sau in particular. This trait is probably not a survival of Chu language as such (which was native to a region further west). It is more likely the survival of an area phonetic feature to which the Chu language in the middle 03c had become subject. (5) A more extended mixing of nasals can be observed in the Myau language, as it is still spoken near the old Chu homeland. This may be the same wide-area phonological trait further developed, though an objection to this is that the related Yau languages do not display the trait.
Of the early and inadvertent case LS 3-4, it is noteworthy that this is the couplet in which the poem's authorial persona is stating his name. We recall that the elements of a Chu minister's name made up a large fraction of the word material collected in our Lexicon. As a suggestive parallel, a seeming trait of the Chvn substrate language is also exemplified (in Shr 137) by a surname, not a word in ordinary usage. It is well known that proper names are frequently conservative within a language; they may hold back from otherwise systematic sound changes and preserve older pronunciations. It would appear that the same conservatism may apply to proper names in substrate languages, which have an exceptional resistance to the pressures of the overlay language and its phonology.
2 Aug 2002 / Contact The Project / Exit to Results Page