Sounds of the Times
Shr 35Problem. The received text of Shr 35C has the line wo3 gung1 bu4ywe4 "my self is not liked:"
The [Latter] Han stone classics were engraved during 175/183. The Shr was the first classic to be engraved, and Shr 35 comes early in the 305-poem Shr collection. That stone version of Shr 35, then, was probably engraved in 175. A surviving fragment of the stone text gives the above line in a different form: as wo3 jin1 bu4ywe4 "I am now not liked:"
Solution. Ogawa (Baxter Handbook 358f) suggests that we have here an assimilation of the final nasal of gung1 "self" to the labial initial of bu4; that is, -ng > -m in the presence of a following b-. In favor of this is that wo3 gung1 "I myself" would probably have been felt as redundant in Latter Han, leaving gung1 "myself" semantically superfluous, and thus vulnerable to errors. We suspect that Ogawa is right.
Interpretation. The stones were engraved in Lwo-yang, in c175. The equation gung1 > jin1 implies that, in that place and time, the two words differed only in their final nasal, so that the assimilation of that final nasal transformed the one into the other in the ear of the engraver who was orally imaging the text as he worked. It is well established that the two words had the same initial (g-). The present facts seem to require that they also had the same vowel. In crude and approximate terms, if we think of jin1 "now" as phonetically gyim in this period, then we must posit gying as the original sound of "myself." Or, if we prefer to posit a nuclear -u- vowel, we have the equation gyung > gyum for "self," so that "now" must have been pronounced gyum.
Whatever the vowel, the two words must have had the same vowel. Unfortunately, no major reconstruction assigns these words the same vowel. Pulleyblank's Lexicon gives respectively kim and kuwng; Baxter's Handbook gives k(r)jim and k(r)jung. The CJT rhyme table of 1067 puts both words in medial grade III, but with gung1 in Table 2 (generally corresponding to Mandarin -ung) and jin1 in Table 5 (early im / vm > Mandarin in / vn). There would thus appear to be a vowel difference in the 11c, just as there is today. The difficulty is to accommodate the vowel identity implied by the Han scribal error, while still giving ground for the vowel difference implied by the CJT placement.
It would be suggestive if some modern dialect gave the same nuclear vowel for both words. But no modern dialect available via the on-line Chinese Character Dictionary does so. Nor do the modern differences lie in the same phonetic direction. The Mandarin ung/in contrast is between front and back vowels. Wvnjou [southern Wu group] has ywong/yang, both back vowels, balanced by Chaujou [southern Min group] eng/im, both front vowels. If we assume that the Chaujou front vowels are due to the influence of medial -i- (or whatever analogous segment put both words into Grade III in the CJT), then we might want to give slightly more credence to the Wvnjou back vowels. This is reinforced by the back vowels which early scholarly loan pronunciations imply for both words. Thus Sino-Vietnamese -ung/-im seems to reflect the same stage as the CJT tables, but Sino-Japanese -yu:/-on may go back to an earlier prototype, and to one which slightly favors the back vowel hypothesis.
We may also check Shr rhyming for a clue as to the behavior of these words before Han. Confining our attention on the poems in the Fvng section of the Shr, we find that jin1 "now" rhymes in Shr 20B with san1 "three," Baxter's sum, and that gung1 "self" rhymes in Shr 36B with jung1 "middle," Baxter's k-lyung. It is also true that in the Ya section, where a wider rhyming practice has been thought to obtain, both of our words can rhyme with words to which Baxter assigns an -i- vowel.
It would seem that all this evidence can be covered by reconstructing something like gium for "today" and giung for "self," leaving it open to local conditions whether the -i- segment is actualized as consonantal -y- or as the nucleus of a glide -iu. We might then have for the rhymes in Shr 20B
sum "three" ~ gyum "now"
and for the Han transcription error in 35C, the assimilation process
gyung > gyum (before b-) "self"
Ya rhyme-sets including words in -i- may be more intelligible if it is assumed that, in that section, the syllable rules give, for "self," not gyung (nuclear u) but giung (nuclear i). That there may have been a difference of syllabic weight, without difference of segmental sequence, in the elite Ya as opposed to the more folklike Fvng, requires to be investigated, but seems not impossible on its face (compare the regional or status indications sometimes conveyed by diphthongization or triphthongization in American English).
This is our best suggestion as of this writing.
13 Jun 2002 / Contact The Project / Exit to Results Page
