152
LAURENT SAGART
In the above chart, an OC final is reconstructed as the corresponding SC final if it is homophonous with the corresponding KHW final, or one of the corresponding KHW finals. In case SC and KHW disagree totally on the pronunciation of a given class of words, no reconstruction is attempted although a separate final must have existed for this class of words in OC.
The main characteristic of the KHW system of final nasal and stop consonants is the merger of the -n/t finals into the ng/k finals. All SC words ending in -n or -t correspond to KHW words ending in -ng or -k. In general, an -n/t final merged into the -ng/k final of the same vowel, resulting in widespread homophony:
*-n/t finals
*ng/k finals
kaengl 'interval' is homophonous with: kaeng 'the 7th celestial stem'
sangl 'new' is homophonous with: #sangl 'sound'
paek3 'eight' is homophonous with: paek3 'hundred'
sak3 'lose' is homophonous with: sak3 'know'
fung3 'style' is homophonous with: fung 'wind'
However, in certain cases, there did not exist a -ng/k final of the same vowel. This led to the creation of new -ng/k finals, which resulted in overcrowding and phonetic realignment: thus the */-in, -it/ finals were changed to /-ing, -ik/, but instead of merging with the original finals /-ing, -ik/ which possibly had a slightly lower /i/, as is the case in SC, pushed them away towards /-ang, -ak/ (lax) with which they eventually merged:
SC: -ing/k; KHW: -ang/k:
kangla 'classic'; sangl 'star'; p'ang2 'level'; sak3 'know'; nak3 'history'
The *-uen/t finals were changed to -üng/k, an innovation in the system of finals which could not result in homophony.
The -on/t finals of SC correspond to KHW -ung/k, as already mentioned. However, the raising of /o/ to /u/ in */-oi, -on, -ot/ is unrelated to the movements of final stops and nasals. It is possible that */-on, -ot/ first merged with */-un, -ut/ before the merger of final dentals and velars took place. A similar situation...