| Despite the significant advancement of research methods and the accumulation of knowledge, some significant issues in the study of Old Chinese persist.;First, this dissertation is a review and revision for methodology. After evaluating research from the Qing scholars to current OC reconstructionists, this dissertation proposes a "complete phonological reconstruction" for OC is enriched with information from different levels of representation in the sound system. Besides the Neogrammarian exceptionless change, other types of sound changes, such as lexical diffusion, are taken into this new reconstruction method. In this new model, borrowing and loanwords data should be analyzed using the theories for modern living languages, such as the theory of different "repair strategies" and the substratum language contact theory. The new model propose to abandon the "similarity condition" and change to the "identicalness condition" which requires that for two syllables to rhyme they must have an identical vowel and coda in the output of the lexical rules. Statistic methods are also reviewed both on their technical and theoretical aspects.;Second, this dissertation applies the new methodology to one of the most resistant issues in Old Chinese (OC): the "stop coda" problem, or yinru duizhuan. Two existing hypotheses fail to explain the phonologically conditioned yinru tongya distribution identified in my study, especially in the following regards: (1) yinru tongya occurs more frequently among syllables that have velar codas (/-k/ and /-g/) than coronal codas (/-t/ and /-d/), and almost never between syllables that have labial codas (/-p/ and /-b/); (2) velar coda syllables with lower vowels have more yinru duizhuan cases than those with high vowels, and among syllables with coronal codas, those with high vowels have more such cases than those with lower vowels. The phonologically conditioned yinru duizhuan distribution in Shijing leads me to propose a lexical diffusion analysis for the yinru duizhuan problem: a final voiced stop coda devoicing rule and a final voiceless stop coda deletion rule occurred in OC, and the diffusional sound change type of the latter created the phonologically conditioned gradient pattern of yinru duizhuan. |
