漢藏比較語言學中的幾個問題

Problems in Sino-Tibetan Comparative Philology

本文檢討漢藏語言的比較研究與漢語上古音構擬的互動關係。理論上，漢藏語的比較研究應該是在漢語上古音構擬完成之後才能進行，必須要到漢語上古音構擬完成之後，才有可能根據構擬的上古漢語與藏、緬語作比較，構擬原始漢藏語。然而由於漢語上古音構擬當中有一些特殊的困難，必須藉助於漢藏語言的比較來提供解決的線索，這才造成這兩個不同的研究層面之間糾結不清的結果。本文討論其中所牽涉的幾個問題，釐清構擬漢語上古音與漢藏語比較研究的關係。

In 1928, the founder of Chinese historical linguistics, Swedish sinologist Bernhard Karlgren, published an article entitled ''Problems in Archaic Chinese,'' in which he enumerated four different means of reconstructing the sound system of Old Chinese. In his article, Karlgren referred to the comparative study of the Sino-Tibetan languages as the most important of the four methods, but then dismissed it at once on the grounds that the field was not mature enough for such researches, leaving it undeveloped in the article. The other three methods delineated by Karlgren were the study of lacunae (holes in patterns) in the phonetic system of Middle Chinese, the study of phonetic compounds in Chinese characters and the study of the rimes of the Shijing. Eighty years have elapsed since the publication of ''Problems in Archaic Chinese.'' Notwithstanding advances in the field made during this period, the primary approaches to the study of Old Chinese phonology continue to be the third and fourth methods mentioned by Karlgren. As for the second method, the focus has shifted from the study of lacunae in the Middle Chinese phonetic system to systematic comparisons between the sound systems of Middle Chinese and Old Chinese. As a result, the reconstruction of the Old Chinese final system has been based on sound change rules inferred from the correspondences between Middle Chinese and Old Chinese rime categories. This paper deals with the first of Karlgren's methods, namely, the comparative study of the Sino-Tibetan languages, and places emphasis on the relationship between this comparative study and reconstruction of Old Chinese. Theoretically speaking, Old Chinese phonetic forms must be reconstructed before they can be used in the comparative study of the Sino-Tibetan languages. However, it has become clear that the correct reconstruction of Old Chinese relies largely upon clues provided by the Sino-Tibetan language comparisons. The reason for this is that the sound relations of compound characters in Old Chinese allow for many different interpretations. There are many ways of reconstructing consonant clusters in order to explain the sound relationships among them. The best solution is to choose the sound values that can accommodate both the phonetic compounds (internal sources) and the comparative evidence from Sino-Tibetan language study (external sources). This paper attempts to justify the reference to comparative evidence in the reconstruction of Old Chinese and discusses the problems involved in this process as well.