Title

On Conditioned Change

Author

Kuang-yu Chang

Genre

Article  

Pages

427-475

Download

Language

Chinese

Key words

Chinese dialects, Conditional change

Abstract

     As the development of the western discipline of historical linguistics has shown us, conditioned change and the comparative method are two sides of the same coin. Following grimm’s law of 1822, there came a sequence of laws: Lottner’s law, the law of the palatals, Grassmann’s Law, as well as Verner’s law, each marking a milestoneleading the comparative method to its maturity in t he later part of the nineteenth century. In rerrsopect, one may be curious about wh at happened in its eastern counterpart? It is safe to say that the majority of what has been said about conditioned change in Chinese historical phonology was by its very nature circumstantial, a hub of disputes. There are several reasons for this. For Karlgren, the great Swedish Sinologist who laid down the most important cornerstones for the reconstruction of Chinese historical phonology, the first objective was to tr anscribe Chinese terminology into a sound system. Unlike his western colleagues in the field of Indo-European languages, who were free of traditional burden, Karlgren did not base his reconstruction purely on the comparative method. In many cases his method was a philogical one, and as such, there was not much to say concerning conditioned change. This situation has remained basically unchanged even in the post-Karlgren period. Due to the relative paucity of Chinese dialect material before 1980, little evidence could be referred to in discussing conditioned change on a solid foundation. This is the second reason why discussion on conditioned changed has remained largely theoretical rather than scientific.

    This study is based on cross-dialect comparison. Most cases in this paper are discussed in a quite straightforward manner; we need only to compare two closely related dialects to ascertain which form serves as the conditioning factor and which form has undergone the conditioned change. But there are also cases where we are required to reinterpret the previously reconstructed forms or to do triangulation on our own before we proceed.

     Reinterpretation is needed in cases where dialects show no sign of distinction for the different categories in ancient texts. An example of this is to be found in *âng and *ång, which have merged as a single category in most modern Chinese dialects, north and south. In order to facilitate the procedure, I have treated them as simply one final. Likewise, Karlgren’s which occurs mainly in third division finals (expecting one occurrence in second division final) can be viewed as a variant to his â (with a back vowel quality of some sort, that is, o~ɔ).

    Triangulation is put into practice when cases arise indicating that a bipartite development is involved. One example of this is found in the Middle Chinese Yinrhyme (*jən), from which derived the various forms in the over-whelming majority of Chinese dialects. In striking contrast are the Hakka and Min dialects which appear to be quite different.

    The past two decades have seen an ever-increasing amount of Chinese dialect material come to light, and yet more can be expected to come. Based on hard facts gathered across China, one can pinpoint exactly what rules put forth in the past have a solid footing as opposed to those which are merely circumstantial. Instead of recapitulating what has been said in the paper, I have deliberately chosen to concentrate on my working principles in this summary. This is because I am of the opinion that in the study of conditioned change, we are often faced with the problem of setting a starting point, and without the guidelines sketched above, the great treasure of Chinese dialect material would in many respects remain inaccessible.