A Reconsideration of the Initials of Karlgren's Archaic Chinese

Abstract

In this further study of the forms of Chinese words in the period of the development of the written characters I have assumed the correctness of the conclusions reached in the earlier article under the same title 1) and later given in somewhat greater detail in my work on the history of the language.2) Differences from Karlgren's representations of the final parts of Chinese words are explained in my article in the Bulletin de la Societe de Linguistique.3) The phonetic symbols here used are the same as in the earlier T'oung Pao article. As all the forms quoted are conjectural reconstructions, it is superfluous to prefix an asterisk, which is already omitted in the earlier paper. If we leave aside for the moment cases in which Karlgren has reconstructed a medial -1as well as others in which -1-, or, as will appear later, an r sound, may reasonably be assumed, the initial compounds with which this article deals are those involving nasals. An examination of the phonetic series in Grammata Serica (GS) containing nasals in conjunction with oral consonants shows that the latter do not occur at random, but each consonant only, or predominantly, in composition with a particular nasal. So we have xm-, but never thmor sm-, and only one solitary, and therefore questionable instance of khm-, in GS 950; th-/tappears with nbut never with -yor -m-. We never find xyin Karlgren's series, but xappears freely in yseries, and there is the same justification for writing xyas for writing thnwhere this attested in nseries, e.g., in GS 622 4, where nand thnare alternative readings of one character. The fact that dn-, gy-, kynever occur is also one having bearing on the evolution of the nasal initials, when this is seen against the freedom with which gl-, dl-, etc., are restored.