摘要
The Sino-Tibetan hypothesis postulating a genetic relationship between Chinese and Tibetan is surveyed, with special attention to the methodological and other scientific shortcomings of a recent attempt to compile a "handlist" of putative Sine-Tibetan lexical comparisons that further purports to substaniate that hypothesis. The hypothetical nature of all postulated early linguistic relationships is stressed; precisely like Indo-European or Altaic, Sino-Tibetan too always remains a hypothesis that may be argued for or against more or less convincingly, but it can never be "proven" to have existed. Phonological correspondences ("sound laws" are important evidence that support a postulated linguistic relationship; but to claim that ghey "prove" such hypotheses is completely to misunderstand their operation, as well as to mistake the powers of the comparative method of the Neogrammarians. Especially is this true when, as in the recent "handlist" that purports to "prove" the Sino-Tibetan hypothesis by constructing "sound laws", the formulations in question actually prove not to be worthy of the name, consisting rather of mere teleological "reconstructive exercises" in which quite imaginary forms of more and more complex configuration are postulated virtually at will, with almost no regard for documentable data. The paper concludes that, for the present at least, available evidence for a Sino-Tibetan linguistic relationship is weakened rather than enhanced by the petitio principii fallacy inherent in the "handlist"; it also studies in some detail the evidence of the Tibeto-Burman numerals, where the simplest (and hence also the most plausible) hypothesis to account for the data points toward early borrowings of these words from Chinese into nearby Tibeto-Burman languages, rather than toward a hypothesis of genetic relationship among all these languages.