論漢藏語比較語言學的方法和依據

Comments on Methodology and Evidence in Sino-Tibetan Comparative Linguistics

本文探討有關漢藏語系語言之間親屬關係研究以及這些語言特徵構擬的認識論和方法論問題。其中主要的問題之一是我們沒有公認的標準，尤其是能經得起統計概率的考核。本文提出一種新的方法，即使用一套固定的特徵來衡量一些語言的親屬關係。這種方法能夠讓我們計算這些語言有共同來源的統計概率。

This paper presents epistemological and methodological problems found in work on the subgrouping of Sino-Tibetan languages and the reconstruction of features of the languages. A key problem is the lack of an accepted standard for judging this work, one that can stand up to statistical evaluation. An alternative methodology that involves using fixed sets of features to give us the statistical probability of common origin is suggested.