Sobre el factor de la casualidad en la comparación lingüística

Abstract

In this paper the author describes and compares two different ways of calculating the factor of chance that interferes with the comparative linguistic practice aimed at establishing genetic relationships among languages or language families. The discussion concentrates mainly on the theoretical and methodological sides of the problem, although some attention is also paid to the connection of the mathematical estimations with the results achieved by the comparison.
   Through the analysis of the probabilistic arguments behind both the "sceptical" (Ringe 1995a) and the "optimistic" (Bengtson & Ruhlen 1994) views of new (especially distant) genetic relationships, the author of this article tries to show the inadequacy (if not the inability) of such attempts to provide a common mathematical basis for calculating the amount of chance linguistic coincidences which are to be ruled out and, consequently, taken into account by comparativists when they attempt to prove the existence of any language family. Therefore, as the only effective means of guaranteeing certain security in the phonetic/semantic matchings among the languages supposed to be related, there remains the identification of the sets of systematic correspondences (and of those regular, genetically significant commonalities that underlie them).