Statistics and comparative studies: "quatitative diachrony"
Résumé
In this paper I discuss how to apply statistics to comparative linguistic studies. I show that the main issue concerning the use of statistics in linguistic stusies lies not in the lack of sophisticated methods, but rather in the extraction of really usefull knowledge from simple statistical data. This paper illustrates how qualitative results can be obtained by applying quantitative criteria. All the domains of linguistic reconstruction can be approached by following the same "ideology" in calculation and analysis. The strategic point in this ideology is that quatitative and qulitative divergences observes in genetically related languages are considered as EQUALLY RELEVANT. This approach is illustrated by concrete results concerning dozens of various languages.