The information of “regularity,” so-called phonological correspondences, morphological and grammatical rules, is always present in the data of languages and dialects, but it cannot necessarily be analyzed in a numerical way that is applied to lexical items and forms in previous linguistics research (e.g., lexicostatistics). Therefore, the shortcoming of the methodologies can potentially prevent previous studies from approaching linguistic issues due to a lack of appropriately quantified information of “regularity.” This paper attempts to provide some possibility regarding how researchers can extract and quantify the regularities in languages and dialects, using the data of Ainu dialects in Asai (1974). We propose principles for extracting the type and regularity from linguistic materials and statistical quantification methods that quantify the type and regularity more appropriately from the viewpoints of both linguistics and statistics. Consequently, the classification result obtained by incorporating the quantified types and regularities further clearly shows the western and eastern group axis comprising the Ainu dialects other than the Samani and Soya dialects of Hokkaido in terms of the northern and southern axis through the Samani and Soya dialects, the classification among the northeastern Hokkaido Ainu dialect group, and the classification among the southwestern Hokkaido Ainu dialect group. Our results have clarified the characteristics of Asai’s (1974) Ainu dialect classification and suggested that regularity information, excluded from Asai’s (1974) classification as regularity while reflecting his viewpoints, can be appropriately extracted and quantified by our proposed methods. Notably, our proposed methods can be applied to any type of information, including the types and regularities thereof. In other words, these methods have the potential to enable researchers to extract and quantify the types and regularities across the present subdivisions of linguistics, phonology, morphology, and syntax, among others. The application of these methods can also facilitate the quantification of vocabulary items and the classification of the other languages (e.g., Native North Americas), which should have more complex phonological and morphological characteristics than the Ainu language.
Ono et al. (Thu,) studied this question.