Quantitative Research Methods for Sub-dialect Division of Jeollanam-do: Using Levenshtein Distance and Phonological Network Analysis

The purpose of our study is to divide the sub-dialects of Jeollanam-do using Levenshtein distance and network analysis. For the first time, we measured the Levenshtein distance on Korean dialect data, and constructed a phonological network which is based on this. Then, clustering analysis was performed to divide the sub-dialects of Jeollanam-do and compared them with the dialect division of Lee Ki-gap(1984, 1986). Through the Levenshtein distance measurement, the closest regions in Jeollanam-do were Yeong-am gunseo and Yeong-am Sijong, Yeong-am gunseo and Yeong-am Haksan, Yeong-am Sijong and Yeong-am Haksan, Wando Wansdo-eup and Wando Gogeum regions. But the farthest regions were Gwang-yang Jinwol and Jindo Imhui, Gwang-yang Jinwol and Jindo Jodo area. This distance was converted into network weights to construct a phonological network in Jeollanam-do. On the phonological network in the sub-dialect of Jeollanam-do, we measured the degree strength centrality, closeness centrality and betweenness centrality. The west of Jeollanam-do tended to be higher in terms of centrality, and the core area of sub-dialect division were judged to be Hampyeong and Yeong-am, and the most heterogeneous sub-dialect was Gwang-yang Jinwol. As a result of clustering analysis on the sub-dialects of Jeollanam-do, the sub-dialects of Jeollanam-do are largely divided into the east and the west, and the east was again divided into two areas, and the west was also divided into two areas. Interestingly, Boseong Ungchi is a unique sub-dialect region in the west, and appears to be a transitional area between the east and the west. This study will be meaningful in that it is the first study that uses Korean dialect phonological data and divide sub-dialects using a quantitative method-Levenshtein distance and network analysis.