The usage rate of Chinese characters in Japanese sentences decreased by the mid-20th century but has become stabilised since then. However, the question remains: why did the usage rate of Chinese characters become stable? This study thus focuses on lexical strata, a factor that was considered in previous research, to conduct a year-by-year analysis of the usage of Chinese characters for each lexical stratum: native Japanese words, Sino-Japanese words and borrowed words in Akutagawa prize-winning works. The results demonstrate that both the proportion of the lexical strata and the usage of Chinese characters for each lexical stratum have not changed substantially in contrast to the change seen prior to the mid-20th century when the usage rate of Chinese characters decreased. Moreover, a secondary analysis of the previous study’s results using the same methodology (a year-by-year analysis) suggests the possibility that the decrease in Sino-Japanese words written in Chinese characters constituted the main reason that the usage rate of Chinese characters also decreased by the mid-20th century. Furthermore, the analysis indicates that relatively fewer native Japanese words were written in Chinese characters because of the overall increase in native Japanese words caused by the growing number of those written in kana.

View full abstract

In this paper, we apply the structural topic model to extract topics from Meian and its sequel Zoku Meian, which are written by different authors, and investigate the changes in topic. As a result, we obtained topics that match the contents of the Meian and Zoku Meian, and visualized the time series changes of the two novels. Based on the analysis results, we concluded that although Zoku Meian is a product of stylistic imitation, the author Mizumura Minae exerted her imagination in the development of the plot and moved the narrative forward. Through this analysis, we proved the applicability and effectiveness of the topic model in detecting the changes of topic in novel.

View full abstract

ELAN is free software that is designed to annotate and transcribe streams of audio and video data. It is developed and maintained by Max Planck Institute for Psycholinguistics and runs on various operating systems. ELAN is widely used by researchers who analyze human behaviors, including but not limited to linguistic activities (e.g., single talk and conversation) and body movements (e.g., sign language and gestures). One main reason for its widespread acceptance is its convenience and ease of use. ELAN is very suitable for spoken/unspoken language researchers who often need to synchronize audio and/or video with transcriptions, replay specific parts of data repetitively, and summarize tagged data, and so on. The present paper introduces the useful functions of ELAN for quantitative and qualitative analyses of spoken languages to those who would like to learn ELAN. Specifically, the present paper introduces (1) the documentation functions of the spoken languages, (2) how to export the tagged data created with ELAN, and (3) the functions linking ELAN to Praat, which is another widely used software in the analysis of spoken languages.

View full abstract