In order to describe the historical changes of "vocabulary" as a set of words, it is necessary to prepare a corpus for this purpose. It is also necessary to devise statistical methods to find significant trends of change using the corpus. In this paper, I introduce the contents and characteristics of the diachronic corpus of the Mainichi Shimbun since the mid-20th century, which I have created and expanded, and report the results of my analysis of the increase and decrease of word usage in the corpus using ridit analysis.

View full abstract

Using data from the 2015 nationwide web survey on dialect consciousness, which was conducted in 2015 and received responses from approximately 10,000 males and females aged 20 and older nationwide, we extracted speaker types using latent class analysis (LCA) and attempted to describe and interpret the characteristics of each type. The seven speaker types extracted by the LCA, in descending order of class size, are: (1) ‘moderate code switchers (between common language (CL) and dialect)’, (2) ‘active dialect speakers’, (3) ‘active code switchers’, (4) ‘CL speakers who grew up in dialect environment’, (5) ‘CL speakers’, (6) ‘indecisive respondent type1’, and (7) ‘indecisive respondent type2’. In terms of the average membership probability, clear regional and age group effects are found among the five major types with large class sizes. The regional effects are interpreted to indicate the regional differences in the consciousness of dialect and CL in contemporary Japanese society, and the age group effects are interpreted to reflect the transition in the consciousness of speakers in postwar Japanese society from ‘CL speakers who grew up in dialect environment’ to ‘CL speakers’ and ‘active dialect speakers’.

View full abstract

This study focuses on five Japanese transitive and intransitive verb pairs (hairu “enter,” ireru “put in,” noru “get on,” noseru “put on,” kowareru /kowasu “break,” and aku /akeru “open,”) and compares the frequency of verb forms that contain six morphemes (-ru (non-past), -ta (past), -te (imperative), -chau (perfective), -nai (negative), and -teiru (continuous)) in spontaneous speech data of children age 1-3 years old their mothers. Although there was no significant difference in the overall frequency of user between intransitive and transitive verbs, intransitive verbs were used significantly more frequently with four morphemes (-ru, -ta, -nai, and -teiru). This trend was observed for both children and mothers, suggesting that children's verb use is influenced by input from their mothers.

View full abstract

This paper outlines a statistical method for checking whether there are differences in the mean values of linguistic data. First, I will give an overview of the definition and applications of mean value. Then, I will discuss some applications of mean values in the field of linguistic research. Second, I will discuss t-test and one-way ANOVA as methods of comparing means based on statistical methods, show how to calculate them, and analyze them based on actual data.

View full abstract