The purpose of this study is to investigate how the personalities of characters in TV anime are conveyed through their choice of words and to elucidate the characteristics associated with each personality type. Initially, the characters' personalities were identified using TEGⅡ (Tokyo University Egograms Ⅱ). Subsequently, their dialogues underwent morphological analysis to extract distinctive words for each type, utilizing log-likelihood ratios. As a result, characters with strict personalities, classified as the CP (Critical Parent) type, exhibited the extraction of second-person pronouns ANTA/ANATA. In addition, words related to sentence endings, such as auxiliary verbs DESU/MASU and sentence-ending particles WA/YO/KASHIRA were identified. Moreover, words expressing negation, such as the auxiliary verb ZU and the adjective NAI, as well as imperative expressions like the verb NASARU, were also extracted. On the other hand, characters with free personalities, classified as the FC (Free Child) type, showed the extraction of interjections like WAA/UWA and adjectives expressing emotions, evaluations, and degrees such as SUGOI/KAWAII/HIDOI. Additionally, words indicating addressing someone with an affectionate suffix like CHAN/SAN and the sentence-ending particle MONO evoking a sense of innocence were extracted. This analysis sheds light on the connection between the personalities of anime characters and the choice of words they use.

View full abstract

This study aims to identify the linguistic elements related to readability of Japanese texts for second language learners, and construct a readability assessment model with high accuracy and explainability that meets the needs of language education. Specifically, 86 linguistic features across 5 categories were extracted from Japanese textbooks with difficulty levels, and readability models were constructed and evaluated. When comparing 4 classification models for automatic difficulty assessment, SVM (Support Vector Machine) showed best performance with an accuracy (ACC) of 0.898 in judging the readability of Japanese texts. Furthermore, feature selection using a stepwise approach identified 35 highly relevant factors to construct a model maintaining 0.880 accuracy while enhancing simplicity and explainability. Additionally, readability scores were quantified into three perspectives for visualization of prediction results. Thus, the readability model developed in this study not only demonstrated high predictive accuracy, but also contributed to explainability desired in the field of language education.

View full abstract

This paper presents an overview of a newly released online vocabulary analysis tool, English/Japanese Word Frequency Table Generator (EJWFTG). EJWFTG automatically generates an integrated vocabulary frequency list from multiple texts (Ishikawa 2024). Section 1 illustrates why an integrated vocabulary list is needed in language research. Section 2 provides a brief survey of EJWFTG and touches upon its main features. Section 3 describes the process of creating a vocabulary list from sample Japanese novel texts. Then, Section 4 presents two examples of utilizing the obtained vocabulary lists for research: a reconsideration of the importance of words based on their frequencies and ranges and an application of a multivariate analytic method to the frequency table for text classification.

View full abstract

In this study, we tasked ChatGPT 4.0 with predicting the semantic categories of basic vocabulary nouns extracted from the “Nihongo kyoiku goihyo (Japanese Language Education Vocabulary List).” The correct data was based on the “Bunruigoihyo (Word List by Semantic Principles) (ver.1.0.1),” and statistical analysis was conducted on the accuracy and reliability of the predictions. The analysis revealed that ChatGPT's predictions and the semantic categories of “sections” in the “Bunruigoihyo (Word List by Semantic Principles)” substantially coincide (with an average agreement rate of κ=.706), yet it was evident that there is a notable tendency for misclassification in nouns related to “Shutai (agents)” and “Katsudo (activities).“ Based on these findings, it is considered that generative AI, as represented by ChatGPT, can become one of the important research tools for quantitative research if used with caution.

View full abstract

This paper discusses the appropriate handling of morphologically analyzed vocabulary data, focusing on technical aspects such as morphological analysis tools, analysis dictionaries, and morphological information. However, when creating vocabulary data, it is important to describe the process of data creation and annotation, as well as knowledge of the target data, in sufficient detail to ensure the reproducibility of the research and enable convincing verification.

View full abstract