This study introduces the Web-Based Corpus Analyzer (WBCA), a browser-based corpus analysis tool designed to support transparent and reproducible corpus research through explicit user-controlled operations. Using a publicly available demo corpus, the study illustrates how WBCA integrates core corpus-linguistic functions—Keyword-in-Context (KWIC), collocation analysis, high-frequency feature extraction, and keyness analysis—within a unified, KWIC-centered workflow. The system supports both plain-text and linguistically annotated corpora while preserving identical analytical logic across modes. All computations are executed locally in the web browser, ensuring data privacy and enabling interactive exploration through filters, sorting, and contextual highlighting. The focus is placed on documenting the analytical workflow, statistical measures, and interface design of the initial public release of WBCA rather than on linguistic interpretation.

抄録全体を表示

This study investigates the use of reporting clauses in the fourteen Wessex novels by Thomas Hardy. A corpus based on the Text Encoding Initiative (TEI) was constructed by manually annotating all instances of character speech, and Python was used to extract the reporting verbs and their adverbial modification. Quantitative analyses show that verbs occur far more frequently than adverbs, while adverbs display greater lexical diversity. The diachronic comparison reveals a reduction in reporting clauses in Hardy's later works. The gender-based comparison further reveals contrasting patterns: female characters tend to follow a rapport orientation (favoring affective and relational expression), while male characters follow a report orientation (favoring assertive or informational expression). These patterns represent overall distributional tendencies in the corpus rather than uniform character traits across all novels. The findings provide a new empirical foundation for the study of speech presentation in English fiction.

抄録全体を表示

This study compares two CEFR-aligned lexical resources, the English Vocabulary Profile (EVP) and the Global Scale of English (GSE), to examine agreement in lexical difficulty assignment and patterns of divergence. Semantic equivalence between definitions of shared words and phrases was identified using a large language model, and CEFR levels for equivalent senses were quantitatively compared. Although strong correlations were found for both words and phrases, exact level agreement was limited to 39.9% for words and 24.9% for phrases. EVP tended to assign higher levels than GSE, particularly for verbs and verb phrases. These results seem to reflect differences in the corpora used for level assignment and suggest the need to distinguish between receptive and productive lexical difficulty indicators.

抄録全体を表示