Analysis of Speech Style and Speaker Attributes in Narrative and Daily Conversational Sentences

2022 Fiscal Year Final Research Report

• Project/Area Number: 18K11991
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 90020:Library and information science, humanistic and social informatics-related
• Research Institution: Future University-Hakodate
• Principal Investigator: Murai Hajime 公立はこだて未来大学, システム情報科学部, 教授 (70452018)
• Project Period (FY): 2018-04-01 – 2023-03-31
• Keywords: 発話 / 文体 / 属性 / 物語 / 日常会話

Outline of Final Research Achievements

In Japanese utterances, even if the content is the same, the style of the utterance changes diversely under the influence of various attributes of speakers and listeners. Based on the corpus of Japanese narrative sentences developed by the applicant and the daily conversation corpus developed by the National Institute for Japanese Language and Linguistics, various attributes that are thought to affect the writing style are added to each utterance sentence. They were added as tags, and the main elements that make up the style of the utterances were analyzed using statistical methods. As a result, several basic stylistic patterns in Japanese utterances were extracted. In addition, the differences between the usages of these stylistic patterns depending various attributes and the differences between stories and daily conversations were also clarified.

Free Research Field

Digital Humanities

Academic Significance and Societal Importance of the Research Achievements

日本語における発話の文体と属性の関係が明らかになることで，コンピュータを用いた話者とその属性の自動推定や，より自然な対話型人工知能の開発などが可能になると期待される．またこれらの成果は国語教育や小説の執筆の補助などにおいても有効に活用可能であると考えられる．