ISCA Archive Interspeech 2024
ISCA Archive Interspeech 2024

Analysis and Visualization of Directional Diversity in Listening Fluency of World Englishes Speakers in the Framework of Mutual Shadowing

Yu Tomita, Yingxiang Gao, Nobuaki Minematsu, Noriko Nakanishi, Daisuke Saito

English is spoken as a lingua franca with a diversity of pronunciations, called accents, and they have been well studied so far. In this study, a diversity of listening behaviors are focused on, and listening disfluencies are measured objectively while listening to World Englishes (WE). When speaker X listens to Y fluently, it does not always mean that Y listens to X fluently. After collecting different passages read aloud by different WE speakers, the collected oral passages are shadowed by the speakers themselves to quantify their listening disfluencies. Results show that, when X listens to Y, X's listening disfluency becomes larger when Y's pronunciation deviates from X's to a larger degree. Further, a method is proposed to visualize simultaneously a) how fluently a speaker listens to WE speakers and b) how fluently the WE speakers listen to that specific speaker. With this visualization, WE speakers are grouped based on their communicability in global contexts.