The Intelligibility of Vietnamese-Accented English to Artificial Intelligence Software and Asian Listeners
Loading...
Authors
Trang Minh Thi Pham, Hanoi University of Science and Technology, Vietnam
Issue Date
2023
Type
Working Paper
Language
Keywords
Alternative Title
Abstract
Given Vietnamese students' limited speaking abilities, this paper aims to offer useful insights to English educators regarding the pronunciation patterns of Vietnamese-accented English by assessing its intelligibility by an artificial intelligence (AI) speech-to-text transcription and Asian human listeners. This research project was conducted in two phases. In the first phase, recordings of two Vietnamese speakers of English were evaluated by Otter, a real-time transcription AI tool. In the second phase, the same recordings were evaluated by 40 Asian human listeners for intelligibility. Additionally, brief interviews were conducted to gather insights into the listeners' responses and their listening experiences. Results revealed a relationship between speaking proficiency and intelligibility, based on both the AI's and Asian listeners' assessment. Pronunciation variations such as sound confusion, omission and the speed at which speech was produced were all contributing factors to the hindrance of speakers' intelligibility. The paper concludes by offering pedagogical recommendations for educators teaching English pronunciation to Vietnamese students.
Description
Citation
Publisher
Hawaii Pacific University
License
Journal
TESOL Working Paper Series
Volume
21
Issue
PubMed ID
DOI
ISSN
2573-1467