An Acoustic Phonetic Analysis of Mandarin English Vowel Spaces

Abstract The objective of this project is to investigate whether or not Mandarin speakers produce English vowels intelligibly. This research replicates the methodology used by Peterson and Barney (1952) and Hillenbrand et al (1995) in their studies of General American English and Midwest English. In this study, 20 Mandarin speakers of English (10 males and 10 females) are asked to read 11 words contain all 11 English phonetic vowels. The participants in the study are college-aged students studying at Saint Cloud State University. They are divided into 4 subgroups based on their genders Length of Residency (LOR) in the USA. The four subgroups are as follows: Five males with LOR \u3c 1 year Five males with LOR \u3e 1 year Five females whose LOR \u3c 1 year Five females whose LOR \u3e1 year Their speech samples were recorded using the same laptop (Model: MacBook Air; System: OS X Yosemite; Version: 10.10.5). The acoustic phonetic software Praat (Boersma & Weenink, 2018) is used for data collecting and measuring. The data is annotated and collected manually. The acoustic correlates measured are: F0, F1, F2, F3, Duration, and Intensity. F1 and F2 are mostly focused in this research because they are the most robust cues for assessing the intelligibility of vowels. These measurements are used to determine whether or not the vowels produced by the participants mask each other. Masking thresholds are based on Koffi (2017). The effect of masking on intelligibility is assessed using Catford’s (1987) Relative Functional Load calculations. Acoustic vowel spaces are created to help visualize how the various vowels produced by Mandarin speakers of English compare with those produced by their GAE counterparts. The analysis shows that intelligibility is severely compromised in the following vowel pairs: [ɪ] vs. [e], [u] vs. [ʊ], and [ɛ] vs. [æ]. Other vowel pairs are challenging but our data indicates that these are the vowels that the 20 participants have hard time producing irrespective of their LOR in the USA. Finally Pedagogical implications and applications are drawn for teaching these vowels to Mandarin speakers of English