Perception and articulation of Cantonese vowel contrasts by native speakers of English and Mandarin

ECS 2021/2022
(PI Jonathan Havenhill)
Early Career Scheme (ECS), University Grants Council (UGC), Hong Kong
Amount: 641,00 HKD

Abstract
This project will investigate how native speakers of English and Mandarin produce and perceive the vowels of Cantonese. Previous work shows that perceiving and producing nonnative vowel contrasts can be challenging for second language learners, with the degree of difficulty depending on the similarity to vowels from the speaker’s native language. It remains unclear, however, whether nonnative speakers rely on acoustic similarity or articulatory similarity, with competing theories arguing for each. Notably, few studies have considered nonnative production of vowels in terms of their underlying articulatory gestures, such as lip rounding and tongue position, instead relying on acoustic measurements, which can be ambiguous.

The study focuses on the Cantonese vowel sounds /i y u/ (as in sı̄ 1 絲 ‘silk’, sȳu 1 書 ‘book’, fů 3 褲 ‘pants’), and /ɛ œ ɔ/ (as in sē1 些 ‘some’, hōe 1 靴 ‘boot’, sō1 梳 ‘comb’). The vowel pairs /i u/ and /ɛ ɔ/ are distinct in both tongue position and lip rounding, with similar vowels found in all three languages. On the other hand, /y/ and /œ/ bear similarity to both /i ɛ/ and /u ɔ/; they are produced with fronted tongue positions like /i ɛ/ and rounded lips like /u ɔ/. Both Cantonese and Mandarin contain /y/, while /œ/ exists in Cantonese, but not Mandarin. English contains neither /y/ nor /œ/, but these vowels are acoustically similar to English vowels in some contexts.

Participants for this study include native English and Mandarin speakers who are advanced learners of Cantonese, as well as those with no prior Cantonese experience. A perception experiment will test participants’ ability to discriminate Cantonese vowel pairs (such as /i/-/y/ and /y/-/u/), and examine how they categorize Cantonese vowels relative to their native and nonnative vowel inventories.

The production experiment will collect acoustic speech recordings, ultrasound images of tongue position, and video of lip rounding. Data will be analyzed to determine whether nonnative speakers more closely match native Cantonese vowels in terms of acoustics or articulation, and whether their production reflects articulatory configurations and timings that are present in their native language.

Results of this project will be used to test theories of phonology and second language acquisition, and provide essential articulatory data that is especially lacking for Cantonese and Mandarin. Findings will pave the way for future studies investigating the use of ultrasound tongue imaging as a means of teaching new sounds to second language learners.