Text

theses

A recent study demonstrated that high-variability perceptual training (HVPT) helped native Cantonese-speaking adults distinguish among English words that differ in the voicing quality of the final phoneme (e.g., unvoiced or voiced, as in tap vs. tab) (Shum, Kit-fong Au, Romo & Jun, 2021). Using the production data from Shum et al.’s (2021) study, we investigated whether the production qualities of their speech were also influenced by training. Research shows that training can make second language (L2) speech more comprehensible, but the impact training has on L2 acoustic features is not fully understood.The present thesis addressed this gap by measuring L2 comprehensibility (Study 1), conducting fine-grained analyses on select L1 and L2 acoustic features of L2 speech (Study 2), and examining the acoustic qualities of the English stimuli (Study 3). While L2 comprehensibility did not improve after training, L2 speakers tended to use acoustic features found in English (i.e., vowel duration) rather than Cantonese (i.e., aspiration) in production.
Further inspection of the data suggested that L2 productions that were acoustically more English- like, were also more comprehensible. Lastly, an examination of the English stimuli revealed that vowel duration was robust, perhaps leading L2 speakers to perceive, and later implement vowel duration differences during production.
These findings suggest that the specific qualities associated with training stimuli can influence the degree to which training is successful. More research is needed to investigate the link between L2 perception and production to fully understand how the perception of robust cues may influence production.

http://dissertations.umi.com/gsnb.rutgers:12433

M.S.

Includes bibliographical references

doi:10.7282/t3-31cq-p340

The author owns the copyright to this work.