Analysis of L2 Learners’ Progress of Distinguishing Mandarin Tone 2 and Tone 3

Yue Sun, Win Thuzar Kyaw, Jinsong Zhang, Yoshinori Sagisaka


Many studies have shown the effectiveness of perceptual training to improve L2 learners’ ability to distinguish Mandarin tones. In this paper, we quantified learners perceptual characteristics on discriminating the most difficult tone pair, Tone 2 and 3 in Mandarin before and after training. L2 learners’ categorical perception is measured by fitting a sigmoid curve to the identification responses with average F0 height be the acoustic dimension. The boundary location of the two tones in L2 learners’ perception space is significantly improved to a higher F0 height after training. Regression analysis indicated that 𝛿F0 and 𝛿t of the initial falling of the concave F0 shape are the key acoustic features for native speakers in discrimination. L2 learners rely on not only the initial fall but also the 𝛿F0 of the final rise to discriminate the tones. A detailed analysis using cognitive measurements reports an increasing attention on the initial fall of the F0 contour for L2 learners after perceptual training. These results confirmed that directing the attention to key acoustic features is essential for L2 learners to improve their categorical perception of novel speech contrasts.


 DOI: 10.21437/Interspeech.2018-1983

Cite as: Sun, Y., Kyaw, W.T., Zhang, J., Sagisaka, Y. (2018) Analysis of L2 Learners’ Progress of Distinguishing Mandarin Tone 2 and Tone 3. Proc. Interspeech 2018, 2549-2553, DOI: 10.21437/Interspeech.2018-1983.


@inproceedings{Sun2018,
  author={Yue Sun and Win Thuzar Kyaw and Jinsong Zhang and Yoshinori Sagisaka},
  title={Analysis of L2 Learners’ Progress of Distinguishing Mandarin Tone 2 and Tone 3},
  year=2018,
  booktitle={Proc. Interspeech 2018},
  pages={2549--2553},
  doi={10.21437/Interspeech.2018-1983},
  url={http://dx.doi.org/10.21437/Interspeech.2018-1983}
}