Article/Report  |   December 2007
Training Japanese Listeners to Perceive American English Vowels: Influence of Training Sets
Author Affiliations & Notes
  • Kanae Nishi
    Indiana University
  • Diane Kewley-Port
    Indiana University
  • Contact author: Kanae Nishi, who is currently at Boys Town National Research Hospital, 555 North 30th Street, Omaha, NE 68131. E-mail: nishik@boystown.org.
Article Information
Special Populations / Cultural & Linguistic Diversity / Speech, Voice & Prosody / Speech
Article/Report   |   December 2007
Training Japanese Listeners to Perceive American English Vowels: Influence of Training Sets
Journal of Speech, Language, and Hearing Research, December 2007, Vol. 50, 1496-1509. doi:10.1044/1092-4388(2007/103)
History: Received July 13, 2006 , Accepted June 2, 2007
Journal of Speech, Language, and Hearing Research, December 2007, Vol. 50, 1496-1509. doi:10.1044/1092-4388(2007/103)
History: Received July 13, 2006; Accepted June 2, 2007
Web of Science® Times Cited: 13

Purpose: Studies on speech perception training have shown that adult 2nd language learners can learn to perceive non-native consonant contrasts through laboratory training. However, research on perception training for non-native vowels is still scarce, and none of the previous vowel studies trained more than 5 vowels. In the present study, the influence of training set sizes was investigated by training native Japanese listeners to identify American English (AE) vowels.

Method: Twelve Japanese learners of English were trained 9 days either on 9 AE monophthongs (fullset training group) or on the 3 more difficult vowels (subset training group). Five listeners served as controls and received no training. Performance of listeners was assessed before and after training as well as 3 months after training was completed.

Results: Results indicated that (a) fullset training using 9 vowels in the stimulus set improved average identification by 25%; (b) listeners in both training groups generalized improvement to untrained words and tokens spoken by novel speakers; and (c) both groups maintained improvement after 3 months. However, the subset group never improved on untrained vowels.

Conclusions: Training protocols for learning non-native vowels should present a full set of vowels and should not focus only on the more difficult vowels.

Acknowledgments
A portion of this article was presented at the 149th meeting of the Acoustical Society of America in Vancouver, Canada, in May 2005. This work was supported by National Institutes of Health (NIH) Postdoctoral Fellowship F32 DC-006313 (awarded to the first author) and NIH Research Grant R01 DC-02229 (awarded to the second author). Manuscript preparation was supported, in part, by NIH Research Grant R01 DC-04300 (awarded to Patricia G. Stelmachowicz). William B. Mills wrote the training program used here. James J. Jenkins and Winifred Strange made significant contributions to the development of the initial research design. The consultation with Charles S. Watson and James D. Miller regarding the feedback system was invaluable. Finally, we wish to thank the study participants.
Order a Subscription
Pay Per View
Entire Journal of Speech, Language, and Hearing Research content & archive
24-hour access
This Article
24-hour access