Second Language Speech Intelligibility Training

Document Type


Presentation Date


Conference Name

Indiana University Linguistics Club Seminar Series

Conference Location

Bloomington, IN

Peer Review



Second-language speech intelligibility training

At Communication Disorders Technology, Inc. (CDT), in Bloomington, IN several English speech-intelligibility training systems have been developed. The earliest were language-specific systems based on phonological error analyses of the accented English speech of native speakers of several different L1's, Mandarin Chinese, Spanish, and Japanese (Rogers and Dalby, 2005). It was believed that what is usually called 'foreign accent' might better be referred to as incomplete mastery of the phonological system (first) and the phonetic details (second) of the target language. The first of these is probably the most important, and perhaps also the most easily remediated. For example, most Mandarin learners of English are difficult to understand because they tend to speak English with an ‘impoverished’ vowel system--one that might be more like their native language than it is like American English. An underlying assumption of the CDT training systems was that learning new phonological contrasts requires intensive training in both production and perception practice through minimal-pairs drill (Dalby and Kewley-Port, 2007). The speech production component of these systems employs feedback to the client that is provided by automatic speech recognition technology. Evidence that such computer-based training can be effective in learning new phonological contrasts will be presented (Burleson and Dalby, 2007).

But why bother? It has been shown that native listeners can adapt to 'foreign accents' quite rapidly and effectively, at least in laboratory experiments (Clarke and Garrett, 2004, Bradlow and Bent, 2008). Since that is so, can it not be assumed that accented speech is just as effective in real communicative situations as native speech? The first evidence that may be relevant to answering this question comes from work done in David Pisoni’s lab at IU with early synthetic speech, where it was found that synthetic speech that was nearly equal in intelligibility to natural speech under ideal listening conditions, was less intelligible when listeners faced simultaneous cognitive tasks of different sorts (see Winters and Pisoni, 2003 for a summary). More direct evidence comes from testing with accented speech that measured intelligibility in sub-optimal listening conditions. Rogers, Dalby and Nishi (2004) showed that short sentences spoken by even highly proficient Mandarin learners of English were degraded in intelligibility more than was native speech when mixed with background noise. More recently, attempts to replicate this result for isolated words have been conducted and will be discussed (Dalby and Rogers, 2009).


Bradlow, A. R. and Bent, T. (2008). Perceptual adaptation to non-native speech. Cognition, 106, 2, pp. 707-729.

Burleson, D. F. and Dalby, J. (2007) Training segmental productions to improve second language intelligibility. Journal of the Acoustical Society of America, Vol. 121, No. 5, Pt. 2, p. 3072 (A).

Clarke, C. M. and Garrett, M. F. (2004). Rapid adaptation to foreign-accented English. J. Acous. Soc. Amer. 116(6) pp. 3647-3658.

Dalby, J. and Rogers, C.L. (2009) Intelligibily of Spanish-accented English words in noise. Journal of the Acoustical Society of America, Vol. 125, No. 4, Pt. 2. p. 2776 (A). Presented at the Second Special Workshop on Cross-Language Speech Perception, Portland, OR, May 2009.

Burleson, D. F. and Dalby, J. (2007) Training segmental productions to improve second language intelligibility. Journal of the Acoustical Society of America, Vol. 121, No. 5, Pt. 2, p. 3072 (A).

Dalby, J. and Kewley-Port, D. (2007). Design features of three computer-based speech training systems. In M. Holland and F.P. Fisher (Eds.) The Path of Speech Technologies in Computer Assisted Language Learning, pp. 155-173. New York: Routledge Publishers

Rogers, C. L. and Dalby, J. (2005). Forced-choice analysis of segmental production by Chinese-accented English speakers. Journal of Speech, Language, and Hearing Research, Vol. 48, pp. 306-322.

Rogers, C.L., Dalby, J. and Nishi, K. (2004). Effects of noise and proficiency on intelligibility of Chinese-accented English. Language and Speech, Vol. 47, part 2, pp 139-154.

Winters, S. J. and Pisoni, D. B. (2003). Perception and comprehension of synthetic speech. Speech Research Laboratory Progress Report No. 26, Indiana University.


speech intlligibility, second language, speech training, computer-based training


Communication Sciences and Disorders | Speech and Hearing Science | Speech Pathology and Audiology

This document is currently not available here.