Spoken English Discrimination (SED) Training and Speech Production in Multilingual Malaysian Speakers
Using high variability phonetics training (HVPT) can improve bilingual speakers’ perceptual performance in discriminating difficult phonemic contrasts in their second language (L2 (e.g., Bradlow, Akahane-Yamada, Pisoni, & Tohkura, 1999)). This perceptual learning has been found to generalize to novel words (Wang & Munro, 2004), novel speakers (Nishi & Kewley-Port, 2007; Richie & Kewley-Port, 2008) and even to speech production (Bradlow et al., 1997). Spoken English Discrimination (SED) training is a computerized individual training program designed to improve bilingual/multilingual speakers’ bottom-up perceptual sensitivity in discriminating difficult (L2) phonemic contrasts. Some of the key features include natural spoken stimuli, stimuli spoken by multiple speakers and background babble that could be either be at a constant level or an adaptive staircase procedure that individualizes the level of background babble. We present two experiments (experiment 1 - SED training, experiment 2 – speech production evaluation) that investigated the effectiveness of the SED training paradigm and whether the effects of SED training can generalize to speech production. The first experiment investigated the potential benefits of using an adaptive procedure compared to a constant level of background babble using minimal pairs of the English /ε/-/æ/ phonemic contrasts as the training materials. It also examined whether SED training would generalize to untrained /ε/-/æ/ minimal pairs and untrained minimal pairs of the English /t/-/d/ phonemic contrasts. The second experiment examined whether SED training lead to improvements in production intelligibility. The results showed that the Adaptive Staircase SED was the more effective training paradigm as it showed greater training benefits and its effect generalized better to the untrained /t/-/d/ phonemic contrast. However, the adaptive staircase design did not generate the same benefits compared to constant level of background babble. The constant SED training paradigm leads to improvements in production intelligibility but only in the /æ/ phoneme. This research could be extended further using different training materials and speakers from different language backgrounds to fully examine the potential benefits of the SED training program in terms of improving phonemic sensitivity and speech intelligibility.