This talk presents the findings of a systematic review of 27 perceptual training studies conducted over the last 40 years which include the testing of generalization and retention of L2 speech learning. It overviews the use of these measures and examines how effective perceptual training is in promoting robust L2 speech learning. It also discusses the benefits and challenges of using these learning robustness evaluation methods.
It is widely acknowledged that second language (L2) speech acquisition is often challenging to adult learners. Certain non-native speech sounds tend to be more difficult to perceive and to produce accurately than others, even after years of experience with the L2. Adult L2 learners are therefore frequently characterized as having not only foreign accent but also accented perception (Strange 1995). Over the last four decades, numerous studies on L2 speech learning have applied training programs to improve the perception and production abilities of L2 learners and thus reduce degree of accentedness with a focus on nativeness or intelligibility (Sakai and Moorman 2018; Thomson and Derwing 2015). However, training studies with different language pairings have yielded complex findings due to the interplay of subject, task, and stimulus variables and assessment procedures (Bohn 2000; Thomson and Derwing 2015). To evaluate the efficacy of a training study, Logan and Pruitt (1995) propose that both generalization and retention of learning need to be examined. This talk presents the findings of a systematic review of 27 perceptual training studies conducted over the last 40 years which include the testing of generalization and/or retention of L2 speech learning. It overviews the use of these measures and examines how effective perceptual training is in promoting robust L2 speech learning. It also discusses the benefits and challenges of using these learning robustness evaluation methods. The limitations of the qualitative review are presented as well as suggestions for future research.
