TRAINING JAPANESE LISTENERS TO IDENTIFY ENGLISH VERTICAL-BAR-R-VERTICAL-BAR AND VERTICAL-BAR-L-VERTICAL-BAR .3. LONG-TERM RETENTION OF NEW PHONETIC CATEGORIES

Monolingual speakers of Japanese were trained to identify English /r/ and /l/ using Logan et al.'s [J. Acoust. Soc. Am. 89, 874-886 (1991)] high-variability training procedure. Subjects' performance improved fr om the pretest to the post-test and during the 3 weeks of training. Pe rformance during training varied as a function of talker and phonetic environment. Generalization accuracy to new words depended on the voic e of the talker producing the /r/-/l/ contrast: Subjects were signific antly more accurate when new words were produced by a familiar talker than when new words were produced by an unfamiliar talker. This differ ence could not be attributed to differences in intelligibility of the stimuli. Three and six months after the conclusion of training, subjec ts returned to the laboratory and were given the post-test and tests o f generalization again. Performance was surprisingly good on each test after 3 months without any further training: Accuracy decreased only 2% from the post-test given at the end of training to the post-test gi ven 3 months later. Similarly, no significant decrease in accuracy was observed for the tests of generalization. After 6 months without trai ning, subjects' accuracy was still 4.5% above pretest levels. Performa nce on the tests of generalization did not decrease and significant di fferences were still observed between talkers. The present results sug gest that the high-variability training paradigm encourages a long-ter m modification of listeners' phonetic perception. Changes in perceptio n are brought about by shifts in selective attention to the acoustic c ues that signal phonetic contrasts. These modifications in attention a ppear to be retrained over time, despite the fact that listeners are n ot exposed to the /r/-/l/ contrast in their native language environmen t.