Training Japanese listeners to identify English /r/ and /l/: IV. Some effects of perceptual learning on speech production

Abstract

This study investigated the effects of training in/r/-/l/ perceptual identification on /r/-/l/ production by adult Japanese speakers. Subjects were recorded producing English words that contrast /r/ and /l/ before and after participating in an extended period of /r/-/l/ identification training using a high-variability presentation format. All subjects showed significant perceptual learning as a result of the training program, and this perceptual learning generalized to novel items spoken by new talkers. Improvement in the Japanese trainees' /r/-/l/ spoken utterances as a consequence of perceptual training was evaluated using two separate tests with native English listeners. First, a direct comparison of the pretest and post-test tokens showed significant improvement in the perceived rating of /r/ and /l/ productions as a consequence of perceptual learning. Second, the post-test productions were more accurately identified by English listeners than the pretest productions in a two-alternative minimal-pair identification procedure. These results indicate that the knowledge gained during perceptual learning of /r/ and /l/ transferred to the production domain, and thus provides novel information regarding the relationship between speech perception and production.

FIG. 1

Percent correct perceptual identification performance for trained (left panel) and control (right panel) subjects at pretest, post-test, and the two tests of generalization. The error bars represent one standard error from the mean.

FIG. 2

Distribution of the Japanese trainees’ identification responses at pretest and at post-test by phonetic environment. The upper panel shows the /r/ stimuli; the lower panel shows the /l/ stimuli. The environments are: i = initial, ic = initial cluster, m = medial, f = final. The error bars represent one standard error from the mean.

FIG. 3

Distribution of preference ratings across all seven response categories by the American English listeners for the Japanese trained (left panel) and control (right panel) subjects’ pre- and post-test productions. A response of “1” indicated that the pretest version was “much better” than the post-test version, “4” indicated no noticeable difference between the pre- and post-test versions, and “7” indicated that the post-test token was “much better” than the pretest version.

FIG. 4

Percent correct performance for trained (left panel) and control (right panel) subjects’ productions as judged by American English listeners in the minimal pair identification task. The open bar represents the full set of pretest tokens, the slashed bar represents the post-test tokens that were included in the perceptual training set, and the solid bar represents the post-tokens that were not included in the perceptual training set. The error bars represent one standard error from the mean.

FIG. 5

Vector plot of individual Japanese subjects’ perceptual identification accuracy (x axis) and production identification accuracy (y axis) from pretest to post-test. Each individual subject’s performance is indicated by a numbered vector. The mean performance is represented by the bold arrow. The diagonal represents the hypothetical vector location and orientation for a perfect correlation between perception and production.