| Description |
This study addressed the role of talker variability in the perception of nonnative contrastive phonemes by adult second language (L2) learners who had no prior knowledge with the target language. Specifically, the study explored how training with varying talkers could affect native English speakers' acquisition of the Arabic pharyngeal-glottal contrast, which is not distinctive in their native language. The present study also examined the effects of task type on learners' word recognition ability. To accomplish this, the present study included two main experiments: Experiment 1 (nonlexical task) and Experiment 2 (lexical task). Sixty adult native speakers of English (with no Arabic experience) participated in the two experiments, 30 subjects in each experiment who were randomly assigned to either a single- or multiple-talker word learning groups. Subjects in the two experiments were presented with nine nonword minimal pairs where six pairs contrasted the Arabic /ħ/ and /h/ phonemes and three pairs included familiar sound contrasts (i.e., /s/ and /ʃ/). The nine nonword pairs were assigned to pictures indicating their meanings and subjects learned the nine nonword pairs in the training phase and were then tested on them later in the testing phase. Findings of Experiment 1 demonstrated a significant effect of training type (p < .001), a significant effect for item type (p < .001), and a significant interaction of training type and item type (p < .001) for subjects in the multiple-talker environment. That is, their performance was more accurate (91.5%) than the single-talker group (67%). The same significant findings were found in Experiment 2 where again, subjects in the multiple-talker training group performed more accurately on test items better than their counterparts in the single-talker training group (single-talker group = 65%; multiple-talker group = 87%). Overall, the results of this experiment provided evidence that multiple-talker training did have a significant effect on the subjects' recognition of the target contrast in a nonlexical discrimination task with above 88% average accuracy. Findings also provided evidence supporting learners' ability to establish lexical representations for the newly learned words that included the target Arabic contrasting phonemes with above 83% average accuracy for only the multiple-talker training group. Even though subjects' scores differed on the two discrimination tasks, this difference was found to be statistically insignificant. That is, subjects' ability to discriminate the novel contrasts was the same on the lexical task as on the nonlexical task regardless of the two tasks' distinct demands. Findings of the two experiments imply that variability in talkers can contribute to acquiring nonnative contrasting phonemes. Results are considered in relation to their implications for understanding the learning process of L2 novel phoneme contrasts and their lexical processing. |
