The Effect of Mandarin Tone on the Perception and Production of English Consonant Clusters

The present study investigates the effect of language-specific features of L1 on the acquisition of L2. It is intended to show how English initial clusters are produced and perceived by Chinese advanced learners whose language does not allow clusters. Previously we found that Chinese advanced learners would simplify the syllable structure of the English clusters: Cantonese, a southern dialect, speakers delete or substitute one consonant whereas Mandarin, a northern dialect, speakers add a vowel [1]. The present paper is to further analyze the acoustic properties of the inserted vowel in the productions of Mandarin speakers and compare them with those of a schwa. In addition, the paper aims to show the perceptual difference between native English speakers and Mandarin speakers. In the production study, minimal pairs contrasting an r-cluster and a schwa, such as /prit/ and /pǩrit/, were collected from five advanced learners of English. Vowel part of the stimuli was extracted and analyzed for formants (F1, F2), pitch, and duration. The inserted vowel part in the clusters was significantly different from a schwa (p<.0001): it was shorter in duration and lower in pitch compared to the latter. Findings suggest that the Chinese learners of English use suprasegmental features to differentiate English syllable structure. The same production data were used in the perception study. Categorical perception tests under three conditions, namely, raw data, normalized duration, and normalized pitch, were given to each of the five native speakers of English and Mandarin Chinese to see whether they perceive the clusters with the inserted vowel differently. Results show a noticeable contrast between the language groups. While the English speakers failed to distinguish the clusters with the inserted vowel from the words with a schwa, the Chinese speakers were able to identify and discriminate the clusters with the high accuracy rate. The accuracy rate by the Chinese speakers, however, dropped drastically from 73-88% to below 50% when pitch was normalized to a level tone. Manipulations had no effect on the perceptions of the vowels by the native English speakers. Based on the experiments, the study suggests that the Mandarin speakers, even at the advanced level, rely on the L1 syllable structure and accommodate the English clusters to the Chinese syllable structure by adding a vowel with a low tone. This could indicate that the tone in Mandarin plays a large role as it is used to add functional meanings in the L2 acquisition. The study also supports the attention-based learning that L2 learning is kin to the salient features of L1.