Frequency-influenced Choice of L2 Sound Realization and Perception: Evidence from Two Chinese Dialects

Frequency-influenced choice of L2 sound realization and perception: evidence from two Chinese dialects Yizhou Lan Department of Asian and International Studies City University of Hong Kong eejoe.lan@gmail.com Abstract The study of second language speech perception usually put L1-L2 phonological mapping as the rule of thumb in predicting learning outcome, and seldom included more fine-grained aspects such as frequency. This study examines how frequency of sounds in L1 may influence L2 segmental production and perception, with examples from English learners native to two Chinese dialects, Cantonese and Sichuanese. Although these two dialects (L1s) have very similar phonological inventory, they produce certain L2 sounds in drastic difference. Productions of English voiceless interdental fricative and central liquid in the onset position were obtained in free speech from the two dialects’ speakers in vast phonological environments. Then, perception tests, including AX and oddity tasks, were done for these two groups of speakers as well. Results showed that the two English sounds were respectively realized as different sounds in Cantonese and Sichuanese L1, which was reflected by both production and perception data. Findings suggest that L2 category formation is frequency-motivated instead of markedness-motivated, and is significantly influenced by the functional load of L1 sound input. Findings further imply that a quantitative and frequency-sensitive learning model is more suitable for L2 sound acquisition. 1 Introduction Second language speech has generally seen as function of linguistic experience. However, how experience shape the formation of phonetic category was understudied. This study addresses a case when speakers from two L1s with similar segmental layout may have different realizations of L2 categories. Although theoretic models in speech learning such were very rich in literature, such as Perceptual Assimilation Model (PAM [1]) and its another version for L2 learners (PAM-L2 [2]) as well as Speech Learning model (SLM, [3]) had addressed different L1 assimilation patterns in learning multiple L2s, few studies had found similar multiple L1s yielding different L2 learning outcomes. PAM and SLM suggest that second language learners will either assimilate the L2 sound categories (or sequence of sounds) to L1 sound categories according to different perceptual distances. Increased exposure to L2 will thus trigger distributive learning of L2 input by forming a new intermediate category between the L1 and L2 in the learner’s common phonetic space [1]. In experience-based models, the positive effect of L2 exposure will increase the chance of distributive learning because the learnability of certain L2 categories should become stable if the input of L2 categories occurs in environments with similar frequency [3]. This paper displays that similar L1 inventories may result in different learning outcomes and argues that this phenomenon is influenced by frequency in similar ways as the native language was (NLM, [5]). The two English sounds under current investigation are the voiceless interdental fricative (/θ/) and the central liquid (/r/). In a pilot study, it was found that Sichuanse speakers replace English /θ/ by /s/ but Cantonese speakers by /f/. PACLIC 28 292 Copyright 2014 by Yizhou Lan 28th Pacific Asia Conference on …

Content truncated to 500 words