Statistical Regularities Affect the Perception of Second Language Speech: Evidence From Adult Classroom Learners of Mandarin Chinese

Abstract

This study investigated how adult second language (L2) learners of Mandarin Chinese use knowledge of phonological and lexical statistical regularities when acoustic information is insufficient for word recognition. A gating task was used to test intermediate L2 learners at two time points across a semester of classroom learning. Native Mandarin speakers (tested once) served as a control group. Mixed-effects modeling revealed that upon hearing truncated speech, L2 learners, like native speakers, identified high token frequency syllable-tone combinations more accurately than low token frequency syllable-tone combinations. Error analysis of correct syllable/incorrect tone responses revealed that native speakers made specific probability-based errors. L2 learners primarily demonstrated more acoustic-based errors but exhibited a trend toward greater probability-based errors during the second test. These findings are interpreted in light of L2 speech learning models that emphasize a statistical learning mechanism.

Do Explicit Instruction and High Variability Phonetic Training Improve Nonnative Speakers’ Mandarin Tone Productions?

“…Unfortunately for nonnative learners, our relatively poor Tone 2 and Tone 3 results are in line with previous lab‐based and classroom‐based L2 Mandarin acquisition research (e.g., Chen et al., ; Everson & Shen, ; Hao, ; Wiener, ; Wiener, Lee, & Tao, ; Yang, , ). Even the L2 Mandarin speakers tested in the present study with over 140 hours of classroom experience still produced Tone 2 and Tone 3 with less than 70% accuracy.…”

“…In contrast, when high variability input was combined with non‐explicit instruction (i.e., learners were unaware of what acoustic cues to attend to and the cues were highly variable), participants produced the proportionally least accurate production for Tone 2, Tone 3, and Tone 4. An explicit instruction approach combined with an adaptive and engaging training paradigm (e.g., Shih et al., ; Wiener, Murphy, et al., ) could therefore initially deliver low variability, single‐speaker input followed by a gradual introduction of high variability, multispeaker speech involving Tone 1 and Tone 4 exemplars before introducing Tone 2 and Tone 3 exemplars. Future research may explore how steadily introducing more variable phonetic input over time affects L2 learners’ productions—including productions in running speech (see Tseng, ; Tseng et al., )—and to what degree tone productions improve given multisyllabic training (Chang & Bowles, ).…”

“…Given the challenges associated with L2 tone acquisition, there is increasing interest in exploring what potential pedagogical strategies may result in more native‐like performance (Duff & Li, ). Most recently, researchers in this area have examined the use of adaptive training systems (Shih et al., ), web‐based training platforms (Godfroid, Lin, & Ryu, ), visuospatial gestures (Morett & Chang, ), and incidental learning videogames (Wiener, Murphy, et al, ), among other approaches. We contribute to this ongoing discussion by investigating how explicit instruction of tone contours and high variability phonetic training affect the production of L2 Mandarin tones.…”

Multi-Talker Speech Promotes Greater Knowledge-Based Spoken Mandarin Word Recognition in First and Second Language Listeners

“…Mandarin provides clear distributions of speech sounds, which are traced and exploited by native and non-native listeners (e.g., Fox and Unkefer, 1985;Wang, 1998;Wiener and Ito, 2015;Wiener and Turnbull, 2016;Wiener et al, 2019). For the majority of speech tokens, there exists a relatively straightforward mapping from the syllable, to the morpheme, to the word, to the written character without the need for intervening derivational morphology Marslen-Wilson, 1994, 1995;Packard, 1999Packard, , 2000Myers, 2006Myers, , 2010; see also Tao, 2015 for corpus evidence).…”

“…To explore whether L2 learners also utilize distributional knowledge of tonal probabilities in Mandarin word recognition and how that knowledge develops over time, Wiener et al (2019) extended Wiener and Ito (2016) by testing L1 English-L2 Mandarin adult classroom learners at two time points roughly 3 months apart (along with a control group of L1 listeners tested once). New test items that were more appropriate for the L2 learners were used.…”