The Application of the Speech Learning Model to the L2 Acquisition of Mandarin Tones

One major hypothesis of the Speech Learning Model (SLM) is that if the sounds in the second language (L2) are dissimilar to the learners’ first language (L1), learners are more likely to form new phonetic categories for these sounds after receiving sufficient L2 input and achieve native-like proficiency. As for the L2 sounds that are similar to the L1, learners often equate them with L1 categories and their perception and production may always deviate from native speakers. To apply the SLM to the learning of tones, this study investigates the way in which the perceived L2-L1 similarity affects the acquisition of Mandarin tones. Three groups of English speakers with no, little, and extensive Mandarin experience judge the English-likeness of the four Mandarin tones. Their ratings are used to make predictions about the L2 acquisition of tones based on the SLM. Then the experienced and inexperienced learners’ tone acquisition is evaluated by a perceptual identification task. The results show that the experienced learners obtain significantly higher accuracy than the inexperienced learners on tones perceived to be less similar to English, but not on the tone perceived to be more similar, generally confirming the SLM predictions. The three groups of English speakers differ to some extent in their L2-L1 similarity ratings, yet the predictions made from the experienced learners’ judgment account for the L2 acquisition outcome more precisely. This study concludes that the SLM can be applied to the learning of L2 tones, and the advanced L2 learners can more critically assess the L2-L1 similarity than those who have little or no knowledge of the L2.