訓練非聲調語言學習者對國語雙音節詞的聲調之聽辨與發聲

Abstract

本研究針對第二語言聲調習得在理論、方法和教學上的議題做探討，透過三種訓練方法，探索非聲調語言初學者學習國語的四個聲調之聽辨與發聲的表現。本研究在探討聲調的可訓練性，特別做了兩次後測，以瞭解訓練成效是否可擴及至新的刺激字(由新的說話者發聲)。十五位在台灣學中文的非聲調語言學習者參與為期兩週的課外訓練課程。受試者可自由選擇訓練組別，過程中需完成四選一的辨識測驗。第一組（A組，人數 = 5）接受聽辨訓練，僅獲得即時的聽覺反饋；第二組（AM組，人數 = 5）也接受聽辨訓練，獲得即時的聽覺和語意反饋(相對應的圖片和英文釋義)；第三組（AV組，人數 = 5）接受聽辨和發聲訓練，並獲得即時的聽覺和視覺反饋(聲調的基頻曲線)，受試者得以比對自己和範例曲線的異同。三組訓練採用相同的刺激字(國語雙音節詞)。聲調訓練為期二~三週，共做三次測驗(前測，第一次後測和第二次後測)，前測和第一次後測內容完全相同，第二次後測採用新的刺激字(由新的說話者發聲)。組間的比較分析顯示，相對於無接受訓練的對照組，有接受訓練的三個組別在國語聲調的聽辨和發聲上均有顯著的進步，結果也同時顯示電腦輔助聲調學習可作為課外訓練課程。進一步探究各組別的表現，則發現僅接受聽辨訓練的受試者(即A和AM組)比同時接受聽辨和發聲訓練的受試者(AV組)有更顯著的進步，換言之，聽辨訓練較能提高第二語言聲調習得的成效。本研究進一步對各聲調正確聽辨和發聲的平均百分比，以及聲調偏誤做一比較分析。最後，根據實驗結果本研究建議：僅接受聽辨訓練足以改善學習者的國語聲調之聽辨與發聲。This thesis aims to address the theoretical, methodological, and pedagogical issues of tonal acquisition in a second language (L2). The present study investigated the effects of three training approaches on the perceptual and production learning of the four Mandarin lexical tones by groups of non-tonal beginning learners. The experiment employed a pretest-posttest paradigm. Fifteen non-tonal learners of Mandarin Chinese in Taiwan received two weeks of training as extracurricular activities. Based on learners’ choices, one group (the A Group, n=5) received perceptual training only with auditory feedback involving four-way forced choice identification tasks with immediate feedback. A second group (the AM Group, n=5) received perceptual training with auditory and meaning-bearing feedback (i.e., corresponding pictures and English equivalents of the stimuli) involving the same identification tasks during training. A third group (the AV Group, n=5) received perceptual and production training with auditory and visual feedback showing pitch contours with which trainees can compare their own productions. The same training stimuli were used in the three training approaches. Following training, a posttest and a generalization test were administered immediately. Pretest, post-test, and generalization test data in perception and production were collected from the three groups and were compared for effectiveness of the three training procedures. Percent correct scores, perceptual sensitivities and production accuracy to each tone, and tonal confusions were also analyzed. The results at post-test showed that the three training groups improved significantly in perceptual accuracy of Mandarin tones as compared with a control group (the C Group, n=6) and perceptual learning also generalized to new stimuli by a new speaker. The three training groups’ production accuracy of Mandarin tones also improved significantly at posttest. More importantly, trainees who received the auditory-only feedback (i.e., the A and AM groups) showed a greater perceptual and productive improvement in identifying Mandarin lexical tones than those who were trained with the audio-visual feedback (i.e., the AV group). The results indicated that the three training approaches are effective and laboratory based training techniques can be implemented in extracurricular activities. These findings imply that the A and AM training approaches employed in the current study facilitate the learning of Mandarin tones and promote tonal modification of listeners’ tonal properties of L2 tones. It is also suggested that training only in perception with auditory-only feedback is sufficient for improvement in both perception and production of Mandarin tones.