台灣華語聲調習得

Abstract

本篇論文是針對六位以台灣華語為母語的嬰幼兒，採長期觀察的方式，研究華語聲調的習得，並詳細描述單音節詞和雙音節詞之中聲調出現順序、頻率、正確率、以及代換模式。本研究同時要用Yip (2002)的標記理論來檢驗各種不同聲調語言中的共通性。\n 本研究一共觀察了有六位年齡在十個月至一歲一個月的嬰幼兒長達八個月。以兩個禮拜一次的頻率收集嬰幼兒和母親或照顧者之間的自然對話。並利用錄製回來的高規格影音檔做譯寫和分析。\n 結果顯示[55]最早出現，也是頻率最高、正確率最高的聲調。而[51]在聲調出現順序、頻率、及正確率都是排在第二。[35]和[21]就比較晚出現，和前面兩個聲調相比，頻率及正確率都較低。輕聲不管是出現順序、頻率、或正確率都排在最後。\n 本研究結果還發現台灣華語中有一特殊的聲調組合[21-35]。在台灣華語中，媽媽對幼兒說話時所使用的「媽媽語」很常把這個聲調組合套用在重疊詞中。而這個聲調組合也被幼兒高度模仿使用。因此本研究認為幼兒發出的這個高頻的聲調組合[21-35]有可能是受到照顧者的影響，並且也認為幼兒並非分開習得此兩個聲調，而是把此兩聲調當作一個整體來習得。\n 最後，將所有跨語言的分析結果拿來檢驗Yip (2002)的標記理論後，我們發現幾乎所有語言都支持平調比曲折調早習得、降調比升調早習得。但是除了泰文之外，沒有語言能支持Yip (2002)提出的低調比高調早習得。因此，語言習得的證據能證明平調、降調、高調比曲折調、升調、低調還普遍。

The purpose of this study is to describe children’s tonal development by analyzing the tone emergence, frequency, accuracy rate, and substitution pattern, based on observing monosyllabic and disyllabic utterances in six Mandarin-speaking children in Taipei, Taiwan. This study also aims to examine several cross-linguistic data the theory of markedness presented by Yip (2002). \nSix subjects are investigated with the age range from 0;10 to 1;6. The data collection is conducted fortnightly by the author and the research team. Based on video and sound files, a set of coding are employed for data analysis. \nThe results showed that the high-level tone [55] emerged the first, and it also ranked as the most frequent and stable tone. Falling tones [51] were consistently ranked in the second place within tone emergence, frequency, and accuracy rate. Rising tones [35] and low-level tones [21] appeared late, and were also less frequent and stabilized later than [55] and [51]. The neutral tone was emerged and stabilized the last appeared and last acquired tone. \nThis study also found the dominated tone combination [21-35] applied particularly in the reduplications of motherese in Taiwan Mandarin. The tone combination [21-35] was proposed to be influenced by motherese, and was acquired as a prosodic whole. \nThe results of this study and all the cross-linguistic data are examined in Yip’s theory of markedness. The first two constraints obtained more evidence that the features of level and falling in tones were suggested to be the unmarked features in tonal languages. Regarding the third constraint, because most of the tone acquisition data indicated that high tones were acquired earlier than low tones, the more unmarked tone feature should be level, falling, and high.