現在位置 首頁>檢索服務>詳目式

全文限國圖網域開放

博碩士論文: 客語兒向言語的字調聲學研究=Acoustic Analysis of the Lexical Tones in Hakka Infant-Directed Speech

  • 作者:張國志(Kuo-chih Chang)(研究生)
  • 其他作者:鄭明中(Ming-chung Cheng)(指導教授)
  • 語文:英文
  • 出版者:國立聯合大學
  • 系所名稱:國立聯合大學
  • 學位類別:碩士
  • 出版日期:2010
  • 畢業學年度:98
  • 頁數:98
    • 主題:語言-其他
    • 關鍵詞:聲調、清晰言語、聲調空間、喉塞音、入聲、兒向言語、tone、clear speech、tone space、glottal stop、checked tone、Infant-Directed Speech
    • 國家:臺灣

    摘要:

    「兒向言語」(Infant-Directed Speech)是成人面對小嬰兒時,自然會說出來的特殊說話方式。「兒向言語」不同於「成人對話」(Adult-Directed Speech),其語音聲學特質有:提升的音高、大幅度的音高變化、誇張的語調等。儘管客語的聲調聲學研究逐漸滙集能量,但其「兒向言語」的聲調分析卻是亟待耕耘的。文獻上針對國語(Liu, Kuhl & Tsao, 2007)和粵語(Xu, 2008)「兒向言語」的聲調聲學研究,發現「兒向言語」整體聲調往上提升,並伴隨聲調的高度發音(hyperarticulation)。值得一提的是,以上研究均未觸及漢語方言中重要的聲調特點,即「入聲」(checked tone)所可能產生的變化。此外,MIPhI Model (Sundberg, 1988)指出,「兒向言語」超音段層次的高度發音,會在嬰孩12個月之後呈現減緩的趨勢。因此,本研究不但是客語「兒向言語」的先聲,更能填補以往「兒向言語」研究在「入聲」方面的不足,並檢視MIPhI Model在聲調語言的適用性。根據張月琴(1995)的聲學研究,四縣客語有六個聲調,分別是陰平(高升)、陽平(低降)、上聲(高降)、去聲(高平)、陰入(高降)及陽入(高平)。配合[i]、[a]、[u]三個元音,本研究共設計了18個雙音節詞,目標字均置於前字,以避免變調產生。藉由Praat語音軟體,本文將16位客家母親的「兒向言語」(嬰兒6~30個月)及「成人對話」予以分析,試著去檢視媽媽對嬰兒說話時,會如何改變聲調的聲學參數-基頻曲線(F0 contour)、基頻均值(F0 mean)、基頻範圍(F0 range)以及基頻斜率(F0 slope)?結果發現,客語「兒向言語」的語音聲學特徵有:聲調提升,聲調音域變寬,聲調升降急遽、聲調曲線明顯等效果。至於「入聲」的變化,因「陰入」與「去聲」同屬「高平」調,在「成人對話」裏,其聲學描繪體現為二條幾乎重疊、帶有微降走勢的基頻曲線;然而「兒向言語」的「陰入」曲線卻分布在較高頻的區域,與「去聲」之間有著明顯的聲學間隔。我們認為,客語「兒向言語」高平調的舒入對比,在基頻相關參數上的變異,肇因於入聲音節結構中的「塞音尾」。在世界眾多的語言裏,「塞音尾」不僅和「喉塞音」(glottal stop)的出現有所關聯,且常為後者所取代(Ladefoged & Maddieson, 1996)。「喉塞音」的發聲特徵是:聲門突然地緊縮(Kingston, 2005; Tang, 2008)。由於聲帶緊張度增強,而產生高調作為「喉塞音」的伴隨現象(Ohala, 1973; Zhu, 2004)。其次,本文試著從超語言(paralinguistic)及語言的角度,去詮釋客語「兒向言語」所做的聲學改變。「維持注意力」及「促進情感溝通」是其主要的超語言功能。語言功能方面,除了感知(perceptual)上,聲調訊號的突顯(salient)之外,「兒向言語」聲調空間(tone space)的擴張,使得整體音位(phoneme)格局間的聲學距離拉大,進一步暗示「兒向言語」可能是一種清晰言語(clear speech),讓嬰兒更容易區分不同的聲調。最後,本研究聲調高度發音的結果,和文獻上聲調習得的年齡相呼應;同時意味著,MIPhI Model在超音段層次上的假設,需考量到印歐語言之外的聲調語言來予以調整。
    Infant-Directed Speech (IDS) is the special way of speaking that is used when caretakers address infants. Compared with Adult-Directed Speech (ADS), IDS has a unique acoustic signature characterized by a higher fundamental frequency (F0), wider F0 range, and exaggerated F0 contours. Beyond these prosodic patterns, studies show that the phonetic units in IDS are hyperarticulated, making the individual sounds of language more distinct from one another. In Sino-Tibetan languages, tone plays a key role in distinguishing the meanings of words and offers essential prosodic information for speech perception. Albeit a small number of IDS studies have explored acoustic modifications in tone languages, research on Hakka IDS is unheard of. In addition, none of previous IDS studies have ever probed one of the most important tonal features of Chinese dialects, namely checked tones. Plus, based on MIPhI Model (Sundberg, 1988), hyperarticulation of IDS on suprasegmental level has a tendency to dwindle after the infant age of 12 months. Hence, this study is the herald of Hakka IDS studies. It can also complement the IDS research on checked tones and examine the applicability of MIPhI Model to tone languages. Based on Chang’s study (1995), Si-xian Hakka has six phonologically distinctive tones: Tone 1 (high-rising), Tone 2 (low-falling), Tone 3 (high-falling), Tone 4 (high-level), Tone 5 (high-falling) and Tone 6 (high-level). 16 mother-infant dyads, with infants 6-30 months of age, joined the recordings. The test stimuli contained 18 disyllable words (6 tones × 3 vowels), with the target lexical tones located in the first syllables. PRAAT was employed to fetch quantitative information of the lexical tones of Hakka on four acoustic cues: fundamental frequency (F0) contour, F0 mean, F0 range and F0 slope. Our results suggested that IDS contributed to elevated F0 mean, expanded F0 range, steeper F0 slope, and distinct F0 contour realization. Paralinguistically, the adjustments mentioned could be linked to emotion-attentional functions of IDS. To be noted, the effect of positive affect in IDS was so prevalent that a rising tendency could be discerned along the contours of Tone 4 and Tone 6. Linguistically, Hakka mothers enhanced the acoustic salience of IDS signal by expanding the F0-related measures. F0 range expansion is indicative of more explicit phonetic realization of Hakka lexical tones. The steeper F0 slopes of rise and fall in IDS were used to aid infants’ tonal identification. As a result, the F0 contours in IDS displayed perceptually more distinct patterns than ADS. Finally, tone hyperarticulation was observed in Hakka IDS. By producing more peripheral tones, Hakka mothers enlarged the acoustic distance between the contrastive tonal categories. This provided infants with perceptually more intelligible and acoustically more distinct speech input and might facilitate the development of phonetic perception in infancy. Of all, the scarce acoustic differences between Tones 4 and 6 in ADS were augmented in IDS due to glottalization of coda stops of high level checked tones. As a result, Tone 6 in IDS disclosed even more pronounced F0 mean, F0 slopes and F0 contours than Tone 4, which is posited to aid infants’ tonal discrimination. In terms of tone hyperarticulation, our results of Hakka IDS fail to show age-related change between 6-18 month olds and 19-30 month olds. This invites adjustments to the MIPhI Model and converges with the findings in literature that mastery for the production of tones was not achieved before age 30 months.

    目錄:

    CHAPTER ONE: INTRODUCTION 1
    1.0 INTRODUCTION 1
    1.1 BACKGROUND AND MOTIVATION 1
    1.2 PURPOSES OF THE STUDY 2
    1.3 RESEARCH QUESTIONS 2
    1.4 SIGNIFICANCE OF THE STUDY 3
    1.5 LIMITATIONS OF THE STUDY 3
    CHAPTER TWO: LITERATURE REVIEW 5
    2.0 INTRODUCTION 5
    2.1 TONE 5
    2.1.1 Tone Languages 6
    2.1.2 The Physiological Mechanism of Tone Production 8
    2.1.3 The Acoustic Correlates of Tones 9
    2.1.3.1 F0 Contour 9
    2.1.3.2 F0 Mean (F0 Height) 10
    2.1.3.3 F0 Slope 11
    2.1.4 The Representation of Tone 11
    2.2 HAKKA 12
    2.2.1 The Syllable Structure of Hakka 13
    2.2.2 The Segments of Hakka 13
    2.2.3 The Rimes of Hakka 14
    2.2.4 The Tonal System of Hakka 15
    2.3 PREVIOUS RELATED ACOUSTIC STUDIES ON INFNAT-DIRECTED SPEECH 16
    2.3.1 Prosodic Patterns 16
    2.3.2 Phonetic Hyperarticulation 18
    2.3.3 Summary 22
    CHAPTER THREE: METHODOLOGY 24
    3.0 INTRODUCTION 24
    3.1 SUBJECTS 24
    3.2 TEST STIMULI 25
    3.3 PROCEDURES 26
    3.4 ACOUSTIC ANALYSIS 27
    3.5 ACOUSTICAL MEASUREMENTS 28
    3.6 STATISTICAL ANALYSIS 30
    3.7 THE GRAPH DRAWING SOFTWARE: MICROCAL ORIGIN 30
    CHAPTER FOUR: RESULTS AND DISCUSSION 31
    4.0 INTRODUCTION 31
    4.1 COMPARISON OF F0-BASED PARAMETERS BETWEEN ADULT-DIRECTED SPEECH AND INFANT-DIRECTED SPEECH 32
    4.1.1 Graphical Comparison of F0 Contours 32
    4.1.2 F0 Mean 36
    4.1.2.1 The F0 Elevation 36
    4.1.2.2 The Sequences of F0 Mean 37
    4.1.3 F0 Range 40
    4.1.4 F0 Slope 41
    4.1.5 Summary 44
    4.2 EXAMINATION OF TONAL HYPERARTICULATION AND TONAL DIFFERENCE AUGMENTATION IN INFANT-DIRECTED SPEECH 46
    4.2.1 Hyperarticulation via Comparison of Tone Space between Adult-Directed Speech and Infant-Directed Speech 46
    4.2.2 Tonal Distance among Tones in Adult-Directed Speech and Infant-Directed Speech 48
    4.2.3 Summary 49
    4.3 GENERAL DISCUSSION 49
    4.3.1 Paralinguistic Roles of Infant-Directed Speech 49
    4.3.2 Possible Linguistic Roles of Infant-Directed Speech 53
    4.3.3 Summary 60
    CHAPTER FIVE: CONCLUSION 63
    5.0 INTRODUCTION 63
    5.1 MAJOR FINDINGS 63
    5.2 CONTRIBUTIONS OF THE STUDY 64
    5.3 SUGGESTIONS FOR FURTHER STUDIES 65
    REFERENCE 66
    APPENDIX A 79
    APPENDIX B 95
    APPENDIX C 96
    APPENDIX D 97