海陸腔客語母音的聲學研究 - 政大學術集成

全文

(1)國立政治大學語言學研究所碩士論文 National Chengchi University Graduate Institute of Linguistics Master Thesis. 治政指導教授:萬依萍博士大立Advisor: Dr. I-Ping Wan. er. io. al. sit. y. ‧. ‧ 國. 學. Nat. 海陸腔客語母音的聲學研究. n. An Acoustic Study of Hai-lu Hakka Vowels. Ch. engchi. i Un. 研究生: 李晉瑋撰 Student: Chin-wei Li 中華民國 101 年 7 月 July, 2012. i. v.

(2) An Acoustic Study of Hai-lu Hakka Vowels. 立. 政 By 治大. ‧ 國. 學 Chin-wei Li. ‧. n. er. io. sit. y. Nat. al. Ch. engchi. i Un. A Thesis Submitted to the Graduate Institute of Linguistics in Partial Fulfillment of the Requirements for the Degree of Master of Arts. July 2012 ii. v.

(3) 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. C 2012 Copyright ○. Chin-wei Li All Rights Reserved. iii. i Un. v.

(4) ACKNOWLEDGEMENTS. 本研究及學位論文是在我的指導老師萬依萍教授的親切關懷和悉心指導下完成的。她嚴謹的科學態度，精益求精的工作作風，深深地感染和激勵著我。從課題的選擇到項目的最終完成，萬老師都始終給予我細心的指導和不懈的支持。三年來，萬教授不僅在學業上給我以精心指導，同時還在思想、生活上給我以無微不至的關懷，在此謹向萬老師致以誠摯的謝意和崇高的敬意。政治大學語言學研究所扎實的教學內容，讓學生在研究所生涯收穫頗豐，在此感謝任課教授們在課程期間的傾囊相授。此外，也感謝惠鈴助教在行政事務上. 治政的幫助，協助我調適這段期間的壓力。大立再者，論文的完成，還要感激我的論文口試委員，王旭老師和林祐瑜老師， ‧ 國. 學. 感謝您們願意親臨政大為我成就這個學位，對論文的內容更是不吝給予學生許多. ‧. 專業的意見，學生十分感激。. sit. y. Nat. 在此，我還要感謝在一起愉快的度過研究生生活的政大語言學研究所所有同. io. er. 學、學長姐、學弟妹，正是由於你們的幫助和支持，我才能克服一個一個的困難和疑惑，直至本文的順利完成。特別感謝我的實驗室同門學長裕台、學姐采君，. al. n. iv n C 實驗室同門學妹涵絜，在論文的過程中，給予我不少的幫助和打氣。 hengchi U. 在論文即將完成之際，我的心情無法平靜，從開始進入課題到論文的順利完成，有多少可敬的師長、同學、朋友給了我無數的幫助，在這裡請接受我誠摯的謝意!最後，我還要感謝一路陪伴我的男友，還有培養我長大含辛茹苦的父母，謝謝你們!. 晉瑋謹致 2012/07/23 iv.

(5) TABLE OF CONTENTS Acknowledgements ....................................................................................................... iv Table of Contents ........................................................................................................... v List of Tables ................................................................................................................ ix List of Figures ................................................................................................................ x Chinese Abstract ........................................................................................................... xi English Abstract ........................................................................................................... xii CHAPTER I. INTRODUCTION ............................................................................... 1 1.0. Introduction ....................................................................................................... 1. 1.1 Hakka Language ................................................................................................ 2. 政治大. 1.2 Hakka Studies .................................................................................................... 3. 立. 1.3 Purpose of the thesis .......................................................................................... 5. ‧ 國. 學. 1.4 Research Questions ........................................................................................... 6 CHAPTER II. LITERATURE REVIEW ................................................................ 10. ‧. 2.0 Introduction to vowel transcription .................................................................... 10. io. sit. y. Nat. 2.1 Hakka Phonological System .............................................................................. 13 2.1.1 Hakka Syllable Structure ............................................................................. 13. n. al. er. 2.1.2 Hakka Consonants ....................................................................................... 14. i Un. v. 2.1.3 Hakka Vowels ............................................................................................. 14. Ch. engchi. 2.1.4 Hakka Tone System .................................................................................... 15 2.2 Literature on the Studies of Hai-lu Hakka Vowels ............................................ 16 2.2.1 The vowel system of Yang (1957) .............................................................. 16 2.2.1.1 Underlying Vowels ............................................................................... 16 2.2.1.2 Surface Vowels ................................................................................... 16 2.2.1.2.1 High Vowels /i/, /ɨ/, and /u/ ......................................................... 17 2.2.1.2.2 Mid Vowels /e/ and /o/ ................................................................ 18 2.2.1.2.3 Low Vowel /a/ ............................................................................. 19 2.2.1.3 Rhyme Structure ................................................................................ 19 2.2.2 The Vowel System of Lo (1990) .............................................................. 22 2.2.2.1 Underlying Vowels ............................................................................. 22 v.

(6) 2.2.2.2 Surface Vowels ................................................................................... 22 2.2.2.3 Rhyme structure.................................................................................. 23 2.2.3 The vowel system of Chen (2000) .......................................................... 25 2.2.3.1 Underlying vowels.............................................................................. 25 2.1.3.2 Surface Vowels ................................................................................... 25 2.2.3.3 Rhyme structure ................................................................................. 26 2.3 Discussion on the Three Studies .......................................................................... 27 2.3.1 Hai-lu Hakka Underlying Vowels ................................................................ 28 2.3.2 Hai-lu Hakka Surface Vowels ...................................................................... 28 2.3.3 Hai-lu Hakka Rhyme Structure ..................................................................... 29. 治政大 ...................................... 31 2.4.1 Introduction for acoustic characteristics of vowels 立 2.4.2 Acoustic studies on vowels of world’s languages ........................................ 33. 2.4 Acoustic Theory and Application on Vowels ...................................................... 30. ‧ 國. 學. 2.5 Acoustic Studies of Hakka ................................................................................... 34 2.5.1 Acoustic Studies of Hakka Spoken in Taiwan .............................................. 34. ‧. 2.5.2 Acoustic Studies of Hakka Spoken in Mainland China ................................ 37. sit. y. Nat. 2.6 Summary .............................................................................................................. 38. io. er. CHAPTER III. METHODOLOGY ......................................................................... 39 3.1 Subjects ............................................................................................................ 39. n. al. Ch. i Un. v. 3.2 Materials ............................................................................................................ 39 3.2.1 Design of testing words .............................................................................. 40. engchi. 3.2.2 Equipments ................................................................................................. 41 3.2.2.1 Stimuli .................................................................................................. 41 3.2.2.2 Recording apparatus............................................................................. 41 3.2.2.3 Acoustic analysis apparatus ................................................................. 42 3.2.2.4 Statistic software and vowel space plotting software .......................... 42 3.3 Procedures ........................................................................................................ 42 3.3.1 Recording..................................................................................................... 43 3.3.2 Acoustic measurements ............................................................................... 44 3.3.2.1 Acoustic measurement of CV ............................................................... 45 3.3.2.2 Acoustic Measurement of CVV ........................................................... 47 vi.

(7) 3.3.2.3 Acoustic measurement of CVVV ......................................................... 49 3.3.2.4 Acoustic measurement of CVC ............................................................ 52 3.3.2.5 Acoustic measurement of CVVC ......................................................... 54 CHAPTER IV. FINDINDS AND DISCUSSIONS .................................................. 57 4.1 Overview of the data ........................................................................................ 57 4.2 Acoustic analysis of monophthongs ................................................................. 58 4.2.1 Formant frequencies of monophthongs ....................................................... 59 4.2.2 Spectrogram of monophthongs .................................................................... 62 4.3 Acoustic analysis of diphthongs ....................................................................... 64 4.3.1 Formant frequencies of diphthongs ............................................................. 64 4.3.2 Spectrogram of diphthongs .......................................................................... 74. 政治大. 4.4 Acoustic analysis of tripthongs ........................................................................ 77 4.4.1 Formant frequencies of triphthongs ............................................................. 77. 立. ‧ 國. 學. 4.4.2 Spectrogram of triphthongs ......................................................................... 78 4.5 Summary .......................................................................................................... 79 CHAPTER V. CONCLUDING REMARKS ........................................................... 81. ‧. 5.1 Summary of the preliminary findings ................................................................ 81. Nat. sit. y. 5.2 Future research ................................................................................................... 82. n. al. er. io. APPENDIX I .............................................................................................................. 84 APPENDIX II ............................................................................................................. 85 APPENDIX III ........................................................................................................... 86. Ch. engchi. i Un. v. REFERNCES ............................................................................................................. 87. vii.

(8) LIST OF TABLES Table 1.1 The 22 consonants of Hai-lu Hakka .......................................................... 14 Table 1.2 The six vowels of Hai-lu Hakka .................................................................. 15 Table 1.3 The tone system of Si-xian and Hai-lu Hakka ............................................. 15 Table 1.4 Underlying vowels in Hai-lu Hakka (Yang, 1957) ...................................... 17 Table 1.5 Surface vowels in Hai-lu Hakka of studies (Yang, 1957) ........................... 17 Table 1.6 Rhyme structures in Hai-lu Hakka (Yang ,1957) ........................................ 21 Table 1.7 Underlying vowels in Hai-lu Hakka (Lo, 1990) .......................................... 22 Table 1.8 Surface vowels in Hai-lu Hakka (Lo, 1990) ................................................ 22. 治政 2000) .................................... 25 Table 1.10 Underlying vowels in Hai-lu Hakka (Chen,大立 Table 1.12 Rhyme structures in Hai-lu Hakka (Chen, 2000) ...................................... 26. Table 1.9 Rhyme structures in Hai-lu Hakka (Lo ,1990) ............................................ 24. ‧ 國. 學. Table 1.13 Allophones of single vowels ...................................................................... 28 Table 1.14 Hai-lu Hakka surface vowels ..................................................................... 29. ‧. Table 2.1 Acoustic measurement of [tʰi55] (meaning, ‘提’ ‘lift’) ............................... 45. sit. y. Nat. Table 2.2 Acoustic measurement of [kʰia55] (meaning, ‘佢’ ‘his’) ............................ 47. io. er. Table 2.3 Acoustic measurement of [tʰiau55] (meaning, ‘條’ ‘a long narrow strip’). 49 Table 2.4 Acoustic measurement of [tʰot55] (meaning, ‘脫’ ‘get off’)........................ 52. n. al. Ch. i Un. v. Table 2.5 Acoustic measurement of [tʰiap55] (meaning, ‘帖’ ‘a handwritten copy’) . 54. engchi. Table 3.1 Mean F1 and F2 values (Hz) of 6 monophthongs of Hai-lu Hakka in all subjects ......................................................................................................................... 59 Table 3.2 Mean F1 and F2 values (Hz) of 11 diphthongs of Hai-lu Hakka in all subjects ......................................................................................................................... 64 Table 3.3 Mean F1 and F2 values (Hz) of 4 triphthongs of Hai-lu Hakka in all subjects ...................................................................................................................................... 78. viii.

(9) LIST OF FIGURES Figure 1.1 Acoustic vowel space ................................................................................. 13 Figure 2.1 The spectrogram of [tʰi55] (meaning, ‘提’ ‘lift’) ....................................... 45 Figure 2.2 The spectrogram of [kʰia55] (meaning, ‘佢’ ‘his’)..................................... 47 Figure 2.3 the spectrogram of [tʰiau55] (meaning, ‘條’ ‘a long narrow strip’) .......... 49 Figure 2.4 The spectrogram of [tʰot55] (meaning, ‘脫’ ‘get off’) ................................ 52 Figure 2.5 The spectrogram of [tʰiap55] (meaning, ‘帖’ ‘a handwritten copy’) ......... 54 Figure 3.1 The distribution of the mean F1 and F2 value of six monophthongs from six subjects in the vowel space .................................................................................... 60 Figure 3.2 The mean F1 and F2 value of six monophthongs from six subjects in the vowel space .................................................................................................................. 60. 政治大 Figure 3.3 The spectrogram of CV structure [tʰi55], [kʰe55], [tsʰɨ55], [kʰa55], [tʰo55], 立 and [tʰu55] .................................................................................................................... 62. ‧ 國. 學. Figure 3.4 The spectrogram of CVC structure [tit55], [tet55], [ʒɨt55], [kat55], [tʰot55], [kut] .............................................................................................................................. 63. ‧. y. Nat. Figure 3.5 The distribution of the mean F1 and F2 value of diphthongs [ie] and [ia] from six subjects in the vowel space............................................................................ 66. sit. n. al. er. io. Figure 3.6 The mean F1 and F2 value of [i], [e], and [a] in diphthongs [ie] and [ia] from six subjects in the vowel space ............................................................................ 67. i Un. v. Figure 3.7 The distribution of the mean F1 and F2 value of diphthongs [io] and [iu] from six subjects in the vowel space............................................................................ 68. Ch. engchi. Figure 3.8 The mean F1 and F2 value of [i], [o], and [u] in diphthongs [io] and [iu] from six subjects in the vowel space ............................................................................ 68 Figure 3.9 The distribution of the mean F1 and F2 value of diphthongs [eu] from six subjects in the vowel space .......................................................................................... 69 Figure 3.10 The mean F1 and F2 value of [e], and [u] in diphthongs [eu] from six subjects in the vowel space .......................................................................................... 69 Figure 3.11 The distribution of the mean F1 and F2 value of diphthongs [ai] and [au] from six subjects in the vowel space ............................................................................ 70 Figure 3.12 The mean F1 and F2 value of [a], [i], and [u] in diphthongs [ai] and [au] from six subjects in the vowel space ............................................................................ 71. ix.

(10) Figure 3.13 The distribution of the mean F1 and F2 value of diphthongs [oi] from six subjects in the vowel space .......................................................................................... 72 Figure 3.14 The distribution of the mean F1 and F2 value of diphthongs [ui] and [ue], and [ua] from six subjects in the vowel space ............................................................. 73 Figure 3.15 The mean F1 and F2 value of [u], [i], [e],and [a] in diphthongs [ui], [ue] and [ua] from six subjects in the vowel space ............................................................. 73 Figure 3.16 The spectrogram of CVV structure [nie31], [kʰia55], [kʰio55], [kʰiu55], [tʰeu55], [pʰau55], [tʰoi55], [pʰui55], [kʰua55], and [pʰai55] ...................................... 74 Figure 3.17 The spectrogram of CVVC structure [tiet55], [tʰiap55], [kiok55], [kʰiuk55], [kuet55], and [kuat55]................................................................................. 76 Figure 3.18 The spectrogram of CVVV structure [iai55], [tʰiau55], and [kuai55]...... 79. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. x. i Un. v.

(11) Chinese Abstract 國. 立政. 治. 大. 學. 語言. 學. 研. 究. 所. 碩. 士. 論. 文. 提要. 研究所別 : 語言學研究所論文名稱 : 海陸腔客語母音的聲學研究指導教授 : 萬依萍博士研究生 : 李晉瑋論文提要內容 : (共一冊，16,041 字，分五章). 政治大本論文為以共振峰頻率資料來研究海陸腔客語母音音質之聲學研究。立語言學的文獻中，以共振峰頻率資料來研究世界上各種語言的母音的音質已 ‧. ‧ 國. 學. 經有相當豐富的文獻，在過去五十年來，前人在客語的語音及音韻系統上已經有豐富的研究成果。但以聲學資料來對海陸客語做描述的文獻相當缺乏。本研究的受試者為居住在新竹縣新埔鎮，六位以海陸腔客家話為母語的人士，其中有三名男性及三名女性。本研究使用了 32 個測試字，包含了單元音、. y. Nat. 雙合元音及三合元音。語音資料是利用 KAY CSL 4100 (KAY Electronics)來做. sit. n. al. er. io. 分析。本研究測試字包含的音節結構有:CV, CVV, CVVC, CVC, 以及 CVVV。測試字在實驗的過程中，被受試者單獨發音，或是放在句子中間。在實驗資料完成分析後，我們以 Origin 6.0 軟體繪出海陸腔客語母音的聲學空間圖。根據供振峰研究資料，本研究得到以下有關海陸腔客語母音音質的表現。首先關於單元音，海陸腔客語共有六個單元音: [i], [e], [ɨ], [a], [o], 以及 [u]。其次，海陸腔客語共有十一個雙元音: [ie], [ia], [io], [iu], [eu], [ai], [au], [oi], [ui], [ue], 以及 [ua]。我們比較單元音及雙元音的聲學空間圖 (vowel space)後，發現相同的母音在雙母音的環境下，由於受到鄰近母音發音位置的影響，在聲學空間上，比起在單母音的時候，變的更高低，或更前或後。. Ch. engchi. i Un. v. 第三，三母音[iai], [iau]和[uai]在聲譜圖上的表現，第一個母音[i]和[u]都比其後的兩個母音，長度來的相對的短一些，表現很像是滑音[j]和[w]。最後，本研究以共振峰聲學頻率的形式描述及記錄了當代海陸腔客語母音的音質，前人的研究中，以楊時逢(1952)最能說明本研究所提供的聲學資料。希望這些資料能對海陸腔客語母音的研究有所貢獻。關鍵詞 : 海陸腔客語；聲學空間圖；母音音質；聲學語音學；共振峰頻率. xi.

(12) Abstract This thesis is an acoustic study that investigates the vowel quality of Hai-lu Hakka vowels with formant frequency data. The acoustic method has been widely applied to the study of vowel quality of languages worldwide. In the past 50 years, Hakka researchers have yielded rich results in phonological system and phonetic description of Hakka, but there are relatively fewer research focus on the acoustic properties and characteristics of vowel phones in Hai-lu Hakka. The subjects of this study included three male and three female native Hai-lu Hakka speakers. Testing items used in this study were 32 syllables that involve monophthongs, diphthongs, and triphthongs. The speech data were analyzed by using KAY CSL 4100 (KAY Electronics). The data discussed in this study were testing items produced in citation form and sentence form, including the following syllable structures: CV, CVC, CVV, CVVC, and CVVV. The vowel qualities of Hai-lu Hakka vowels were measured and analyzed, and the acoustic vowel space of Hai-lu Hakka is plotted by the software Origin 6.0. Several results concerning the vowel quality of Hai-lu Hakka vowels were reported based on the formant frequency data. Firstly, there are six monophthongs in the vowel system of Hailu Hakka: [i], [e], [ɨ], [a], [o], and [u]. Secondly, there are eleven diphthongs in the vowel system of Hai-lu Hakka : [ie], [ia], [io], [iu], [eu], [ai],. 立. 政治大. ‧. ‧ 國. 學. y. Nat. sit. n. al. er. io. [au], [oi], [ui], [ue], and [ua]. By comparing the relative position of the vowel in a diphthong and the corresponding vowel as in a monophthong, we found that the second vowel in a diphthong tend to be higher or lower, more fontal or back, and it is possibly due to the coarticulatory influence of the adjacent phones.the Thirdly, as shown in the spectrogram of the three tripthongs [iai], [iau] and [uai], the duration of [i] in [iai55], [i] in [tʰiau55] and [u] in [kuai55] are relatively shorter than the rest two. Ch. engchi. i Un. v. vowels in the same syllable, as [ai] in [iai55], [au] in [tʰiau55] and [ai] in [kuai55]. In these CVVV syllables, [u] and [i] are similar to glides or so-called semi-vowels, or approximants [j] and [w] as shown in the spectrogram. Finally, the vowel system in Yang (1957) is more suitable for accounting for the data in this study. Hopefully, the vowel formant data presented in this study will contribute to the study of vowels in Hakka. Keywords: Hai-lu Hakka; vowel space; vowel quality; acoustic phonetics; formant frequency. xii.

(13) CHAPTER I INTRODUCTION 1.0 Introduction This thesis attempts to investigate the vowel quality of Hai-lu vowels from an acoustic investigation. Under this study, the dialect spoken in Hsinchu County of Taiwan will mainly be explored. The focus of this paper attempts to investigate the. 政治大. phonetic properties of all vowel phones of Hai-lu Hakka’s phonetic system.. 立. The acoustic method has been widely applied to the study of vowel quality of. ‧ 國. 學. languages worldwide. Ladefoged and Maddieson (1996) employed the acoustic. ‧. method in the sound of the world’s languages. Other surveys of vowels by acoustic. Nat. er. io. sit. y. approach in the world’s language, includes Danish, (Fischer-Jorgensen, 1972); Dutch, (Pols, Tromp & Plomp, 1973), American English (Peterson & Barney, 1952; Olive,. al. n. iv n C U Clark 1993; hHillenbrand, e n g c h iGetty,. Greenwood & Coleman,. & Wheeler, 1995;. Ladefoged, 2006), British English (Wells, 1963), Japanese (Chiba & Kajiyama, 1941), Korean (Yang, 1992; Yuen, 2001), Russian (Halle, 1959) Swedish (Fant, 1973), Beiging Mandarin (Howie, 1976; Wu 1986, Zee 2000, 2001; Zee & Lee 2001), Cantonese (Zee, 1999),Taiwanese Southern Min (Myers & Tsay 2003), Taiwan Mandarin (Liu, Zeng & Cao, 1999; Fon, Chiang & Chueng, 2004; Pan, Li, Lee, Huang & Tsou, 200), and Hakka (Liang, 2004; Huang, 2004; Deng, 2006; Zee & Lee, 1.

(14) 2008; Zee & Lee, 2009). Traditionally, in research of vowels in world’s languages, vowels are usually described with the tongue position (relative height and backness) and the roundness of lips as an articulatory aspect. However, some subtle differences in vowel quality are hard to capture with the conventional classification. Besides conventional analysis, the method of acoustic analysis provides researchers with an objective and precise. 政治大. way to describe the quality of vowels. With the help of modern acoustic analysis. 立. apparatus and by converting the physical attributes of vowels into concrete digital. ‧ 國. 學. numbers, acoustic analysis is an objective and precise way to describe the quality of. ‧. vowels. The experimental results can reveal more concrete realities of Hakka vowel. y. Nat. er. io. sit. phones based on scientific analyses.. This chapter is organized as follows. Section 1.1 introduces the Hakka language.. al. n. iv n C h e n g cstudies previous phonological h i Uon Hakka. Section 1.2 reviews. vowels, including. Si-xian Hakka and Hai-lu Hakka. Section 1.3 presents the purpose of this thesis. Research questions are shown in section 1.4. 1.1 Hakka Language Hakka, one of the major Chinese subdivisions or varieties and is spoken natively by the Hakka people in southern China and the island of Taiwan and throughout the diaspora areas of East Asia, Southeast Asia and around the world. Hakka is not 2.

(15) mutually intelligible with Mandarin, Wu, Minnan, or other branches of Chinese. It is most closely related to Gan, and is sometimes classified as a variety of Gan. Because of its original usage in scattered isolated regions where communication is limited to the local area, the Hakka language has developed different variants or dialects. The Hakka language has numerous variants or dialects, spoken in Guangdong, Fujian, Jiangxi, Guangxi, Sichuan, Hunan, Guizhou, including Hainan and Taiwan.. 政治大. In Taiwan, Hakka is the third largest speech group. The sub-dialects of Hakka in. 立. Taiwan can be classified into two main systems: Si-xian Hakka and Hai-lu Hakka. In. ‧ 國. 學. this study, we mainly focus on the Hai-lu Hakka only spoken by native speakers living. ‧. in the Hsinchu area. Hai-lu Hakka is mainly spoken in northern Taiwan, in particular,. y. sit er. io. 1.2 Hakka Studies. Nat. Hsinchu County.. al. n. iv n C h e n ghave In the past 50 years, Hakka researchers i U rich results in phonological c h yielded. system and phonetic description of dialects of Hakka, but there are relatively less research focus on the acoustic properties and characteristics of vowel phones in Hai-lu Hakka. In accounting for the phonological system of Hakka, autosegmantal phonology and Optimality Theory are adopted in Hakka linguistic research (Hsiao, 1991, 1994; Chung, 1992, 1994, 1995). Among previous linguistic studies on Hakka spoken in 3.

(16) Taiwan, most of them have focused on Si-xian dialects of Hakka. Other varieties of dialects, such as Hai-lu (Chen 2001), Chao-an (Chen, 2000), Rao-ping (Hsu, 2002, 2005), Yong-ding (Lee, 2003), Chong-lok, Da-pu, have received relatively less attention from researchers. Some attempts have so far been made at Hakka phonetic and phonological studies with acoustic or experimental approaches (Liang, 2004; Huang, 2004; Deng, 2006;. 政治大. Cheng et al., 2009; Zee & Lee, 2008, 2011). Liang (2004) conducted acoustic. 立. research to describe the acoustic properties of consonants and vowels of Si-xian. ‧ 國. 學. Hakka in his thesis. Huang (2004) investigated the physical realities of the Hakka. ‧. tones and vowels by taking acoustic and statistical approaches. Cheng et al. (2009). y. Nat. er. io. sit. explored the systematicity and variability of the vowel pattern in Sixian Hakka from an acoustic perspective. Zee & Lee (2008) conducted studies on the effect of vowel. al. n. iv n C U and frequencies, h and e nanalysis g c hofi tone. duration on formant. tone sandhi in Hakka. dialect, both based on the acoustic data from Hakka spoken in Mainland China. Zee & Lee (2011) conducted acoustic studies on the Yongding Hakka vowels. In the past 50 years, Hakka researchers have yielded rich results in phonological, syntax, morphological and semantic studies in dialects of Hakka, but there are relatively less research focuses on the acoustic properties and characteristics of vowel phones in Hai-lu Hakka. Therefore, the purpose of the present study is to fill the 4.

(17) research gap. 1.3 Purpose of the Thesis The purpose of this thesis aims to discover the acoustic characteristics of Hai-lu Hakka vowels. Many of the features required for linguistic descriptions of vowels have been established for some time. In research of vowels in world’s languages, vowels are usually described with the tongue position (relative height and backness). 政治大. and the roundness of lips as an articulatory aspect (Bell, 1867; Sweet, 1877; Jones,. 立. 1956). However, some subtle differences in vowel quality are hard to capture with the. ‧ 國. 學. conventional classification. Besides conventional analysis, the method of acoustic. ‧. analysis provides researchers with an objective and precise way to describe the quality. y. Nat. er. io. sit. of vowels. From an acoustic phonetic point of view, Steven and House (1955) and Fant (1960) point out that the most important articulatory characteristics of vowels are. n. al. the point of maximum. Ch constriction. eofnthe hi g cvocal. i Un. v. tract. With the help of modern. acoustic analysis apparatus and by converting the physical attributes of vowels into concrete digital numbers, acoustic analysis is an objective and precise way to describe the quality of vowels. The introduction to the acoustic method on vowel quality is provided in Ladefoged’s (1971, 1995, 1996, 2000, 2006) series of introductory books to phonetics. The concept of vowel quality includes two main features: height and backness that 5.

(18) contrast one vowel with another and four other minor features such as rhotacization, rounding, advanced tongue root, and nasalization (Ladefoged, 2006). The acoustic feature of vowel quality is mainly represented by the frequencies of the first two formants, the first formant (F1) and the second formant (F2). Generally, the first formant frequency decreases when the height of the vowel goes up and the second formant frequency decreases when the backness of the vowel moves backward.. 政治大. Since conventional Hakka phonological studies cannot provide a precise and. 立. objective way to describe and analyze Hai-lu Hakka vowels, we adopt the method of. ‧ 國. 學. acoustic analysis to describe the quality of vowels. This study is a phonetic analysis of. ‧. the single vowels, diphthongs and triphthongs in Hai-lu Hakka. We hope that the. y. Nat. n. al. er. io. vowel system.. sit. formant frequency data provide a basis for the transcription of the Hai-lu Hakka. 1.4 Research Questions. Ch. engchi. i Un. v. This study aims to investigate the nature of Hakka vowels by acoustic analysis. We try to answer the following research questions:. (1) How many single vowels in Hai-lu Hakka are there under study? A vowel sound whose quality doesn't change over the duration of the vowel is called a monophthong. Monophthongs are sometimes called "pure" or "stable" vowels. All languages have monophthongs and many languages have diphthongs, but 6.

(19) triphthongs or vowel sounds with even more target qualities are relatively rare cross-linguistically. Nearly all languages have at least three phonemic vowels, usually /i/, /a/, /u/ , and very few languages have fewer, though some have been argued to have just two, /ə/ and /a/. With the acoustic data, we aim to discover how many monophthongs are there in Hai-lu Hakka under study.. (2) How many diphthongs in Hai-lu Hakka are there under study? Diphthongs are. 政治大. types of vowels where two vowel sounds are connected in a continuous, gliding. 立. motion. They are often referred to as gliding vowels. Most languages have a. ‧ 國. 學. number of diphthongs, although that number varies widely, from only one or two. ‧. to fifteen or more. With the acoustic data, we aim to discover how many. y. Nat. n. er. io. al. sit. diphthongs there are in Hai-lu Hakka under study.. i Un. v. (3) What are the three vowels represented in the spectrogram? In phonetics,. Ch. engchi. a triphthong is a monosyllabic vowel combination involving a quick but smooth movement of the articulator from one vowel quality to another that passes over a third. A spectrogram is a time-varying spectral representation (forming an image) that shows how the spectral density of a signal varies with time. Also known as spectral waterfalls, sonograms, voiceprints, or voicegrams, spectrograms are used to identify phonetic sounds, to analyze the cries of animals or many other. 7.

(20) fields. With the acoustic data, we aim to discover what the three vowels are represented in the spectrogram.. (4) What position should each Hai-lu Hakka vowel be located in a vowel space? Peter Ladefoged (1993) recommended use of plots of F1 against F2 – F1 to represent vowel quality. In the fourth edition of his book (Ladefoged, 2001), he changed to adopt a simple plot of F1 against F2, and this simple plot of F1 against F2 was. 政治大. maintained for the fifth (and final) edition of the book (Ladefoged, 2006).. 立. Hayward (2000) compares the two types of plots, and she concludes that plotting. ‧ 國. 學. of F1 against F2 – F1 "is not very satisfactory because of its effect on the placing. ‧. of the central vowels", so she also recommends using a simple plot of F1 against. y. Nat. al. er. io. sit. F2. As a matter of fact, this kind of plot of F1 against F2 has been used by. n. analysts to show the quality of the vowels in a wide range of languages. In this. Ch. engchi. i Un. v. study, we adopt the simple plot of F1 against F2 as plotting the vowel space of Hai-lu Hakka.. For these objectives to be achieved, the thesis is structured as follows. Chapter two reviews the literature that is relevant to the topic of this thesis. Chapter three focuses on the methods adopted in the present study, including the details of the subjects in section 3.1. The design of the test words and equipments used in the study are shown in section 3.2. Procedures of the experiment and the acoustic measurement 8.

(21) are presented in section 3.3. Chapter four presents the findings and discussion of the results of the study. Section 4.1 provides overview of the data. Section 4.2 presents the detailed data and spectrogram of monophthongs, Section 4.3 gives the detailed data and spectrogram of diphthongs, and the detailed data and spectrogram of triphthongs are shown in 4.4. Section 4.5 is the summary of the findings. Concluding remarks are shown in chapter five.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 9. i Un. v.

(22) CHAPTER II LITERATURE REVIEW This review of literature focuses on the previous studies that are related to the topic of thesis. This review includes the literature on the history of transcription of vowel system, the phonological system of Hakka, three studies of Hai-lu Hakka vowels, and acoustic studies on Hakka.. 政治大. This chapter is divided into five main sections. Section 2.0 gives an introduction. 立. to the transcription of vowel. In section 2.1, we give a brief summary of the. ‧ 國. 學. phonological system of Hakka, including consonants, vowels, and tone. Section 2.2. ‧. provides an introduction to the phonological research on the Hai-lu Hakka vowels. y. Nat. er. io. sit. conducted by three researchers (Yang, 1992; Lo, 1990; Chen, 2000). The discussion on the three studies is provided in 2.3.. al. n. iv n C U theory and approach. The In section 2.4, an introduction ishoffered e n gforc hthei acoustic. acoustic theory and approach have been widely applied to linguistic studies of speech sounds cross-linguistically. In section 2.5, acoustic phonetic studies on dialects of Hakka spoken in Taiwan and Mainland China will be introduced. Acoustic research on dialects of Hakka, including Liang (2004), Huang (2004), Cheng et al. (2009), and Zee & Lee (2008, 2011) will be summarized. Summary of this chapter and motivation of this thesis are 10.

(23) included in section 2.6. 2.0 Introduction to vowel transcription Description of phonetics is concerned with the physical properties of speech sounds or signs (phones): their physiological production, acoustic properties, auditory perception, and neurophysiological status. On the other hand, phonology concerns itself with systems of phonemes, abstract cognitive units of speech sound or sign. 政治大. which distinguish the words of a language. A phoneme is the smallest contrastive unit. 立. in the sound system of a language. A phone is one of many possible sounds in the. ‧ 國. 學. languages of the world. Phones that belong to the same phoneme, such as [t] and [tʰ]. ‧. for English /t/, are called allophones.. y. Nat. er. io. sit. Many authors agree that vowel transcription is more difficult than consonant transcription (Ball, 1991, 1993; Butcher, 1989; Howard & Heselwood, 2002b). Vowel. al. n. iv n C identification is more difficult andhproblematic consonant identification (Cutler, i U e n g c hthan. Smits, & Cooper, 2005). Due to the constraints on perception imposed by individual phonology (Best & Tyler, 2007), there is greater difficulty with transcription of vowels than with that of consonants (Ball, 1991, 1993; Butcher, 1989; Howard & Heselwood, 2002b). The phonetic transcription of vowel sounds is based on the International Phonetic Alphabet (IPA) cardinal vowels. Cardinal vowels refer to a system of auditory and 11.

(24) articulatory standard reference points whose symbols embody information about the fundamental vowel parameters of height, fronting and rounding. The IPA system provides the tools for describing a speaker’s vowel system without reference to accent or language, and it is valuable in its ease of interpretation by anyone trained in its use (Howard & Heselwood, 2002b). Phonetic transcription is the ideal form for establishing the characteristics of a speaker’s productions.. 政治大. The IPA provides the framework for transcribing vowels, and each individual. 立. transcript should be interpreted in the context of speech patterns of the community.. ‧ 國. 學. One way to represent such community patterns is through a standard phonemic. ‧. description of the dialect, and the other way is to examine the detailed phonetic. y. Nat. er. io. sit. transcriptions of speech with reference to this phonemic foundation. Either way, they should satisfy the requirement of representation but must also be based on the. al. n. iv n C h esymbols principles of the IPA which require n g c hto ibeUselected that best describe the articulatory and auditory quality of the speakers. Acoustic analyses of speech data can empirically provide objective information about the speech signal. The interpretation of vowel acoustic analyses often relies on the correlations between the first two vocal tract resonant frequencies and the articulatory parameters as vowel height and fronting. Figure 1.1 illustrates the traditional vowel map with major IPA cardinal vowel positions indicated. The 12.

(25) traditional vowel map is plotted by the values of formant 1 (F1) and formant 2 (F2) plotted on a graph with appropriately oriented and scaled axes. Figure 1.1 Acoustic vowel space. 政治大. 立. ‧. ‧ 國. 學. Figure 1.1 is an acoustic vowel space showing how monophthongs can be. sit. y. Nat. io. n. al. er. objectively represented in the F1/F2 plane. With this graphical representation of. i Un. v. acoustic data, we can provide objective foundation for comparing the relationships. Ch. engchi. between vowel productions in actual spoken language and the abstract IPA cardinal vowel positions. 2.1 Hakka Phonological System This section describes the phonological system of Hai-lu Hakka, including syllable structure, consonants and vowels of Hakka. 2.1.1 Hakka Syllable Structure As one of Chinese languages, Hakka shares the same model of syllable structure 13.

(26) with most Chinese dialects. The structure contains four components: Initial, Medial, Nucleus and Ending. Initial refers to the first consonant (except the glides [j] and [w]) in a syllable. Medial is the prevocalic glides. Nucleus is always a vowel or sometimes a syllabic nasal. Ending of a syllable in Hakka refers to nasals [m, n, ŋ], stops [p, t, k], or glides [j] and [w] (Chung, 2004). The rhyme is comprised by medial, nucleus and ending. There are at least sixteen possible syllable shapes, V, VC, VG, VN, GV, GVG,. 政治大. GVC, GVN, CV, CGV, CVC, CVN, CVG, CGVG, CGVN, CGVC, where G is a glide. 立. and N a nasal (Chung, 2004).. ‧ 國. 學. 2.1.2 Hakka Consonants. ‧. In Si-xian Hakka, there are 18 consonants. There are 22 consonants in. y. Nat. er. io. sit. Hai-lu Hakka, by adding two palatal fricatives [ʃ] and [ʒ], and two alveolo-palatal affricates [tʃ] and [tʃʰ]. Table 1.1 shows the 22 consonants in Hai-lu Hakka.. n. al. i n C U h eHakka Table 1.1 The 22 consonants of Hai-lu ngchi. v. Bilabial Labio-dental Alveolar Alveolo-palatal Palatal Velar Glottal Plosive. p pʰ. Fricative. ʃʒ ts tsʰ. m. Lateral Glide. k kʰ. fv. Affricate Nasal. t tʰ. hɦ. tʃ tʃʰ ŋ. n l. w. j. 2.1.3 Hakka Vowels The six single vowels in Hakka sound system are [i], [e], [ɨ], [a], [o], and [u]. The high central unrounded vowel [ɨ] occurs only after the dentals [ts-], [tsʰ-], [s-] and the 14.

(27) alveolo-palatals [tʃ-], [tʃʰ-], and [ʃ-], for example, [sɨ] 字‘word’ (Chung, 2004). Table 1.2 lists the six vowels of Hai-lu Hakka. Table 1.2 The six vowels of Hai-lu Hakka Front. Central. Back. High. i. ɨ. u. Mid. e. o. Low. a. 2.1.4 Hakka Tone System. 政治大. The tone system of Hai-lu Hakka is introduced in this section. The description of. 立. tone system of Hakka follows the convention of tone letters of Chao (1930). The tone. ‧ 國. 學. letter system of Chao (1930) divides the pitch into five levels, with the lowest being. ‧. assigned the value 1, and the highest the value 5.. y. Nat. Hai-lu. Tone value 33. 53. 11. 55. 31. 13. 55. 31. al. n. Si-xian. Ch. engchi. er. io. Table 1.3 The tone system of Si-xian and Hai-lu Hakka. sit. The tone system of Si-xian and Hailu Hakka is given in Table 1.3.. i Un. v. 33 3. 3. 5. 5 Hakka preserves all of the entering tones of Middle Chinese. In Hakka literature,. an entering tone, or a checked tone, is not a tone in the phonetic sense, but rather describes a syllable that ends in a stop consonant, such as stops [p, t, k]. 15.

(28) 2.2 Literature on the Studies of Hai-lu Hakka Vowels Over the past few years, some studies have been made on the phonological system of Hakka. In the following sections, three phonological studies of Hai-lu Hakka vowels are summarized and compared. Among the three authors, there are several different hypotheses concerning the surface and underlying form of Hai-lu Hakka vowels. Yang (1992), Lo (1990) and Chen (2000) will be introduced in following sub-sections.. 立. 政治大. 2.2.1 The vowel system of Yang (1957). ‧ 國. 學. In the book of Yang (1957) The Hakka Dialect of Tao-yuan, Taiwan, Yang. ‧. introduces the phonological system and vocabulary of Si-xian and Hai-lu Hakka. y. Nat. io. sit. spoken in Tao-yuan area, Taiwan. His observation and investigation come from his. er. field study in Tao-yaun area in 1957. The vowel system of Hai-lu Hakka from Yang’s. n. al. i n C U h e n sub-sections. work will be summarized in the following gchi. v. 2.2.1.1 Underlying Vowels The underlying vowels in the phonological system of Hai-lu Hakka proposed by Yang (1957) are presented in Table 1.4. The six vowels in Hakka sound system are /i/, /e/, /ɨ/, /u/, /o/, and /a/.. 16.

(29) Table 1.4 Underlying vowels in Hai-lu Hakka (Yang, 1957) Front. Central. Back. High. i. ɨ. u. Mid. e. o. Low. a. 2.2.1.2 Surface Vowels According to Yang (1957), the surface vowels in Hai-lu Hakka are listed in Table 1.5. The ten diphthongs include [ie, ia, io, iu, eu, ua, ai, au, oi, ui] and the three. 政治大 Table 1.5 Surface vowels in Hai-lu Hakka of studies (Yang, 1957) 立 Monophthong ieɨaou triphthongs include [iai, iau, uai].. ‧ 國. Triphthong. 學. Diphthong. ie ia io iu eu ua ai au oi ui iai iau uai. ‧. 2.2.1.2.1 High Vowels /i/, /ɨ/, and /u/. y. Nat. al. er. io. sit. According to Yang, high vowels are pronounced differently in various. n. environments. The high front vowel /i/ is pronounced as the lower [ɪ] as a single. Ch. engchi. i Un. v. vowel, and the tongue position of /i/ is close to the place between [i] and [e]. In diphtohongs, it is also pronounced as a lower variant [ɪ]. Compared with the other Hakka dialect mainly spoken in Taiwan area -- Si-xian Hakka, the vowel [i] is pronounced much closer to the cardinal vowel [i]. The high central unrounded vowel /ɨ/ has two surface variants, [ɿ] and [ʒ]. When /ɨ/ is after the dentals /ts-/, /tsʰ-/, /s-/, the vowel is [ɿ]. When /ɨ/ is after the alveolo-palatals /tʃ-/, /tʃʰ-/, /ʃ-/, it becomes the fricative [ʒ]. 17.

(30) The high back rounded vowel /u/ has two surface variants: [ʊ] and [u]. The tongue position of [ʊ] is between the high rounded vowel [u] and the mid back rounded vowel [o] as in the following types of rhyme structure combination: [iu], [eu], [au], [ui], [uŋ], [ut], [uk], [iau], [iuŋ], [iut], and [iuk]. The 60 possible combinations of rhyme structure of Hai-lu Hakka from Yang (1957) is provided in section 2.2.1.3. According to Yang (1957), the tongue position of [u] is near the cardinal vowel [u] in. 政治大. the following types of rhyme structure: [un] and [iun].. 立. 2.2.1.2.2 Mid Vowels /e/ and /o/. ‧ 國. 學. Similar to high vowels, Yang thought Hakka mid vowels have surface variants in. ‧. different environments.. y. Nat. er. io. sit. The mid front unrounded vowel /e/ has two variants: [e] and [ ]. The tongue position of the vowel is close to that of a cardinal vowel [e] as in the following rhyme. al. n. iv n C U [ ] in the following types of h e nfront structures: [eu] and [en]. It is a lowered h ivowel g cmid combination: [ie], [em], [ep], [et], [iet] and [uet]. As for the mid back rounded vowel /o/, there are two variants: [ɔ] and [o]. [ɔ] appears as in the following types of rhyme structure: [oi], [on], [ion], [oŋ], and [ioŋ]. The tongue position of [o] is lower than [o] but higher than [ɔ], and it appears as in the following types of rhyme structure: [io], [ok], and [iok]. On the other hand, in Si-xian Hakka, /o/ is always pronounced as a sound [o] between [o]and [ɔ]. 18.

(31) 2.2.1.2.3 Low Vowel /a/ Similar to high vowels and mid vowels, Yang thought Hakka low vowel has surface variants in different types of combination. The low central vowel /a/ has four allophones: [a], [ä], [ ], and [ɑ]. The tongue position of the vowel is near the cardinal vowel [a] as in the following two types of rhyme structure: [an] and [uan]. [ä] is a more centralized vowel, and this variant. 政治大. appears as in the following types of rhyme structure: [ia], [ua], [ai], [iai], [uai], [am],. 立. [iam], [ap], [iap], [at], [uat], [ak], and [iak]. The vowel [ ] is lower than [e] and [ɛ],. ‧ 國. 學. and it appears as in [ian]. In rhyme structures [aŋ], [iaŋ], and [uaŋ], the position of the. ‧. tongue is close to the cardinal vowel [ɑ]. Nat. sit. y. 2.2.1.3 Rhyme Structure. n. al. er. io. As mentioned in 2.1.1, the syllable structure of Hai-lu Hakka shares the. Ch. i Un. v. same model of syllable structure with most Chinese dialects. The structure contains. engchi. four components: Initial, Medial, Nucleus and Ending. Initial refers to the first consonant (except the glides [j] and [w]) in a syllable. Medial is the prevocalic glides. Nucleus is always a vowel or sometimes a syllabic nasal. Ending of a syllable in Hakka refers to nasals [m, n, ŋ], stops [p, t, k], or glides [j] and [w] (Chung, 2004). The rhyme is comprised by medial, nucleus and ending. There are at least sixteen possible syllable shapes, V, VC, VG, VN, GV, GVG, GVC, GVN, CV, CGV, CVC,. 19.

(32) CVN, CVG, CGVG, CGVN, CGVC, where G is a glide and N a nasal (Chung, 2004). Yang (1957) lists 60 types of possible rhyme structures in Hai-lu Hakka as listed in Table 1.6. As for Si-xian Hakka, Yang lists 61 possible rhyme structures by adding [ieu]. The possible rhyme types of Hai-lu Hakka are listed in Table 1.6.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 20. i Un. v.

(33) Table 1.6 rhyme structures in Hai-lu Hakka (Yang ,1957) CV. CVV/CVC. CVVV/CVVC. i. ie. iai. e. ia. iau. ɨ. io. iam. a. iu. ian. o. im. iaŋ. u. in. iap. ip. iak. it. ion. eu. ioŋ. em. iun 治政大iuŋ iet. en. ɨp. iok. ɨt. iuk. ai. uai. au. uan. am. uaŋ. y. an. uat. aaŋ l ap. uet. io. sit. Nat. n. er. ‧ 國. iut. 學. et. ‧. 立ep. at. Ch. engchi. ak oi on oŋ ot ok ui ua un uŋ ut uk 21. i Un. v.

(34) 2.2.2 The Vowel System of Lo (1990) Lo (1990) proposed different phonetic and phonological description of Hakka spoken in Hsinchu County. The vowel system of Hai-lu Hakka in Lo’s study will be summarized in the following. 2.2.2.1 Underlying Vowels According to Lo (1990), the underlying vowels in Hai-lu Hakka are listed in. 政治大 Table 1.7 Underlying vowels in Hai-lu Hakka (Lo, 1990) 立 Front Central. Table 1.7. The six vowels in Hakka sound system are /i/, /e/, /ɨ/, /u/, /o/, and /a/.. ‧ 國. ɨ. i. Mid. 學. High. u. e. o. a. ‧. Low. Back. sit. y. Nat. 2.2.2.2 Surface Vowels. n. al. er. io. According to Lo (1990), the surface vowels in Hai-lu Hakka are listed in Table. Ch. i Un. v. 1.8. The eight diphthongs include [ie, ia, io, iu, eu, au, ua, ue] and the four triphthongs include [iai, iau, ioi, uai].. engchi. Table 1.8 Surface vowels in Hai-lu Hakka (Lo, 1990) Monophthong. ieɨauo. Diphthong. ie ia io iu eu au ua ue. Triphthong. iai iau ioi uai. Lo’s (1990) generalization and analysis on Hakka phonological system and phonetic properties are summarized as follows. Concerning the high front unrounded vowel /i/, the rhyme structure combinations /ien/ and /iet/ are in complementary distribution with /iam, iaŋ/ and /iap, iak/. In Hai-lu Hakka, the actual pronunciations 22.

(35) of /ien/ and /iet/ are [ian] and [iat]. The rhyme combination /iai/ is rare that it only occurs in colloquial conversation, thus there is no corresponding word for this vowel. The high central unrounded vowel /ɨ/ has two variants: [ɿ] and [ə]. The alveolar apical [ɿ] occurs after the dentals /ts-/, /tsʰ-/, /s-/, and the schwa [ə] occurs in [əm, ən, əp, ət]. They are in complementary distribution. Concerning the high back rounded vowel /u/, four kinds of rhyme structure. 政治大. combination, [uat], [uet], [uot], and [uok], are disappearing gradually. For example,. 立. the word [kuet] (meaning, country) is pronounced [ket] in Si-xian Hakka.. ‧ 國. 學. 2.2.2.3 Rhyme structure. ‧. As mentioned in 2.1.1, the syllable structure of Hai-lu Hakka shares the. sit. y. Nat. same model of syllable structure with most Chinese dialects. The structure contains. n. al. er. io. four components: Initial, Medial, Nucleus and Ending. Initial refers to the first. Ch. i Un. v. consonant (except the glides [j] and [w]) in a syllable. Medial is the prevocalic glides.. engchi. Nucleus is always a vowel or sometimes a syllabic nasal. Ending of a syllable in Hakka refers to nasals [m, n, ŋ], stops [p, t, k], or glides [j] and [w] (Chung, 2004). The rhyme is comprised by medial, nucleus and ending. There are at least sixteen possible syllable shapes, V, VC, VG, VN, GV, GVG, GVC, GVN, CV, CGV, CVC, CVN, CVG, CGVG, CGVN, CGVC, where G is a glide and N a nasal (Chung, 2004). According to the analysis in Lo (1990), there are 63 types of rhymes in Hai-lu. 23.

(36) Hakka. The possible rhyme types are listed in Table 1.9. Table 1.9 Rhyme structures in Hai-lu Hakka (Lo ,1990) CV. CVV/CVC. CVVV/CVVC. Syllabic nasal. i. ie. ien. m. e. ia. iet. n. ɨ. io. iau. ŋ. a. iu. iam. o. im. iaŋ. u. in. iap. ip. iak. it. ion. eu. ioŋ. iuŋ iuk. ɨt. uen. ɨm. uet. ɨn. uan. au. uaŋ. io. al. uat. n. am an aŋ ap. Ch. engchi. at ak on oŋ ot ok ue ua un uŋ ut 24. y. ɨp. ‧. iut. 學. et. Nat. ‧ 國. ep. sit. 立. en. 政 iok治大 iun. er. em. i Un. v.

(37) uk 2.2.3 The vowel system of Chen (2000) Chen (2000) provides a description of phonological system and lexicon of Hai-lu Hakka in his field work in Zhudong, Hsinchu County. His observation of vowels of Hai-lu Hakka will be introduced in the following sub-sections. 2.2.3.1 Underlying vowels The six underlying vowels of Hai-lu Hakka in Chen (2000) are presented in Table. 政治大. 1.10. The six vowels are /i/, /e/, /ɨ/, /u/, /o/, and /a/.. 立. High. Central. i. ɨ. Back u. e. Low. o. ‧. Mid. Front. 學. ‧ 國. Table 1.10 Underlying vowels in Hai-lu Hakka (Chen, 2000). a. y. Nat. er. io. sit. 2.2.3.2 Surface Vowels. According to Chen (2000), the surface vowels in Hai-lu Hakka are listed in Table. n. al. Ch. engchi. i Un. v. 1.11. The nine diphthongs include [ie, io, iu, eu, ai, au, oi, ui, ue, ua] and the three triphthongs include [iau, ioi, uai]. Table 1.11 Surface vowels in Hai-lu Hakka (Chen, 2000) Monophthong. ieɨauo. Diphthong. ie io iu eu ai au oi ui ue ua. Triphthong. iau ioi uai. As for the Hai-lu Hakka vowels, Chen makes the following observation. When the single high vowels [i] and [u] occur as the last segment in the word, they are pronounced with higher tongue position. The high central unrounded vowel /ɨ/ occurs 25.

(38) only after the dentals /ts-/, /tsʰ-/, /s-/ and the alveolo-palatals /tʃ-/, /tʃʰ-/, and /ʃ-/. Regarding the combination of diphthongs, the mid vowel [-e] only goes with the high back vowel [-u] as in [eu] and [ue]. The mid back vowel [-o] only goes with [-i], as in [io] and [oi]. 2.2.3.3 Rhyme structure As mentioned in 2.1.1, the syllable structure of Hai-lu Hakka shares the same model of syllable structure with most Chinese dialects. The structure contains. 政治大 four components: Initial, Medial, Nucleus and Ending. Initial refers to the first 立. ‧ 國. 學. consonant (except the glides [j] and [w]) in a syllable. Medial is the prevocalic glides.. ‧. Nucleus is always a vowel or sometimes a syllabic nasal. Ending of a syllable in. sit. y. Nat. Hakka refers to nasals [m, n, ŋ], stops [p, t, k], or glides [j] and [w] (Chung, 2004).. n. al. er. io. The rhyme is comprised by medial, nucleus and ending. There are at least sixteen. Ch. i Un. v. possible syllable shapes, V, VC, VG, VN, GV, GVG, GVC, GVN, CV, CGV, CVC,. engchi. CVN, CVG, CGVG, CGVN, CGVC, where G is a glide and N a nasal (Chung, 2004). In Chen’s (2000) observation and instigation of Hai-lu Hakka spoken in Jhudong, Hsinchu County, there are 59 types of rhyme structure, as listed in Table 1.12. In Chen’s observation, some rhyme combinations have few examples in Hakka vocabulary, such as [-ioi], [-iot], [-ep], [-uen], [-uat], [-uak], and [-uaŋ]. These rhyme combinations have only one or two real word examples in the lexicon.. 26.

(39) Table 1.12 Rhyme structures in Hai-lu Hakka (Chen, 2000) CVVV/CVVC. CV. CVV/CVC. i. ia. iet. e. io. ien. ɨ. iu. iau. a. im. iap. o. in. iak. u. ip. iam. it. iaŋ. eu. ioi. em. ion. am. iut. an. iuk. aŋ. uen. Nat. uai. at. uan. io. ap. al. uaŋ. n. ak oi. on. Ch. uat. e n guak chi. y. iuŋ. ‧. au. 學. ‧ 國. iun. sit. 立. ai. 政 iot治大 iok ioŋ. ep et. ŋ. er. en. Syllabic nasal. i Un. v. oŋ ot ok ui ua un uŋ ut uk 2.3 Discussion on the Three Studies Among the previous phonological studies summarized as in 2.2, three 27.

(40) researchers agree that there are six underlying vowels in Hai-lu Hakka. When it comes to the complicated rhyme structure of Hai-lu Hakka, all the three researchers provide different opinions of rhyme structure. Their analyses differ mostly in the number and the phonetic properties of surface vowels of Hai-lu Hakka. 2.3.1 Hai-lu Hakka Underlying Vowels First, concerning Hai-lu Hakka underlying vowels, all three scholars agree that. 政治大. there are six underlying vowels in the phonological system of Hai-lu Hakka. Namely,. 立. the high front unrounded vowel /i/, the mid front unrounded vowel /e/, the high mid. ‧ 國. 學. unrounded vowel /ɨ/, the low mid unrounded vowel /a/, the mid back rounded vowel. ‧. /o/, and the high back rounded vowel /u/.. y. Nat. n. al. er. io. Table 1.13. sit. Those different ideas on the allophones of the underlying vowels are provided in. Ch. Table 1.13 Allophones of single vowels Underlying vowels. engchi. Surface vowel phone variants. i Un. v. Yang (1957). Lo (1990). Chen (2000). /i/. [i] [ɪ]. [i][j]. [i]. /ɨ/. [ɿ][ʒ]. [ɿ][ə]. [ɨ]. /e/. [e][ ]. [e]. [e]. /a/. [a]. [a]. /o/. [a][ä][ ][ɑ] [o][ɔ][o]. [o]. [o]. /u/. [u] [ʊ]. [u]. [u]. 2.3.2 Hai-lu Hakka Surface Vowels Basically, they all agreed that there are monophthongs, diphthongs, and triphthongs in the vowel system of Hai-lu Hakka. However, all the three researchers 28.

(41) provide different opinions of the number and the phonetic properties of surface vowels of Hai-lu Hakka. The comparison of the three authors is offered in Table 1.14. Table 1.14 Hai-lu Hakka surface vowels Yang (1957). Lo (1990). Chen (2000). Monophthong i e ɨ a u o Diphthong. ie io iu eu ai au ua. ia oi ui. ue. oi ui ue. Triphthong. iau uai. iai. iai ioi. ioi. For Hai-lu Hakka surface vowels, the three researchers Yang (1957), Lo (1990),. 政治大. and Chen (2000) agree that there are six single vowels: [i], [e], [ɨ], [a], [u], and [o]. As. 立. for diphthongs, they agree that there are seven diphthongs, namely [ie], [io], [iu], [eu],. ‧ 國. 學. [ai], [au], and [ua]. In addition to the seven diphthongs, Yang (1957) and Chen (2000). ‧. all think that there are [oi] and [ui], Lo (1990) and Chen (2000) agree that there is [ue],. y. Nat. er. io. sit. and Yang (1957) proposes another diphthong [ia].. As for triphthongs, three researchers agree that there are two triphthongs, namely. n. al. Ch. engchi. i Un. v. [iau] and [uai] in the phonological system of Hai-lu Hakka. Yang (1957) proposes that there is still another triphthong [iai]. Unlike Yang (1957), Chen (2000) thinks that the third triphthong is [ioi]. Lo (1990) proposes that there are four triphthongs: [iai], [ioi], [iau], and [uai]. 2.3.3 Hai-lu Hakka Rhyme Structure Since the three authors make different claims about the number of surface vowels of Hai-lu Hakka, they also disagree in the number of types of possible rhyme 29.

(42) combination. According to Yang’s studies (1957), combined with nasals and stops, there are 60 types of rhymes in Hai-lu Hakka, and 61 types in Si-xian Hakka, by adding “ieu.” According to the analysis in Lo (1990), there are 63 types of rhymes in Hai-lu Hakka. In Chen’s (2000) observation of Hai-lu Hakka spoken in Jhudong, Hsinchu County, there are 59 types of rhyme structure. Chen (2000) observed that some rhyme combinations are disappearing gradually, and this claim is not mentioned. 政治大. in Yang (1957), and Lo (1990) does mention that some rhyme combinations occur. 立. 學. ‧ 國. only in colloquial vocabulary.. 2.4 Acoustic Theory and Application on Vowels. ‧. In this section, first, an introduction to acoustic theory and the basic principle of. y. Nat. er. io. sit. acoustic analysis of speech, the acoustic characteristics of vowel quality and the relationship between the vowel quality and the formant frequencies will be introduced. n. al. in 2.4.1.. Ch. engchi. i Un. v. Secondly, studies on speech sound with the acoustic method on vowels of different languages in the world will be introduced in 2.4.2. The acoustic method of measuring the first two formants as the vowel quality has been widely applied to the study of languages worldwide, and the surveys of the complete set of surface vowel in the world’s languages, including French, German, and Taiwanese Southern Min, Mandarin, and Hakka. 30.

(43) 2.4.1 Introduction for acoustic characteristics of vowels This section introduces vowel quality and the acoustic characteristics of vowels. A vowel is produced with the vibration of the vocal tract and the resonance of the oral cavity, and sometimes involves the resonance of the nasal cavity, too. A vowel is articulated without major constrictions and obstructions. The quality of a vowel depends on its overtone structures. A vowel simultaneously contains a number of. 政治大. different pitches, and these characteristic overtones are called formants of the vowels.. 立. The lowest is the first formant (F1), the second lowest the second formant (F2), and so. ‧. ‧ 國. 學. on.. Formants are the resonances of the vocal tract. The air in the vocal tract is set in. y. Nat. er. io. sit. vibration by the pulses of air from the vocal folds. Every time they open and close, the air in the vocal tract above them will be set in vibration. Since the vocal tract has a. al. n. iv n C h evibrate complex shape, the air within it will i Uthan one way. We can consider n g cin hmore. the body of air behind the raised tongue to be vibrating in one way, and the body of air in front of it to be vibrating in another (Ladefoged, 1993, 1996, 2000, 2001, 2006). There are some rules for describing the relation of vowels, formants and tongue positions. The frequency of F1 is lowered by any constriction of in the front half of the vocal tract, and the greater the constriction the more F1 is lowered. The frequency of F2 tends to be lowered by a back tongue constriction, and the greater constriction 31.

(44) the more F2 is lowered. The frequency of F2 is raised by a front constriction, and the greater the constriction is, the more F2 is raised. Vowels are largely distinguished by the first two formants. In general, F1 varies mostly with tongue height and F2 with tongue advancement. Low vowels have higher F1 frequency and high vowels have lower F1 frequency. Back vowels have a lower F2 and typically a smaller F2－F1 difference, whereas front vowels have a relatively higher F2 frequency and a larger F2. 政治大. －F1 difference. Rounded vowels have a lower F2 frequency compared to their. 立. unrounded counterparts. Higher formants, e.g., F3 and F4, vary a great deal from. ‧ 國. 學. speaker to speaker. Though they are not uniquely determined for each speaker, they. er. io. sit. y. Nat. Ladefoged, 1993).. ‧. are certainly indicative of a person’s voice quality (Bao, 1984; Kent & Read, 1992;. However, it is important that these values are offered to indicate trends but not. al. n. iv n C h e nmay absolute values, because formant values from speaker to speaker. Factors hi U g cchange such as age and sex are important because they determine vocal tract length. Generally speaking, formants of children are highest in frequency, women’s are intermediate, and formants of men’s are lowest in frequency (Peterson & Barney, 1952; Bao, 1984; Kent & Read, 1992). Other variables that affect acoustic properties. include speech style (Johnson, Flemming & Wright, 1993) and dialect (Peterson & Barney, 1952; Holden & Nearey, 1986). 32.

(45) Gender differences can also be observed from acoustic vowel spaces (Hillenbrand et al., 1995). As a tendency, male vowel quality has lower formant frequencies than its female congener. Another apparent difference is that the female vowel space is larger than the male space, that is, the female vowel qualities stake out a larger acoustic area (Childers & Wu 1991, Peterson & Barney 1952, Wu & Childers 1991). This phonetic difference between males and females may result from their. 政治大. different vocal tract length. The average length of the adult female vocal tract (the. 立. distance from the vocal folds to the lips) is on average 14–14.5 cm, and the average. 2.4.2 Acoustic studies on vowels of world’s languages. Nat. y. ‧. ‧ 國. 學. male vocal tract is 17–18 cm.. er. io. sit. Acoustic phonetics is a technical area of linguistics. It is to study the sound waves made by human vocal organs for communication. Phoneticians depict and. al. n. iv n C h eand analyze sound waves using machines n gcomputer c h i Uprograms. It is concerned with speech sounds as physical events with measurable properties such as duration and energy. In Acoustic phonetics studies, the spectrogram of speech sounds has been frequently used since 1950s. The acoustic method is a good tool in investigating language variation, especially the variation in vowel. Since vowels are not distinct as consonants are, the concrete method of acoustic analysis provides an objective solution to record and analyze the variants of vowels. 33.

(46) The acoustic method of measuring the first two formants as the vowel quality has been widely applied to the study of languages worldwide. Ladefoged and Maddieson (1996) employed the acoustic method in the sound of the world’s languages. Other surveys of vowels by acoustic approach in the world’s language, includes Danish, (Fischer-Jorgensen, 1972); Dutch, (Pols, Tromp & Plomp, 1973), American English (Peterson & Barney, 1952; Olive, Greenwood & Coleman, 1993; Hillenbrand, Getty,. 政治大. Clark & Wheeler, 1995; Ladefoged, 2006), British English (Wells 1963), Japanese. 立. (Chiba & Kajiyama, 1941), Korean (Yang, 1992; Yuen, 2001), Russian (Halle, 1959). ‧ 國. 學. Swedish (Fant, 1973), Beiging Mandarin (Howie, 1976, Wu 1986, Zee 2000, 2001;. ‧. Zee & Lee 2001), Cantonese (Zee, 1999),Taiwanese Southern Min (Myers & Tsay. y. Nat. er. io. sit. 2003), Taiwan Mandarin (Liu, Zeng & Cao, 1999; Fon, Chiang & Chueng, 2004; Pan, Li, Lee, Huang & Tsou, 200), and Hakka (Liang, 2004; Huang, 2004; Deng, 2006;. n. al. Ch Zee & Lee, 2008; Zee & Lee, 2009).. engchi. i Un. v. 2.5 Acoustic Studies of Hakka Some research had been dedicated to the acoustic studies on Hakka. In this section, the previous acoustic studies on dialects of Hakka spoken in Taiwan and Mainland China will be introduced. 2.5.1 Acoustic Studies of Hakka Spoken in Taiwan In this section, acoustic studies in Taiwan will be introduced. In this section, we 34.

(47) focus on types of Hakka dialect spoken in Taiwan area, mainly Si-xian Hakka in different areas. Former acoustic research in studying speech sounds focus on various topics, including consonants and vowels (Liang 2004 for consonants and vowels of Si-xian Hakka spoken in Meinung Cheng et al. 2009 for vowel pattern of Si-xian Hakka), tones and vowels (Huang 2004 for tones and vowels of Si-xian Hakka spoken in Meinung).. 政治大. Liang (2004) conducted an acoustic research in his thesis. He describes the. 立. acoustic characteristics and set up models of acoustic nature of Hakka vowels and. ‧ 國. 學. consonants. His study focused on the Si-xian Hakka spoken in Meinung. The acoustic. ‧. cues are the first and second formant frequencies, Voice Onset Time (VOT), release. Nat. er. io. sit. y. burst, turbulence noise, nasal murmur and formant transition. The results of Liang’s (2004) study show that the Hakka vowel /i/ is the most front and highest vowel (even. al. n. iv n C higher than /u/), and the vowel /ɨ/hiseancentralhand high vowel (approximate to the gc i U height of /u/), and the vowel /a/ is a central and low vowel. With regard to vowel. height and F0 frequency, the higher vowel is, the higher F0 frequency it shows in the analysis. In addition, compare to the unchecked tones, the tongue positions of six vowels on checked tones seem to have the tendency of neutralization. Liang (2004) concluded that, generally, the Hakka vowels /i, e, a, o, u/ are similar to the English vowels /i, e, a, o, u/ (Ladeforged & Maddieson, 1996), and only Hakka /a/ is produced 35.

(48) more front than English /a/. Huang (2004) investigated the physical realities of the Hakka tones and vowels by taking acoustic and statistical approaches. He targeted the Hakka spoken at Meinong, a township in the Kaohsiung County. The quantitative information of the citation and sandhi tones of Hakka come from four acoustic cues: fundamental frequency (F0), F0 slope, duration, and intensity. With regard to vowels, he also. 政治大. examined the influence of vowel height on the fundamental frequency and the gender. 立. differences of the four cues.. ‧ 國. 學. Deng (2006) conducted an acoustic study on the Hakka palatalized fricatives.. ‧. The consonant characteristics are explored with the friction duration time, LPC. y. Nat. io. sit. frequency and formant transition to gain the consonant specifics among the fifteen. er. Mandarin, Si-xian Hakka Hai-lu fricative consonants. Deng’s (2006) study utilized the. al. n. iv n C U She utilized Praat system field study to collect sounds amonghthe e nsixgHakka c h i dialects.. to fetch the information of acoustic cues such as the fundamental frequency, the noise duration time, the intensity and the LPC frequency for statistical analyses. Cheng et al. (2009) explored the systematicity and variability of the vowel pattern in Sixian Hakka from an acoustic perspective. In terms of social variables as gender and age, Cheng et al (2009) found that females’ vowel space is comparatively larger than males’. Furthermore, research results show the apical vowel [ɨ] moves 36.

(49) gradually in the direction toward [i]. The younger the informant is, the greater approximation to [i]. Cheng et al. (2009) provided phonological explanations for this gradual pattern change from four perspectives: markedness, vowel system, speech perception and language evolution. 2.5.2 Acoustic Studies of Hakka Spoken in Mainland China This section introduces the acoustic studies of Hakka spoken in Mainland China.. 政治大. Zee & Lee conducted a series of acoustic studies on Yongding Hakka vowels. They. 立. examine the effect of vowel duration on formant frequencies, and analysis of tone and. ‧ 國. 學. tone sandhi (Zee & Lee, 2008), and the vowel space and the gender differences of the. Nat. y. ‧. three vowels /i a ɨ/ of Yongding Hakka (Zee & Lee, 2011).. er. io. sit. Zee & Lee (2008) conducted studies on the effect of vowel duration on formant frequencies, and analysis of tone and tone sandhi in Hakka dialect, based on the. al. n. iv n C h Mainland acoustic data from Hakka spoken in e n g cChina. h i UIn their studies on the effect of vowel duration on formant frequencies, they examine whether more schwa-like formant frequencies as vowel duration is shortened. Zee & Lee (2011) conducted studies on the Yongding Hakka vowels. Yongding Hakka has a vowel system of three vowel phonemes /i a ɨ/. Their results show for both genders, F2 for [a] is significantly higher for Yongding than other Chinese dialects. 37.

(50) 2.6 Summary In previous studies, the acoustic approach had been widely applied to the investigation of speech sounds of languages in the world. Some studies have been dedicated to the acoustic properties of speech sounds of Hakka dialects, but most of the former acoustic research in studying speech sounds focus on the phonological system of Si-xian dialect.. 政治大. Since there are scarcely any studies focusing on the acoustic studies of Hai-lu. 立. Hakka segments, the aim of this present study is to investigate the acoustic nature of. ‧ 國. 學. Hai-lu Hakka vowels, and hope to provide reference materials for further research on. ‧. speech cognition, production, perception, pathological diagnosis and correction of. y. Nat. n. al. er. io. sit. speech disorders, and Hakka language teaching.. Ch. engchi. 38. i Un. v.

(51) CHAPTER III METHODOLOGY 3.1 Subjects A total of six subjects were invited to participate in the study. The subjects were aged from fifty-three to seventy-eight years old. Three of them were male and the other three subjects were female. All of them were Hakkaese and were brought up in. 政治大. Hakkaese families, and they speak Hai-lu Hakka as their mother tongue. Their. 立. birthplaces and living places are within Xinpu town, Hsinchu County. As for the. ‧ 國. 學. language background of the subjects, all of the subjects are fluent in using Hai-lu. ‧. Hakka, and one female subject is fluent in using Si-xian Hakka, too. All of the. y. Nat. er. io. sit. subjects can speak Taiwan Mandarin fluently. Most of them can understand Taiwanese Southern Min, but they cannot speak Taiwanese Southern Min fluently.. al. n. iv n C U but most are considered not hen Four of the subjects have learned some English h g c inischool fluent by self evaluation. For the detailed data and language background of the subjects, please see Appendix I. 3.2 Materials This section presents the materials in the present study, including design of testing words, and equipments.. 39.

(52) 3.2.1 Design of testing words There were 32 test items in total in this study. The 32 test items include five types of syllable structure: CV, CVV, CVVV, CVC, and CVVC. During the experiment, the testing words were produced in two forms: the citation form and the sentence form. In the citation form, the test words were produced in individual syllables with several seconds of interval between syllables. As for the sentence form,. 政治大. the test words were produced in the short sentence [tshiaŋ24 ŋiam33 tshut5 ___ lia24. 立. kai11 sɨ33 loi55] (Please say ____ this word), and there are intervals between these. ‧. ‧ 國. 學. words.. Almost all of the testing words have the tone value [55], except for the words:. Nat. er. io. sit. y. [nie31] (meaning, ‘蟻’ ants). Tone [55] is chosen for consistency purpose, and the other reason is that it is the most common tone value which has the most possible. al. n. iv n C h emore There are syllables that n gpossible chi U. syllables in the lexicon.. include a greater. variety of combinations of segments in the syllable final and could also be written down in Chinese characters. The rationale behind the selection of the same tone value for test items is because studies such as Howie (1976), Tsao & Yang (1984), and Hoole & Hu (2004) found that tones do have influences on the vowel quality. For example, Howie (1976) found syllables with high tones have higher vowel quality. For the detailed test words involved in this study, please see Appendix II. 40.