University of Macau / Faculty of Arts and Humanities. / Department of Chinese

碩士 / 國立臺灣師範大學 / 國文學系 / 101

碩士 / 國立交通大學 / 語言與文化研究所 / 91 / Tonal coarticulation occurs in combinations of any tone types (Shen, 1990), except for the combination of Tone 3s, which requires factors more than tonal coarticulation. When a Tone 3 morpheme combines another syllable, it either becomes 21 when it is followed by a non-T3 morpheme; or if the Tone 3 morpheme is followed by another Tone 3 morpheme, the tonal value of the first Tone 3 becomes 35. When two Tone 3 morphemes are linked together, regardless of whether the semantic relationships of the two morphemes are related or not, tonal coarticulation occurs; moreover, the first syllable is usually generalized as Tone 2. In this case, it is generally assumed that Mandarin Tone 2 and sandhi Tone 3 are perceptually the same. In previous researches, such as phonetic production experiments, perceptual identification experiments were conducted in order to clarify the mystery of third tone sandhi in Mandarin. In all perception tests, native Mandarin speakers identified sandhi Tone 3 as Tone 2; while in production tests, subjects produced sandhi T3 slightly differently from Tone 2 but without strong significant statistic difference. Therefore Tone 2 and sandhi Tone 3 are neutralized according to the results of previous production and perception studies. However, in a categorization test, Taiwanese Mandarin speakers tended to categorize sandhi Tone 3 as Tone 3 rather than Tone 2 (Peng, 2000). Thus, there remain questions about the neutralization of Tone 3 in sandhi form. The asymmetric results among the tests that Peng conducted could have arisen from the stimuli used which were real words. When exposed to real words, lexical tones fixed in subjects’ mind could influence subjects’ choices in their memory rather with the real deeper mental process. Moreover, in Concept Formation (Jeager, 1986), which was a paradigm used in Peng’s categorization test, before the combination of T3 + T3 (would perceptually be recognized as T2 + T3) was presented as test items in the test session, subjects were exposed to T2 + T3 first in the learning session. This might also cause an effect on Peng’s work. Therefore, in this study, we conducted two psychological experiments. First, in CF test, we avoided preforming the ambiguous category of the sandhi combination (T2 + T3 and T3 + T3) and included a semantic judgment step in the experiment to force the subjects to undergo a deeper mental lexical processing. To avoid the orthography interference, half of the testing items were nonsense, and word frequencies of all test items are properly controlled. Second, a reaction time survey based on priming effect was also tested. The subjects must make responses as fast as possible to the characters appearing on a PC screen during the test, and nonsense words were also used. Our basic assumption is that since sandhi Tone 3 was said to be statistically not significantly different form Tone 2 both in production and identification tests, this neutralization should also be found in the psychological research.

華語三聲變調的運作範疇(domain)一直以來皆是漢語音韻學家激烈探討的主題，不同研究對於運作範疇的定義皆不同，而過去研究也留下了一些尚待解決的問題，例如分類詞(Classifier)以及介係詞(Preposition)如何與其他音節形成三聲變調的運作範疇。本篇論文藉由觀察語料庫中324句最短為6音節，最長為24音節且由數個三聲音節所組成的中文句子，來檢視台灣華語三聲變調若以Nespor & Vogel (1986)所提出之韻律階層(Prosodic hierarchy)為範疇來運作是否能得出自然且多樣的聲調組合。以下分析主要根據韻律階層的順序分成以下四部分：音韻詞(Phonological word)、附著詞組(Clitic group)、音韻詞組(Phonological phrase)、語調詞組(Intonational phrase)。而作者假設三聲變調有兩種運作方式，一種為循環(Cyclic)，一種為同時(Simultaneous)。循環運作時，三聲變調可先以音韻詞邊界為範疇，再以附著詞組邊界為範疇，再以音韻詞組邊界為範疇，最後再以語調詞組邊界為範疇。同時運作時，三聲變調則只以語調詞組邊界為範疇。語料庫分析結果顯示，中心語(Head)之非遞歸邊(Nonrecursive side)的分支性(Branchingness)需被納入音韻詞組的定義中以得出正確的聲調組合。再者，語調詞組的定義也需將語意(Semantic)以及韻律(Metric)的因素納入考量，以限制語調詞組邊界的形成。經過修正，韻律階層理論不僅能成功描述分類詞以及介係詞之三聲變調現象更可預測華語三聲變調的多樣性。 / Mandarin tone sandhi has been argued by different scholars to apply to different phonological domains such as the foot, the phonological phrase etc. This thesis tackles three main issues left by Shih (1986) and Hsiao (1991): the classifier, the preposition and the under-generation problem. In revising the framework of Nespor & Vogel (1986)’s Prosodic Hierarchy, this thesis proposes a possible explanation of the problems. This thesis constructed a corpus of 324 sentence tokens consisting of 6 to 24 third tone syllables. The tokens were recorded from four Taiwan Mandarin native speakers at a speech rate of at least 200 beats per minute. Through the observation of the tone patterns collected, it is found that the original definitions of the phonological phrase and the intonational phrase given by Nespor & Vogel (1986) are not able to describe the Taiwan Mandarin data. This thesis provides a revision of the prosodic hierarchy by taking into account the branchingness of the nonrecursive side on the phonological phrase level and by incorporating Selkirk (1984)’s Sense Unit Condition on the intonational phrase level. Finally, a foot formation condition is also proposed in order to restrict the minimal length of a phonological phrase or an intonational phrase.

博士 / 國立清華大學 / 語言學研究所 / 92 / This dissertation investigates the nature of tone sandhi by focusing on the issue of unpredictable tone sandhi operation directionalities that has being attracting much attention lately. Based on data of five Chinese dialects, including Tianjin, Boshan, Sixian-Hakka, Chengdu and Beijing Mandarin, and the Kuki-Chin language of Hakha-Lai, an intriguing correlation between tone sandhi operation directionalities and normal vs. misapplications is found. In the tone sandhi patterns that are direction-sensitive, target-to-trigger rule application directionality would produce misapplication outputs (outputs with tonal changes that are not properly conditioned (i.e., overapplication) or failure of tonal changes when properly conditioned (i.e., underapplication)), while trigger-to-target directionality would produce outputs of normal application (outputs with tonal changes that are properly conditioned). It is argued in this dissertation that over- and underapplications in tone sandhi, like those observed in reduplications and paradigms, are identity effects. They are forced by the desire for a tonal output to be more like a tonal base it prosodically relates to. The desire to achieve identity (captured by the output-to-output correspondence constraint) forces tone sandhi to operate in the target-to-trigger direction, leading to misapplications. Prosodically related outputs would however sometimes fail to correspond. If preserving identity would produce forms that contain highly marked sequences (captured by the markedness constraints) or forms that involve tonal changes taking place at the prominent position (captured by the positional faithfulness constraint), the desire for identity would be sacrificed. In that case, tone sandhi operates in the reverse direction and the resultant outputs are those of normal application. Thus, the directionalities are predictable through the interactions of the markedness constraint/positional faithfulness constraint and the output-to-output correspondence constraints, where the markedness constraint/positional faithfulness constraint must dominate the output-to-output correspondence constraint. The investigation of the issue of directionality discloses an important feature of tone sandhi. In tone sandhi, identity preservation between prosodically related outputs is important. The output-to-output correspondence relation may force a tonal output to deviate from the canonical surface patterns of the language, so that it becomes more like a tonal base to which it prosodically relates. Identity preservation is highly respected in tone sandhi, unless this would produce forms that are highly marked or forms that involve tonal changes in the wrong position.

碩士 / 國立政治大學 / 語言學研究所 / 105 / Mandarin tone sandhi has been argued by different scholars to apply to different phonological domains such as the foot, the phonological phrase etc. This thesis tackles three main issues left by Shih (1986) and Hsiao (1991): the classifier, the preposition and the under-generation problem. In revising the framework of Nespor & Vogel (1986)’s Prosodic Hierarchy, this thesis proposes a possible explanation of the problems. This thesis constructed a corpus of 324 sentence tokens consisting of 6 to 24 third tone syllables. The tokens were recorded from four Taiwan Mandarin native speakers at a speech rate of at least 200 beats per minute. Through the observation of the tone patterns collected, it is found that the original definitions of the phonological phrase and the intonational phrase given by Nespor & Vogel (1986) are not able to describe the Taiwan Mandarin data. This thesis provides a revision of the prosodic hierarchy by taking into account the branchingness of the nonrecursive side on the phonological phrase level and by incorporating Selkirk (1984)’s Sense Unit Condition on the intonational phrase level. Finally, a foot formation condition is also proposed in order to restrict the minimal length of a phonological phrase or an intonational phrase.

Esta dissertação trata dos processos fonológicos de elisão, ditongação e degeminação no falar de São Paulo. Além dos trabalhos clássicos sobre esses processos de sândi vocálico externo, em Português Brasileiro, serão apresentados dados de um experimento feito para este trabalho. De acordo com os dados do corpus desta dissertação, houve preferência pela aplicação da elisão e não da ditongação, nos casos em que ambos os processos eram possíveis. Os números referentes à elisão confirmaram a afirmação de Bisol com relação à aplicação categórica desse processo quando a vogal a ser elidida é [a]. No entanto, houve alta aplicação de elisão de [u], indicando um favorecimento pela elisão e não pela ditongação. Houve ainda algumas ocorrências de elisão de vogal [coronal], quando esta vogal era precedida por uma consoante que partilhava os mesmos traços com ela. Considerando-se os contextos para ditongos crescentes vs ditongos decrescentes, observou-se uma forte preferência pelo ditongo crescente nos dados coletados do dialeto de São Paulo. Finalmente, com relação à posição do contexto de aplicação dos processos, na seqüência de três vogais adjacentes (V1V2V3), observou-se que o contexto V1+V2 favorece a ocorrência de elisão, e não da ditongação. O processo de elisão, no corpus desta dissertação, foi mais aplicado quando a vogal a ser elidida estava na fronteira de grupos clíticos, podendo estar ou na fronteira de sintagmas fonológicos ou dentro de um mesmo sintagma. / This dissertation deals with the phonological processes of vowel elision, diphthongation and vowel degemination used by speakers from São Paulo. Besides presenting some classic studies about the so called processes of external sandhi in Brazilian Portuguese, this dissertation will also present new data recorded in order to carry on the analysis. Data showed that vowel elision is more productive than diphthongation in contexts within which both processes were possible to be applied. Besides, they confirmed Bisol\'s hypothesis that there is categorical use of vowel elision when the vowel (to be elided) is [a]. Nevertheless, elision of vowel [u] was also productive, indicating the preference for vowel elision over diphthongation. There were also some occurrences of coronal vowel elision, when this vowel and its preceding consonant shared phonological features. Also, it could be noted a strong preference for rising diphthongs, not for the falling ones, in the data collected in São Paulo. Finally, it was observed that the sequence \"first vowel + second vowel\" - V1+V2 - (in a sequence formed by three adjacent vowels - V1+V2+V3) favors the use of elision over diphthongation. Vowel elision, according to the data collected in order to carry on the analysis presented in this dissertation, was more productive in clitic group boundaries, within the phonological phrase as well as in phonological phrase boundaries.

Esta dissertação trata dos processos fonológicos de elisão, ditongação e degeminação no falar de São Paulo. Além dos trabalhos clássicos sobre esses processos de sândi vocálico externo, em Português Brasileiro, serão apresentados dados de um experimento feito para este trabalho. De acordo com os dados do corpus desta dissertação, houve preferência pela aplicação da elisão e não da ditongação, nos casos em que ambos os processos eram possíveis. Os números referentes à elisão confirmaram a afirmação de Bisol com relação à aplicação categórica desse processo quando a vogal a ser elidida é [a]. No entanto, houve alta aplicação de elisão de [u], indicando um favorecimento pela elisão e não pela ditongação. Houve ainda algumas ocorrências de elisão de vogal [coronal], quando esta vogal era precedida por uma consoante que partilhava os mesmos traços com ela. Considerando-se os contextos para ditongos crescentes vs ditongos decrescentes, observou-se uma forte preferência pelo ditongo crescente nos dados coletados do dialeto de São Paulo. Finalmente, com relação à posição do contexto de aplicação dos processos, na seqüência de três vogais adjacentes (V1V2V3), observou-se que o contexto V1+V2 favorece a ocorrência de elisão, e não da ditongação. O processo de elisão, no corpus desta dissertação, foi mais aplicado quando a vogal a ser elidida estava na fronteira de grupos clíticos, podendo estar ou na fronteira de sintagmas fonológicos ou dentro de um mesmo sintagma. / This dissertation deals with the phonological processes of vowel elision, diphthongation and vowel degemination used by speakers from São Paulo. Besides presenting some classic studies about the so called processes of external sandhi in Brazilian Portuguese, this dissertation will also present new data recorded in order to carry on the analysis. Data showed that vowel elision is more productive than diphthongation in contexts within which both processes were possible to be applied. Besides, they confirmed Bisol\'s hypothesis that there is categorical use of vowel elision when the vowel (to be elided) is [a]. Nevertheless, elision of vowel [u] was also productive, indicating the preference for vowel elision over diphthongation. There were also some occurrences of coronal vowel elision, when this vowel and its preceding consonant shared phonological features. Also, it could be noted a strong preference for rising diphthongs, not for the falling ones, in the data collected in São Paulo. Finally, it was observed that the sequence \"first vowel + second vowel\" - V1+V2 - (in a sequence formed by three adjacent vowels - V1+V2+V3) favors the use of elision over diphthongation. Vowel elision, according to the data collected in order to carry on the analysis presented in this dissertation, was more productive in clitic group boundaries, within the phonological phrase as well as in phonological phrase boundaries.

No description available.

No description available.