Orientador: Gladis Massini-Cagliari / Banca: José Sueli de Magalhães / Banca: Célia Marques Telles / Banca: Maria Bernadete Marques Abaurre / Banca: Cristina Martins Fargetti / Resumo: Esta tese objetiva estudar os processos fonológicos de sândi vocálico externo - ditongação, elisão de V1, crase e não realização ou apagamento de V2 - no Português Arcaico (PA) em 200 Cantigas de Santa Maria (CSM), século XIII, à luz das teorias não lineares. As CSM são uma coleção de 420 cantares em louvor da Virgem Maria compiladas em galego-português por Afonso X (1221 - 1284). Partimos da escansão, da contagem das sílabas poéticas dos versos das CSM e verificamos a localização dos acentos, elucidando dúvidas acerca da delimitação da palavra em PA e da resolução de encontros vocálicos em juntura vocabular. A metodologia busca abstrair da escansão dos versos em sílabas poéticas e do padrão de metrificação presente nas CSM os limites entre as palavras da língua daquela época, trazendo informações necessárias para uma pesquisa sobre a prosódia de línguas em relação às quais o contato com os falantes nativos não é possível. Os contextos de aplicação rítmicos e fonotáticos dos processos de sândi vocálico externo constituem o cerne de nosso estudo, que se expande, também, para a análise da ocorrência do hiato. Os resultados obtidos revelam que, na primeira fase da história da língua portuguesa, os processos fonológicos que desfazem a estrutura de hiato nos dias de hoje começaram a atuar ainda no PA. Ademais, esses eram expressivos no PA, por exemplo: a elisão, se comparada com o hiato (a permanência das duas vogais envolvidas no contexto de juntura vocabular), contém uma margem de aplicação muito próxima à manutenção de vogais, isto é, tanto a elisão quanto a manutenção do hiato eram produtivos naquele momento da língua. Por outro lado, a ditongação, a crase e o apagamento ou não realização de V2 consistiam em processos marginais, de contextos de aplicação bem mais restritos. Esses fatos se devem, muitas vezes, às restrições de natureza rítmica, prosódica ... / Abstract: This thesis aims to study the processes of external phonological vocalic sandhi in 13th century Archaic Portuguese (AP). These processes are diphthongization, V1 elision, crasis and deletion or nonproduction of V2 and will be analyzed in 200 Cantigas de Santa Maria (CSM) from the point of view of the Nonlinear Phonology theories. The CSM are a collection of 420 chants in praise to Virgin Mary compiled in Galician- Portuguese by Alfonso X (1211-1284). We started from the scansion of the CSM verses, verifying the location of stresses and clarifying doubts about the word's delimitation in AP. The methodology aims to abstract the limits between words by from the scansion of the verses in poetic syllables and from the metrification patterns present in the CSM. The rhythmic and phonotactic contexts of application of the processes of external phonological vocalic sandhi constitute the core of our study, which extends itself to the analysis of hiatus occurrences. In the first phase of Portuguese language the results achieved reveal that the phonological processes which undo the hiatus structure nowadays started to act already in AP. Besides, those processes were expressive in AP, for instance: the elision has an application margin close to the vowels maintenance when compared to hiatus (the permanence of two vowels which are involved in word junctioncontext). Both elision and hiatus were more productive at that moment of the language than diphthongization, crasis and deletion or nonproduction of V2, which were marginal phonological processes, with more restricted application contexts. These facts are due to rhythmic, prosodic and phonotactic restrictions in AP. This thesis also studies the restructuring of the syllabic margins and of the nucleus, with emphasis to ... / Doutor

Orientador: Gladis Massini-Cagliari / Banca: José Sueli de Magalhães / Banca: Ester Mirian Scarpa / Banca: Cristina Fargetti / Banca: Daniel Soares da Costa / Acompanha CD-ROM / Resumo: Esta tese objetiva analisar a ocorrência de processos fonológicos de cunho estilístico, isto é, processos considerados não esperados, a partir de um corpus do Português Medieval: as Cantigas de Santa Maria. Trata-se, portanto, de verificar os possíveis condicionamentos linguísticos envolvidos na sua realização, sobretudo no que se refere a fenômenos de sândi - crase elisão e hiato - e paragoge. As Cantigas de Santa Maria são atribuídas à lavra de Afonso X de Castela, o rei "Sábio". Esta coleção é composta por 420 cantares em louvor à Virgem Maria, de quem o monarca é adorador. Todavia, para o presente estudo, foi feito um recorte no que se refere ao cancioneiro mariano, sendo coletados, assim, os processos estilísticos atuantes nas cem primeiras cantigas. Entende-se como processo não esperado aquele que deixa de atuar em contextos em que sua realização é esperada, o que inclui casos típicos de opacidade (cf. KIPARSKY, 1985). Desta forma, pode-se dizer que esta pesquisa tem como objetivo, ainda, descrever as regras opacas dos referidos fenômenos. O objetivo principal é, contudo, fazer uma análise do tratamento que vêm recebendo tais formas tidas como desviantes, no âmbito de teorias fonológicas, desde o estruturalismo até modelos mais recentes, como a teoria da Otimalidade. O que foi possível constatar é que as teorias, de modo geral, encontram sérias dificuldades de inserir processos estilísticos na gramática da língua, sendo a Fonologia Lexical (KIPARSKY, 1982, 1985; MOHANAN, 1986; PULLEYBLANK, 1986) o modelo que mais se aproxima de uma solução, à medida que descreve tais fenômenos e os enquadra na gramática fonológica. Por conta disso, esta tese defende que esses casos não podem ser relegados à margem da gramática. Deste modo, para descrição e análise das regras de sândi e paragoge, lançaremos... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: This thesis aims to analyze the occurence of stylistic phonological processes occurrence, or processes considered unexpected, from a corpus of Medieval Portuguese: The Cantigas de Santa Maria. Possible linguistic constraints involved in their performance are analyzed particularly as regards the phenomena of sandhi - crasis, elision and hiatus - and paragoge. The Cantigas de Santa Maria are attributed to Alfonso X of Castile, the king "Wise". This collection is composed of 420 cantigas in praise of Virgin Mary. However, for the present study a necessary cutout was made with regard to the Marian repertoire, being collected stylistic processes which operate in the first hundred cantigas. It is considered as an unexpected process the rule which do not act in contexts where its occurrence is expected, which includes typical cases of opacity (cf. Kiparsky, 1985). So, we can say this research aims also to describe the rules of said opaque phenomena. However, the main objective is to analyze the treatment of deviant forms in the context of phonological theories from structuralism to newer models like Optimality Theory. It was possible to note that theories generally have serious difficulties in inserting stylistic processes in grammar being Lexical Phonology (KIPARSKY, 1982, 1985; MOHANAN, 1986; PULLEYBLANK, 1986) the theory that most approaches a solution, as it describes such phenomena and fits them in phonological grammar. Because of this, this thesis argues that these cases can not be relegated to the margins of grammar. Thus, for description and analysis of sandhi and paragoge rules we will use the model proposed by Lexical Phonology (KIPARSKY, 1982, 1985, MOHANAN 1986; PULLEYBLANK, 1986), as well as studies on syllable structure - like Selkirk (1982) and Collischonn (2005) -, since this analysis requires detailed syllabic structure... (Complete abstract click electronic access below) / Doutor

Orientador: Gladis Massini-Cagliari / Banca: Daniel Soares da Costa / Banca: Flaviane Romani Fernandes Svartman / Resumo: O objetivo principal desta pesquisa é estudar o direcionamento da adjunção de clíticos fonológicos no Português Arcaico (daqui em diante PA), a partir das Cantigas de Santa Maria de Afonso X, o rei Sábio, a fim de se chegar à determinação do direcionamento da cliticização e a pistas da formação de constituintes prosódicos. Trata-se de averiguar a possibilidade de se considerar o grupo clítico como constituinte prosódico relevante no PA. Para comprovar tal possibilidade três fatores foram considerados: as pistas que vêm da música, as que vêm da estrutura poética e o processo de sândi. Como a origem e a evolução dos fenômenos prosódicos do Português ainda são, em grande parte, um dos pontos mais inexplorados da história da nossa língua, a descrição dos fenômenos prosódicos e de sua relação com os processos segmentais de um período passado desta língua (no caso, o PA) constitui uma contribuição importante no sentido de elucidar mais completamente a história da Língua Portuguesa. A metodologia deste trabalho baseia-se no mapeamento dos pronomes oblíquos e reflexivos clíticos. Desta forma, a pesquisa visa uma análise quantitativa e qualitativa dos dados, uma vez que, a partir da quantificação da ocorrência desses pronomes e do seu posicionamento, é possível chegar a afirmações quanto à formação de constituintes prosódicos superiores. As conclusões a que chegamos estão baseadas em fundamentos que levam em consideração aspectos rítmicos das cantigas. Os resultados obtidos através das investigações realizadas acerca da possibilidade de os clíticos do PA assumirem proeminência poética e musical apontam para a consideração da tonicidade do clítico no nível lexical, no momento histórico investigado. Desta forma, o clítico só poderia ser adjungido a um constituinte prosódico que preservasse... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: This research aims to study the direction of clitic pronoun adjunction in Archaic Portuguese (AP) (13th century), analyzing the texts of the reminiscent religious medieval cantigas (420 Cantigas de Santa Maria, compiled by Alfonso X, the Wise). The objective is to determine the cliticization direction, in order to find clues to the formation of superior prosodic units. This study intends to show the possibility of the clitic group be considered a relevant prosodic constituent for AP. To prove this possibility we have considered three factors: the metrics of the verses, the music and the sandhi phenomena. As the origin and the evolution of Portuguese prosodic phenomena are one of the most unexplored points of our linguistic history, the description of prosodic phenomena and its relation to segmental processes in a past period of the language represents an important contribution in the sense of enlightening specific points of Portuguese history. The methodology is based on mapping all unstressed pronouns in the cantigas. The analysis is quantitative and qualitative; departing from the quantification of the occurrence of accusative, dative and reflexive pronouns and its positioning, the purpose is to find clues on the formation of superior prosodic constituents. Our considerations are based on rhythmic aspects of the cantigas. Based on the metrics and on the music we could verify that the clitics could have poetic and musical prominence, so we could suggest their tonicity in a lexical level, in that period of time. In this way, clitics only could be attached to a prosodic constituent that preserved the word tonicity in the lexical level, in other words, the clitic group. Concerning sandhi phenomena, we could conclude that, despite the fact that clitcs in the cantigas are unstressed, they behaved as independent prosodic... (Complete abstract click electronic access below) / Mestre

Made available in DSpace on 2015-05-14T12:43:22Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 579925 bytes, checksum: 8fd8e9523ce46237ba5ccd88da11c49e (MD5) Previous issue date: 2008-03-31 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / This current work investigates the elision‟s sociolinguistic behavior upon the João Pessoa native dialect. Elision is one of sandhi‟s extern vocalic phenomenon observed through all languages and it means fading the low vowel /a/ when it is followed by a different vowel. (ex.: menina humilde> meninumilde). This process analyses is through the theorical-methodological scope that the Quantitative Sociolinguistics comprehends. On the construction of this work it has been used a corpus of spoken language that integrates the Projeto de Variação Lingüística do Estado da Paraíba VALPB (Linguistic Variation Project of Paraíba State), composed by eighteen informers, divided by sex (masculine and feminine), age range (teenagers, adults, elders) and academic time spent (none, five to eight years and more than eleven years). The results found show that the linguistic constraints sise of the first word from a sentence, stress, prosodic constituents, vowel‟s quality present themselves as the most relevant on the elision appliance, therefore, reveling themselves as a process without great social determinants. / Este trabalho analisa o comportamento sociolingüístico da elisão no dialeto pessoense. A elisão é um dos fenômenos de sândi vocálico externo observado nas línguas e consiste no apagamento da vogal baixa /a/ quando esta for seguida de uma vogal diferente (ex.: menina humilde> meninumilde). A análise deste processo está fundamentada no modelo teórico-metodológico que compreende a Sociolingüística Quantitativa. Para a realização da pesquisa, foi utilizado um corpus de língua falada, que integra o Projeto de Variação Lingüística do Estado da Paraíba (VALPB), composto por dezoito informantes, estratificados de acordo com o sexo (masculino e feminino), a faixa etária (jovens, adultos e idosos) e anos de escolarização (nenhum ano, cinco a oito anos e mais de onze anos). Os resultados obtidos mostram que as restrições lingüísticas extensão da primeira palavra da seqüência, acento, tipo de palavra, constituintes prosódicos e qualidade da vogal apresentam-se como as mais relevantes à aplicação da elisão, revelando-se, portanto, como um processo sem grandes determinantes sociais.

博士 / 國立交通大學 / 電信工程系 / 92 / In this dissertation, a new prosody modeling approach for Mandarin speech is proposed. It explicitly takes several major affecting factors as companding factors (CFs) and estimates all model parameters by an EM algorithm. Two important prosodic features, syllable duration and pitch contour, are discussed based on a reading style, 5-speaker, manually segmented, microphone-speech database. In the syllable duration modeling, a multiplicative model is proposed. Affecting factors considered include speaker, utterance, prosodic state, tone, and base-syllable. Besides, the three Tone 3 patterns (i.e., full tone, half tone and sandhi tone) are also properly considered via using three different CFs to separate their affections on syllable duration. Experimental results showed that the variance of the syllable duration was greatly reduced from 180.17 to 2.52 frame2 (1 frame = 5 ms) by the syllable duration modeling to eliminate effects from those affecting factors. Two extensions of the duration modeling method are also discussed. One is the use of the same technique to model initial and final durations. The other is to replace the multiplicative model with an additive one. Besides, the syllable duration modeling method is applied to an automatically-segmented, 500-speaker, telephone-speech database. In the syllable pitch contour modeling, we take the mean and shape of syllable log-pitch contour as two basic modeling units and consider several affecting factors that contribute to their variations. Affecting factors considered include speaker, prosodic state, tone, and syllable initial and final classes. Experimental results showed that RMSEs of 0.362 ms and 0.373 ms in the reconstructed pitch period were obtained for the closed and open tests, respectively. Lastly, a hybrid method which incorporates the above statistical prosodic models with the linear regression method is proposed to predict syllable duration and pitch contour parameters for Mandarin text-to-speech (TTS). Experimental results showed that the hybrid method outperforms the conventional linear regression method. Besides, detail analyses of those two prosodic models are also performed. All inferred values of the affecting factors of duration and pitch models agree well with our prior linguistic knowledge. Automatically labeled prosodic states also provide useful cues to determine the prosodic phrase boundaries including those occurred at inter-syllable locations with and without punctuation marks. Moreover, the studies of three Tone 3 patterns, the relationship among syllable, initial, and final durations, the comparison between multiplicative and additive model, and the use of telephone-speech corpus make us better understand the characteristics of Mandarin syllable duration and confirm the effectiveness of our method. The syllable pitch contour model also gives a quantitative and more complete description to the coarticulation effect of neighboring tones instead of the conventional qualitative descriptions of some few tone sandhi rules. So the proposed prosodic modeling approach is a promising one for Mandarin speech.

Gegenstand der Arbeit sind verschiedene Aspekte der Morphosyntax und Tonologie im Buli, einer in Ghana (Upper East Region nahe der Grenze zu Burkina Faso) gesprochenen Gursprache des Niger-Kongo-Phylums, die bisher nur unzureichend beschrieben wurde. Die grammatische Beschreibung gibt einen umfangreichen Einblick in das sprachliche System des Buli und umfasst Kapitel zu den folgenden Themen: Phonologie, Nominalklassensystem, Pronominalsystem, Komplexe Nominalphrasen und nominale Syntagmen, Verbsystem. Dies geschieht in erster Linie anhand des in Wiaga gesprochenen Buli, wird aber durch die Bezugnahme auf andere Dialekte sowie auf verwandte Gursprachen ergänzt, beispielsweise hinsichtlich der Bewertung tonaler Divergenzen in Nominalklassensuffixen oder in Assoziativkonstruktionen. Die autosegmental angelegte Tonstudie verfolgt keine theoretischen Ziele, sondern hat sich unmittelbar aus dem Bedarf einer adäquaten Beschreibung der synchronen Sprache ergeben, in der Ton sowohl in lexikalischer als auch in grammatischer Hinsicht ganz zentrale Aufgaben übernimmt. Dabei wurden neben regelmäßigen Ton–Sandhi-Erscheinungen (Tieftonausbreitung) auch interessante Phänomene an den Schnittstellen der Phonologie zur Syntax und zur Semantik-Pragmatik vorgefunden, z.B. die tonale Kongruenz bestimmter Verbformen mit der Diskursrolle des Subjekts (+/- Diskurspartizipant, d.h. 1./2. Person vs. 3. Person) oder das Vorkommen eines Grenztons an äußerungsfinalen Morphemen, zu denen auch die Klassensuffixe indefiniter Substantive und enklitische Objektpronomen am Verb gehören. Im Rahmen der morphosyntaktischen Analysen wurde unter anderem ein aus typologischer Sicht spannender nominaler Kompositionstyp identifiziert, der sich neben seinen strukturellen Eigenschaften auch durch eine spezifische possessive Semantik auszeichnet und vermutlich an der Entwicklung von attributiven (qualifizierenden) Adjektiven und eines von zwei Zahlwörtern für ‘eins’ beteiligt war. / This work deals with different aspects of the morphosyntax and tonology in Buli, a Gur language of the Niger-Congo-Phylum, which is spoken in Ghana (Upper East Region near the border to Burkina Faso) and has so far been described insufficiently. The grammatical description provides an extensive insight into the linguistic system of Buli and contains chapters concerning the following topics: phonology, noun class system, pronominal system, complex noun phrases and nominal syntagmata, and verb system. It is primarily based on the Buli variant spoken in Wiaga and supplemented by reference to other dialects as well as to related Gur languages, for instance with regard to tonal divergencies in noun class suffixes and in associative constructions. Following an autosegmental model, the tonal study does not aim at theoretic enhancement, but results directly from the need of an adequate description of the synchronic language in which tone plays an important role in lexicon and grammar. Besides regular Tone-Sandhi (Low tone spreading), interesting phenomena at the phonology/syntax interface as well as at the phonology/semantics-pragmatics interface were observed – e.g. some verb forms displaying tonal agreement with the discourse role of the subject (+/- discourse participant, i.e. 1st / 2nd vs. 3rd person) or the appearance of a boundary tone on utterance-final morphemes, including the noun class suffixes of indefinite nouns and the enclitic object pronouns at the verb. In the course of the morphosyntactic analysis, a typologically remarkable nominal compound type was identified that is distinguished by structural features as well as by specific possessive semantics and can be assumed to be involved in the development of attributive (qualifying) adjectives and one of two numerals with the meaning ‘one’.

Nell’attuale contesto di aumento degli impatti antropici e di “Global Climate Change” emerge la necessità di comprenderne i possibili effetti di questi sugli ecosistemi inquadrati come fruitori di servizi e funzioni imprescindibili sui quali si basano intere tessiture economiche e sociali. Lo studio previsionale degli ecosistemi si scontra con l’elevata complessità di questi ultimi in luogo di una altrettanto elevata scarsità di osservazioni integrate. L’approccio modellistico appare il più adatto all’analisi delle dinamiche complesse degli ecosistemi ed alla contestualizzazione complessa di risultati sperimentali ed osservazioni empiriche. L’approccio riduzionista-deterministico solitamente utilizzato nell’implementazione di modelli non si è però sin qui dimostrato in grado di raggiungere i livelli di complessità più elevati all’interno della struttura eco sistemica. La componente che meglio descrive la complessità ecosistemica è quella biotica in virtù dell’elevata dipendenza dalle altre componenti e dalle loro interazioni. In questo lavoro di tesi viene proposto un approccio modellistico stocastico basato sull’utilizzo di un compilatore naive Bayes operante in ambiente fuzzy. L’utilizzo congiunto di logica fuzzy e approccio naive Bayes è utile al processa mento del livello di complessità e conseguentemente incertezza insito negli ecosistemi. I modelli generativi ottenuti, chiamati Fuzzy Bayesian Ecological Model(FBEM) appaiono in grado di modellizare gli stati eco sistemici in funzione dell’ elevato numero di interazioni che entrano in gioco nella determinazione degli stati degli ecosistemi. Modelli FBEM sono stati utilizzati per comprendere il rischio ambientale per habitat intertidale di spiagge sabbiose in caso di eventi di flooding costiero previsti nell’arco di tempo 2010-2100. L’applicazione è stata effettuata all’interno del progetto EU “Theseus” per il quale i modelli FBEM sono stati utilizzati anche per una simulazione a lungo termine e per il calcolo dei tipping point specifici dell’habitat secondo eventi di flooding di diversa intensità. / In the current environmental context, it emerges the need to understand all the possible effects of the increase of the anthropic impact and the “Global Climate Change” on the ecosystems, considered as the main users of services and essential functions upon which whole economic and social textures are based. The forecast study of the ecosystems collides with the high complexity of them and with the lack of integrated observations. The model approach seems to be the most appropriate for the analysis of the complex dynamics of the ecosystems and the contextualization of experimental results and empiric observations. The reductionist-deterministic approach, which generally is used in the implementation of the models, has not been able to reach higher levels of complexity within the structure of the ecosystem. The biotic factor is the component that best describes the complexity of the ecosystem, because of the high dependency upon the other components and their interactions. In this thesis, it is suggested a new stochastic and model approach based on the employment of a naive-Bayesian compiler operating within a fuzzy environment. The combined employment of fuzzy logic and naive-Bayesian approach is useful to process the high degree of uncertainty of the ecosystems. The obtained generative models, called Fuzzy Bayesian Ecological Models (FBEM) are able to modelize the states of the ecosystems as functions of the high number of interactions. FBEM models have been used to understand the environmental risk for the intertidale habitat of sand beaches in case of coast flooding events forecasted from 2010 to 2100. The application has been developed within the EU “Theseus” Project, in which the FBEM models have been used also for a long term simulation and for the computation of the tipping points specific of the habitat depending on flooding events of different intensity.

碩士 / 國立中央大學 / 客家語文研究所 / 98 / This thesis is concerned with Raoping Hakka lexicon in Hsinchu County and is divided into seven chapters. Chapter one is an introduction that explains the motive, goals, and methodology of the research. The second chapter is the literature review of Raoping Hakka. Chapter three describes the phonology of the four dialects: Liuchia六家 in Chubei City, 關西Kuanhsi Township, 紙寮窩Chihliaowo in Chunglin Township, and 湖口Huk’ou Township. The fourth and fifth chapters compare the differences in the lexicon and morphology of these Raoping Hakka dialects. Chapter six deals with horizontal comparison between Raoping Hakka in Hsinchu and three closely-related dialects: Zhao’an Hakka in Lunpei崙背, Raoyang 饒洋in Guangdong, and Hailu dialect in Hsinchu. The last chapter is the conclusion, and the lexicon of the four dialects is listed in the appendix. There are only minor differences in the phonology among these four dialects. The first one is citation tone value of yangqu: Liuchia shows high level but the other three are mid-rising. Complex tone sandhi phenomena are typical of Raoping Hakka. Hsinchu Raoping, Raoyang and Zhao’an share many words in common. Two of these words ‘ŋiaŋ53’ and ‘teu11’ may reveal the history between Hakka and the畲She. Raoping Hakka dialects in Hsinchu show the signs of language contact with adjacent Hailu dialect but still preserve their unique vocabulary and characteristic phonological traits transmitted from their ancestors hundreds of years ago in varying degrees.

碩士 / 國立新竹教育大學 / 臺灣語言與語文教育研究所 / 97 / The study investigates how the speakers process different dialects in Taiwanese by using Event-related Potential（ERP）. The experiment manipulates seven conditions containing corresponding differences between Chang-chou and Chhuan-chou dialects in Taiwanese. The picture is shown to the subjects as a prime and induce their expected word. Then a auditorily presented word is played to the subjects a target. There are three conditions manipulated for the target: Chang-chou variant for the picture shown as the prime, Chhuan-chou variant for the picture shown as the prime, and a word which does not match the prime. The subjects reaction time and PMN amplitude are measured. The results show that the native speakers of Chang-chou dialect rarely change their native accent, but the native speakers of Chhuan-chou dialect do. Their behaviors show the following three patterns: （1） Compared with Chang-chou variants, when the subjects heard Chhuan-chou sounds, their reaction time becomes shorter and their PMN amplitude becomes lower, as in the conditions of ionn and iunn. （2）No differences are found in the reaction time or PMN amplitude in hearing Chang-chou and Chhuan-chou correspondents, as in the conditions of i and u. （3）Compared with Chang-chou variants, when the subjects heard Chhuan-chou sounds, their reaction time becomes longer and their PMN amplitude becomes larger, as in the tone sandhi condition of Shang. In summary, these results suggest that the changing stage from Chhuan-chou dialect to the Chang-chou dialect is not the same in the seven conditions. For most of the cases, no differences are found in the reaction time or PMN amplitude in hearing Chang-chou and Chhuan-chou correspondents. This indicates that the two variants are competing and that the Chhuan-chou variants are psychologically real to the subjects, even though many of the Chhuan-chou variants are not found in the production data collected by field methods. Finally, we suggest that the exposures of hearing other dialects in daily lives can affect the process and mechanism of phonetic recognition

碩士 / 國立臺灣大學 / 語言學研究所 / 96 / This study investigated how gender and dialectal differences influence the realization of high tonal targets in Tone 1 and Tone 4 in Taiwan Mandarin and how the focus status interacts with dialectal variations. 32 native speakers of Taiwan Mandarin, balanced in gender, 12 from Taipei, 10 from Taichung, and 10 from Kaohsiung, were recruited to perform reading tasks. Target Tone 1 and Tone 4 syllables occurred in isolation and were also embedded in carrier sentences at the initial, medial, and final positions. For Tone 1, the F0 of the midpoint of the target syllable was measured; for Tone 4, the F0 maximum of the target syllable was examined. Results showed that, for female speakers, the FB0B for target syllables in Tone 1 was the highest for the Taipei dialect, followed by the Kaohsiung dialect, with the Taichung dialect as the lowest among the three dialects. For Tone 4, the Taipei dialect was higher in F0 than the other two dialects. However, female speakers of different dialects set their pitch frame similarly. For male speakers, the FB0B of both target tones in the Taipei dialect and the Kaohsiung dialect were similar. But the Taipei and the Kaohsiung dialects were still higher than the Taichung dialect. As for pitch setting, Kaohsiung male speakers set their pitch frame similarly as Taipei male speakers. Both dialects were higher than Taichung male speakers. In addition, dialectal variations interacted with gender and focus status as well. For male speakers, the trends found for dialectal differences of non-focus items were identical to that of focus items; however, for female speakers, dialectal variations were found for focus items but not for non-focus items. This study found that the dialect formed in Taichung was lower in F0 height than the dialect in Kaohsiung, which could be accounted for by tone sandhi rules and the tonal system of Min dialects in Taichung.