本篇論文旨在從優選理論（Optimality Theory）的架構觀點分析海陸客語中 的三個音韻現象：變調（Tone Sandhi）、介音（Pre-nuclear Gliding）及音節連併（Syllable Contraction）。 　　在變調方面，由於海陸客語中上聲與陰入聲助數詞不因環境而變調，在利用之前文獻解釋時必須另外設定條件來說明海陸助數詞的特殊變調行為，本文建議利用優選論的制約來分析，本文所提出的一組制約，可以成功的解釋海陸客語中的上聲與陰入聲變調現象。 　　在介音方面，本文選擇以端木三所提出的音節結構為海陸客語音節結構，因為海陸客語中介音所呈現的行為較似起首子音的次發音特徵，因此本文亦提出了一組制約解釋海陸介音在音節中的歸屬。 　　在音節連併方面，本文檢驗了之前文獻所提出的方法，發現無法全面解釋海陸音節連併的現象，因此，提出一組關於音段與聲調的連併之制約將海陸音節連併現象以優選理論呈現。 　　本文是第一篇嘗試以優選理論解釋海陸客語音韻現象的研究，希望能從不同於之前文獻的角度討論海陸客語音韻的問題。之前理論運作的缺陷與不足，在本文的分析下，均可以成功的以優選理論之制約層級來說明。 / This thesis aims to explore three issues in Hai-lu Hakka Phonology in terms of Optimality Theory (OT). The three issues are tone sandhi, pre-nuclear gliding and syllable contraction. 　　Hai-lu Shang Toned and Yin-ru Toned numerals preserve their base tones in any case. Thus, previous studies would need to add an additional condition for the special tone sandhi behaviors of Hai-lu numerals. This thesis offers a set of OT constraints to explain tone sandhi in Hai-lu. The OT constraints can successfully operate Hai-lu tone sandhi without additional conditions. 　　Pre-nuclear glides in Hai-lu behave like the secondary articulations of the onsets. This thesis suggests Duanmu's (1990) syllable structure for Hai-lu syllables. A set of constraints are proposed to explain the pre-nuclear gliding in Hai-lu syllables. 　　The approaches previous studies suggest for syllable contraction are not across-the-board in Hai-lu. Thus, a set of segmental and tonal constraints are proposed to explain Hai-lu syllable contraction. The OT constraints operate syllable contraction in Hai-lu without difficulties. 　　This thesis is a pioneering study which aims to examine Hai-lu phonology in terms of OT. With OT constraint rankings, this thesis explains the remaining problems from previous studies successfully.

碩士 / 國立政治大學 / 英國語文學研究所 / 83 / This thesis deals with neutral tone sandhi in Mandarin Chinese at the phrasal and lexical levels. Previous research of the phrasal neutral tone, e.g., R. Cheng (1989), has focused only at the syntactic level. In this research, we employ prosodic theories (Selkirk (1984 & 1986), Nespor and Vogel (1986), etc.) to condition phrasal neutral tone sandhi, and propose a set of focal phrasing principles and a couple of relevant constraints. In addition, we discuss the rhythm of neutral tone under the framework of the beat-counting theory (Hsiao, 1991 & 1994) to see in what ways the theory has to be modified. We propose a pair of revised principles of functor beat assignment and directional beat adjunction, along with a backup rule of the default beat addition. As to neutral tone at the lexical level, we incorporate morphological template theory (McCarthy and Prince (1986)), and lexical phonology theory (Kiparsky (1973 & 1982), Mohanan (1982 & 1986), etc.). We propose that lexical neutral tone target a trochaic foot template, while a default tone target an iambic foot template. Observing the relevant phonological and morphological rules, we find that the neutral- toned data in the literature can be somehow generalized by way of level ordering.

碩士 / 國立高雄師範大學 / 英語學系 / 92 / This thesis mainly examines five issues within the framework of Optimality Theory (Prince and Smolensky 1993, McCarthy and Prince 1993): the nature of the changed Tone 2，　Tone 2 Sandhi (T2S), Tone 3 Sandhi (T3S), the interaction with T2S and T3S, and individual differences in applying T2S. T2S refers to the phenomenon in which the contour of final Tone 2s produced by Southern Min (SM) speakers in Taiwan become falling when not preceded by Tone 3s. T2S is treated as a weakening/neutralization process in final positions in this thesis. As for the nature of changed Tone 2s, phonetic evidence shows that the changed Tone 2 does fall. Evidence from perception and its interaction with T3S imply that changed Tone 2 is phonologically Tone 3. For T2S, the constraint *Rising ]# , which disallow rising contour in final positions, is proposed to be the triggering constraints. To avoid violating *Rising ]# , a final Tone 2 is forced to change. Moreover, T3S receives a more detailed analysis in this thesis. Two essential problems －(a) why is it the second Tone 3 changes instead of the first one?, and (b) why must it change to a Tone 2 but any other tone? －are successfully dealt with by introducing HeadFaithTone and FaithNulearTone, respectively. Previous attempts failed to solve these problems. As for the two patterns of the interaction between T2S and T3S, the difference lies essentially in the ranking of OCP(Whole, L), Dep-T, and *Rising ]# . When OCP(Whole, L) >> Dep-T >> *Rising ]# , T2S fails to apply. When *Rising ]# , OCP(Whole, L) >> Dep-T, T2S applies and triggers T3S. Besides, Individual differences are successfully accounted for by different rankings of the same constraint set. This thesis shows that OT provides a consistent and systematic analysis for the three issues mentioned above.

碩士 / 國立新竹教育大學 / 臺灣語言與語文教育研究所 / 95 / Abstract A debate in Southern-Min language tone sandhi is what derived by rule or stored in mental. No studies have ever tried to demonstrate the nature of the issue with evidences in neurophysiology. We approach the issue by analyzing the records of Event-related brain potentials (ERP) in spoken word priming paradigm of 16 Taiwaneses. Our results elucidate the psychological processes that underlie the N400 component suggest the psychological reality of Southern-Min language tone sandhi and, in line with model of tone snadhi processing which assume that a dual-syllable whole word is a treatment unit during semantic comprehensive processing. Processing of tone sandhi in sentential context aside, however, for assignments, there still an unidirectional transform mechanism is working. Another, analysis of RP (recognition potential) shows that the minimum duration signal necessary to discriminate tone value of a spoken word is about 200ms after onset of stimuli, identified up to 300ms. After all, the implication of our results for the linguistic processes underlying the N400 is biphasic, not only associated with lexical access ( pre-lexical phonological analysis stage) but also post-lexical integration processes.

碩士 / 玄奘大學 / 應用外語系碩士班 / 98 / In the thesis, it discusses AA, AAA, AABB, ABAB patterns of the reduplicated adjectives through Mora theory mainly and Mora theory as a tool find out the original weight or the product of other possible existed weight. The thesis argues the tone sandhi could produce the syllable weight and hypothesize that the tone sandhi could produce a Mora. However, the one Mora is not real value and the hypotheses only wants to prove that the tone sandhi could really have the product of the syllable weight. The purpose of the thesis wants to prove that the tone sandhi could produce the syllable weight to change the semantic intensity of the different reduplicated adjectives. However, the thesis solely has a reasonable hypothesis to prove the syllable weight increases by the tone sandhi and it merely needs a virtual Mora value, not exact one. Key words: the reduplicated adjective, Mora theory, tone sandhi

碩士 / 國立交通大學 / 外國語文學系外國文學與語言學碩士班 / 102 / Phonetic similarity and allophonic sandhi rules were regarded as the possible triggers to tone changes. This paper investigated the tone changes in Hailu Hakka produced by speakers above 50 and under 30 years old. Level tones triplet 55-33-11 with similar pitch contours were examined, as well as sandhi related tone pairs, including 5-3 and 24-33. According to tone sandhi rules, tone /24/->[33]/_σ and tone /5/->[3]/_σ, sandhi tones surface at the initial position of disyllabic words/phrases. Four native Hailu Hakka speakers above 50-year-old and four speakers under 30-year-old produced disyllabic words with target tones produced at either the initial or the final positions of a disyllabic word/phrase. Results showed that tones 55, 33 and 11 with phonetic similar F0 contours were produced with significantly different f0 and duration by speakers above 50 years old. However, speakers under 30 years produced tones 55 and 33 with similar duration. Regarding sandhi related tones 3 and 5, beside f0 of sandhi tones 3 and 5 produced by speakers under 30 years old, both duration and f0 contours of tones 3 and 5 were significantly different from each other for both age groups. Both age groups, produced indistinguishable duration and F0 contours between sandhi tones 24 and 33. In the position where no sandhi rules took place, duration of tones 24 and 33 produced by speakers under 30 years old were not significantly different from each other. It is proposed that F0 of sandhi tones 3 and 5 produced by speakers under 30 years old were merged, but not in the production of speakers above 50 years old. Moreover, the tone merger phenomenon is even more prominent among F0 of sandhi tones 24 and 33 which were merged in the production of both age groups. Besides, the direction of tone merging was also corresponding to the sandhi rules in Hailu Hakka. Tone /5/[3] was merging to tone /3/[3] and tone /24/[33] was merging to tone /33/[33]. Allophonic sandhi rule is affecting Hailu Hakka tone change.

本文調查年輕世代 (1988 至 1995 年間出生者) 閩南語的變調範域 (Tone Sandhi Domain)，即調群 (Tone Group)。文中以「傳統調群」表文獻記載的變調範域，以「新調群」表本文重新調查的變調範域。論文的研究立基於音韻句法互動之上，並持間接指涉假設觀點，從韻律音韻學的角度出發，認為句法音韻間存有一韻律介面，且調群為此介面上的韻律單位：音韻詞組 (Phonological Phrase, ϕ)。 本文發現新調群與傳統調群的劃分相異。前人研究指出，調群邊界 (#) 標於音韻詞組右端，音韻詞組為一最大投射 XP，且不可為附加語及附著語。新調群僅部分可以傳統音韻詞組界定，其他則對應至句法上的主要語 X，以及附加語最大投射 XP。此外，新調群的邊界為選擇性地出現，具語言變異。新調群出現四種於傳統調群未見的劃分形式：依調群邊界前的語法單位分為主要語 X 類別：(1) V # NP # (2) V # QP # 及附加語 XP 類別： (3) QP # A # (4) AdvP # V #。 本文提出調群新定義，並以優選理論中的 ROE 模型進行分析，提出三個新制約：Align-R (XP-Fnc, ϕ)、ϕ-min 及 Align-R (X-head, ϕ)。Align-R (XP-Fnc, ϕ) 用以區分功能性投射與詞彙性投射；ϕ-min 限制音韻詞組至少含兩音節；Align-R (X-head, ϕ) 解釋調群邊界標註於主要語 X 右端。分析時著重制約 Align-R (XP-Fnc, ϕ) 、ϕ-min 及 Align-R (X-head, ϕ) 與切分線 | (Critical Cut-Off Line) 的關係。整體制約排序為：Ident-T-Rϕ | Align-R (XP-Fnc, ϕ) >> ϕ-min >> Align-R (X-head, ϕ) >> *T >> Align-R (XP-Lex, ϕ)。除信實性制約 Ident-T-Rϕ 之外，其他標記性制約皆列於切分線之下，預測語言變異。 / The goal of this thesis is to explore the tone sandhi domain, tone group (hereafter, TG), of Taiwan Southern Min of younger generation. A corpus is built and analyzed within the framework of Prosodic Theory, a derived theory from the Phonology-Syntax Interface Theory. In the literature, TG boundary # was marked at the right edge of an XP which is neither an adjunct nor a clitic. However, the TG boundary # of younger generation is marked at the right edges of a maximal XP, an adjunctive XP and an X-head. Moreover, language variation is observed in the TG parsing: the presence of TG boundary # is optional. There are four TG parsings not observed in the literature: (1) V # NP # (2) V # QP # (3) QP # A # (4) AdvP # V #. TG is redefined as a Phonological Phrase which syntactically corresponds to a maximal XP, an adjunct XP, an X-head but not to a clitic XP. The four TG parsings are then analyzed with Rank-Ordering Model of EVAL (ROE), a subtheory within the Optimality Theory domain which particularly addresses language variation. According to ROE, constraints on the right side of critical cut-off line participate in the prediction of variation. Three constraints involving in variation are proposed: Align-R (XP-Fnc, ϕ), ϕ-min and Align-R (X-head, ϕ). Align-R (XP-Fnc, ϕ) distinguishes lexical projection and functional projection, ϕ-min requires a Phonological Phrase to be minimally disyllabic, and Align-R (X-head, ϕ) predicts # to occur at the right edge of an X-head. The analysis centers around the three constraints’ relation with the critical cut-off line. The constraint rankings are: Ident-T-Rϕ | Align-R (XP-Fnc, ϕ) >> ϕ-min >> Align-R (X-head, ϕ) >> *T >> Align-R (XP-Lex, ϕ). Except for the faithfulness constraint Ident-T-Rϕ, the other markedness constraints are arranged at the right side of the cut-off line in order to predict language variation.

INTRODUCTION: The habitual physical activity (PA) induce desirable changes in plasma lipid levels, especially the increase of HDL-cholesterol (HDL) and reducing triglycerides (TG). There are similarities between the findings of studies on the advantage that causes PA in plasma lipid levels, however there is still a necessary to explain the role of the frequency and intensity of PA on this benefit. Such information is important for making more specific recommendations for PA to increase HDL and decrease LDL and TG. OBJETIVO: Estimate the contribution of PA to the distribution of HDL, LDL and TG among male and female participants of the baseline of the Longitudinal Study of Adult Health (ELSA Brasil) and to investigate if the association between PA and HDL; PA and TG; PA and LDL is modified according to the different assessment parameters of PA. METHODS: Cross-sectional study with participants aged between 35 and 69 years, both sexes. Data collection was conducted through interviews, anthropometric measurements, and laboratory tests. Participants who used drugs that influence the levels of HDL, LDL and TG were excluded. Physical activity was assessed by a modified International Physical Activity Questionnaire (IPAQ) modified. Three different parameters of PA were considered in the analysis: intensity (mild, moderate and strong); weekly time >150 minutes (WHO criteria) and total time PA in minutes. The analysis was performed by univariate and multivariate linear regression using the logarithmic transformation of TG. For inclusion in the multiple linear regression model the variables that were associated with p value <0,20 was chosen and remained in the final models variables that remained associated with the level of p<0,05, after all adjustments. Multiple linear regression was initially adjusted for sex and age, and then by possible confounding variables. RESULTS: Among the 15.105 participants in the Study ELSA-Brasil, 12,688 were included in this study, and 54.8% were female and the mean age was 50 years (± 8.12), 50% reported white skin, 52% had a higher education. The mean BMI was 26,8kg / m2 (± 4.76), 14% were smokers and 69%, users of alcoholic drink, being 8% excessive alcohol users. The mean values of HDL and LDL were, respectively, 49mg/dL and 134mg/dL for men and 60 mg/dL and 133mg/dL for women. Median values of TG were 131mg/dL for men and 100 mg/dL for women. Concerning weekly physical activity during leisure time, 78% of individuals engaged in light physical activity; 13%, and 9% in moderate, or intense PA, regardless of gender. The LDL proved ssociado the intensity of PA in univariate analysis, but lost significance after adjustment estatítica by sex and age. The final linear regression models showed that the levels of HDL increase with greater intensity of PA, which are higher in practice over 150min / week and grow by directly with the total time of PA. The intensity of PA explained 0.05% of the total distribution of HDL in the study population. The PA >150 min/week and total time explained respectively, 0.11% and 0.16% of the distribution of HDL. After adjustments, TG levels decreased significantly with increasing intensity of PA and with longer PA. The intensity of PA explained 0.19% of the distribution of TG in the population studied, since PA as recommended by WHO explains 0.03% of this distribution, and the total time PA explains 0.08%. Our results reinforce the importance of regular practice of physical activity for increasing HDL and reducing TG, but was not associated with levels of LDL. The best explanation for the intensity distribution of HDL and TG levels, but the relative contribution of PA to the general distribution of HDL and TG levels in the study population was small. Among the different parameters investigated the intensity of PA contributed more to explain the distribution of HDL and TG than the WHO recommendation of practice> 150 min/week. But it is noteworthy that our results reinforce the notion that PA practice as recommended by the WHO brings benefits on levels of these lipids. / INTRODUÇÃO: A prática habitual de atividade física (AF) induz alterações desejáveis nos níveis plasmáticos de lipídeos, em especial no aumento da HDL-Colesterol (HDL) e na redução dos triglicérides (TG). Existem similaridades entre os achados dos estudos em relação ao benefício que a AF provoca nos níveis plasmáticos de lipídeos, entretanto ainda há necessidade de aclarar o papel da frequência e da intensidade da AF sobre esse benefício. Tal informação é importante para tornar mais específicas as recomendações de AF com vista a aumentar a HDL e diminuir a LDL e o TG. OBJETIVO: Estimar a contribuição da AF para a distribuição dos níveis de HDL, LDL e TG entre homens e mulheres participantes da linha de base do Estudo Longitudinal de Saúde do Adulto (ELSA Brasil) e investigar se a associação entre AF e HDL; AF e LDL e AF e TG varia segundo diferentes formas de classificação de AF. MÉTODOS: Estudo transversal com participantes com idade entre 35 e 69 anos, ambos os sexos. A coleta de dados foi realizada através de entrevistas, medidas antropométricas e exames laboratoriais. Foram excluídos participantes que fizeram uso de medicamentos que influenciam nos valores de HDL, LDL e de TG. A atividade física foi avaliada pelo International Physical Activity Questionnaire (IPAQ) modificado. Três parâmetros diferentes de AF foram considerados na análise: intensidade (leve moderada e vigorosa); tempo semanal > 150 minutos (recomendação da Organização Mundial da Saúde) e tempo total de AF em minutos. A análise foi feita por regressão linear uni e multivariada utilizando-se a transformação logarítmica do TG. Para inclusão no modelo de regressão linear múltipla foram escolhidas as variáveis que apresentaram associação com p valor <0,20, tendo permanecido nos modelos finais as variáveis que se mantiveram associadas ao nível de p <0,05, após todos os ajustes. A regressão linear múltipla foi ajustada inicialmente por sexo e idade, e posteriormente por possíveis variáveis de confusão. RESULTADOS: Dos 15.105 participantes do Estudo ELSA-Brasil, 12.688 preencheram os critérios de inclusão, sendo 54,8% do sexo feminino e com idade média de 50 anos (±8,12), 50% relatou cor de pele branca, 52% possuía nível superior de escolaridade. A média do IMC foi 26,8kg/m2(±4,76), 14% eram fumantes e 69%, usuários de bebida alcóolica, sendo 8% usuários excessivos de álcool. Os valores médios de HDL e LDL foram, respectivamente, 49mg/dL e 134mg/dL para homens e 60mg/dL e 133mg/dL para mulheres. Os valores medianos dos TG foram 131mg/dL para homens e 100mg/dL para mulheres. Quanto à prática semanal de atividade física no lazer, 78% dos indivíduos praticavam AF leve; 13%, moderada e 9%, vigorosa, sendo similar a distribuição da atividade física por gênero. O LDL mostrou-se ssociado à intensidade da AF na análise univariada, mas perdeu significância estatítica após ajuste por sexo e idade. Os modelos finais de regressão linear mostraram que os níveis de HDL crescem com a maior intensidade da AF, são maiores nos que praticam mais de 150min/semana e crescem diretametne com o tempo total de AF. A intensidade da AF explicou 0,05% da distribuição total dos níveis de HDL na população estudada. AF>150 min/semana e tempo total explicaram respectivamente, 0,11% e 0,16% da distribuição do HDL. Após ajustes, os níveis de TG reduziram significativamente com o aumento da intensidade da AF e com o maior tempo de AF. A intensidade da AF explicou 0,19% da distribuição dos TG na população estudada, já AF segundo recomendação da OMS explica 0,03% dessa distribuição, e o tempo total de AF explica 0,08%. Nossos resultados reforçam a importância da pratica regular de atividade física para o aumento do HDL e redução dos TG, mas não foi associada aos niveis de LDL. A intensidade explicou melhor a distribuição dos niveis de HDL e TG, mas a contribuição relativa da AF para a distribuição geral dos niveis de HDL e TG na população de estudo foi pequena. Entre os diferentes parâmetros investigados a intensidade da AF contribuiu mais para explicar a distribuição de HDL e TG do que a recome ndação da OMS de pratica >150 min/semana. Mas, vale ressaltar que nossos resultados reforçam que a prática de AF como recomenda a OMS traz benefícios sobre os niveis desses lipídios.

Introduction: The language comprehension is a complex skill, and the factors associated with disorders affecting this skill have not yet been clearly established. Age, schooling, general health conditions, depression and quality of life are the most common factors investigated. Objectives: 1) To describe the performance in the shorter version of the Token Test for language comprehension of an elderly population residing in a low income community in Belo Horizonte, Southeast Brazil; 2) to investigate the extent to which socio-demographic factors, quality of life and physical and mental health explain this performance; and 3) to estimate the prevalence of disorders in the ability of language comprehension based on the cut-off points set by De Renzi in 1978.I Methods: We selected 405 people aged 60 or more residents in an area located in the southwest of the city of Belo Horizonte. The participants were 373 older people of whom 333 accomplished the evaluation test for language comprehension. The interview included sociodemographic and health factors, the Mini-Mental test, the shorter version of Token Test for language comprehension, the quality of life questionnaire SF-12, and the questionnaire for common mental disorders GHQ-12. The performance in the shorter version of the Token Test was described using measures of central tendency and the chi-square test in men and women according to age and education. Multiple linear regression analysis was used to investigate the association between these variables and the final score in the language comprehension test. The prevalences of mild, moderate and severe changes in language comprehension were estimated for all participants by gender, age and schooling. Results: The mean scores in the shorter version of the Token Test was 23 out of 36 points (standard deviation: 6.4) and did not vary between men and women, increased with education and reduced with increasing age. The final model of multiple linear regression included age, education, knowing how to read and write, cognitive status, hypertension, self evaluation of degree of comprehension in conversational situations, and the mental component of SF12 scale for quality of life. Together, these factors explained 62% of the variability in final score of the language comprehension test in the elderly participants. The prevalence of changes in the language comprehension in the population was 69%, 22% were classified as mild, 35% as moderate and 13% as severe. Conclusion: The results confirm that language comprehension is a complex and multidimensional ability, influenced by factors related to the background, the general health and subjective conditions. They also suggest that disorder in language comprehension is a frequent problem, largely preventable, which deserves further investigation among Brazilian elderly. Moreover, the cut-off points used to discriminate the presence and severity of language disorders based on the Token Test need to be validated for a population with lower education, as the participants in the present study. / Introdução: A compreensão da linguagem oral é uma habilidade complexa, e ainda não foram estabelecidos fatores associados a essa habilidade, tais como a idade, condições gerais de saúde, depressão e qualidade de vida. Objetivos: 1) descrever o desempenho de idosos na versão reduzida do Teste de Token para compreensão oral em uma população idosa residente em uma comunidade de baixa renda em Belo Horizonte, Minas Gerais; 2) investigar em que medida fatores sócio-demográficos, a qualidade de vida, e a saúde física e mental explicam este desempenho; e 3) estimar a prevalência de alteração na habilidade de compreensão oral com base nos pontos de corte estabelecidos por De Renzi em 1978.I Métodos: Foram selecionados 405 indivíduos com 60 anos ou mais residentes na área de abrangência de um Centro de Saúde, localizado na região sudoeste da cidade de Belo Horizonte. Participaram do estudo 373 idosos dos quais 333 realizaram o teste de avaliação da compreensão da linguagem oral. A entrevista incluiu fatores sócio-demográficos e de saúde, o Mini-exame do Estado Mental, a versão reduzida do Teste de Token para compreensão oral, o questionário de qualidade de vida SF-12 e o de transtorno mental comum GHQ-12. O desempenho na versão reduzida no Teste de Token foi descrito utilizando medidas de tendência central e o teste qui-quadrado para a pontuação final em homens e mulheres e segundo a escolaridade e faixa etária. Utilizou-se o modelo de regressão linear múltipla para investigar a associação entre essas variáveis e o desempenho no teste de compreensão oral. A prevalência de alterações da compreensão oral foi estimada para o conjunto dos participantes, por sexo, faixa etária e escolaridade. Resultados: A média da pontuação na versão reduzida do Teste de Token foi de 23 em 36 pontos (Desvio Padrão: 6.4), não variou entre homens e mulheres, aumentou com a escolaridade e reduziu com a faixa etária As variáveis que se mantiveram associadas diretamente à pontuação final no Teste de Token, no modelo final de regressão linear, foram a idade, escolaridade, saber ler e escrever, estado cognitivo, hipertensão referida, autopercepção sobre compreensão de situações conversacionais e componente mental da qualidade de vida. Juntos, estes fatores explicam 62% da variabilidade da compreensão oral nos idosos participantes. A prevalência de alteração na compreensão oral na população estudada foi de 69%, sendo 22% de transtorno leve, 35% moderado e 13 % severo. 30% por cento dos idosos não apresentaram nenhum comprometimento da compreensão. Conclusão: Os resultados confirmam que a compreensão oral é uma habilidade complexa e multidimensional, sendo influenciada por fatores relacionados ao background, condições gerais de saúde e conceitos subjetivos. Sugerem também que este é um problema freqüente, em grande parte evitável, e que necessita ser melhor investigado em idosos brasileiros. Além disso, os pontos de corte utilizados para discriminar a presença e gravidade do transtorno na compreensão oral no Teste de Token necessitam ser validados e adequados a uma população com baixa escolaridade, como a participante no presente estudo.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O objetivo desta tese é enriquecer o campo do trânsito religioso investigando a associação da religião com a saúde das pessoas e com seus hábitos de vida principalmente o cigarro. A tese foi dividida em duas partes: a primeira visa identificar a associação entre a autopercepção da saúde, a religião e o trânsito religioso. A segunda entre religião, trânsito religioso e o hábito de fumar. Para tanto, foram analisados dados transversais do Estudo Pró-Saúde realizado no Rio de Janeiro-RJ no ano de 1999. As religiões foram categorizadas de acordo com os critérios do Instituto Brasileiro de Geografia e Estatística (IBGE) e o trânsito religioso derivou da comparação entre religião de criação e religião relatada em 1999. Os resultados evidenciaram que 62% dos participantes mantiveram-se na religião de criação, 26% mudaram de religião e 12% mudaram para sem religião. O trânsito religioso foi marcado por um crescimento de kardecistas e do grupo sem religião. As pessoas que perceberam a sua saúde regular ou ruim apresentaram chance 40% mais elevada de ter mudado de religião, quando comparadas àquelas que a perceberam como boa ou muito boa (artigo 1). A maior parte das religiões apresentaram-se negativamente associadas ao consumo de cigarros quando comparados às pessoas sem religião ajustadas por variáveis sociodemográficas, relacionadas à saúde e transtorno mental comum. Os pentecostais e protestantes históricos apresentaram uma maior associação negativa com o consumo de cigarros e apenas a religião afro-brasileira apresentou uma chance mais elevada de consumo. As pessoas que mudaram de religião apresentaram uma chance 40% mais elevada de ser um ex-fumante quando comparadas a quem não mudou de religião (artigo 2). Para esclarecer as associações observadas na presente tese, é necessário a realização de estudos posteriores com emprego de outras metodologias, especialmente com o delineamento longitudinal.