UNIVERSIDADE ESTADUAL PAULISTA
"JÚLIO DE MESQUITA FILHO"

Faculdade de Ciências e Letras - Campus de Araraquara
  Agenda Pós-Graduação - Lingüística e Língua Portuguesa

Aluno(a) Mariane Carvalho
Titulo A VARIAÇÃO MELÓDICA DA FALA: um estudo da correlação entre uma análise acústica e uma análise auditiva da entoação
Orientador(a) Prof. Dr. Luiz Carlos Cagliari
Data 27/02/2013
Resumo RESUMO

O objetivo desta dissertação de mestrado é estudar alguns aspectos da correlação entre dados acústicos e dados obtidos através de modelos de análise auditiva da entoação, a fim de verificar se a análise acústica se distancia ou não do reconhecimento auditivo. Para isso, investigamos, através de alguns dos principais métodos de descrição da entoação, do ponto de vista perceptivo (fonologia funcional) e do ponto de vista acústico (modelo autossegmental), qual a relação entre a percepção e a realidade física do som na análise do fenômeno da entoação. Nossa intenção é provar que essa correlação é tanto possível quanto viável. Para alcançar esse resultado, fizemos, primeiramente, uma descrição individual das duas abordagens eleitas para essa comparação. Uma das abordagens, diz respeito ao estudo auditivo desenvolvido por Halliday (1970) e adaptado por Cagliari (1982/2007) para a análise dos dados do português brasileiro. A outra é de base acústica e segue o modelo de análise autossegmental, desenvolvido por Pierrehumbert (1980). Feito isso, propomos uma comparação entre os dois modelos, trazendo à baila suas semelhanças e diferenças, com o intuito de chegarmos a um possível paralelo entre eles. O resultado foi satisfatório e mostrou que, apesar de formalmente diferentes, é possível constatar semelhanças entre eles. Partindo desse princípio, fizemos uma detalhada análise acústica por meio do programa computacional PRAAT e, auditiva, (seguindo a metodologia de Halliday e Cagliari). A análise descreve a estrutura fonológica da entoação de um pequeno trecho do livro História sem fim (sd). A gravação foi feita por um informante adulto da cidade de Araraquara. O resultado mostrou, mais uma vez, uma correspondência entre as formas fonológicas, e provou a aproximação das análises acústica e auditiva. Além disso, também foi nosso objetivo mostrar que a correlação entre as duas análises (acústica e auditiva) permite uma melhor explicação da variação melódica da língua, bem como sua correlação com outros níveis da gramática (sintaxe, semântica e pragmática). Com relação a isso, concluímos que a entoação só é descrita corretamente quando os fatores sintáticos e semânticos são levados em consideração, pois os padrões entoacionais revelam as intuições dos falantes e ouvintes, sendo parte do componente fonológico da gramática da língua.

Palavras-chave: Análise acústica. Análise auditiva. Fonologia entoacional. Halliday. Pierrehumbert.













ABSTRACT

The aim of this dissertation is the analysis of the correlation between acoustic and auditory data, in order to verify whether the acoustic analysis follows or not the auditory recognition. For this, we investigated, with some of main methods of description of intonation, the perception (functional phonology) and the acoustics (autossegmental model), what is the relation between perception and physical reality in the analysis of intonation. Our intention is to prove that this correlation is as much possible as feasible. We started with the description of the two approaches chosen for our comparison. One of them is concerned with the auditory approach. This theory was developed by Halliday (1970), according to Cagliari (1982/2007), who adapted Halliday’s model to describe the intonation of Brazilian Portuguese. The other, is the acoustic approach developed by Pierrehumbert (1980) autossegmental approach. We propose a comparison between the two theories, bringing up their similarities and differences in order to reach a possible comparative parallel between them. The results showed that, despite the formal differences, similarities between the two approaches. Based on this idea, we made a detailed acoustic analysis of our data with the software PRAAT and an auditory analysis of the same data (following the Halliday’s and Cagliari’s methodology). The analyses describe the phonological structure of intonation of a short passage from the book História sem fim (The Neverending Story (sd). The recording was performed by an adult informant from the city of Araraquara. Once more, the results showed a good correspondence between the phonological forms merged from the acoustic and auditory analyses. In addition, our goal was to show that the correlation between the two approachs (acoustic and auditory) allows a better explanation of language melody variation, as well as it relation with other levels of the grammar (syntax, semantics, and pragmatics). In this regard, we conclude that the intonational phenomenon is better described only when the syntactic and semantic factors are considered, because intonational patterns reveal the intuitions of speakers and listeners, being part of the phonological component of the grammar of the language.
Keywords: Acoustic analysis. Auditory analysis. Intonational phonology. Halliday. Pierrehumbert.
Tipo Defesa-Mestrado
Texto Completo

APG 2.0
Copyright 2014 (c) UNESP - Faculdade de Ciências e Letras do Campus de Araraquara