Estudo acústico comparativo da pronúncia de alunos de língua inglesa com falantes nativos americanos.

Abstract

In this dissertation, we aimed to develop an acoustic-comparative analysis between English students from Núcleo de Línguas (NL) of Federal University of Espírito Santo (UFES) in Vitória - ES and native American speakers from Minnesota-USA, to verify their productions of high vowels: [i:] - [ɪ], [u:] - [ʊ], in minimal pairs, in order to obtain a diagnosis about the oral production of each phoneme, as well as, the probable similarities and sound differentiations between the vowels and their pairs, through the ascertainment of frequency and duration values. The acoustic study of the sound productions was based on the acoustic theory of speech production developed by Fant (1970), called "Linear Source-Filter Theory" (KENT, READ, 2015, p. 16), which allowed understanding the acoustic-articulatory relations and the fundamental procedures for an acoustic analysis of speech production. The acoustic characteristics of the high vowels, produced by the subjects in this research, are evaluated through parameters of vocalic quality, which allow us to distinguish measurements in the duration of the segments in time per second, as well as in the F1 and F2 values (in Hertz), which differ between two studied languages. These frequency values make it possible for us to verify, through F2, the forward/backward movement and, through F1, the height of the tongue body during the articulation of the segments, as well as to measure the vowel sustain time in the sound production range which, in LI, are markedly different between the minimal pairs. Regarding the research methodology, five (5) female subjects were invited: two Americans, who were born in the United States of America, in the state of Minnesota, and three students from NL(UFES). All this sound material was recorded in a professional studio, in order to ensure accuracy of the acoustic data for delimiting the phonetic competence value for frequencies and duration. We did not use, for data analysis, each absolute value of formants and time in the range of high vowel production, but the average value calculated by the Praat software (BOERSMA, WEENINK, 2021) itself. In the discussion of the results, we addressed the detailed analyses of each minimal pair through the vocalic productions of each research subject, graphically demonstrating the sound phenomena separately, with the intention of visualizing the numerical frequency data. We consider that this comparative study in acoustic phonetics could highlight specific events of speech sounds focused on the production, combination, description and representation of sound data produced by the corpus, contributing to reflections regarding the formal teaching of pronunciation to English learners.