Tendências prosódicas de construções de discurso reportado em fala espontânea

Abstract

This thesis claims the postulation of prosodic aspects in formal pairing of grammatical constructions instanced in speech modality. According to Construction Grammar theoretical bases (GOLDBERG 1995, 2006), it is argued that grammatical constructions are systematic collections of linguistic form-function pairings learned in everyday language and therefore they can be mapped to real and contextualized speech data. Specifically, the goal is to describe and analyze prosodic tendencies in reported speech constructions of spontaneous speech corpus samples of Brazilian Portuguese (Rio de Janeiro – standard dialect NURC/RJ). By the finding that the prosody adds information to the strictly linguistic content of the message, other theoretical contributions are increased about the melody of speech, as Grice and Baumann (2007), Madureira (2005) e Pierrehumbert (1980). The reported speech is discussed based on Rocha (2004), which pointed out the contribution of prosody in distinguishing constructional types of reported speech. Occurrences of discursive phenomenon are subjected to PRAAT program (BOERSMA & WEENINK, 2011), by which prosodic information is analyzed and correlated to the systematization and processing of the data. The results show that there is prosodic information such as related to boundary tone, average fundamental frequency (F0) and the rate of change of F0, which may be considered tendencies for the construction of reported speech in general. Based on the ToBi notation system (Tones and break indices), the tendencies are: the introducer unit reported speech ends with a boundary tone L%, preceded by the pitch accent H * + L; the measure of the average F0 introducer unit, as compared to the measure of average F0 reported unit, is higher; the measure of the rate of change of F0 introducer unit, as compared to the measure of the rate of change of F0 reported unit, is lower. These results were achieved by an intra- and cross-speakers analysis. In addition, when adopting the distinguish constructional types of reported speech, it appears that the direct speech with dicendi verbs tends to receive one or more emphasis; the direct speech without dicendi verbs tends to have a pause between the introducer unit and the reported unit; the indirect speech has no pause separating the two units; the free indirect speech displays the highest average F0 in reported unit and not in ―introducer‖ unit.