Modalidade em perspectiva: estudo baseado em corpus oral do Português Brasileiro

Abstract

This study deals with the description of modality in a corpus of spontaneous Brazilian Portuguese spoken language, more specifically, the Minas Gerais variety. Modality, in a more concise definition, is the evaluation of a conceptualizer, which relativizes the uttered locutivematerial in terms of degree of certainty, possibility, necessity, capability, and volition.However, there is no consensus on a precise definition of such category, explained by a number of reasons: (a) modality is a field of study for logics and linguistics that utilizes various methodologies; (b) the categories is both differentiated and interrelated to others, suchas time, aspect, mode; (c) this notion is mistaken and overlaps those of illocution, attitude, and emotion. Based on the principles of the Speech Act Theory, in Corpus and Computational Linguistics, this work aims at describing the behavior of different modality markers and proposing a modality annotation scheme for a C-ORAL-BRASIL minicorpus. CORAL-BRASIL is a corpus of Brazilian Portuguese spoken language, the fifth branch of CORAL-ROM, a comparable corpus, representative of the four main European Romance Languages (Italian, Spanish, French, and Portuguese), prosodically segmented in utterances and their informational sub-units. The utterance is the minimal reference unit pragmatically interpretable. To better fit the purposes of this research, I utilized the informal part of the CORAL-BRASIL corpus, a minicorpus composed of 20 texts of three interactional typologies,divided into private and public: 7 monologues 6 private and 1 public ; 7 dialogues 5private and 2 public ; and 6 conversations 4 private and 2 public ones. For the treatment in this thesis, 1.197 modal indexes, distributed in 1.046 utterances, codified and, later on,semantically annotated through the MMAX2 software were utilized. After the quantitative and qualitative analysis, we can conclude that (i) the category of modality behaves distinctively in written and spontaneous oral language; (ii) the modal indexes are widely utilized in dialogic situations, be them public or private; (iii) only a few informational units can be modalized: Comment, Parenthetical, Topic, Locutive Introducer; (iv) the Comment unit is the most modalized and there is no realization restriction for this IU; (v) the epistemic modality is the most frequent one among the types of meaning; (vi) verbs are the preferred strategy for modality marking.