A distribuição dos nomes repetidos, dos nomes recategorizados, dos pronomes plenos e dos pronomes nulos em função anafórica em Português Brasileiro

Abstract

The main objective of this dissertation is to explain how repeated names, recategorized names, overt pronouns and null pronouns are distributed in the co-referencial process, filling anaphoric centers in the subject position in Brazilian Portuguese (BP). The study was based on the Centering Theory (GROSZ; SIDNER, 1986; GROSZ; JOSHI; WEINSTEIN; 1995) and on the works that explain the effects of processing coreferencial forms, such as the Repeated Name Penalty (GORDON; GROZS; GUILLIOM, 1993) and the Overt Pronoun Penalty (GERLOMINI LEZAMA, 2008). The working hypothesis was that, in some contexts, the repeated name would be the most frequently used form, and this would be the evidence that this form is more easily processed than the overt pronoun or the null pronoun. Data were collected from the analysis categories which were based on the main psycholinguistic effects on anaphoric processing. Two types of analyses were conducted for that purpose: (i) analysis of written corpus, containing texts written in the standard Portuguese language and (ii) analysis of speech corpus, containing texts spoken in the standard norm. The number of co-referencial forms in authentic use in the language was analyzed. In order to analyze the occurrences of these forms, the number of repeated names, recategorized names, overt pronouns and null pronouns in the subject position were counted, taking into account the distance between the co-referential expression and its referent, the type of sentence or paragraph in which the resumption occurs, the syntactic position of the referent, the type of verb in the sentence prior to the resumption and the type of article in the noun phrase when introducing the referent. The combined analyses of the results collected in both corpora suggest that there are contexts in which the repeated name is more frequently used, and that the corresponding form to be used is directly linked to two factors: the distance between the antecedent and the co-referent form in number of words and the textual arrangement in which the co-referent form appears.