A influência da retomada e da distância sintática no processamento de pronomes plenos e nulos em português brasileiro

Abstract

In this dissertation we investigated how repeated names, overt pronouns and null pronouns are corefentially processed in Brazilian Portuguese (PB). Gordon, Grosz & Gilliom (1993) have shown that repeated names are more difficult to be processed compared to the overt pronouns vying for the subject anaphoric position of an entity previously introduced in discourse, creating an effect that it has appointed of "repeated name penalty "(RNP). In turn, Gerlimini-Lezama (2008) tested the same type coreference to expand the theory for the use of the null pronoun in Spanish, which is a pro-drop language, unlike English, and in turn, has shown that, at least this language, null pronouns are more easily processed than repeated names, creating the effect that he named "overt pronoun penalty" (OPP). In PB, RNP has been the subject of discussion among many scholars and quite reaffirmed by Leitão (2005) and colloboraters (Queiroz & Leitão, 2008; Leitão & Simões, 2011, inter alia). Taken together, these studies have corroborated acceptance of the effect of RNP in PB not only in the subject position, but also the object position. However, Maia & Cunha Lima (2011, 2012) presented evidence contradictory to previous studies on the RNP in PB, indicating the occorence of OPP in PB. We work with the hypotheses that the distance has a significant role during processing, giving the name repeated the status of reference expression at the expense of overt and/or null pronouns for both simple subject and composed subjects. The analysis of the experimental results suggest that (i) at the critical region the null pronouns are processed more easily than overt pronouns and repeated names, and (ii) the distance between the antecedent and its referring expression plays no significant role in the coreferential processing.