Investigação teórico-metodológica sobre a penalidade do nome repetido em português brasileiro

Abstract

This work investigated the anaphoric processing of repeated names and full pronouns in Brazilian Portuguese (PB), focusing on theoretical and methodological aspects related to the effect of Repeated Name Penalty (PNR), which consists of an increase in the cost of processing repeated names when compared with pronouns in the establishment of coreferencing in certain syntactic and discursive contexts. This effect was initially found and called the “repeated-name penalty” by Gordon et al. (1993), when conducting a study in English to test the prediction of the Centering Theory (GROSZ et al., 1983, 1995), which assures that the pronoun is the preferred anaphoric form to retake the most prominent antecedent in the discourse. In addition, the PNR was confirmed in other studies conducted in English (GORDON and CHAN, 1995; CHAMBERS and SMYTH 1998; KENISSON and GORDON, 1997) and in other languages such as Chinese Mandarin (YANG et al., 1999) and French (ERNST, 2007). In Brazilian Portuguese, there are differences in the PNR research results, based on the assumptions of Gordon et al. (1993), some attested to the occurrence of this penalty (LEITÃO, 2005; QUEIROZ and LEITÃO, 2008; LEITÃO and SIMÕES, 2011; GONDIM and LEITÃO, 2012; LEITÃO, RIBEIRO and MAIA, 2012; LIMA, 2014; BARBOSA, GONDIM and LIMA, 2016) and others pointed out that this effect does not exist in our linguistic system (MAIA and CUNHA LIMA, 2011, 2012; MAIA, 2013; LIMA, 2015). We observed that among these contrasting studies in PB there are differences in the experimental materials used that involve several factors, both linguistic and methodological. Based on this, we conducted a series of five experiments through the self-monitoring reading technique, manipulating, isolating / controlling and relating some factors, trying to observe if they influenced the occurrence of PNR in order to help clarify the divergences between PB studies. In this multifactorial perspective, in the experiments we manipulate the following factors: Type of Resumption (repeated name, full pronoun); Number of antecedents (one antecedent, two antecedents); Type of sentence (juxtaposed sentences, coordinated sentences), and we controlled the factors: type of final task (task that focused on the antecedent, task that did not focus the antecedent); Form of segmentation / measurement of reading time (stimuli divided into several segments with measurement of reading time only of the anaphoric element, stimuli divided into segments corresponding to sentences with gauging the reading time of the entire sentence that contained the anaphora). The results pointed to a possible interaction between these factors influencing the occurrence or not of PNR, strengthening our general hypothesis that this penalty is a multifactorial effect, which occurs due to the joint action of several factors.