Variação linguística e processamento: investigando o papel da distância entre sujeito e verbo na realização da concordância verbal variável no PB

Abstract

This dissertation investigates the processing of variable verbal agreement in Brazilian Portuguese, focusing on cognitive and perceptual aspects of this phenomenon. BP exhibits two general rules for number agreement: (i) redundant plural agreement both in subject and verb (As meninas brincam aqui) and, (ii) non-redundant agreement, in which one, the plural morpheme is mandatory in the subject (usually carried out in D) and may be omitted in the other elements (As meninaØ brincaØ aqui). Previous studies on variable agreement, developed in the context of Variationist Sociolinguistic, have listed a set of factors that would favor the occurrence of each of these rules. There are, however, gaps in the literature regarding the processing of this variation. The theoretical background of this research integrates the language model formulated by the Generative Theory, and processing models developed in the context of Psycholinguistics. In this dissertation we aim to evaluate: (i) to what extent speakers would process both rules equivalently? And (ii) what is the relevance of linear distance between subject and verb in the occurrence of these rules? Three experiments were conducted in order to explore these two questions. The first one, based on a self-paced listening task, aimed to investigate how the speakers deal with the two agreement rules. Experiment 2, conducted through an elicited production by repetition task, aimed to evaluate the role of linear distance in the realization of the agreement. The second goal of this experiment was to evaluate the possible correlation between linear distance and extra-linguistic variables, specifically, level of education. Finally, Experiment 3 – implemented by a self-paced reading "maze" task – is a follow up of the Experiment 2, and aimed to deepen the study of the role of linear distance. Taken together, the results suggest a differentiated treatment of the two agreement rules, with a higher processing cost in the case of morphologically non-redundant agreement. Regarding the role of linear distance between subject and verb, a greater distance between these elements favors the occurrence of the non-redundant rule, indicating that this factor would have an important role in the processing of verbal agreement. The results concerning to the level of education reveal subtle differences between the two groups investigated.