Processamento de ambiguidade lexical em português brasileiro PB: um estudo psicolinguístico

Abstract

This study aims to investigate, in light of Experimental Psycholinguistics, whether there are differences in the processing of lexical ambiguity, with regard to the typologies that constitute it (Cançado, 2018), that is, whether homonymous and polysemous words would present homogeneous or heterogeneous processing data. This doctoral thesis seeks to contribute new evidence to Experimental Psycholinguistics and the Study of Semantics. Broadly speaking, lexical ambiguity refers to the dual possibility of interpreting words and is constituted by two semantic processes: homonymy and polysemy. The study's hypothesis is that the processing of lexical ambiguity occurs differently between the typologies, with homonymy being a less costly semantic process than polysemy, due to the strong immediate semantic commitment (Frazier; Rayner, 1990, 1999). Furthermore, it is expected that the type of meaning (dominant versus subordinate) will also act as a main effect in the processing. The methodology used in the study was divided into two stages. Initially, a normative study, with an experimental design 2, with Typology (Homonymy versus Polysemy) as the independent variable and the percentage of occurrences of sentences with one meaning or another as the dependent variable. The task consisted of an elicited production, via Google Forms, which sought to verify which meanings of the ambiguous words would be classified as dominant (higher percentage values) or subordinate (lower percentage values). In the second stage, the studies were conducted online, using the Eyelink 1000 eye-tracking device, in order to verify the real-time processing of the words. Two experiments were conducted using the eye-tracking technique and a forced- choice task, and the difference between them lies in the syntax of the sentences containing the experimental items. Experiment 2 presented the disambiguating context after the critical element, and Experiment 3 presented it before. Both experiments had two types of analysis, one with an experimental design 2 and one with an experimental design 2x2. In the experimental design 2, there was one independent variable (Typology: Homonymy versus Polysemy) and several online and offline chronometric and non-chronometric measures. The dependent variables analyzed were IA_Dwell_Time, IA_First Fixation, Sentence_RT, among others. The 2x2 experimental design included two independent variables (Typology: Homonymy versus Polysemy; Meaning: Dominant versus Subordinate) and the same dependent variables. The empirical evidence suggests that the processing of ambiguous words shows differences regarding typology, whether homonymy or polysemy, but the data are not conclusive regarding the less costly semantic process. Furthermore, the results suggest that the syntactic position of the disambiguating context and the type of meaning (dominant versus subordinate) evoked influences the processing of sentences with ambiguous words.