A partícula -se nos contextos de construção da voz média e da voz passiva sintética em português brasileiro: um estudo experimental

Abstract

In linguistic studies, there is a deep discussion regarding the concept of middle voice (or medial voice) in Brazilian Portuguese (BP). Parallel to this, there is normative grammar, which proposes a simplified scheme of just three voices, in which the subject appears represented as an agent, in the active voice, as a patient, in the passive voice, or as an agent and patient at the same time, in the reflective voice. This study presents, with theoretical foundation in the words of Camacho (2003), a construction characterizing the middle voice in BP that occurs through the use of the clitic -se. This construction also applies to the synthetic passive voice, which means that there is a formal similarity between the cases, in addition to making it possible for an attribution of meaning to occur that is not tied solely to a single linguistic aspect – syntactic or morphological, for example. Therefore, this analysis aims to observe how the relationship between this clitic occurs in the formation of these voices, as well as analyzing the factors that can contribute to the sentence being interpreted as passive or as average. The analysis took place experimentally through the online technique of self-paced reading of sentences that included the two structures of verbal voices, observing the influence of the type of subject - inanimate or animate - on the interpretation of the constructions. Two experiments were carried out with the same sentences, but with different questions. In the first experiment, the question was just a probe, to maintain attention on the sentences. In the second, the subject chose a more average or more passive option, in the second, an interpretation question to check voice preference. The experiments were applied to adult individuals, with complete or incomplete higher education. It was expected that verbs falling into the passive voice category would be more easily processed and would have more preference in interpretations to the detriment of the middle voice category, as this category seems to require more effort at the time of processing. However, the results showed passive voice verbs obtaining greater reading time in the critical segment. In terms of response time to the question, the passive voice, as well as the animated subject, had shorter reading times. With the results of this research, we intend to contribute to the conceptualization of average voice in BP, in addition to suggesting ways to better differentiate between the two categories analyzed.