COMPREENSÃO LEITORA E A GERAÇÃO DE INFERÊNCIAS EM LÍNGUA ESTRANGEIRA: LEITURA NO PAPEL X LEITURA NA TELA DIGITAL

Abstract

Abstract: Within the perspective of Psycholinguistics studies, the investigation of inference generation in reading comprehension is considered rather recent. Even though there are several studies which investigate inference generation, this does not indicate that the possibilities of research in this area are done. The proof is the present research, of qualitative and quantitative nature, which verifyed the differences in the process of reading comprehension of texts in English in two different modes of presentation: paper and digital screen. The participants in this study, carried out in October, 2016, were five undergraduate proficient English students, from 2nd to 4th grade in the English major, of a public university located in the central-south region of Paraná. In order to accomplish our main goal, an argumentative text was read on paper and an expository text was read on digital screen. The different text types have been chosen in order to verify whether the text types influence (CALDART, 2012) the inference making process. Considering the study of Baretta (2008), the pause protocols (CAVALCANTI, 1987; 1989) – which consist of verbalizing everything that comes to our minds while reading – were used, in the version adapted by Tomitch (1995). After being transcribed the pause protocol reports made it possible to ascertain the role of online inference generation in reading. We triangulated the data from the pause protocols with the participants‘answers to comprehension questions, retrospective questionnaire and reader‘s profile, so as to confirm or refute our hypothesis. We observed that our data indicates subtle differences in the inferential process not being those exactly related to the modes of presentation, but due to the (unexpected) behavior of the readers while reading in the digital support. Thus, corroborating the study of Caldart (2012), we observed that what may have influenced the inference generation of the participants, was the different text types used. Furthermore, we found that the participant who generated the most inferences during reading did not obtain the highest scores in the comprehension tests, as well as the participant who infered the least did not obtain the lowest ones, as hipothesised. We also noted that other processes, such as off-line inferences, may have influenced the comprehension of the five participants, a fact that does not bring clear results to our research, but possibilities for further research.