Movimentos oculares e prosódia de leitura oral: análise dos marcadores prosódicos gráficos na leitura de alunos do 5 º ano do Ensino Fundamental

Abstract

The present Doctoral Thesis investigates the processing of Prosodic Graphic Markers - MPGs (PACHECO, 2003) in read-aloud. It aims to examine the processing of oral reading in Brazilian Portuguese (BP) taking into account the MPGs: Comma (VG), Period (PT) and Colon (DP) according to the BP intonational system defined by Cagliari (1981), which presents tonal patterns delimited by a Tonal Group (GT) consisting of a Salient Tonic Syllable (TT), Tonic Component (CT) and Pretonic Component (CPT). The methodology used combines acoustic and cognitive analysis. For the acoustic analysis, we recorded the oral production during reading of 58 5th grade students from the municipal school in Ponta Grossa (PR), all with ages between 9 and 12 years old, classified around the Reading Fluency variable, with a total of 29 students in the Fluent Group (FL) and 32 in the Low Fluency Group (FP). In all, 928 data were analyzed (58 students x 16 target sentences). Variables were controlled for taking the measures of intonational contours: Intensity (in dB) and Fundamental Frequency (f0) (in Hz) of CPT and CT, Duration (in ms) of TT and Pause (in ms). For the cognitive analysis, we employed the Eye Tracker Technique to take measurements of the eye movements of the 58 participants. The eye variables controlled were: Total Reading Time (TTL), Number of Fixations (NF), Number of Saccades (NS), and Mean Fixation Time (MTF). For the relationship between ocular and acoustic data, we applied tests combining the statistical variables and the stimuli: Test 1 (the number of fixations and duration of the internal pause), Test 2 (the number of revisits and duration of the internal pause), Test 3 (the number of fixations and duration of the final pause), and Test 4 (the number of revisits and duration of the final pause). The data was composed of three stimuli, corresponding to the reading of three different types of texts: Text 1, complex with the markers DP (2 target sentences), VG (3 target sentences) and PT (3 target sentences), Text 2, simple with the markers VG (3 target sentences) and PT (1 target sentence) and Text 3, simple without punctuation, containing 4 target sentences: i) 3 target sentences with VG value and ii) 1 target sentence with PT value. Data processing comprised the following steps: i) eye data extraction in BeGaze software; ii) segmentation of audio signals in Praat software; iii) extraction of the means of the physical parameters of the segments; iv) tabulation of the acoustic and eye relationship data; v) application of quantitative tests. In the prosodic analysis, the participants present in all the markers evaluated, inconsistency regarding the reduction or maintenance of CT and CPT in f0 and Intensity and no elongation of TT. The comparison of the averages of the eye variables between the FP and FL groups proves that the more complex the text, the greater the difference between the fluency groups. The results of the correlation analysis indicate that there is a significant correlation between the variables internal pause and fixations in test 1 (in texts 1 and 2) and internal pause and revisits in test 2 (in texts 2 and 3), which characterizes difficulty in reading processing due to internal pauses, revisits and fixations within the sentence; and between the variables End Pause and Fixations in test 3 (in text 1) and end pause and revisits in test 4 (in text 2), which means that when the duration of the amount of fixations and revisits increases, there is an increase in the duration of the end pause, which may prove the dwell-time effect (HIROTANI; FRAZIER; RAYNER, 2006; RAYNER, 1998), in which readers remain at the end of a clause until the resolution of the present clause. Overall, the data showed that there is a strong correlation between fixation, revisit, and pause duration in the processing of punctuation marks during spoken reading.