Análise de redações em língua portuguesa a partir do perfil lexical : contribuições do processamento linguístico-computacional

Abstract

The writing of texts depends on the knowledge of the vocabulary that will be used, and the writing quality is referred to the ability to use various words in the productions. This work deals with the evaluation of texts written by high school students considering lexical analyses based on human correction and computational processing. Thus, the general objective of this work is to demonstrate the quality of the vocabulary contained in compositions through the lexical-grammatical profile drawn and the comparison between the human and linguistic-computational corrections made by software. With regard to specific objectives, the research aims to: constitute a córpus of high school compositions; propose a methodology for identifying the vocabulary quality of texts; draw a lexical-grammatical profile from the processing of the córpus; propose analyses based on readability metrics; to establish comparisons between the human correction of texts and the linguistic-computational. To support the theoretical foundation, this research will mainly use the works developed within the scope of Lexicology (ANTUNES, 2012; BARBOSA E SOUZA 2016; BIDERMAN, 1999, 2006; GUERRA; ANDRADE, 2012; GENOUVRIER, PEYTARD, 1985; ROCCO, 1981), From Córpus Linguistics (BERBER SARDINHA, 2004, 2010, 2012; BIBER; REPPEN, 2015; HALLIDAY, 1993; SINCLAIR, 2004) and of the Interculturality (BARBOSA,2009; BARBOSA; SOUZA,2016). On relation with the material, it was necessary a essay corpus from high school, a electronic form for the evaluation of the essays, A software that processes static linguistic: Wordsmith tools, a list of Portuguese words, the electronic dictionary Houaiss, a printer virtual Doro writer, a online static calculator and the sample of the calculation, regarding the procedures we proceed with a lexical way study on the essays and the correction of the texts with 10 Portuguese writing teachers. Using the lexical way of study, we understand the relative sum of two percentage complements: the one with a fundamental vocabulary and the one with out a fundamental vocabulary (a more formal vocabulary). The most formal vocabulary has a variation of 6.42% to 25.29%. From the 10 teachers participating, 7 was influenced by the presence of a more formal vocabulary on the essays. That indicates a strong positive correlation on the evaluation on who’s correcting it. The results shows that the Portuguese teachers and writing teachers needs to pay attention on the teaching of vocabulary having in view the vocabulary choices and the increased of lexical repertoire.