Análise multidimensional de um corpus multinível de aprendizes de Língua Inglesa

• Autor(a) principal: Silva, Cícero Soares da
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da PUC_SP
• Texto Completo: https://repositorio.pucsp.br/jspui/handle/handle/26500
• topic: CNPQ::LINGUISTICA, LETRAS E ARTES::LINGUISTICA::LINGUISTICA APLICADA; Linguística de corpus; Níveis de proficiência; Análise multidimensional; Corpus linguistics; Learner corpus research; Profiency levels; Multidimensional analysis

The goal of the present research was to identify the dimensions of variation from patterns of co-occurence of lexicogrammatical elements of the langauage used by English learners. The Multilevel Corpus of Brazilian Learners of English as a Foreign Language - COBRA-7 was used. Compiled by W.M. Dantas in 2012 (DANTAS, 2012), the COBRA-7 corpus comprises 2516 school essays of different topics, totalling 571,564 words (tokens) and 19,800 types, distributed in six proficiency levels: basic I, basic II, pre-intermediate, intermediate, high-intermediate and advanced based on the Common European Framework of Reference for Languages (CEFR). Learner Corpus Research (LCR) is aimed at investigating electronic collections of authentic textual data produced by Foreign Language or Second Language learners. Data are gathered according to explicit design criteria for a particular purpose for Foreign Language Acquisition/Foreign Language Teaching. The data are coded using objective criteria such as text source, learnel level, L1 background, among others (GRANGER, 2004, p.4). The methodology used in this research was Functional Multidimensional Analysis (MDA) (BIBER, 1988; 2009; BERBER SARDINHA, VEIRANO PINTO, 2014, 2019), a corpus-based method. MDA can reveal the underlying parameters of variation in a corpus, that is, the grammatical features co-occurring in the texts. The analysis carried out sought to reveal the functional dimensions of variation present in the COBRA corpus. Through a series of ANOVAs, it was possible to determine to what extent the dimensions are good predictors of the variation across independent variables such as “school”, “level” and “gender”. The results showed significant statistical variation exists between learning levels with respect to the dimensions. It is hoped that this study will stimulate the development of other studies using Multidimensional Analysis (MDA), mainly in the field of Learner Corpus Research (LCR) especially from na MDA perspective