Investigando colocações em um corpus de aprendiz

• Ano de defesa: 2014
• Autor(a) principal: Valdenia Carvalho e Almeida
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Colocações; Linguística de Corpus; Corpus de aprendiz; Língua inglesa Estudo e ensino Falantes estrangeiros; Lingua inglesa Gramatica; Corpora (Lingüística); Linguística Processamento de dados; Linguística de corpus
• Link de acesso: http://hdl.handle.net/1843/MGSS-9PHMJ6

The aim of the present study was mainly to investigate how Brazilian learners of English use collocations with the high frequency verbs get, give, make and take, such as get a job, give a chance, take advantage, when compared to native speakers of English. For this purpose, the study was based on Corpus Linguistics, a study area which collects and explores electronic corpora with the aim of studying language in its natural context of use by using computational tools. More specifically, the research presented here was based on the area of learner corpus linguistics. A learner corpus is a collection of systematic and computerized texts produced by learners of a language. As this kind of corpus allows the access to learners's interlanguage it is frequently used for the purpose of descriptive studies, with the aim of identifying aspects of language use in which learners liken or differ from native speakers. In other words, learner corpora allow a deeper knowledge of the learners's interlanguage and can contribute to the area of language learning and teaching. The analysis of the interlanguage can reveal some of the learners's linguistic difficulties and, once aware of their needs teachers can better plan and define their pedagogical actions. Several studies from the area have its origin in the ICLE project (International Corpus of Learner English). The study investigated the learner corpus Br-ICLE (Brazilian Subcorpus of the International Corpus of Learner English) with 159,000 words, and two general corpora, BNC (British National Corpus) with 100 million words and COCA (Corpus of Contemporary American English) with 425 million words. The computer tool AntConc 3.2.4 (ANTHONY, 2011) was used to extract and analyze the data. By using this tool the collocations were identified and analyzed. Besides investigating how Brazilian learners use collocations the research created some activities to promote linguistic awareness of the use of collocations. These activities were done by some university students of English. The results indicate that learners have problems producing and using the collocations studied. They often make mistakes related to the verb and its collocates and to other lexical items, such as prepositions and determiners. The results also showed some inadequacy of academic vocabulary of Brazilian learners. They often do not use collocations that are typical of the academic register, which may be explained by some deficient academic vocabulary. To compensate it, Brazilian learners frequently use a high-frequency verb where an academic verb was more appropriate (make tasks instead of perform tasks). Based on the results, language teaching should place some emphasis on the lexical and grammatical patterns of a language. Based on the deviations identified in the Br-ICLE corpus and on the principles of Data driven learning (DDL) we created some tasks to promote language awareness. Besides the activities to promote language awareness, the study also brings the lists with all the collocational deviations and a list with academic collocations that appear to be problematic for Brazilian learners. The pedagogical implications of the study indicate that teachers should promote learners's linguistic awareness of collocations and thus, learners should notice that learning vocabulary is not only learning new words but to become familiar with their possible combinations (WRAY, 2002).