Corpus building and the use of corpora at lower levels of language learning

Abstract

The use of linguistic corpora is not a common practice in the teaching of Hungarian as a foreign language. Although the use of these large text collections has a number of benefits of which the present article can present only a few, teachers or students are often reluctant to consult them. The principal reason for this seems to be that most existing corpora were built with the expectations and aims of linguists and researchers in mind. This article argues that corpora can only be relevant and effective for language teaching purposes when they are built in accordance with the learners’ needs. An example of pedagogical databases is the MagyarOK corpus for Hungarian, available among the public corpora on the Sketch Engine website. In the second half of the article, the content of this corpus and the methodological considerations behind its construction are briefly described.