A child-oriented corpus-based study of semantically related verbs using sketch engine

Laburpena

This paper presents a corpus-based study on four pairs of high-frequency verbs: say and tell, do and make, see and look, and go and come, which are semantically related but do not show the same grammatical or collocational behavior, and therefore can easily be confused in their usage by Spanish young learners (YLs) of English. The study is child-oriented in two senses: First, the corpus used, a corpus of Children’s Short Stories (CSS), was compiled with texts addressed to children, so they display specific linguistic features. Secondly, the research has been conducted with YLs in mind, so the choice of the pairs of common verbs studied is based on their difficulty for Spanish-speaking YLs and the data-driven learning (DDL) activities proposed are directed to YLs (aged 10-12). The methodology comprises the selection of verbs, their grammatical patterns, and an inventory of the strongest collocates for each category, based on word sketches and the typicality score (Log Dice) provided by Sketch Engine (SE) text analysis software. By identifying the most frequent semantically closely related verbs and their use in grammatical and phraseological structures, and by providing EFL student teachers with this information, awareness can be raised about a) how to find vocabulary especially worth teaching in primary education because they are high-frequency words, simple in meaning but difficult in usage, and b) how to teach these words, including direct and indirect corpus-aided language learning, like DDL, simplified and adapted for YLs, and the inclusion of authentic examples in the teaching materials.