Abstract

How can the effects of corpora on the language learning process be effectively assessed? This is an old question which is, however, just as important now as it was ten or twenty years ago. Does corpus use aid students in a measurable way? How does corpus use affect students’ subsequent language use? Is there a marked qualitative difference between the work of students who make use of corpora in their studies and those who do not? Does corpus use necessarily lead to an improvement in students’ language production? This research reports an ongoing study into phraseological production in advanced learner writing (Philip 2005a; 2005b; 2006/forthcoming); in particular, it investigates corpus use and non-use in a semi-structured writing task. Unlike other studies in the EFL literature, this research examines multiple versions of a single base text, making it possible to observe what effect the use of corpora has on the production of phraseological units in text, rather than in gap-fill sentences or other, more traditional types of elicitation experiment, such as those found in Deignan et al. (1997 and Boers (2000). The rationale behind the present approach is influenced by the use of parallel corpora in translation studies in which multiple versions of single texts – in two or more languages – can be visualised and examined as KWIC concordances. The data presented here has been gathered from five groups of advanced learners between March 2005 and June 2006. The learners, following advanced (C1) general English courses at the University of Bologna. Four of the groups have used corpora during the course of their studies; of these four groups of corpus users, two have used corpora extensively, and have been taught how to carry out advanced searches (multiple nodes, wildcards, and node plus tag). The fifth group, having had no exposure to language corpora, serves as a control.