Learning Italian verb + noun collocations through Data-driven learning: first insights into the role of semantic transparency

This paper presents the first findings of a study on the role that semantic transparency in verb + noun collocations may play in evaluating the effectiveness of data-driven learning (DDL). It is based on a controlled between-groups experiment conducted in 8 classes of Chinese learners of Italian at the University for Foreigners of Perugia, Italy. Phraseological competence data was collected at 4 points in time over a timespan of 13 weeks. The phraseological competence test was evenly divided into multiple-choice and gap-fill items, aimed to elicit definitional and transferable knowledge respectively. Overall, control groups appear to fare consistently better than experimental groups, though the latter seem to have better retention rates. Semantic transparency was found to have a significant role in the development of phraseological competence, for both conditions under consideration. In particular, the effect of semantic transparency was found to be more prominent in relation to the transferable knowledge of collocations.