Combining corpora and video-based descriptions

Abstract This paper explores the methodological potential of combining corpus-based and video-based resources in Contrastive Analysis within the theoretical framework of Cognitive Linguistics. Focusing on the breaking domain and the verbs used to describe events in this domain in English, Italian, and Swedish, we integrate data from the TenTen corpora and IMAGACT, a multimodal linguistic resource, to examine how different linguistic resources contribute to semantic analysis. We argue that using multiple data types enhances methodological robustness. We show that the use of a video-based language resource provides a consistent and reliable tertium comparationis and allows the researchers to observe how actions are performed and to identify the characteristics of the action participants. These insights are combined with corpus data, which offers richer information about the verbs’ use in a variety of contexts. By combining different data types, we gain a more comprehensive understanding of verbs and action semantics.