IMAGACT4AL. Una ontologia per immagini dell’azione per l’apprendimento del lessico verbale di base delle lingue seconde

In many languages, the most frequent action verbs are “general”, i.e. they can refer to a series of pragmatically different physical actions (e.g. turning the postcard upside down vs. turning right at the crossroad). These verbs are difficult to translate both for language learners and for automatic translation systems. IMAGACT is a multilingual ontology of action that identifies the action concepts by means of prototypical scenes (1010 in total), and link them to the correspondent verbs in different languages (4 in the first public release; around 20 in the current version). In this paper we present: (a) the extension procedure needed to map the ontology on new languages; (b) the IMAGACT research interface, which comprehends 3 main functions (Dictionary, Compare and Gallery); (c) an experiment regarding the use of IMAGACT in a L2 acquisition scenario (Chinese students learning Italian).