1163 Translating Action Verbs using a Dictionary of Images: the IMAGACT Ontology Alessandro Panunzi°, Irene De Felice*, Lorenzo Gregori°, Stefano Jacoviello † , Monica Monachini*, Massimo Moneglia°, Valeria Quochi*, Irene Russo* °University of Florence, *ILC CNR (Pisa), † University of Siena alessandro.panunzi@unif.it, irene.defelice@ilc.cnr.it, lorenzo.gregori@unif.it, stefano.jacoviello@gmail.com, monica.monachini@ilc.cnr.it, moneglia@unif.it, valeria.quochi@ilc.cnr.it, irene.russo@ilc.cnr.it Abstract Action verbs have many meanings, covering actions in diferent ontological types. Moreover, each lan- guage categorizes action in its own way. One verb can refer to many diferent actions and one action can be identifed by more than one verb. The range of variations within and across languages is lar- gely unknown, causing trouble in all translation tasks. IMAGACT is a corpus-based ontology of action concepts, derived from English and Italian spontaneous speech corpora, which makes use of the uni- versal language of images to identify the diferent action types extended by verbs referring to action in English, Italian, Chinese and Spanish. This paper presents the IMAGACT search interface and the various kinds of linguistic information the user can derive from it. IMAGACT makes explicit the vari- ation of meaning of action verbs within one language and allows comparisons of verb variations wit- hin and across languages. Because the action concepts are represented with videos, extension into new languages beyond those presently implemented in IMAGACT is done using competence-based judgments by mother-tongue informants, without intense lexicographic work involving underdeter- mined semantic descriptions. Keywords: Action verbs; Image ontology; Multilingual dictionary; Computer-aided translation 1 Introduction In all language modalities, action verbs bear the basic information that should be understood in order to make sense of a sentence. Moreover when we communicate, we have to refer to actions very often. Native speakers do not have a problem fnding the right verb for a specifc action in their own langu- age. However, in a foreign language, they often have difculty choosing the appropriate verb. The rea- son is that the more common action verbs, in their own meaning, refer to many diferent actions: in this sense, they are “general” verbs. Moreover, each language categorizes actions in its own way. These facts imply that there are not one-to-one translation relationships between diferent general verbs in diferent languages (Majid et al. 2007; Kopecka & Narasimhan 2012). If we take the English verb to 1 / 8 1 / 8 1 / 8 1 / 8 1 / 8 1 / 8