FEATURES AND CLASSES OF THE EVENT STRUCTURE: TOWARDS A COMPUTATIONAL REPRESENTATION

Abstract

Research currently being conducted in the area of  Natural Language Processing is still far from achieving deep levels of understanding of language. To build intelligent systems that deal with the representation of meaning, it is necessary in the field of language technologies, to create wide coverage semantic resources. The main goal of our research is the establishment of classes for representing events in a computational system. The basic unit of representation is thefeature;  we have  considered four specific features:  dynamism, telicity, duration  and gradual change.  From the combination of these semantic features we have established a set of event classes that  allows us to characterize verbal behavior. To establish these classes we took into account possible event changes that a verbal unit may suffer because of the  context, thus representing  the compositionality of event  meaning. Therefore we have  considered theprototypicity of verbal senses, and the sensitivity of the different classes to the context. The total 2 JUAN APARICIO, IRENE CASTELLÓN Y MARTA COLL-FLORITANU.FILOL.ESTUD.LINGÜÍST., 1/2011, pp. 1-29, ISSN: 2014-1408number of defined classes is classified into two groups, simple classes: states,  processes andpoints, the combination of which gives rise to the complex classes: culminations, accomplishments and graduals