Towards a Computational Treatment of Aktionsart

Abstract

In the area of ​​Natural Language Processing (NLP), when creating intelligent applications, semantic processing is essential. However, research currently being conducted in ​NLP is still far from achieving deep levels of understanding of language. The main goal of our research is the representation of Aktionsart (how the event expressed by a verb is construed as unfolding over time). One of the basic difficulties presented by the semantic processing of language is establishing classes, mainly due to the gradual nature of meaning, and the high incidence of context in the interpretation of the different units. In this work we focus on the presentation of the lexical aspectual classes of our proposal. The total number of defined classes is classified into two groups, simple classes: states, processes and points, the combination of which gives rise to the complex classes: culminations, accomplishments and graduals. This presentation will take place both from the theoretical point of view, and its computational implementation.