Analyzing verbal subcategorization aimed at its computational application

Abstract

The verb is one of the most important lexical components: it includes information
regarding the necessary components that make up sentences and their features. This is
precisely the domain of the analysis of subcategorization. However, specifying the sub-
categorization of each verb is a difficult task, mainly because of the following reasons: first, because the distinction of the semantic values and the alternations in each verb is
problematic; and second, because of the presence of certain phenomena such as ellipsis,
unspecification (of general and specific elements), and dependencies between Cases.
   This work presents the following: after having reviewed the complex phenomena that are involved in verbal subcategorization, and contextualized these in our research area (i.e., in computational linguistics), we explain the procedure adopted to analyze 100 selected verbs, where Levin (1993) has been taken as point of departure. Once the research has been completed, we have defined what we have considered as subcategorization, namely, all the semantic/syntactic value(s) that we have defined for each verb (ssv), the set of outstanding elements in each ssv, their semantic specifications, and their Case realizations. Thus, we have tried to provide a coherent proposal as a base for grouping verbs depending on the goal.