Reassessing Constructions in the ARTEMIS Parser

Abstract

The aim of this study is to reexamine the status of constructions in ARTEMIS (Automatically Representing TExt Meaning via an Interlingua-based System), a Natural Language Understanding prototype that seeks to provide the syntactic and semantic structure of a given fragment in a natural language. The architecture of ARTEMIS has been designed to conform to the tenets of the Lexical Constructional Model (LCM), a theory in which constructions are a central tool for the linguistic description of languages. However, since ARTEMIS is a computational device, there are many formalization requirements which involve the adaptation of the LCM, a process which necessarily leads to reconsidering several issues, as are: (i) what counts as a constructional structure; (ii) how constructions contribute to parsing operations in ARTEMIS; and (iii) the location and the format of constructional patterns.