A two-level semantics for french expressions of motion

Abstract

Developing suitable representations for formalizing time and space knowledge has always been of a great importance in Artificial Intelligence (AI) and cognitive science. We here present a new way to conjoin these two problems. From the linguistic study of motion (which is the best concept available to associate space and time at the lexical and phrase levels), realized by Laur (1991), we construct a system to represent the spatio-temporal semantics of motion. This linguistic analysis consists of a semantic classification of the French motion verbs and spatial prepositions and of the elaboration of compositional rules between the semantic classes of these verbs and these prepositions. Our system, based on a two-level semantics representation, allows to formally represent the results drawn by the linguistic part and to perform some kind of natural spatio-temporal reasoning.