Grid for Decoding Motion Encoding

Abstract

The need of cumulability of data and replicability of resultus in motion event description has led to the construction of a prototype of an annotation tool named Modeg (Motion DEcoding Grid). This classification grid allows to encode the most relevant formal and semantic elements implied in the linguistic expression of motion events. Modeg identifies a set of comparative concepts and their possible values that can be annotated with respect to their expression in four loci (satellite, adnominal, verb, noun) as well as to their distribution in the sentence. The software on which Modeg is tested should be considered not a final product, but rather as a tool developed to demostrate the applicability of the criteria we have identified. The results that can be obtained with Modeg are shown with examples mostly relates to parallel corpora analysis.