Semantic analysis and thematic annotation

Abstract

This contribution aims at investigating some methods, resources and tools devoted to the semantic analysis and the thematic annotation. The first part, devoted to the paradigmatic axis, describes available lexico-semantic resources for the classical languages, which belong to accomplished or on-going projects. Starting from Minozzi's Latin WordNet, the structure of multilingual lexico-semantic networks mapped on the original (American English) Princeton WordNet will be discussed and criticized. The automated procedures to create the basis for a new Ancient Greek WordNet from bilingual dictionaries (mainly: the LSJ) will be illustrated and the on-going project named Homeric Greek WordNet, validated by students and scholars, will be presented. Furthermore, the difficulties to map the conceptual nodes related to the ancient world on a modern semantic network will be discussed. The second part of the contribution, devoted to the syntagmatic axis, is focused on the semantic and thematic annotation of classical and biblical texts. The top-down approach to the annotation of themes and motifs in the Memorata Poetis Project is illustrated and pros and cons are discussed. In that project, devoted to the study of multilingual and multicultural intertextuality, a taxonomy of thematic labels established a priori is shared by all the members of the project. Finally, the bottom- up approach of Euporia is discussed. In this approach, folksonomies are created by the annotators, and the labels are grouped and organized in ontologies a posteriori, during an incremental process of revision.