Proposta de filtros cognitivos a partir dos aportes da Teoria da Valência Verbal e dos Papéis Qualia para mapear os sentidos dos verbos que expressam as relações semânticas em Tesauros

Abstract

The complexity of human communication gives an inaccuracy condition to language processing before logical language, once it may entail situations of ambiguity, synonymy, homonymy and polysemy. In this context, in the studies of Information Science (IS), Linguistics gains importance due to its relevance in the development of services such as treatment, normalization, translation, construction and dissemination of information, as well as the refinement of thesaurus, a kind of Knowledge Organizational System (KOS). In addition to the processing of natural language, the group of relationships present in a KOS is, for itself, vast; once it covers different notions amongst concepts and terms, being them associative, of equivalence or hierarchic, making the work of IS researchers interested in investigations about thesaurus construction utterly hard. Hence, the creation of semantic markings and sharing patterns in the present Information Society becomes prominent and, together with the difficulty in enumerating all the conceptual relations, it justifies the necessity of studies with focus on the argumentation factors of these relations. Therefore, in order to contribute with the treatment given to the relations between the concepts, this dissertation analyses the semantic value of the cognitive schemes invoked by verbs. For this purpose, this study relies on the contributions of Verbal Valence Theory, that highlights the necessary arguments to make the verb meaningful, and of Qualia Structures, that allows the refinement of semantic relations. Another point of support of the study is the field of Agriculture Intensification, previously studied by the mentor of this dissertation in her doctoral thesis. Beginning from collected sample and the principles of Corpus Language, a meaning restriction methodology proposal has been developed for the three relations found in the field at issue (<afeta>, <influencia> and <causa>). The methodology relied on a step-by-step approach that has shown the criteria for the construction of the corpus, the technology used for its automatic compilation, the analysis that indicated the attribution of verbal valence and the Qualia roles of the three relations and, lastly, the validation of the real sense of each one of them in a thesaurus in the field of Agriculture Intensification, thus confirming the contribution of the arguments in the identification of the semantic relations between objects.