Authors:
Yetian Wang
1
;
Ramona Kühn
2
;
Randy Allen Harris
3
;
Jelena Mitrović
2
;
4
and
Michael Granitzer
2
Affiliations:
1
David R. Cheriton School of Computer Science, University of Waterloo, 200 University Avenue West, Canada
;
2
Faculty of Computer Science and Mathematics, University of Passau, Innstraße 43, Passau, Germany
;
3
Department of English Language and Literature, University of Waterloo, 200 University Avenue West, Canada
;
4
Institute for Artificial Intelligence Research and Development of Serbia, Fruškogorska 1 21000 Novi Sad, Serbia
Keyword(s):
Ontology, Rhetorical Figures, Knowledge Representation, Computational Rhetoric, Language Modelling.
Abstract:
Formal ontologies for rhetorical figures have been developed to improve the computational detection for different applications in the area of Natural Language Processing, such as hate speech and fake news detection, argumentation mining, and sentiment analysis. The existing ontologies all model different aspects of rhetorical figures, thus creating a variety of formalisms and in the worst case, creating incompatibilities and contradictory representations. In this paper, we focus on figures of perfect lexical repetition and their representation in three ontologies in three different languages: The Ploke ontology, the Serbian RetFig, and the German GRhOOT ontology. We combine those ontologies to benefit from synergy effects and create a multilingual, coherent, robust, and modular ontology for rhetorical figures of perfect lexical repetition.