How to Represent the World: Ontology-Controlled Natural Languages

Abstract

As far as human knowledge about the world is commonly given in NL expressions and as far as universal ontology is a general science of the world, the examination of its impact on natural language science and technology is among the central topics of many academic workshops and conferences. Ontologists, knowledge engineers, lexicographers, lexical semanticists, and computer scientists are attempting to integrate top-level entity classes with language knowledge presented in extensive corpora and electronic lexical resources. Such a deep quest is mostly motivated by high application potential of reality-driven models of language for knowledge communication and management, information retrieval and extraction, information exchange in software and dialogue systems, all with an ultimate view to transform the World Wide Web into a machine-readable global language resource of world knowledge, the Onto-Semantic Web. One of the practical applications of integrative ontological framework is to discover the underlying mechanisms of representing and processing language content and meaning by cognitive agents, human and artificial. Specifically, to provide the formalized algorithms or rules, whereby machines could derive or attach significance (or signification) from coded signals, both natural signs obtained by sensors and linguistic symbols.