Reasoning on the Semantic Web needs to reason both on ontology-based assertions and on ontologies themselves

ABSTRACT In this arti le, we present a method whi h aims at using ontologies to perform di erent types of reasoning. This method, alled operationalization, proposes to automati ally adapt the representation of the axioms to the type of reasoning the KBS in whi h the ontology is integrated is dedi ated to. This pro ess is based on the des ription of the operational goal of the KBS through a s enario of use. So, we argue for the distin tion between the problems of knowledge representation in ontologies, that are built at the on eptual level, and the problems related to reasoning with ontologies, that o ur at the operational level. Moreover, be ause reasoning on the Web requires to reason both on domain knowledge and on ontologies, for instan e for evaluation or alignment purposes, we propose an extension of the operationalization method that permits to reason on ontology by operationalizing a meta-representation of the language used to represent ontologies. The language we use, alled OCGL, is based on the Con eptual Graphs model and a tool, alled TooCoM, implements this language and the operationalization method.