Will My Ontologies Fit Together? A preliminary investigation

In realistic applications, it is often desirable to integrate different ontologies into a single, reconciled ontology. Ideally, one would expect the individual ontologies to be developed as independently as possible, and the final reconciliation to be seamless and free from unexpected results. This allows for the modular design of large ontologies and facilitates knowledge reuse tasks. Few ontology development tools, however, provide any support for integration, and there has been relatively little study of these issues at a fundamental level. Understanding at this fundamental level would help us predict, for example, what logical consequences to expect from the integrated ontology, and whether the integration of ontologies preserves some desirable properties of its parts. To the best of our knowledge, the problem of predicting and controlling the consequences of ontology integration has has been tackled only in [4]. The authors propose a set of reasoning services (with decidability and complexity results) to check whether, through integration with other ontologies, desirable properties of an ontology have been destroyed. In this paper, we propose first steps towards a different approach, the socalled normative approach. We specify certain properties that one would like to preserve in the integration and devise a set of restrictions that, when adhered to, guarantee to preserve these properties. Thus, while the approach of [4] determines the preservation of desirable properties ex post, our methodology prescribes some restrictions that guarantee the preservation of desirable properties. We introduce two ‘integration scenarios’ that, we believe, capture some of the common practices in ontology engineering, and postulate desirable properties that should be satisfied by the integrated ontology. We provide syntactic restrictions on the use of shared vocabulary that guarantee the preservation of