Within philosophy, ontology is the study of things that exist; the objects, properties, categories and relations that make up the world. A definition of ontology as it is conceived by philosophers is:
… the science of being in general, embracing such issues as the nature of existence and the categorial structure of reality. … Different systems of ontology propose alternative categorial schemes. A categorial scheme typically exhibits a hierarchical structure, with ‘being’ or ‘entity’ as the topmost category, embracing everything that exists (Honderich, 1995).
This approach to the definition and use of ontology has also been successfully applied in information systems, for example in comparing and evaluating data modelling frameworks (Milton, 2000; Milton and Kazmierczak, 2004).
Philosophers distinguish between reference ontologies, which aim to determine the fundamental categories and categorial structures, and domain specific ontologies, where a particular reference ontology is applied to a certain problem domain. An analogy could be made between the way that ontologies and data models are conceptualised, as shown in Table 3.1, “Ontologies and data models”.
Table 3.1. Ontologies and data models
|
Ontology |
Data model |
|---|---|
|
Reference ontology |
Data modelling language |
|
Domain specific ontologies |
Data models |
|
Specific reality |
Database implementation of specific database instances |
Reference ontologies, such as that of Bunge (1977; 1979), Chisholm (1996), Basic Formal Ontology (Smith, 1978; Smith and Mulligan, 1983) and DOLCE (Gangemi, et al., 2002) are concerned with the most general categories of what there is in the world. They deal with concepts such as thing, individual and property. In the analogy proposed in Table 3.1, “Ontologies and data models”, data modelling languages such as entity-relationship modelling are proposed as analogous to reference ontologies.
At this level, particular reference ontologies are used to create a domain specific ontology directed towards an aspect of reality, in the way that particular data modelling languages are used to create data models of a particular project, such as a university student records system or a banking system. Early artificial intelligence (AI) focus was at this level.
At this point we note that AI researchers have a particular view of ontology, referring to it as
… an engineering artefact, constituted by a specific vocabulary used to describe a certain reality, plus a set of explicit assumptions regarding the intended meaning of the vocabulary words (Guarino, 1998).
As previously mentioned, ontologies developed within the AI community are often directed to a specific domain of knowledge in a specific context, and are intended to be implemented or defined within a specific software artefact. Examples are KIF (Genesereth and Fikes, 1992), Ontolingua (Gruber, 1992; 1995), and OIL (Fensel, et al., 2000).
Recent work by AI researchers around the Semantic Web and the IEEE Standard Upper Ontology (IEEE, 2003), and DOLCE (Gangemi, et al., 2002) is encouraging in that they all recognise the central role of higher level ontologies in information systems.
It is the higher level reference ontologies and the basis for the selection of a particular reference ontology for the creation of a domain specific ontology, directed towards information systems research, that is the concern of this paper.