On Oct 22, 2009, at 6:00 PM, Adrian Walker wrote:
For modern usage, I'd rate Ed Barkmeyer's definition of 'ontology' as the most satisfactory so far.
Here it is:
What makes written knowledge an "ontology" is that the language has a grammar and an interpretation of the grammatical constructs that is suitable for automated reasoning.
Otherwise (if less colorfully) put, what makes written knowledge an ontology is that it is expressed in a logic (i.e., a formal language + model theory) with a complete proof theory.
To forestall what seems an inevitable misreading by some in this forum: Note I said that ontologies are expressed in a logic. A logic is a representational medium for an ontology. A logic is not itself an ontology (except perhaps in a trivial sense).
ps: On the definition of a logic above, see Part A of Barwise and Feferman, Model Theoretic Logics
(Springer Verlag, 1985). I should note that this fantastic resource (which IIRC originally cost well over $150) has been put online and made available at no cost to the public: http://tinyurl.com/ylojq4p