Semantic Core and Ontologysowa@watson.ibm.com
Date: Tue, 7 Jul 92 16:06:56 EDT
Cc: CG@cs.umn.edu, INTERLINGUA@ISI.EDU, KR-ADVISORY@ISI.EDU, SRKB@ISI.EDU
Subject: Semantic Core and Ontology
Yes, my note only addressed the issues of the core language, and
ontology must also be an extremely important part of the project.
Even if we all use the same version of logic, we can't communicate
unless we have a common vocabulary (i.e. types, relations, etc.).
So when we say that the IRDS normative language is being "based on"
conceptual graphs, that only means that we are using CGs as a common
syntax. The choice of predicates necessary for expressing all of
the content is an extremely important matter. And those predicates
could be equally well expressed in many different syntaxes.
In fact, I think that one reason why we are all so willing to agree
on the common semantics for the underlying logic is that we realize
that the most serious issues are in ontology. And we can express
the same ontologies in KIF, CGs, KRSS, SNePS, SUMM, etc.