Conceptual graphs: was: Is "Model" part of the RDF model?


I'll play the  role of target audience :-) Conceptual Graphs (CGs) are the
brainchild of John Sowa, and have been around for some time (since the
1980s) they are an extension of the existential graph logics first put
forward by CS Pierce. There was a fair bit bit of AI research a few years
back regarding its use for modelling natural language and general knowledge

The key web page is : where you can also
find a very well written draft ansi proposal. The most interesting part of
that draft is the way that logical quantifiers are dealt with and section
9.2 which describes 6 basic graph operations on CGs from which rules of
inference can be built (something similar would be way cool for RDF, but
this is drifting into rdf-logic territory I think). The draft is well worth
a look.

The key book is Knowledge Representation by John Sowa, which is a super book
if you're working with RDF or Knowledge/Rule based systems.

It does surprise that more people haven't made the connection betwen RDF and
CG graphs, or that we don't have (afaik) a Xmlised form of CGs (the CG
standard mentioned includes the bnf for anyone feeling adventurous ...). For
example, a mapping of RDF to CGs would mean that RDF can be mapped directly
onto predicate calculus, which means that anything describable by RDF (such
as XLink) could have a predicate calculus form.

-Bill de hÓra

Received on Monday, 23 October 2000 12:27:00 UTC