Re: CCAT Conceptual Catalogue and Ontologies

sowa@ranch.poly.edu (John F. Sowa)
Date: Fri, 9 Sep 94 16:12:29 -0400
From: sowa@ranch.poly.edu (John F. Sowa)
Message-id: <9409092012.AA02504@ranch.poly.edu>
To: cg@cs.umn.edu, fritz@rodin.wustl.edu
Subject: Re: CCAT Conceptual Catalogue and Ontologies
Cc: srkb@cs.umbc.edu
Sender: owner-srkb@cs.umbc.edu
Precedence: bulk

I agree with Pat Hayes that "we need a model theory for CGs that
is precise (and stable) enough to allow some metamathematics to
be done."

The term "conceptual graph" has become in Wittgenstein's sense a
"natural concept" that is defined by family resemblance, not by
a set of necessary and sufficient conditions.  There is a fairly
consistent, clean first-order subset based on my 1984 book, there is
a wider, ill-defined set based on all the things that I have said
"could be done" with CGs, and there is a still wider set of all
the extensions that everybody else has added in all the CG-related
papers published in the past ten years.

On the Saturday after ICCS'94, Michel Wermelinger and I went over
a number of the issues that we have discussed concerning syntax and
semantics.  I am in the process of writing that up and so is Michel.
Next week, I will be going to Seattle to work with the X3T2 gang,
including Mike Genesereth and our dear gadfly Matt Ginsberg.

Unless something drastic happens in Seattle, we will continue with
the ongoing agreement to develop a firmly defined subset of CGs
with a model theoretic semantics that is identical to the KIF
semantics.  That subset will include the first-order subset of the
1984 book and as much of the metamathematical stuff as we can
get well-defined and agreed to by all (or at least a majority)
of those concerned.

This subset will not preclude anyone else from experimenting with
extensions and proposals for extensions.  But we want to have a
firmly defined core syntax and semantics by December of this year.
Michel said that as soon as we get the syntax firmly defined, he
will modify his current YACC parser to accept it.  That code will
be distributed via ftp to anyone who wants to "fill in the blanks"
of the parser with their favorite semantics -- which we hope will
conform to the definition that will be available at the same time.

Meanwhile, I plan to send another update to the email lists
after the Seattle meeting to let people know what progress we
are making.

John Sowa