[Top] [All Lists]

Re: [ontac-forum] RE: ontac-forum Digest, Vol 13, Issue 2

To: ONTAC-WG General Discussion <ontac-forum@xxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Thu, 18 May 2006 08:48:24 -0400
Message-id: <446C6D18.6060908@xxxxxxxxxxx>
Cathy, Chris, and Leo,    (01)

Bundy was using the word "reasoning" in a very general
(i.e., vague) sense:    (02)

AB> "... in a logic-based representation the predicates
 > and functions, their arities and their types, may all
 > need to change during the course of reasoning...."    (03)

CL> I just want to say - as someone who did hands-on work
 > in a logic-based representation for 2 years - how on
 > earth is this to be effected?    (04)

Bundy would certainly admit that no known method of deduction
could ever change the predicates, functions, arities, and
types.  However, what he was proposing is a method of
metalevel reasoning about problems and theories.  That
reasoning process would determine what kinds of changes
should be made to the theories themselves, before they
are used to solve the problems by deduction.    (05)

CM> It is the default for the languages (or "dialects") of
 > Common Logic -- to whose development John has been a major
 > contributor -- that predicates and function symbols denote
 > relations and functions that do not have a fixed arity.
 > (The predicates and function symbols can therefore take
 > varying numbers of terms as arguments.)  Granted, arity-freedom
 > isn't quite the same as fixed arities *changing*, as suggested
 > in the quote above, but the effect of the latter could be
 > simulated via the former simply by having axioms that determine
 > the logical behavior of the same arity-free predicate with
 > respect to varying numbers of arguments.    (06)

As a matter of fact, I made a suggestion along those lines in
the discussion period after Bundy's talk.  He had been using
KIF, and I suggested that he consider CLIF as a more flexible
representation, which could facilitate the revisions.    (07)

In any case, the process of revision is not carried out
by deduction within a theory, but by a metalevel stage of
reasoning about a theory and its applicability to a given
problem.    (08)

LO> But it depends on what is meant by needing "to change during
 > the course of reasoning". That could suggest on the fly automated
 > kr learning, which I don't think anyone can handle yet.    (09)

Yes, indeed.  Bundy explained some of the methods they were using,
which are discussed further on the web site.  But he also admitted
that this is a long-term research project, which will undoubtedly
take many years and many more projects and researchers to accomplish.    (010)

John    (011)

Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki: 
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG    (012)
<Prev in Thread] Current Thread [Next in Thread>