Re: A Single Foundational Logic for the Semantic Web

From: "Seth Russell" <seth@robustai.net>
Subject: Re: A Single Foundational Logic for the Semantic Web
Date: Tue, 30 Apr 2002 14:13:27 -0700

> From: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
> 
> > Even formulae like the following have to exist in all models
> >
> > _:b rdf:type rdf:Statement .
> > _:b rdf:subject _:b .
> > _:b rdf:predicate rdf:type .
> > _:b rdf:object n3:falsehood .
> >
> > this is, more-or-less,
> >
> > { this a n3:falsehood }
> >
> > Now does this formula belong to log:Truth?  If it does, then it is a true
> > formula, and thus is a falsehood, because it belongs to n3:falsehood, and
> > thus it can't belong to log:Truth.  If it does not, then it is a non-true
> > formula, and thus is a truth, because it does not belong to n3:falsehood,
> > and thus it must belong to log:Truth.
> 
> How about ?
> 
> {{ this a n3:falsehood } a  n3:NonSense }
> 
> Me thinks it's possible that binary logic is not going to be very useful on
> the semantic web, just because of the kinds of things you are bringing up.
> 
> Seth Russell
> Logic is Great, Survival better!

Multi-valued logic does not do better here, unless, *maybe*, if you have a
very weak multi-valued logic.  The problematic formualae are just more
complex.

peter

Received on Tuesday, 30 April 2002 18:44:49 UTC