Re: yet another non-entailment

From: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>
Subject: Re: yet another non-entailment
Date: Tue, 27 Aug 2002 14:00:36 +0200

> 
> 
> >I'm not proposing anything complicated.  I would be happy with a format
> >that is something like
> >
> >DESCRIPTION:
> >
> ><text>
> >
> >RATIONALE:
> >
> ><text>
> >
> >PREMISE:
> >
> >???
> >
> >CONCLUSION:
> >
> >???
> 
> I think nearly all this information is available (in a machine readable
> form) and it's a small matter of programming to present it in the fashion
> you suggest.
> 
> What we haven't had is "rationale" could you give an example, maybe for one
> of your suggested tests (I guess any of the examples of a non-entailment in
> Pat's model theory would suffice).

See the example at the end of my message.

> My fear is that the rationales would end up as vacuous (e.g. "Logical
> consequence of the model theory"; this rationale is, in the medium term, a
> requirement - if it isn't true either the model theory or test will need to
> be changed).

Rationale would be things like

``This has been disputed ....''

``This shows how the treatment of domains of properties differs from that
in RDFS.''  (For an example that shows that a subclass of a domain is also
a domain.)

> Jeremy

peter

Received on Tuesday, 27 August 2002 08:11:24 UTC