Re: On the definition of "ontology"
vet@cs.utwente.nl (Paul van der Vet)
Date: Fri, 6 Oct 1995 10:49:14 +0100
From: vet@cs.utwente.nl (Paul van der Vet)
Message-id: <199510060949.KAA08928@ethanol.cs.utwente.nl>
To: phayes@ai.uiuc.edu
Subject: Re: On the definition of "ontology"
Cc: srkb@cs.umbc.edu
X-Sun-Charset: US-ASCII
Sender: owner-srkb@cs.umbc.edu
Precedence: bulk
At 10/4/95, Pat Hayes wrote:
> At 2:30 PM 10/4/95 +0100, Paul van der Vet wrote:
> > [...] when you develop a
> >knowledge-based system you'll have one intended model in mind.
>
> This raises an interesting issue (which may be getting too philosphical for
> skrb, but here goes anyway.)
I didn't mean to raise *that* issue, of course. But it's near weekend,
I've finished the stuff I should have finished two weeks ago, so let's
indulge ...
> What one 'has in mind' is presumably expressed
> there in some mental representation. Now, how do we know that this
> mentalese representation in fact has unique models? All the lessons of
> logic would suggest that it usually doesn't; that usually there will be
> nonstandard models of our thoughts. (This will almost certainly be true,
> for example, if theoremhood in our mentalese is RE and it is given a
> semantics which reflects this accurately.)
>
> Now, if this is the case, then what warrant can we have for insisting that
> there is "a" model that we have in mind? We have no special access to
> models of our mental language: we can only think of things IN that
> language.
Let's, for the sake of argument, take the leap of faith for granted and
assume we have some device to read our mind.
- Would we obtain a theory in logic or a theory plus model? If the
latter, it is again a logical theory since it is a bundle of thoughts
(at least, so I understand Pat). Tarski could pull the trick of having
the recursion `bottom out' at the level of natural language (the
"absolute" meta-level). We are prevented from doing the same because
language originates in thought and there we go again. In other words,
the idea of model-theoretic semantics for thoughts (*if* they are
present? expressed? in a logical theory) may be incoherent.
- There are semantics other than model-theoretic such as update
semantics and Kowalski's closely related `logic without models' idea.
The incoherence objection doesn't apply, but in adopting these
semantics we loose some of the advantages of model-theoretic
semantics.
Paul.
---------------------------------------------------------------------
Paul van der Vet Phone +31 53 89 36 94 (until Oct. 10)
Knowledge-Based Systems Group Phone +31 53 489 36 94 (after Oct. 10)
Dept. of Computer Science Fax +31 53 33 96 05 (until Oct. 10)
University of Twente Fax +31 53 489 29 27 (after Oct. 10)
P.O. Box 217 Email vet@cs.utwente.nl
7500 AE Enschede
The Netherlands
---------------------------------------------------------------------