uos-convene
[Top] [All Lists]

Re: [uos-convene] Relating ontologies: structuring and development tools

To: Michael Gruninger <gruninger@xxxxxxxxxxxxxxx>
Cc: Upper Ontology Summit convention <uos-convene@xxxxxxxxxxxxxxxx>
From: Chris Menzel <cmenzel@xxxxxxxx>
Date: Thu, 9 Mar 2006 13:19:44 -0600
Message-id: <20060309191944.GR931@xxxxxxxx>
On Thu, Mar 09, 2006 at 01:55:58PM -0500, Michael Gruninger wrote:
> ...the methodology is:
> 
> 1. Specify a class of structures
> 2. Existence Theorem : show that this class is nonempty
> 3. Classification Theorem: classify the structures in this set up to 
>    isomorphism
> 4. Show that each structure in the class is a model of the theory
> 5. Show that each model of the theory is elementarily equivalent
>    to a structure in the class.    (01)

That's nice and clear.    (02)

> Pedantically speaking, these are relative consistency proofs, since
> the structures are typically defined with respect to classes of
> graphs, geometries, ordered sets, groups, semigroups etc. which we
> assume exist and are nonempty.    (03)

Well, yeah, when we're talking about interesting theories, "consistent",
of course, typically means "can be proved in ZFC to have a model". :-)    (04)

> The classification theorem is not always necessary, but it does give
> the best "picture" of the possible models of the theory, alerting
> people to models that are potentially unintended. It also gives more
> insight into the theory.  For example, you could say that a group is
> any structure that satisfies the group theory axioms, but this doesn't
> tell us much;     (05)

Indeed, doesn't even tell you there *are* any.    (06)

> you get much more sense of the models from the classification theorem
> and structure theorems that tell you how to decompose groups into
> subgroups.    (07)

Agreed, of course!    (08)

On a purely point of technical fun, I'm not sure why you call step 5
"Axiomatizability".  Let me follow your methodology.  My theory is True
Arithmetic (i.e., the set of sentences of number theory true in the
standard model).    (09)

Step 1: My class is the class N* of structures isomorphic to the
standard model (or, more concretely, to omega with the usual definitions
of ordinal addition and multiplication).    (010)

Step 2: ZF proves this class is nonempty (by the axioms of infinity and
separation).    (011)

Step 3: Done, by definition of N*.    (012)

Step 4: Done, by the definitions of True Arithmetic and N*.    (013)

Step 5: Done, by the completeness of True Arithmetic and the definition
of elementary equivalence.    (014)

BUT: True Arithmetic is not (recursively) axiomatizable, by Gödel.    (015)

-chris    (016)

 _________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files: http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit    (017)
<Prev in Thread] Current Thread [Next in Thread>