uos-convene
[Top] [All Lists]

Re: [uos-convene] Lattice of theories

To: John F.Sowa <sowa@xxxxxxxxxxx>
Cc: Upper Ontology Summit convention <uos-convene@xxxxxxxxxxxxxxxx>
From: Christopher Menzel <cmenzel@xxxxxxxx>
Date: Mon, 13 Mar 2006 15:27:46 -0600
Message-id: <AE281DA8-B360-4B2D-8534-60C0D3E7BE48@xxxxxxxx>
John wrote:    (01)

> The point I have been emphasizing even more than the
> lattice of all possible theories is the need for a
> *registry* of actually specified and tested theories.
>
> Among the relationships to be recorded in the registry
> are implication/entailment, analogy (or whatever you
> want to call it), and many, many others, including
> who defined them, used them, tested them, etc.    (02)

An excellent idea.    (03)

> We should also agree on some standard terminology
> for the relations of theories:
>
>  1. I have never liked the word "subsumption" -- partly
>     for the reason that it can be interpreted in many
>     ways (as you did in your note).  I'd rather just
>     call it implication or entailment.    (04)

But that's too narrow, as it only works for theories that share a  
common language (or where the language of one is properly included in  
that of another).    (05)

>  2. That's an excellent reason for getting rid of the
>     word "subsumption":  "... ZF subsumes Peano Arithmetic".
>     By using Goedel numbering, you could also say that
>     arithmetic subsumes all sorts of theories.    (06)

??  In the sense I intended, ZF subsumes PA in that, under an  
appropriate mapping of the language of PA into that of ZF, every  
axiom of PA is a theorem of ZF.  Sure enough you can represent the  
axioms of any theory T you please in PA as Gödel numbers, and  
moreover you can even represent *that* a given sentence A is a  
theorem of T (by representing proofs in T as Gödel numbers), but you  
aren't thereby guaranteed (indeed it will almost never be the case)  
that there is a reasonable translation of A into the language of PA  
that is *itself* a theorem of PA.  That's what's special about  
relative interpretability.    (07)

> That may
> be theoretically interesting, but hopelessly confusing
> for any serious discussion of practical problems.    (08)

Well, I certainly agree with that.  Of course, it has nothing to do  
with relative interpretability, either.    (09)

>  3. Relative interpretability is a big mouthful to say or type,    (010)

Well, I have no particular affection for the *term*, it's the *idea*  
I'm concerned about.    (011)

>     and it's hard to explain without getting
>     into lots of technical issues.    (012)

I guess you could say the same thing about the calculus vis-a-vis  
engineering!    (013)

>     I use analogy because
>     that's what Bohr did when he renamed the sun and earth
>     the proton and electron for his model of the H atom.    (014)

Sure, but now we're back to quibbling. :-)    (015)

> Just one other point:
>
> CM> Is not subset the arc relation in the lattice under that
> > understanding?
>
> It is true that if theory T1 implies T2, then T2 is a subset
> of T1 and vice versa.    (016)

Vice versa??    (017)

> However, that can be confusing, since
> (a) the directions of the two relations are opposite, and
> (b) implication is a logical relation (which is the main point)
> and subset is a secondary issue that distracts attention.    (018)

Well, I was only talking extensionally among friends. :-)    (019)

-chris    (020)

 _________________________________________________________________
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    (021)
<Prev in Thread] Current Thread [Next in Thread>