Re: possible semantic bugs concerning domain and range

From: pat hayes <phayes@ai.uwf.edu>
Subject: Re: possible semantic bugs concerning domain and range
Date: Fri, 27 Sep 2002 22:10:25 -0500

> I am cross-posting this to the RDF core WG for its interest.
> 
> >From: pat hayes <phayes@ai.uwf.edu>
> >Subject: Re: possible semantic bugs concerning domain and range
> >Date: Wed, 25 Sep 2002 12:13:40 -0500
> >
> >[...]
> >
> >>  >>  These can all be expressed using my notion of range and
> >>  >>  rdfs:subClassOf or rdf:type. The important inferences about ranges -
> >>  >>  notably, the kind that arise from an association of a datatype with a
> >>  >>  range - apply in both semantics, but require more care to state in
> >>  >>  yours.
> >>  >
> >>  >Do these inferences actually work?  I thought that RDF Core had decided
> >>  >that they didn't work in the presence of super-properties.
> >>
> >>  No, the problem was the possibility of an XML datatype value space
> >>  being included in another when the datatype mappings were
> >>  incompatible, and we basically decided to punt on that one. I don't
> >>  think superproperties pose any problem, but maybe I missed something:
> >>  what do you see as the problem there?
> >
> >Suppose that you have the following:
> >
> >    foo rdfs:subPropertyOf bar .
> >    foo rdfs:range xsd:[integer union string] .
> >    bar rdfs:range xsd:[string union integer] .
> >
> >    john foo 10 .
> >
> >how is 10 to be interpreted?
> 
> integer. YOu can in this case also INFER that john bar 10, and now 
> you can infer that the literal in that second triple means a string. 
> So now, John's foo is ten and his bar is "10".

So the above implies that john has two bar's, ten and "10"?  

> However, if you have some way to keep track of things a bit more 
> finely you might also be able to infer from the first triple plus the 
> datatype that:
> 
> john foo _:x .
> (and  _:x is ten)
> 
> never mind for now how you know that second fact, suppose you do 
> somehow - and then you can use subproperty reasoning to conclude that
> 
> john bar _:x
> (and _:x is ten)
> 
> from which you know that john's bar is also ten. So John has to have 
> two bars: ten and "10", if you can somehow manage to construct this 
> chain of reasoning. But he still only has to have one foo.

So, I guess that john does have two bar's, ten and "10".

However, I'm completely confused about this notion of ``keeping track of
things a bit more finely''.  How does this fit into the RDF model theory?
Is the RDF model theory a kind of quantum theory, where events only happen
if they are noticed?

> All this is admittedly weird, but then so was your example.

What was weird about the example?  All it has is two properties, in a
subproperty relationships, and two ranges, both fairly simple XML Schema
datatypes.   I can forsee this exact situation arising quite frequently in
a datatype extension of RDF that allows XML Schema union datatypes.

> And it is 
> quite internally coherent.

I find this line of reasoning completely unsupportable.  Why should john
end up with two bar's?

> Pat

peter

Received on Saturday, 28 September 2002 07:16:00 UTC