Re: Rule-based approach to conclude owl:intersectionOf

> <#A> rdf:type <#X> .
> <#B> rdf:type <#X> , <#Y> .
> <#C> rdf:type <#X> , <#Z> .
>
> =>
>
> <#X> rdfs:subClassOf <#Y> , <#Z> .

X = { A, B, C }, Y = { B } and Z = { C }. X is not a subclass of either 
Y or Z, but the other way around.

----

Assuming you had said
	<#Y>, <#Z> rdfs:subClassOf <#X>
I would still have to disagree with you. From the RDF Concepts and 
Abstract Syntax document:

	To facilitate operation at Internet scale, RDF is an open-world 
framework that
	allows anyone to make statements about any resource.

	In general, it is not assumed that complete information about any 
resource is
	available.

Your conclusion is that, because every instance of Y **that you know 
of** is also an instance of X, that this holds in general. However, the 
following statement could also hold:
	<#D> rdf:type <#Z>
On the face of it, using the same reasoning, we now "know" that Z is 
*not* a subclass of X. But adding yet another statement,
	<#D> rdf:type <#X>,
restores our "conclusion".

In essence,
	<#Y> rdfs:subClassOf <#X>
states that Y will always be a subclass of X, regardless of what 
additional information we learn, and we therefore cannot deduce it from 
a few instances of the use of X and Y.

Chris

Received on Tuesday, 5 April 2005 12:20:24 UTC