Re: skolems: visible differences?

Testing my understanding of the skolem constant proposal for  
anonymous individuals:

If we adopted the treatment of bnodes as skolem constants, would it  
be the case that we no longer had the tree-shape restriction for  
related anonymous individuals?

If that is the case, could we simply have the syntax allow turtle  
style bnode syntax _:xxx for individual names and have these be  
considered anonymous individuals in the way RDF people are used to  
writing them? As I understood it, the avoidance of giving names to  
anonymous individuals was to make it impossible to express, in the  
abstract syntax, anything but tree-shape relationships.

Coupling this with serializing skolem constants as bnodes, it would  
seem that this would advance our ability to have more RDF graphs be  
considered valid OWL, affirmatively addressing Issue 46, yes?

-Alan

On Dec 10, 2007, at 9:32 AM, Peter F. Patel-Schneider wrote:

>
> From: Jeremy Carroll <jjc@hpl.hp.com>
> Subject: skolems: visible differences?
> Date: Mon, 10 Dec 2007 12:22:10 +0000
>
> [...]
>
>> ===
>> So I have some questions:
>> If we agree to specify the use of Skolems, what visible difference  
>> does
>> it make (e.g. in terms of tests)
>
> [...]
>
> Well in the RDFS-compatible semantics almost any entailment from a
> consistent ontology to a conclusion that incorporates OWL syntax
> involving bnodes would become false.
>
> This does not, of course, affect the direct model-theoretic semantics.
>
> peter
>

Received on Saturday, 12 January 2008 23:52:19 UTC