Monadic predicates in RDF

At 09:30 AM 6/6/02 -0400, Thomas B. Passin wrote:
> > >   Fred isChocolateLover 'true' .
> > >
> > > can be expressed as:
> > >
> > >   Fred rdf:type ChocolateLover .
> > >
> > > #g
> >
> > This is true. Or, err
> >
> >   "" rdf:type True .
> >
>  But does it really amount to the same thing?  With TRUE/FALSE, you can say
>"I definitely know that Fred is NOT a ChocolateLover" (poor Fred, what he's
>missing!).  Without it, you have to infer "I can't find anything that says
>that Fred IS a ChocolateLover, so he must not be one".

Well, you can always introduce:

     Fred rdf:type ChocolateHater .

or some similar.  Then, the syntactic constructs allow similar options to 
be formed:

- say nothing about Fred's chocolate preference
- assert that Fred is a chocolate lover
- assert that Fred is not a chocolate lover
- assert that Fred both is and is not a chocolate lover

#g

>Isn't this this similar to the existential stataus of a bnode, whose
>existence has to be inferred because it is connected to other nodes?  Most
>people are saying that there really is a difference between such a node and
>an identified resource.  So how is it that inferring FALSE is suppsed to be
>the same as asserting it?

Well, now you touch on the whole negation/open worlds/closed worlds 
debate.  In general, RDF philosophy tends to eschew the negation-as-failure 
approach, though it may have enclaves of utility.  I think that if you can 
validly infer a result that is no different from asserting it.

But, lurking around the corner, I think there are modalities to be 
considered, which may colour this view somewhat.

#g


-------------------
Graham Klyne
<GK@NineByNine.org>

Received on Thursday, 6 June 2002 10:24:23 UTC