Re: unreification

ssarkar@ayushnet.com wrote:

> > [ident1, s, p, o]
> > [ident2, Rx, reifies, ident1]
> > [ident3, Rx, isA, Lie]
> >
> > but then if  ident1 gets forgotten, ident3 suddenly forgets what it's all about
> > (yuck).
>
> This seems to me a trick for truth value maintenance
> over the whole set of sentences involved in reification.
> I think that the notion of atomicity in reified
> sentences should be considered and a syntax needs
> to be in place.  A <start> and <stop> or <begin>
> and <end> can bring atomic notions of reified
> sentences in RDF just like in high level programming
> languages.

I'm not sure what you mean by 'atomic notion', but i assume that your are referring
to the use of  an 'ident' to allow a statement to be the object of other statements.
This, of course, is not allowed in RDF serialization or in the RDF model.  But it
certainly is allowed internal to a application that reads and writes RDF.  So we do
have a dilemma: How does an application express externally (hopefully in RDF) such
knowledge that is not explicitly in the RDF model - for example, the knowledge that a
collection of statements hold within a certain context?  If I reify each of the
collected statements and then assert of those reifications that they belong in a
certain context, assuming  that the reifications stand for the originals, then I have
no way to put the reifications themselves in some other perspective (yuck).

I think the solution, as you have hinted, is to allow  statements themselves to
contain idents and be objects of other statements and to have some way to communicate
that in RDF serialization.

<signature format="mime/topic">
topic: Seth Russell
isHavingProblemsWith:  Context Browser

topic: Context Browser
description: a tool for viewing knowledge in perspective - in context
</signature>

Received on Thursday, 16 November 2000 14:57:14 UTC