W3C home > Mailing lists > Public > www-rdf-logic@w3.org > June 2001

Re: Reification as nesting

From: Brian McBride <bwm@hplb.hpl.hp.com>
Date: Tue, 19 Jun 2001 19:07:26 +0100
Message-ID: <3B2F94DE.45A2ECCC@hplb.hpl.hp.com>
To: pat hayes <phayes@ai.uwf.edu>
CC: Drew McDermott <drew.mcdermott@yale.edu>, www-rdf-logic@w3.org
Hi Pat,

This mailing list is WG only.  Please don't include non-WG members on
messages to this list - it only invites them to reply to the list.

pat hayes wrote:
[...]
> I took it that the 'graph' described in M&S was not an entity, but a
> way of thinking about RDF as a whole. A kind of graphical abstract
> syntax for the entire language, rather than a syntactic construct.

Others milage may vary, but M&S, in the section on reification, talks
about whether a statement is in the model or not.  That gives us
the opportunity to formalize that concept, ref:

  http://www.w3.org/2000/03/rdf-tracking/#rdfms-graph

[...]
> The issue however is whether these sets can themselves be
> incorporated into sentences. Adding that ability changes the language
> profoundly.

Right.  And I haven't go that one clear in my mind, but unfortunately
I don't think it takes us that far.  But I'd rather not get into
that till we have a formal semantics or we'll just misunderstand each
other (= i'll misunderstand you).

> 
> BTW, a problem with the above (for me) is that when used in the ways
> I want to use them, many of the triples inside such a set are not
> statements - they don't have truthvalues or express a proposition.
> Maybe this is just a terminological quibble, but it does have a
> nonterminological consequence when trying to define a coherent
> semantics, because we need a way to distinguish the 'real' statements
> from the mere parts of other, larger, statements.

Whilst I once hoped this might be a way to do what I think you are
suggesting, I don't think its sufficient unless we have a really
bent semantics.  But please can we have a formal semantics for basic
RDF before we get into this.

> 
> >N3 introduces a syntax for embedding the respresention of such a set
> >of statements (N3 calls them contexts) inside the representation of another,
> >a feature RDF/XML current lacks.
> 
> Well, sorry again to quibble,

Forgive me, but if you were sorry, you wouldn't do it. :)

> but N3 didnt *introduce* such a syntax.
> If anyone did, it was probably one of Peirce, Frege or Aristotle,
> although one could make a pretty good case that it was the first
> proto-human who ever thought of writing down speech, probably in
> ancient Sumeria.

I'm just back from an hour standing in the corner and that dunce's cap
is a very tight fit.

> 
> BTW, I have yet to discover what exactly a 'context' is in N3. Is N3
> supposed to be syntactic sugar for a particular way of writing RDF,
> or is it a different language altogether?

I will defer to those who know n3 better than I do, however I too find
contexts in N3 confusing.  When I last looked they were defined to be
sets of statements in N3.  What confused me was that this was said to
be the same thing as a bag of reified statements in RDF/XML.

Brian
Received on Tuesday, 19 June 2001 14:09:17 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:40 GMT