Re: use/mention and reification

On 2002-01-23 23:06, "ext Dan Brickley" <danbri@w3.org> wrote:

> ... if *everyone* had faultless access to the meaning of
> each and every URI name, I wouldn't have my current concerns about
> reification.

Quite so. My own puzzlement was how treating reification as
quotation such that subjects, predicates, etc. are expressed
as literal URIs rather than URI labeled resources helps us
in either determining the degree of misunderstanding or
rectifying it. I see it as a matter of interpretation, not
of syntax. What is missing is the distinctions that would
enable one to employ specific interpretations.

It seems that we really have three types of expression
being discussed here:

  1. assertion    (John asserts) the sky is blue.
  2. statement    (Mary says) the sky is blue.
  3. quotation    (Joe utters) "the sky is blue".

Assertion corresponds to the usual way of asserting
statements in RDF:

  <rdf:Description rdf:ID="#sky">
     <is rdf:resource="#blue"/>
  </rdf:Description>

Non-asserted statements appear to correspond to the
present method of statement reification:

  <rdf:Statement>
     <rdf:subject rdf:resource="#sky"/>
     <rdf:predicate rdf:resource="#is"/>
     <rdf:object rdf:resource="#blue"/>
  </rdf:Statement>

Quoted statements seem either unsupported at present
or at best could be treated as RDF/XML fragments
in literals.

Perhaps one approach to addressing these three types
of statements is to give them formal expression in
RDF:

  <rdfs:Class rdf:about="rdf:Statement"/>

  <rdfs:Class rdf:about="rdf:Assertion>
     <rdfs:subClassOf rdf:resource="rdf:Statement/>
  </rdfs:Class>

  <rdfs:Class rdf:about="rdf:Quotation>
     <rdfs:subClassOf rdf:resource="rdf:Statement/>
  </rdfs:Class>

Then, we can express the distinction between statement
types by rdf:type:

  <rdf:Assertion>
     <rdf:subject rdf:resource="#sky"/>
     <rdf:predicate rdf:resource="#is"/>
     <rdf:object rdf:resource="#blue"/>
     <rdfx:authority rdf:resource="#John"/>
  </rdf:Assertion>

  <rdf:Statement>
     <rdf:subject rdf:resource="#sky"/>
     <rdf:predicate rdf:resource="#is"/>
     <rdf:object rdf:resource="#blue"/>
     <rdfx:authority rdf:resource="#Mary"/>
  </rdf:Statement>

  <rdf:Quotation>
     <rdf:subject rdf:resource="#sky"/>
     <rdf:predicate rdf:resource="#is"/>
     <rdf:object rdf:resource="#blue"/>
     <rdfx:authority rdf:resource="#Joe"/>
  </rdf:Quotation>

where the serialized form

  <rdf:Description rdf:ID="#sky">
     <is rdf:resource="#blue"/>
  </rdf:Description>

is simply a shorthand or synonym for the
more explicit rdf:Assertion form.

Then, applications can differentiate as needed between
assertions, statements, and quotations, with the
default likely being only considering assertions for
queries.

Yet, in all cases, we have a consistent graph representation
which also allows us examine, based on the intersection
of resource nodes in all statement types, how much different
folks views/beliefs correlate.

Eh?

Patrick

--
               
Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email: patrick.stickler@nokia.com

Received on Thursday, 24 January 2002 10:40:08 UTC