W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > October 2001

RE: Properties not predicates (was Re: PRIMER: draft data model section)

From: Pat Hayes <phayes@ai.uwf.edu>
Date: Thu, 25 Oct 2001 10:28:33 -0500
Message-Id: <p0510108ab7fdd93147d5@[205.160.76.193]>
To: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>
Cc: w3c-rdfcore-wg@w3.org
>  > It was for this, if for nothing else, that I wanted us to have a
>>  model theory.
>>  Please, please, lets try to stick with the concepts we have a formal
>>  mathematical definition for in the model theory.  We have nodes
>>  and arcs and
>>  labels.  What does reification mean in those terms?
>>
>
>It's just triples, (i.e. nodes, arcs and labels).
>
>We can reify an arc (A) in the graph:
>- it's reification is a node
>- that node has four arcs leaving it
>    - one, labelled <rdf:subject>,  pointing to the source of A,
>    - another labelled <rdf:object>, pointing to the pointy end of A,
>    - another, labelled <rdf:type> pointing to <rdf:Statement>,
>    - and the last one labelled <rdf:predicate> pointing to a node with the
>same label as the label of A.
>
>That's it, there is no more.

OK, but that does not mean what the M&S says reification ought to 
mean, if those four arcs are interpreted according the RDF rules.

Suppose A is the triple

aaa  bbb  ccc .

What that means, we are all agreed, is that the property called 'bbb' 
has the value called 'ccc' when applied to the thing called 'aaa', 
right? These labels are being USED to REFER TO things outside the 
graph - resources and properties with extensions, things like that. 
Suppose for example we are interpreting 'aaa' to refer to me, 'ccc' 
to refer to a rather large integer, and 'bbb' to refer to the 
relation between US residents and their SS numbers.

Suppose the reification is a node we can refer to by the nodeID 
_:reif.  If the arc labelled <rdf:subject> points to the source of A, 
then that arc says that a relation called 'rdf:subject' has the value 
called 'aaa' when applied to _:reif. But the value called 'aaa' is 
me, Pat Hayes, this thing typing on the keyboard, made of flesh and 
blood. So this reification says that the subject of _:reif is a 
person. It also says that the rdf:object of _:reif is a number. So 
this isn't about the parts of the triple at all; it is about the same 
things that those parts were about when they were being used in the 
triple, because they are *still* being used in that triple. The very 
same node cannot mean two things at once. [Added later. Well, maybe 
it can. If that is really what people want reification to be like, 
then we can put that meaning into the MT; but it will require major 
scaffolding to put it there, rather like the machinery needed for 
datatyping.]

In order to get the M&S notion of reification, you cannot USE the 
parts of A themselves to refer to the parts of A, since they are 
already being used to refer to something else. You need to have three 
new things in the reification which somehow, by some magical process, 
are known to REFER TO  the parts of A. That process is nowhere 
described, and that lack of any connection between the actual triple 
A, and the things in the reification of A that are supposed to refer 
to the parts of A, is what makes reification as described in the M&S 
so mysterious.


>            ================
>
>It's like praying: you open your mouth and you say things. Descriptions of
>praying that go further than that open up more questions than answers.
>
>The single question of "why?" is more addressable than any more
>philosophical issues.
>
>Perhaps I miss the point about reification, just as some would say that I
>miss the point about praying ...
>
>I have found this thread unedifying :(.
>
>My vote seems to go for no Model Theory concerning reification; it's a
>schema thing.

What is that supposed to mean?

Look, model theory is just being clear about meanings. If you have 
something central in the language that is outside the semantic 
theory, you are just admitting to confusion. If you (or anyone) can 
come up with a coherent story about what this kind of reification is 
supposed to mean, I can build that meaning into the MT.  But right 
now, there doesn't seem to be a coherent story.

Pat

-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes
Received on Thursday, 25 October 2001 11:28:38 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:41:12 EDT