# Re: Interpretation of RDF reification

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Wed, 22 Mar 2006 18:33:05 -0500 (EST)
Message-Id: <20060322.183305.98841654.pfps@research.bell-labs.com>

```
From: "McBride, Brian" <brian.mcbride@hp.com>
Subject: RE: Interpretation of RDF reification
Date: Wed, 22 Mar 2006 22:48:34 -0000

>
>
>
> >
> > Ah, I see. This was very clear. So X represents a stating,
> > and there can be many statings of the same statement, in the
> > sense that I could state it today, and you may have stated it
> > yesterday, giving us two statings of one statement?
>
> Yes.
>
> >
> > > What do you intend to do with this "marriage relationship"?  To
> > > understand your use case, I'd like to know what sort of statements
> > > you'd like to make about the relatioship.
> >
> > Basically, the use case is to be able to state marriages as a
> > single triple (for convenience, say), while still retaining
> > the ability to talk about the marriage, and retaining the
> > connection between the triple and the marriage node.

Ah, well, this is *one* of the things that is lacking in the RDF account of
reification.  There is no connection between a reification and the triple it
reifies, and certainly no connection to an occurence (whatever "occurence"
might mean here) of the triple.

> I would need help from a logician like Peter or Pat on that one.
>
> Roughly, it would seem that there is the concept of the property,
> married, a relation which relates pairs of resource.
>
> Then there is the concept of Marriage which is a class.  For simplicity
> lets say there is a subclass of marriages, conventional marriages,
> Cmarriage, to allow for variations.  Cmarriages have a husband property
> and a wife property, a start date property and end date property and may
> have other properties.
>
> It seems like there should be some property that relates the class
> Married to the property married.  It would seem that husband and wife
> properties are distinguished in relation to Married and married from
> properties like startDate and endDate.

Well, if you want to go to FOL or an expressive ontology language you could do
most of this, but RDF hasn't anywhere near the expressive power required.

> I personal[l]y have not come across this before, but it feels like it
> should be a common pattern that ontologists would be familiar with.
> Maybe even a pattern that could be documented, as in the work of the OEP