W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2004

Re: Troublesome relations

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Tue, 19 Oct 2004 09:53:10 -0400 (EDT)
Message-Id: <20041019.095310.11484503.pfps@research.bell-labs.com>
To: tpassin@comcast.net
Cc: www-rdf-interest@w3.org

From: "Thomas B. Passin" <tpassin@comcast.net>
Subject: Re: Troublesome relations
Date: Tue, 19 Oct 2004 08:55:36 -0400

> 
> Reto Bachmann-Gmuer wrote:
> 
> > I'm probably just missing the problem, but can't just any n-order 
> > predicate like predPQ(a, b, c) be reduced to _1(resPredPQ,a), 
> > _2(resPredPQ,b), _3(resPredPQ,c)?
> 
> No, not when the n-order predicate inherently represents a relation 
> between two or more things.  In your decomposition, any one of the three 
> reduced expressions could be changed or falsified without affecting the 
> other two.  But when the original predicate concerns related things, one 
> can't be changed without affecting the value of the other(s) as well.

Hmm.  I don't see what the problem is here.  It seems to me that it is just
as valid to argue that all relationships should be so represented, even
binary ones from the domain, leaving only the ``structural'' predicates
_1, _2, etc.  If you are worried about malformed relationships or changes,
then include axioms that prevent this.  Of course, this presumes that your
formalism can have such axioms.

> Say that there is a marriage between John and Sally.  The marriage as a 
> concept (or a legal state, for that matter) is concerned with a 
> relationship between the two people.  It cannot be carved up into binary 
> relationships without doing serious damage to the basic concept.  John 
> can be "married", and Sally can be "married" (binary predicates).  But 
> the thing that brings them into the relationship, the marriage, 
> inherently involves the two of them, and that is not binary.

This is not a particularly good example, as
	married(John,Sally)
looks rather binary, at least to me.  A different example, such as 
	married(John,Sally,Fido)
(for some trisexual species) would be more appropriate.

However, even then I don't see any impediment to representing this trinary
relationship as

	_:a rdf:type ex:marriage .
	_:a ex:groom ex:John .
	_:a ex:bride ex:Sally .
	_:a ex:dog ex:Fido .

or

	_:a rdf:type ex:marriage .
	_:a _1 ex:John .
	_:a _2 ex:Sally .
	_:a _3 ex:Fido .

_:a then represents the trinary relationship between John, Sally, and Fido.


> Cheers,
> 
> Tom P

Peter F. Patel-Schneider
Bell Labs Research
Received on Tuesday, 19 October 2004 13:46:22 GMT

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