[om-list] Relation Aspects (corrected)

[Mark Butler]

Relation Aspects

As I mentioned before, I have implemented my data model using a rule 
that there is only one relation between any two concepts.  The relation 
itself is an abstraction - the nature of the relation is represented by 
various evalative predicates from various commentators about the 
relation.  Some might deny the existence of the relation all together, 
some might affirm some aspects, but not others, and so on.

This has some very interesting consequences that I haven't become 
entirely used to yet.  For example, I might be tempted to create a class 
called Belief that is an sub class of Relation.  However, in order to 
guarantee that only one relation exists between any two concepts, the 
relation object cannot be characterized by programmatic inheritance.  
Characterizing a relation as only a belief jumps the gun, because that 
relation may have other aspects besides belief.   

A relation between any two persons may have several independent aspects, 
including parenthood, friendship, belief, love, trust, loyalty, and so 
on.  All of those aspects have to be treated independently of the 
relation object itself under the rule I have adopted. That doesn't mean 
that the real world relation between two entities does not have 
objective character, but rather that  the character of concepts and 
relations should be represented explicitly rather than implicitly.

   A relation aspect is very similar to a one of the common definitions 
of the word relationship - e.g. the friendship between A and B.   
However, relation aspects are also identical with the concept of an 
relation attribute, so no additional classes are needed.

By the way, it seems to me that any symmetrical relationship must be an 
emergent aspect of two asymmetrical relationships.  For example, 
friendship between A and B is symmetrical, but it cannot exist on its 
own, rather it is contingent on A's regard for B and B's regard for A.  
If either aspect is missing, the relation between the two cannot be 
characterized as a true friendship. 

I like to think of a relation betwen two objects as consisting of light 
flowing from one to the other and vice versa.  One wouldn't normally 
confuse a photon travelling from A to B with one from B to A.  Any 
transitive attribute (or aspect) of the relation must be characterized 
in each direction independently.  It also seems that any symmetrical 
aspect must be an emergent property of a one or more transitive 
aspects.  Can a symmetrical relationship be grounded any other way?

 - Mark


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://six.pairlist.net/pipermail/om-list/attachments/20031015/5c4f831b/attachment.html