Re: Blank nodes semantics - existential variables?

On 6/29/2020 4:17 PM, thomas lörtsch wrote:
> I used a rethoric figure, I think it’s called reductio ad absurdum. I want to say that assuming all different existentials may point to one and the same thing leads to rather non sensical interpretations like the one you just gave. Which is why I think it’s very unintuitive.

Another way to think about this subject is that it is all about the 
identification of nodes, whether bnodes or not.  There are basically two 
ways to identify a "thing":  by its intrinsic identifier (its URI for 
rdf graphs) or by its properties.  In fact, I don't see why a node's URI 
identifier couldn't be considered just another property.

Identifying a node by its collection of properties may give a unique 
identity or it may not, depending on those properties.  Since any given 
collection of properties must be finite - we can't write down an 
infinitely large rdf graph - and since presumably there could be an 
indefinitely large number of unique nodes, it may be that most bnodes in 
a specific graph will not be unique under an open world assumption.  But 
that does not imply that any specific bnode is or is not unique.

In view of the above, there is no point in thinking about a default case 
for the distinctness of different bnodes.  This would have to be 
established in each case for the actual bnodes by their actual properties.

 > what are the chances that I just repetaed my self versus that I want 
to express that there are two different things that Bob has and that are 
worth mentioning?

How can we know what those chances are, even if we were willing to use 
probability to establish logical conclusions?  Suppose, for example, 
that a witness to a store robbery reports seeing a large man breaking 
the store's window at 11 AM, and another witness reports seeing an 
average size man running out the store's door at 11:05.  Each of them 
has seen a man, but were they both the same man?  It is impossible to 
tell without more information.  They might have seen the same person, 
they might not.  It's not a matter of probability, it's a matter of 

Tom P

Received on Monday, 29 June 2020 21:35:59 UTC