Re: From the Semantics TF: rdf:states

On 6 Aug 2024, at 10:09, Thomas Lörtsch <tl@rat.io> wrote:

° It has been suggested that the mapping between RDF-star and LPGs should use rdf:states instead rdf:reifies; but this introduces the possibly incomprehensible behaviour that distinct edges with the same label (type) between the same node identifiers will correspond to a unique triple in the edge, e.g.,
:e111 rdf:states <<( :john :worksFor :A )>>.
:e222 rdf:states <<( :john :worksFor :A )>>.
:e111 :sal 100K.
:e222 :sal 200K.

will induce the unique triple in the graph:
:john :worksFor :A.

I still don’t know what you mean by "induce". And by extension I don’t understand what you see as problematic in this example.

You said it doesn’t means that it entails, but then you provided an explanation using reification vocabulary. That didn’t help me understand either what you mean by "induce", because reification was the thing that I had asked you to explain first, and you explained it by introducing the term "induce". So we are going in circles here.

In this context, “induce” can be sen as synonym to “entails”.

What matters here is the substance of the argument; this was argued strongly by Souri.
According to Souri’s example (see the minutes of the meeting), if you have two distinct edges between the same pair of nodes:

:e111 rdf:states <<( :john :worksFor :A )>>.
:e222 rdf:states <<( :john :worksFor :A )>>.
:e111 :sal 100K.
:e222 :sal 200K.

then only a unique triple will be entailed in the graph::

:john :worksFor :A.

Souri argued that the above unique triple in the graph together with the two original distinct edges generates a clash of intuitions.

cheers
—e.