Re: Decision from the Semantics TF: liberal baseline

On 8 Jan 2025, at 14:16, Doerthe Arndt <doerthe.arndt@tu-dresden.de> wrote:

As a matter of fact, while the reif3 entailment pattern does provide this special case for rdf:refies, I don't think that the semantic condition on RDF interpretation does!... It says

    <x, [I+A](rdf:Proposition)> ∈ IEXT([I+A](rdf:type))
          if ∃ y . <y,x> ∈ IEXT([I+A](rdf:reifies))

and IEXT([I+A](rdf:reifies)) only covers asserted triples with the rdf:reifies predicate.

Interesting. You are right. That is a good argument to drop the „super-range“ :)  (but let’s see what other say)

My bad.
Indeed the definition does not do what was meant to do: it just work for asserted triples with the rdf:reifies predicate.
At this point, I will revert the definition for triple terms being of type rdf:Proposition to hold only for triple terms appearing as subject or object in asserted triples.
I hope that now the document https://github.com/w3c/rdf-star-wg/wiki/RDF-star-%22liberal-baseline%22 is consistent.
Can you please check it?

I understand that the gist of this thread is that subjects, predicates, objects in a triple term (at any level of nesting) have the same denotation — namely the value of [I+A](.) — as if they were appearing as subjects, predicates, objects in top-level asserted triples, but nothing else; this is “transparency”.
If those subjects, predicates, objects are mentioned ONLY within triple terms, then they will not have any inferred property at all (including metamodelling properties).
The only inferred properties that those subjects, predicates, objects in a triple term (at any level of nesting) may have, come from other asserted top-level triples mentioning them.

Would you agree?
—e.