Re: Triple terms in subject position

On 7 Jul 2025, at 14:36, Pierre-Antoine Champin <pierre-antoine@w3.org> wrote:
On 03/07/2025 18:13, Colby Russell wrote:
Simple solution: recognize that reified triples belong to purview of
literals--i.e., they _are_ literals.  (Literally.  That's _exactly_ what
they are--and always have been--or *should* have been, at least...)

I sympathize with this position, and I would have advocated for it if we dealt only with ground<https://www.w3.org/TR/rdf12-concepts/#dfn-ground> triple terms.

Even with only ground triple terms, a literal-based approach could not deal with transparent triple terms, as it is now in RDF 1.2.
That is,  IRIs and literals within a triple term do denote resources and these contribute to the overall meaning of the triple term.
So, your proposal heads in a completely different direction.
cheers
—e.


Unfortunately, triples (and therefore triple terms) can contain blank nodes,
and this prevents us from considering arbitrary triple terms as literals.

  best

Instead of accommodating reification through special syntax, they need
only a standard datatype to tag them appropriately.

The correct disposition for this issue follows straightforwardly from
this--an example where the relationship between policy and mechanism is
policy from mechanism--since literals can only appear in the object
position, then any given reified triple, being that which is expressed
as a literal, must also.  (At least for as long as RDF imposes this
constraint on literals of any type--a decision which may need to be
revisited.)