On Thu, Jul 24, 2014 at 4:03 AM, Richard Light <[email protected]> wrote:
> > I must say that I'm not so sure that named graphs are going to be > particularly useful for implementations of the CRM. As I understand it > (and I don't claim to be an RDF expert), the idea of quads was invented so > that "naked" RDF assertions could be given a "context". The problem I have > always had with that idea is that you only get one shot at it (i.e. you can > only assign one context to any given triple). > A triple is a true proposition*; duplicates are redundant (A and A <-> A). However, there can be multiple speech acts asserting that the proposition is true. There are ways of giving semantics to named graphs that enable that; however, the semantics of named graphs were deliberately left underspecified (a decision that was not uncontroversial). In the end, what was published was a Working Group Note listing some of the possibilities that were argued for - see: http://www.w3.org/TR/rdf11-datasets/ . There are other possible semantics that named graphs might have ; for example, the name in a named graph might denote some graph containing the reified forms of the statements in the graph part of the named graph. This differs from the quotational semantics given in §3.7 of the note cited above given the presence of blank nodes - ("one does not simply quantify into quoted contexts!"). Since the CRM does not require that the propositional content of an propositional object be true, it might be possible to avoid these questions by dealing with Graphs (as sets of propositions), and assertions of the contents of those Graphs directly . Simon * which is why, now that RDF 1.1 make any triple will an ill-typed literal false, any graph that contains such triple is inconsistent.
