I am well aware of the way blank nodes are used, that they represent a locally-declared node with no IRI given. However, I am not sure how to interpret an actual logical "meaning" of a blank node. Does it have an identity? Is it different from any other blank node? Is it only an existential quantifier? What can be inferred from the corresponding knowledge base?
Let's say we have this simple document:
<iri1> <p> [ <p> <iri2> ] .
<iri1> <p> [ <p> <iri2> ] .
If the two blank nodes here only serve as existential quantifier, can it reduced to a single statement? Or does it mean there are (at least) two distinct entities, both in the same relation to <iri1> and <iri2>?
What about this?
<iri1> <p> [ <p> <iri2> ] .
<iri1> <p> <iri3> .
<iri3> <p> <iri2> .
Is the first line here meaningful?
My understanding is that unless the RDF document contains some sort of contradition when used with a logical framework (like OWL), then all the things not implied by the document can still be considered possible. This should also extend to owl:sameAs, and with that in mind it can be said that a blank node is just a regular node but with an "IRI" different from any other regular IRI.
The problem with this interpretation is that it implies the opposite thing that I assumed at first – since <iri1> and <iri2> in the first document are not explicitly made distinct via owl:differentFrom, they "could" be one and the same thing (another document is free to link them). However, nothing can link the two blank nodes, and since people don't usually link everything with owl:differentFrom, we can know these blank nodes are distinct from each other (and every other node in existence), since they can never be unified.
Which assumption is correct?