Add new terminology document

[niklasl]

The draft resulting from the RDF-star Triple-Edge subgroup meeting on 2024-01-12.

To be furthered worked on (in this draft PR online, and/or during subsequent meetings).

• Do we want this as a new document (and not to be worked into the existing rdf-terminology.md)?
• Is the title and file name OK as is?

[gkellogg]

Perhaps we should continue to use reification terminology and call this a Triple Token.

[afs]

This is the key piece of terminology. It is not going to be easy to find the right word(s).

A triple is a thing in the RDF data model. We have been saying it is a "type" in the sense of a abstract descriptive concept. Is "triple" a class? It hasn't been described like that in RDF to-date. https://www.w3.org/TR/rdf12-concepts/#section-triples

If occurrence is part of the RDF data model, the "token" quality seems more like a named occurrence. A named occurrence is a resource (or a blank node that is mapped to resources by "A", the bnode mapping). There can be several named occurrences (resources) for one occurrence.

When we can define terminology, we do need to consider if the word is already in use in RDF. "Instance" has been suggested but in RDF that brings in "an instance of a class".

Currently, I prefer a name that is not taken from this abstract theoretical framework; rather one that is closer to how it it is used in RDF. Words like "occurrence", "usage", "mention" - all of which bring their own implication-baggage, including they should be "possible occurrence" etc because of itself, it doesn't assume the triple is in-use/asserted anywhere. I doubt we'll find a perfect word.

[TallTed]

re: "instance", also consider "instance data"

[hartig]

I am proposing an alternative term here: "Triple Membership". I see two advantages of this term:

1. It can be used as a noun (like it is also possible for terms such as "triple occurrence" and "triple instance"). In other words, it allows us to talk about making statements about a membership of a triple (in some graph), which is not possible for the term "assertion" because the word "assertion" may also be understood as the action of asserting.
2. It does not carry any meaning with respect to semantics. In contrast, the term "asserted triple" already implies something regarding the semantics. (Note that I am not saying that we should revisit the question of whether a triple that is a member/element of a graph is considered as asserted within this graph. I am just saying that the term "membership" allows us to talk about triples in graphs in a semantically neutral way. The property of a triple being asserted in a graph is something that we define based on the property of a triple being a member of a graph.)

[niklasl]

I think this is a valuable distinction. A graph is defined as a set of triples, so certainly each triple is a member of that set. With that, we should also:

1. State that "When a triple is a member of a graph, the statement it represents is asserted."
2. Define "Asserted Statement" as "The statement is asserted with respect to the meaning of the graph."

[afs]

Text from RDF 1.2 Concepts

This hasn't changed from RDF 1.1:

https://www.w3.org/TR/rdf12-concepts/#resources-and-statements

Asserting an RDF triple says that some relationship, indicated by the predicate, holds between the resources denoted by the subject and object.

New text in RDF 1.2 Concepts (added 2023-04-27): w3c/rdf-concepts@1d02fff

https://www.w3.org/TR/rdf12-concepts/#section-triples:

An asserted triple is an RDF triple that is an element of an RDF graph.

[niklasl]

Thank you @afs. Then we should not change the terminology (i.e. keep "Asserted triple" and definition; in fact we should just reference the definition that is in 1.2).

Looking at https://www.w3.org/TR/rdf12-concepts/#entailment:

An RDF triple encodes a statement—a simple logical expression, or claim about the world.

and again at https://www.w3.org/TR/rdf12-concepts/#resources-and-statements:

This statement corresponding to an RDF triple is known as an RDF statement.

I see three aspects to this:

1. Abstract syntax: the triple is an element of the graph.
2. Formal semantics: the triple is asserted.
3. Expression: the meaning of the statement (i.e. the relationship in the universe of discourse) that the asserted triple encodes (within the context of the graph, given a chosen interpretation).

(In suggesting "asserted statement" I conflated 2 and 3. I'm not sure if I differentiate 1 and 2 correctly.)

[afs]

Do we want this as a new document (and not to be worked into the existing rdf-terminology.md)?

IMO The scope of this document is terminology for, and in support of, the strawman proposal.

rdf-terminology.md is the defined terms in "concepts" and "semantics" (The URL fragments are #dfn-). The link text "RDF 1.1 ..." should "RDF 1.2 ..." because the links are to github WG documents.

So keep this new document; it is a working document for the WG.

Some terminology may be added to rdf-terminology.md when the output of the strawman proposal goes into the RDF specs and if rdf-terminology.md extends beyond the lifetime of this WG.