No connection between propositions and facts in model-theoretic semantics #144

rat10 · 2025-07-17T12:29:05Z

Add clarification that annotations on propositions can't annotate facts according to the model-theoretic semantics, but that an operational semantics may nudge users towards assuming such a connection.

Preview | Diff

…tics add clarification that annotations on propositions can't annotate facts according to the model-theoretic semantics, but that an operational semantics may nudge users towards assuming such a connection

Add clarification that annotations on propositions can't annotate facts according to the model-theoretic semantics, but that an operational semantics may nudge users towards assuming such a connection.

franconi · 2025-07-17T13:04:44Z

The RDF Semantics spec never mentions the notion of reification; the spec only introduces a special rdf:reifies property, whose only characteristic is to have a range as the special rdfs:proposition class. (Observe that Appendix D is going to be axed from this document). So, in this context your added paragraph is not appropriate.

pchampin · 2025-07-17T13:42:25Z

The RDF Semantics spec never mentions the notion of reification

It does, but only for the old-style reification. So the use of "reification" in this PR is not consistent.

Similarly, the use of "(reified) proposition" is problematic. Does it refer to (1) a proposition that happens to have an incoming rdf:reifies relationship, (2) the denotee of the reifier at the other end of this relationship, or (3) the reifier itself. Grammatically, I would assume (1), but from previous discussions with @rat10, I assume it is rather (2) or (3). Either way, I disagree with this sentence: by definition, facts are propositions (the set of facts F(I) is a subset of the set of propositions IPR(I)). In interpretation (1), the connection between a proposition and a fact of the same form is the identity. In interpretation (2) and (3), the connection between the denotee of the reifier and the fact (which is identical to the proposition) is captured in IEXT(I(rdf:reifies)).

I have concerns about the 2nd part of the PR (discussion about "identification" vs. "denotation"), but they depend on which of the interpretations above – (1), (2) or (3) – is correct...

franconi · 2025-07-17T13:44:56Z

The RDF Semantics spec never mentions the notion of reification

It does, but only for the old-style reification. So the use of "reification" in this PR is not consistent.

Appendix D will disappear from the Semantics document.

franconi · 2025-07-17T13:53:14Z

Similarly, the use of "(reified) proposition" is problematic. Does it refer to (1) a proposition that happens to have an incoming rdf:reifies relationship, (2) the denotee of the reifier at the other end of this relationship, or (3) the reifier itself. Grammatically, I would assume (1), but from previous discussions with @rat10, I assume it is rather (2) or (3). Either way, I disagree with this sentence: by definition, facts are propositions (the set of facts F(I) is a subset of the set of propositions IPR(I)). In interpretation (1), the connection between a proposition and a fact of the same form is the identity. In interpretation (2) and (3), the connection between the denotee of the reifier and the fact (which is identical to the proposition) is captured in IEXT(I(rdf:reifies)).

This is exactly what I said in my comment to the PR in Concepts #220.

pfps · 2025-07-17T15:46:09Z

I propose closing this PR without merging it.

rat10 · 2025-07-17T15:56:35Z

Similarly, the use of "(reified) proposition" is problematic. Does it refer to (1) a proposition that happens to have an incoming rdf:reifies relationship, (2) the denotee of the reifier at the other end of this relationship, or (3) the reifier itself. Grammatically, I would assume (1), but from previous discussions with @rat10, I assume it is rather (2) or (3). Either way, I disagree with this sentence: by definition, facts are propositions (the set of facts F(I) is a subset of the set of propositions IPR(I)). In interpretation (1), the connection between a proposition and a fact of the same form is the identity. In interpretation (2) and (3), the connection between the denotee of the reifier and the fact (which is identical to the proposition) is captured in IEXT(I(rdf:reifies)).

This is exactly what I said in my comment to the PR in Concepts #220.

In that comment you gave an explanation that seems helpful. Would you turn that into a PR as an addition to the Semantics spec, or should I give it a try?

franconi · 2025-07-17T16:02:59Z

This is exactly what I said in my comment to the PR in Concepts #220.

In that comment you gave an explanation that seems helpful. Would you turn that into a PR as an addition to the Semantics spec, or should I give it a try?

I believe that such comment is not appropriate in the Semantics document.

pfps

I am against this change.

franconi

This PR should be cancelled.

TallTed · 2025-07-17T22:47:13Z

This is exactly what I said in my comment to the PR in Concepts w3c/rdf-concepts#220 (comment).

In that comment you gave an explanation that seems helpful. Would you turn that into a PR as an addition to the Semantics spec, or should I give it a try?

I believe that such comment is not appropriate in the Semantics document.

Might it be appropriate in another of our 20ish documents (e.g., Concepts)? If it aids comprehension of the intended meanings of things, such that both @franconi and @rat10 agree on the meaning of both the comment (in Concepts) and the text in Semantics that is informed by the comment?

franconi · 2025-07-17T22:59:03Z

I believe that such comment is not appropriate in the Semantics document.

Might it be appropriateu in another of our 20ish documents (e.g., Concepts)? If it aids comprehension of the intended meanings of things, such that both @franconi and @rat10 agree on the meaning of both the comment (in Concepts) and the text in Semantics that is informed by the comment?

You may remember (it’s in the minutes) that in the meeting last week I volunteered to craft the core of a sort of F.A.Q. note/document, stemming from the examples file we already have and from potential misunderstandings (like ph) and from clarification requests such as @rat10.

lisp · 2025-07-18T01:09:56Z

The RDF Semantics spec never mentions the notion of reification; the spec only introduces a special rdf:reifies property, whose only characteristic is to have a range as the special rdfs:proposition class. (Observe that Appendix D is going to be axed from this document). So, in this context your added paragraph is not appropriate.

in the semantics document, the combination with "reifies" in relation to "Proposition" among the axiomatic triples establishes reification as a relevant notion.
should the editors choose to excise the word "reification" entirely from the semantics document that will not improve the exposition.

TallTed

Grammar, punctuation, and such

TallTed · 2025-07-25T18:07:33Z

spec/index.html

@@ -710,7 +710,7 @@ <h3>Properties of simple entailment and satisfiability</h3>
    <p class="fact"> If E contains an IRI which does not occur anywhere in S,
      then S does not simply entail E.</p>

-    <p>The following semantic properties relate triple terms and triples asserted in a graph, and they introduce a general definition of satisfiability.</p>
+    <p>The following semantic properties relate triple terms, triples asserted in a graph and reified triples, and they introduce a general definition of satisfiability.</p>


Suggested change

<p>The following semantic properties relate triple terms, triples asserted in a graph and reified triples, and they introduce a general definition of satisfiability.</p>

<p>The following semantic properties relate triple terms, triples asserted in a graph, and reified triples, and they introduce a general definition of satisfiability.</p>

TallTed · 2025-07-25T18:16:27Z

spec/index.html

@@ -1969,8 +1969,52 @@ <h2 id="non_semantics">RDF reification, containers and collections</h2>
    processes to check formal RDF entailment. For example, implementations may decide
    to use special procedural techniques to implement the RDF collection vocabulary.</p>

+
+  <section id="TTerms">
+    <h3>RDF 1.2 reification - triple terms and reifiers</h3>


Suggested change

<h3>RDF 1.2 reification - triple terms and reifiers</h3>

<h3>RDF 1.2 reification — triple terms and reifiers</h3>

TallTed · 2025-07-25T18:17:46Z

spec/index.html

+  <section id="TTerms">
+    <h3>RDF 1.2 reification - triple terms and reifiers</h3>
+    <p>
+	To repeat nomenclatura:


Suggested change

To repeat nomenclatura:

To repeat our nomenclature:

TallTed · 2025-07-25T18:18:06Z

spec/index.html

+	To repeat nomenclatura:
+	<ul>
+		<li>an `rdfs:Proposition` subsumes multiple kinds of triples: 
+	    		abstract triples, asserted triples and reified triples


Suggested change

abstract triples, asserted triples and reified triples

abstract triples, asserted triples, and reified triples

TallTed · 2025-07-25T18:18:43Z

spec/index.html

+		</li>
+		<li>an abstract triple is encoded as a triple term</li>
+		<li>an asserted triple is also called a fact, 
+	    		and also known as an RDF statement 


Suggested change

and also known as an RDF statement

and is also known as an RDF statement,

TallTed · 2025-07-25T18:20:13Z

spec/index.html

+	    never entails that triple as a fact.
+	Neither does a fact entail a reification of that triple.


Suggested change

never entails that triple as a fact.

Neither does a fact entail a reification of that triple.

never entails that triple as a fact,

nor does a fact entail a reification of that triple.

TallTed · 2025-07-25T18:21:28Z

spec/index.html

+	From that follows that in a strict interpretation of the model-theoretic semantics of RDF 1.2
+	    an assertion on a reified triple (denoted by a reifier) 
+	    can never be an assertion on a fact asserting that same triple.


Suggested change

From that follows that in a strict interpretation of the model-theoretic semantics of RDF 1.2

an assertion on a reified triple (denoted by a reifier)

can never be an assertion on a fact asserting that same triple.

It follows that, in a strict interpretation of the model-theoretic semantics of RDF 1.2,

an assertion on a reified triple (denoted by a reifier)

can never be an assertion on a fact asserting that same triple.

TallTed · 2025-07-25T18:23:15Z

spec/index.html

+	A looser interpretation of that connection 
+	    as one of <a href="https://w3c.github.io/rdf-semantics/spec/#dfn-identify">identification</a>, 
+	    not <a href="https://www.w3.org/TR/rdf12-concepts/#dfn-denote">denotation</a>, 
+	    as applied in RDF 1.2 Concepts, RDF 1.2 Primer and the RDF 1.2 note on triple terms (tbd),
+	    establishes an operational semantics of such a connection between reification and fact 
+	    as convention and best practice.


Suggested change

A looser interpretation of that connection

as one of <a href="https://w3c.github.io/rdf-semantics/spec/#dfn-identify">identification</a>,

not <a href="https://www.w3.org/TR/rdf12-concepts/#dfn-denote">denotation</a>,

as applied in RDF 1.2 Concepts, RDF 1.2 Primer and the RDF 1.2 note on triple terms (tbd),

establishes an operational semantics of such a connection between reification and fact

as convention and best practice.

A looser interpretation of that connection

as one of <a href="https://w3c.github.io/rdf-semantics/spec/#dfn-identify">identification</a>,

not of <a href="https://www.w3.org/TR/rdf12-concepts/#dfn-denote">denotation</a>,

as applied in RDF 1.2 Concepts, RDF 1.2 Primer, and the RDF 1.2 note on triple terms (tbd),

establishes an operational semantics of such a connection between reification and fact

as convention and best practice.

TallTed · 2025-07-25T18:24:47Z

spec/index.html

+	This design was chosen to facilitate assertions on asserted triples, a.k.a. "statements about statements",
+	    while keeping the model-theoretic semantics of RDF 1.2 simple 
+	    and upholding a safe distance from modal logic complications.


Suggested change

This design was chosen to facilitate assertions on asserted triples, a.k.a. "statements about statements",

while keeping the model-theoretic semantics of RDF 1.2 simple

and upholding a safe distance from modal logic complications.

This design was chosen to facilitate assertions on asserted triples,

also known as "statements about statements",

while keeping the model-theoretic semantics of RDF 1.2 simple

and upholding a safe distance from modal-logic complications.

TallTed · 2025-07-25T18:25:17Z

spec/index.html

  <section id="Reif">
-    <h3>Reification</h3>
+    <h3>RDF 1.0/1.1 reification - statement quad reification </h3>


Suggested change

<h3>RDF 1.0/1.1 reification - statement quad reification </h3>

<h3>RDF 1.0/1.1 reification — statement quad reification </h3>

rat10 added 3 commits July 17, 2025 13:57

no connection between propositions and facts in model-theoretic seman…

b5c741b

…tics add clarification that annotations on propositions can't annotate facts according to the model-theoretic semantics, but that an operational semantics may nudge users towards assuming such a connection

Merge pull request #1 from rat10/rat10-patch-1

48d6170

Add clarification that annotations on propositions can't annotate facts according to the model-theoretic semantics, but that an operational semantics may nudge users towards assuming such a connection.

fixed wording

1624502

rat10 mentioned this pull request Jul 17, 2025

definition of reifiers is non-normative and seems vague w3c/rdf-star-wg#169

Open

pfps added the spec:enhancement Change to enhance the spec without affecting conformance (class 2) –see also spec:editorial label Jul 17, 2025

rat10 mentioned this pull request Jul 17, 2025

Annotations on assserted triples are based on operational semantics w3c/rdf-concepts#220

Open

pfps requested review from pfps and franconi July 17, 2025 16:25

pfps requested changes Jul 17, 2025

View reviewed changes

franconi reviewed Jul 17, 2025

View reviewed changes

rat10 marked this pull request as draft July 23, 2025 15:54

add definition of reification, move comment into appendix

8505cea

TallTed suggested changes Jul 25, 2025

View reviewed changes

pfps added the ms:CR Milestone: Candidate Recommendation label Jul 31, 2025

	<p>The following semantic properties relate triple terms, triples asserted in a graph and reified triples, and they introduce a general definition of satisfiability.</p>
	<p>The following semantic properties relate triple terms, triples asserted in a graph, and reified triples, and they introduce a general definition of satisfiability.</p>

	<h3>RDF 1.2 reification - triple terms and reifiers</h3>
	<h3>RDF 1.2 reification — triple terms and reifiers</h3>

	abstract triples, asserted triples and reified triples
	abstract triples, asserted triples, and reified triples

	and also known as an RDF statement
	and is also known as an RDF statement,

		never entails that triple as a fact.
		Neither does a fact entail a reification of that triple.

	<h3>RDF 1.0/1.1 reification - statement quad reification </h3>
	<h3>RDF 1.0/1.1 reification — statement quad reification </h3>

No connection between propositions and facts in model-theoretic semantics #144

Are you sure you want to change the base?

No connection between propositions and facts in model-theoretic semantics #144

Uh oh!

Conversation

rat10 commented Jul 17, 2025 • edited by pr-preview bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

franconi commented Jul 17, 2025

Uh oh!

pchampin commented Jul 17, 2025

Uh oh!

franconi commented Jul 17, 2025

Uh oh!

franconi commented Jul 17, 2025

Uh oh!

pfps commented Jul 17, 2025

Uh oh!

rat10 commented Jul 17, 2025

Uh oh!

franconi commented Jul 17, 2025

Uh oh!

pfps left a comment

Choose a reason for hiding this comment

Uh oh!

franconi left a comment

Choose a reason for hiding this comment

Uh oh!

TallTed commented Jul 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

franconi commented Jul 17, 2025

Uh oh!

lisp commented Jul 18, 2025

Uh oh!

TallTed left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

rat10 commented Jul 17, 2025 •

edited by pr-preview bot

Loading

TallTed commented Jul 17, 2025 •

edited

Loading