◆ Semantic Web Catalogue

Problem thread

Statements about statements

How do you say who made a statement, when, and how sure they are, in a data model built from single statements?

issue
data model
also trust
status
largely settled
steps
7, 1999–2026
threads
all problem threads

RDF is made of single statements, but trusting data means knowing where each statement came from. RDF's original answer, reification, was verbose and little used. Named graphs let a whole set of statements be named, signed and described, which nanopublications and the PROV vocabulary built on. Nested triples offered a compact way to annotate one statement, and became triple terms in RDF 1.2.

Lineage

  1. 22 Feb 1999
    proposes standard ★ seminal
    RDF

    Reification: to make a statement about another statement, build a model of the original statement as a new resource and attach properties to it. [src]

  2. May 2005
    proposes writing ★ seminal
    Named Graphs, Provenance and Trust

    Treats the Semantic Web as a set of RDF graphs, each named by a URI, so statements can describe whole graphs. Publishers can sign their graphs, and consumers decide which graphs to trust by their content, their description and the task at hand. [src1][src2]

  3. 2010
    proposes writing ★ seminal
    The Anatomy of a Nanopublication

    The nanopublication: a single assertion published together with its provenance and publication details, each as a named graph, so it can be cited and credited on its own. [src1][src2]

  4. 30 Apr 2013
    settles standard ★ seminal
    PROV

    W3C's vocabulary for provenance: who made something, from what, and how. PROV-DM is the model and PROV-O its OWL ontology. [src1][src2]

  5. 2017
    responds writing
    Foundations of RDF* and SPARQL*
    ↳ responds to RDF

    Standard reification blows up data size and complicates queries. Nest triples instead, so a statement can carry metadata such as a certainty score, as edge properties do in graph databases. [src]

  6. 21 May 2024
    proposes standard
    RDF Dataset Canonicalization

    A single canonical form for any RDF dataset, so the same statements always hash and sign the same way and a signature can travel with the data. [src]

  7. 7 Apr 2026
    settles standard
    RDF 1.2

    A triple used as the object of another triple is a triple term; with rdf:reifies, one can make statements about statements, including statements that are not themselves asserted. [src]

Open now

Under each question, the answers people have put forward or built. The catalogue gives no odds on them.

  • Which of reification, named graphs and triple terms will publishers actually use for provenance? [src]
    RDF 1.2 adds triple terms alongside the older mechanisms; its core specification became a Candidate Recommendation in April 2026.
    • Triple terms, now in the standard. RDF 1.2
      RDF 1.2 lets a statement be the object of another statement.[src]
    • Named graphs, one graph per claim with its provenance. The Anatomy of a Nanopublication
      Nanopublications wrap each assertion, its provenance and its publication details in separate named graphs.[src]
  • How will signatures on RDF data be checked in everyday use? [src1][src2][src3]
    Canonicalization (2024) and Data Integrity proofs (2025) make signing possible; the Named Graphs paper's picture of consumers applying trust policies still depends on tools that do it.
    • Canonicalize the data, then sign it. RDF Dataset Canonicalization
      RDF Dataset Canonicalization (2024) gives each dataset one form to hash, and Data Integrity (2025) attaches proofs to it.[src1][src2]

Decision points

Dated moments when a decision or a deadline can change the outcome.

  1. 30 Apr 2027
    RDF & SPARQL Working Group charter ends charter

    The W3C group finishing RDF 1.2 is chartered until this date.[src]

How to contribute