SKOS

What it is

SKOS (Simple Knowledge Organization System) is a W3C Recommendation for representing controlled vocabularies, taxonomies, and thesauri as Linked Data. It’s the foundational vocabulary for expressing how concepts in one knowledge organization system map to concepts in another — and every subsequent mapping standard (including SSSOM) builds on it.

Published by W3C in 2009 as a Recommendation. Still the default vocabulary for ontology mapping in the Semantic Web world.

The 5 mapping relations

SKOS defines exactly five predicates for expressing correspondence between concepts in different vocabularies:

Predicate	Meaning
skos:exactMatch	The two concepts can be used interchangeably
skos:closeMatch	The concepts are similar enough to substitute in some retrieval contexts
skos:broadMatch	The target is broader than the source
skos:narrowMatch	The target is narrower than the source
skos:relatedMatch	The concepts are associated but neither broader nor narrower

That’s the whole vocabulary. It’s deliberately minimal — the W3C chose “lightweight” over “expressive.”

Why the minimalism is a problem for crosswalks

The 04-10 foundation synthesis identified five specific gaps that make SKOS insufficient for compliance crosswalking:

1. No confidence scores — every SKOS mapping is a bare triple with no way to say “we’re 60% sure of this match”
2. No provenance — no fields for author, review date, tool, or justification
3. Only 1:1 correspondences — compliance mappings are inherently many-to-many (a NIST control often maps to multiple ISO controls)
4. No negation — you can’t explicitly assert that two concepts are not equivalent
5. Deliberately minimal formal semantics — Cohen (2013) analyzed this limitation in detail; SKOS trades precision for simplicity

None of these gaps matter for lightweight thesaurus alignment (the SKOS design target). All of them matter for audit-grade compliance crosswalks.

How later standards fix the gaps

• SSSOM keeps SKOS predicates as the base vocabulary but wraps them in mandatory mapping_justification + optional confidence, author_id, mapping_date, mapping_tool, and a predicate_modifier that supports negation
• STRM / NIST IR 8477 (coming) replaces the SKOS predicate set entirely with 5 set-theory relationships (equivalent-to, subset-of, superset-of, intersects-with, no-relationship) — keeping SKOS-style simplicity but with precise semantics
• OSCAL Control Mapping Model (coming) adopts the STRM predicates and adds machine-readable fields for method, confidence, and coverage

In Crosswalker

SKOS is the historical baseline Crosswalker’s edge semantics have to improve on, not a target to adopt directly. The synthesis identifies the right stack as: STRM predicates for the relationship type, SSSOM metadata fields for the provenance envelope. SKOS predicates remain valid as an export format — if a vault ever needs to emit RDF for interop with Semantic Web tooling, SKOS is the lowest-common-denominator predicate vocabulary, reachable via the YAML-LD / LinkML bridge path discussed in the 04-10 synthesis.

Links

• W3C SKOS Reference — the 2009 Recommendation
• SKOS Primer — human-readable introduction
• Cohen 2013 — “The Unrealized Promise of SKOS” — the gap analysis
• SSSOM — the modern extension