OWL 2
OWL 2 is the ontology language for the semantic web: a vocabulary for class expressions, restrictions and formal semantics, with profiles that trade expressivity for reasoning that stays computable.
What it is
OWL 2 is an ontology language built on RDF, with a formal semantics rather than only a naming convention. Where RDFS lets a graph name classes and properties, OWL lets it state how classes are defined — as unions, complements, or restrictions on their members. Three profiles restrict the language: EL for large ontologies where classification must stay fast, QL for query answering over large data, and RL for rule-based reasoning.
What it gives the knowledge layer
It gives the knowledge layer a schema that can be checked, not merely published: a reasoner can classify a graph, derive memberships that were never asserted, and detect inconsistencies between an ontology and its data. The profiles make that decidability trade-off explicit and selectable.
Where the spec lives
The language is introduced in the OWL 2 Document Overview, which maps the specification set — structural specification, profiles, mapping to RDF graphs, conformance and test cases. The OWL 2 Primer is the working introduction.
What it does not solve
Formal semantics do not make content current or correct: an ontology can be internally consistent and still describe a domain that changed last quarter. Reasoning cost also rises with expressivity, which is why the profiles exist. And a schema says what terms mean, not where a claim came from — that is PROV-O.
FAQ
No. RDF needs no schema, and RDFS covers light typing and hierarchies. OWL is for when the requirement is more than description: restrictions, cardinality, disjointness.
They trade expressivity for predictable computation: EL for large ontologies where classification must stay fast, QL for query answering over large data, RL for rule-based reasoning.
It can: an ontology plus individuals lets a reasoner derive new memberships and detect inconsistencies, depending on the profile and the engine.
The ontology is the schema half of it. A knowledge graph is the ontology together with the instance data that uses its terms, which is why the ontology limits what can be asked.
Sources
- OWL 2 Web Ontology Language Document Overview — W3C Recommendation; the map of the OWL 2 specification set.
- OWL 2 Web Ontology Language Primer — W3C Recommendation; the working introduction to OWL 2 constructs.