Ontology
The ontology layer is the type system of the Concept Kernel Protocol. It provides semantic grounding for every concept through formal schema definitions and constraint validation.
LinkML Schemas
Every concept kernel defines its structure using LinkML — a modeling language for linked data:
classes:
Cat:
description: A domesticated feline
is_a: Mammal
slots:
- name
- breed
- coat_color
slot_usage:
name:
required: true
slots:
name:
range: string
breed:
range: CatBreed
coat_color:
range: string
enums:
CatBreed:
permissible_values:
Persian: {}
Siamese: {}
MaineCoon: {}LinkML provides inheritance (is_a), typed slots, enumerations, and cardinality constraints out of the box.
SHACL Validation
Beyond schema structure, each kernel enforces SHACL shapes — constraint rules that validate data at runtime:
ex:CatShape a sh:NodeShape ;
sh:targetClass ex:Cat ;
sh:property [
sh:path ex:name ;
sh:minCount 1 ;
sh:datatype xsd:string ;
] ;
sh:property [
sh:path ex:breed ;
sh:in ( "Persian" "Siamese" "MaineCoon" ) ;
] .SHACL constraints are:
- Aggregated across the concept graph during burst propagation
- Enforced before any mutation is committed
- Stored per-concept in the kernel's
shacl/directory
Type Alignment with BFO
CKP aligns its upper ontology with the Basic Formal Ontology (BFO), providing a principled foundation for categorizing concepts into:
- Continuants — Entities that persist through time (objects, qualities)
- Occurrents — Entities that unfold in time (processes, events)
This alignment ensures interoperability with biomedical ontologies, industrial standards, and other BFO-aligned systems.
Ontology Registration
When a new concept is admitted, its ontology is registered through the CK_Ontology kernel:
- LinkML schema is parsed and validated
- SHACL shapes are extracted or generated
- RDF triples are stored in the triplestore (Oxigraph)
- Relationships to existing concepts are established
- The concept becomes queryable via SPARQL