Abstract:
This paper presents the Hebrew Manuscripts Ontology (HMO), developed within the Mapping Hebrew Manuscripts
(MHM) project, as an evaluated domain ontology and transformation workflow for representing Hebrew
manuscripts as Linked Open Data. The problem addressed is that rich Hebrew-manuscript metadata in the
National Library of Israel remain largely locked in MARC records, limiting entity-level querying, cross-record
reasoning, and future interoperability with external knowledge graphs. HMO responds with a domain-specific
model aligned with CIDOC CRM and LRMoo that combines three elements: structural granularity through
Bibliographic Unit, Codicological Unit, and Paleographical Unit (BU-CU-PU); an event-centric representation
of production, transfer, and ownership; and an explicit epistemological layer for source attribution and status,
with certainty support defined at schema level. The paper evaluates this contribution at three levels: six fully
instantiated pilot manuscripts chosen to cover key structural edge cases; schema-level validation over 37 SPARQL
checks executed on the released ontology files and controlled vocabularies; and a larger-scale feasibility run of
the conversion workflow over 10,000 catalog records. Within the pilot, HMO supports research questions that
are difficult to ask in MARC alone, such as retrieving manuscripts with multiple hands, identifying codices with
more than ten codicological units, tracing transfer-of-custody chains, and distinguishing textual witnesses from
bibliographic works. The current release follows a MARC-only population policy for external identifiers: Wikidata,
VIAF, GeoNames, and owl:sameAs slots are defined at schema level but are not yet populated in the pilot RDF,
so interoperability is demonstrated here as alignment readiness rather than completed entity linking. The ontology
(OWL/TTL), SHACL shapes, pilot RDF, crosswalk, validation materials, and conversion pipeline are included in
the repository materials accompanying this paper for direct inspection. The contribution claimed here is therefore
an evaluated and inspectable semantic framework, not a claim that full corpus-scale reconciliation or community-
wide uptake has already been achieved.
