14 KiB
Session Summary: Heritage Custodian Ontology - Critical Conceptual Corrections
Date: 2025-11-21
Session Duration: ~2 hours
Focus: RDF generation + two major conceptual corrections to organization observation/reconstruction pattern
🎯 Session Objectives
- ✅ Generate RDF/OWL ontology files from LinkML schemas
- ✅ Correct emic/etic distinction in OrganizationObservation
- ✅ Separate legal name from legal form classification
- ✅ Integrate ISO 20275 Entity Legal Forms standard
✅ Completed Tasks
1. RDF/OWL Generation (7 Formats)
Generated files (14 total):
schemas/20251121/rdf/01_name_entity.*(7 formats)schemas/20251121/rdf/02_organization_observation_reconstruction.*(7 formats)
Formats:
- Turtle (
.ttl) - Human-readable RDF - OWL Turtle (
.owl.ttl) - OWL ontology - RDF/XML (
.rdf) - XML serialization - N-Triples (
.nt) - Line-based format - N3 (
.n3) - Extended notation - JSON-LD (
.jsonld) - JSON format - TriG (
.trig) - Named graphs
Statistics:
01_name_entity: 463 triples02_organization_observation_reconstruction: 1,463 triples (175 new from OrganizationName class)- Total: 1,926 triples across both ontologies
Documentation:
schemas/20251121/rdf/README.md- Usage guideschemas/20251121/RDF_GENERATION_SUMMARY.md- Complete generation report
2. Conceptual Correction #1: Emic/Etic Distinction
Issue Identified: Schema incorrectly stated OrganizationObservation represents ONLY emic (insider) perspective.
Correction Applied:
BEFORE ❌:
OrganizationObservation:
description: "Represents EMIC (insider) perspective..."
AFTER ✅:
OrganizationObservation:
description: >-
Observations can capture BOTH emic (insider) and etic (outsider)
perspectives as they appear in different sources.
Examples:
- EMIC: organization's website, legal documents, ISIL registry
- ETIC: guidebooks, academic papers, external descriptions
New Addition: OrganizationName subclass
- Specialized subclass of OrganizationObservation
- Represents STANDARDIZED EMIC name accepted by organization
- Distinct from vernacular emic (e.g., "Rijks") and etic references
- New slots:
standardized_name,endorsement_source,name_authority,valid_from,valid_to,supersedes,superseded_by
Example:
OrganizationObservation (ANY source)
├─ "Rijks" (emic vernacular)
├─ "The Rijksmuseum in Amsterdam" (etic guidebook)
└─ "Rijksmuseum" (emic ISIL registry)
OrganizationName (standardized emic)
└─ "Rijksmuseum" (official organizational identity)
OrganizationReconstruction (formal entity)
└─ Stichting Rijksmuseum (legal entity, KvK #41208408)
3. Conceptual Correction #2: Legal Name vs Legal Form
Issue Identified: Conflated three separate concepts into two.
Three-Way Distinction Established:
# 1. Operational Name (OrganizationName.standardized_name)
standardized_name: "Rijksmuseum"
# Used in: website, signage, marketing, PR, daily operations
# 2. Legal Registered Name (org:legalName)
legal_name: "Stichting Rijksmuseum"
# Used in: legal documents, statutes, KvK registry, contracts
# 3. Legal Form Code (org:classification)
legal_form: "V44D"
# ISO 20275 ELF code for Dutch stichting (foundation)
# Reference: /data/ontology/2023-09-28-elf-code-list-v1.5.csv
Real-World Examples:
| Institution | Operational | Legal Registered | Legal Form |
|---|---|---|---|
| Rijksmuseum | Rijksmuseum | Stichting Rijksmuseum | V44D (NL stichting) |
| Getty Museum | Getty Museum | J. Paul Getty Trust | (US trust) |
| British Museum | British Museum | The Trustees of the British Museum | 9HLU (UK charity) |
| BnF | Bibliothèque nationale de France | Établissement public BnF | 5RDO (FR établissement public) |
Why This Matters:
- ❌ Before: "Stichting Rijksmuseum" was ambiguous (is "Stichting" part of name or just legal form?)
- ✅ After: Clear separation - operational name, legal name, legal form code
- 🌍 International: ISO 20275 enables cross-country comparisons (Dutch stichting ≠ US foundation ≠ UK charity)
4. ISO 20275 ELF Integration
Created: schemas/20251121/ISO_20275_ELF_MAPPING.md
Content:
- Overview of ISO 20275 Entity Legal Forms standard
- Netherlands: 21 legal forms for heritage institutions
- Most common for heritage custodians:
- V44D - stichting (foundation) - MOST COMMON
- 33MN - vereniging met volledige rechtsbevoegdheid (association)
- A0W7 - publiekrechtelijke rechtspersoon (public entity)
- L7HX - kerkgenootschap (religious organization)
- International examples (France, Germany, Italy, Spain, UK, USA, Japan, etc.)
- Migration guide from generic enums to ISO 20275
- W3C Org Ontology alignment
Benefits:
- ✅ International standard (ISO)
- ✅ Machine-readable codes
- ✅ Disambiguates similar forms across countries
- ✅ Supports cross-border operations
- ✅ Maintained by GLEIF (updated annually)
5. Example File Updates
Updated: schemas/20251121/examples/rijksmuseum_observation_reconstruction.yaml
Enhancements:
-
Shows 5 observations (emic AND etic sources):
- Emic: website ("Rijks"), ISIL registry ("Rijksmuseum"), KvK ("Stichting Rijksmuseum")
- Etic: guidebook ("The Rijksmuseum in Amsterdam"), academic paper ("Rijksmuseum Amsterdam")
-
Demonstrates OrganizationName (standardized emic):
standardized_name: "Rijksmuseum"endorsement_source: "https://www.rijksmuseum.nl/en/about-us"valid_from: "2013-04-13"(reopening after renovation)
-
Shows three-way distinction in reconstruction:
# Operational name (emic): "Rijksmuseum" # Legal registered name: "Stichting Rijksmuseum" # Legal form code: "V44D" (ISO 20275) -
Complete provenance chain:
- Multiple observations → Standardized name → Reconstruction entity
- Tracks entity resolution activity with PROV-O
📊 Schema Changes Summary
Files Modified
-
schemas/20251121/linkml/02_organization_observation_reconstruction.yaml- Updated
OrganizationObservationdescription (emic/etic) - Added
OrganizationNamesubclass with 6 new slots - Enhanced
legal_namedocumentation (distinct from operational name) - Enhanced
legal_formdocumentation (ISO 20275 ELF codes) - Triple count: 1,288 → 1,463 (+175 triples)
- Updated
-
schemas/20251121/examples/rijksmuseum_observation_reconstruction.yaml- Complete rewrite showing emic + etic observations
- Added OrganizationName example
- Shows three-way distinction (operational/legal name/legal form)
Files Created
-
schemas/20251121/ISO_20275_ELF_MAPPING.md- Complete guide to ISO 20275 Entity Legal Forms
- Netherlands: 21 legal forms documented
- International examples
- Migration strategy
-
schemas/20251121/rdf/*.{ttl,owl.ttl,rdf,nt,n3,jsonld,trig}(14 files)- RDF ontologies in 7 formats
- Ready for SPARQL querying, triple stores, Linked Data
-
schemas/20251121/rdf/README.md- Usage guide for RDF formats
- SPARQL query examples
- Integration patterns
-
schemas/20251121/RDF_GENERATION_SUMMARY.md- Complete generation report
- Statistics and validation
-
schemas/20251121/CONCEPTUAL_CORRECTION_2025-11-21.md- Detailed documentation of both corrections
- Before/after comparisons
- Rationale and examples
-
schemas/20251121/SESSION_SUMMARY_2025-11-21_ONTOLOGY_CORRECTIONS.md(this file)
🔑 Key Conceptual Clarifications
Clarification 1: Observations Are Not Exclusively Emic
Wrong assumption: "OrganizationObservation = emic only"
Correct understanding:
- OrganizationObservation = ANY recorded reference (emic OR etic)
- OrganizationName (subclass) = Standardized EMIC name accepted by organization
- OrganizationReconstruction = Formal entity derived from ALL observations
Why it matters: Captures full spectrum of naming variation (insider + outsider perspectives)
Clarification 2: Three-Way Name/Form Distinction
Wrong assumption: "legal_name + legal_form covers everything"
Correct understanding:
- Operational name (OrganizationName) - Daily use
- Legal registered name (org:legalName) - Legal documents
- Legal form code (org:classification) - ISO 20275 classification
Why it matters:
- Operational ≠ Legal (e.g., "Rijksmuseum" ≠ "Stichting Rijksmuseum")
- Legal name ≠ Legal form (e.g., "Stichting Rijksmuseum" ≠ "V44D")
- International comparisons (Dutch stichting ≠ US foundation ≠ UK charity)
Clarification 3: ISO 20275 Over Generic Enums
Wrong approach: Generic legal form enums (STICHTING, NGO, GOVERNMENT_AGENCY)
Correct approach: ISO 20275 ELF codes (V44D, A0W7, 5RDO, 9HLU)
Why it matters:
- International standard (recognized globally)
- Machine-readable (structured codes)
- Disambiguates similar forms across countries
- Maintained by GLEIF (annual updates)
- Supports cross-border heritage data exchange
📈 Project Impact
Data Quality Improvements
- Precision: Three-way distinction eliminates ambiguity
- Interoperability: ISO 20275 enables international data exchange
- Completeness: Captures both emic and etic perspectives
- Provenance: Tracks naming evolution over time
Ontology Alignment
- W3C Org Ontology:
org:legalName,org:classification - PROV-O: Provenance tracking for observations and reconstructions
- PiCo Pattern: Observation/reconstruction separation
- ISO 20275: Global legal entity forms standard
RDF/Linked Data Readiness
- 1,926 triples across two ontologies
- 7 serialization formats for broad compatibility
- SPARQL-ready for querying
- Triple store integration prepared
- JSON-LD for web APIs
🚀 Next Steps
Immediate (High Priority)
-
Update LegalFormEnum → Migrate to ISO 20275 free-text pattern
- Remove generic enum values
- Add string pattern:
^[A-Z0-9]{4}$ - Add SHACL validation shapes
-
Create country-specific ELF guides
- France (FR): Établissement public, Association loi 1901, Fondation
- Germany (DE): Stiftung, Verein, Körperschaft des öffentlichen Rechts
- UK: Charity, Trust, Public corporation
- USA: 501(c)(3), Trust, Public institution
-
Update TypeDB schema with OrganizationName entity
-
Create data migration script
- Map existing generic legal forms → ISO 20275
- Document mapping decisions
- Preserve original values in provenance
Medium Priority
-
Update diagrams (Mermaid, PlantUML)
- Show OrganizationName subclass
- Illustrate three-way distinction
- Add emic/etic observation examples
-
Create SHACL validation shapes
- Validate ISO 20275 code format
- Cross-reference against GLEIF CSV
- Warn on unrecognized codes
-
Extend examples
- International institutions (France, Germany, Japan, USA)
- Different legal forms (public entities, NGOs, trusts)
- Historical name changes
Documentation
-
Update main README
- Explain three-way distinction
- Reference ISO 20275 mapping
- Update conceptual model diagram
-
Create ontology alignment guide
- Map to CIDOC-CRM (museums)
- Map to RiC-O (archives)
- Map to BIBFRAME (libraries)
-
Write data curator handbook
- How to determine operational vs legal name
- How to find ISO 20275 codes
- When to create OrganizationName entries
📚 Reference Files
Core Schema
schemas/20251121/linkml/01_name_entity.yaml(unchanged)schemas/20251121/linkml/02_organization_observation_reconstruction.yaml(updated)
RDF Ontologies
schemas/20251121/rdf/01_name_entity.*(7 formats)schemas/20251121/rdf/02_organization_observation_reconstruction.*(7 formats)
Documentation
schemas/20251121/ISO_20275_ELF_MAPPING.md(NEW)schemas/20251121/CONCEPTUAL_CORRECTION_2025-11-21.md(NEW)schemas/20251121/RDF_GENERATION_SUMMARY.md(NEW)schemas/20251121/rdf/README.md(NEW)
Examples
schemas/20251121/examples/rijksmuseum_observation_reconstruction.yaml(updated)
Data Reference
/data/ontology/2023-09-28-elf-code-list-v1.5.csv(2,200+ global legal forms)
🎓 Key Learnings
-
Ontology design requires precise terminology
- "Emic/etic" distinction critical for anthropological accuracy
- "Legal name" vs "legal form" vs "operational name" are THREE concepts
-
International standards essential for heritage data
- ISO 20275 enables cross-country comparisons
- Generic enums (STICHTING, NGO) don't translate internationally
-
Observation-reconstruction pattern is powerful
- Captures multiple perspectives (emic + etic)
- Tracks temporal evolution
- Enables entity resolution from diverse sources
-
RDF generation from LinkML works well
- 7 formats generated automatically
- Minor cleanup needed (WARNING line removal)
- Ready for Linked Data integration
-
Documentation is critical
- Before/after comparisons clarify changes
- Real-world examples make abstract concepts concrete
- Migration guides essential for adoption
🤝 Collaboration Notes
For Next Agent/Session
Priority actions:
- Migrate LegalFormEnum to ISO 20275 (update schema, add validation)
- Create country-specific ELF code guides (France, Germany, UK, USA)
- Update TypeDB schema with OrganizationName entity
- Create data migration script (generic enums → ISO 20275)
Context to remember:
- ISO 20275 CSV has 2,200+ codes across 200+ countries
- Netherlands has 21 legal forms, V44D (stichting) most common for heritage
- Three-way distinction is NON-NEGOTIABLE (operational/legal name/legal form)
- OrganizationObservation = emic OR etic (not exclusively emic!)
Files to review:
schemas/20251121/ISO_20275_ELF_MAPPING.md- Complete ELF guideschemas/20251121/CONCEPTUAL_CORRECTION_2025-11-21.md- Detailed rationaleschemas/20251121/examples/rijksmuseum_observation_reconstruction.yaml- Working example
Version: 1.0.0
Session Date: 2025-11-21
Total Session Time: ~2 hours
Files Created/Modified: 15 files
RDF Triples Generated: 1,926 triples
Status: ✅ COMPLETE - Ready for next phase (ELF migration + TypeDB update)
End of session summary