Should "Event" Be a Subtype of "Work"? An Ontological Analysis
Recommendation: Keep Event as a Separate Primitive
Section titled “Recommendation: Keep Event as a Separate Primitive”Every major formal ontology, knowledge graph, and metadata standard maintains events and works as categorially distinct. The distinction (endurant/perdurant, continuant/occurrent) is considered foundational. No major ontology makes events a subtype of works.
Key Findings
Section titled “Key Findings”1. Major Ontologies — Unanimous Separation
Section titled “1. Major Ontologies — Unanimous Separation”| Ontology | Events | Works/Objects | Relationship |
|---|---|---|---|
| BFO | Occurrent | Continuant | Disjoint — separate mereological relations |
| DOLCE | Perdurant | Endurant | Disjoint categories |
| SUMO | Process | Object | Sibling categories under Physical |
| UFO | Perdurant (UFO-B) | Endurant (UFO-A) | Required UFO-AB to relate them |
| CIDOC-CRM | E5 Event (Temporal Entity) | E71 Human-Made Thing (Persistent Item) | Separate branches of hierarchy |
| Schema.org | Event (under Thing) | CreativeWork (under Thing) | Sibling types |
| Dublin Core | ”Non-persistent, time-based occurrence” | Persistent resources | Distinct categories |
| Wikidata | subclass of “occurrence” | subclass of “intellectual work” | Separate hierarchies |
| IFLA-LRM | Not modeled | Work → Expression → Manifestation → Item | Events outside FRBR scope |
2. CIDOC-CRM: The Key Insight
Section titled “2. CIDOC-CRM: The Key Insight”Creation (E65) is an EVENT that PRODUCES a Thing (E71). The creation event and the created thing are distinct:
- The symphony premiere (event) ≠ the symphony (work)
- The premiere happened once. The symphony persists.
- Events have causal coherence. Works have attributive coherence.
3. Greek Etymology — Ergon
Section titled “3. Greek Etymology — Ergon”ἔργον covers both deed (activity) and product (creation). Aristotle distinguished these senses. This supports having BOTH primitives, not collapsing them.
4. What Events Have That Works Don’t
Section titled “4. What Events Have That Works Don’t”- Participants (co-present during unfolding) vs. creators/audiences
- Non-persistence (happened, now over) vs. persistence
- Temporal parts (beginning, middle, end that unfold) vs. structural parts that coexist
- Causal coherence vs. attributive coherence
- Occurrence semantics (“it happened”) vs. expression semantics (“it was published”)
- FRBR chain (express_as, adapt, translate, cover) makes no sense for events
5. The Philosophical Dissolution Tradition
Section titled “5. The Philosophical Dissolution Tradition”Whitehead, Quine, Goodman, Lewis dissolve the distinction — but by making EVERYTHING events/processes, not by making events a subtype of objects. This isn’t useful for a knowledge graph.
6. Industry Perspective
Section titled “6. Industry Perspective”Event management uses “production” language, but also recognizes events are ephemeral experiences, not persistent artifacts. The Experience Economy positions events alongside goods and services — distinct categories.
Sources
Section titled “Sources”- Stanford Encyclopedia of Philosophy, “Events” (2025 revision)
- CIDOC-CRM v7.1.3 (2024)
- BFO 2.0 / ISO 21838-2
- DOLCE (Laboratory for Applied Ontology)
- SUMO (Ontology Portal)
- UFO (Guizzardi et al.)
- Schema.org type hierarchy
- Dublin Core DCMI Type Vocabulary
- IFLA-LRM (2017)
- Wikidata Q1656682, Q17537576
- Stout, “The Category of Occurrent Continuants”
- Baratella, “Are There Occurrent Continuants? A Reply”
- Guarino, Baratella & Guizzardi, “Events, their names, and their synchronic structure”