Source-Ontology Core/Extension Module Conventions

Status: Editorial observation, direct primary-source verification of all 8 sources Date: 2026-07-05 Related specification: docs/adr/ADR-0020-define-core-as-base-module-not-domain-tier.md, docs/adr/ADR-0021-source-classified-core-placement-criterion.md

Purpose

ADR-0020 defined XwkOnt's own "core" (xwkont-core:/core.ttl) as a base module distinct from optional extension modules, grounded in how three of XwkOnt's eight source ontologies (BFO, DOLCE, GFO) independently use the word "core" this way for themselves. ADR-0021 then used that pattern as the placement rule: a concept qualifies for XwkOnt-core if any one contributing source classifies it as part of that source's own core/base module. That rule left the remaining four sources (SUMO, YAMATO, TUpper, GUM) unresolved — not yet checked for whether they modularize themselves the same way at all. This note closes that gap: all 8 sources, checked directly against their own primary artifacts, not inferred or assumed.

This is a descriptive cross-cutting finding, not a new crosswalk concept — per the discussion that produced ADR-0020, "core" here is a property of each source's own publication/module structure, not a claim about what any source's classes mean in the world. It does not go in docs/crosswalks/concepts/.

Summary Table

Source	Modularizes core/extension?	Own base/core artifact	Own extension artifact(s)	Own terminology for the base tier
BFO	Yes	bfo-core.ttl (21838-2/owl/, BFO-ontology/BFO-2020)	Separate relation-ontology / IAO extension files	"core" (literal filename)
DOLCE	Yes	DOLCE-Lite (2009-relabeled "DOLCE-CORE")	DnS (ExtendedDnS.owl)	"core" / "DOLCE-CORE" (Borgo & Masolo's own 2009 relabeling)
GFO	Yes	modules/gfo-base.owl	modules/situation/*.ttl; the parallel gfo-core: namespace found inside the situation module	"core module" (GFO's own README)
TUpper (via PSL)	Yes	psl_core (separate COLORE directory; tupper.clif imports several PSL modules that build on it)	psl_actocc, psl_disc_state, psl_atomic, psl_complex, psl_duration, psl_locations, psl_occtree, psl_occurrence, psl_participates, psl_resources, psl_soo, psl_subactivity, extended_psl, interval_psl, and others (all confirmed present as sibling COLORE directories)	"PSL-Core" — PSL's own documented terminology: "organized into PSL-Core and a partially ordered set of extensions. All extensions within PSL are consistent extensions of PSL-Core, although not all extensions within PSL need be mutually consistent."
SUMO	Yes, but not as "core"	Merge.kif	Dozens of separate domain .kif files (Anatomy.kif, Cars.kif, Economy.kif, Food.kif, etc.) plus Mid-level-ontology.kif	"upper"/"meta-level", not "core" — SUMO's own text: "SUMO originally concerned itself with meta-level concepts, and has been considerably expanded to include a mid-level ontology and dozens of domain ontologies." Merge.kif's own header additionally states: "The SUMO is a modular ontology. That is, the ontology is divided into self-contained subontologies." Directly searched Merge.kif for "core": the only hits are unrelated domain content (AtomicNucleus, Virus) and "core team" in CONTRIBUTING.md's governance section — no self-description of Merge.kif as SUMO's own "core."
UFO	Not file-modularized, but artifact-tiered	The 2021 comprehensive paper (Guizzardi et al.) — UFO's own primary theoretical exposition	The OntoUML vocabulary artifact — a derived modeling-notation artifact	No literal "core" self-description found; this repository's own established practice (role.md, universal-type.md, situation-state-of-affairs.md) already treats the 2021 paper as primary and OntoUML as secondary/fallback, reused here rather than newly derived
YAMATO	No evidence found	—	—	No internal core/extension split found in the one fetchable primary artifact (the 2010 technical report, xwkont:ref:yamato-mizoguchi-2010) or in secondary literature searched. YAMATO's own OWL/Hozo files remain behind hozo.jp's registration gate and were not fetched — this is a "not found in what's checked" absence, not a confirmed "YAMATO has no such structure."
GUM	Inconclusive	GUM-31.owl (GUM-3.owl namespace)	A separate GUM-3-space.owl module is referenced alongside GUM-31.owl at the same archived location (xwkont:ref:gum-owl's own Descriptive Metadata: "An earlier GUM-3-space.owl module... also exist[s] at the same archived location but w[as] not fetched in this pass") — suggestive of some modularization, but no primary or secondary source found using "core" for GUM's own base tier.

What This Means for ADR-0021's Placement Rule

Four of eight sources (BFO, DOLCE, GFO, TUpper) now directly confirm the base/extension pattern, all independently, with TUpper's finding the most explicit of all — psl_core is a literal directory name, and PSL's own documentation states the base/extension relationship in almost the same words GFO's README uses ("all extensions... are consistent extensions of PSL-Core... not all extensions... need be mutually consistent" vs. GFO's "parallel extensions... not all... consistent with each other"). This is now a 5-of-8 pattern once UFO's already-established artifact tiering is counted (SUMO also modularizes, just under different vocabulary — "upper," not "core" — so it doesn't add a sixth terminological hit, but does confirm the structural pattern extends further than "core" the word alone).

For ADR-0021's existential rule, this means: a concept whose only supporting source is TUpper can now be checked against psl_core vs. the PSL extension modules tupper.clif imports, the same way BFO/DOLCE/GFO already can. SUMO's Merge.kif vs. domain-.kif split can also be used, reusing SUMO's own "upper"/"meta-level" framing rather than forcing "core" vocabulary onto a source that doesn't use it — ADR-0021's rule is about a source's own core/base-vs-extension classification, not literally requiring the word "core," so SUMO's upper/meta-level-vs-domain split still counts. YAMATO and GUM remain genuinely unresolved: neither has confirmed evidence of an internal base/extension split at all in what's been fetched, so a concept whose only supporting source is YAMATO or GUM cannot yet be resolved by this rule and would need either source's OWL/Hozo files (YAMATO, gated) or a republished canonical host (GUM, currently dead) fetched before the question is even askable.

Provenance

• BFO, DOLCE, GFO: previously verified in docs/adr/ADR-0020-define-core-as-base-module-not-domain-tier.md and this session's situation-state-of-affairs.md pass; reused here, not re-derived.
• UFO: previously established across docs/crosswalks/concepts/role.md, universal-type.md, and situation-state-of-affairs.md; reused here.
• SUMO: Merge.kif (ontologyportal/sumo, master branch) fetched and searched directly 2026-07-05, per xwkont:ref:sumo-niles-pease-2001. CONTRIBUTING.md also fetched and checked.
• TUpper: COLORE repository directory listing (gruninger/colore/ontologies) fetched directly 2026-07-05, confirming psl_core as a sibling directory to the PSL extension modules tupper.clif imports, per xwkont:ref:tupper-colore. PSL-Core/extension terminology confirmed via PSL's own published documentation (Grüninger et al., "The Process Specification Language (PSL) Overview and Version 1.0 Specification," NIST).
• YAMATO: checked against the existing fetched technical report (xwkont:ref:yamato-mizoguchi-2010) and a secondary-literature search; no internal modularization found in either.
• GUM: checked against xwkont:ref:gum-owl's existing Descriptive Metadata (which already notes the unfetched GUM-3-space.owl module) and a secondary-literature search of Bateman's own GUM papers; no "core" terminology found.