Skip to content

ADR-0015: Expand Source-Ontology Scope to Include GFO, YAMATO, TUpper, and GUM

  • Status: Accepted
  • Date: 2026-07-02

Context

Since the first concept crosswalk, XwkOnt's crosswalk practice has treated four source ontologies — BFO, DOLCE, SUMO, UFO — as the project's working scope. This was never a single explicit enumeration in README.md or docs/FOUNDING_PRINCIPLES.md (neither names specific source ontologies; both describe scope generically as "the world's major foundational ontologies"). It was, instead, a de facto scope established by repeated practice: every concept crosswalk's Source Ontology Correspondence table, docs/STANDARDS_BASELINE.md, docs/evaluations/meta-ontology-standards-evaluation.md, and multiple ontology-spec documents (docs/ontology/core-ontology.md, core-glossary.md, core-axioms.md, core-validation.md) all consistently name the same four.

xwkont:ref:borgo-galton-kutz-2022 (Borgo, Galton & Kutz, 2022, "Foundational ontologies in action," the editorial introducing an Applied Ontology special issue) covers seven foundational ontologies via common modeling cases: BFO, DOLCE, GFO, GUM, TUpper, UFO, and YAMATO. continuant-occurrent.md flagged the resulting gap as uncertainty-004 on 2026-07-01, during the primary-source verification pass, deliberately deferring the scope question rather than deciding it at the same time.

This research (2026-07-02) investigated this question. Its planned first step — directly verifying xwkont:ref:borgo-galton-kutz-2022's full text — was blocked: the paper is behind Cloudflare/paywall protection on every mirror checked (publisher host, DOI redirect, institutional repository, author self-archive pages), not merely difficult to parse. That is recorded in docs/references/ref-borgo-galton-kutz-2022.md's "Full-text access attempt" note; the reference remains candidate, not reviewed.

Lacking the editorial's own comparative framing, this research instead investigated each of the four non-included ontologies directly:

  • GFO (General Formal Ontology, Herre et al., Onto-Med research group) — actively maintained; OWL files (gfo.owl, gfo-basic.owl) in a public GitHub repository (Onto-Med/GFO); a GFO 2.0 successor has been in development since 2020 (Onto-Med/GFO-2.0). Directly fetchable in the same way BFO's and DOLCE's OWL files already were.
  • YAMATO (Yet Another More Advanced Top-level Ontology, Mizoguchi, Osaka University, since 1999) — single-maintainer, but has its own dedicated Applied Ontology 2022 paper (Mizoguchi & Borgo, 10.3233/AO-210257) separate from the editorial, and hosted OWL/Hozo files at hozo.jp.
  • TUpper (Grüninger, Ru & Thai) — now part of ISO/IEC 21838-4:2023 (Top-Level Ontologies, Part 4), the same standards family BFO occupies as ISO/IEC 21838-2. Has its own Applied Ontology 2022 paper (10.3233/AO-220263).
  • GUM (Generalized Upper Model, Bateman, University of Bremen) — a linguistically-motivated ontology mediating between natural language and domain knowledge, a different orientation from the other seven's general-purpose top-level framing. Stable OWL DL identifier at purl.org/net/gum2. Has its own Applied Ontology 2022 paper (10.3233/AO-210258).

None of the four lack a fetchable primary source; unavailability is not a reason to exclude any of them. The maintainer, presented with this research, decided to add all four rather than a subset or none.

Decision

XwkOnt's source-ontology scope is expanded from four to eight: BFO, DOLCE, SUMO, UFO, GFO, YAMATO, TUpper, GUM.

This is a scope expansion, not a redefinition — the existing four remain exactly as verified in the primary-source verification pass; nothing about their content, confidence values, or identifier conventions changes. The four new sources join the same crosswalk methodology (docs/methodology/crosswalk-methodology.md), the same per-source identifier-convention practice (docs/methodology/primary-source-verification.md §2 — each source's own native ID scheme, not a forced uniform one), and the same primary-source-verification bar that pass established (direct fetch and extraction, not secondary-source paraphrase) before any crosswalk row can cite them.

Crosswalk-content work itself (adding GFO/YAMATO/TUpper/GUM correspondences across all 8 existing concept crosswalks, and the reference records their primary sources require) is not performed by this ADR. It is tracked as future work in TODO.md. This ADR settles the scope question; it does not itself do the ~4x expansion of crosswalk content that follows from it.

Scope

This ADR amends the project's de facto source-ontology scope as reflected in practice across docs/STANDARDS_BASELINE.md, docs/evaluations/meta-ontology-standards-evaluation.md, and the ontology-spec documents' illustrative source lists. Those documents' existing four-source phrasing is historical/contextual prose (describing what was true when each was written), not a live scope declaration, and is not edited by this ADR — this ADR itself is the authoritative scope statement going forward, consistent with the ADR/README precedence order in docs/PROJECT_LIFECYCLE.md. continuant-occurrent.md's uncertainty-004 is resolved by this ADR and updated accordingly.

This ADR does not change ADR-0009's mapping-assertion schema, ADR-0013/ADR-0014's confidence vocabulary, or docs/INFORMATION_ARCHITECTURE.md's crosswalk template structure — a ninth column in the Source Ontology Correspondence table (or four new rows, depending on template layout) is a content change within the existing template, not a template change.

Rationale

Three considerations favored all four over a subset or none:

  1. Reuse Before Introduce, applied to TUpper specifically: TUpper's ISO/IEC 21838-4:2023 status makes it, if anything, a stronger case for inclusion than SUMO already is (SUMO has no ISO standardization and is the crosswalk's consistently lowest-confidence source). Excluding a ratified standard in the same family as BFO's own ISO/IEC 21838-2 would be inconsistent with why BFO was included in the first place.
  2. No fetchability blocker for any of the four: the original hesitation (the earlier deferral) was about workload, not availability. This research confirmed all four have real, direct primary sources — three with their own dedicated 2022 Applied Ontology papers beyond the blocked editorial, one (GFO) with the most straightforward GitHub-OWL fetch path of any source ontology in the project including the existing four.
  3. Coherence of the comparison set: GFO, YAMATO, TUpper, and GUM are exactly the four ontologies the Borgo/Galton/Kutz special issue treats as peers of BFO/DOLCE/UFO (and, per SUMO's low-confidence-but-included precedent, GUM's different linguistic orientation is not disqualifying — SUMO already establishes that a structurally different source ontology belongs in the crosswalk if it's independently well-evidenced). Adding a subset would require a principled reason to exclude specific ones, and no such reason survived this research — TUpper is arguably the strongest candidate, not the weakest, undercutting a TUpper-only alternative on the grounds that it would exclude equally well-evidenced peers for no clear reason.

Consequences

Positive

  • Crosswalk coverage now matches the actual comparative literature (xwkont:ref:borgo-galton-kutz-2022's own scope), closing the mismatch uncertainty-004 flagged.
  • TUpper's inclusion strengthens, not weakens, the project's "Reuse Before Introduce" posture — a ratified ISO standard is exactly the kind of source XwkOnt should prioritize crosswalking.
  • Each of the four has an independently fetchable primary source already identified, so the eventual verification work (unlike the blocked editorial itself) is not expected to hit the same Cloudflare/paywall wall.

Trade-offs

  • Crosswalk-content workload roughly doubles: 8 existing concepts × 4 new sources = up to 32 new Source Ontology Correspondence rows, each needing the same primary-source-verification rigor applied to the original four, plus new docs/references/ref-*.md records for each new source's primary artifacts (GFO's OWL files, YAMATO's paper/OWL, TUpper's ISO standard and/or paper, GUM's paper/OWL). This is substantial future work, not completed by this ADR.
  • YAMATO's single-maintainer status is a real, if modest, continuity risk relative to the institutionally-backed GFO and the ISO-standardized TUpper — worth noting in YAMATO's eventual reference record, not a reason to exclude it now.
  • xwkont:ref:borgo-galton-kutz-2022 itself remains unverified (full text still inaccessible); the BFO/DOLCE framing claim it's currently cited for in continuant-occurrent.md is unaffected by this ADR and remains an open citation gap.

References

  • docs/crosswalks/concepts/continuant-occurrent.md (uncertainty-004, now resolved)
  • docs/references/ref-borgo-galton-kutz-2022.md (full-text access attempt, 2026-07-02)
  • docs/methodology/primary-source-verification.md (per-source identifier convention practice, to be extended to the four new sources)
  • docs/STANDARDS_BASELINE.md, docs/evaluations/meta-ontology-standards-evaluation.md (historical four-source context, not edited by this ADR)
  • ISO/IEC 21838-4:2023 (TUpper)