Methodology

An account of how this ontology was extracted and validated, the choices made in mapping legal categories to the Basic Formal Ontology, and why the source text appears only as fragments. Read this before citing or critiquing the artifact.

Source text and rights

This ontology was extracted from A Structural Ontology of the Law by David R. Koepsell, forthcoming from Palgrave. The book is not yet published. The ontology — its classes, relations, and formal structure — is released openly under CC-BY 4.0. The book's text remains under copyright and under contract.

For that reason, this site differs from its companions in one respect. Where the Spinoza and Leibniz sites display full public-domain source passages, this site shows only short fragments of the source text: a keyword plus minimal surrounding context, capped at a small number of words, as scholarly provenance markers. The fragments indicate where in the work each formal commitment arises. They are not a substitute for reading the book, and they are too short to reconstruct the book's prose. When the book is published, the full argument will be available there.

This is a deliberate decision by the author, who holds the underlying rights and judges these short provenance fragments to be fair scholarly use that promotes rather than substitutes for the forthcoming work.

Pipeline

The ontology was constructed using BFO-Agent, the same dialogue architecture used to produce the companion ontologies. The architecture combines large language model extraction with reasoner-based validation:

Extraction

The book was chunked by chapter and section and fed sequentially to the proposer (Anthropic Claude). For each chunk, the model identified candidate structural claims about legal entities, proposed BFO-typed class structures to formalize them, and recorded the source span establishing provenance. The extraction covered twenty chapters.

Validation

Each proposal was tested against the existing ontology by the HermiT description-logic reasoner. Proposals producing inconsistency, or violating BFO's standard disjointness axioms, were rejected automatically. The artifact was further subjected to convergence and grounding evaluations, including tests of whether independently-extracted adjacent domains (contract theory, legal personhood, digital personhood) converged on compatible structures.

Commit and finalization

Accepted proposals were merged into the working ontology and committed. The ontology was then canonicalized — class names and relations normalized to a consistent scheme — and finalized as a frozen artifact, archived at Zenodo with DOI 10.5281/zenodo.19713357.

BFO mapping decisions worth flagging

Legal categories pose distinctive challenges for an upper ontology built primarily for the natural sciences. The decisions below are tentative and would benefit from scholarly review.

Norms, rules, and laws

Legal norms were typed predominantly as generically dependent continuants — abstract structures that can be inscribed in many physical carriers (statute books, databases, memories) without being identical to any of them. This captures the sense in which a law is the same law whether printed or recited, and persists across re-codification.

Obligations, rights, powers, duties

The Hohfeldian incidents — claim, duty, privilege, power, liability, immunity, and their correlatives — were rendered as specifically dependent continuants inhering in the relations between legal persons. A duty is not a quality of a person in isolation; it is a relational structure linking obligor and obligee. The correlativity of rights and duties is preserved structurally.

Legal persons

Legal personhood was treated as a role (BFO realizable entity) borne by entities that may or may not be natural persons. This allows the ontology to represent corporations, states, and other artificial persons as bearers of the legal-person role without committing to their being persons in any thicker metaphysical sense. The role is realized in legally significant acts.

Institutions and offices

Institutions were typed as a kind of object aggregate or generically dependent continuant depending on context, with offices treated as roles that persist across the particular individuals who occupy them. The distinction between an office and its occupant — central to legal and political theory — is preserved.

Acts and events

Legally significant acts (contracting, legislating, adjudicating, conveying) were typed as processes, with their participants drawn from the bearers of the relevant roles. The ontology distinguishes the act from its legal effect: the effect is a change in the relational structure of obligations and powers that the act brings about.

Jurisdiction and validity

Jurisdiction was rendered as a relational structure delimiting where and over whom a body of norms has force. Validity — a norm's membership in a legal system — was treated structurally rather than modally, since OWL 2 DL cannot express the modal and hierarchical relations (a norm is valid because enacted under a higher norm) that a full theory of legal validity requires. Readers interested in the Kelsenian hierarchy of norms should treat this ontology as a catalogue of the relevant categories, not a structural rendering of validity-conferral.

Limitations

Single rendering. The ontology represents one structural reading of law, by one author. Different theorists would produce different ontologies. Treat this as a starting point for disagreement, not a settled result.

Reasoner caveats. HermiT validates consistency under standard OWL 2 DL semantics. Much of legal theory involves modal notions (validity, obligation, permission), defeasibility (rules with exceptions), and hierarchy (norms authorizing norms) that OWL 2 DL renders only approximately or not at all.

Source fragments only. As described above, the source-text view shows only short provenance fragments, not full passages. This is a deliberate limitation pending publication of the book.

Jurisdiction-neutral. The ontology describes the categories of legal structure in general, not the content of any particular legal system. It is not a model of U.S. law, or any other body of law.

Proposer biases. The language model used for extraction has prior associations with legal concepts from training data. Where the model misreads a passage, the misreading propagates to a proposal. Author review and reasoner validation caught most such errors but not all.

References

Arp, R., Smith, B., and Spear, A. D. (2015). Building Ontologies with Basic Formal Ontology. MIT Press.

Glimm, B., Horrocks, I., Motik, B., Stoilos, G., and Wang, Z. (2014). HermiT: An OWL 2 reasoner. Journal of Automated Reasoning 53(3), 245–269.

Hohfeld, W. N. (1913). Some fundamental legal conceptions as applied in judicial reasoning. Yale Law Journal 23(1), 16–59.

Koepsell, D. R. (forthcoming). A Structural Ontology of the Law. Palgrave.

Reinach, A. (1913). Die apriorischen Grundlagen des bürgerlichen Rechtes. Jahrbuch für Philosophie und phänomenologische Forschung 1.