Downloads
The namespace IRI content-negotiates: Accept: text/html returns
this page, any RDF media type returns the Turtle. Files are also served
directly, with CORS enabled so browser-based tooling can fetch them.
| Resource | File | Type | Size |
|---|---|---|---|
| Companion ontology | arith.ttl | text/turtle | 8.5 kB |
curl -H "Accept: text/turtle" https://epistemic-ontology.net/arith
Namespace
@prefix arith: <https://www.epistemic-ontology.net/arith#> .
The companion owl:imports the Record Ontology, so a reasoner
fetching this document will resolve https://www.epistemic-ontology.net/record
and get the base vocabulary back.
Why a companion rather than an extension
The Record Ontology owns the operation face of the formal: rec:Inference,
premises, conclusions, force. What it never had was the object face — the
things inferences are about. This companion supplies that at representative
scope, and it is deliberately attached, never absorbed: the same pattern the
base ontology applies to every domain vocabulary. A domain plugs in through
warrant; it does not get merged into the core.
Scope is representative on purpose: integers, variables, the four operations,
equality. Not analysis, not set theory, not proofs-as-objects. Names align with
the OpenMath arith1 content dictionary by dc:source pointer — never by
import, which is the SKOS lesson applied again.
Two disciplines
The reasoner checks shape, never value. 2 + 2 = 4 is not a DL entailment;
OWL 2 DL is decidable precisely because it excludes nearly all mathematics.
Evaluation and exercise live in the engine, which compiles a described record
into a computer-algebra system and runs it.
Expressions are Records composed of Records. arith:Expression is a
subclass of rec:Record, and operands are rec:composedOf parts. So the
patchwork structure the base ontology already models carries mathematical
content without any new mereology.
The definitional hierarchy
The four operations sit in a hierarchy that bottoms out at the Peano ground — mathematics' own derivation web, formal all the way down. That makes it the limiting case of the knowledge-automaton argument: a web whose interior is entirely derivable, and therefore entirely puncturable, regenerating from the ground by deduction alone.
Exercise, compiled from the description
The point of describing a formal record in full — sides, angles, relationships, the whole thing coming together — is that the description can be run. A compiler walks the expression subgraph into sympy, so the exercise is derived from the description rather than hand-written beside it. Two consequences:
- Closed-world well-formedness is a compile error. A missing operand is caught by the compiler, because OWL cannot see absence.
- Performative provenance: every exercise act is logged at a moment, so earned warrant has a genealogy, and each record carries a lifecycle standing — unexercised → confirmed / failed → exorcised.
The worked pair lives with the Record Ontology examples: the 3-4-5 triangle, described and exercised from its description, and the 2 = 1 pseudo-proof, which fails at its hidden division by zero and is expelled.
Cite
Cite the namespace IRI, not a file path: the IRI is the permanent identifier and will keep resolving as serializations change.
@misc{arithcompanion2026,
title = {The Arithmetic Companion to the Record Ontology},
author = {Hinchley, Ron},
year = {2026},
version = {0.1.0},
url = {https://epistemic-ontology.net/arith}
}