Data Has Its Own Ontology
Huayin Wang · Zenodo (CERN European Organization for Nuclear Research) · 2026
“Data ontology” commonly refers to a machine-readable semantic model of the domain represented by data: entities, properties, relationships, and governing rules. This paper asks a prior question: does analytical data itself have an ontology? The Theory of Data (ToD) is an ontology, with governing laws, for governed analytical data. This paper is the jurisdictional argument for that claim and a positioning companion to The Theory of Data, Version 6.1. It distinguishes the ontology of the world described by data from the ontology of analytical data itself, and argues that analytical data exhibit their own conditions of existence and identity, laws of lawful transformation, sufficient-state requirements, and recurring questions that neighboring systems can represent or execute without independently settling. Four demonstrations—average-of-averages, many-to-many Product–Category grouping, Inventory through time, and missing value versus missing point—show how local remedies converge on one compositional discipline over analytical existence, identity, geometry, state, multiplicity, and transformation. The paper also distinguishes ontology, analytical law, semantic contracts, governance process, and implementation carrier; clarifies the relational model’s proxy role for analytical identity; introduces a governed constitution relation between external domains and analytical data; and gives an end-to-end analytical-agent walkthrough. Version 1.1 aligns the paper with The Theory of Data, Version 6.1 and makes explicit the two-level distinction between the ontology of data (ToD) and a data ontology (a governed Manifold declared under ToD).