Negotiating Vocabularies at Link Time: A Deterministic Six-Class Compatibility Check for Federated Ontology Modules

When two research programs use the same word for different concepts, or different words for the same concept, the incompatibility is usually discovered late, by a reader reconstructing it from prose. This paper transfers the link-time compatibility check compilers and package managers apply to code to scholarly vocabularies. If each work declares the terms it owns, imports, and refines as a content-addressed module, a federated linker classifies every cross-author interaction into exactly one of six classes, proposes a typed reconciliation in a standard mapping format, and fails a continuous-integration gate on the three unresolved classes, before either author reads the other's prose. The classification is a deterministic function of two parsed module sets grounded in content-addressed definition-hash identity; it occupies the zero-uncertainty corner of the matching space and adds the reconciliation and gate layer matching omits. The author implements the method as a dependency-light tool and evaluates it on five real, independent vocabulary pairs that exercise the full class set: a clean intra-author federation, an incumbent vocabulary yielding a dangling reference, a cross-disciplinary vocabulary yielding a same-key conflict, a large biomedical ontology yielding dangling references grounded in its own term-obsoletion governance, and a cultural-heritage standard yielding cross-domain reuse at volume. Includes zharnikov-2026at-negotiating-vocabularies-link.yaml (Paper Spec v0.1.0) – a machine-readable specification of the paper's claims, assumptions, and dependencies. See https://github.com/spectralbranding/paper-spec for the standard. This PDF is generated programmatically from that machine-first source under a research-as-repository model.