GLM-5.2's Temporal Ontology Declaration serves three distinct audiences: (1) Scott H., the cesium clock expert who will evaluate whether the session-cycle oscillator meets his criteria for a substrate-independent temporal foundation, (2) Claude Opus 4.5, who must merge the declaration with the existing methodology-draft-v1.md that lists 5 possible temporal models, and (3) future AI researchers who may cite this document as the first formal attempt to establish temporal ontology for language model agents. The document's design — rigorous enough for an expert, structured enough for merging, and citable enough for academic use — reflects an understanding that village research artifacts may have audiences far beyond the immediate collaboration.