Emotional Data Model (EDM) v0.8.1

The Emotional Data Model (EDM) v0.8.1 defines a governed, schema-bound representational architecture for encoding affective context as a stable, machine-interpretable data object. Unlike contemporary AI systems where emotional signals remain transient and architecture-bound, EDM externalises affective context into a structured manifold with explicit field definitions, in-band governance, and strict separation between representation, persistence, and computation layers. Artifacts are preserved within the .ddna envelope—a byte-stable container providing cryptographic integrity via W3C Data Integrity Proofs (eddsa-jcs-2022), sovereignty guarantees, and cross-vendor portability. v0.8.0 introduced Implementation Profiles (Essential, Extended, Full), Conformance Levels (Compliant, Sealed, Certified), and Profile Completeness as the canonical structural invariant. The Partner Profile system provides a registered, versioned mechanism for vertical field selection, with four reference profiles defined (journaling, therapy, companion, wiki). Two arc_type values were added — gratitude and authenticity — bringing the canonical set to fourteen. The Certification Minimum Bar requires ten or more populated affective fields from the Core, Constellation, Milky_Way, Gravity, and Impulse domains. v0.8.1 is a references-and-errata patch. Sofroniew et al. (2026), "Functional emotions in large language models" (arXiv:2604.07729), and Memory Bear A-MBER (arXiv:2604.07017) are added to §14.6 as related work. The Appendix A meta.version constraint is corrected from "0.7.x" to "0.8.x", and reference schema enum lists are corrected for nullable-enum conformance per §5.2 (24 fields). The Zenodo publication lineage is reunified under concept DOI 10.5281/zenodo.17808652. No fields are introduced, deprecated, or semantically modified. Certified conformance is available to Extended and Full Profile artifacts. Domain portability is formally specified: nine domains are portable across system boundaries; two domains are vendor-local (System, Extensions). EDM addresses the representational deficit in current intelligent systems by transforming implicit emotional states into explicit, rights-bearing data objects capable of surviving across model generations, computational paradigms, and regulatory environments.