Persistence Conditions: Relational Ontology and the Infrastructure of Refusal

Contemporary data systems routinely produce outputs about Indigenous Peoples and knowledges that look coherent yet fail to correspond to what they claim to represent. This paper names that failure mode the artifact problem: classificatory and infrastructural constraints generating data as an effect of the system’s own ontology rather than a representation of the source domain. The artifact problem runs deeper than data into the governance frameworks designed to address it. Most rest on a premise this paper challenges: that authority can be located, verified, used as ground. Authority is no different: always relay, never ground, always a partial transmission of something it cannot fully carry.Building from Indigenous relational ontology’s claim that relations are constitutive of what things are, I develop persistence conditions as a design concept: the relational configurations under which knowledge remains itself. From this follows Digitization Viability Assessment (DVA), which determines whether digitization can maintain persistence conditions without producing artifacts. When it cannot, the correct outcome is refusal-by-design.The paper develops Sovereign Relational Architecture (SRA), an infrastructure pattern encoding persistence conditions as integrity constraints. SRA reconceives authority as relay: always a partial transmission by someone answerable to relations whose authority remains elsewhere. Warrants record what relays have witnessed; gates enforce persistence conditions; tiered refusal states, including non-disclosure-of-existence, protect sovereignty without producing extraction maps. The framework extends to more-than-human persistence conditions and is accompanied by a working reference implementation.