RDIV / ANSWER E-002 · EVIDENCE
What Is an Evidence Graph?
An evidence graph connects source objects to claims, entities, events, contradictions, and findings so investigative reasoning can remain traceable.
Fix the object first.
An evidence graph is a structured reasoning map. Nodes represent objects such as documents, claims, people, organizations, events, and findings. Edges represent explicit relationships such as supports, contradicts, mentions, controls, precedes, or derives from.
The visible label is not the whole system.
Complex investigations fail when the narrative becomes detached from the record. A graph forces the analyst to expose where a conclusion came from, which sources are independent, and where contradictions or missing links remain.
Trace the burden.
- Graph structures are common tools for representing relationships; the evidentiary value depends on the discipline of the underlying source and edge records.
- Traceability is a methodological property, not proof that every edge is correct.
- Different investigations need different ontologies and edge vocabularies.
- Automated entity resolution can create false edges and therefore requires human verification for consequential relationships.
Where the work adds something.
R/IV’s graph doctrine is intentionally asymmetric: admitted source objects populate evidence; model output may assist analysis but must not become evidence merely because the system generated it.