Documentation

Docs

Information layers

The strict separation between preserved material, extracted facts, calculations, interpretation, hypotheses, and human judgment.

AnankeNet keeps six information layers separate. Each layer has a different source, meaning, and level of certainty.

LayerMeaningMust not be presented as
EvidencePreserved source material and its recorded originAn analytical conclusion
ObservationA fact extracted directly from evidenceA probabilistic inference
DerivedA value calculated from observationsA source fact
AnalysisAn explainable signal or relationshipProof of intent or conduct
HypothesisA possible explanation requiring reviewAn established conclusion
Human reviewAn accountable assessment of the recordAutomatic ground truth

Why the distinction matters

A numerical score can appear precise while remaining uncertain. A relationship can be real while its meaning is unknown. A hypothesis can be useful while still being wrong.

The system therefore retains the layer of every important record. Interfaces and exports should make those labels visible rather than blending everything into a single narrative.

Unknown is a valid result

Missing information remains missing. Insufficient information is represented as insufficient. The system should never invent a value simply to complete a record or force a decision.

A model prediction is not an observation. A correlation is not proof. A risk score is not evidence of criminal conduct.