The Epistemic Lineage Engine

Knowledge systems usually preserve selected artifacts and outcomes more reliably than the transformations, alternatives, environments, observations, and authority conditions that produced them. The Epistemic Lineage Engine (ELE) is proposed as a substrate for making those changes reconstructable. Its central unit is a lineage episode binding predecessor state, transformation, successor state, observation, result, branch relations, and evidence. The source correctly distinguishes lineage from judgment, but several claims require qualification. A snapshot graph can be part of lineage; replay is not always deterministic; observations can have effects; undo does not erase external consequences; alternatives cannot always be retained indefinitely; and no engine can prove that every causally relevant influence was captured. ELE should therefore govern declared lineage closure and its limits rather than claim complete causality. This paper formalizes an episode envelope, separates exact replay, deterministic replay, equivalence-class replay, procedural reenactment, and historical reconstruction, and adds privacy, retention, redaction, authority, environment, and irreversible-effect requirements. It integrates governed semantic project lineage, Reasoning by Way of Receipts, the Three Hash Rule, Semantic Debt, Governance Coverage Gaps, GMB, and the universal GSM. ELE does not decide truth, approval, legitimacy, or authorship. It can preserve the materials from which such decisions are inspected. This paper reaches Specification and proposes configuration criteria; it does not establish a running engine, complete lineage capture, or deterministic replay.