Jnana

Jnana is a candidate governed semantic compilation architecture that separates canonical meaning, admissibility law, constitutional amendment constraints, procedural execution, external effects, structural receipts, and derived projections. Its central question is not merely whether computation can run, but whether a proposed transition may be admitted, lowered, executed, allowed to affect a domain, committed to canonical state, and retained as precedent under declared authority and evidence. The source paper presents this architecture through a biological metaphor: MCIR as semantic genome, MCIR-Law as developmental law, PEIR as regulated expression, Domain Effects as an immune membrane, ReceiptLog as structural memory, and state projections as phenotype. The metaphor is pedagogically strong, but it is not evidence of implementation or conformance. This revision makes the boundary contracts explicit, separates architecture from implementation, qualifies claims of semantic identity and negative-power guarantees, adds ambiguous-effect and recovery states, and locates Jnana within the larger Governed Semantic Kernel, Constitutional Computing, Governed Meaning Base, and semantic-language stack. Jnana's strongest contribution is layered constitutional containment. A bounded semantic root and decidable admission regime can govern potentially expressive execution only if execution cannot silently widen meaning or authority, external effects cross explicit membranes, receipts remain independently checkable, and every conformance claim is limited to the verified vertical slice. Refusal is a valid outcome. So are deferment, explicit gap, and unresolved effect status. Persistence is earned by evidence; it is never inferred from narrative continuity.