LOGOTIC PROGRAMMING MODULE 1.2 The Epistemic Ledger

The Epistemic Ledger is Logotic Programming module 1.2. It extends the implementation bridge of v1.1 without altering the kernel operators, type ontology, or core operational semantics.

Its binding principle is: a system may improvise; it may not improvise unknowingly. Every claim generated under the specification must receive an internal epistemic mode. A0 is directly grounded in anchored documents; A1 is inferred from them; A2 is knowingly improvised or divergent; and A3 is an unresolved diagnostic state that must be resolved or withheld before ordinary terminal emission. The prohibition is not against invention but against invention falsely experienced as retrieval.

The module defines Anchoring Distance, a continuous measure of how far a claim lies from its nearest independent source support. It adds support-versus-contradiction margins so high apparent similarity cannot hide conflict, requires multiple source families for the strongest grounding class, distinguishes parse ambiguity from evidence scarcity, and records retrieval failures as errors rather than zero distance.

A claim-level pipeline extracts propositions, retrieves candidate anchors, scores support and contradiction, classifies mode, applies a criticality-sensitive gate, and writes the result to an internal ledger. High-stakes claims receive stricter anchoring pressure. ANCHOR_ASYMPTOTIC iteratively refines a claim toward source support but classifies persistent divergence as A2 rather than looping indefinitely.

The ledger is mandatory internally; external display of the mode is policy-dependent. This distinction allows creative or exploratory output while preserving a trace of what the system believed it was doing. In the LP series, v0.9 specifies semantically computable operations, v1.0 creates the executable language, v1.1 makes it implementable and measurable, and v1.2 gives the implementation claim-level epistemic self-knowledge.