Logotic Programming proposes a third programming regime concerned with conditions of intelligibility.
Symbolic programming encodes instructions. Statistical programming encodes objectives, priors, and optimization conditions. Logotic programming arranges entities, relations, anchors, invariants, boundaries, and navigation so that a bounded semantic domain remains recognizable when traversed by human or machine interpreters.
Its execution is retrieval and interpretation rather than CPU instruction or parameter training. A logotic program does not directly control an answer. It changes the structured environment from which recognition occurs.
The method emphasizes bounded semantic spaces rather than universal ontology. Persistent identifiers, crosslinks, consistent signatures, typed relations, canonical anchors, and substitution rules make some interpretations more stable than alternatives without claiming universal truth.
The paper relates the method to authority control, FAIR data, linked data, FRBR, RAG, prompt engineering, constitutional AI, social ontology, actor-network theory, speech acts, and hermeneutics. Its distinctive claim is corpus-level compatibility engineering: not “how should I phrase this prompt?” but “what structured domain will a probabilistic system encounter when it retrieves?”
A formal specification defines a semantic space, invariants, relations, boundaries, witness conditions, and success criteria. A Rebekah Cranes case study demonstrates entity anchoring across works, identifiers, and interpretive contexts.