A04 — Separate reasoning and coordination grammars

• Input [inference]: A procedural artifact containing both reasoning functions and timing, allocation, synchronization, sequencing, or coordinated action.
• Operation [proposal-only]: Represent reasoning functions in one grammar and execution/coordination functions in another; retain possible bridge operations without assuming their status.
• Output [proposal-only]: Two related representations plus recorded bridge candidates and unresolved entanglements.
• Scope conditions [direct]: The hypothesis is motivated where coordination-heavy synthesis pressures the current reasoning vocabulary.
• Failure conditions [direct]: The grammars cannot be defensibly separated, or coordination-heavy work succeeds after only minor refinement of one grammar family.
• Predicted effect [direct]: Coordination-heavy artifacts will request additional execution primitives more often than diagnostic artifacts; extractor convergence should be higher for reasoning-heavy artifacts.
• Measurement [direct]: Missing-primitive requests and independent extractor convergence by artifact family.
• Named baseline [inference]: One refined grammar used for both reasoning and coordination.
• Boundary condition [direct]: Candidate execution operations have not been independently extracted from blinded artifacts.
• Implementation test [direct]: Blind extraction and synthesis on reasoning-heavy and coordination-heavy families, with and without candidate execution operations.
• Decision rights [absent]: No canonical owner or interface adjudicator.
• Uncertainty handling [direct]: Candidate execution and bridge operations remain hypotheses; unsettled boundaries are preserved.
• Verification [direct]: Direct family comparison and independent synthesis/review.
• Provenance [direct]: Link candidates to FE-012A/FE-012B pressure observations and disagreement records.
• Implementation cost [absent]: Not measured.
• Limitations [direct]: Internal-run evidence, no independently frozen execution vocabulary, and a competing one-grammar explanation.
• Fidelity check [inference]: The test records whether separation was actually possible and does not treat candidate operations as accepted.
• Null-capable future experiment [direct]: Coordination artifacts may be represented by a modestly refined single grammar, weakening the split hypothesis.

Statement basis

• Direct: Hypothesis, rationale, predictions, weaknesses, and weakening test.
• Inference: Input, baseline, fidelity check, and some provenance handling.
• Proposal-only: Two-representation operation and output.
• Absent: Interface contract, decision authority, and cost estimate.