RFR-2026-005: Reasoning-versus-coordination grammar test

Research opportunity

Test whether reasoning and execution grammars are separable, overlapping, or a single context-sensitive grammar.

Background

FE-012B failures cluster in coordination-heavy domains and the theory predicts vocabulary pressure there.

Origin documents and trace

• research/experiments/FE-012B-primitive-synthesis/methodological-review.md, section: Problems that resisted synthesis; Overall Assessment
• research/theory/theory-of-framework-engineering-v0.1.md, section: Predictions; Open Questions
• research/theory/falsification-criteria.md, section: Theory Weakens If

Specific assumption challenged: Missing coordination terms are reusable primitives rather than domain-specific vocabulary.

Supporting evidence: Underwater robotics, lunar maintenance, and cathedral restoration resisted synthesis; candidate terms include synchronize, schedule, sequence, allocate, and coordinate.

Unknowns

• Whether coordination primitives generalize
• Grammar boundary
• Domain versus artifact-family confounding

Dependencies

• RFR-2026-002
• RFR-2026-007

Suggested REP and methodology

Create REP-2026-005. Use a crossed sample of reasoning/coordination intensity and domains; independently extract operations; compare one-grammar, two-grammar, and context-sensitive models.

Expected outputs

• Balanced corpus extension
• Competing grammar models
• Model-selection report

Success criteria

Predictions and model-selection criteria are frozen before coding, and a simpler competing model can win.

Execution recommendation

• Recommended agent: theory-testing agent
• Estimated effort: high
• Expected knowledge gained: Refines or falsifies a central theoretical boundary.
• Frontier Score: 392
• Score inputs: knowledge gain 5/5; impact 4/5; cross-project reuse 4/5; scientific importance 5/5; dependency cost 4/5; implementation difficulty 4/5.
• Score calculation: 5 × 4 × 4 × 5 − 4 − 4 = 392.
• Score status: structured expert judgment, not measured utility.