research-document
Hypothesis Review Board
Hypothesis Review Board
| Hypothesis | Current Confidence | Evidence For | Evidence Against | Dependencies | Competing Hypotheses | Decision | Next Action |
|---|---|---|---|---|---|---|---|
| H001 Procedural Invariance | Moderate | H003; H015 | H012; H013 | keep | Run ECR-000001 | ||
| H002 Representation Independence | Low-Moderate | H004; H008; H015 | H012; H014 | keep | Compare layers in ECR-000001 | ||
| H003 Multi-Model Convergence | Moderate | H015 | H013; H014 | keep | Evaluate cross-model agreement | ||
| H004 Repeatability | Low-Moderate | H015 | H013; H014 | keep | Extend repeatability after FE-013 | ||
| H005 Procedural Grammar | Low-Moderate | H001; H015 | H008; H012; H013 | keep | Evaluate FE-013 and ECR-000001 | ||
| H006 Control Flow | Moderate | H005; H015 | H014 | keep | Compare loops and branches | ||
| H007 Constraint Preservation | Low | H006; H015 | literal comparator sufficiency | keep | Test constraint layer | ||
| H008 Procedural AST Recovery | Low | H005; H006; H015 | forced formalism | keep | Compare AST quality | ||
| H009 Clarity Relevance | Low-Moderate | H001; H002; H007 | research-only insight | keep | Track product relevance only | ||
| H010 EDF Relevance | Low | H006; H011 | unrelated execution theory | keep | Track execution relevance only | ||
| H011 Reasoning/Execution Separation | Low | H006; H010 | H005 | keep | Compare profiles across artifacts | ||
| H012 Vocabulary Bias | Moderate | None | H001; H003; H005 | keep | Compare vocabulary-free runs | ||
| H013 Recognition Bias | Moderate | None | H003; H005; H015 | keep | Track recognition rate | ||
| H014 Prompt Robustness | Low | H015 | H003; H006 | keep | Plan prompt-robustness run | ||
| H015 Measurement Instrument Reliability | Moderate | None | H014; H013 | keep | Review schema compliance | ||
| H016 Structural Recognition | Low | H013 | H013; H002 | keep | Compare P001C to P001D graph-only | ||
| H017 Procedural Compression | Low | H002; H015 | H002 | keep | Compare EXP-002 representations | ||
| H018 Procedural Isomorphism | Very Low | H002; H015 | domain-semantic dependence | keep | Compare EXP-003 domain variants |
ECR-000001 Limited Review Pending
- Ready for limited review: H003, H012, H015, H007, H008
- Blocked/deferred: H013, H005, H006, H009, H010
- Required before confidence update: human review of workspace findings
ECR-000002 Recognition Sensitivity Planned
- Primary focus: H013, H003, H015
- Secondary coverage: H001, H002, H005, H006, H007, H008, H011, H012, H014, H016
- Exploratory only: H009, H010
- Required before confidence update: manual execution, comparison, and human review
ECR-000003 Representation Sensitivity Planned
- Primary focus: H016, H017, H018
- Supporting coverage: H013, H003, H015, H002
- Required before confidence update: phase-gated manual execution, comparison, and human review
EDR Gate
- No evidence run may update confidence directly.
- Evidence must pass through an accepted Evidence Decision Record before hypothesis review.
- Next experiments should be pulled from the research queue after EDR review.