research-document
Evidence To Theory Map
Evidence To Theory Map
Purpose: Show how each major experiment feeds into theory statements, predictions, and future work.
FE-008
Observed findings
- Identity-Capability separation reduced overlap pressure in artifact characterization.
- Artifact characterization appears more stable when observable properties are recorded before classification.
Theory statements supported
- Identity-Capability separation improves characterization.
- Framework Engineering currently functions as a characterization methodology.
Predictions generated
- Independent reviewers should converge more reliably when identity and capability are separated.
- Artifact characterization should remain more auditable than single-label classification.
Future experiments required
- Larger blind reviewer comparison
- Comparison against alternative characterization schemes
FE-011A
Observed findings
- Structured redesigns improved traceability and evidence handling.
- Added structure also increased complexity and overhead.
- The gains may be partly explained by structure in general rather than by Framework Engineering specifically.
Theory statements supported
- Structured redesign improves traceability.
- Framework Engineering currently functions as a redesign methodology.
- Several provisional design heuristics may improve reasoning quality.
Predictions generated
- Frameworks separating observation, interpretation, uncertainty, and verification will often produce more auditable outputs.
- Simplicity and usefulness may degrade if added structure exceeds the actual problem need.
Future experiments required
- Independent evaluator runs
- Controls using non-FE structured baselines
- Human participant validation
FE-012A
Observed findings
- A small primitive vocabulary reconstructed many reasoning-heavy artifacts with moderate or strong recognizability.
- Vocabulary growth slowed after the middle of the artifact set.
- Representational and hierarchy-oriented artifacts reconstructed less cleanly.
Theory statements supported
- Procedural reasoning may be represented by a finite primitive vocabulary.
- Reasoning-heavy artifacts may compress more cleanly into a primitive grammar than other artifact types.
Predictions generated
- Independent extractors will converge more strongly on reasoning-heavy artifacts than on coordination-heavy artifacts.
- Primitive stabilization will weaken outside familiar procedural families if the current vocabulary is incomplete.
Future experiments required
- Independent extractor comparison
- Merge-disagreement audit
- Expansion into coordination-heavy artifact sets
FE-012B
Observed findings
- The frozen primitive vocabulary synthesized coherent procedural structures in many novelty problems.
- Coordination-heavy and phased-maintenance problems created missing primitive pressure.
CompareandPrioritizeremained structurally important rather than collapsing away.
Theory statements supported
- Primitive vocabulary can synthesize coherent procedural structures.
- Reasoning grammar and execution grammar may be distinct.
- Framework Engineering is currently stronger at characterization and redesign than at full framework generation.
Predictions generated
- Coordination-heavy artifacts will require additional execution primitives more often than diagnostic artifacts.
- A generic reasoning grammar may transfer further than a coordination grammar.
Future experiments required
- Independent synthesis and review
- Direct tests of candidate execution primitives
- Reasoning-heavy versus coordination-heavy comparison studies
Traceability Summary
FE-008 contributes most strongly to characterization theory.
FE-011A contributes most strongly to redesign and heuristic theory.
FE-012A contributes most strongly to reconstruction and primitive vocabulary theory.
FE-012B contributes most strongly to synthesis limits and grammar-splitting hypotheses.