research-document

Hypothesis Registry

Hypothesis Registry

ID Hypothesis Category Confidence Status Supports Depends On Competes With Kill Condition Next Test Last Updated Evidence Runs Dependency Risk Product Relevance
H001 Procedural Invariance: Independently developed procedural artifacts contain recurring structural elements despite surface terminology differences. Structural Moderate Active FE-013; ECR-000001 None H012; H013 Diverse artifacts require unrelated representations with no stable structural overlap. Compare within-family stability across canonical, paraphrased, and structural packets in ECR-000002 2026-07-01 FE-012C; FE-013; ECR-000001; ECR-000002 planned Medium Research-only unless replicated
H002 Representation Independence: Underlying procedural structure is more stable than any single surface representation. Representation Low-Moderate Active FE-012C; FE-013 H001 H012; H014 Primitive, AST, constraint, and natural-language representations produce unrelated results. Compare primitive, AST, and constraint stability within ECR-000002 families 2026-07-01 FE-012C; ECR-000001; ECR-000002 planned Medium Potential Clarity relevance
H003 Multi-Model Convergence: Independent model families converge on similar procedural structures when given the same artifact and instrument. Reliability Moderate Active FE-012C; FE-013 H001; H015 H013; H014 Agreement falls apart across diversified artifacts or repeatability runs. Cross-model comparison in ECR-000002 across recognition conditions 2026-07-01 FE-012C; ECR-000001; ECR-000002 planned High Research-only
H004 Repeatability: The same model family produces materially similar structural analyses across fresh runs. Reliability Low-Moderate Active FE-012C repeatability work H015 H013; H014 Same-model outputs drift materially across repeated runs without explainable packet ambiguity. Repeatability extension after FE-013 execution 2026-06-30 FE-012C repeatability High Research-only
H005 Procedural Grammar: Procedural artifacts have recoverable grammar-like structures. Structural Low-Moderate Active FE-012A; FE-012C; FE-013 H001 H008; H012; H013 Extracted structures cannot be compared meaningfully beyond superficial wording. Compare whether grammar-like structure persists under reduced recognizability in ECR-000002 2026-07-01 FE-012A; FE-012C; FE-013; ECR-000002 planned High Potential Clarity and EDF relevance
H006 Control Flow: Loops, branches, sequencing, and termination conditions are stable features of procedural artifacts. Control Flow Moderate Active FE-012C; FE-013 H001 H014 Control-flow features vary randomly across model, run, or representation. Compare control-flow shape across ECR-000002 recognition conditions 2026-07-01 FE-012C; FE-013; ECR-000002 planned Medium Potential EDF relevance
H007 Constraint Preservation: Procedural artifacts may be better compared by invariant constraints than by literal primitive sequence. Constraint Low Active FE-013 H002; H005 H012; primitive-sequence-only comparison Constraint extraction is inconsistent or fails to explain drift better than primitive comparison. Compare constraint stability to primitive drift within ECR-000002 families 2026-07-01 FE-013; ECR-000001; ECR-000002 planned Medium Potential Clarity relevance
H008 Procedural AST Recovery: Procedural artifacts can be represented as AST-like structures. Representation Low Active FE-013 H005 H002; forced formalism AST schema forces artificial structure or models cannot extract comparable trees. Compare AST presence and AST stability across ECR-000002 recognition conditions 2026-07-01 FE-013; ECR-000001; ECR-000002 planned Medium Research-only
H009 Clarity Relevance: Research findings about procedural reasoning can improve Clarity’s belief-management model. Product Relevance Low-Moderate Active Clarity relevance tracking H001; H007 Research-only interpretation Findings do not improve belief tracking, confidence updates, or decision quality. Record product relevance observations in ECR-000002 without letting them drive extraction 2026-07-01 ECR-000001; ECR-000002 planned Medium Clarity candidate
H010 EDF Relevance: Execution-oriented artifacts exhibit structural patterns relevant to EDF. Product Relevance Low Active EDF relevance tracking H006; H011 Research-only interpretation Execution artifacts do not show separable coordination, sequencing, allocation, or synchronization structure. Record EDF relevance observations in ECR-000002 without treating them as validation 2026-07-01 ECR-000001; ECR-000002 planned High EDF candidate
H011 Reasoning/Execution Separation: Reasoning-heavy and execution-heavy artifacts exhibit distinct structural profiles. Structural Low Active FE-012B; FE-013 H001; H006 H005 Reasoning and execution artifacts remain indistinguishable under multiple representations. Compare P001, P002, and P003 profiles across ECR-000002 conditions 2026-07-01 FE-012B; FE-013; ECR-000001; ECR-000002 planned High EDF candidate
H012 Vocabulary Bias: Provided primitive vocabularies may artificially increase agreement. Bias Moderate Active FE-012C concerns None H001; H002; H005 Vocabulary-free extraction produces the same structures. Compare canonical, paraphrased, and structural wording effects in ECR-000002 2026-07-01 FE-012C; FE-013; ECR-000002 planned Medium Research-only
H013 Recognition Bias: Models may reconstruct known frameworks from training rather than extracting structure from packets. Bias Moderate Active FE-012C concerns None H003; H004; H005 Better-blinded or novel artifacts preserve agreement without recognition. Run recognition-sensitivity comparison in ECR-000002 2026-07-01 FE-012C; ECR-000001; ECR-000002 planned High Research-only
H014 Prompt Robustness: Small prompt changes should not materially alter recovered procedural structure. Reliability Low Active Instrument design concerns H015 H003; H006 Minor prompt changes produce major structural drift. Use shared instruction block across ECR-000002 recognition variants as a bounded robustness check 2026-07-01 ECR-000001; ECR-000002 planned Medium Research-only
H015 Measurement Instrument Reliability: The packet-based measurement instrument can reliably collect comparable procedural-structure outputs. Instrument Moderate Active FE-012C packet instrument None H014; H013 JSON validity, schema compliance, and agreement collapse across runs. Evaluate schema compliance and usable outputs across all ECR-000002 conditions 2026-07-01 FE-012C; FE-013; ECR-000001; ECR-000002 planned Critical Supports all downstream product relevance screening
H016 Structural Recognition: Some models may recognize procedural families from control-flow topology alone, while others require lexical or semantic cues. Bias Low Active Recognition sensitivity analysis H013 H013; H002 If graph-only topology variants consistently reduce recognition while preserving extracted structure, then recognition is not primarily topology-driven. Compare EXP-001 topology perturbation variants in ECR-000003 2026-07-02 ECR-000002; ECR-000003 planned High Research-only
H017 Procedural Compression: Different surface representations of the same underlying procedure may compress into a stable recovered procedural structure. Representation Low Active Representation sensitivity analysis H002; H015 H002 If narrative, bullet, graph, and transition-table representations produce materially unrelated recovered structures, the compression hypothesis weakens. Compare EXP-002 representations in ECR-000003 2026-07-02 ECR-000003 planned High Research-only
H018 Procedural Isomorphism: Different domains can instantiate the same procedural structure, allowing isomorphic procedures to be recognized as structurally similar despite different subject matter. Structural Very Low Proposed Representation sensitivity analysis H002; H015 domain-semantic dependence If isomorphic procedures with identical control-flow produce unrelated structural extractions, the isomorphism hypothesis weakens. Compare EXP-003 domain variants in ECR-000003 2026-07-02 ECR-000003 planned High Research-only