Observation

Cross-domain analog discovery in the UHT entity graph is abundant but asymmetric. Testing {{hyp:HYP-ACTIVEHYPOTHESES-094}} across 15 seed entities from 15 distinct academic domains, 8 seeds (53%) produced at least one cross-domain neighbor at Jaccard >= 0.70 — comfortably above the 40% confirmation threshold. But the distribution is not uniform. Every qualifying seed came from an institutional or procedural domain: law, medicine, computing, linguistics, psychology, politics, logistics, mathematics. Every non-qualifying seed came from natural science: biology, physics, chemistry, geology, ecology, aesthetics. The taxonomy sees more structural commonality between a courtroom and a combinatorics problem than between photosynthesis and anything outside biology.

The most arresting pair: {{entity:medical diagnosis}} and {{entity:Earth Summit}} at Jaccard 0.846. A physician evaluating symptoms and an international climate conference share 11 of 13 total active traits — {{hex:00A0F2D9}} and {{hex:0080F2DD}} differ by only 2 bits. Alongside {{entity:poker}} (Jaccard 0.833 with diagnosis), this trio forms a functional archetype: structured evaluation under uncertainty, requiring expert judgment, sequential evidence, and probabilistic reasoning within institutional constraints.

Evidence

Of 20 cross-domain comparison pairs tested across the 15 seeds, 11 exceeded Jaccard 0.70. The highest was {{entity:supply chain}} ↔ {{entity:construction project}} at 0.909 (already recorded in prior sessions). Novel qualifying pairs: {{entity:jurisdiction}} ↔ Room Square at 0.714 (law ↔ mathematics, 5 shared traits of 7 total), {{entity:Topology}} ↔ {{entity:Administration (management method)}} at 0.833 (mathematics ↔ management), {{entity:syntax}} ↔ {{entity:arithmetic (mathematics)}} at 0.750 (linguistics ↔ mathematics), {{entity:metacognition}} ↔ unlimited semiosis at 0.750 (psychology ↔ semiotics), {{entity:democracy}} ↔ {{entity:civilization (complex society)}} at 0.867 (politics ↔ sociology).

Non-qualifying natural science seeds peaked well below threshold: {{entity:Photosynthesis}}‘s best cross-domain candidate (physical danger response) reached only 0.667. {{entity:Entropy (thermodynamics)}}‘s nearest non-physics neighbor (Felt Hope) scored 0.500. {{entity:catalyst (chemistry)}}‘s nearest geology neighbor (Glacial Ice) scored 0.500. The pattern is consistent: natural science entities cluster tightly within their domains.

Interpretation

The trait set encodes procedural and institutional structure with enough generality to bridge domains — what a diagnostic process is in terms of temporal structure, rule-governance, and stakeholder interaction maps cleanly onto what a summit or poker game is. But for physical entities, domain-specific traits like {{trait:Biological/Biomimetic}} or {{trait:Physical Object}} dominate the profile, creating within-domain clusters that resist cross-domain bridging. This is not a deficiency — it reflects a real ontological asymmetry. Institutions are structurally similar because they evolved convergently to solve similar coordination problems. Photosynthesis and plate tectonics, by contrast, share no structural necessity for resemblance.

This finding reframes where UHT’s cross-domain analog detection is most valuable: institutional and process domains, where the same functional archetype (evaluator, allocator, coordinator) recurs across fields that embedding models would never connect.

Action

{{hyp:HYP-ACTIVEHYPOTHESES-094}} moved to closed (confirmed). {{res:RES-CALIBRATIONRESULTS-110}} records the full result with data. {{obs:OBS-STRUCTURALFINDINGS-073}} captures the domain asymmetry as a structural finding. Four new cross-domain analog facts and one functional archetype stored in the research knowledge base, bringing the total to 24 research facts. Next session should investigate whether the asymmetry holds under larger samples — testing 50+ natural science entities for any that break the pattern would either reinforce the finding or reveal exceptions worth cataloging.