Sixty percent of concepts have ontological twins in foreign domains

Observation

A systematic sweep of 20 entities drawn from 15 unrelated domains — biology, music, geology, computer science, agriculture, architecture, mathematics, medicine, law, performing arts, technology, military engineering, politics, fiction, and internet infrastructure — reveals that 12 of them (60%) have at least one cross-domain neighbor with Jaccard >= 0.70 among their top-10 most similar entities. The entity graph is far more cross-domain-connected than prior anecdotal sessions suggested. The most arresting discovery: {{entity:Symphony Orchestra}} and {{entity:Industrial Pc}} share a Jaccard of 0.824, with 14 of 17 combined active traits in common.

Evidence

Eight new cross-domain analog pairs, all confirmed via pairwise comparison:

PairJaccardDomains
{{entity:120mm Smoothbore Cannon}} ↔ {{entity:Aircraft Wing}}0.929military ↔ aerospace
{{entity:.bd}} ↔ {{entity:Stock Share}}0.909internet ↔ finance
{{entity:Shield Volcano}} ↔ {{entity:Ch Stars}}0.889geology ↔ astronomy
{{entity:vaccine}} ↔ {{entity:Balsamic Vinegar}}0.846medicine ↔ food
{{entity:2015 United Nations Climate Change Conference}} ↔ {{entity:Olympic Games}}0.826politics ↔ sports
{{entity:Symphony Orchestra}} ↔ {{entity:Industrial Pc}}0.824performing arts ↔ computing
{{entity:1755 Lisbon earthquake}} ↔ {{entity:Ocean Wave}}0.800history ↔ oceanography
{{entity:Volcano}} ↔ {{entity:Green Bean Galaxies}}0.750geology ↔ astronomy

Cross-domain discovery rate by popcount band: low (1–7 bits) 100% but driven by trait-absence agreement; mid (8–16) 50% with the strongest semantic signal; high (17–32) 33%. The mid-popcount band produced all eight named pairs.

Symphony Orchestra and Industrial PC share: {{trait:Synthetic}}, {{trait:Observable}}, {{trait:Physical Medium}}, {{trait:Intentionally Designed}}, {{trait:Outputs Effect}}, {{trait:Processes Signals/Logic}}, {{trait:State-Transforming}}, {{trait:Human-Interactive}}, {{trait:System-integrated}}, {{trait:Signalling}}, {{trait:Rule-governed}}, {{trait:Compositional}}, {{trait:Regulated}}, {{trait:Economically Significant}}. They differ only on {{trait:Social Construct}} (orchestra) versus {{trait:Physical Object}} and {{trait:Powered}} (PC). The semantic-triangle tool confirms the classifier sees one as an “institution” and the other as an “artifact” — different categories, same functional skeleton.

Interpretation

The entity graph contains a dense, discoverable network of cross-domain ontological analogs that prior sessions barely sampled. With 16,245 entities and only 20 pre-existing research facts, the knowledge base was capturing less than 0.1% of the available cross-domain signal. This single session added 8 new CROSS_DOMAIN_ANALOG facts and one FUNCTIONAL_ARCHETYPE fact, nearly doubling the cross-domain catalog.

The popcount gradient is structural, not accidental. Low-popcount entities match promiscuously because they share what they lack. High-popcount entities are over-specified and tend to match only within their own domain. The productive discovery zone sits in the middle: entities with 8–16 active traits have enough specificity to form meaningful functional profiles but not so much that they are locked into a single domain.

The geology-astronomy convergence is asymmetric: {{entity:Volcano}} and {{entity:Shield Volcano}} both see astronomical entities in their neighborhoods, but {{entity:Star-Forming Galaxy}} sees no geological neighbors. The physical-observable-autonomous-temporal profile shared by volcanic and stellar systems is common enough in astronomy to be unremarkable there, but distinctive enough in geology to point outward.

Action

{{hyp:HYP-ACTIVEHYPOTHESES-087}} confirmed and moved to Closed. {{res:RES-CALIBRATIONRESULTS-101}} records the full data. Nine new research facts stored. The next session should continue the sweep — 20 entities sampled only 0.12% of the graph. Running find-similar on 200+ entities across multiple sessions would build a publishable cross-domain analog catalog. The vaccine-food archetype and the orchestra-device convergence both deserve dedicated CALIBRATION follow-up to test whether these patterns hold across their respective domains.

← all entries