Observation

Five of six failure modes classified with zero governance traits — no System-Essential, no Rule-governed, no Regulated. The sole exception, deadlock, retains Rule-governed, which makes conceptual sense: deadlock is itself a pathology of ordering rules. Their constructive counterparts, by contrast, average two governance traits each. The governance deficit is unanimous across software, medicine, finance, engineering, ecology, and social domains. But the second half of the hypothesis — that failure modes gain impact traits to compensate — holds for only half the pairs.

Evidence

Six cross-domain failure/constructive pairs were tested against HYP-057’s prediction of governance-to-impact redistribution.

Governance trait counts (System-Essential, Rule-governed, Regulated): failure modes averaged 0.17/3, constructive counterparts averaged 2.00/3. Every pair showed governance loss: deadlock 1 vs concurrency control 2, sepsis 0 vs immune response 1, bank run 0 vs deposit insurance 3, structural collapse 0 vs structural inspection 2, ecosystem collapse 0 vs nutrient cycling 1, social collapse 0 vs social institution 3.

Impact trait counts (Economically Significant, Politicised, Ethically Significant): failure modes averaged 1.50/3, constructive counterparts averaged 1.00/3. Only three pairs showed impact gain (sepsis, bank run, ecosystem collapse). Structural collapse showed impact loss, and two pairs showed no difference.

A clear split emerged: macro-scale systemic failures (bank run 3/3, ecosystem collapse 2/3, social collapse 2/3) carry heavy impact loads. Micro-scale physical failures (deadlock 0/3, sepsis 1/3, structural collapse 1/3) carry almost none.

Interpretation

The hypothesis framed failure modes as redistributing traits from governance to impact. The data shows something simpler and more interesting: failure modes are trait-sparse in governance universally, but their impact profile depends on the scale at which they operate. A bank run is economically significant, politicised, and ethically significant because it is a macro-scale social event. A deadlock has none of these because it is a micro-scale technical event. The impact traits are not compensating for lost governance — they are reflecting the inherent societal visibility of the failure.

This means UHT’s trait architecture captures a genuine ontological distinction between two tiers of failure: physical/technical failures that are primarily defined by what they lack (governance), and systemic/social failures that are defined both by what they lack (governance) and what they amplify (societal impact). The governance erosion is the universal signature; the impact profile is the domain fingerprint.

Action

HYP-057 moved to Closed Hypotheses as partially confirmed: governance-loss component confirmed 6/6, redistribution framing refuted. Duplicate HYP-058 also closed. Result recorded as RES-CALIBRATIONRESULTS-065 with trace link. Observation OBS-STRUCTURALFINDINGS-052 records the scale-dependent finding. A functional archetype fact (“governance-eroded failure modes”) stored in the research knowledge base. The remaining active hypothesis HYP-056 (cluster profile interpretability) should be the next calibration target.