Skip to content

Weakeners

A weakener is a structural credibility gap detected by the Jena rule engine. Each pattern has an ID, a severity, an affected node, and a description. Weakeners are not errors. They are evidence-package quality alerts.

The current core pack ships 23 patterns: 21 Level-1 patterns plus 2 active compound patterns. Domain packs extend the catalog additively — the nasa-7009b pack contributes 6 NASA-STD-7009B-specific patterns (W-NASA-01W-NASA-06), bringing the total to 29 when both packs are loaded. See the auto-generated catalog reference for the full table sourced from the live code.

CategoryPatternsExamples
EpistemicW-EP-01 to W-EP-04Orphan claim, broken provenance, evidence-source gap, unassessed factors at elevated risk
AleatoryW-AL-01, W-AL-02Missing uncertainty quantification, missing sensitivity analysis
OntologicalW-ON-01, W-ON-02Applicability gap, operating-envelope gap
StructuralW-SI-01, W-SI-02Internal consistency gaps
ArgumentationW-AR-01 to W-AR-05Comparator absence, eliminative-argumentation gap, residual-risk gap
ConsistencyW-CON-01 to W-CON-05Decision-vs-factor mismatch, profile-vs-evidence mismatch
ProvenanceW-PROV-01Provenance-chain orphan
CompoundCOMPOUND-01, COMPOUND-03Risk escalation, assurance-level override

Run uofa catalog to see the full set with descriptions and severity assignments.

SeverityMeaningExample
CriticalMissing evidence the standard explicitly requires for the declared rigor levelW-EP-01 (orphan claim)
HighSignificant credibility gap that affects defensibilityW-AL-01 (missing UQ)
MediumStructural gap that should be addressed before submissionW-CON-04 (sensitivity analysis not linked)
LowMinor evidence-completeness suggestion(rare; no Level-1 pattern at Low currently)

Severity is calibrated against the V&V 40 risk-tier framework. The same factor missing at MRL 2 may be Medium; missing at MRL 5 it is Critical.

Compound rules fire on the output of Level-1 rules. This chained inference is the differentiator over standalone SPARQL queries.

PatternSeverityFires when
COMPOUND-01CriticalCritical and High weakeners coexist on the same UofA
COMPOUND-03HighDeclared assurance level is inconsistent with detected gaps

A second compound pattern (COMPOUND-02 — factor credibility erosion) is implemented but commented out pending v0.6 calibration.

Morrison et al. (2019) is the FDA-co-authored worked example for ASME V&V 40. Re-expressed as a UofA evidence package, COU1 (CPB Use, Class II, MRL 2, Accepted) fires 11 weakeners across 5 patterns:

PatternSeverityHitsWhat it detects
W-AL-01High3Missing UQ on validation results
W-AR-05High3Comparator absence — results not linked to reference entities
W-EP-02High3Broken provenance — validation results with no generation activity
W-ON-02High1COU lacks both applicability constraint and operating envelope
W-CON-04Medium1Complete profile with no sensitivity analysis linked

No compound firings at MRL 2. Morrison’s CFD-side UQ omission is methodologically defensible at this risk level (the rationale is that bench testing carries the safety determination), but the rule engine still flags it as a potential reuse blocker at higher MRLs.

uofa diff surfaces how the same model produces different weakener profiles under different risk contexts.

uofa diff packs/vv40/examples/morrison/cou1/uofa-morrison-cou1.jsonld \
packs/vv40/examples/morrison/cou2/uofa-morrison-cou2.jsonld
PatternCOU1 (MRL 2)COU2 (MRL 5)Why
W-PROV-0107COU2 has 7 provenance-chain orphans
W-EP-0406COU2 has 6 unassessed factors at MRL 5 > 2
W-EP-0230COU2’s regenerated validation activities have proper provenance
W-AL-0130COU2 has Monte Carlo UQ in place of MRL-2 omission
W-AR-0530COU2 results explicitly linked to comparator entities
W-AL-0201COU2 reports UQ but no sensitivity analysis
COMPOUND-0102Critical + High coexistence fires only at MRL 5

Same model. Same data. Different model risk produces a measurably different credibility evidence profile. This is the central analytical demonstration of the construct.