Epistemic Systems & AI Governance
The Correction Trilemma: A Diagnostic Heuristic for Trade-offs in Epistemic Systems
Introduces registration, traceability, and receptivity as three core capacities required for self-correcting epistemic systems. The paper frames recurring failure modes as Corrective Chaos, Elite Dogmatism, and Analytical Paralysis.
Overview
This work gives NourScene a strong vocabulary for research integrity, AI governance, institutional correction, and metascience. It is especially relevant to organizations using AI, data systems, expert review, or public knowledge infrastructures.
Core Capacities
- Registration: the system detects and formalizes counterevidence.
- Traceability: the system can localize errors to sources, components, or processes.
- Receptivity: the system can implement corrective action after diagnosis.
Failure Patterns
- Corrective Chaos: response without reliable causal grounding.
- Elite Dogmatism: optimized systems that fail to register emerging anomalies.
- Analytical Paralysis: detected and diagnosed errors that do not lead to action.
