Principlev1

Evaluate rival experts by assessing the structure and

Evaluate rival experts by assessing the structure and transparency of their arguments, agreement from other experts, appraisals from meta-experts, evidence of interests and biases, and their track records of calibrated prediction.

Why This Is a Principle

Derives from The performance of an agent is bounded by the accuracy of (agent performance bounded by world model accuracy), Systematic Overconfidence Taxonomy (overconfidence), and Domain-Specific Calibration Development (calibration from feedback). This is Goldman's framework operationalized—it prescribes specific evaluation criteria that follow from axioms about knowledge reliability and calibration. Actionable and general.

Source Lessons

L-0374

Contradictions in expert advice

When experts disagree the disagreement itself contains information about the limits of current knowledge. Expert contradiction is not a failure of expertise — it is a map of where the evidence runs out, where hidden variables lurk, and where your own epistemic work must begin. The most dangerous response is not confusion but premature certainty: picking one expert and ignoring the other destroys the signal the disagreement was carrying.

Connections

Derived From (3)

AxiomSystematic Overconfidence Taxonomy AxiomDomain-Specific Calibration Development AxiomThe performance of an agent is bounded by the accuracy of

Back to Concepts