The Value of Normative Information

Abstract

The assumption that acquiring normative information is instrumentally valuable imposes strict structural constraints on decision-making under normative uncertainty. Standard decision theories and straightforward aggregation principles, such as Expected Choiceworthiness, fail to value normative information because they conflate unconditional and conditional evaluations. Guaranteeing that learning complete normative truths is instrumentally worthwhile requires a two-level framework that aggregates the first-order conditional evaluations of competing theories. Furthermore, extending this instrumental value to partial normative information necessitates an additively separable aggregation function. However, endorsing the instrumental value of normative information is inconsistent with the classical principle that empirical information is always instrumentally valuable, provided that risk-averse frameworks like risk-weighted expected utility remain epistemically possible. Because risk-averse theories can rationally mandate avoiding empirical evidence, metanormative hedging transmits this resistance to evidence into actual choice, demonstrating that normative and empirical uncertainty are structurally distinct. Ultimately, theories of decision-making must either abandon the universal value of empirical information, reject the value of normative information, or dismiss non-additive first-order theories entirely, while avoiding regress by treating metanormative uncertainty as ordinary normative uncertainty. – AI-generated abstract.