Ethicists have thus far not paid much attention to uncertainty, very often concentrating on highly idealized hypothetical situations where both empirical (e.g. the state of the world, the spectrum of possible decisions and their consequences, the causal connections between events) and normative (the content of norms, value scales) matters were clearly defined and well-known to the decision-maker. In this article, which stems from a project on different types of decisions under uncertainty related to the rapid progress in biomedical research, I analyze some situations of normative uncertainty, cases when an agent must make a decision, but does not know which choice is correct, for example, because he/she has contrary intuitions about the permissibility of available decisions. The view termed comparativism claims that in such cases the appropriate decision depends not only on the credences that one assigns to different norms, but also on how much possible decisions are worth taking in the light of these norms. I analyze a few cases of normative uncertainty, and a specific counter-argument against the current versions of comparativism, showing that under normative uncertainty this view imposes risk neutrality, although it permits us to have different risk attitudes under empirical uncertainty. I also argue that a precautionary approach to situations of normative uncertainty is overly simplistic.