Logical and Indexical Uncertainty — LessWrong