Jeffrey conditionalization: against caution
摘要
Borut Trpin has raised a novel objection to Jeffrey conditionalization (JC). He argues that JC can lead to increasing confidence in a false hypothesis, based on non-misleading observations. But I show that on a standard way of understanding whether observations are misleading, that the observations in his case would be misleading. I show why Trpin’s response to this objection is inadequate, and use two probabilistic models to show that (1) the observations actually are misleading and (2) if they are made non-misleading, JC leads to certainty in the true hypothesis. Finally, I prove that the cases that he is worried about cannot arise in general.