Michael Rescorla (University of California, Los Angeles)
Kolmogorov's theory of conditional probability generalizes the ratio formula so as to define conditional probabilities when the conditioning proposition has probability 0. Recently, several authors have suggested that Kolmogorov's theory can illuminate conditionalization on propositions with initial probability 0. I pursue a version of this strategy, articulating a diachronic norm that I call Non-factive Kolmogorov Conditionalization (NKC). NKC encompasses numerous scientifically and philosophically important scenarios, including numerous scenarios where the agent becomes certain of a false proposition that has initial probability zero. I highlight key advantages that NKC offers over all rival credal update rules.
