> Wouldn’t weighing evidence play a key role in assessing relative likelihoods of two possible causes of an event that had already occurred?

Absolutely, yes. However, it would not affect the priors.