rorty comments on Deference for Bayesians