Rational Beliefs and Bayesian Learning : a Note