Bayesian learning in negotiation