CONVERGENCE OF ADAPTIVE LEARNING AND EXPECTATIONAL STABILITY: THE CASE OF MULTIPLE RATIONAL-EXPECTATIONS EQUILIBRIA