Learning Principles in Paired-Associate Lists