Gravar e-mail: On the nature of relations learned in Pavlovian conditioning