Reply to Bereiter's statement on Kohlberg's cognitive-developmental view