Critiquing Human Judgment Using Knowledge-Acquisition Systems