What is the difference between “active” and “inactive” problem recognition?