When coming up with inference rules one may “think” that a rule holds when in fact it does not. In order to show that a rule does not hold you must find (using your imagination) at least one instance in which the given functional dependencies hold and in which the “supposedly implied” functional dependencies do not hold.