The principle here is similar. Black’s queen is loose. As a matter of course you also study the constraints on the enemy king’s movement and ask whether you have any sort of mating threat in view. Again White nearly has a mate on the bank rank: if the Black queen moves forward from c7, RxR is likely checkmate. In effect the queen is stuck where it is; the protection it provides the knight is illusory, so the knight is as good as loose. Look for a square from which White’s queen can attack the knight and queen at the same time, taking advantage of the queen’s paralysis. The search leads to Qf7. If Black replies QxQ, White plays RxR#. So instead Black plays Qc8 or Qb8, and White takes the knight. (If Black then plays QxQ, once more White can play RxR, which amounts to mate after Black uselessly brings his queen back. Here as in the previous problem, the queen is committed to guard duty on the back rank.) Or 1.Qf7 can be met with 1. ...Kb8, which gives Black's queen a guard. But then White has 2. RxR, QxR; 3. QxN, still winning a piece.
Think of our current theme this way. We have been studying double attacks; these are simultaneous attacks on two vulnerable points at the same time. A loose piece is a vulnerability, as is the king. If you attack them at the same time, you generally win one or the other. We also saw that even a mere square can be a vulnerability if the queen would deliver checkmate by landing on it; attacking such a square—and thus creating a mating threat—can be as good as attacking the king itself for purposes of creating a double attack. Now we are adding still another point on the board that qualifies as a vulnerability: a queen that is defending against mate. Once you see this—once you identify a way that you could deliver mate if it weren’t for the enemy queen—then attacking the queen becomes itself a kind of mating threat. The usual worry, which is that if you attack the queen it will bite back, is out of the picture.