Empirical proof and disproof are, then, unobtainable. But this conclusion will seem less depressing if we remember that prayer is request and compare it with other specimens of the same thing.
We make requests of our fellow creatures as well as of God: we ask for the salt, we ask for a raise in pay, we ask a friend to feed the cat while we are on our holidays, we ask a woman to marry us. Sometimes we get what we ask for and sometimes not. But when we do, it is not nearly so easy as one might suppose to prove with scientific certainty a causal connection between the asking and the getting.
Your neighbour may be a humane person who would not have let your cat starve even if you had forgotten to make any arrangement. Your employer is never so likely to grant your request for a raise as when he is aware that you could get better money from a rival firm and is quite possibly intending to secure you by a raise in any case. As for the lady who consents to marry you—are you sure she had not decided to do so already? Your proposal, you know, might have been the result, not the cause, of her decision. A certain important conversation might never have taken place unless she had intended that it should.
Thus in some measure the same doubt that hangs about the causal efficacy of our prayers to God hangs also about our prayers to man. Whatever we get we might have been going to get anyway. But only, as I say, in some measure. Our friend, boss, and wife may tell us that they acted because we asked; and we may know them so well as to feel sure, first that they are saying what they believe to be true, and secondly that they understand their own motives well enough to be right. But notice that when this happens our assurance has not been gained by the methods of science. We do not try the control experiment of refusing the raise or breaking off the engagement and then making our request again under fresh conditions. Our assurance is quite different in kind from scientific knowledge. It is born out of our personal relation to the other parties; not from knowing things about them but from knowing them.
Our assurance—if we reach an assurance—that God always hears and sometimes grants our prayers, and that apparent grantings are not merely fortuitous, can only come in the same sort of way. There can be no question of tabulating successes and failures and trying to decide whether the successes are too numerous to be accounted for by chance. Those who best know a man best know whether, when he did what they asked, he did it because they asked. I think those who best know God will best know whether He sent me to the barber’s shop because the barber prayed.
C.S. Lewis, “The Efficacy of Prayer,” originally from The World’s Last Night and Other Essays, here from The Essential C.S. Lewis (Touchstone, 1996) 380.