A redated post.
This is from Lewis's essay "On Obstinacy of Belief," to which I have linked. This is Lewis's classic response to the argument from wish fulfillment. It isn't a matter of riding Lewis's coattails here, it is simply pointing out that if you want to use a wish fulfillment argument against Christianity, this is the response you've got to answer. If you want to defend the argument from design, you've got to answer Hume. If you want to defend the evidentialist objection, you've got to answer Plantinga. If you want to defend psychological egoism, you've got to answer Butler. If you want to defend the wish fulfillment argument, you've got to answer Lewis.
Lewis's claim is that wish fulfillment arguments can be made on all sides, so they are pretty much useless to any side in particular.
There are of course people in our own day to whom the whole situation seems altered by the doctrine of the concealed wish. They will admit that men, otherwise apparently rational, have been deceived by the arguments for religion. But they will say that they have been deceived first by their own desires and produced the arguments afterwards as a rationalization: that these arguments have never been intrinsically even plausible, but have seemed so because they were secretly weighted by our wishes.
Now I do not doubt that this sort of thing happens in thinking about religion as in thinking about other things; but as a general explanation of religious assent it seems to me quite useless. On that issue our wishes may favour either side or both. The assumption that every man would be pleased, and nothing but pleased, if only he could conclude that Christianity is true, appears to me to be simply preposterous.
If Freud is right about the Oedipus complex, the universal pressure of the wish that God should not exist must be enormous, and atheism must be an admirable gratification to one of our strongest suppressed impulses. This argument, in fact, could be used on the theistic side. But I have no intention of so using it. It will not really help either party. It is fatally ambivalent. Men wish on both sides: and again, there is fear-fulfilment as well as wish-fulfilment, and hypochondriac temperaments will always tend to think true what they most wish to be false.
Thus instead of the one predicament on which our opponents sometimes concentrate there are in fact four. A man may be a Christian because he wants Christianity to be true. He may be an atheist because he wants atheism to be true. He may be an atheist be-cause he wants Christianity to be true. He may be a Christian because he Wants atheism to be true. Surely these possibilities cancel one another out? They may be of some use in analysing a particular instance of belief or disbelief, where we know the case history, but as a general explanation of either they will not help us. I do not think they overthrow the view that there is evidence both for and against the Christian propositions which fully rational minds, working honestly, can assess differently.