It is tempting sometimes, when defending a position, to sweep certain cases under the rug, saying they are exceptional, or that they are troublesome for the competing positions. This may sometimes be an appropriate thing to do, but it is something a philosopher or theologian should not feel too good about. If my semantic theory is incompatible with, say, some obvious fact about the Liar Paradox, my semantic theory is, well, false. Sometimes one may hope that some small tweak will get one out of the problem, or that one can add an exception clause. But that is a dangerous thought, for it might be that the problem does in fact show that a completely different approach is needed.
Take the bearing that cases of small children, as well as of those the severely mentally retarded, have for accounts of salvation. Let us suppose that one theologizes and concludes that belief that Jesus is Lord is necessary for salvation. But then one is faced with the case of small children who, as far as we can tell, lack the conceptual resources to have the belief that Jesus is Lord. What can one do? Well, one could bite the bullet, and say that no small children are saved. One could mitigate this, then, by invoking a doctrine of limbo. But in fact nobody that I know makes this move. Or one could say that God can miraculously make it possible for a small child that he foreknows will soon die to have such a belief. I have heard one or two people be friendly towards this answer—I myself think it is not that implausible.
But the usual thing to say is just to make an exception of the case, and say that the belief that Jesus is Lord is necessary for salvation, except in the case of children and mentally retarded individuals. This is not a solution one should feel good about, though. First of all, as soon as one starts introducing some exception clauses, one should start worrying whether there aren't more. Secondly, one worries about continuity between exceptional cases and non-exceptional one (intellectual maturation seems to be a continuous process). But, perhaps most importantly, one should look for a unified theory that takes care of the exceptional cases.
There are, in fact, unified theories for this problem. Here is one. What is necessary for salvation is that one believe the Christian teaching to an extent proportionate to one's capacities. One thing that appeals to me about this formulation is that it does not simply let infants off the hook. It lets them off the hook, while increasing the requirements for those whose capacities are greater than average. More is asked of those to whom more is given.
And this unified theory, then, very naturally leads to further questions. What if someone's lack of an ability to believe is not due to an internal gap, but due to an external one? Does it, after all, matter whether one is incapable of belief due to internal causes, such as lack of intellectual capacity, or due to external causes, such as brainwashing or not having heard the gospel from someone who wasn't distorting it through living a life incompatible with it? Thus, a unified theory has the advantage of raising further questions, and leading perhaps to the solution of other difficulties.