Abortionists sometimes cite popular belief in universal infant salvation as a wedge tactic to taunt Christians: If you believe all babies are heavenbound, why do you oppose abortion? This is meant to generate a dilemma: logically, you should either support both or oppose both.
John Piper recently posted on this subject:
Given the cards he dealt himself, I think he played his hand fairly well. That said:
i) Speaking for myself, I'm dubious about universal infant salvation. All the world's worst people used to be cute little kids. I can't help mentally rewinding the clock. Go back in time from what they are to what they were.
Seems arbitrary to say that if you die at seven you fly to heaven, but if you die at nine you fry.
We see children as they are, not as they will be. At least initially. Sometimes we live long enough to see how they turn out–for better or worse.
So I doubt a key premise of the argument. But even if I didn't, I don't think the argument goes through.
ii) If this poses a dilemma at all, it only poses a dilemma for freewill theists rather than Calvinists. The unstated premise of the argument is that people can lose their salvation. Hence, if somebody is now saved, killing him now is the way to seal his salvation. If salvation can be lost, it is risky to live another day. To play it safe, die when you are saved. The longer you wait, the greater the risk that you will died unsaved.
Incidentally, the logic of that argument is hardly confined to infants. It would apply just as well to born-again adults.
iii) But, of course, Calvinism rejects the operating premise. What ensures your salvation is not when you die, but election–which is unalterable. Not, in the first instance, what happened in time, but what happened in eternity. The elect can't lose their salvation. You either have it or you don't.
From a Reformed standpoint, nothing you do can change the number of the elect. In the classic formulation of the Westminster Confession: "These angels and men, thus predestinated, and foreordained, are particularly and unchangeably designed, and their number so certain and definite, that it cannot be either increased or diminished" (WCF 3:4).
iv) But it might be argued that this misses the point. The claim is not that we retroactively cause God to elect more people if more babies die in the womb. The claim, rather, is that if more (elect) babies die in the womb, then that's how God predestined the end-result all along. Our alternate course of action (i.e. aborting elect babies) is the consequence of God's foreordination, rather than God's foreordination as the consequence of our alternate course of action.
v) There is, however, a basic problem with that argument. It's a counterfactual scenario. As such, it doesn't refer to the world in which you and I actually live, but to an alternate timeline.
But even if you believe in universal infant salvation vis-a-vis the actual world, you can't just switch to an alternate timeline, yet assume everything else remains the same. Even if your thought-experiment only changes on variable, that's just a thought-experiment. You can conjecture that God might do it that way, but it's not as if you have given God a blueprint which he must follow.
Suppose there's a possible world in which some people kill their children in the superstitious belief that doing so will ensure their salvation. It doesn't follow that in fact raises the number of the elect. For in that alternate timeline, God may not elect all dying infants, even if he does so in this world.
vi) Furthermore, even if you subscribe to predestinarian universal infant salvation, that doesn't imply that more people are ultimately elect. It may simply mean a greater percentage of the elect die in infancy, and fewer in adulthood–even though the overall number is exactly the same. The sum is the same. All that's different is how the elect are distributed by time of death. Whether more die younger or older.