William Rowe, in Can God be Free? (2004), gives us three propositions
A) There necessarily exists an essentially omnipotent, essentially omniscient, essentially perfectly good being who has created a world.
B) If an omniscient being creates a world when there is a better world that it could have created, then it is possible that there exists a being morally better than it.
C) For any creatable world there is a better creatable world. (pg. 120)
If B and C are true, then A is false. We don’t want that. C looks plausible to me so I will not dispute it. I know there are those who would have no problem of denying C and argue that there is a best possible world but God does not have to create that one. Again, this looks intuitively false to me. It seems to me that if there is a best possible world, God, if He were to create, should choose that one. But if this is true, then if He were to create, He would have no freedom to choose any other world. God is not free. Now, I do think we can avoid C by arguing that there is a best possible set of worlds and that set consists of an infinite number of worlds. So God is free to choose from that set. That seems to get rid of that problem.
But let’s suppose that C is true. Is B true? Rowe says,
For suppose a being selects a world W1 to create when there is a better world W2 it could have created instead. Surely it is logically possible that there be a being whose degree of moral goodness is such that when confronted with worlds W1 and W2, either of which it has the power to create, it will choose to create W2, the better world. And this would then be a better being than the being whose degree of goodness permitted it to select the less good world to create when it could have easily created the better world. (pg. 112)
It’s a good argument but I think we can reject Rowe’s principle. Arntzenius, Elga, and Hawthorne, “Bayesianism, Infinite Decisions, and Binding” (2003), argued that when we are dealing with a finite case, we are to pick the most dominant option. For example, if I have a choice of finitely many outfits to impress an honorable person, the rational thing to do is to choose the best outfit to wear. However, this does not work when it comes to infinite options. They argued,
Whenever one has no best option, there is no univocal answer to the question, “What should I do?” Suppose, for instance, that God offers to let you live any finite time of your choosing. Assuming that your utility is an increasing bounded function of the length of your life, there is no answer to the question: what life-span should you choose? Where there is a lowest upper bound on your utility, one could perhaps give useful vague guideline: pick a large number. By picking a large number, you can come very close to the lowest upper bound on your utility. So you should pick a very large number…In addition to the normative issue, there is something of a motivational puzzle here. What exactly would cause you to ask for one lifespan rather than another? But this puzzle is nothing new. We are already used to the idea that, pace Buridan, an ass confronted with equally attractive bales of hay will go to one of them rather than die of indecision. (16-17)
Now, suppose we take A/E/H’s example of Satan cutting an apple into infinitely many pieces labeled by natural numbers. Eve may take whatever piece she wants. If she takes a finite amount of pieces, then she does not suffer. If she takes infinitely many of the pieces, then she is expelled from the garden. Her first priority is to stay in the garden and her second is to eat as many pieces as she can. Satan reasons she should eat the first apple because if she takes the first one, it is a finite number and she will not be thrown out of the garden. Even if she takes an infinite amount, she will still enjoy eating the pieces so she should definitely take apply #1. But Satan reasons the same way for #2, #3, ad infinitum. Of course if she accepts them all, she will be thrown out of the garden. Yet, as A/H/E have argued, the rational thing for Eve to do is to take a very large finite number of pieces.
Now, suppose in W5 Eve is the most rational human person in that world and the most rational she can get. In W5, she is put in the situation with Satan. She picks 4340 pieces although she could have picked 4341. Is she at fault? It seems that she is not. Could she be more rational if she picked 4341? Again, no. Is it possible that there is a more rational person than her? No. It seems that we cannot judge her degree of rationality by simply seeing how many pieces they picked. At the very least, we cannot judge whether there is a person more rational than her by simply looking at the choice she made. The reader can see how this can be applied to Rowe’s argument. Rowe argued that the Expressive Thesis, the goodness of an agent’s actions is expressive of the agent’s goodness, is related to B (pg. 100). However, because there are an infinite number of possible worlds, this gives us the ability to see that the expressive thesis cannot be applied to God creating a world. The expressive thesis, like the dominant option theory, might be applicable to finite choices, but not necessarily to infinite choices. For example, Billy and Sally see that there are an infinite number of people drowning. There is a machine where they can press a number and an angel would save them from drowning. The machine can only accept a finite number. Billy pressed 8245 and Sally 8643. But Billy is very much like St. Francis of Assisi and Sally is a known murderer. Here we see that Sally is not morally better than Billy because she saved more people. We cannot reduce moral goodness by the action of the person. So, if God creates W456 and He could have created a better world, He is not at fault for creating W456.