The Cosmological Euthyphro Dilemma (herein: CED) asks whether God’s reasons for Her creation of the universe originate within God, independently of God, or if God possesses no reasons for Her creation. As I have previously argued, each possible response is problematic both for theism and for arguments commonly used in support of theism; I launched the argument first as a response to the Argument from Contingency; second, defended it from criticism; and, third, argued that CED can be used to construct a new atheistic ontological argument. Maximus Confesses posted a number of significant challenges to my argument in the comments section to one of my previous posts on this topic, so I thought I would post my response to his argument.
Maximus explains that he denies the possibility of libertarian free-will, but does not deny that God possesses compatibilist free-will. Thus, as I understand him, he defends the possibility that God possesses some sort of compatibilist free-will. I would have said that this cannot work, because God would still behave the same way in all possible worlds and that most compatibilists would at least expect there to be true counterfactuals concerning God’s actions, but, as we will see, Maximus anticipates this move and provides an apt response.
For Maximus, free actions are those which are “rationally informed”. Maximus conceptualizes God as time-independent:
“Consider God, being an agent acting without any tense or spacial predication, actualizes all points of time from time T1 onward. There is no deliberation going on since God does not need to deliberate in order to garner more information for a more informed decision, what God is doing is necessarily actualizing space-time in a perfectly informed fashion.”
All of this seems perfectly comprehensible. God might possesses reason R1 for creating our universe and R2 for not creating our universe. Supposing that |R1|>|R2|, God would know that |R1|>|R2| and would not need to deliberate between R1 or R2. Instead, God’s timeless recognition that |R1|>|R2| would explain why God creates the universe (or, for any other action, why God would do that action). Maximus goes on to provide a helpful quote from Anselm concerning the sort of free-will that God might possess (from Anselm’s On Free Will):
There is a free will that is from itself, which is neither made nor received from another, which is of God alone; there is another made and received from God, which is found in angels and in men [sic]. That which is made or received is different in one having the rectitude which he preserves than in one lacking it. Those having it are on the one hand those who hold it separably and those who hold it inseparably. The former was the case with all the angels before the good were confirmed and the evil fell, and with all men prior to death who have this rectitude.
Here, Anselm appears to have chosen a particular horn in the dilemma: God’s reasons for Her actions necessarily originate within God (they are “neither made nor received from another”). In this sense, God’s reasons are distinct from creaturely reasons (i.e. the reasons of humans or of angels do not originate from our essences nor is our free-will originate from us intrinsically, but was provided to us by God). But given what Maximus has said about Anselm, we may wonder how it is that God acts differently in other possible worlds.
Here, Maximus suggests that we examine Buridan’s Ass. Imagine a hungry donkey, placed equidistant between two equally valuable sources of food. The donkey has just as much reason to pursue one food source as to pursue the other, so, if the donkey only acts on the basis of reasons, the donkey will starve. But our intuition is that the donkey will not starve because the donkey will choose one of the food sources without reason. Thus, it seems that free-will must involve a component of randomness. It is strange that Maximus points us to this possibility, given that Maximus has previously stated that he rejects libertarian free-will and that the actions produced by free-will are those which are “rationally informed”. Traditionally, compatibalists have been pressed to respond to Buridan’s Ass by responding that the donkey will starve; it is only the libertarian who is willing to say that free actions may be reasonless. So it appears that Maximus has jumped to the other fork in the dilemma.
Be that as it may, his response does not suitably defang CED. Maximus continues by positing that while God would not possess random reasons within Herself, She is capable of creating a “random actualizer” that would do the work for Her:
If there are two possible universes with a negligible difference, are we to suppose God irrational to choose one over the other? I should say not, it is possible for God to necessarily create a random actualizer to do the work for him. It would pick at random a possible world that he finds worthy of creation. This would make this world contingent through a necessary random chooser. Any effect it would have would be explicable by the random chooser.
Maximus’s example can be developed further. Suppose that there are multiple worlds, all of which God would find equally valuable, and that God must choose among these worlds. God has timeless recognition of God’s inability to act without reasons, so God creates a random actualizer to do the work that She cannot do Herself.
There are a number of worries that come about as a result.
First, while I don’t have a problem with the notion that God cannot do that which would be logically impossible for God to do (such as actions that contravene Her nature), we ordinarily do not suppose that God lacks some sort of power that is possessed by some creature. Yet that’s exactly what this solution would demand: God cannot act randomly, but God can create a random actualizer that does act randomly. At the very least, that poses problems for God’s sovereinty, ultimacy, and aseity.
Second, given that God could have created some other possible world (that is, which possible world God creates is contingent upon the results of the random actualizer), any evils which appear in the possible world God creates were not necessary for bringing about some greater good; whatever greater good they accomplish could have been brought about in some other way in any of the possible worlds that are equally good. Thus, any actual evil would be gratuitous; as most theistic and atheistic philosophers agree that gratuitous evil is inconsistent with God’s existence, this possibility grants us a reductio of theism.
But it gets worse. Consider the collection of all possible and impossible worlds; call this C. The random actualizer chooses among some subset (or subcollection) of this collection, with the condition that all members of the subset are equally valuable from God’s perspective; call this subset S. We may imagine that the worlds in C can be ranked in order of their value from God’s perspective. At any given ranking, there may be multiple worlds at that rank; this would explain why there are multiple worlds in S. Nonetheless, it seems plausible that there are worlds better than the world we live in (Leibniz’s statement that we inhabit “the best of all possible worlds” seems false). If so, then our world is not a member of S. But then we again have a reductio, since, if our world is not in S, our world would not be possible and would consequently not be actual. Yet our world is actual.
Furthermore, the random actualizer was meant to save the notion of metaphysical possibility. This falls into trouble, for it seems plausible that there are worlds of less value than our own which are metaphysically possible. Consider, for example, a world filled with nothing but endless pain and torment. Surely, such a hellish world is not a member of S. Yet no world which is not a member of S would be metaphysically possible. So the random actualizer does not save metaphysical possibility, at least as it is normally discussed, after all.
Maximus provides another possible way to save metaphysical possibility other than the random actualizer. Perhaps God creates all possible worlds that are of equal value, leaving us with a kind of modal realism. In that case, some notion of possibility and necessity would be preserved, at least as well as those notions are preserved in other forms of modal realism (i.e. such as David Lewis’s). However, as with the random actualizer, does not preserve the same notion of metaphysical possibility. Again, only the worlds in S would be metaphysically possible and not the full set of worlds that we would normally take to be metaphysically possible. Furthermore, this possibility does not preserve the notion that there are counterfactual truths concerning alternative ways in which God could have acted; so this possibility does not preserve God’s free-will (in either libertarian or compatibilist senses).