My last post examined how Daniel Johnson describes Calvinism. This post concerns the section of Johnson’s article, “Calvinism and the Problem of Evil: A Map of the Territory,” that addresses whether God can intentionally cause evil. Although I raised some objections in the previous post, those objections are not infectious for how Johnson addresses this charge against Calvinism.
The argument that there is a problem
The argument is this (my reconstruction):
- God intentionally causes others to do morally wrong actions. [Assumption of Calvinism]
- It is always wrong to intentionally cause others to do morally wrong actions. [Assumption]
- God intentionally causing others to do morally wrong actions is itself morally wrong. [1,2]
- It is impossible for God to do morally wrong actions. [Assumption]
- It is not true that God intentionally causes others to do morally wrong actions. [3, 4]
- God intentionally causes others to do morally wrong actions and God does not intentionally cause others to do morally wrong actions. [1, 5]
- Therefore, God does not intentionally cause others to do morally wrong actions. [1, 6]
I have reconstructed the argument Johnson addresses in the form of a reductio, staying true to the spirit of the argument. The reason I constructed the argument this way is that it makes evident a choice that has to be made in order to avoid the contradiction in 6. Either 1, 2, or 4 has to be rejected. All parties to the dispute accept 4. So either 1 or 2 has to be rejected. Johnson rejects 2 whereas the critic of Calvinism rejects 1.
In order to decide which choice of rejection is better, we can ask what arguments there are in favor of 1 or 2 and we can ask whether there are reasons to reject 1 or 2. The Calvinist critic rejects 1 because she accepts 2 and 1, 2, and 4 lead to a contradiction. By that reasoning, the Calvinist could reject 2 because 1, 2, and 4 lead to a contradiction. Why isn’t the choice of rejecting 1 or 2 arbitrary? The answer is that the Calvinist critic has an argument for 2. Johnson spends much of this section addressing the argument for 2. If the argument for 2 is no good and Johnson has a good argument for 1, then Johnson will have succeeded in showing that the objection to God’s intentionally causing someone to do some bad act is not morally problematic.
The argument for premise 2
Why think it is always wrong to intentionally cause another to do moral evil? The Calvinist critic’s answer, Johnson writes, is that we have a moral judgment that humans ought not do that to each other, even if the intention is for some good. In fact, I think this moral judgment would be made to any sentient beings capable of free agency, including aliens if there be any. But our moral contextual playground, God aside, is predominantly about other humans (although it also encompasses how humans treat animals and nature). And it is in this context that our judgments are formed and modified over time. The Calvinist critic then extends the relevant judgment to God: God is a sentient being capable of free agency, thus if we ought not do this to each other, then nor should God do that to us.
Although Johnson notes that this argument, if correct, would also undercut some libertarian views, he rightly in my opinion does not resign his discussion but goes on to examine the argument for premise 2. The argument for premise 2 is really an argument by analogy. It is an argument by analogy because it takes a moral context – moral judgments about other humans – finds a link that God shares – freedom – and then extends that context to God.
There are a couple ways Calvinist can respond. First, find morally significant disanalogies. There are several disanalogies that Johnson raises: God is not subject to any external law and hence cannot do wrong in causing evil; God is the creator of human beings and in a sense has ownership over them whereas other humans do not have ownership over each other; God can employ mean-end reasoning due to his omniscience that we could not employ reliably. Some Calvinists have not been satisfied with pointing out these differences and have sought to distinguish between different types of divine actions. On this approach, the trick is to find some metaphysically robust distinction between different types of divine actions such that one type applies to the good actions of humans and the other type applies to the bad actions of humans; then argue that the second type of divine action is exculpatory.
On this second approach, there’s also lots of ways one might do this. One might suggest that God allows evils by withholding his power to produce as much good. This was Edwards’ view. Alternatively, one might suggest that God’s power is exerted in different ways. One way is to sustain the world; another way is to be present in the world through acting in a special way (e.g., in becoming Incarnate; in causing a powerful wind to create a path for the exodus out of Egypt). Although God transcendently causes all human actions, he does not act as an external determining cause. Where there is only transcendent causation, God is said to allow something; if God acts as an external determining cause, God is said to do something. This is Crabtree/McCann’s view.
Johnson’s view is that these views are not tenable because there are Scriptural counterexamples to them. Matthew 6:26-32 is a counterexample to the Crabtree/McCann view. In this passage, God is said to feed the birds and clothe the grass of the field. Jesus’ point in this passage is that God is determining what is happening so that we should not be anxious. And God provides for us through the actions of other humans. This is incompatible with saying that God merely allows (i.e. acts as transcendent sustaining cause) humans to act in various ways without also acting as a determining power. Passages like Genesis 50:20 and Exodus 4:1 are problematic for Edwards’ view as well, for they show God is actively doing something that involves human agents acting in morally wrong ways.
On Johnson’s view, the distinction between doing and allowing are context-sensitive ways of highlighting the analogies between human and divine actions. They are not parts of a metaphysically robust distinction. When the Reformed speak about God doing something, they are highlighting some analogies; and when they speak about God allowing something, other analogies come into focus.
Concerning Johnson’s Response
In my view, Johnson is right that the Crabtree/McCann view is not strong enough to account for the passages noted. One of the payoffs of their view, were it correct, is that it would be much harder to argue God is blameworthy because God only has a sustaining causal role. But once they allow that God is an external determining cause of human actions, the motivation for doing and allowing as they describe that distinction dissipates. The Edwardsian view, recall, was that the distinction between allowing and doing is metaphysically robust. The problem for the Edwardsian view was that God is said to provide for others, and that can come through God doing things which entail that humans do bad things. So this distinction, it is thought, does not do the work it needs to in the Joseph story. The difference between Johnson and me, I think, is that Johnson denies there’s a metaphysically robust difference and I do not. Or rather, I might not deny that difference depending on what counts as a metaphysically robust difference between types of actions.
I agree with Johnson that ascriptions of allowing and doing to God are contextually sensitive ways to highlighting the analogies. This is a nice way of putting the matter that I had not previously considered. To say God allows or permits something is to highlight that God’s relationship to the event is one whereby God is not blameworthy; he didn’t do something that would morally impugn him. So I agree that this distinction is significant in the context of moral judgments and we probably don’t need to follow John Calvin in tossing out the distinction. I also agree with Johnson that the Edwardsian view cannot play the role as outlined above. After all, God is doing something in the case of Joseph’s abduction according to Joseph.
Where I disagree, I think, is about whether there is a metaphysically robust distinction that Edwards was after that can do some important work. In my view, there is. If we consider some of the literature on manipulation (e.g., see work by Al Mele, Patrick Todd, Michael McKenna, Derk Pereboom, among others), one of the more popular objections to compatibilism is that we would not be blameworthy for doing A were we manipulated by, say, some goddess Diana, to do A. Without going into too many details, a “hardline” response is to say that we might be blameworthy, but it depends on how we were manipulated.
Applying this to the topic of God allowing or doing something, you might think that God does causally different things when causing agents to do good things or bad things. For example, suppose that when God intentionally causes someone to perform the right action, he is repairing, or making more whole, some part of the agent (e.g., morally fine-tuning the reasons-responsive mechanism) that leads the agent to do the action on her own. When God causally intends the agent do some bad thing, he refrains from providing as much common grace and the agent does something bad on her own. I’m not providing a general theory here. And I am not suggesting that we should accept a theory of allowing where God is not performing some actions himself. 1 Rather, I am suggesting that Johnson’s contextual view is compatible with the view that there is some metaphysically robust difference in how God interacts with humans in causing them to do good actions or bad actions. Or if calling this a “metaphysically robust distinction” is not preferable, at least what God is doing to these agents or how God is causally interacting with them differs depending on the type of action the agent performs.
Holding onto this difference might play two different roles. First, it might be thought to help ward off the criticism that God is acting in a blameworthy way. But I doubt it. For any action God does, whether that is doing something or allowing, if it it thought to necessitate some human action, then so long as one can apply the moral judgment under discussion, it will not be convincing. A second way in which this difference between how God acts in causing good human actions versus bad human actions might be thought to be useful is in allowing humans to be praiseworthy or blameworthy. And here I think we can get some mileage, for it can be part of the story about how humans can act on their own even while caused, and we can retain the ascription of moral responsibility. Johnson wasn’t discussing this second issue though.
Is Johnson’s response to the argument for the second premise adequate?
I think so. The differences between God and humans should cause us to doubt that we can so easily extend our moral judgments to God without considering special revelation. Just forget about Calvinism for a moment. In a position of power, it would be wrong for me to command someone to sacrifice one’s son on a mountain. It is hard for me to think of cases where another human being would be justified in making this command. If I didn’t already have the story of Abraham and I were to extend moral judgments as easily to God, I might have objected that a perfectly good God could not do this, as Richard Dawkins has. But because I already have that story, I do not as easily make the judgment that a perfectly good God could not make that command. 2 What we do is tell Dawkins that our moral judgments when applied to God, unaided by special revelation, require caution. Then we go on to give an explanation of why God might do this in the case of Abraham with the knowledge that God did do it (that’s the faith seeking understanding approach at work here).
Now come back again to the Calvinism issue. There is a general moral judgment applicable to humans, and then it gets applied to God. Humans cannot intentionally cause others to do harm; therefore God cannot. Why is it fair game to tell Dawkins to use caution, but then suggest that for whatever passages like Genesis 50 mean, they can’t mean what the Calvinist says they do?3 Obviously, it isn’t fair. If someone has convinced himself that it is pretheoretically more clear that God cannot intentionally cause another human to do an evil than that God can command someone to sacrifice one’s own son, that person is in the grip of a theory for which no evidence to the contrary is likely to be adequate.
Note by the way that this is not the skeptical position that we should never make moral judgments about God on the basis of our common morality. Rather, it is just the view that our moral judgments applied to God are defeasible. And that defeasibility based on how different God is sufficient to question the argument for premise 2.
But also notice that in responding to the argument for premise 2, Johnson provided arguments for premise 1 of the original argument. So if 1 is true, then 5 in the original argument is false. But 5 depends on 3 and 4. 4 is not open to dispute here; and 3 depends on 1 and 2 in the original argument. So given that 1 is true (supported by passages like Genesis 50), premise 2 – the claim that it is always wrong to intentionally cause others to do morally wrong actions – has to go.
- You might think that omissions are events, in which case you omit to A just in case you do B. ↩
- I still might want to explain how a perfectly good God could make that command. ↩
- In fact, that is the common response I’ve received from Calvinist critics to passages like that, who then go on to say they are not sure how to understand those passages. ↩