Causal Connections, Logical Connections, and Skeptical Theism: There Is No Logical Problem of Evil

Abstract

In this paper, I consider Sterba’s recent criticism of skeptical theism in context of his argument from evil. I show that Sterba’s criticism of skeptical theism shares an undesirable trait with all past criticisms of skeptical theism: it fails. This is largely due to his focus on causal connections and his neglect of logical connections. Because of this, his argument remains vulnerable to skeptical theism.

1. Introduction

James Sterba (2019a, 2019b) argues that known facts about evil are logically incompatible with the existence of God—an omnipotent, omniscient, and perfectly good being. Moreover, he claims that his argument is immune to criticisms stemming from skeptical theism. The project of this article is to show that his criticism of skeptical theism is unsuccessful. In essence, I will argue that Sterba focuses on the known causal connections of evils when he should be focusing on (un)known logical connections of evils, and that this is the downfall of his argument. In Section 2, I will lay out the background assumptions operating in Sterba’s argument from evil and will consider his objections to skeptical theism, arguing that his objections ultimately fail. And in Section 3, I will consider the upshot of Sterba’s argument from evil. In short, it (arguably) has an effect on those who offer a theodicy in response to arguments from evil—it is relevant to those who try to identify God’s actual reasons for allowing evil—but it has no effect on skeptical theists. Therefore, his argument remains vulnerable to skeptical theism.

2. Sterba’s Argument from Evil

James Sterba (2019a, 2019b) argues that certain known evils are impermissible for God to allow. And since God would not allow an evil that is impermissible, it follows that God does not exist. What renders these evils impermissible are certain constraints on God: there are certain conditions that must be met for God to allow evil, and these conditions, argues Sterba, are not met in some cases. These constraints play a crucial role in Sterba’s argument, since they rule out certain kinds of responses to his argument from evil. As such, I will offer a brief outline of these constraints prior to laying out Sterba’s argument.

2.1. Sterba’s Constraints

Sterba endorses the widely held outweighing constraints for God to permit evil. That is, he holds that an evil E is permissible for God to allow only if it is either (i) required for a greater good or (ii) required for the prevention of a worse evil—any evil for which (i) or (ii) does not hold is impermissible for God.1 Sterba (plausibly) thinks more than (i) or (ii) is required for an evil to be permissible for God—there is a further constraint on God’s actions beyond (i) and (ii). In particular, Sterba holds that, for an evil to be permissible for God, it must also be in line with the Pauline Principle (PP), which prohibits one from performing an action that is wrong-in-itself to bring about good consequences (Sterba 2019b, p. 177). Crucially, the PP is not absolute: there are exceptions to the PP, such as cases in which performing an action is the only way to avoid a worse evil (Sterba 2019b, p. 177). For example, consider the following case:

TORTURE: You and three friends were hiding from a would-be torturer. Since the torturer was unable to find you and your friends, she found another person to torture instead, call her Sarah. She begins torturing Sarah—say, by reading her passages from the dreaded Hs: Habermas, Heidegger, and Hegel—and you are able to jump from your hiding spot and prevent her from engaging in this torture. If you were to do so, Sarah would be set free. However, it would result in you and your three friends being tortured.

In the case of TORTURE, thinks Sterba, you may allow Sarah to be tortured since that is the only way to avoid a much worse evil (i.e., you and your three friends being tortured). So, there are some cases in which it is permissible to allow evil to avoid a worse evil—the PP does not hold absolutely. Let us say that when the PP does not hold for an action and yet the action is permissible, that the PP is violated in a justified way. With this in mind, we may say that the PP adds another constraint to God in allowing evil: (i) or (ii) must hold for an evil and they must hold in a way that either (a) does not violate the PP or (b) violates the PP in a justified wayAnd so there are (at least) three constraints on God with respect to allowing evil.2

From here, Sterba pressures the theist by focusing on particularly bad evils that occur in the final stages of life for the victim. By isolating these cases, Sterba can (in some sense) plausibly claim that certain purported greater goods are not candidates for justifying God permitting these evils. For example, in the final stages of suffering, it might be dubious to think that any character development occurs, thereby ruling out ‘soul-building’ theodicies that suggest that one reason God allows (at least the possibility of) evil is to build character. Additionally, says Sterba, it cannot be that free will explains these evils here, since the victim plausibly has a claim to having a right to not undergo such evils.

So, some common theodicies are (arguably) rendered irrelevant due to the particular evils Sterba focuses on. What adds more kindle to Sterba’s argument is that in paradigm cases in which the PP is violated in a justified way, it is due to our human limitations. For example, in TORTURE, the reason that it is permissible to violate the PP is because of human limitations: you are not able to prevent your friend from being tortured without making it such that you and your three friends are tortured. And this point can be generalized: all clear justified violations of the PP involve human limitations. Indeed, Sterba says that:

[n]one of the exceptions to the Pauline Principle that are permitted to agents, like ourselves, due to our limitations of power, would hold of God. This means that the Pauline Principle’s prohibition of intentional doing evil would be even more absolute in the case of God than it is our selves.

(Sterba 2019b, p. 177)

And so he says, in comparable situations, God would “always be able to prevent both moral evils.” (Sterba 2019b, p. 178). He infers from this that:

God, unlike ourselves, is never justified in permitting significant and even horrendous evil consequences of one immoral action so as to prevent the greater evil consequences of another immoral action.

(Sterba 2019b, p. 178, emphasis mine)

To illustrate Sterba’s point, consider again TORTURE: while you cannot prevent Sarah from being tortured (an evil) because it would result in you and your three friends being tortured (a worse evil), God—obviously enough—could save Sarah without it resulting in you and your three friends being tortured (e.g., he could whisk all five of you up to heaven). So, while you would be justified in permitting Sarah’s torture on account of this fact, God would not. Furthermore, this applies to all possible actions, according to Sterba. This thesis is crucial for Sterba’s argument from evil: if it is false that God is never justified in permitting significant evil to prevent worse evil, then—as I will discuss in the section below—Sterba’s argument is vulnerable to skeptical theism.

2.2. Sterba’s Argument Stated

We are now in a position to consider Sterba’s argument from evil. Before doing so, recall the constraints Sterba puts on God for allowing evil: for an evil to be permissible, it must either (i) result in a greater good or (ii) be required to prevent a worse evil and (iii) they (i.e., (i) or (ii)) must hold in a way that is either (a) does not violate the PP or (b) violates the PP in a justified way. Moreover, Sterba argues that all paradigm cases in which the PP is violated in a justified way are due to human limitations, and so they will not help God out. Given this, we may state Sterba’s argument as follows:

(1)

Goods that could be provided to us are of just two types. They are either goods to which we have a right or goods to which we do not have a right.3

(2)

With respect to goods to which we have a right, such as freedom from a brutal assault, God would never be causally stuck, as we sometimes are, in situations where we can only provide some with such goods by not providing others with such goods.

(3)

Since then God would be facing no causal or logical constraints with respect to providing us with such goods, God should always have provided us with such goods and thereby prevented the evils that would otherwise occur.

(4)

But this clearly has not happened because there are significant and especially horrendous evil consequences of immoral actions that God, if he exists, would have to be permitting, and this is logically incompatible with God’s existence, unless there is a justification for God’s permitting those consequences to provide us with goods to which we do not have a right.

(5)

Now with respect to such goods [i.e., goods that we don’t have a right], God would also never be causally constrained by lack of resources, as we sometimes are, and thereby be unable to provide us with such goods without permitting the significant and especially horrendous consequences of immoral actions to be inflicted on us.

(6)

Since then God would be facing no causal or logical constraints with respect to providing…us with such goods, God should always have provided us with such goods without permitting the significant and especially the horrendous consequences of immoral action to be inflicted on us.

(7)

But that clearly has not happened because there are significant and even horrendous consequences of immoral action inflicted on us which, if God exists, would have to have resulted from God’s widespread permission of just those consequences, and that is logically incompatible with God’s existence (Sterba 2019b, pp. 184–85).

What are we to make of this argument? The first thing to note here is that Sterba’s argument makes two invalid inferences: both the inference from premise (2) to (3) is invalid and the inference from premise (5) and (6) is invalid—and for the same reason. Both premise (2) and premise (5) make a claim about God not being causally constrained in a particular way, and both (3) and (6) claim that, therefore, God is not casually or logically constrained in that particular way.4 To see why this is a problem, consider a right that Sterba says we have: a right to be free from brutal assault, such as described in TORTURE. I conceded above that there is no doubt that God could prevent Sarah’s torture and the torture of you and your three friends, and this means that he is not causally constrained on this matter. However, it does not follow from this that he is not logically constrained. To illustrate this, consider the following case:

ZUES: God created a powerful creature, Zeus. And God has made an agreement with Zues that he may create a mini-world as he sees fit—God promises not to interfere with Zeus’s world.

Now, suppose that TORTURE took place within the context of ZUES. That is, suppose that Sarah’s torture is taking place within a world created by Zues—a world that God has agreed not to interfere with. Let us call this world ZUES > TORTURE. In ZUES > TORTURE, God is not causally constrained with respect to Sarah. However, he is logically constrained: while he could (causally) prevent the torture of Sarah and you and your three friends (say, by zipping you five away to heaven), his contract with Zues logically constrains him, in the sense that causally intervening and preventing Sarah’s torture logically entails violating a different constraint (namely, that of keeping his promise to Zues). And so while God is not causally constrained in ZUES > TORTURE, he is logically constrained. In other words, ZUES > TORTURE shows us that causal and logical constraints come apart—one can be free of causal constraints but not of logical constraints. (That he is logically constrained does not mean that it is impossible for him to act. Rather, it means that it is impossible for him to act without (in this case) violating a contract.5)

Or, consider a less bizarre (but equally clearly false) example:

AUTONOMY: Humans have a right to autonomy, understood as a right to act as we see fit. And this right is absolute: it’s always wrong to violate no matter what.

Now, suppose that AUTONOMY is true, and that Sami is torturing Sally. If that is the case, then while God may causally prevent Sami from torturing Sally (say, by whisking her away to heaven), he is logically constrained from doing so: he has the power to intervene, but since intervening involves violating an absolute right, he is logically constrained from preventing the torture.6 (Again, that he is logically constrained does not mean that it is impossible for him to act. Instead, it just means that (in this case) it would involve violating an absolute right.7)

Alternatively, we may understand this point axiologically. Suppose that violating someone’s autonomy is much more evil than we normally understand it. Indeed, suppose that the true disvalue of autonomy violations is so intense that violating one’s autonomy is axiologically worse that allowing someone to be tortured. If that is the case, then, again, God will not be causally constrained with respect to preventing Sarah’s torture, but he will be logically constrained, since doing so would logically entail violating Sami’s autonomy—a far more valuable good. Again, this shows that causal and logical constraints come apart. (Again, it is not that God’s causally constrained here. Instead, it just means that intervening would involve making matters axiologically worse.8)

No doubt Sterba could argue that (a) our world is not the result of this God–Zeus contract, (b) a right to autonomy is not absolute, and (c) violating Sami’s right to autonomy is not axiologically worse than allowing Sarah to be tortured; however, that is beside the point. While they are false, they suffice to illustrate the invalidity of his inference from premise (2) to (3) and from premise (5) to (6). In other words, they show that lacking causal limitations does not entail lacking logical limitations, and for this reason, Sterba’s argument is invalid. Of course, the argument can be repaired and made valid, and below I will consider how such an argument fares.

2.3. Sterba’s Argument Repaired

In order to fix Sterba’s argument, we need to find a way to show that God (at least probably) is not logically constrained with respect to certain evils. Fortunately for us, Sterba has provided a different version of his argument in his book that is not (at least obviously) invalid.9 For this version, Sterba asks us to consider the (well-known) case from Dostoevsky (1984) in which a child accidentally causes the dog of a powerful and evil General to go lame. The General locks the child up overnight, and releases a pack of dogs on the child the next morning. The dogs tear the child to pieces in front of the child’s mother. This is doubtless an instance of evil. And, Sterba uses this case for the concrete version of his argument, which he states as follows:

(8)

God’s permission of the evil consequences of the General’s action could not be a morally acceptable means to prevent some other greater evil consequences of an immoral action. This is because God, being all-powerful, could always prevent the evil consequences of any action, as needed, by just sufficiently restricting the external freedom of the evildoer in each case. Hence, this is just, I claim, what God morally should do.

(9)

Neither could God’s permission of the morally evil consequences of the General’s action be a morally justified means to secure some good to which we are not entitled. This is because the greatest good to which we are not entitled that God could morally provide us with would be a Godly opportunity for soul-making, and to make the provision of that good, and other such goods to which we do not have a right, conditional on God’s permission of significant and especially horrendous evil consequences of immoral actions, like the General’s, would lead to morally perverse incentives for us and for God as well. In addition, making the provision of a Godly opportunity for soul-making, and other such goods to which we do not have a right, conditional on the permission of significant and especially horrendous evil consequences of immoral actions, like the General’s, would not be morally justified because we do not have a right to such goods, and so clearly their provision could not be conditional on the violation of anyone’s rights, especially when there are countless other ways that these goods could be provided that are not morally objectionable. (Sterba 2019a, pp. 96–97, emphasis mine).

Where Sterba goes astray is the italicized portion of premise (8): while it may be true that God can prevent the causal evil consequences of an action, it is the logical evil (and good) consequences that matter. For example, it may be that by preventing the causal evil consequences in the case of the General and the child, that there are worse evils that logically follow from doing so. And so, Sterba needs to show that (at least probably) there are not any such logical entailments. If he can do this, then he will have the tools to repair the argument as it is originally stated in the previous section.

To show that there (at least probably) are not any (great) goods or (terrible) evils logically connected to preventing the child’s death at the hands of the General’s dogs, he must make something like the following inference:10

(10)

We recognize no evils that are logically entailed by God preventing the child’s death by dogs.

(11)

Therefore, probably, there are no such evils.11

But this is where skeptical theism becomes relevant. Skeptical theists (roughly) think that the fact that we don’t know of any good or evil logically connected to some state of affairs is not good reason to think it is likely that there is no such connection.12 And so skeptical theists will pressure Sterba to offer justification for the inference from premise (10) to (11). And Sterba is aware of this. Indeed, he considers that it might be argued that “for all we know, it could be just logically impossible for God to prevent the evil consequences of both immoral actions in such situations.” (Sterba 2019b, p. 178).13 What does Sterba make of this response? He asks:

Could there be entailment relations between such goods and permitting the consequences of other evils that would render it logically impossible for God to prevent both evil consequences? Yet notice how strange such entailment relations would be. Here we are dealing with situations where we lack the causal power to prevent the evil consequences of both immoral actions and we appeal to the lack of causal power to justify why we permit the lesser evil consumes to prevent the greater evil consequences. Now…we are imagining that it is logically impossible for God to present the consequences of both immoral actions that are just causally impossible for us to prevent. Right off, that would make God impossibly less powerful than ourselves.

(Sterba 2019b, pp. 178–79)

He concludes from this that it must be:

that God can always prevent the horrendous evil consequences of both actions in contexts where we, due to our limited causal power, can only prevent the evil consequences of one of them.14

(Sterba 2019b, p. 179)

What are we to make of Sterba’s argument against skeptical theism here? First, it is worth pointing out that the fact that if X were true, it would be strange is not strong evidence for ~X. And so, the charge of strangeness from Sterba does not seem to be doing much work. But there’s more at play in Sterba’s argument here than just strangeness. While his argument is not exactly clear here, the thought seems to be something like this. Consider again Sarah’s torture, call it T, as described in TORTURE above. If you prevent T in TORTURE, then bad consequence, call it C (you and your three friends being tortured), comes about. The reason you are morally permitted to allow T is because preventing T would result in C—a much worse evil, which makes your allowing T a justified violation of the PP. But to say that God may permit T must mean that while he causally could prevent C, he logically could not do it, and that seems to suggest that God is not nearly as powerful as we thought he is. This is because while we cannot prevent both T and C, it is due to contingent reasons. But on this view, it is logically impossible for God to prevent T and C. That seems to make God much less powerful than we’d like to think. And this is why we should not think God’s logically constrained in these cases. In effect, this has the result that skeptical theism is false, and therefore that it will not undermine Sterba’s argument.

The problem with Sterba’s argument here is that he focuses on only known causal connections, thereby neglecting possible unknown logical connections. For example, one does not need to claim that God’s permitting T is justified because if he did so, C would follow and he could not prevent it. Instead, it might be that God’s permitting T would result in a different consequence C*, and it would be logically impossible for God to prevent T without allowing C*. Consider again ZUES>TORTURE (see above, Section 2.2): if God prevents T, then it logically follows that he is violated his contract with Zeus, call this consequence Z. So if God prevents T, then a rights violation occurs, namely Z. Notice that this is not the case for you: if you were causally able to prevent T and C, Z would not follow, since you never agreed to a contract with Zeus. Indeed, God could prevent both T and C just like we could if we were more powerful. It is T and Z that he cannot prevent both of. But then the fact that God cannot prevent T and Z does not—in any serious sense of the term—make him less powerful than us; thus, this charge of strangeness melts away. The lesson here is that what logically constrains God with respect to preventing an evil does not need to be the causal consequences we know of that would result from our preventing the evil. Once we recognize this, all charges of strangeness and powerlessness melt away.

Of course, if Sterba was right that there being a logical constraint on God here would result in him being “impossibly less powerful” than us, then he would be able to hold that it is improbable that there are such constraints. But Sterba’s wrong, and so he cannot use this method to rule out God being logically constrained from preventing an evil. As far as I can tell, Sterba provides no other reason for thinking that God is not logically constrained. And this leaves us with no reason at all to think that God is not probably logically constrained with respect to preventing evil. And this means that Sterba’s argument is vulnerable to skeptical theism: he must make an inference like that from premise (10) to (11). But this is precisely the inference that skeptical theism blocks. And hence Sterba’s argument about evil is vulnerable to skeptical theism. (Here it is worth noting here that Sterba does not contest that skeptical theism, if true, undermines this type of inference. That is (presumably) why he argues that it is false rather than ineffective15).

2.4. Will Other Constraints Help?

One might want to object that the issues I raise above are only issues because I rejected several constraints that he argues in favor of. That is, in footnote 2 above, I said that Sterba holds that there are the following two additional constraints on God (or anyone): (a) an action is morally justified only if it is “reasonably acceptable” to all those affected, and (b) an action should not be permitted if “significant and especially horrendous evil consequences of immoral actions [occur] simply to provide other rational beings with goods they would morally prefer not to have.” (Sterba 2019a, p. 128). Call these moral constraints. Perhaps one would argue that these moral constraints, if accepted, would undercut the argument I gave above.

However, even if these moral constraints were accepted—even if (a) and (b) are real constraints on God (or anyone)—Sterba’s argument remains undermined. This is because we are not in a position to know whether there are evils for which (a) and (b) are not satisfied. That is, we are not in a position to know whether there is evil that some persons would not reasonably accept or evil that some persons would not morally prefer. This is because in order to know that, we need to know the reason that an evil was allowed (if there is any): whether a person would reasonably accept or morally prefer some evil will be influenced in a large part by whether God is logically constrained in the manner portrayed in the above sections (i.e., whether for some evil, there are evils or goods logically connected to the prevention of it). For example, if preventing some evil would result in an evil far worse than the one prevented, then one might reasonably accept the evil or morally prefer it.16 What’s key here is what that logical constraint is (if there is one at all)—whether one would reasonably accept an evil or morally prefer it will largely depend on the nature of the logical constraint. However, I showed above that Sterba has given us no reason to think that God is not logically constrained with respect to evil. Furthermore, he also hasn’t given us any reason to think that if there is a logical constraint, that it is one that would not result in everyone reasonably accepting or morally preferring the evil we find in the world. But this leaves us in a state of agnosticism about whether Sterba’s moral constraints (i.e., (a) and (b)) are satisfied: unless we know that either there is no such logical constraint or what the logical constraint is, we cannot know whether an evil is reasonably acceptable or morally preferable. Of course, Sterba points out that it may be difficult to imagine what these logical constraints might be. But that is not good reason to think there are not any such constraints. And hence even if Sterba makes use of these additional moral constraints on God (i.e., (a) and (b)), it will not help his argument from evil.

3. Lessons Learned

What we have learned here is that Sterba’s constraints (if accepted) will have an effect on those doing the project of theodicy: they add more obstacles (such as the PP) that those advocating theodicy must hurdle. But his constraints (and argument) just does not affect skeptical theists: since the PP admits exceptions, we still need to consider whether God permitting an evil is required to prevent a worse evil (or bring about a great good) in a way that justifiedly violates the PP, and skeptical theism comes into play here. And while Sterba raises some objections to skeptical theism, the objections don’t consider unknown logical connections—they only consider known causal connections. This is his downfall. Perhaps there are other objections in the area that threaten skeptical theism. But as things stand, Sterba’s objection to skeptical theism shares the same fate as all past objections to skeptical theism: it fails