Animal Suffering and the Laws of Nature

Abstract

Two recent atheistic arguments from evil have made much of natural evil and the suffering of animals in their case contra theism. The first argument is that of James Sterba. Sterba’s argument is an incompatibility argument premised on the claim that there are actual events logically incompatible with the existence of God. The second is that of Michael Tooley, who erects his argument in part on the claim that failing to prevent the suffering of animals cannot be justified by appeals to the great value of regular and predictable laws of nature, nor to the desirability of divine hiddenness. This article examines the arguments of Sterba and Tooley and contends that both are self-undermining. Each of the arguments employs premises that provide reason for thinking that other premises found in their arguments are false. Prior to a discussion of the two arguments, we explore the nature of incompatibility arguments, and examine three assumptions that lurk in the background of discussions of the problem of evil.

1. Introduction

Concerning the moral status of animals, Jeremy Bentham proclaimed in the late 1700s that “…the question is not, Can they reason? nor, Can they talk? but, Can they suffer?” (Bentham 1789, chap XVII, note 1). The moral elevation of sentience extended the purview of the Problem of Evil. In Voltaire’s Candide (1759), attention was focused exclusively on human suffering. Animal suffering is mentioned in Parts X and XI of Hume’s Dialogues Regarding Natural Religion (1779), but human suffering is the primary focus. Two centuries post-Hume, however, the fact of animal suffering itself fueled arguments from evil. In 1979, William Rowe developed an influential evidential argument from evil, employing a case of a fawn caught in a forest fire, ignited by lightning, as evidence contra theism (Rowe 1979). As a case of natural evil, apparently connected to no greater good, the suffering of Rowe’s fawn eluded resolution via theodicies like the Free Will Theodicy or the Soul-Making Theodicy. Natural evil is any suffering originating from natural causes. Suffering flowing out of the Black Death of the 1340s, the 1755 Lisbon earthquake, the eruption of Mount Tambora in 1815, or the Covid virus of 2020 would be cases of natural evil. Suffering resulting from predation, cancer, or other diseases, would also be cases of natural evil. Moral evil is contrasted with natural evil. Moral evil is any suffering originating from the actions, or culpable inactions, of moral agents. Natural evil and moral evil are mutually exclusive and jointly exhaustive concepts. Natural evil, then, is any suffering that is not moral evil.1 It is evil that results from non-agential processes and forces. While there may be hard cases in which it is not clear if the evil is natural or moral, it cannot be both. Two recent atheistic arguments from evil have made much of natural evil and the suffering of animals in their case contra what we might call Classical Theism. There are varieties of theism—think of J.S. Mill’s idea of a finite deity, or the movement known as Process Theism, which differ from Classical Theism. Classical Theism is the intersection of Judaism, Christianity, and Islam.2 More abstractly, Classical Theism is the proposition that there is an agent who is omnipotent, omniscient, morally perfect and the creator of the universe. In short, that God exists.3 A theist is anyone who accepts that proposition. An atheist denies it.

The first argument is that of James Sterba, who contends that “…in the case of God, it is the absence of any law-like prevention of the significant and especially horrendous consequences of natural evil in our world that is logically incompatible with God’s existence” (Sterba 2019, p. 166). Sterba’s argument is an incompatibility argument premised on the claim that there are actual events logically incompatible with the existence of God. The second is that of Michael Tooley, who erects his argument in part on the claim that failing to prevent the suffering of animals cannot be justified by appeals to the great value of regular and predictable laws of nature, nor to the desirability of divine hiddenness (Tooley 2019, p. 23).4 These arguments provide interesting variations of Philo’s second of four conditions specified in Hume’s Dialogue XI, “whence arises the misery and ill of the universe”:

…a capacity for pain would not of itself produce pain if it weren’t for something else, namely the world’s being governed by general laws; and this seems to be in no way necessary for a very perfect being. It is true that if each thing that happens were caused by an individual volition on God’s part, the course of nature would be perpetually broken, there would be no dependable regularities, and so no man could employ his reason in the conduct of life. But if some such volitions threatened to have that effect, mightn’t other particular volitions remedy this inconvenience? In short, might not God exterminate all misfortune, wherever it was to be found, and make everything all good, through judiciously placed individual volitions, and thus without any preparation or long chains of causes and effects?

(Hume 1779, p. 206)

Like Philo, Sterba and Tooley both contend that a God created world, governed by divine particular volitions, would lack the natural evil of animal suffering.

In what follows, I examine the arguments of Sterba and Tooley and contend that both are self-undermining. Each of the arguments employs premises that provide reason for thinking that other premises found in their arguments are false. Prior to a discussion of these arguments, we will explore the nature of incompatibility arguments, and examine three assumptions that often lurk in the background of discussions of the problem of evil.

2. Incompatibility Arguments

Two propositions are logically incompatible just in case they take opposite truth-values.5 An incompatibility argument is an argument contending that certain propositions are logically incompatible, perhaps contrary to appearances. Some incompatibility arguments are what we might call “internal” incompatibility arguments as they allege that a particular theory, ideology or set of doctrines is incoherent as the theory, ideology or set of doctrines implies a contradiction. An “external” incompatibility argument is any that alleges that a particular theory, ideology or set of doctrines implies a proposition incompatible with a known fact about the world. Contrasted with incompatibility arguments are what we might call improbability arguments. An improbability argument contends that a particular proposition is probably false, given the truth of certain other propositions. Improbability arguments seek to muster strong evidence against their target proposition, while incompatibility arguments seek to show that their target—a particular theory, ideology or set of doctrines—implies a contradiction. Atheistic improbability arguments seek to show that theism is likely false based on appeals to evidence. An atheistic internal incompatibility argument purports to show that theism is incoherent, as it contains contradictory propositions; while an atheistic external incompatibility argument contends that theism is logically incompatible with a known fact of the world.

To show that two propositions, A and B, are logically incompatible even if they do not appear to be, a third proposition, C, must be conjoined to A and B, such that the set of (A and B and C), implies an explicit contradiction. To demonstrate logical incompatibility, this third proposition, which we might call “the probe”, must be a necessary truth. A necessary truth is a proposition true in all possible circumstances; it is true no matter what.6 Why must the probe be a necessary truth?

A necessary truth is required as a necessary truth entails only truths. Employing a necessary truth as the probe ensures that any contradiction following from the set of (A and B and C) follows from A and B alone, since C, the probe, is a necessary truth. The role of the probe is to make it clear that a contradiction follows from the conjunction of A and B as the two are incompatible. The probe itself should play no role in generating the contradiction. Its role is to indicate the presence of a contradiction. To employ as a probe anything other than a necessary truth raises the prospect of “false positives,” giving the appearance of two propositions being incompatible when in fact they are compatible. Consider:

1. There are two animals in the doghouse.

2. All the animals in the doghouse are mammals.

Clearly, (1) and (2) could both be true, so they are logically compatible. But if we add a probe which is, let us suppose, contingently true, a problem arises:

3. There is an odd number of mammals in the doghouse.

The set consisting of (1), (2) and (3) is inconsistent. Since the probe itself should play no role in generating the contradiction, we erroneously seem to have indicted (1) and (2) of incompatibility, even though we know they are compatible. We have a false positive indicating logical incompatibility. So, not just any proposition will do as a probe in the derivation of an explicit contradiction. A necessarily true proposition is required.

Suppose it is unclear whether propositions, P and Q, are logically compatible, and we seek to determine whether they are. The standard way of demonstrating compatibility is by constructing a model. A model consists of a third proposition (or set of propositions), R, which is itself logically possible and which is consistent with both P and Q. One then conjoins R with P (or alternatively with Q). If the conjunction or set of (P and R) entails Q, then the set (P and Q and R) is consistent or possible. And if a set is consistent, then so too are all its subsets. Hence, P and Q are compatible. It is important to note that proposition R need only be possible, it need not be plausible, likely, necessary, or even true. Demonstrating logical compatibility via a model is less demanding than demonstrating logical incompatibility via a probe as the former requires only a third proposition that is logically possible, while the latter requires one that is logically necessary.

Within regard to the Problem of Evil, atheistic incompatibility arguments are typically collected under the rubric of the Logical Problem of Evil, while atheistic improbability arguments are collected under the rubric of the Evidential Problem of Evil. Theistic responses to the Logical Problem of Evil distinguish theodicies from defenses, with the latter as attempts to defeat internal incompatibility arguments for atheism by providing a model in which the suspect propositions—say:

4. God exists,

and,

5. Evil occurs,

could be true. Recall that a model need not consist of true propositions. A theodicy is often taken to mean a plausible account of moral reasons which would justify God in permitting the evils found in the world. Theodicies are more ambitious than defenses, as they aim for plausibility, while the latter aim for compatibility. In addition to defenses and theodicies there are, of course, refutations. Simply put, refutations are nothing more than objections to specific arguments. Refutations proceed along either of two broad routes. Arguing that a premise of the target argument is false, or at least, not well supported; or, arguing that the premises of the target argument do not support its conclusion. As regards responses to the problem of evil, refutations are less ambitious than a defense, as they seek only to show that a particular atheistic argument is unsound. Refutations attempt no justification for the appalling suffering we see around us, nor do they seek to demonstrate compatibility. The argument of this paper is a refutation of two atheistic arguments from evil. Theodicies and defenses are left to others.

3. Three Assumptions

In his 1958 novel, The Mackerel Plaza, a satire on mid-twentieth century liberal Protestantism, novelist Peter De Vries has the minister, Andrew Mackerel, proclaim to his flock that “it is the final proof of God’s omnipotence that he need not exist in order to save us.” Mackerel’s proclamation would be a very high view of omnipotence if it were coherent. There are two assumptions about the idea of omnipotence, which, while coherent, are suspect. These suspicious assumptions lurk often in the background of formulations of the problem of evil. While neither Sterba nor Tooley embrace these assumptions, an examination of them is relevant for our discussion. The first assumption is that:

ASP-1: it is likely that God can do X if humans have the causal power to do X.7

This is a peculiar assumption as it is obviously false. For example, humans have the causal power to fatigue themselves but a superior being of the sort worshipped by theists, an agent who is omnipotent, could not fatigue himself.8 Or again, humans have the causal power to engage in moral wrongdoing, but a morally perfect being could not.9 An omniscient being could not inadvertently bring about an unintended consequence, while humans obviously can. The relevance of (ASP-1) as regards the problem of evil is found with the idea of an outweighing good or justifying reason that plays a role in discussions of the argument from evil. If there is a justifying reason for God to allow an evil E, then either God is within his moral rights allowing E, or there is some good, that outweighs E, and requires that God allow E. Indeed, there may be suffering that is within a human’s power but not within God’s to prevent without the loss of an outweighing good (Wykstra 1984, pp. 75–76). For example, there will be evils, theists insist, that God must permit if humans are to have the opportunity themselves to prevent evil. With this in mind, we see that an improvable world could be among the best possible worlds as there could be evils which humans could prevent that God cannot prevent, but which humans fail to prevent. Additionally, if (ASP-1) were true, then every version of the soul-making theodicy will fail. We can understand soul-making as, in large part, moral development. The idea is that humans must have the opportunity and freedom to mature morally, as individuals, via the prevention or amelioration of the suffering of others. Moral maturation requires actions on the part of humans, as, for example, in the cultivation of moral virtues, such as charity, bravery or kindness. One may be disposed toward charity but without engaging in charitable actions, one will not have the virtue of being charitable. Indeed, moral maturity and human soul-making require actions that seek to provide relief to a sufferer but are costly to the moral agent. Contemporary discussion of soul-making theodicies began with John Hick in the 1960s (Hick 1966). While soul-making presumably involves more than moral maturation, I will emphasize moral development in what follows.

A second assumption of interest is that:

ASP-2: It is improbable that God’s power is ever limited.

Given our reasons for denying (ASP-1), we have good reason for denying (ASP-2) as well. Additionally, suppose it is a divine goal in creation that humans would have the opportunity to make a real difference in the amount of value in the universe, say by their choices and actions in forming their characters. If God allows the moral space necessary for the determination of one’s character, think of the soul-making theodicy, then God’s power would be limited by that divine goal.

Finally, it will facilitate our discussion if we note the falsity of a third common assumption even though it is not concerned with omnipotence:

ASP-3: Moral evil occurs only because persons engage in morally impermissible, or rationally suboptimal, decisions or actions, or negligent inactions.

This assumption is false as moral evil results not just from immoral actions and choices but can also result from the morally permissible and rationally optimal actions of persons. To see this, let us adapt a case from Robert Nozick (Nozick 1974, p. 263). Suppose we have a world with twenty-six males, named: A, B, C, D and so on down to Z. In this world there are also twenty-six females, named conveniently: A’, B’, C’, D’ and so on down to Z’. The males are ranked from A (the highest) to Z (the lowest) by their possession of traits found desirous by females, and all agree on the ranking. So, A has the greatest aggregate of those traits, B has the second most, C the third most and so on down to Z who has the least amount. The same is true of the females, as A’ has the greatest aggregate of those traits desirous to males, B’ the second most, C’ the third, and so on down to Z’ who has the least. Again, all agree on the ranking. Persons have the freedom and right in this world to choose whether to marry and with whom they marry. All wish to marry. Naturally, A is the most attractive to the females, while A’ is the most attractive to the males. Unsurprisingly, to the bitter disappointment and unhappiness of the rest of the population, A and A’ marry. Left with their second choice, B and B’ voluntarily marry; and so on down to Z and Z’. There has been no wrongdoing in this world, and no suboptimal decision, yet unhappiness, disappointment, frustration, heartbreak, and perhaps even despair, all intrinsically undesirable states, may all be present.10 So, even though there is moral evil present, contrary to (ASP-3), it is not suffering brought about by wrongdoing.

One might object that the suffering involved in our Nozick case is an artifact of a kind of scarcity—a scarcity of diverse preferences—as the preference ranking of all the males are the same, but only one will have his top preference satisfied.11 All others will not. The same is true of the females—only one of the females will have her top preference satisfied. All others will not. This scarcity results in suffering, which, according to this objection, is an instance of natural evil and not moral evil. If the preferences varied over the respective populations of males and females, rather than being uniform, there would be no suffering.

Does this objection succeed? It does not. The first thing to notice is that a scarcity can be an artifact of human decision and action rather than a natural process. In our case, there is a sufficient number of potential mates such that each person can be married. There is no lack of resources. If there is a scarcity, it is due to human choice. Second, persons are not captives to their preferences. Even if preference voluntarism, like doxastic voluntarism, is false, it does not follow that person have no control at all. Just as one can indirectly, or in a roundabout way, control at least some of her beliefs, so too one can modify, revise, and even discard a preference. Various kinds of therapies and our own experiences show that we have at least indirect control over our preferences, even if we lack direct control. Third, with at least some of our preferences, there is a close connection with reason and deliberation. Deliberating about alternatives can generate, revise or change our preferences and our rankings. Deliberation about our preferences can be efficacious. Fourth, satisfaction or not of any given preference in our Nozick case is a function of the conscious decisions of others. Agents make decisions and those decisions can impact others. Finally, notice that the size of the population is arbitrary as intrinsically undesirable states (unhappiness, disappointment, frustration, heartbreak and despair) can result as long as one person’s preference is frustrated because of a decision another makes. Unrequited love may generate unhappiness, disappointment, frustration, heartbreak, and perhaps even despair, but whether a love is reciprocated depends on the decision of another.

Seeing that the third assumption is false is important as it is too often assumed that moral evil originates exclusively from the wrongful actions of moral agents. But moral evil is not limited to wrongful actions. Moral evil and natural evil are jointly exhaustive of the possibilities and mutually exclusive. Every case of suffering then is brought about either by a moral agent or a natural process or event. In our Nozick case, unhappiness, disappointment, frustration, heartbreak and despair, all intrinsically undesirable, result from the morally permissible and rational actions of moral agents. There are no immoral or suboptimal actions bringing about the suffering. The decisions and actions of agents are not instances, in the relevant sense, of natural processes or events, so they are not generators of natural evil.12 While moral evil typically results from immoral actions, it can result from moral actions as well. Moreover, seeing that (ASP-3) is false shows us that there could be evil—intrinsically undesirable states—even in possible worlds in which no human ever engages in wrongdoing or makes a suboptimal decision. The common charge that God could and should have brought about a possible world in which no moral agent ever engages in wrongdoing is underdeveloped given the falsity of (ASP-3).

4. Constrained Intervention and Animal Suffering

Sterba’s argument is an external incompatibility argument, which contends that the suffering of animals resulting from natural events like forest fires, diseases, and earthquakes would not occur if God existed (Sterba 2019, p. 189). Sterba presents nine alleged necessary truths stipulating moral requirements of natural evil prevention common to moral agents, whether individual or collective (an individual moral agent or a political state), and whether natural or supernatural (human or divine). For our purposes, the fourth of these nine is relevant and is distilled into the following principle:

6. for any moral agent S, S is morally required to prevent all the significant or

horrendous natural evil that she can when doing so harms no human and causes

no greater harm to animals.

Although not explicit in (6), the principle should be understood as incorporating an order of application such that political states have the immediate duty of preventing significant or horrendous natural evil, and individual humans have that duty only when political states have failed to act. God would be required to act if political states and individuals have failed to prevent a particular instance of significant or horrendous suffering. This order of application ensures that individual humans could engage in soul-making actions. Sterba holds that (6) is binding on any moral agent, whether individual human, political state, or God.

There is an important qualification on (6) that we should note: Sterba holds that humans have a right to moral development and soul-making (Sterba 2019, pp. 83–84, 91).13 He also holds that denying persons a good for which they have a right is a kind of harm. God then is obligated to provide humans with soul-making opportunities. With this point in mind, we can revise (6) to read:

6’. for any moral agent S, S is morally required to prevent all the significant or

horrendous natural evil that she can when doing so harms no human and

causes no greater harm to animals and allows for the opportunity of

soul-making.

Sterba recognizes with (6’) and the other natural evil prevention moral requirements, that the creation of perverse incentives is ripe (Sterba 2019, pp. 163–64). The creation of perverse incentives is an instance of the phenomenon of unintended consequences. Unlike perverse incentives, not every unintended consequence is undesirable or negative for the agent or others, so “undesired effects are not always undesirable effects” (Merton 1936, pp. 894–904). Perverse incentives however are not just undesired and unforeseen, but, importantly, undesirable for the agent. Every unintended consequence of an agent’s actions, which is undesirable for the agent, is a case of self-sabotage. Self-sabotage occurs when one’s plans or behaviors backfire, whether foreseen or not. If God were to act every time that humans failed to prevent a case of significant or horrendous suffering, by ending the suffering, a pattern would be detectable. With a detectable pattern, an incentive for humans to forgo seeking to prevent significant or horrendous suffering would loom—why take taxing steps to bring about a certain event X if one can ensure that X obtains by simply doing nothing? If human moral maturity and soul-making are divine goals, then a constant intervention to mitigate significant or horrendous suffering whenever political states and individual humans have failed to act would create a perverse incentive, as it would incentivize persons to forego costly moral actions, thus, undercutting the presumed divine goal. Sterba attempts to defuse the threat of perverse incentives via his idea of constrained intervention (Sterba 2019, pp. 163–64).

Constrained intervention is the idea that God would prevent or mitigate significant or horrendous suffering when political states and individual humans have failed to do so. This divine intervention may be via the miraculous or via divinely engineered natural means—a well-timed fire-suppressing rainstorm, say, or the extinction of a recently mutated pathogen. It may be that some divine interventions would go unnoticed, but Sterba allows that many constrained divine interventions would be obvious—thus the threat of perverse incentives. Sterba suggests three conditions that we should therefore expect as constraints on the morally required divine interventions, to avoid the threat of perverse incentives (Sterba 2019, pp. 163–64):

CI-1: For any significant or horrendous natural evil E preventable by S, if S can

successfully prevent E and does so, then God does not intervene.

CI-2: If S seeks to prevent E but has only partial success, then God would

intervene and successfully finish what S had started.

CI-3: If S does not seek to prevent E, assuming that God will prevent E, then

God will prevent only some but not all of E’s bad outcomes.

Divine intervention then is not morally required on every occasion, but only when the created agent’s efforts fall short of preventing all the significant or horrendous suffering. But when a human yields to the temptation of letting God do all the work and thereby takes no action to prevent suffering, God would rescue only some of those threatened. The differential between (CI-2) and (CI-3) is detectable by humans, Sterba holds, and would motivate humans, or many humans, to do all they can, whether collectively or individually, to prevent or mitigate significant or horrendous natural evil. The idea of constrained intervention might be distilled as:

CI: as regards the prevention of significant or horrendous natural evil, divine action would be calibrated with human efforts as outlined in (CI-1), (CI-2) and (CI-3), to avoid any perverse incentive if God exists.

Proposition (CI) will serve as an implicit premise in Sterba’s external incompatibility argument contra Theism.

With proposition (CI) in hand, we can now reconstruct Sterba’s argument contra theism (Sterba 2019, pp. 184–89):

6’. for any moral agent S, S is morally required to prevent all the significant or

horrendous natural evil that she can when doing so harms no human and causes

no greater harm to animals and allows for the opportunity of soul-making. So,

7. Constrained intervention is morally required of God. But,

8. Constrained intervention is obviously not operative as there are many instances of significant or horrendous animal suffering that God could prevent without harming humans or other animals. So,

9. God does not exist.

Given the idea of constrained intervention, Sterba’s argument presents a formidable challenge to theistic belief.

But is the idea of constrained intervention problem-free? It is not, as there is a tension between premises (6’) and (7). Briefly put, Sterba holds that God would intervene to prevent significant or horrendous suffering but would do so in only a way that furthers the opportunity for human moral development and soul-making and avoids perverse incentives. That is, constrained intervention would be public or detectable, yet would be calibrated for human moral development and soul-making. Premise (6’) implies that there must be space for free human moral development and soul-making, but the mechanism which avoids perverse incentives, referenced in premise (7), the public detection of God’s intervention, would eliminate the space necessary for that development. In short, if (6’) is true, then (7) probably is not.

Let us develop this objection a bit. Premise (6’) implies that if God exists, then God’s prevention of significant or horrendous suffering is limited by the need to leave moral space for human moral development and human soul-making. God would be morally restrained from preventing all significant or horrendous suffering as doing so would leave no opportunity for human moral development and soul-making. But notice that the idea of constrained intervention in (7), as a way of blunting any perverse incentive, implies human recognition of the pattern flowing out of (CI-1)—(CI-3). Human agents would realize that if they seek to prevent significant or horrendous animal suffering but are unable to complete the task, God providentially or miraculously completes it. Human agents would also realize that if they fail to rescue animals from significant or horrendous suffering, preventable suffering results as God would not honor their inaction by completing a task that they should have completed. And it seems clear enough, if one can detect divine intervention, so that one knows that God intervenes, then one can reasonably infer that God exists. Sterba’s argument, then, consists of propositions that imply that God must leave moral space for human moral development and soul-making if God exists; and that God’s existence is knowable as a way of thwarting perverse incentives if God exists. A knowledge of God’s existence however would result in an evaporation of the space necessary for free moral development in much the same way that crime decreases in those areas known to be under closed-circuit TV surveillance.14 Theistic belief would be coerced and not free if the existence of God were clear, obvious or manifest for all to grasp (See, for example, (Hick 1966, pp. 255–61, 275, 318–36); (Murray 1993, pp. 27–38); and (Murray 2002, pp. 62–82); and (Jordan 2008)).

Sterba’s Just State analogy is relevant here (Sterba 2019, pp. 12, 49–69). No one would consider a state engaged in constant, pervasive and intrusive surveillance just. Any state ignoring a robust distinction between the private realm and the public, where only the latter is legitimately within the purview of governmental scrutiny, is not plausibly described as just.15 The loss of liberty alone would count against a surveillance state being a just state. Indeed, it is not just liberty that is lost in a surveillance state, but autonomy and agency are compromised if not lost altogether as well. Consider that the inmates in Bentham’s proposed panoptic ward would lack any real opportunity to weigh choices and act on decisions that they know run counter to what is expected given that they are under constant surveillance. Uncoerced moral development and soul-making is not possible in the Panopticon.

One might object that God’s existence being known threatens human moral development and soul-making only if that knowledge is accompanied by the belief that God punishes wrongdoing or rewards right doing.16 But if it were known that God would not punish wrongdoing or reward right doing, then Sterba’s constrained intervention is rescued. This objection contends that an essentially morally perfect agent could prevent natural evil if the agent engaged in constrained intervention and was morally indifferent toward punishing or rewarding human actions as appropriate.

The problem with this objection is that it forfeits the idea of moral perfection. Consider again Sterba’s Just State analogy. No one would hold that a state that treated criminal behavior the same as lawful behavior was acting justly. If a state neither punished where appropriate, nor rewarded where appropriate, the state would be morally indifferent and would not be just. Moreover, a state that did not make it known that certain behaviors were criminal and due punishment would also fall short of justice. Likewise, an essentially morally perfect agent could not be indifferent toward human behavior, such that the agent neither punishes Stalin, nor rewards Mother Teresa. If an agent S did not punish where appropriate or reward where appropriate (assuming ability), then there is good reason to doubt that S is morally perfect.

Since the idea of constrained intervention found in (7) is effectively equivalent to a divine Panopticon, then the space necessary for human moral development and soul-making required by (6’) would be lost.17 Given this, let us modify Sterba’s idea of constrained intervention so that any divine intervention to prevent significant or horrendous suffering is not a detectable matter. To see how this might work, we turn to Michael Tooley’s external incompatibility argument.

5. Animal Suffering and “God-Willing” Laws

Tooley’s argument consists of two parts, with the first part composed of thirteen premises and the second, twenty-five premises. The first part is our focus here (Tooley 2019, pp. 9–23). The goal of the first part is to show that appeals to human free will, human soul-making, the regularity of the laws of nature, and alleged desirability of divine hiddenness, are all irrelevant to the problem of animal suffering. Let us begin by outlining the first part of Tooley’s argument:

10. Animals suffer. And,

11. No animal is a moral agent. And,

12. Only moral agents can deserve to suffer. So,

13. No animal deserves to suffer. And,

14. Humans do not have souls. And,

15. No human or animal now alive is identical with any that had previously died. So,

16. Soul-making theodicies cannot explain the suffering of animals. And,

17. We have good scientific accounts of the natural mechanisms that cause natural evil. And,

18. These accounts do not involve any agent causing natural evil. So,

19. The suffering of animals is case of genuine natural evil. And,

20. The prevention of the suffering of animals would not impede human freedom. And,

21. The prevention of the natural evil of animal suffering does not require the loss of regular and predictable laws, or the hiddenness of God. So,

22. No “natural religion” theodicy or defense provides a satisfactory answer to this incompatibility argument.

Part one of Tooley’s argument is extremely ambitious as it would be sound only if several major philosophical problems have been solved. For example, proposition (14) implies that Cartesian dualism is false. Proposition (15) implies that reincarnation doctrines of afterlife are all false. Proposition (20) is curious as it would be an odd world in which humans enjoy the range of freedom that they do in the actual world yet cannot harm any animal even when they freely choose to do so.

Let us focus on proposition (21)—the claim that God could prevent animal suffering with no loss of regular and predictable laws of nature. This proposition might be seen as a philosophical codification of Philo’s assertion that:

a being who knows the secret workings of the universe might easily, by particular volitions, turn all these happenings to the good of mankind and make the whole world happy, without revealing himself in any operation. A fleet whose purposes were useful to society might always meet with a fair wind. Good rulers might enjoy sound health and long life. Persons born to power and authority might be endowed with good temperaments and virtuous dispositions. A few outcomes such as these, regularly and wisely brought about, would change the face of the world; and yet they would no more seem to disturb the course of nature or thwart human conduct than does the present arrangement of things where the causes are secret, and variable, and complex.

(Hume 1779, pp. 206–7)

Proposition (21) rests in part on premises (17) and (18). These premises assert that science generates knowledge that implies no agent causing what otherwise appears as natural evil. Put another way, (17) asserts that science gives us no reason to think that natural evil is nothing but moral evil incognito, while (18) holds that science gives us to reason to think that natural evil is not reducible to moral evil. We should not overlook Tooley’s appeal to science in (17) and (18), as those two premises entail that science and human inquiry are sufficiently competent to discover that no agent is causing what otherwise appears as purely natural forces resulting in natural evil. Put another way, we can know that the law-like regular and predictable operations observed in nature are laws of nature and not the particular volitions or actions of agents.

Tooley’s assertion in (21) that God could prevent animal suffering with no loss of regular and predictable laws is built upon his idea of a “god-willing” law (Tooley 2019, p. 18). Put simply, Tooley holds that a “god-willing” law has the form that:

Whenever a natural event of type F happens, and God does not will that it not

be followed by an event of type G, the event of type F will causally give rise to

an event of type G.

(Tooley 2019, p. 20).18

A “god-willing” law is not a natural law if by a natural law is meant a law describing only natural forces and natural causation and incorporates no theological condition. A “god-willing” law may be regular and predictable and, in those ways, mimic laws of nature, but a “god-willing” law is not a natural law, as a “god-willing” law contains a theological condition stipulating a divine volition.19 Tooley holds that human investigators very probably could neither discover nor detect that regular and predictable laws are in fact “god-willing” rather than natural laws:

If there were an omnipotent and omniscient being, all of the suffering and deaths due to natural disasters and to viruses and diseases could have been prevented if such a being so chose… No human person would ever know that this had been done unless the deity chose to communicate that fact to humans. An appeal to the claimed desirability of the hiddenness of God does nothing to block, accordingly, any well-formulated version of the argument from evil.

(Tooley 2019, p. 23)

Unlike Sterba’s discoverable divine constrained intervention, Tooley holds that humans lack the competency to detect divine volition or divine activity. In this way, Tooley seeks to argue that in a God created world, God could bring it about that both that there would be no natural evil and could have done so without sacrificing science.

Notice that Tooley’s argument proceeds by claiming both that God could, if God existed, replace natural or scientific laws with “god-willing” laws which would mimic natural laws and would prevent all cases of natural evil. And that if God existed and had implemented “god-willing” laws, humans would probably be incapable of discovering or detecting that the law-like operations are “god-willing” laws, as such laws would operate regularly and predictably and would seem like natural laws, except there would be no deadly earthquakes or viruses or forest fires or other natural mechanism resulting in suffering and death. The argument asserts in (17) and (18) that we can know via science that natural evil is real as there are no supernatural agents causing the earthquakes, viruses, forest fires and the like. So, according to (17) and (18), we can know that there are no agents via their particular volitions, wantonly or carelessly, bringing about evil that appears to us as natural evil but is in fact moral evil. Yet, the reasoning in support of (21) requires that, if God existed, then there would be no natural laws but “god-willing” laws, but we could not discover that fact. A “god-willing” law involves particular volitions of a supernatural agent and not general and regular natural laws.20 So, Tooley’s asserts both that human science can discover that the operative regular and predictable laws are in fact natural laws and are not particular divine volitions of the “god-willing” type; and that human science could not discover that regular and predictable law-like operations in nature are in fact particular volitions of the “god-willing” laws and not natural laws. But arguably if science can detect not-P, then it can detect P (for any empirical matter P). If science can tell us the polyp was not cancerous, then it could also tell us that it was cancerous if it had been. Tooley’s argument however denies this principle, as it implies that science is sufficiently competent to ensure us that the operative laws are natural and not supernatural, yet if the laws were supernatural, science would lack the competency to discover this. This asymmetry undercuts Tooley’s argument. If (21) stands then (17) and (18) fall. But if (17) and (18) stand, then (21) does not. Either way, the argument undermines itself.

6. Conclusions

So far, I have argued that the incompatibility arguments of Sterba and Tooley, built on the idea that a theistic universe would not incorporate general natural laws, but instead a regime of predictable but particular providential volitions, both fail as each is self-undermining: one part of their argument would render another part likely false.21 Of course, even if the foregoing is sound, the argument contributes nothing toward resolving the problem of natural evil.22 This has been an exercise in refutation and not theodicy. Without going into any of the detail necessary, presumably the problem of natural evil, if resolvable, would require a principled and well-argued case that the following seven propositions, or propositions relevantly similar, were true:

A. The opportunity for humans to discover, generate, and disseminate knowledge, including scientific knowledge, would be a plausible divine goal in creation, in addition to the divine goal that humans can morally develop.

A1. If the opportunity for humans to discover, generate, and disseminate scientific knowledge is a divine goal, then the world could not be a world that operates on particular divine volitions rather than regular and general natural laws.

A2. A world with regularities and laws of nature as complex as that of the actual world is not surprising if the opportunity for humans to discover, generate, and disseminate scientific knowledge is a divine goal.23

A3. Given (A), (A1) and (A2), natural evil is not surprising.

B. For any person S, and any amount of evil, e, if e seems appropriate or just right for divine purposes, such that any amount less than e or greater than e would seem insufficient or excessive, then S has a reason to neither mitigate nor prevent any evil.

B1. Given (B), natural evil that seems excessive is not surprising.

C. Given (A)—(B1), that the distribution of evil is ambiguous is not surprising.

At most, the foregoing has provided some support of (A1). The problem of natural evil remains a formidable challenge to a theistic commitment even if the arguments of Sterba and Tooley contribute little to that challenge.24