The Book of Job is famous for its discussion of the problem of evil. But it raises another problem for classical theism, which has been under-appreciated: the problem of divine domination.Footnote 1 The problem, in short, is that if classical theism is true, the divine-human relationship is dominative; dominative relationships are unjust; therefore, if classical theism is true, the divine-human relationship is unjust.
Domination is a type of power over another. It is typically understood as a power possessed by a stronger party, which the weaker party is unable to control or resist. Because the weaker party cannot control or resist the stronger party’s power, the weaker party is at the stronger party’s mercy. The weaker party, according to many theories of domination, suffers a kind of unfreedom. The stronger does what he can; the weaker suffers what he must.
One attractive feature of domination theories is that they capture the common intuition that benevolent mastery is unjust. Slaveholding is a paradigmatic form of domination; the slave has little to no power to control or resist his master and so is at his master’s will. In other words, the master can interfere arbitrarily with his slave. This interference is arbitrary because it is controlled almost entirely by the master’s will. This seems to be part of the injustice of slaveholding. Theories of domination explain why though a benevolent master is preferable to a cruel one, even benevolent mastery is unjust. Even the slave with a benevolent master suffers a morally significant form of unfreedom for being completely at his master’s mercy.
Other theories of freedom do not seem to capture this intuition quite as well. For example, a liberal theory according to which one is unfree only if another actually interferes in his choices does not identify something unjust about benevolent, non-interfering, mastery. Neither does a perfectionist theory, according to which freedom consists in choosing the good, or virtue. Slaves can choose the good and develop virtue; many have.
According to classical theism, God is essentially all-powerful, sovereign, personal, omnipotent, and a se (independent from anything else). If such a being exists, he dominates humans in virtue of these properties. This is because he is in a position to treat humans however he pleases. Humans are completely at his mercy, unable to control or resist his exercises of power. He is akin to a master, even if he is a perfectly benevolent one. Divine domination may also be a problem for non-classical theisms that posit a very powerful God. However, because the problem is clearest for classical theism, I will focus on classical theism in this article. I will have more to say about domination and the divine-human relationship in the second section.
This article uses the Book of Job to introduce the problem of divine domination (the first section), formalizes and describes it in further detail (the second section), addresses unsuccessful responses to it (the third section), then concludes by suggesting what a more promising response to the problem might require (the final section).
Job’s complaints
Job, our protagonist, is virtuous: ‘blameless and upright’ (Job 1:1).Footnote 2 He’s well-off: ten happy children and great wealth (Job 1:2–3). He is pious, performing sacrifices for himself and his children, just in case they ‘have sinned and cursed God in their hearts’ (Job 1:5).Footnote 3 And he enjoys a stellar reputation.Footnote 4 All is well with him.
But his happiness is interrupted by a bet God makes with Satan, of which Job remains unaware throughout the book (Job 1:6–12).Footnote 5 To test Job’s righteousness, God permits Satan to kill Job’s children, destroy his property, and smite him with painful blisters all over his body (Job 1:13–19; 2:1–7).Footnote 6 Job discerns that God is causing his suffering. Still, and notwithstanding his wife’s advice to ‘curse God and die’, Job refuses to even accuse God of wrongdoing (Job 1:20–22; 2:9–10).
Job’s friends, Eliphaz, Bildad, and Zophar, learn of his misfortune. After mourning with him in silence for a week, they discuss his plight. The bulk of the book consists of their conversation. Anticipating punishment theodicies, Eliphaz supposes in Job 4 that God must be punishing Job’s wickedness. Bildad echoes this in Job 8.Footnote 7 Anticipating defeat of evil theodicies, Bildad predicts that if Job repents, God will restore his losses. His suffering will be overcome (Job 8:1–7, 20–22).Footnote 8 Anticipating sceptical theism, Zophar reminds Job that God’s reasons are inscrutable – a claim God later affirms (Job 11:7–9; 38–39).Footnote 9 God also suggests that he is not required to prevent creaturely evils (Job 38–40:2, 40:6–41).Footnote 10
However, in chapter 9, Job raises a problem that no one in the book addresses. His response to the punishment theodicy merits quoting at length:

Job had no opportunity to defend himself against the charges; he didn’t even know he was on trial. He learns about his indictment, trial, and sentencing only while serving his punishment! He can’t defeat his prosecutor in court, because his prosecutor is also his judge: ‘I must appeal to my accuser for my right’ (Job 9:15).Footnote 11
Accused of a crime, the prosecutor his judge, sentenced and punished before getting to defend himself: no wonder Job is aggrieved. This complaint seems to be that he is dominated in this legal system. One fundamental check against domination in a legal system is the right to a fair trial (Arena Reference Arena2012, 50–70; Crummett Reference Crummett2020). This affords the accused a fair opportunity to defend themself against their accuser. The accuser’s interests are represented by a prosecutor. It’s natural to think that fairness precludes the same person from being both prosecutor and judge. If they are, then the judge is highly unlikely to be impartial. Their role of prosecutor would give them extra reasons to favour the accuser over the defendant. And the prosecutor would be too powerful, since as judge they would be able to unilaterally rule in favour of their client against the defendant. Still, one needn’t accept this characterization; I’m merely trying to represent Job’s complaints. Job cannot defend himself and his judge is also his prosecutor. He’s not getting a fair trial (Seow Reference Seow2013, 547). He therefore lacks a meaningful way of contending with or resisting the power of his accuser. In his view, he’s dominated.
Job complains that God’s properties exacerbate his domination. Suppose that Job lived in a society where the same human was both prosecutor and judge. The conflation of these offices might make this society’s legal system dominative. Still, Job could theoretically fight or flee from the prosecutor-judge. By contrast, God’s rule is irresistible. God is all-powerful; no human has successfully contested with him (Job 9:4–9). Indeed, no one could even stop him to demand an account from him (Job 9:11–12, 32). He is sovereign; there is no other authority to adjudicate complaints against him (Job 9:15, 33). God’s nature and position preclude Job from defending himself or appealing for help.
Job holds God responsible for Job’s suffering while granting that there may be more proximate causes (Crenshaw Reference Crenshaw, Barton and Muddiman2001, 338–339). Since God is sovereign, nothing happens but by his leave. If ‘the earth is given into the hand of the wicked’, it must be by God’s licence: ‘If it is not he, who then is it?’ (Job 9:24).
Job may also take God’s goodness to make the problem worse, not better. He says:
19 If it is a matter of justice, who can summon him?
20 Though I am innocent, my own mouth would condemn me;
though I am blameless, he would prove me perverse.
One common interpretation of this passage is that Job is accusing God of spuriously seeking a guilty verdict against him (Pope Reference Pope1965, 73; Habel Reference Habel1975, 53; Habel Reference Habel1985, 193). However, this conflicts with Job’s insistence on God’s righteousness in chapters 1 and 2.Footnote 12 Another possibility is that God’s goodness worsens Job’s problem.
Though this might initially seem counterintuitive, there are at least three ways it might be true. First, someone who takes God to be perfectly good cannot in good conscience accuse him of injustice: a perfectly good God is incapable of injustice. Contending that God’s punishment is unjust might violate Job’s conscience. Insofar as righteous people (as Job is) tend to strongly prefer not to violate their own consciences, God’s perfect goodness might chill Job’s protest. And insofar as protest is typically one means of resisting a powerful would-be oppressor, chilling Job’s protest weakens one of Job’s ways of mitigating his domination.
Second, Job hopes to demonstrate his innocence. However, God’s all-surpassing goodness undermines human claims of innocence. Compared to God, none is righteous. Job anticipates this problem, asking, ‘How can a mortal be just before God?’ (Job 9:2). And Bildad later worries: ‘How can a mortal be righteous before God? How can one born of woman be pure?’ (Job 25:4. See also Job 35).
This is one of the interpretive challenges the book poses, a paradox it invites us to inhabit.Footnote 13 On the one hand, Job is repeatedly described as righteous; on the other, he and others suggest that humans can be neither just nor righteous, at least before God. I don’t claim to have a way to resolve the tensions these passages pose. I only mean that some of them seem to license our locating this problem in them.
Of course, even if Job is in fact not just, righteous, or pure, this doesn’t mean that Job has done anything to warrant punishment. But this brings us to the third way that God’s goodness may make things worse for Job: if Job’s standing to contend with God is predicated on his innocence, then his complaint is self-undermining. Complaining costs him his innocence and therefore his standing to complain. His ‘own mouth’ condemns him. Job’s friends seem to confirm this concern as well, claiming that Job’s complaints somehow retroactively justify his punishments (e.g., Job 22; 35).Footnote 14 And some translations of the Hebrew Bible even present Job as conceding this: ‘my complaint is rebellious’, he says in Job 23:1, according to the Syriac, Vulgate, and Targum.
Job’s complaints introduce the problem of divine domination. They also present two lessons about its relationship to the problem of evil. First, both problems are rooted in God’s attributes. God’s power and position as sovereign, Job seems to say, are sufficient to render Job dominated. In the following section I will argue that Job is right to raise these concerns, especially once we account for God’s personality, omnipresence, and aseity. Second, the problem of evil makes the problem of divine domination salient. Job’s suffering occasions his complaint about God’s domination. I will say more about both lessons in what follows.
I will also highlight two connections between the problems that the book of Job doesn’t make explicit. By the end of this article, I hope to have shown (1) that divine domination could exist independently of God actually causing human suffering (viz., God would dominate us even if God were a perfectly benevolent master), and (2) that some familiar responses to the problem of evil exacerbate the problem of divine domination. Many philosophers of religion believe that sceptical theism and greater good theodicies provide adequate responses to the problem of evil. However, if my argument succeeds, the more satisfactory one finds these two responses, the more troubled one should be by the problem of divine domination.
Divine domination
Job complains that the legal framework his friends have proposed to explain God’s actions is unjust because it is dominative. However, we can generalize his objection, removed from the legal context in which he presents it:
Premise 1 | The divine-human relationship is dominative;
Premise 2 | Dominative relationships are unjust;
Conclusion | Therefore, the divine-human relationship is unjust.
Job complains that he cannot meaningfully contend with God. This is a core feature of dominative relationships. And Job objects that this is a deficiency of his relationship with God. He is identifying something similar to what theorists of domination mean in calling dominative relationships unjust. My genre and terms are different from Job’s, but my reformulation is, I think, faithful to the spirit of his complaint. God’s domination of humans is presumptively unjust and therefore requires justification.
Assumptions
I pose divine domination as a problem for classical theism, according to which God is essentially all-powerful, omnipresent, a se, sovereign over the universe, and personal (Feser Reference Feser, Fuqua and Koons2023; Lebens Reference Lebens, Fuqua and Koons2023; Koons Reference Koons, Fuqua and Koons2023; Leftow Reference Leftow1998; Stump Reference Stump2013; Wainwright Reference Wainwright, Taliaferro and Meister2010; Williams Reference Williams, Diller and Kasher2012). Most classical theists agree that each of these is a divine property, though God’s personality is somewhat contested (Plantinga Reference Plantinga1980, 46–47; Feser Reference Feser2017, 189–199; Stump Reference Stump, Hazony and Johnson2018; Page Reference Page2019; Spencer Reference Spencer, Fuqua and Koons2023. The impression one gets is that most classical theists either endorse divine personality or would, if they could). I’ll argue that if classical theism is true, God dominates humans in virtue of his essential properties.
While divine domination may also be a problem for non-classical theisms, it is beyond the scope of this article to investigate this in detail. Divine domination is, I will argue, at least a problem for classical theism. It may also be more of a problem for non-classical theisms that are more like classical theism than for more distant non-classical theisms. One might, for instance, deny that God is very powerful. Or one might deny that God has a will. This might circumvent the divine domination problem, as it’s common to think that only an entity with a will can dominate.
I’ll also assume that domination is unjust. Theorists of domination agree on this, even while disagreeing with one another about precisely what makes domination unjust (McCammon Reference McCammon and Zalta2018). Indeed, a common strategy for disputing some account of domination is to argue that the account implies that some benign relationship is dominative (e.g., Kramer Reference Kramer2003, 135–143; Friedman Reference Friedman, Laborde and Maynor2008; McCammon Reference McCammon2015; Shapiro Reference Shapiro2016, 20–24; Simpson Reference Simpson2017).
We do not need a detailed account of what makes domination unjust to recognize that it is unjust and to have a feel for why it is unjust. Nearly all theories of domination agree that slavery is the paradigmatic form of domination, and that slavery is unjust in virtue of being dominative. Therefore, if I can convince you that under classical theism the divine-human relationship involves the features that make slavery dominative, then as long as you grant that slavery is unjust because it is dominative, you should also be able to accept that the divine-human relationship is unjust.
The divine-human relationship is dominative
Domination has as many theories as it has theorists. No one account would satisfy everyone, and spatial limits prevent me from testing the divine-human relationship against every existing formulation of domination. My discussion can’t be comprehensive. So, we will have to get at the problem in a different way.
I propose to first test the divine-human relationship against Phillip Pettit’s account of domination. I choose Pettit’s because it is the canonical account in the literature. To be clear, I do not mean for my argument to hinge on the veracity of Pettit’s; my aims are more ecumenical. Instead, I think that Pettit provides a helpful way to give a feel for what domination is, why it is unjust, and why the divine-human relationship is dominative. After using Pettit’s account to give a feel for the problem, I’ll pose the issue in more general terms at the end of this section.
According to Pettit, someone dominates another ‘to the extent that (1) they have the capacity to interfere (2) with impunity and at will (3) in certain choices that the other is in a position to make’ (Pettit Reference Pettit1996, 578. See also Pettit Reference Pettit1997, 52). Power to interfere with impunity and at will is arbitrary:
The only brake on the interference that they can inflict is the brake of their own untrammelled choice or their own unchecked judgment, their own arbitrium: ultimately, as it may be, their own capricious will (Pettit Reference Pettit1996, 580).
Another way Pettit puts it is that ‘the dominating agent is not forced to track the avowed interests of the individual interfered with’ (Pettit Reference Pettit2005, 93).
Pettit’s definition captures the intuition that benevolent mastery is dominative. It is the very capacity to interfere at will that makes relationships dominative. Benevolent masters might never actually interfere with their slaves or might only interfere in ways that benefit their slaves. Nonetheless, a benevolent master dominates his slaves because he is in a position to treat his slaves as he pleases. His slaves could do nothing to block his interference, should he decide to interfere against their interests.
By the lights of Pettit’s definition, God dominates humans. God is all-powerful.Footnote 15 If anyone can interfere with humans’ choices, he can. The vast gap in power between God and humans means that no human could force God to track that human’s interests or block God’s interference. God’s power to kill Job’s family and destroy his property surely constitutes a capacity to interfere in Job’s life and decisions. Job is powerless to stop him. Even if God only ever interfered in ways that benefitted those with whom he interfered, this would be solely a matter of divine discretion: God can interfere with humans ‘with impunity and at will’.
Furthermore, according to classical theism, God is the supreme ruler of the universe. As Job complains, there is no third party to whom humans can appeal to force God to repair the harm they have suffered or to restrain God from harming humans (Job 9:15, 32–33).
Insofar as God is the all-powerful, supreme ruler of the universe, his capacity to interfere with humans is and must be unconstrained by anything outside his will. Since God can interfere with humans with impunity and at will, he dominates humans. This domination is far more comprehensive than any imaginable case of inter-human domination.
Pettit also identifies several markers of dominative relationships. One is that dominated parties cannot meaningfully contest with their dominators (Pettit Reference Pettit1997, 63). Another is what Pettit calls ‘common knowledge’ (Pettit Reference Pettit1997, 58–59). Everyone paying sufficiently close attention to the relationship, including its parties, will recognize it as dominative (Pettit Reference Pettit1997, 59–61). A third is that the dominated party is incentivized to ‘toady and fawn, bow and scrape, placate and ingratiate – in a word, abase themselves’ to remain in the dominator’s good graces (Pettit Reference Pettit1997, 61). By contrast, undominated people ‘can look others in the eye without reasons for fear or deference that a power of interference might inspire’ (Pettit Reference Pettit2012, 84). Let us call these markers of dominative relationships the contestability test, the common knowledge test, and (using Pettit’s parlance), the eyeball test (Pettit Reference Pettit2012, 84).
Applying Pettit’s tests to the divine-human relationship strengthens our reasons to think it is dominative. Consider the contestability test. Job uses legal imagery to express the impossibility of contesting with God: God is prosecutor and judge. No trial can be fair when the same person plays both roles. But Job also offers a more general case: the vast gap between God’s power and humans’ makes meaningful contestation impossible (Job 9:4–7).
Next, consider the common knowledge test. In a dominative relationship, both parties will know who dominates and who is dominated. Throughout the book of Job, God calls Job his slave (י֣עַבְדִּ) (e.g., Job 1:8; 2:3; 42:7–8). Job also compares himself to a slave (albeit rhetorically) (Job 7:2). Jewish, Christian, and Islamic Scriptures frequently describe humans as slaves of God, sometimes in the mouth of God; other times in the mouths of humans.Footnote 16 So do many prominent Jewish, Christian, and Muslim theologians (Byron Reference Byron2003; El-Sharif Reference El-Sharif2012; Hezser Reference Hezser2024; Sheikh Reference Sheikh2019). The salience of given names like Abdullah (Arabic for slave of Allah), Obadiah (Hebrew: slave of Yahweh), and Theodoulos (Greek: slave of God) in these traditions further suggests acknowledgment that some humans are slaves of God.Footnote 17
Various theistic traditions depict God and many of his worshippers describing humans as God’s slaves. The master-slave relationship is a paradigmatic case of domination (Wall Reference Wall2001; Lovett Reference Lovett2010, 34–54; Gourevitch Reference Gourevitch2014; McCammon Reference McCammon2015, 1030ff., Reference McCammon and Zalta2018; Hasan Reference Hasan2021). The widespread characterization of God’s rule as mastery suggests a recognition that it is dominative. The common knowledge test invites us to take theists seriously when they describe themselves and others as slaves of God. Maybe theists are mistaken, and humans are not slaves of God. However, if classical theism is true, God’s worshippers might be best positioned to know what it is like to relate to him. Given this assumption, the ubiquity of the slave description among theists is notable because it lines up so neatly with what theists would say if God did dominate humans.
Finally, the eyeball test. Job complains that though innocent, he cannot look God in the eye; he can only implore God’s mercy (Job 9:15). Augustine provides an even more vivid example of the eyeball test (though if you didn’t know better, you might have attributed what follows to Schopenhauer or Nietzsche):
[Humanity] was reduced to … infirmity and subjection to decay, which left it as weak as a spider when it was cast out of paradise. That was when God’s slave was ordered to undergo a beating. Consider when our whipping began … Adam has endured a whipping in all those who have been born since the dawn of the human race; Adam is whipped in all of us who are alive today; and his whipping will continue in all who come after us. Adam is the human race under the whip, and many have so hardened themselves that they do not even feel their lacerations … All your life on earth is your beating. Mourn then, as long as you live … whether you are enjoying good fortune or are beset by troubles; cry to God … Cry out to the hand that beats you, ‘Have mercy on us, O Lord, have mercy!’ Is not this the plea of someone being whipped: ‘Have mercy on us, O Lord, have mercy?’ (Augustine Reference Ramsey and Boulding2004, 36–37).Footnote 18
I mention Augustine not because he is exceptional among theists, but because he is not. Like many theists, he describes humans as God’s slaves. He imagines God as holding the power of the whip over humans. For Augustine as for us, the whip symbolized a master’s power. He deploys this symbol liberally, describing God as violently punishing humans. The only recourse for humans is to bow and scrape, beg and plead, appeal to God’s merciful will.
Pettit’s account of domination implies that God dominates humans. God’s power and sovereignty give us pro tanto reasons to think he is in a position to interfere with humans at will. The contestability, common knowledge, and eyeball tests confirm the hypothesis.
I’ve used Pettit’s account to give a feel for the case that God dominates humans. Now we can step back and make the point more generally. Nearly all theorists of domination – including Pettit’s critics – agree that mastery is a paradigmatic form of domination. And they agree that this is so because mastery involves vast imbalances of power, near-total control of the master over the slave, and few meaningful checks on the master’s power (Pettit Reference Pettit1996; Lovett Reference Lovett2001; Richardson Reference Richardson2002, 29–32; Laborde and Maynor Reference Laborde, Maynor, Laborde and Maynor2008; Shapiro Reference Shapiro2012, 307–311; Gourevitch Reference Gourevitch2014; Blunt Reference Blunt2015; McCammon Reference McCammon2015, Reference McCammon and Zalta2018; Gädeke Reference Gädeke2020; Hasan Reference Hasan2021).
The divine-human relationship amplifies the dominative features of slavery. The gap in power between God and humans is immeasurably greater than that between a master and his slaves. And God’s control over humans is more comprehensive than any human master’s.
Theorists of domination argue that checks on relationships that limit how the empowered party can treat the disempowered party can help guarantee that the relationship is non-dominative. Two especially important checks are exit (Lovett Reference Lovett2010, 38–52; Pansardi Reference Pansardi2013; Taylor Reference Taylor2017; Drugge Reference Drugge2021) and contestation (Pettit Reference Pettit1997, 185–201; Laborde Reference Laborde2008, 149–172; Farrell Reference Farrell2020). A’s capacity to interfere with B is limited if B can leave the relationship or meaningfully contest with A. Contestation comes in many forms, including demands for accountability, appeals to higher authority, protest, reform, resistance, and revolution.
When all goes well, these checks give B ways of controlling A’s interference. They enable those who would otherwise be dominated to block would-be dominators’ attempts to interfere. Slaves might revolt and overthrow their masters, as they did in Haiti between 1791 and 1803. Reformers might change the laws to make slavery illegal, as they did in Great Britain in 1834. People might flee the slave society, as did tens of thousands of Underground Railroad ‘riders’ in the eighteenth- and nineteenth-century United States.
However, none of these strategies is viable with God. God’s power precludes meaningful contestation with him. As Peter Geach (Reference Geach1973, 8) puts it, ‘God is not just more powerful than any creature; no creature can compete with God in power, even unsuccessfully’. God’s sovereignty over the universe means that no third party could check his rule. He can’t even in principle be removed from his position. And God’s omnipresence means no one can exit his jurisdiction.
If slavery is dominative, the divine-human relationship is too.
Objections
I’ve argued that the divine-human relationship is dominative and therefore unjust. Here I consider four objections: (1) human childlikeness; (2) God’s goodness; (3) God’s justice; (4) God’s greatness.
Human childlikeness
I’ve said that we should take theists seriously when they describe themselves and others as slaves of God. But this is not the only way that theists describe the divine-human relationship. Many take humans to be God’s children (Proverbs 14:6; Isaiah 64:8; Malachi 2:10; Matthew 23:9; Ephesians 4:5–6; 1 John 3:1. The Qur’an treats such language with more scepticism. See, e.g., Qur’an 5:18.) Moreover, some take being a child of God and being a slave of God to be closely related (Garnsey Reference Garnsey1996, 222–225; Lactantius Reference Bowen and Garnsey2003, 230).
These parallels aren’t entirely surprising; parents are far more powerful than their young children. And like slaves vis-à-vis their masters, young children have little power to contest their parents’ interference. Young children are, prima facie, dominated by their parents.
And this is to children’s benefit. A child’s flourishing often requires that she is governed by someone more powerful, whose interference she cannot block and whose rule she cannot exit. Any other arrangement might inhibit her development, or worse. Children need parents to be able to sometimes interfere at will with their children for their benefit.
Given this, one might object that like children, humans need someone far stronger and wiser to govern them. And like children with parents, humans’ inability to contend with God’s interference is to our benefit. Successful contestation could result in significant self-harm. Humans should welcome divine domination as a profound and fundamental benefit.
The problem with this objection is that I have not argued that divine domination yields no benefit from humans; rather, I argue that it is unjust. And domination can be both beneficial and unjust. Here’s an example. Suppose an antebellum southerner finds an abandoned black child in the woods. He saves the foundling from exposure by bringing her into his household as a slave. He gives her food, water, and shelter. He dominates her and the relationship between them is unjust. To be sure, she benefits from the domination. But these benefits don’t make the relationship between them just.
They wouldn’t make her slavery just even if this were the best life she could reasonably hope for. Suppose the master is benevolent and knows that every other available master would be crueller. Emancipating the foundling would put her at risk of capture and enslavement by a crueller master. Smuggling her to the North would involve similar risks. So, he keeps her as his slave and rules her as gently as possible. In this scenario, the foundling benefits from her relationship with her master more than she would in any other relationship she might hope for. Even still, the benefits don’t make his mastery just. A relationship can be beneficial – it could even be the best arrangement the dominated person could reasonably hope for – while remaining dominative and therefore unjust. The divine domination problem remains.
Let’s modify the objection: if parent-child relationships are benign, as they seem to be, then either they aren’t dominative, or domination is sometimes benign. If God’s rule over humans is like parental rule, then it is similarly either not dominative or benignly dominative. Either way, there is no problem of divine domination.
My response to this objection denies that the parent-child relationship is inherently benign. Most societies have social practices and institutions to protect children from arbitrary parental interference. The intuition that the parent-child relationship is inherently benign may rely on taking such protections for granted. But in the absence of any such protections, parental rule is dominative: children are unable to resist arbitrary parental interference. And when dominative, parental rule is unjust. Representations of the risks unchecked parents pose to their children abound in literature: think of Pap Finn in Huckleberry Finn, Alphonso in The Color Purple, or Sweetness in God Help the Child. In The Color Purple, for instance, Alphonso – stepfather to Celie – repeatedly rapes Celie before selling her as a child bride to the book’s main antagonist, Mr ___. Alphonso can do this in virtue of his power over Celie as her parent, and because Celie lacks any power to resist him. The parent-child relationship is not inherently benign (Godwin Reference Godwin2011; Godwin Reference Godwin2020). Rather, it can be benign when children are protected from unrestrained parental power.
Thom Brooks (Reference Brooks2024) has recently argued that children can enjoy non-domination in a republican civil society. While children are at risk of parental interference, non-domination is compatible with non-arbitrary interference. According to the republican theory of government, interference is non-arbitrary when it is ‘the product of discursive control by the general citizenry’ (Brooks Reference Brooks2024, 46). Therefore, parental interference needn’t imply that children are dominated, so long as it is the product of discursive control of a democratic society of free and equal citizens (Brooks Reference Brooks2024, 44–46). In other words, if parental interference is controlled by laws that the public agrees to, it is not arbitrary and doesn’t imply that children are dominated.
Perhaps Brooks is right that external checks on parents can protect children from parental domination. Still, no solution in the vicinity seems to be available in the divine domination case. In a democratic society of free and equal citizens, the general citizenry exercises discursive control over laws, even if children don’t (or can’t) exercise this control themselves. But humans are in a categorically different position with respect to God. God’s omnipotence and sovereignty mean that none of us, even collectively, can exercise any discursive control over God.
My argument does not imply that in the absence of external checks on parental rule, parents ought not to procreate. But it does imply that absent any such checks, all procreation would at best be morally tragic. By morally tragic, I mean that even if procreation were on balance the right course of action, it would still incur significant moral harms. To see why, consider three scenarios:
Parental Domination | Parents dominate children.
No Procreation | Humans don’t procreate. Therefore, no children are dominated.
Protection | Parents govern children non-dominatively.
Whatever benefits children enjoy in Parental Domination, the relationship is unjust because it is dominative. Protection produces the same benefits non-dominatively. It is thus preferable to Parental Domination. Brooks’s argument implies that Protection is possible in a republican society, since he takes appropriate checks on parental rule to make it possible for parents to govern children non-dominatively.
What if Protection is impossible?Footnote 19 If a version of moral absolutism, according to which one must never directly bring about an injustice (regardless of the consequences) is true, then No Procreation is preferable to Parental Domination. If absolutism is false and the benefits of Parental Domination outweigh its injustice, Parental Domination is preferable. Even so, all procreation in Parental Domination would be morally tragic. Births that were good overall would introduce an injustice.
Now compare the following cases:
Divine Domination | God dominates humans.
No Creation | God doesn’t create humans. Therefore, no humans are dominated.
Protection* | God governs humans non-dominatively.
Protection* realizes the benefits of Divine Domination without its injustice. If it is possible, then, Protection* is preferable to Divine Domination. However, as Job attests, God’s properties seem to make Protection* impossible. Unlike human parents, God can’t be subjected to the rule of someone else who can control his actions. God cannot but dominate humans.
Therefore, just as under Parental Domination procreation is at best morally tragic, under Divine Domination creation is at best morally tragic. Creating humans introduces an unjust relationship between humans and God. If absolutism is true, No Creation is preferable to Divine Domination: failing to create humans is better than creating and dominating. If it is false and Divine Domination’s benefits outweigh its injustice, creating humans is morally tragic. Whatever good each human is or enjoys, her creation imposes on her the injustice of divine domination.
God’s goodness
Next, consider an objection from divine benevolence. Perhaps God’s benevolence is perfectly robust such that God can’t interfere with humans in unjust ways. And if unjust divine interference is impossible, then the divine-human relationship isn’t dominative (Morriss Reference Morriss2012, 14–15; Crummett Reference Crummett2021, 146).
At first dint, this objection may appear irrelevant. Only external checks on the empowered person mitigate domination (Pettit Reference Pettit1996; Wall Reference Wall2001, 219–220; Lovett Reference Lovett2012; Pansardi Reference Pansardi2013; McCammon Reference McCammon2015, 1043–1047; Gädeke Reference Gädeke2020; Brooks Reference Brooks2024). Benevolent masters still dominate their slaves because slaves have no way of controlling their masters’ interference. Similarly, if the only check on God’s interference with humans is his own benevolence, God still dominates humans. He remains able to interfere with humans at will; humans cannot block his interference. Domination is a matter of the empowered person’s position, not how she uses it.
However, we can understand the objection in a different way. It’s natural to think that the injustice of domination has something to do with the risk of arbitrary interference against her interests to which the dominated person is exposed (Kramer Reference Kramer2003; Krause Reference Krause2013; McCammon Reference McCammon2015; Pansardi Reference Pansardi2013; Pettit Reference Pettit1996). A relationship is unjust qua dominative only if the less powerful party can do nothing to stop the more powerful party from interfering with her against her interests. If she cannot, the more powerful party enjoys arbitrary power over her. However, if there is no risk of arbitrary interference against her interests, if the more powerful party could never arbitrarily interfere with her against her interests, then perhaps there is no injustice.
Understood like this, the objection aims to disrupt the inference from God’s properties to the injustice of the divine-human relationship:
Premise 1 | God is perfectly robustly benevolent.
Premise 2 | Benevolence entails a commitment against arbitrarily interfering with someone against her interests.
Premise 3 | For any property P, if God is perfectly robustly P, then it is impossible for God to act contrary to commitments entailed by P.
Conclusion | Therefore, it is impossible for God to arbitrarily interfere with someone against her interests.
If arbitrary divine interference against humans’ interests is impossible, then God’s rule puts humans at no risk. The divine-human relationship is either not dominative, or if dominative, benign.
In the first section of this article, we encountered Job taking his suffering to be a reason to doubt God’s perfect benevolence. He described God as harming humans for sport: ‘When disaster brings sudden death, he mocks at the calamity of the innocent’ (Job 9:23). This is one way the problem of evil makes the problem of divine domination salient. People suffering grave harms seem entitled to doubt God’s perfect benevolence.
Nevertheless, for the sake of argument let us grant Premise 1. I am, after all, posing an objection to classical theism, and classical theists typically take God to be perfectly benevolent.
Even so, we should deny the conjunction of the second and third premises. For one thing, it is unlikely that humans can make any necessary, non-formal, inferences about God’s actions from his essential properties. If God’s actions are informed by his reasons for action, then while humans may be able to infer that God will always act benevolently from God’s perfectly robust benevolence, we cannot know what benevolent action in a particular situation will be. This is because we cannot know all of God’s reasons. We must be open to benevolence sometimes involving interfering arbitrarily with humans against their interests. Knowledge of God’s essential properties is insufficient. But God’s reasons seem to be opaque to humans, as God himself says to Job. This is the main insight of sceptical theism: God’s reasons are beyond our ken (Bergmann Reference Bergmann2001, Reference Bergmann, Flint and Rea2009; McBrayer Reference McBrayer2010; Hendricks Reference Hendricks2023, 49–70. See also Murphy Reference Murphy2017, Reference Murphy and Draper2019).
That benevolence towards some person is compatible with arbitrarily interfering with her, against her interests, is especially clear when competing goods are at stake. Suppose that my family and I are in the state of nature. Therefore, neither of my two young children, Jack and Diane, enjoy the civil protections that Brooks thinks would shield them from my domination. Each has a legitimate interest in consuming a scarce, non-sharable, good. Diane, the stronger sibling, seizes the good by force. It is compatible with my benevolence towards Diane that I take the good from her and give it to Jack. Despite my benevolence towards Diane, I act against her interests. Since it seems that benevolence towards someone is compatible with arbitrarily interfering against her interests, it is plausible that in a world beset by scarcity and tragedy, God benevolently sometimes acts against one person’s interests to promote other goods (including others’ goods). This is, of course, the lesson of greater good theodicies (Collins Reference Collins, McBrayer and Howard-Snyder2013; Langtry Reference Langtry1998; MacDonald Reference MacDonald2023; Stump Reference Stump2010).Footnote 20
God’s perfectly robust benevolence does not, therefore, seem to imply the impossibility of his arbitrarily interfering against some person’s interests. Indeed, sceptical theism and greater good theodicies teach us that it couldn’t. Divine benevolence doesn’t solve the problem of divine domination.
Instead, God’s power, position, and inscrutability expose humans to the risk that he interferes with them against their interests. In response, many theists organize spiritual practices around this risk. Here is how the New Testament’s Epistle of James captures this idea:
Come now, you who say, ‘Today or tomorrow we will go to such and such a town and spend a year there, doing business and making money’. Yet you do not even know what tomorrow will bring. What is your life? For you are a mist that appears for a little while and then vanishes. Instead, you ought to say, ‘If the Lord wishes, we will live and do this or that’ (James 4:13–15).Footnote 21
If my argument is right, James’ striking warning is inevitable. No one can securely make even short-term plans. God could, at will and with impunity, demand anyone’s life at any moment (or interrupt one’s plans in some other way). Humans must be uncertain about the future and should always appeal to God’s goodwill even in making short-term plans.
This uncertainty motivates some theists to devote significant time and energy trying to anticipate God’s plans (Butler Reference Butler1979; Francis Reference Francis2011; Jeffers Reference Jeffers2007; Klingshirn Reference Klingshirn2021; Ness Reference Ness1990; von Stuckrad Reference von Stuckrad2000). A tough row to hoe, if humans can’t access God’s reasons.
God’s power, position, and inscrutable will also incentivize humans to bow and scrape before, and flatter and cajole him. We saw this in the quote from Augustine in the second section of this article. Many theists call these attitudes and behaviours humility and piety. Their critics have preferred ‘monkish virtues’ (Hume Reference Hume1902, 270), ‘slave morality’ (Nietzsche Reference Nietzsche and Faber1990, 83, 155–156), and the outward projection of humans’ inward yearnings (Feuerbach Reference Feuerbach and Evans1881, 250, 281). We needn’t decide who is right to notice that despite God’s benevolence, theists often behave before him in ways that would in other contexts be considered servile.
Finally, recall Job’s concern that someone who believes God to be unimpeachable cannot in good conscience accuse God of injustice. Challenging and contending with God might initially appear to be potential means of mitigating divine domination. However, to challenge or contend with God requires that one accept as possible that God ought to have done otherwise. One cannot do this without also denying God’s robustly perfect benevolence. God’s robustly perfect benevolence therefore seems to make it wrong to resist divine interference via challenge and contestation. Someone hoping to avoid wrongdoing, once she realizes this, is therefore strongly disincentivized against challenging or contending with God. Knowledge of God’s robustly perfect benevolence deepens, rather than ameliorates, the problem of divine domination.
Marilyn Adams argues that though God’s position as sovereign generates a default position of flattery, the story of Job teaches that God leaves some room for humans to challenge and even blaspheme him. Indeed, in the face of ‘horrendous suffering’, Adams says, ‘blasphemy is inevitable’ (Adams Reference Adams2003, 38–39, 44). Similarly, while John Roth grants that God’s sovereignty makes him a master, he argues that humans are also licensed to protest (apparent?) divine injustices (Roth Reference Roth and Davis2001, 13. See also, Reference Roth2004; Rea Reference Rea2019). Perhaps so, though Stephen Davis worries that Roth gives up on God’s perfect goodness (Davis Reference Davis and Davis2001, 20–21). And one might have the same worry about Adams.
If – and it is a big if – Adams and Roth are right, then God’s benevolence may not make the problem of divine domination worse as I, following Job, suggest. If God permits blasphemy and protest, a person’s knowledge of God’s perfection doesn’t jeopardize his standing to challenge God. But it doesn’t mitigate the problem, either. Blasphemy, protest, and other forms of human speech don’t seem to block or control God’s interference in any way. The divine-human relationship remains dominative.
God’s greatness
I have used the example of master’s rule over the slave to motivate the intuition that domination is inherently unjust. And I’ve argued that God dominates humans by arguing that the dominative features of inter-human slaveholding are amplified in the divine-human relationship. But perhaps mastery gets us off on the wrong foot in thinking about the divine-human relationship. A better analogy might be between humans and some animals. Rats bred for scientific experiments are subjected to treatment that would be considered torture if performed on humans. They are injected with diseases, shocked, drowned, and despite being highly social, isolated from other rats.
Humans arbitrarily interfere with these rats and are not forced to track their interests. Whether rats are dominated may depend on whether they have a will (a question I am not prepared to answer). But they plainly suffer something at least analogous to domination. To avoid begging the question, I’ll put ‘domination’ and its cognates in scare quotes when talking about human-rat relations.
If the human-rat relationship is benign, then humans ‘dominating’ rats is not unjust. What could explain why dominating other humans is unjust, while ‘dominating’ rats is not? Perhaps humans’ elevated status compared to rats entitles us to treat rats in ways that we may not treat other humans. And perhaps in the same way, God’s greatness – his elevated status compared to humans – entitles him to dominate us. Suppose further that a greater status difference implies a wider range of entitlements the superior party enjoys over the inferior. This might explain the intuition that humans may do to rats things we may not do to other primates, and to fruit flies things we may not do to rats. If classical theism is true, the status difference between God and humans is far greater than that between any creatures. Perhaps this is part of what Bildad is doing in suggesting in Job 25:6 that humans are no more than worms and maggots before God (Job 25:6). God thus enjoys more entitlements over us than we could ever enjoy over any other creature.
If successful, this objection disrupts the inference from the injustice of inter-human slavery to the injustice of divine domination. However, this objection is committed to a dubious explanation of what makes domination unjust.
An example adapted from H.G. Wells’s The Time Machine will illustrate why. Suppose that in one million years, two species sharing the common ancestor homo sapiens, evolve. Eloi are beautiful and fine, noble and virtuous, intelligent and gentle. Morlocks are ugly and coarse, base and vicious, dumb and brutish. But Morlocks can recognize and therefore complain when they are harmed or mistreated, or when they perceive themselves as dominated. Would the Eloi enslaving the Morlocks be just? Wells’s Time Traveller does not think so (Wells Reference Wells1895, 112–119).Footnote 22 Neither do I.
If you agree with us, you should doubt that status differences justify domination. The idea that God’s greatness justifies divine domination, or makes God’s rule over humans non-dominative, is therefore specious.
God’s justice
I’ve argued that the implications of (b) and (c) preclude (a):
(a) The divine-human relationship is just.
(b) Classical theism is true.
(c) Dominative relationships are inherently unjust.
A classical theist wishing to resist my argument by affirming (a)–(c) would need to explain why despite God’s overwhelming power he is not able to interfere with us arbitrarily, or why his ability to do so is non-dominative. In previous sections I argued that two prima facie plausible explanations don’t work: first, that God is precluded from such by his benevolence; second, that God’s greatness makes his ability to arbitrarily interfere with us non-dominative. But perhaps someone who affirmed (a)–(c) could supply an alternative way to reconcile them.
Another way to respond to my argument would be to argue that (a) must be true because God is perfectly just. Therefore, we must instead reject either (b) or (c). I suggested earlier two ways of rejecting (b): deny that God is very powerful or deny that God has a will.
But classical theists cannot reject (b). And if one cannot consistently affirm that God is perfectly just and reject (a), then the only other option is to reject (c). While this option is technically available to the classical theist, it’s worth making its stakes clear: denying that dominative relationships are inherently unjust implies denying that intra-human slavery is inherently unjust insofar as it is dominative. Bite the bullet but invest in crowns.
Conclusion
God dominates humans in virtue of his essential properties. Job’s complaints showed us that God’s power and sovereignty are enough to get the problem going, though formalizing the argument and considering objections to it revealed that God’s omnipresence and aseity also contribute to the problem.
The book of Job suggests several connections between the problems of evil and divine domination, which my analysis has highlighted. Both problems are grounded in God’s properties. As a result, like the problem of evil, the problem of divine domination puts pressure on God’s decision to create. The problem of divine domination suggests that creating humans generates an injustice. Furthermore, the problem of evil makes the problem of divine domination salient: suffering at the hands of a powerful agent can enlighten the sufferer to her domination.
I have also clarified two novel aspects of divine domination’s relationship to the problem of evil. First, divine domination could obtain independently of suffering. If benevolent mastery is unjust, then the divine-human relationship would be dominative even if there were no human suffering. Finally, two familiar responses to the problem of evil (greater good theodicies, sceptical theism) exacerbate the problem of divine domination.
This is a significant problem for theism. In creating humans, God creates an unjust relationship; creation is at best a moral tragedy. Absent a rectification of the divine-human relationship’s injustice, the divine-human relationship is more unjust than slavery.
To be clear, I pose the problem of divine domination not as a defeater for classical theism, but as an invitation to classical theists to take it seriously. In that spirit, let me end by gesturing towards what I think a solution to the problem would require. Earlier, I noted that external checks on parental rule may allow parents to govern their children while protecting children from parental domination. I claimed that classical theism cannot avail itself of an analogous solution to the problem of divine domination: unlike human parents, God is sovereign of the universe and thus subject of none.
However, God ceding to humans the power to control against his potential interference might be able to approximate the sort of concession that occurs when parents and children enter civil society. A solution along these lines would need to explain how God gives humans this power, which humans can exercise it, how they can successfully block divine interference, and the consequences of a successful block. A classical theist might need it to do all this while also preserving divine goodness, power, sovereignty, and aseity. No small feats. But such a concession, if possible, could reconfigure the divine-human relationship such that humans could control God’s potential interference. It might solve the problem of divine domination.
Acknowledgements
This article owes its existence to Brian Lee and Jeffrey Stout, whose provocations set it in motion. For discussion of and feedback on earlier versions of this article, I am very grateful to Jane Abbottsmith, Alexia Alkadi-Barbaro, Paul Billingham, Joseph Clair, Marilie Coetsee, Ryan Davis, Ryan Darr, David Decosimo, Jessica Flanagan, Jill Frank, Charles Guth, Justin Hawkins, Jennifer Herdt, Anne Jeffrey, Michelle Kosch, Brian Lee, Alexander Livingston, Austen McDougal, Joseph Moore, Mary Nickel, and audiences at the Cornell Political Theory Workshop, the First Fridays Faith and Politics Group, the Villanova Frontiers of Pluralism Conference, the Yale Religious Ethics Colloquium, and two anonymous reviewers for Religious Studies.