The Relevance of God to Metaethics

Abstract

I argue in this paper that moral realism is more likely to be true if God exists than if God does not exist. I will first argue that, without the existence of God, objective moral facts would be queer, but that the queerness problem is solved if God exists. I will then go on to argue that no being could be God unless that being has authority over all created rational beings just as morality does, which explains why the existence of God is relevant to metaethics.

1. Introduction

I shall argue in this paper that moral realism—the view that there are objective moral facts and properties—is more likely to be true if God exists than if God does not exist.

I shall argue that objective moral facts, which prescribe what sort of people we should be, how we should live our lives and what we should or shouldn’t do, seem queer entities, but that this apparent queerness disappears if moral facts derive from God in some way, such as from his commands or attitudes, etc. What seems queer is that a mindless universe could prescribe how we should live, but there does not seem to be anything queer about God prescribing how we should live.

If morality derives from God as suggested then this would imply that God has authority over all created rational beings just as morality does, and I will go on to argue that this is necessarily the case. No being could be God unless that being has authority over all created rational beings.

I do not argue in this paper that either theism or moral realism is true. Nor do I argue that moral realism would be impossible if theism turns out not to be true. My aim is simply to argue that moral realism is more plausible if theism is true than if atheism is true.

2. The Queerness Argument

According to moral realism, there are objective moral facts and properties. An act of sadistic torture has the property of “wrongness” attached to it. An act of feeding a starving child has the property of “goodness”, and perhaps sometimes the property of “obligatoriness”, attached to it. Many acts have the property of “permissibility” attached to them.

Russ Shafer-Landau (2004, p. 87) distinguishes between moral properties and moral principles. Whereas moral properties attach to actions, so that in a world without actions there would be no moral properties, moral principles—for example, the principle that if an act of sadistic torture were to occur it would be wrong—exist regardless of whether or not there is any agent capable of performing the action in question. Such principles are objective in the sense of being “stance-independent”; they are “not made true by virtue of their ratification from within any given actual or hypothetical perspective” (Shafer-Landau 2003, p. 15).1 He thinks that the existence of moral principles is simply a brute fact about the world, and needs no explanation.

He looks for an analogy with logical standards and physical laws. According to Shafer-Landau,

there isn’t anything that makes such things true—they simply are true. Propositions in these domains are irreducible, and so there won’t be any illuminating explanation of what makes them true. One can say, of course, that what makes these standards true is the nature of logic or of the physical universe, but if this is taken as at all illuminating, the moral realist too can avail herself of the response: correct moral standards are true in virtue of the nature of morality, the way morality works, the ‘moral universe’, or some such locution. . . . If some standard is true, irreducible, and to be construed realistically, then nothing makes it true; its truth is not a creation, but instead a brute fact about the way the world works.

(Ibid., pp. 47–48)

There does seem to me to be an important disanalogy here though. Imagine a world where there are no physical objects but where the physical laws are exactly as they are in the real world. These laws tell us how the world would have worked had it contained physical objects. For example, we know that in such a world two objects would attract each other with a certain force, depending on their masses and the distance between them.

But now imagine a world where there are no moral agents but where the moral principles are exactly as they are in the real world. These principles do not appear to tell us anything about how the world would have worked had it contained moral agents. The principle that if an act of sadistic torture were to occur then it would be wrong is consistent both with a world where moral agents frequently torture others and also with a world where moral agents never torture others. The moral principles do not tell us how moral agents would behave, but only how moral agents should behave. Unlike physical laws, they appear to be, not a brute fact about the way things are, but a brute fact about the way things should be.

Or imagine two possible worlds, identical to one another in terms of their fundamental laws except that one contains objective moral principles and the other does not. In both these two worlds, over a long period of time, agents evolve to have certain emotional reactions towards specific types of behaviour, even imposing certain standards of behaviour on one another and using words like “right” and “wrong” to describe certain actions. Perhaps people refrain from torture in both of these two worlds, and torture is typically described in both worlds as “wrong”. What is the difference between the two worlds?

The obvious answer, it seems to me, is that in one of these worlds the inhabitants act in certain ways and are required by the universe to act in these ways, whereas in the other possible world the inhabitants act in exactly the same way but are not required by the universe to do so (although they may be required by one another). Here, then, is the relevant difference between moral standards and the laws of physics. The laws of physics tell us that the world is a certain way, such that any object in the world will behave in a certain way; objective moral standards tell us that the world is meant to be a certain way, such that human beings are required to behave in a certain way. But there seems something strange about a mindless universe requiring us to be one way rather than another.

This is, in effect, the queerness argument that Mackie raises against objective moral values. Mackie argues that if there were objective moral values they would be metaphysically peculiar, because “they would have to be intrinsically action-guiding and motivating” (Mackie 1977, p. 49)—so peculiar, in fact, that it would be preferable to abandon belief in such things and embrace a moral error theory.

It is important to see that Mackie’s concern is not primarily that moral facts would be non-natural; mathematical facts might also be thought of as non-natural, and even queer, in this sense (cf. Mackie 1977, p. 39), but as Hattiangadi points out,

even if we postulate objective, sui generis, mathematical facts, we are not postulating facts that are categorically prescriptive. In contrast, if we suppose that there are normative facts, we are postulating sui generis, objective facts that are also categorically prescriptive. Putative normative facts would be utterly unlike any other facts with which we are acquainted, including putative non-natural facts about mathematics. Thus, to postulate sui generis normative facts or properties is to postulate entities that are inherently queer.

(Hattiangadi 2007, pp. 49–50)

What is allegedly queer about objective morality, then, is that it is (or would be) objectively and categorically prescriptive.

But what does Mackie mean when he says that moral facts are “objectively prescriptive” (Mackie 1977, p. 48)? One thing he appears to think is that the wrongness of some action will necessarily provide us with both a reason and a motive to avoid that action. Other error theorists deny that they must provide us with a motive, but agree that they must provide us with a reason, and this is what they see as queer (cf. Garner 1990, pp. 142–43; Joyce 2001, pp. 30–31; Olson 2014, pp. 103–38).

But what is so peculiar about moral facts providing us with a reason for action? The problem seems to be that, at least according to one common understanding of what it means to have a normative reason, our reasons are contingent on our desires, goals or purposes. Moral principles, on the other hand, apply to us irrespective of our desires, goals and purposes. Hence Mackie says:

Objective wrongness, if there is such a thing, is intrinsically prescriptive or action-guiding, it in itself gives or constitutes reason for not doing the wrong action, and this holds also for some, if not all, other moral features. To say that they are intrinsically action-guiding is to say that the reasons that they give for doing or for not doing something are independent of that agent’s desires or purposes.

(Mackie 1982, p. 115)

I assume that when Mackie says that I have a reason to act in a certain way, he is implying that I would be irrational for not acting in that way (at least unless I have another reason, which is just as strong, to act in some other way). This suggests that if we are morally required to act in some way then we must also be rationally required to act in this way. However, some moral realists, such as David Brink (1984, p. 114), deny that we must have a reason to do what is morally required (cf. Foot 1972). If Brink is right then one might think the queerness problem disappears: Mackie is simply wrong to insist that moral features provide us with reasons for action.

But the queerness remains. In denying that morality necessarily provides us with reasons, Brink is not denying that it makes requirements of us; as he says, “While moral requirements apply to us independently of our antecedent desires and interests, they give us reasons for action conditional on their promoting our interests and desires” (Brink 1992, p. 7, emphasis mine). Brink thus distinguishes between moral requirements and moral reasons.

But as I have argued previously (Taylor 2020), the real queerness of morality lies in the fact that we are required to act a certain way, irrespective of whether or not we have reason to do what is required of us.2 Actions would have “an intrinsic to-be-done-ness or not-to-be-done-ness” attached to them (Mackie 1982, p. 116; cf. Copp 2000, p. 130). It is this that is queer. There is nothing strange about human beings expressing prescriptions or recommendations to other human beings, but there does seem something mysterious about the universe prescribing how we are to live.

3. The Appeal to God

I maintain that theism could resolve the queerness problem. This was also Mackie’s view, even though he was an atheist. In fact, it is interesting that, when he argues against objective morality he concedes that this might be defensible if theism could be defended (Mackie 1977, p. 48), whereas when he argues against theism he concedes that this might be defensible if objective morality could be defended (Mackie 1982, pp. 115–16). He clearly thinks that theism and objective morality stand or fall together.

Why should the existence of God make morality less queer? Mackie points out that

we can understand a human thinker, either as an agent or as a critic, seeing things as to be done or not to be done, where this is a reflection or projection of his own purposiveness; hence if we are to explain an intrinsic to-be-done-ness or not-to-be-done-ness, which is not such a reflection or projection, it is natural to take this as an injection into reality made by a universal spirit, that is, something that has some analogue of human purposiveness.

(Mackie 1982, pp. 116–17, emphasis in original.)

God would be just such a “universal spirit”.

Mackie’s argument, then, goes something like this. If, as Mackie himself believes, atheism is true, then what has happened is that human beings have adopted certain attitudes or values, projected these attitudes onto the world and then come to believe that these attitudes or values are an objective feature of the world. If, on the other hand, theism is true, then what has happened is that God has certain attitudes or purposes which he “injects” into reality, so that when we (correctly) judge something to be wrong we are recognizing a real objective feature of the world. He does not say exactly what he means by “injection”, but it is clear that he means something analogous to human “projection”.

There are similarities between Mackie’s view and the view of Elizabeth Anscombe. Anscombe (1958, p. 6) thinks that the concept of moral obligation arose within a divine law concept of ethics, and makes no sense in a world that has abandoned theism. She draws an analogy with criminality; the concept of criminality arose within the concept of criminal law and criminal courts, and would make no sense in a world which has abandoned criminal law or courts. Just as it would make no sense to talk about criminals once we have abandoned criminal law and criminal courts, so it makes no sense to talk about obligation once we have abandoned theism; in both cases, we would be left with “the interesting [situation] of the survival of a concept outside the framework or thought that made it a really intelligible one” (ibid.).

Anscombe thought we should adopt a more Aristotelian concept of ethics. But Mackie thinks that this would fail to solve the queerness problem. An Aristotelian concept of ethics implies that there is an end, or goal, of human life; and he thinks such an approach to ethics tends to equivocate between the descriptive claim that humans would in fact find satisfaction if we pursued some goal, and the prescriptive claim that humans have a “proper” goal, that we ought to be pursuing whether we are or not. He thinks that the prescriptive claim only makes sense if God exists, so that “the true purpose of human life is fixed by what God intended (or intends) men to do and to be” (Mackie 1977, pp. 46–47; cf. Mackie 1980, pp. 38–39).

Hence, even if the purely descriptive claim that a particular way of life would enable us to develop our capacities or find satisfaction might make sense without the existence of God, the prescriptive claim that we are required to live that way of life only makes sense if we are required by God to do so; in this case, God’s requirements “would add an objectively prescriptive element to what otherwise were hard, descriptive truths, but in a quite non-mysterious way: these would be literally commands issued by an identifiable authority” (Mackie 1977, pp. 230–31).3

I think Mackie is right to think that theism, if it could be defended, would be the best way to resolve the queerness problem. I think it fares better than some of the other solutions that have been offered.4

For example, Lillehammer (2013, p. 4206) suggests that the problem is “targeted at a strong form of ethical non-naturalism”, such that is “unlikely to be accepted by many of Mackie’s contemporary critics”. Mackie (1977, pp. 41–42) himself anticipates this type of response, and acknowledges that we might seek to avoid the force of his argument by replacing his own language with the more everyday terms that people use to describe moral judgments; but as he points out, the problem only seems to be resolved because “the claim for moral authority may pass unnoticed”. The problem is the authority of morality, and we cannot evade the problem by adopting language that masks this authority. Hence his point that Aristotelian ethics is vulnerable to the queerness problem so long as it entails an objectively “proper end” for humans.

Miller (2019, sct. 3) discusses a possible solution which involves acknowledging that morality “binds an agent” whilst denying that this gives the agent a reason for action. I have already shown why this does not solve the problem; the problem is that morality binds the agent in the first place.

But suppose that it is argued that there is nothing queer about having good reasons, and that being morally required to do something just is having an overriding reason to do it. Again, Mackie anticipates this response; replacing his own language with “talk about good reasons for or against possible actions” (Mackie 1977, p. 41) is just one of the ways in which he suggests we might seek to evade the problem. But when we condemn Hitler’s actions, we are doing much more than saying he had good reason not to kill Jews; we are condemning him for being immoral, not for being irrational. As R. M. Adams (1999, p. 138) points out, when we do something morally wrong we have something to feel guilty about, whereas when we do something irrational we merely have something to feel silly about.

Perhaps we could simply bite the bullet, accept that moral facts are mysterious but deny that this justifies rejecting them. I think this is in fact the best non-theistic realist response to the queerness problem. As Platts points out:

The world is a queer place. I find neutrinos, aardvarks, infinite sequences of objects, and (most pertinently) impressionist paintings peculiar kinds of entities; but I do not expect nuclear physics, zoology, formal semantics or art history to pay much regard to that.

(Platts 1980, p. 72, emphasis his)

Mackie himself does not propose adopting an error theory unless we can explain how belief in objective morality “has become established and is so resistant to criticisms”—which he explains in terms of projecting attitudes and social demands onto the world (Mackie 1977, pp. 42–43). But even with this explanation, it is still open for a moral realist to accept moral facts in spite of their queerness.

Nevertheless, I think theism does provide a solution to the queerness problem which is better than the solutions just outlined, because it would solve the specific problem relating to the authority of morality in a way that these non-theistic responses fail to do. This is because God, if he exists, is himself authoritative—a point I will argue for later on—which means that the authority of morality has its source in the authority of God. Of course, my discussion of these other solutions has been very quick (and I do not claim to have discussed all possible solutions), and so I want to emphasize that I am not arguing that atheistic moral realism is impossible. My aim is merely to show that theism makes (or would make) moral realism more plausible than it would otherwise be.

4. The Relationship Between God and Morality

Suppose, then, that Mackie is right to think that theism could provide a solution to the queerness problem; how might ethics relate to God? It is clear from the previous section that Mackie thought objective morality might involve divine commands, but it would be misleading to think of him as suggesting that a theist should adopt a straightforward Divine Command Theory. He clearly does not think that God could make just anything right or wrong (Mackie 1982, pp. 114–15). Given his suggestion that morality must be related in some way to what would enable us to develop our capacities, it may be that he has in mind a sort of combination of Natural Law Theory and Divine Command Theory—perhaps the sort of theory attributed by Mark Murphy to Duns Scotus, Suarez and Locke (Murphy 2025, sct. 1.4). Although there may be some plausibility to this type of theory, it is not the only way to relate ethics to God, and is not the way I will develop in what follows.

One might instead adopt the type of Divine Command Theory developed by Robert Adams. Adams’ version of Divine Command Theory is specifically a theory about moral obligation, rather than moral properties generally. He denies that moral goodness, for example, depends on God’s commands. Instead, moral goodness involves resemblance to God. There are, however, different ways in which one thing can resemble another—for example, a caricature does not resemble in the same way that a portrait does—and so resemblance to God is not sufficient in order for something to be good. Specifically, in order to be good, something must resemble God in such a way that the resemblance can serve as a reason for God to love the thing (R. M. Adams 1999, pp. 28–38).5

God, of course, resembles God supremely. But Adams’ semantic account of moral terms allows him to deny that God would be good if, for example, he were cruel, even though a cruel God would clearly still resemble himself. Borrowing from the causal regulation theory associated with Putnam and Kripke, Adams argues that, in determining what the term “good” refers to, we must consider the type of things we use the word “good” to describe. Just as we can only conclude that the term “water” refers to H₂O if most of the stuff that we have, historically, called “water” is constituted by H₂O, in the same way we can only conclude that the term “good” refers to resemblance to God if most of what we have, historically, called “good” resembles God. I have elsewhere suggested a potential objection to Adams’ semantic account, but I don’t consider this to be decisive (Taylor 2016). Adams thinks that goodness is, at least to some extent, recognizable; we might sometimes be mistaken, but cannot be totally mistaken, about what type of things are good. For example, it could not be good to love cruelty, and so if God loves cruelty then goodness would not resemble God after all. God would not be good if he loved cruelty (R. M. Adams 1999, pp. 15–17; 46; cf. R. M. Adams 1979).

This account of moral goodness enables Adams to develop a modified Divine Command Theory of obligation, according to which being obligatory consists in being commanded by a specifically good God, and being wrong consists in being contrary to what is commanded by a specifically good God (R. M. Adams 1999, p. 250).6 His theory is a property-identity theory; he thinks it is metaphysically necessary, though not an analytic truth, that the property of wrongness “is (i.e., is identical with) the property of being contrary to the commands of a loving [or good] God” (R. M. Adams 1979, p. 76).

Adams denies that a good God must command or forbid all the things that God in fact commands or forbids. For example, he thinks that it is consistent with God’s goodness that his commands relating to euthanasia could have been somewhat different (R. M. Adams 1999, pp. 255–56). Hence Adams is open to the idea that there are possible worlds where something which is wrong in this world might be obligatory. Adams therefore appears to reject the claim that the moral strongly supervenes on the non-moral—i.e., the claim that two possible worlds identical in terms of non-moral features must be identical in terms of moral features. This does not mean, however, that just anything could be right or wrong; it may be consistent for a good God to have given slightly different commands but it does not follow that it is consistent for a good God to have given radically different commands.

I do not think, therefore, that it is a problem with Adams’ theory if the moral fails to strongly supervene on the non-moral; more problematic is whether his theory entails that the moral fails to weakly supervene on the non-moral. The moral would fail to weakly supervene on the non-moral if, in any world, either two things could differ in their moral features without differing in their non-moral features, or the moral features of something could change over time without any change in its non-moral features. Murphy (2002, p. 90) thinks that it could be rational for God to give a command to one person but not another, even though there is no relevant non-moral difference between them, in which case one would have a moral obligation that the other would lack. In addition, as R. M. Adams (1999, p. 270) acknowledges, commands are given at a specific moment in time, which means that, with the giving of the command, the moral features of something changes even though there appears to be no change in its non-moral features.

It is important to be clear exactly what the problem is here. One might argue that God’s command is one of the things that the moral must supervene on, and so the fact that God has given a command is a relevant non-moral feature of the situation which explains the different moral features (cf. Blackburn 1993, p. 146). The problem is that, according to Adams’ theory, the property of wrongness and the property of being forbidden by God are identical to one another. It looks, then, as if the property of being forbidden by God is a moral, not a non-moral, property. We can still say, of course, that the property of wrongness supervenes on the property of being forbidden by God (any property supervenes on itself) but what we want to say is that moral properties supervene on non-moral properties.

There are ways for Divine Command Theorists to respond to this concern. For example, it might be argued that although the property of being forbidden by God is identical to the property of wrongness, it is itself a non-moral property—just as we might argue that although water is identical to H₂O, water is not a chemical property.7 Alternatively, since Adams insists that God’s will would not be sufficient, in the absence of a command, for something to be morally obligatory, it is open to him to respond that being willed by God is a non-moral property, and that God’s commands (and hence moral properties) supervene on God’s will. Or one might adopt a slightly different theory, where obligation relates to another feature of God’s mental state.

For example, perhaps the property of wrongness is identical, not to the property of being forbidden by God, but to the property of being disapproved of by God—where God’s attitudes of approval and disapproval are analogous to the type of human attitude that expressivists are concerned with (cf. Taylor 2023, pp. 232–35). If moral error theory is true, it is plausibly this type of attitude that we project onto the world and mistake for objective moral properties, and if theism is true it is plausibly this type of attitude that God might “inject” into the world.

Blackburn (1973, p. 122) points out that this type of attitude must be held towards something on the basis of the natural features of the thing in question; it is not possible for somebody to have different attitudes towards two things that are believed to be identical in non-moral respects. If this is so with humans there seems no reason why it shouldn’t be so with God too (and of course, being omniscient, if God believes two things to be identical in non-moral respects then they are identical in non-moral respects). The potential problem of weak supervenience just mentioned in relation to Adams’ account is therefore avoided. Even if God could give different commands to two people whose situations are identical in non-moral respects, he could not approve of one person and yet disapprove of another whose situation is identical in non-moral respects. In addition, unlike commands, which are given at a particular point in time, God’s attitudes are plausibly held eternally, which means that the moral status of something cannot change over time unless there is some change in its non-moral features.

This is not to deny, of course, that God’s commands have an important role to play in our moral duties. On the plausible assumption that God disapproves of us disobeying his commands, it would be wrong for us to do what God has forbidden us to do. But because the property of wrongness is not now identical to the property of being forbidden by God, the fact that God has forbidden something is a non-moral (though of course morally relevant) consideration—just as a parent’s command to a child is a non-moral consideration which is relevant in explaining why two children in otherwise identical situations might have different obligations, assuming children have a moral obligation to obey their parents.

Whereas Adams adopts a Divine Command Theory only for moral properties of the obligation family, I think the sort of account I am suggesting could be adopted for non-obligatory moral goodness too.8 For example, if the property of wrongness is identical to the property of being disapproved of by God, then the property of moral goodness might be identified with the property of being approved of by God. (I would be morally obligated to do something if, in addition to God approving of me doing it, he also disapproves of me not doing it.)

No doubt many will object to the claim that God himself is good in virtue of the fact that he approves of himself, although I do not think this is as problematic as many think. According to expressivism, when I say I am morally good I am expressing approval of myself. But if I approve of myself, this is most likely because I have, or think I have, those qualities which I approve of. Similarly, if God approves of himself it is presumably because he has those qualities that he approves of, and none that he disapproves of. Furthermore, I think a good argument can be made for the claim that God necessarily has none of those qualities of which he disapproves. If so, then God could not, in any possible world, approve of something that God, in this world, considers morally abhorrent, like torture.9

Nothing I have said entails that we need to believe in God in order to determine what is right or wrong. If we have the capacity to come to know moral truths, I see no reason why this capacity would not be just as effective if not more so (for atheists as well as for theists) if morality derives from God in the way that I have suggested.10

Although I have offered a theory somewhat different from that offered by Adams, it is unimportant from the point of view of this paper which theory one prefers. What is important is that, whether we think in terms of God’s commands or his attitudes, they must carry the type of authority that morality carries. But why should we think of either his commands or his attitudes as authoritative? I will argue that the answer lies in the kind of being that God is. I will argue in Section 5 that, in order for some being to be God, that being must be a fitting object of worship, and that this entails being authoritative over all created rational beings. I will then expand on this by arguing in Section 6 that a being could not be “that than which none greater can be conceived” unless that being were authoritative over all created rational beings.

5. God as a Fitting Object of Worship

James Rachels famously gave an argument for atheism which begins with the premise that “any being who is not worthy of worship cannot be God” (Rachels 1971, p. 325), and then goes on to argue that the primary function of worship is that it is a way of acknowledging that we are God’s subjects—it is a way of “recognizing him as having an unqualified claim on our obedience” (ibid., p. 334). He emphasizes:

We cannot recognize any being as God, and at the same time set ourselves against him. The point is not merely that it would be imprudent to defy God, since we certainly can’t get away with it; rather, there is a stronger, logical point involved—namely, that if we recognize any being as God, then we are committed, in virtue of that recognition, to obeying him.

(Rachels 1971, p. 322)

I want to make two initial points. Firstly, Rachels seems to use the words “obedience” and “allegiance” more or less interchangeably. For example, he says that “to apply the title ‘God’ to a being is to recognize him as one to be obeyed”, before going on to say that “to recognize any being as God is to acknowledge that he has unlimited authority, and an unlimited claim on one’s allegiance” (ibid., p. 333, emphasis mine). The word “allegiance” has the advantage that it is not as narrow as the word “obedience”—we can show allegiance to God either by obeying his commands or by living and acting in ways that he approves of, and by not living and acting in ways that he disapproves of. I think allegiance is therefore the better word.

Secondly, the logical implication of recognizing some being as God, I believe, is that we recognize this being as having a claim on our allegiance, rather than that we are thereby committed to giving him our allegiance. I assume Satan could recognize God as one to whom it is appropriate to submit whilst refusing to submit to him. I think the case is analogous to recognizing moral facts; it is a logical implication of recognizing some action as wrong that we recognize this action as one to be avoided, but we need not thereby be committing ourselves to avoid it.

Rachels goes on to argue that absolute submission to such an authoritative figure would be incompatible with human autonomy, and an abdication of moral decision-making, and he therefore concludes that there can be no being having an unqualified claim on our submission—and hence no God. I do not think that Rachels’ argument succeeds in showing that God does not exist, but this is not my concern in this paper.11 My concern here is to argue that he has a more or less correct concept of God as being, necessarily, worthy of worship, and in worshipping God we are acknowledging that he has an unqualified claim on our allegiance.

Rachels’ claim that no being could be God unless that being is a fitting object of worship is a widely held view, accepted by both theists and atheists. Stephen Clark (1982, p. 341) argues that “The being God is, is one almighty, everlasting, omniscient and omnibenevolent being, but His being, exactly, God lies in being accounted worshipful”. Similarly, Nicholas Everitt (2004, p. 15) thinks that God is, by definition, among other things, “an appropriate object of worship”, and Richard Gale (1991, pp. 7–8) thinks that “being eminently worthy of worship and obedience” is a “hard-core” property of God—it is a non-negotiable feature of God such that the term “God” could not refer to a being who was not worthy of worship and obedience. Hugh Burling (2019, p. 343) defends the view that “the reference of ‘God’ is determined by the implicit sense of the term, the definitive description ‘the being that is worthy of our worship’”.12

Even the word “god” with a small “g” seems to imply an object of worship. Although the term might sometimes be used in other senses, such as when it is used to describe human beings who display certain characteristics, in its usual, and most obvious, sense, the term “god” describes an object of worship. For example, when, in the Old Testament, we read of the “gods” of the nations, it is clear that these are seen as objects of worship. When Isaiah pours scorn on those who “bow down to a block of wood”, it is taken for granted that to make a “god” is to make an object of worship (Isaiah 44:9–20). Similarly, in the New Testament, when St. Paul finds in Athens an altar to “an unknown God”, he takes it for granted that this is an object of their worship—something that they worship as unknown (Acts 17:23).

So “being an object of worship” seems to be a central meaning of the term “god”. In this sense, as St Paul tells the Christians at Corinth, there are many “gods” because there are many objects of worship; but there is only one who is worthy of being worshipped and so there is only one true God (1 Cor 8:4–6). There are many objects of worship, but only one fitting object of worship.

But we might still wonder, if there is some being who is God, and is therefore a fitting object of worship, what it is that makes him a fitting object of worship. As Bayne and Nagasawa (2006, p. 303) put it, “What might the basis of God’s worshipfulness be?”

A plausible answer is that God is a fitting object of worship for the very reason that he is God. In other words, there are certain attributes in virtue of which some being is God, and it is in virtue of these attributes that this being is a fitting object of worship. Some being is God if that being is such that none greater can be conceived, and similarly, some being is a fitting object of worship if that being is such that none greater can be conceived. Just as God is God in virtue of his excellences, so God is a fitting object of worship in virtue of his excellences.

Bayne and Nagasawa object to this idea, on the basis that the attributes in question “seem neither individually nor jointly sufficient to ground worshipfulness” (ibid., p. 307). Take omniscience for example; if God is omniscient then he has much greater knowledge than we do, but we do not generally consider somebody worthy of worship simply because they know more than we do. The same is true with God’s other attributes; we do not consider somebody worthy of worship simply because they are more powerful than we are, for example.

However, as Murphy (2017, pp. 130–31) argues, in order for some being to be worthy of worship, there must be “a massive inequality” between the object of worship and the potential worshipper, where this inequality is “of a certain sort of value”. This massive inequality has to do with God being “absurdly greater than we are, greater in terms of pure perfection”; it cannot, however, be “a difference simply in the degree to which some valuable feature is realized”; instead, there must “be a sense in which the worshipped and the worshipper are not even playing the same game, not rightly placed on a common scale”. (See also Crowe 2007, p. 473.)

Murphy (2017, p. 131) emphasizes that “God’s goodness is intrinsic goodness, and ours merely goodness by participation; it is not merely that God has greater goodness than we do, but that ours is a meager and reflected type”. The same is true of all God’s attributes. For example, the intrinsic nature of divine knowledge is such that, whereas we gain knowledge by learning things, God knows everything immediately, without the need to learn.13

As well as the intrinsic nature of God’s properties, Murphy also points out that our goodness is a participation in, or reflection of, God’s goodness. If we want an analogy, we could perhaps think about the light of the sun and that of the moon; the point is not merely that the sun is much brighter than the moon, but rather that the moon has no brightness of its own but simply reflects the sun’s light.

Bayne and Nagasawa are therefore right that we should not consider somebody worthy of worship simply because they are more knowledgeable, powerful, etc., than we are, but they are mistaken in thinking that the attributes in virtue of which God is worthy of worship are shared by humans to a lesser degree. God’s power and knowledge is qualitatively different from our power and knowledge. It is highly plausible, therefore, to claim that if some being has certain attributes in virtue of which that being is God, then that being is worthy of worship in virtue of those same attributes. I think Murphy (2017, p. 132) is therefore right to claim both that “to be worthy of worship entails being absolutely perfect”, and that “being absolutely perfect entails being worthy of worship”.14

As mentioned earlier, Rachels sees worship as an acknowledgement that we are God’s subjects, and that he has an unqualified claim on our obedience or allegiance. Clark also links worship with obedience; immediately after the sentence from Clark quoted above, he continues: “When Elijah asked the people of Israel whether Baal or the LORD were to be their God, he was asking which they would obey and worship” (Clark 1982, p. 341). The implication is that, if any being is worthy of our worship, then that being is also worthy of our allegiance. I think that Rachels is correct in thinking that God, if he exists, must necessarily be a fitting object of worship; but is he also correct in thinking that this is an acknowledgement that he necessarily “has unlimited authority, and an unlimited claim on one’s allegiance” (Rachels 1971, p. 333)?

Murphy does not think so. According to Murphy:

Worshipfulness requires that one think of God as unsurpassably great and that one have dispositions, attitudes, and desires with respect to God that are fitting dispositions, attitudes and desires to have with respect to one that is unsurpassably great. But we cannot say that God’s being unsurpassably great entails that God has authority.

(Murphy 2002, p. 136)

Elsewhere, Murphy (2017, p. 132) suggests that “typical of appropriate worship of God are actions that express glorification of God, praise of God and our self-abasement in relation to God”. Clearly, though, Murphy thinks this stops short of expressing allegiance or submission to God.

I think this is a mistake. It seems to me that we cannot sincerely worship God whilst living our lives in ways that incur his disapproval. It is difficult to conceive how somebody can genuinely express glorification of God, for example, whilst remaining unmoved by what is important to God. As Isaiah suggests, worship that involves rituals but does not involve living the type of lifestyle that brings honour to God is a sham (Isaiah 29:13); and when Jesus quotes this passage, he makes it clear that the type of lifestyle that honours God involves submission to God’s commands (Mark 7:6–8). Worship involves being deferential towards God, which includes being deferential to his attitudes and commands. I think Rachels is therefore right to say that the ritual aspects of worship signify something more important; we worship God by living our lives in ways that honour him, such as by looking after the vulnerable (cf. James 1:27)—or even, pace Murphy (2017, p. 133), by keeping a promise to referee a very complex journal submission.

I think, then, that the sort of “dispositions, attitudes, and desires” that are fitting to have with respect to God include the disposition to submit to him, the desire to act in ways that he approves of and the attitude that we will put him first in our lives.

Now, Murphy might grant, at least for the sake of argument15, that if one has the sort of dispositions, attitudes and desires which it is fitting to have towards one who is unsurpassably great, then “one will submit oneself entirely to God’s governance, and will thus be under divine authority” (Murphy 2002, p. 137). But his suggestion now is not that in worshipping God we are recognizing that we are already under divine authority, but that in worshipping God we are putting ourselves under divine authority. Even if we have a moral obligation to worship God, thereby putting ourselves under God’s authority, we would still not be subject to God’s authority if we refuse to fulfil our obligation to worship him—just as somebody with a moral obligation to join the army does not come under the authority of a commanding officer if he or she refuses to join the army.

I agree that in worshipping God we are in effect agreeing to submit to him, but this in itself does not tell us whether or not we are already under his authority. When a boy joins the Boy Scouts in the UK, he promises, among other things, to do his duty to the King and to keep the Scout Law. Whatever duty he has to the King presumably precedes his promise to fulfil it, whereas his duty to keep the Scout Law is created by his promise. So the question is whether, as Rachels (1971, p. 334) says, worshipping God is “a way of recognizing him as [already] having an unqualified claim on our obedience”, or whether it is a way of giving him an unqualified claim on our obedience. I think it is the former, but in order to show this I need to say more on why we should think God is authoritative.

6. God’s Authority

According to Murphy (2002, p. 18), the thesis that God has authority over all created rational beings says, in its strongest modality, that “Necessarily, if A is a created rational being, then God has authority over A”; and Murphy interprets this as:

Necessarily, if A is a created rational being and φ-ing is an action not otherwise excluded, then God’s command that A φ constitutively actualizes a decisive reason for A to φ.

(ibid.)16

Murphy insists that the commands would have to “constitutively actualize” a reason; in other words, God’s command must be the reason (or at least part of the reason) why we should do what is commanded. If God simply commands us to do what we already have a decisive reason to do, perhaps because it is antecedently obligatory, and his command does not give us any additional reason to do it, then our reason is not constitutively actualized by his command but by the antecedent obligation. Similarly, if God commands us to do something which we previously had no reason to do, and the only reason we have to do what he says is a fear that he will punish us if we disobey, then our reason is not constitutively actualized by his command but by the fear of punishment. In this case, God’s command might cause us to have a reason which we previously did not have, but it would not constitutively actualize the reason (Murphy 2002, pp. 15–16).

I take Murphy’s account to be a specific case of a fairly standard account of authority generally, whether we are talking about divine or human authority; hence he makes a general comment that “to be under another’s authority is for that other’s dictates to constitute decisive reasons for action for one” (ibid., p. 126). Similarly, Raz (1979, p. 12) defends what he calls “the simple analysis” of authority, according to which “X has authority over Y if his saying, ‘let Y φ’, is a reason for Y to φ”. For Raz, then, authority is a “normative power” (ibid., p. 7). The concept of authority as the ability to impose duties on others is, according to Christiano (2012, sct. 1.1), held by most political theorists.17

I think a couple of clarifications are in order. Firstly, as already indicated, it might be God’s attitudes of approval and disapproval, rather than his commands, that constitutively actualize reasons for us. Secondly, although God’s disapproval of a particular action may well be part of our reason for avoiding the action in question, there are presumably non-moral features of the action which incur God’s disapproval (for example, it might cause unnecessary suffering), and God may well approve of us being motivated directly by those non-moral features rather than solely by his disapproval (and may disapprove of us not being motivated directly by those features). These features may be a reason for us to avoid the action in question.18 Nevertheless, it is God’s disapproval which explains why these features are morally relevant.

Murphy discusses several arguments that might be used to support the authority thesis, of which the most persuasive, in my view, is the idea that practical authority (or something like it) is a divine perfection—that is, something without which God would not be such that none greater can be conceived.

One might think this is intuitively the case; as Murphy (2002, p. 60) puts it, “if we think of two beings arrayed in the traditional divine perfections, yet one of these beings possesses authority whilst the other lacks it, we may suppose that the latter being falls short with respect to perfection”. Yet Murphy rejects this idea.

Practical authority cannot be a divine perfection, according to Murphy, because every divine perfection, at least if it is capable of degree,19 must have an intrinsic maximum and must be possessed by God to the maximum degree. Power, for example, is clearly capable of degree, in the sense that it is possible for somebody to have a greater or a lesser degree of power, so in order for power to be a divine perfection God must have power to the maximum degree possible. If it were possible that God could have had more power than he does in fact have, then a being greater than God could be conceived—namely one identical to God in every respect except for being more powerful. In order for God to be maximally great, he must be “maximally excellent in every possible world with respect to every perfection” (ibid., p. 60).20

But Murphy insists that practical authority has no intrinsic maximum. This is because the amount of practical authority God has depends on how many beings there are in the world over which God could have authority. Suppose that God has authority over all created rational agents, and suppose that the number of created rational agents is n; then God has authority over n rational agents. But it is logically possible for God to have created n+1 rational agents, in which case God would have authority over n+1 rational agents. However many created rational agents there are, it is always possible that there could have been more, and so however much practical authority God has, it is always possible that he could have had more. It follows that practical authority has no intrinsic maximum, and therefore cannot be a divine perfection.

Furthermore, if God had not created anything at all, he would not have had any authority at all, which means that God’s authority must be contingent—whereas God’s divine perfections are possessed essentially.

An obvious response to Murphy is that, although “being authoritative” might not itself be a divine perfection, “being such that if there were any rational creatures, then [God] would be authoritative over them” might be a divine perfection; but Murphy finds this response unpersuasive (Murphy 2017, p. 140). This is somewhat puzzling, though, given that he accepts a similar idea in a different context—that of being creator.

It is a central belief about God that he is creator. The Nicene Creed describes him as “maker of heaven and earth, of all that is, seen or unseen”. Yet it seems impossible that “being creator” could be one of the divine perfections. The problem is that (at least on most views) God is not necessarily a creator. God was not logically compelled to create anything at all, and if he had chosen not to create anything he would not have been a creator. But if “being creator” is a divine perfection, then it would have been logically impossible for God to fail to be a creator; a divine perfection is something that God possesses essentially. Hence, “being creator” cannot be a divine perfection.

But although “being creator” cannot be a divine perfection, Murphy (2011, pp. 8–10) thinks that “being creator of all else that exists” is a divine perfection. Being creator of all else that exists is consistent with nothing else existing. What it is incompatible with is the existence of anything else which has not been created by God. If “being creator of all else that exists” is a divine perfection (and I agree with Murphy that it is), then God is necessarily “maker . . . of all that is, seen or unseen” (so long as this is understood de dicto), even though he is only contingently “maker of heaven and earth”.21

Now if, as seems plausible, we should distinguish between “being creator” and “being creator of all else that exists”, then I see no reason why we shouldn’t also distinguish between “being authoritative” and “being authoritative over all created rational beings”. The latter is consistent with nothing else existing, but it is incompatible with the existence of any created rational being that is not under God’s authority.

I think then that Murphy’s argument fails. For it to be necessarily true that, in a world containing created rational beings, God has authority over them, it is not necessary for authority to be a divine perfection; it is sufficient that “being authoritative over all created rational beings” should be a divine perfection. But although Murphy’s argument fails, this does not prove that “being authoritative over all created rational beings” is a divine perfection. So is there anything I can say to support this idea?

Murphy (2011, pp. 10–11) thinks that “sovereignty” is a divine perfection and that being sovereign does not just entail, but includes, being creator of all else that exists. I think it is at least plausible that sovereignty also includes being authoritative over all else that exists. Murphy defines “sovereignty” as follows:

For some being to be sovereign over a domain is for things in that domain to be dependent on that being for their existence/actuality and to have their character be controlled by facts about that being.

(ibid.)

Now, clearly this is a theological definition of sovereignty; when we talk about human sovereignty, we tend to refer to a person or group of people ruling over those who live in a particular territory. Those living in the territory do not depend on the sovereign for their existence or have their characters controlled by facts about the sovereign. “Sovereign” in the more usual sense of the word tends to be associated with having authority over others. This need not, of course, mean total authority, in every area of life, but sovereigns are traditionally understood to have the authority to make laws that govern particular areas of life. So if sovereignty is a divine perfection, the question is whether the concept of sovereignty as involving authority over created rational beings is included in the concept of sovereignty in Murphy’s sense.

It seems to me that being sovereign in this sense is included in being sovereign in Murphy’s sense. Assuming the domain over which God is sovereign in Murphy’s sense includes everything that exists, then moral standards must be included in the things over which God is sovereign in Murphy’s sense. Hence, moral standards must be included in the things which are dependent on God for their existence and which have their character controlled by facts about God (cf. Wainwright 2005, p. 127). But if this is the case, then it looks as if God is necessarily authoritative over all created rational beings.

It would certainly strike me as odd if a human government has authority over the inhabitants of a nation but God does not have authority over the inhabitants of the world. Of course, political anarchists and political obligation theorists disagree on whether human governments do have authority over citizens. But even if they do not, the lack of authority on God’s part would still seem to imply some deficiency in God—and if they do, it would seem to imply that there is something lacking in God which is not lacking in some humans. If, as Raz (1979, p. 7) thinks, authority is a normative power, then there would seem to be some power that God lacks, and this would seem especially puzzling if it is true that some humans do have this power.

Suppose, though, that the very concept of imposing duties by command is incoherent. If it is conceptually impossible for any being to have this ability, then even God could not have it; God would not then have authority over created rational beings. But God would still be such that none greater could be conceived because no other possible being could have this authority either. Although Murphy himself clearly does not think there is anything incoherent about God having authority (Murphy 2002, p. 7), somebody else might raise this objection.

But I think it very unlikely that there is anything incoherent about having authority over others. Perhaps parents, as a matter of fact, do not have any authority over their children, but this does not mean that those who think they do are conceptually confused. Or consider this: many nations have tax laws which appear to obligate employers to take a proportion of their employees’ salaries to pay to the state. Now, it would usually be considered morally wrong for employers to give part of their employees’ salaries to a third party, so it looks as if the state is telling them to do something which would otherwise be morally wrong. Whether or not employers are justified in doing this, then, seems to depend on whether or not the state has authority. Perhaps the state does not have authority; perhaps employers are in fact guilty of stealing and perhaps employees would even be morally justified in using reasonable force to prevent this theft. My point is simply that it does not seem incoherent to think that the state has authority, and that employers are morally justified in doing what would otherwise have been wrong.22

If authority over all created rational beings is a normative power, and there is nothing incoherent about it, then God’s omnipotence would seem to be undermined if God does not have this power. Murphy (2002, pp. 48–50) suggests, however, that God’s omnipotence might be undermined if God does have this normative power. There may be some good involved in willingly submitting to God, and so God may choose to create a world where rational agents willingly submit to God’s authority rather than a world where rational agents are already under God’s authority. But, as my Boy Scout analogy demonstrates, there seems no reason why we cannot willingly submit to pre-existing authority. Alternatively, perhaps we are obligated by God’s commands but can willingly take on the obligation to also do whatever God requests. If God is necessarily authoritative over all created rational beings, then it is true that he lacks the power to create rational beings who are not under his authority; but this is no more problematic than the fact that he lacks the power to create rational beings who have totally private thoughts, if he is omniscient. Many theists think there is some good involved in telling God what we are thinking; this is certainly not precluded by the fact that God already knows this. Similarly, willingly submitting to God’s authority is not precluded by the fact that we are already under his authority.

I think therefore we should conclude that being authoritative over all created rational beings is a divine perfection. This means that no being who is not authoritative over all created rational beings could be God. To say that some action is prohibited by God is already to say that it is prohibited by a legitimate authority, and there is therefore no need to insist that we must have a prior moral obligation to obey God (cf. Wainwright 2005, p. 136). To ask, on learning that some action is prohibited by God, why we should obey God makes no more sense than to ask, on learning that some action is wrong, why we should avoid wrong actions.

Given that God is authoritative over all created rational beings, the moral queerness problem is resolved. As Mackie (1977, pp. 230–31) himself says, the “objectively prescriptive element” of morality would be “quite non-mysterious” if the type of theory I am defending is true, since morality would derive from the commands (or attitudes) of “an identifiable authority”.

7. The Existence of God

Now, it might be objected that my argument fails because God does not exist. It might be thought that the existence of God is itself just as queer, if not more so, than non-theistic moral facts.

The first thing to note is that, even if God does not exist, this would not affect my argument at all. Atheists as well as theists can agree (as indeed Mackie does) that moral realism is, or would be, more plausible if God exists than if he does not. If the existence of God is unlikely, there is always the option of adopting a moral error theory. My argument is neutral on whether or not God in fact exists.

One can even agree with this claim whilst at the same time holding that atheistic moral realism is more plausible than theistic moral realism. The plausibility of theistic moral realism depends both on the plausibility of moral realism if theism is true and also on the plausibility of theism, whereas the plausibility of atheistic moral realism depends both on the plausibility of moral realism if atheism is true and also on the plausibility of atheism. So even if theism would make moral realism substantially more plausible than it would otherwise be, if theism itself is extremely implausible then atheistic moral realism could still end up more plausible than theistic moral realism (although it might still be less plausible than moral error theory).

For example, Sterba (2024, pp. 1–3) argues that the existence of God is logically impossible, because it is incompatible with all the horrendous evil in the world. But I think Sterba is wrong. Briefly, his argument relies on a premise that an all-good, all powerful God must, necessarily, adhere to three “necessary moral requirements”, which he refers to as “exceptionless minimal components of the Pauline Principle never to do evil that good may come of it” (ibid., p. 1). But it is not clear to me that Sterba’s three requirements are in fact necessary moral requirements that an all-good, all powerful God must adhere to.

Two of these requirements relate to the necessity of preventing horrendous evil consequences of immoral actions (so long as certain criteria are met). But I think there is a big difference between performing immoral actions that lead to horrendous evil consequences in order to attain some great good, and failing to prevent horrendous evil consequences of others people’s immoral actions in order to attain some good. It may be that it is always wrong to do evil that good may come of it, but only sometimes wrong to fail to prevent evil that good may come of it (even when Sterba’s criteria are met).

In fact, even Sterba does not think that an all-good, all-powerful God should prevent all evil; it is only horrendous evil that he thinks he must necessarily prevent (ibid., p. 3). But once one concedes that such a God might fail to prevent “significant” evil it becomes difficult to see how it is logically incompatible for such a God to fail to prevent “horrendous” evil.

One might argue that God would need much greater reasons to justify failing to prevent horrendous evil than those needed to justify failing to prevent significant evil; but even if this is so, it does not follow that we would know, or even be able to know, what those reasons might be. M. M. Adams (1989, p. 305) thinks these reasons are such that we are “cognitively, emotionally, and/or spiritually too immature to fathom”.

Furthermore, Adams (ibid., p. 307) suggests ways in which God might not merely “engulf” horrendous evils (the “beatific face-to-face intimacy with God” will make the victim’s life worthwhile on the whole), but also “defeat” them, by “integrating participation in horrendous evils into a person’s relationship with God”. But even if her Christian Universalism is false, and some wicked people will never experience the beatific intimacy with God, I do not think it follows that theism would be logically incompatible with the existence of horrendous evil.

So, regardless as to whether or not the existence of horrendous evil makes God’s existence unlikely, I do not think it proves that God’s existence is logically impossible. My own view is that theism is not just logically possible but is actually true, and since I also believe that theism would make moral realism substantially more plausible than it would otherwise be, I think that theistic moral realism is substantially more plausible than atheistic moral realism. Furthermore, I agree with Mackie (1982, pp. 115–16) that moral realism, if true, would provide evidence for God’s existence. Nevertheless, my aim has not been to argue that God exists, but to argue for the more modest claim that moral realism is more plausible if God exists than if he does not.

8. Concluding Comments

I have argued that morality seems queer because it seems queer that a mindless universe could make demands on us. There is nothing queer, however, about God making demands on us, and I have therefore concluded that, since God is necessarily authoritative over us, the queerness problem is resolved if morality derives from God in some way—such as being identified with his commands or his attitudes. This explains the relevance of God to metaethics. The authority of morality has its source in the authority of God.

I want to close with a few final comments.

A common misunderstanding of the type of theistic ethics that I have discussed in this paper is that it implies that those who do not believe in God cannot be moral. But I am not aware of any Divine Command Theorist who believes this, and nor should they.23 Both theists and atheists have moral convictions and attitudes that can, to some extent, reflect God’s own attitudes, so that they “do by nature things required by the law”, thus showing that God’s laws are “written on their hearts” (Romans 2:14–15). If moral realism is true then even moral error theorists can do what is morally required; and in exactly the same way, if Divine Command Theory is true then even atheists can do what is divinely required. And just like theists and moral realists, atheists and moral error theorists can act on genuinely moral motives.

Even if it turns out that there is no God and no objective morality, it does not follow that we would suddenly become horrible people. If, as Frans de Waal (2006) for example thinks, we have evolved to be sympathetic to others then presumably we might go on being sympathetic to others. Moral error theorists do not typically abandon their concern for others. But human beings are also very good at causing an enormous amount of unnecessary suffering, and if moral realism is discovered to be false then, however much we might condemn those who do this, we presumably cannot consistently say that what they are doing is objectively wrong.

Finally, I have not argued that moral realism is impossible if God does not exist. We may well be able to have objective morality without God’s existence. But I think the queerness argument shows that moral realism is more plausible if God exists than if he does not.