Sterba on Divine Commands and Fairness

Abstract

James Sterba has recently argued against Divine Command Theory (DCT). Sterba also offers, as a preferable alternative to DCT, a metaethical account which he has developed over a number of years (culminating in Sterba’s 2013 book), which attempts to ground ethics in rationality. Both his case against DCT and his alternative metaethical theory are worth considering, but in this paper I will argue that both fail. This is not of course to say that DCT is either true or the most promising account of metaethics, or that the project of trying to ground ethics on rationality is a dead-end. I do conclude, however, that Sterba’s own unique approach, which involves trying to derive morality from rationality by appeal to a non-question-begging requirement, is doomed to failure, and also that his new arguments against DCT are ineffective.

1. Introduction

James Sterba’s recent work is far reaching and I have addressed some of it elsewhere (Molto 2024, 2022). In the current paper, I will restrict myself to three topics. The first thing I will do in this paper will be to argue that Sterba has failed to adequately defend a key component of his own metaethical account, namely that the most basic of all moral norms, the principle of fairness, is derivable from a norm of rationality, namely the principle of non-question-begging. Indeed, I will argue that this key claim is false and that therefore Sterba’s novel theory of metaethics fails. The second thing I will do in this paper will be to evaluate Sterba’s arguments against Divine Command Theory (DCT). I will try to show that these arguments against DCT also fail, because Sterba does not provide decisive rebuttals to well-known responses that have been offered against arguments of the sort he presents. The final thing I will do in this paper is to briefly explore a further objection to DCT, which is implicit in some of Sterba’s comments about DCT, but which he does not systematically press. I will argue that this is the most powerful objection, but that it at most shows that DCT is an unpromising theory of metaethics for those not already committed to the existence of the classical Abrahamic God (henceforth, just ‘God’).

A great deal is going on in Sterba’s most recent contribution to this journal (Sterba 2024). In it, we find a restatement of his well-known argument against the existence of God, further arguments against DCT which do not assume that his earlier argument succeeds, a rehearsal of his own metaethical account and arguments intended to show that this account fits nicely with evolutionary theory. One result of the density of issues at play here is that Sterba’s intended goal is left somewhat opaque. For example, it is not entirely clear if Sterba’s central concern in the first half of his paper is to undermine moral arguments for the existence of God, or to disprove any possible metaethics which would ground morality in divine commands (that is, DCT narrowly construed), or to disprove any possible metaethics which depends on God in any capacity (which we might term ‘DCT’ in a broader usage of that term), or whether the discussion of DCT is entirely aimed at problematizing a possible alternative to his own account. In any case, there is too much here for me to adequately address all of these issues. I will say nothing about the logical problem of evil here and nothing about evolution. I intend to consider only Sterba’s positive account and his new arguments against DCT, by which I mean those that do not involve the problem of evil. With respect to the former, with which we will begin, I will also concern myself only with one part of it, the one I consider to be most original and interesting. The component of Sterba’s own account that interests me is the claimed entailment between the principle of fairness and the principle of non-question-begging.

2. From Rationality to Fairness

Briefly, Sterba’s positive account is as follows. Sterba thinks that all of morality is ultimately based on a single norm, namely the principle of fairness:

treat all relevant interests fairly.

(Sterba 2024, p. 9)

This claim can certainly be objected to, but as this is not the most interestingly novel part of the account, I will not pause to consider objections here. What is especially novel and interesting is what comes next.

Sterba considers fairness to be a matter of adjudicating appropriately between different interests, most saliently between egoistic interests and altruistic interests. He then reasons as follows:

Now, when we rationally assess the relevant interests in conflict cases, three solutions are possible:

• Egoistic interests always have priority over conflicting altruistic interests.
• Altruistic interests always have priority over conflicting egoistic interests.
• Some kind of compromise is rationally required. In this compromise, sometimes egoistic interests have priority over altruistic interests, and sometimes altruistic interests
• would have priority over egoistic interests.

Once the conflict is described in this manner, the third solution can be seen to be the one that is rationally required. This is because the first and second solutions give exclusive priority to one class of interests over the other, and only a question-begging justification can be given for such an exclusive priority. Only by employing the third solution—sometimes giving priority to egoistic interests and sometimes giving priority to altruistic interests—can we avoid a question-begging resolution.

(Sterba 2024, p. 10)

From this, Sterba concludes, that the principle of fairness is rationally obligatory, assuming a commitment to the principle of non-question-begging. Begging the question is of course a fallacy and so, in general, it is not rational to arrive at conclusions by begging the question. The upshot is that, according to Sterba, morality can be grounded in principles of rationality.

There is much to be said for this proposal, and the ideal of grounding morality in rationality holds genuine appeal as well as a long history, but nevertheless I will argue that one of the links in this account breaks down. Specifically, I think the argument presented in the passage above fails and, moreover, that not only has no good reason been presented by Sterba for thinking that the principle of fairness can be derived from the principle of non-question-begging, but in fact, there are good reasons for thinking that it cannot be so derived.

Sterba, in the passage quoted above, has reasoned that “…only a question-begging justification can be given for such an exclusive priority <of one set of interests over the other>” (Sterba 2024, p. 10). But why should this be the case? Why should we think that the only route to an exclusive priority of interests which conflicts with the principle of fairness is by way of begging the question? In his earlier, and more thorough, treatment of this topic, Sterba gives his reasons in more detail. Sterba divides his argument into two separate moves. The first move is as follows.

Noting that both egoism, which favors exclusively self-interested reasons, and altruism, which favors exclusively altruistic reasons, offer only question-begging starting points, we took as our non-question-begging starting point the prima facie relevance of both self-interested and altruistic reasons to rational choice. The logical inference here is somewhat analogous to the inference of equal probability sanctioned in decision theory when we have no evidence that one alternative is more likely than another. Here we have no non-question-begging justification for excluding either self-interested or altruistic reasons as relevant to rational choice, so we accept both kinds of reasons as prima facie relevant to rational choice.

(Sterba 2013, pp. 45–46)

Thus, according to Sterba, we may infer from the fact that egoism and altruism do not have non-question-begging arguments in their favour, the only non-question begging starting point is to accept that both egoistic and altruistic reasons are prima facie relevant to rational choice.

The second move is as follows:

Once, however, both self-interested and altruistic reasons are recognized as prima facie relevant to rational choice, the second step of the argument for the compromise view offers a nonarbitrary ordering of those reasons on the basis of the rankings of self-interested and altruistic reasons imposed by the egoistic and altruistic perspectives, respectively. According to that ordering, high-ranking self-interested reasons have priority over conflicting low-ranking altruistic reasons, other things being equal, and high-ranking altruistic reasons have priority over conflicting low-ranking self-interested reasons, other things being equal. There is no other plausible nonarbitrary ordering of these reasons.

(Sterba 2013, p. 46)

I will take issue with both of these moves. With respect to the first move, it does not follow from the premise that the only arguments in favour of egoism and altruism are question-begging arguments to the conclusion that we either should or must recognize both egoistic and altruistic reasons as prima facie relevant. The analogy with decision theory is insightful here, but not helpful for Sterba’s case, I think. It is true that, according to the principle of indifference, if we are evaluating two mutually exclusive and (as far as we know) exhaustive options, for which we have no evidence either for or against, then via the principle of indifference, we should evaluate each as having a 0.5 probability. However, quite apart from the fact that the principle of indifference is controversial (opposition to the principle stretches at least as far back as John Maynard Keynes (1921), who first named the principle), it also does not support Sterba’s conclusion in the current case.

The point of the principle of indifference is that, if it seems to us that there are only two options and we have no rational reason to prefer one over the other, then we should assign them the same likelihood. Of course the same goes if we have three or more options and no rational reason to evaluate any as more probably than any other. The problem with the analogy to the case at hand is that the options that Sterba has presented are not exhaustive. We have the two extremes of egoism and altruism and the compromise position that Sterba is endorsing, which he has characterized as allowing “the prima facie relevance of both self-interested and altruistic reasons to rational choice”. But if each of these occupies a place in logical space, then so does the inverse of Sterba’s compromise position: that neither self-interested nor altruistic reasons are relevant to rational choice.

Admittedly, we might have difficulty in entertaining the possibility that our own interests fail to give us relevant rational reasons, but we must accept that it is possible that they do not. Moreover, if we were to appeal to the incorrigibility of some sort of egoism as a reason to rule out this possibility, then we would have to admit that the same incorrigibility would speak in favour of egoism over altruism as well. If Sterba is right that we cannot have compelling non-question-begging arguments in favour of a general priority of egoistic interests over altruistic interests, then I do not see what sort of compelling non-question-begging arguments we could have for egoistic interests being relevant to rational choice at all. The same goes for altruism. If all of our arguments for altruism over egoism are necessarily question-begging, then I would expect that all of our arguments for the significance of altruism to rational choice would also be question-begging. The upshot is that the appeal to the principle of indifference is undermined. In other words, if we are setting aside any unargued preference we have for one possible position over another, then we must allow equal weighting to all logically possible positions, and this includes the position on which neither egoistic nor altruistic interests are relevant to rational choice.

Thus, it seems to me, that any attempt to compromise between egoism and altruism is just begging the question against a position which denies that either egoistic or altruistic reasons are relevant. An analogy with the principle of indifference in this case would have to find a way to compromise between these three positions and it is difficult to imagine what that would look like, certainly it does not involve simply granting prima facie relevance to both categories of reasons. Thus, the first move does not work in my opinion.1

The second move in the argument fares no better, I believe. This move takes us from recognizing the prima facie relevance of both egoistic and altruistic reasons to a particular compromise between them, which Sterba claims is non-arbitrary and leads to the principle of fairness. Now I accept that the compromise that Sterba offers is non-arbitrary, but I deny that it is the only non-arbitrary compromise between egoism and altruism available.

The problem that Sterba is confronting us with is how to adjudicate between two categories of reasons when we agree that both categories matter, but when we do not know which matters more and we cannot come to know this in the future. Certainly one thing we could do is to adopt a policy of fairness, balancing the two categories on the assumption of their equal relevance. But the fact that this policy is fair cannot serve as an independent reason for adopting it at this stage, because Sterba’s claim is that rationality leads to fairness, so the argument cannot presuppose that being fair is part of being rational. I hold that fairness makes more onerous demands of us than rationality does and that, therefore, there are rational policies that are unfair. If I am right about this, then the second move in Sterba’s argument will not work. To see whether I am right, we will need a bit more detail.

Sterba introduces the possibility of ranking reasons within each category. He then claims that rationality demands that we not privilege the lower-ranking reasons of one category over the higher-ranking reasons of the other category (Sterba 2013, p. 46). I think it is true, at least arguably, that if we have assumed that both categories of reasons are relevant but that we do not know how relevant each category is in comparison with the other category, then it would be irrational to privilege lower-ranking reasons of one category over higher-ranked reasons of the other. This is also in accordance with fairness. So far, rationality does demand fairness. However, Sterba goes on:

Of course, there will be cases in which the only way to avoid being required to do what is contrary to your highest-ranking interests is by requiring someone else to do what is contrary to his or her highest-ranking interests. Some of these cases will be so-called lifeboat cases, as when two individuals are stranded in a lifeboat that has only enough resources for one to survive. Although such cases are surely difficult to resolve (maybe only a chance mechanism, like flipping a coin, can offer a reasonable resolution), they surely do not reflect the typical conflict between our relevant egoistic and altruistic interests. Typically, one or the other of the conflicting interests would rank significantly higher on its respective scale, thus permitting a clear resolution.

(Sterba 2013, pp. 44–45)

This is where I believe rationality and fairness can come apart. Whatever typical conflicts between egoistic and altruistic reasons might look like, it is clearly the case that we can and do adjudicate between egoistic and altruistic reasons which are, at least plausibly, of the same ranking. These are not all lifeboat cases. There are life and death situations, where we must choose whether to preserve my own life at the expense of someone else’s. But there are also lower-ranked conflicts, for example, when I must adjudicate my own financial interests over someone else’s financial interests. Then there are utterly trivial cases, where I must adjudicate whether I should prefer my own interests in having the last piece of pie over my sibling’s interest in having it. In fact, these sorts of conflicts are an essential part of our moral landscape and clearly contribute to whether or not we identify a policy as fair or unfair.

The trouble for Sterba’s position is that, though fairness might make demands of us concerning how we adjudicate between reasons in same-ranking cases, rationality does not. Here is an example of a policy which is rational but not fair.

The Unfair Compromise: we should flip a coin now in favour of either responding to egoistic or responding to altruistic reasons in the case of same-ranking conflicts and abide by the outcome forever, or until more information about the values of the different reasons comes to light. Let a random decision-making process determine who should be systematically advantaged.

The Unfair Compromise represents a rational compromise between egoism and altruism, given the assumption that both egoistic and altruistic reasons matter prima facie and without any non-question-begging argumentative resources for determining which of these interests matters more. It is a non-arbitrary compromise in that it does not pick one side of the debate over the other for advantage, because each category has equal chance of coming out systematically advantaged.

If there are doubts about the rationality of this policy, consider the following analogy. Imagine you are in a gameshow where you are confronted by a series of choices between two prizes taken from two categories. You are given an ordering of their value: A1 is more valuable than A2, which in turn is more valuable than A3 and B1 is more valuable than B2, which is more valuable than B3. You have no information about the comparative value of the As and the Bs. You might reason that, confronted by a choice between A1 and B3, you should take A1.2 But what should you do when confronted with the choices between A1 and B1, A2 and B2, or A3 and B3? Even setting aside the arbitrary options of preferring the As or the Bs without any reason for doing so, reason allows you at least these two options: alternate between the categories or flip a coin to determine which category to prefer. Both are rational, the second of these, I suggest, is unfair (or, if you prefer, leads to unfairness).

But is The Unfair Compromise actually unfair? As Sterba points out, these decisions are difficult. We are not always inclined to think that someone who seeks to preserve their life at the expense of someone else’s has acted immorally. The point though is that morality should have something to say about such issues. It is not neutral and it certainly is not neutral about systematically privileging some interests over others. I believe that it is not fair or moral to privilege one’s own interests over the interests of others in every case where the interests are equal, even if we do so on the basis of flipping a coin between egoistic and altruistic interests. If I am right about this, then fairness is something over and above certain applications of the principles of rationality. Thus, if I am right, even if we grant that there are no non-question-begging arguments in favour of either egoism or altruism, and we grant that both offer prima facie reasons for action, still reason alone does not compel us towards the principle of fairness.

3. Divine Command Theory

So much for Sterba’s positive account of metaethics. We turn now to the negative component of Sterba’s recent work, his arguments against DCT. A curious feature of Sterba’s case against DCT is just how broadly he seems to be interpreting that the theory. DCT is usually understood as the thesis that moral norms are grounded in divine commands. Sterba does characterize DCT this way at a number of points, including in the paper’s abstract (Sterba 2024, p. 1) but at the same time there are some suggestions that his concern is really with any meta-ethical theories that involve God in any capacity, rather than merely with DCT in the narrow sense. For one thing, he uses Thomas Aquinas as an example of the sort of view he is targeting. Aquinas’s metaethics is only a version of DCT if that doctrine is interpreted very broadly.

In any case, Sterba believes he can show that DCT is false several times over. First, Divine Command Theory entails the existence of a God (i.e., the God of classical monotheism), but Sterba believes he can prove that no such God exists. Second, even if the argument for atheism fails, Sterba believes he can show that no such God could play the metaethical role the DCT requires God to play. We shall not address the first of these issues, as Sterba’s argument for atheism has been examined thoroughly (see the 2021 and 2022 special issues of this journal in particular). What is new is the second attack, that even if there is a God, DCT must be false. I will lay out these arguments, for there are several, and then argue that they are unproved.

Even in the sections of the paper devoted to establishing this thesis, we find several different argumentative strands, including several distinct considerations against DCT, several possible responses on behalf of DCT, and further objections to these responses. To simplify, the main objections seem to be as follows:

• Objection 1. Humans, if DCT were true, would be “heavily circumscribed” in their role of interpreting divine commands, being unable to test established divine commands against any external standard (Sterba 2024, pp. 4–5).
• Objection 2. Divine commands can come into conflict, in which case humans would have no way to resolve this conflict (Sterba 2024, p. 5).
• Objection 3. DCT claims that God’s commands are conveyed by instances of special revelation, but as people can only be bound by laws that they know of and have reason to accept, then anyone unfamiliar with special revelation will not be bound by moral norms (Sterba 2024, p. 6).
• Objection 4. God could command anything at all and therefore anything at all could have been morally obligatory (Sterba 2024, p. 6).

Presumably Sterba thinks each of Objections 1–4 are problematic, but I am not sure that Sterba considers any of these objections to be new or considers any of them to be knock-down on their own. He does, though, seem to believe that they collectively disprove DCT and he also says that the fourth objection is probably the strongest.

In brief, I do not think that objections 1–4 individually or collectively disprove DCT in either the narrow or broad sense of ‘DCT’. I think these objections, at most, suggest that some auxiliary hypothesis must be added to DCT to ward off the objections. But the sorts of auxiliary hypothesis that could do this work are reasonably well explored and not defeated by anything offered by Sterba in his article. We can see this be briefly taking the objections in turn.

Objection 1 is a bit of an odd one in that it is not totally clear what the objection is, or how it militates against DCT. Sterba introduces this objection by making an analogy between God, according to DCT, and a human legislator. His concern seems to be that, in contrast with human legislators, whose power to legislate is constrained by a human judiciary which can judge whether the resulting legislation accords with a higher constitutional legislation, no such human oversight exists vis-à-vis God’s legislation of morality according to DCT. I have to admit, I am not sure why that should matter. DCT is hardly the only realist moral theory according to which humans lack the power to overturn moral norms. I suspect Sterba does not think much rides on this objection to DCT and we shall move right on to potentially more decisive objections.

Objections 2, 3, 4 are all clearer. However, I suggest, possible responses are salient in each case. Objection 3 is perhaps problematic for versions of DCT that are accompanied by the claim that the contents of divine commands are solely accessible through special revelation. Certainly there are theological traditions that would seem to want to say something close to this. But there are many theological traditions that would explicitly deny this and it is clearly not entailed by DCT itself.

To give just one example of such a tradition, we may point to the Roman Catholic teaching on moral epistemology. The following is a well-known quote taken from the documents of the Second Vatican Council and quoted in the Catholic Catechism:

In the depths of his conscience, man detects a law which he does not impose upon himself, but which holds him to obedience. Always summoning him to love good and avoid evil, the voice of conscience when necessary speaks to his heart: do this, shun that. For man has in his heart a law written by God; to obey it is the very dignity of man; according to it he will be judged. Conscience is the most secret core and sanctuary of a man. There he is alone with God, Whose voice echoes in his depths.

(Gaudium et Spes, §16)

The claim here is that the moral law exists in each human heart. Importantly, it pre-exists any particular engagement with special revelation. Moreover, it is not merely religious traditions that avoid this problem by denying that moral knowledge comes primarily through special revelation, this claim can be found explicitly in the philosophical literature as well (see for example, Adams 1999, p. 364). Given this, Objection 3 clearly does not undermine DCT itself, but at most poses problems for some versions of it.

The defenders of DCT will typically respond to Objection 2 by claiming that no divine commands do conflict and, in some relevant use of the word ‘possible’, it would not have been possible for any to have done so. A similar response to Objection 4 is well-known: in some relevant sense of ‘possible’, it was not possible for God’s commands to have been fundamentally of a different kind than what God has in fact commanded. We can discuss Objections 2 and 4 together.

The relevant sense and range of ‘possible’ here is up for grabs, as is any background theory of God which explains why these things would not be possible, and there are various theories on offer here. It is sometimes admitted that it is logically possible that God could have commanded us, for example, to torture children, (for example, by Nelson Pike 1969), but there is some other use of ‘possible’ on which this is impossible. But according to other views, it is not even logically possible for God to command something evil (e.g., Wierenga 1989, pp. 204–5). Thomas Talbott (1988) even seems to hold that it is logically impossible for God to have commanded us to do evil and yet, at the same time, God could have commanded us to do evil!

The standard explanation for why it is not possible for God to command us to do evil, in whatever sense of ‘possible’, is an appeal to the essential nature of God and the relationship between this nature and what God commands. In brief, divine commands are necessarily reflections of God’s nature and cannot conflict with it, and God’s nature is essential, so God has the same nature in every possible world. Thus, God cannot give commands in any possible world which conflict with the nature God has in this world. This view comes in different varieties too. So, for example, the Classical Theist tradition generally holds that God’s will is identical with God’s nature, in virtue of the doctrine of divine simplicity, and so naturally what God chooses to command must reflect God’s nature, because they are in fact identical (for a 20th Century defence of this sort of position, see Stump and Kretzmann 1985). This view tends towards the position that all of God’s commands are necessary (again, in some but usually not all senses of ‘necessary’, see Stump and Kretzmann 1985, pp. 366–67). By contrast, views which are not committed to divine simplicity, do not have the same pressure to maintain that aspects of the divine nature are essential, still less that all God’s commands are necessary. There are various views on what God might have chosen to do differently, but it remains a widely held view that, again in at least some sense of ‘possible’, it would not have been possible for God’s commands to have been fundamentally different in their moral profile from what they in fact are.3

The important point is just that appeals to the essential nature of God are pretty standard for replying to Objection 4 and, assuming God’s nature is consistent (though see Beall 2023 for a dissenting voice on this), also Objection 2. Sterba does very briefly consider versions of DCT that involve appeals to God’s nature. This is all he has to say on the topic:

What then about grounding ethics not on the God’s commands but on God’s nature? Here I contend that the standard of goodness, and especially the standard of moral goodness, must be a norm, a requirement that one ought to act or be in a certain way. In the case of morality, I claim that the ultimate norm can be expressed as treat all relevant interests fairly. By contrast, the God of traditional theism, if he exists, would be just a concrete rational being, not an abstract norm.

(Sterba 2024, p. 6)

On the basis of a footnote in Sterba’s (2024, ft 8) paper, it seems that his main target in this passage is Robert M. Adams, whose “modified” DCT (Adams 1999) is probably the best-known version of the theory of the past century. On Adams’s account, the good is grounded in God’s nature (he describes God as serving the role of a platonic form of goodness), but moral rightness is grounded in the “commands of a supremely good and loving God”. It is essential for this grounding relation to hold that God be “supremely good and loving”, so God’s nature is built into the account of moral rightness (which is another example of appealing to God’s nature to escape the above arbitrariness objections to DCT), but moral rightness itself is grounded in divine commands, not divine nature.

Sterba seems to think that not merely rightness, but also goodness must be reduced to something normative. Now, it would require a great deal more argument than Sterba provides to establish this claim, given that there are many metaethical accounts which do claim to ground goodness on features of the universe that are not norms (moral naturalism being alive and well). Additionally, this objection only impacts a particular component of Adams’s account, which is inessential to DCT generally. An appeal to God’s nature is likely part of the best available solution to the arbitrariness objection to DCT, but this appeal in no way requires that any normative notion be grounded in God’s nature, it simply explains why (at least some of) God’s commands had, in some sense, to be what they are.

In Section 4, I want to consider one final and closely related objection to DCT, that lies in the background of Sterba’s paper.

4. Reasons-Giving Commands

I think that the best objection to DCT, or at least to some versions of it, is one that is clearly underlying some of Sterba’s reasoning but is not presented explicitly as an objection in the paper. This objection is that DCT fails as an account of metaethics because humans would not be bound by putative moral laws purely on account of those having been stipulated by a God. In other words, the objection goes, there is nothing about divine commands in themselves that makes them suitable to play the role of ultimate grounds for morality, given the special normative nature of the latter. This is closely connected with the argument from Sterba that we were discussing just now. In both cases, with respect to grounding goodness on God’s nature and grounding rightness on God’s commands, we might reasonably wonder whether we have given sufficient explanation for the essential normativity of these concepts.

Although Sterba does not consider this objection directly, he does consider and reject one possible response to this objection (Sterba 2024, pp. 5–6), which suggests that he does think that an objection along the lines just sketched is a problem for DCT. The attempted response that Sterba considers is that God is the creator of all humans and that that is sufficient for God’s commands to have binding force on each human. In this regard, Sterba argues, correctly in my opinion, that God’s having created humans would not suffice on its own to give God’s commands binding force. However, this is a moot point as few serious academic defenders of DCT are likely to be tempted by this move in the first place. Adams, for example, does not think this. Rather than to appeal God’s having created humans as the grounds of the binding force of God’s commands, Adams instead offers an account of moral obligations as arising from social relationships. On Adams view, it is necessary, though not sufficient, for our having significant moral reasons to obey God’s commands that the social bond we share with God is rightly valued, that we esteem God, and that what God commands is itself good (Adams 1999, pp. 244–46, as we have seen Adams avoids accusations of circularity by offering different grounds for goodness and rightness/obligatoriness).

Returning to the objection at hand, the worry arises because moral norms have a unique nature, which serves to restrict the possible candidates for grounding. What is this nature? Some hold, though it is controversial, that this nature is captured by one or other of the versions of internalism about morals. As David Brink expresses the general idea here, ‘Internalism is the a priori thesis that the recognition of moral facts itself either necessarily motivates or necessarily provides reasons for action’ (Brink 1984, p. 113). If one or other of these versions of internalism is true, then God’s having created humans would fail to explain this unique nature of morality, and would not serve as a promising candidate for grounding.

For the sake of argument, let us grant that it is part of the concept of a moral norm that, if there were such things as objective moral norms, they would be intrinsically reasons-giving for humans (note, I think reasons internalism is more plausible than motivational internalism, but I believe everything that follows could be said about either). By contrast with our assumption about moral norms, the concept of a divine command is such that, if there are such things as divine commands, they would not necessarily be intrinsically reasons-giving. On this basis, we might conclude that, given their reasons-giving nature, moral norms can only be grounded in other intrinsically reasons-giving features of the universe. This may seem to be a problem for DCT. I will now argue that, even if this is true, we cannot conclude from it that DCT is false.

To see why DCT might still be true, even given the conceptual difference between moral norms and divine commands, we must get a bit more clarity on what DCT is claiming. Nobody, I think, would claim that “moral norm” and “divine command” are intentionally equivalent expressions. I think any version of DCT according to which “moral norms” and “divine commands” were a priori identical would be deeply implausible. But grounding is not identity and it need not be a priori. So the question is not whether the concept of a moral norm is somehow linked with the concept of a divine command, but rather whether those concepts are somehow incompatible in a way that makes the grounding thesis implausible.

It is a popular, though not universal, view that grounding relations are necessary (see Jon Erling Litland 2015, for recent opposition). If fact P is grounded in fact Q, then fact P is grounded in fact Q in every possible world in which P and Q obtain. So, if we assume that morality must be grounded in something which is reasons-giving, then we would expect it to be essentially reasons-giving. Still, this does not mean that all divine commands must be essentially reasons-giving.4 To see this, contrast the two following questions:

Q1: Must all commands of an omnipotent being be reasons-giving for humans?

Q2: Is an omnipotent being able to command something which is intrinsically reasons-giving for humans?

I maintain that it is possible for the correct answer to the first question to be no and for the correct answer to the second question to be yes simultaneously.

Sterba no doubt thinks that the answer to Q1 is no, and I will not challenge this. But does this force a negative answer to Q2 as well? Not at all. Even if the concept of omnipotence, or for that matter the conjunction of omnipotence, omniscience, omnibenevolence and so on, does not imply that all divine commands be reasons-giving, omnipotence might still imply the possibility of God giving some commands that are reasons-giving. An omnipotent house builder is not required as a matter of conceptual truth to build only eternal houses, but it certainly is a conceptual truth that an omnipotent house builder is capable of building eternal houses.

So long as the concept of an intrinsically reasons-giving command is coherent, then it follows from the meaning of “omnipotent” that an omnipotent God could give such commands. So, in trying to answer Q2, we are asking whether or not the concept of an intrinsically reasons-giving command is coherent. I propose that we should assume that it is unless we have some reason to think that it is incoherent. I do not think it has ever been shown to be incoherent. There are, admittedly, familiar objections to the existence of intrinsically reasons-giving facts or properties, which might be thought to be equally powerful objections to the existence of intrinsically reasons-giving commands. The most famous of these of course is J.L. Mackie’s contention that reasons-giving facts or properties would be “queer” entities (Mackie 1977, pp. 38–42). But the queerness objection is not an argument for conceptual impossibility, but rather an argument that the existence of such entities in our world is implausible.

Indeed, far from using the queerness objection as ammunition against DCT, Mackie held that the existence of intrinsically reasons-giving facts (if there were such things, which he denied) would best be explained by their having been created by God:

[O]bjective intrinsically prescriptive features, supervening upon natural ones, constitute so odd a cluster of qualities and relations that they are most unlikely to have arisen in the ordinary course of events, without an all-powerful god to create them. … the more intrinsically puzzling something is, the more it requires, to explain it, something whose power is limited only by logical necessity.

(Mackie 1982, pp. 115–16)

The problem that the queerness objection possess for DCT, then, is whether the purported implausibility of such “queer” entities would undermine DCT insofar as that position is committed to the existence of such entities. Whether it does depends on what makes “queer” entities implausible in the first place. There are several ways in which Mackie’s queerness objection might be cashed out (see in particular Shepski 2008, p. 375): first, that accepting intrinsically reasons-giving norms would involve an objectionable ontological profligacy, second, that such norms are objectionable because they would be mysterious, and thirdly that intrinsically reasons-giving norms are objectionable because they would be sui generis.

Once again, none of these worries suggest that the existing of such norms is an impossibility. But more than that, to the believer in an omnipotent, omniscient God, these objections lose much of their bite. If someone already believes in the omniscient, omnipotent and clearly supernatural God of Abrahamic theism, for example, they can hardly be expected to take it as a knock against their metaethical theory, that the commands of such an entity turn out to have something mysterious or, perhaps it might be better to say, supernatural, about them. Moreover, why think it unlikely that such a being would generate commands which are sui generis? If God is omnipotent, then giving commands which are sui generis in their force is not more difficult for God than giving commands which are not sui generis. Is there any good reason for thinking God would have a preference against generating sui generis features of the universe, like reasons-giving commands? I think not. Given that, for the theist with prior commitments to both the existence of God and to moral realism, it is unlikely that positing supernatural, sui generis intrinsically reasons-giving commands would constitute an ontological profligacy either. Thus, the familiar Mackian objections to moral facts are less telling against a theistic ethics which includes an independent commitment to a supernatural being who is omnipotent and creates ex-nihilo, than against a non-theist naturalist ethics (see Yoney 2023, pp. 19–20 for similar considerations).

Of course, these responses to Mackian queerness worries are only dependent on a prior commitment to the existence of God. And this brings us back to Sterba’s intentions in attacking DCT in the first place. If his goal is to undermine potential moral arguments for the existence of God, then this purpose might be served by highlighting the fact that divine commands are not reasons-giving for humans as a matter of conceptual truth. So, for example, in response to those who would argue that DCT offers the best metaethics on the market, it can certainly be objected that it does not provide a compelling conceptual reduction of moral notions at all (a fact that many contemporary DCT theorists, like Adams, accept). Moreover, it is also true that the metaphysical commitments of DCT (first the commitment to God, second the commitment to some reasons-giving divine commands) are enormous. There are plenty of reasons why the non-believer would not find DCT independently plausible.

However, Sterba does not say he is simply trying to undermine positive arguments for DCT, or arguments for the existence of God which involve DCT as a premise. Rather, Sterba claims to be able to disprove DCT and to be able to do so quite independently of his other arguments against the existence of God. In this goal, it seems to me that Sterba fails. I do not think Sterba has demonstrated that we cannot ground goodness in a divine nature, given how terse his discussion of this point is. But even if we assume he is right about this, there is surely a version of DCT available which does not involve grounding goodness in a divine nature, but does ground rightness in a set of reasons-giving divine commands, which may or may not be a subset of divine commands generally. So long as the concept of a reasons-giving command is a coherent one, then we must grant that it is possible for an omnipotent being to give such commands. For that matter, there is a version of DCT, at least on the broad conception of DCT, which dispenses with divine commands entirely and simply posits reasons-giving moral properties, created ex nihilo by God.

5. Conclusions

I have not tried to argue for DCT in this paper. I have also not tried to argue against metaethical accounts that seek to ground morals in rationality. There are obvious and significant advantages to the latter which the former lacks, most saliently, that any such account provides an immediate answer to the question ‘why should I be moral?’ What better answer could there be than ‘because it is rational to be moral’? I have however argued that Sterba’s own account does not work. Fairness is something over and above rationality, or at least over and above the principles of rationality that Sterba has involved in his account.

As for DCT, it is not disproven. Some of Sterba’s arguments lead on to further and deeper worries, including worries about the coherence of various aspects of theology (whether the theory of divine simplicity or divine timelessness is defensible, whether an omnibenevolent God can enjoy any sort of free will, whether an omnibenevolent God can simultaneously be an omnipotent God, and many others). There are also worries about metaphysical parsimony. From the perspective of the agnostic, the theoretical benefits of metaethical accounts which ground morals in rationality compared the costs incurred by DCT in terms of metaphysical commitments, may well make the former a more promising research programme. Perhaps establishing this was Sterba’s central goal in his recent discussion of DCT. However, from the perspective of the theist, who has already paid the bulk of metaphysical costs incurred by DCT, no compelling reason has been provided to abandon it.