Moral value

Here are two key quotations.

The claim that God's ‘holy and loving nature supplies the absolute standard against which all actions are to be measured’ might naturally be construed as a move in moral epistemology, offering a recipe for settling the question whether some person or action is good. ‘If you want to know whether this person or that pattern of behaviour is good’, the recipe says, ‘then look at God and ask yourself whether this is what God is like, or whether this is the sort of thing a being with God's moral character might do’. But this isn't at all what Craig has in mind, since (as noted above) he is here interested only in moral ontology. He means to be giving an account of the way in which moral facts are ‘grounded’ in God's ‘moral nature’.

How might God's moral nature provide the ‘ground’ of moral values? Craig explains.

It seems, then, that moral values are to be identified with God's moral attributes – and, as the context makes clear, this is supposed to be analogous to the way in which (on the received view) water is identified with H₂O and heat with the energy of molecular motion.

This may help explain the kind of ‘identification’ Craig has in mind – though not perhaps the way in which these identities are discovered.Footnote ⁴ Before moving on, though, we need to get much clearer about what is being identified with what. For example, love is a prime moral value, but I take it that Craig doesn't mean to identify our love with God's love. So what does he mean? His appeal to the standard meter bar analogy may give us a clue. He seems to be thinking of the standard meter bar as a paradigm in the following sense. The degree to which an object approximates to being a meter long just is the degree to which it is similar in length to the standard meter bar.Footnote ⁵ If God's moral nature is a paradigm in (roughly) this sense, then Craig can say that the value of our love – its goodness – consists in its resemblance (however partial and fragmentary) to God's love. More generally, he can say that what it is for us to be morally good (to the degree that we are so) is the degree of our resemblance in relevant respects to God, where the attributes on his list of divine virtues tell us what at least some of the relevant respects are. If that's right, then we can capture Craig's idea in a snappy slogan: goodness is God-likeness.

We have been speaking about the moral goodness of persons and their characters. But in the first passage quoted above, Craig says that God's moral nature is the absolute standard against which actions are to be ‘measured’. How is this to be understood? Perhaps the idea is that morally good actions are actions of a kind that a person who relevantly resembles God might do. Once again, the claim is about what it takes for an action to be good, and not what it takes for us to recognize its goodness. We don't have to do any measuring or comparing in order for our actions to be of a kind that a God-like person would do.

It appears, then, that Craig's ontology of moral value makes two related ‘informative identifications’.

In this way, God's moral nature is said to be the ultimate foundation or ground of the moral goodness both of persons and of their actions. God-likeness (in relevant respects) is what makes persons and actions morally good.

Moral duty

So much for the ontological foundation of moral values. But not every morally good act is a moral duty. For example, it might be good – though not a duty – to join Doctors Without Borders. Craig believes that theism offers a ‘perspicuous account’ of what makes some things morally obligatory and others morally wrong (Craig & Kurtz et al. (Reference Craig, Kurtz, Garcia and King2009), 172).

Craig's theory of moral duty is explicitly built on his theory of moral value. It assumes that God has a ‘perfect moral nature’ – that God is ‘by nature loving, generous, just, faithful, kind, and so forth’, and that God's commands ‘flow necessarily from his moral nature’ (Craig & Kurtz et al. (Reference Craig, Kurtz, Garcia and King2009), 30). Only given this pair of assumptions are God's commands said to constitute our duties, turning the morally good into the morally required.Footnote ⁷

One might wonder about the phrase ‘flow necessarily from his moral nature’. Does it mean that each divine command is necessitated by God's moral nature – that God's moral nature makes it impossible for him not to command what he does in fact command? Or does it mean merely that it is necessary that all divine commands flow from God's moral nature, where the ‘flow from’ relation is understood in a weaker sense? Craig doesn't say. But whatever the details, it's clear that the main point of the claim that God's commands ‘flow necessarily from his moral nature’ is to head off a familiar objection to the divine command theory. It will be convenient to refer to it as ‘the arbitrariness objection’. It goes something like this.

Some philosophers think the arbitrariness objection is decisive (Shafer-Landau (Reference Shafer-Landau2004), 80–81). But Craig thinks his version of the divine command theory is completely untouched by it. To see why, consider the duty to be generous to those in need. On Craig's account, we can endorse all three of the following claims.

Given (A), it might be thought that there is nothing objectionably arbitrary about God's commanding generosity. Given (B), the goodness of God's reason for issuing this command is rooted in his moral nature; it is not therefore independent of God. (C), finally, assures us that it is God's command, and not merely the goodness of generosity, that raises it to the level of a moral imperative.

Many questions remain. Could God have failed to command generosity? Could generosity have failed to be a duty? Just what degree of generosity is required? And why did God choose to require just that degree of generosity rather than some other? If there is no reason, then at least a limited version of the arbitrariness objection might still get a bit of traction. It isn't at all clear to me how Craig would deal with these issues. But I have other fish to fry, and I am going to assume, at least for the sake of argument, that Craig has qualified the divine command theory sufficiently to immunize it against the arbitrariness objection.

Here is another familiar objection to the divine command theory to which Craig thinks his version is immune.

Craig's response is that on his theory moral obligation is constituted by the commands of a God who is essentially just and loving. It is therefore impossible for him to issue such commands, and so we don't need to worry about what would be the case if he did.

You might wonder whether this really gets to the heart of the problem. Even if such commands are incompatible with God's nature, isn't it still true that according to the divine command theory eating our children would be morally obligatory if – per impossibile – God commanded it? And isn't this something Craig should be worried about? He doesn't think so. He says that it ‘makes no sense’ to ask whether it would be ‘right to eat our children’ if God commanded it (Craig & Antony (Reference Craig and Antony2008) ).

According to Craig, then, it is either vacuously true both that it would be – and that it would not be – morally obligatory to eat our children if God commanded it, or else the scenario is so completely incoherent that it doesn't matter what truth-value we assign this counterfactual.

I am not persuaded that this is the only – or the best – way to handle the objection. Many counterfactual conditionals with impossible antecedents – counterpossibles, as I shall call them – seem to me to be non-vacuously true or false, and the assignment of truth-values in such cases need not be arbitrary. To take just one example, it seems to me that if – per impossibile – a completely truthful and omniscient being said that two-plus-two is five then two-plus-two would be five. What's driving my intuition in this case is that the antecedent does not entail the consequent merely in virtue of being impossible. It also does so because of the way in which its content is logically related to the consequent. Whatever a completely truthful and omniscient being says must be true, and what we have here is a straightforward substitution instance of that pattern.Footnote ⁸

So what if God commanded us to eat our children?Footnote ⁹ Remember that for Craig God is, necessarily, a perfect being. If that is understood, then it really doesn't matter to Craig's position whether it's impossible for a perfect being to command such a thing. Why? Because if a perfect being commanded it, the being would have a morally sufficient reason for doing so; and if – per impossibile, perhaps – a perfect being had a morally sufficient reason for commanding us to eat our children, we should do it. If I am right about this, then Craig's divine command theory escapes refutation – not for the reason he gives, but rather because the alarming-sounding counterpossibles implied by it turn out to true!Footnote ¹⁰

What's so special about being God-like?

Given fairly standard assumptions about God's moral nature, Craig may be able to deal with the worry about horrific divine commands. And given that he has a separate (though still God-based) account of moral values, he may be able to fend off the arbitrariness objection to his divine command theory. But what of the two ‘informative identifications’ that lie at the heart of his moral ontology? Should we accept the view that God's moral character is the ultimate standard of moral goodness, and that moral goodness in finite creatures is the degree to which their characters resemble God's? Can Craig get away with saying that goodness just is God-likeness? I am not sure that he can.

What's so special about being God-like, anyway? Why is being like God the standard of moral goodness? For this idea to have content or plausibility, it must be spelled out in terms of the characteristics that are included in God's moral nature. Perhaps the following formulation would get the job done.

The trouble is that this makes it look as if love and generosity and justice and the rest are doing all the work in the proffered account of moral goodness, leaving God no significant role to play.

A simple Euthyphro-like dilemma may help to clarify the alternatives here. Is God good because he is loving, generous, just, faithful, kind, and so forth? Or are these attributes good-making because God has them? On the first alternative, moral goodness supervenes directly on these marvellous traits of character. Anyone who possesses sufficiently many of them to a sufficiently high degree is morally good, and – let it be noted – this will be so whether or not there is a God. Even if there were no perfect being, love and justice and the rest could still be constitutive of the moral ideal, and it could still be the case that persons are morally good to the degree to which they realize that ideal in their lives. On a view like this, the existence of God plays no role at all in the ontology of moral value.

So what about the other alternative? Could Craig get away with saying that love and generosity and justice and faithfulness and the rest are good-making only because there is a God who is loving and generous and faithful (etc.) to the maximum possible degree? This alternative seems incredible to me. It implies that if there were no God who perfectly exemplified them, these properties would count for nothing. A person could be as fair-minded and loving and generous and faithful as you please and still fail to be morally better than a cruel and malicious person. Why would a discerning and consistent atheist have to think a thing like that?

Is there a way to slip between the horns of this dilemma? Well, some theists may wish to say that God is God's moral nature. It would then be open to them to say that God is the ultimate standard of moral goodness, in which case God's existence would be required for the existence of goodness. That God is identical to his moral nature is of course entailed by the classical version of the doctrine of divine simplicity, according to which God is identical to each of his attributes, and each of them is identical to all the others. Despite the efforts of some able philosophers, I have never been able make sense of this doctrine, and to judge from his pronouncements on the subject, neither can Craig (Craig (Reference Craig2007) ). So I won't pursue that possibility further here.

In rejecting the second horn of my Euthyphro-like dilemma, I have put a lot of weight on intuitions about how things would be even if God did not exist. This may give some readers pause. Like most philosophically inclined theists, Craig holds that the non-existence of God is metaphysically impossible. It is, according to him, a necessary truth both that God exists, and that God is perfect in love and justice and so on. If that's right, then my conclusions are based on intuitions about counterpossibles. Might this give us a reason to discount them?

In an earlier context, we saw that Craig thinks we should disregard counterpossibles entirely. We needn't worry about whether eating our children would be obligatory if God commanded it, he says, because such horrific commands are incompatible with God's moral nature – and because nothing of significance hangs on the way we deal with counterfactuals whose antecedents are impossible. The upshot of our discussion, however, was that Craig is quite wrong about this. Moreover, Craig may well be committed to the non-vacuous and non-trivial truth of at least one proposition that (by his lights) must count as a counterpossible.

To see this, recall that the second of Craig's central claims is:

T2 is formulated in the indicative mood, which muddies the waters a bit. Maybe what Craig really has in mind is, ‘If you think that God does not exist, then you should conclude that morality lacks a sound foundation’.Footnote ¹¹ But why would Craig say this if he didn't think he could show that God is absolutely necessary to the foundations of morality – a claim that is quite naturally expressed by a subjunctive version of T2: If theism were false, morality would not have a sound foundation. Does Craig think that this counterpossible is only vacuously true? Or that it is inconsequential whether we say it is true or false? One wouldn't have thought so.Footnote ¹²

Without God, Craig says, we are stuck with a naturalistic world-view on which ‘moral values do not exist but are mere illusions of human beings’ (Craig & Kurtz et al. (Reference Craig, Kurtz, Garcia and King2009), 176). I would be very surprised to learn that he does not also think that if there were no God, moral values would not exist and would be mere illusions of human beings. After all, Craig explicitly claims that the moral worth of human beings is entirely dependent on the existence of GodFootnote ¹³ and that without God human life could have no moral significance.

Of course, this is just what you should expect if you make the two ‘informative identifications’ that lie at the heart of Craig's theory of the foundations of morality. If God's moral nature is the ultimate standard of moral goodness, and if goodness is God-likeness, then if there were no God nothing would be morally good. A generous person would not be better (in that respect) than a stingy one. A just judge would not be better than a partial one, a loving mother would not be better than an uncaring one, a faithful friend would not be better than a fickle one. And so on through the whole list of moral virtues.

But what if you haven't yet joined Craig in making the two ‘informative identifications’ that land you with these implications? What if you're still trying to decide whether you should accept them? Is there any other way in which it could be shown that morality would be illusory in a Godless world? And if so, would that give us a reason to accept the view that goodness is God-likeness?

Nothing special about humans in a Godless world

Craig does in fact have quite a bit to say on behalf of the view that morality would be illusory without God. Consider first his vigorous defence of the claim that human flourishing would have no moral significance in a Godless world.

There are some pretty remarkable inferences here. The atheist is supposed to draw the conclusion that human flourishing isn't morally significant from the premise that there is nothing special about human beings. And he is supposed to derive that premise from the following facts.

• Humans are tiny compared to the universe.

• They haven't been around very long.

• They owe their existence to mindless natural processes.

• They die after a short time.

• Eventually all of them will be permanently dead.

I fail to see why anyone should move from these premises to the conclusion that there is nothing special about human beings. If one were to draw up a list of things that make us special, it would probably include things like these. Humans are (or can be) self-conscious, capable of rational reflection and deliberation, of making plans and carrying them out. They fall in love, they have children, form family bonds, and care for one another. Some of them write poems or compose symphonies or discover proofs of deep mathematical theorems. Others understand and appreciate those poems and symphonies and theorems. Non-human animals share some, though by no means all, of these characteristics; and none are shared by rocks. So why aren't characteristics like these – all of which could be found in a Godless universe – sufficient to make us ‘special’?Footnote ¹⁴ That we are the ‘accidental by-products’ of mindless natural processes, or that we haven't been around very long, or that we won't be around all that much longer, or that we are tiny in comparison with the universe is entirely beside the point. What matters to our worth is what we are – not how we got here or how long we will be here. If that's right, then no matter how much angst an atheist may experience in the face of a mindless, unplanned, unguided, silent universe, the unvarnished facts of her condition do not deprive her of worth.

What we are in a Godless world

I am certain that Craig would agree that what we are matters to our moral worth. But in a Godless world he thinks that we ourselves would be mere ‘specks of dust’. What's required for moral worth would be missing – viz., a divinely implanted immaterial mind or soul. Without that, Craig says, we are ‘just animals’, or perhaps ‘just molecules’ organized in a certain way. Consider the following remarkable passage.

Why should a materialist – whatever her views about God – think that such a ‘rearrangement’ of molecules has no moral significance? Probably Craig is thinking along the following lines. Given materialism, a little girl is merely an arrangement of molecules. But mere arrangements of molecules can have no moral significance. So a consistent materialist must conclude that little girls have no moral significance. To this argument, the proper response is simply to point out that not all arrangements of molecules are created equal.Footnote ¹⁵ When molecules are arranged in such a way as to constitute a living human organism having a first-person point of view, the person thus constituted has special moral significance. And – note well – she has it in virtue of the very same special-making properties highlighted in the last section. That she has a material mind is of no consequence. If it makes her self-aware, capable of rational reflection, and so on, there is no reason to deny that she has moral worth.

I haven't mentioned the ‘material constitution’ account of personsFootnote ¹⁶ to endorse it, but only to make it clear that there are materialist options on which persons are special. Craig has yet to give any reason for thinking that the absence of an immaterial soul entails that there are no morally relevant differences between human persons and other animals, or (for that matter) between animals and inanimate objects. If a bomb rearranges a rat's molecules, that's a bad thing too – not as bad as rearranging a little girl's molecules, but still bad. It isn't okay to do just anything to a rat, on the ground that it's ‘only a soulless animal’ or that it's ‘only a bunch of molecules arranged in a certain way’.

What difference would the presence of an immaterial soul make, anyway? As far as I can see, it could make a moral difference only if having one is necessary for having properties like those I have already emphasized – self-awareness, a capacity for rational reflection, and so on.Footnote ¹⁷ Clearly, Craig has his work cut out for him here. First, he needs to show that these special-making features cannot be possessed by beings with material minds. Second, he needs to show that they would be possessed by beings having immaterial minds. And third, he needs to show that immaterial minds could not emerge in the ordinary course of nature but must be implanted by God.

‘The temptation of speciesism’

Craig puts a lot of weight on another, rather different, line of argument. If there is no God, he says, we are stuck with the depressing view that human morality is merely a ‘pattern of behaviour’ that evolved in response to ‘socio-biological pressures’. It is useful ‘for the perpetuation of the species’, but it isn't ‘objectively true’ or in any way superior to patterns of behaviour exhibited by other animals (Craig & Antony (Reference Craig and Antony2008) ). One of Craig's examples is slavery.

To think that there is something special about human morality (according to which slavery is wrong) is to ‘succumb to the temptation of speciesism – an unjustifiable bias in favour of our own species’. Indeed, it is to suffer from ‘delusions of moral grandeur’ (Craig & Antony (Reference Craig and Antony2008) ).

Well, what if (as Craig suggests) ants were ‘endowed with rationality’? There's a counterpossible for you! If ants were endowed with rationality, they would not be ants. But let's play along for a moment. If these (very special) ants had a moral code, it might be one that (like that of our ancestors) condoned the slavery of (equally special) aphids. But so what? Craig's suggestion appears to be that it would be a ‘speciesist’ mistake for an atheist to hold that our morality (which forbids slavery) is objectively superior to the slavery-permitting morality of these imaginary ants. But why think a thing like that? Human slavery is not practised in most parts of the world today because some of our ancestors discovered that enslaving other persons is evil. If we reject some of the moral principles of Craig's imaginary ants, we do so on precisely the same grounds as those on which our ancestors rejected some of the moral principles of slave-owning humans. The only ‘speciecists’ in Craig's little thought-experiment are the ant-persons who enslave aphid-persons.

Here is another case in which Craig claims to find a ‘speciesist mistake’. This time the issue is the immorality of torture.

There are at least two things to say in response. First, torturing animals, including mice, is a very bad thing (for us) to do. What makes it bad is that it involves making the victim suffer severe pain. A real ‘speciesist’ would be morally blind to the badness of wantonly causing members of another species to experience severe pain. In the second place, cats are not capable of rational reflection and are not capable of knowing that it's bad to inflict needless pain. This judgement does not reflect an unreasonable prejudice in favour of our species. It is the recognition of a plain and obvious fact about an entirely relevant difference between human persons and cats. It is on account of this difference that we are subject to moral censure and blame when we torture other sentient creatures, whereas cats are not.

Evolution and the origin of morality

The thought that evolution and social conditioning are the source of our capacity for moral behaviour troubles Craig in another way.

My own first thought is to cheer for the baboons. If these sophisticated primates have developed a sort of proto-moral sense, good for them! That's not Craig's point, however. He wants us to see that ‘given a naturalistic worldview’ cooperative and self-sacrificial behaviour is ‘just the by-product of biological evolution and social conditioning’ (my italics). And this, he seems to think, implies that such patterns of behaviour would not be objectively correct. To make this point, Craig quotes atheistic evolutionist Michael Ruse who (as Craig puts it) ‘reports’ that morality is ‘just an aid to survival and reproduction and any deeper meaning is illusory’ (Craig & Antony (Reference Craig and Antony2008), Craig & Kagan (Reference Craig and Kagan2009), Craig & Kurtz et al. (Reference Craig, Kurtz, Garcia and King2009), 31–32).

It appears that Craig intends to give a conditional endorsement of Ruse's position. This, he seems to be saying, is what a clear-headed atheist should think. The trouble is that Ruse is not simply ‘reporting’ a well-known fact about biology. He is saying something about the implications of what biology tells us. We have to decide whether Ruse is right in drawing out those alleged implications. Does he have a decent argument for saying that what biologists have discovered about the origins of the human moral sense should lead anyone to believe that morality is an illusion?

This is not the place for a survey of Ruse's work on this subject. But it may be worth looking at the argument of the Ruse article regularly cited by Craig (Ruse (Reference Ruse1994) ). The first premise of Ruse's argument appears to be that Darwinian evolution selects for a certain amount and degree of cooperative behaviour among kin. From this, Ruse leaps to the conclusion that mutation and natural selection ‘explain morality’. He then argues as follows.

Ruse's conclusion: there are no objective moral truths. In that sense, he repeatedly says, morality is an illusion – though he claims that there is great biological utility in our believing that it's not.

There are at least two problems with this argument. First, it is obvious (to me, at least), that there is not – nor is there likely to be – a complete Darwinian explanation of most of what we now call morality. Introducing a bias in favour of cooperative behaviour among the members of one's own group scarcely begins to account for the moral judgements we actually make. To pick just a few examples, it does not account for anyone's thinking that killing strangers is wrong, that enslaving members of another race or tribe is wrong, that treating women as inferiors is wrong, or that cruelty toward non-human animals in the pursuit of trivial human ends is wrong. It has taken centuries of history and conversation and argument and rational reflection on the relevant facts for us (or at least many of us) to see that these things are wrong. We can be grateful that evolution gave us a big brain and with it a capacity for a lot of things, including moral reflection. But we have had to use what evolution has given us to arrive at the insights that form the core of what most of us think of as morality.

In the second place, it is question-begging to assume without further argument that there could be a world identical to ours in other respects but in which suffering (say) is not bad. A typical moral realist would claim that moral properties supervene on natural ones in a way that renders such a world impossible. Only because Ruse has already decided (on some other grounds, I presume) that there are no objective moral facts, does he think that such a world is possible.

Ruse undoubtedly thinks that objective moral truths would be ‘redundant’ in yet another sense. They contribute nothing to the evolutionary explanation of moral beliefs, and do no explanatory work in accounting for the human moral sense. But even if this point is granted, the argument needs a further premise – perhaps something like this: if the truth of a proposition contributes nothing to the best causal explanation of our belief that it is true, then that proposition is not true. We needn't look far to find a counterexample. It is hard to see how the truth of this proposition could contribute anything to the best scientific explanation of anyone's belief that it is true.

One final point about Craig's use of Ruse's reflections on evolution and morality is worth making. It concerns the overall structure of Craig's moral argument for the existence of God. To make his argument stick, Craig needs to convince us that there are objective values and obligations, and he needs to do this in a way that does not simply presuppose the truth of theism. Consequently, he must find a way of rejecting the sort of naturalistic ‘explaining away’ of morality attempted by Michael Ruse. He seems to be doing just this when he warns against the ‘the genetic fallacy’, which he defines as ‘trying to invalidate an idea by showing how it originated’.

If this is right, however, it undermines the evolutionary part of Craig's own case for T2. ‘Trying to invalidate an idea by showing how it originated’ is not obviously less problematic if there is no God, and Craig surely does not think that rationality requires anyone to commit a genetic fallacy. At the very least, then, he owes us an explanation of why Ruse-type ‘explaining away’ is not fallacious if atheism is true. So far, he has not discharged this obligation.

That doesn't necessarily mean that he couldn't do so. More sophisticated and challenging versions of the evolutionary argument against moral realism can be found in the current literature. Some able philosophers argue that evolutionary explanations of morality lead to moral scepticism. That chance and natural selection have moved us even a small way in the direction of correct moral beliefs seems to them to be too great a coincidence to be credible.Footnote ¹⁸ Some see this as a reason to reject moral realism (Street (Reference Street2006) ), others as a reason to adopt the hypothesis that evolution is divinely guided and that God has written the moral law on the human heart (or at least on some human hearts!).

There is a lot to think about here, and I don't claim to have settled all the relevant issues.Footnote ¹⁹ But even if it should turn out that God is needed to guide evolution to the point where knowledge of basic moral principles is possible for us, it would not follow that God is needed to provide the sort of ontological ground Craig is after.Footnote ²⁰ So let's return to that issue. Can Craig give us any real reason to prefer a theistic account of the ontological foundation of morality? To think that theism gives the only possible (or even the best) ontological explanation of basic moral truths?

Does morality require a foundation outside itself?Footnote ²¹

Why are love and justice and generosity and kindness and faithfulness good? What is there in the depths of reality to make them good? My own preferred answer is: Nothing further. If you like, you may say that they are the ultimate standard of goodness. What makes them the standard? Nothing further. Possessing these characteristics just is good-making. Full stop. Is there some problem with this? Some reason to press on, looking for a ‘deeper’ answer that only theism can provide?

It's not obvious that there is. No matter what story you tell about the ontological ground of moral value, you must at some point come to your own full stop. If you say that love is necessarily good because God necessarily exists and loves and because God's moral nature is the ultimate standard of goodness, then we can ask what makes God's moral nature the ultimate standard. It would be unwise to respond, ‘because it includes love and justice and the rest’, since that would confine us to a small and entirely unenlightening circle of ‘explanations’.Footnote ²² At some point, you are simply going to have to bite the bullet and say, ‘That's just how it is’.Footnote ²³

So which is the correct stopping point? The non-theological one, according to which goodness supervenes directly on love and justice and the rest, rendering the detour through theology entirely unnecessary, has at least the virtue of simplicity. However, Craig gives what he takes to be a decisive reason for rejecting such a view in the following passage.

Let's get one small point out of the way quickly. There is nothing ‘atheistic’ about the view to which I am attracted – it is available to atheists, but it is equally available to theists. A theist can say that God is good because, among other things, he possesses the good-making property of being loving. Indeed, I believe this is what theists should say. It is by far the simplest and most straightforward way to give content and significance to the claim – central to traditional theism – that God is good.

So what is Craig's argument against ‘atheistic moral realism’? In the passage just quoted, he makes it clear that there is no room in his ontology for abstract moral properties. Values like love or justice can exist only as properties of individual persons. Otherwise, Craig says, they would be groundless, ‘floating in an unintelligible way’. It seems clear to me that this objection misses its intended target. The pertinent issue here isn't whether uninstantiated moral properties can exist. It is whether – in a Godless universe – goodness is present in whatever instances of love and justice might exist in that universe. So far, then, Craig has done nothing to show either (a) that love and justice could not be instantiated in a Godless universe or (b) that goodness would not be present if they were.

That's not quite the end of the matter, however. While pressing his ‘speciesist’ charge (critically discussed above), Craig asks, ‘How do these strange non-natural moral properties come to supervene on the members and actions of our species?’ (Craig & Antony (Reference Craig and Antony2008) ). Craig here appears to be gesturing in the direction of a ‘queerness argument’ specifically directed against the view that moral properties supervene upon, without being reducible to, non-moral ones. Unfortunately, Craig does not elaborate the point; so my response will be brief.

However the natural/non-natural distinction gets made, it's clear that on Craig's own view the goodness of creatures is non-natural. It consists in resemblance (in relevant respects) to God, who is himself a paradigmatically non-natural being. Which respects are relevant is fixed by God's moral attributes, which themselves must be non-natural. So the problem Craig means to be raising here can hardly be that he objects to the non-natural. Instead it has to do with the way in which goodness ‘supervenes’ on non-moral properties. The moral realist has failed to tell us how goodness supervenes on love and justice and the rest.

It is just this demand for a ‘further explanation’ that I have already questioned. It is (I say) every bit as reasonable to ask in virtue of what God is good as to ask in virtue of what human love is good-making. If we are told that God is good in virtue of being loving and just and so on (and what else is there to say?), this merely brings us back to the tight and wholly unilluminating circle of explanations to which I have already called attention.

It may be said that God's moral attributes just are the ultimate standard of goodness. But how is this is any more satisfying than saying that love (for example) just is good-making? As far as I can see, building God and God's attributes into the account of moral values merely complicates things and replaces one set of puzzles with another.Footnote ²⁴

Even if I am right in thinking that we don't need to ground moral values in God, it may still be thought that we need God to account for moral duty. Moral duties (it may be said) must be constituted by commands in order to have imperative force, and a perfectly good God is the only adequate source of such commands. However, it seems to me that a non-theist who embraces moral realism is not without resources at this point. There are many options, of which I'll mention just two.

The first is simply to deny that duties must be constituted by commands in order to have imperative force. There are, after all, lots of normative laws that do not require a lawmaker. If, for example, you know that two propositions are inconsistent with one another, and you also know that one of them is true, then you should not accept the other.Footnote ²⁵ Nobody thinks we need a ‘divine command’ to back up this rule. I see no reason why it should be different for moral rules. If this is right, then the way is open for the non-theist to say that basic moral duties are fundamental moral facts and (like moral values) require no further foundation or ground.

But suppose something further is desired. Here is another option. It is a variant of the ideal spectator theory. Even an atheist might consistently identify duties with commands that would be given by a perfect being. That might not settle every question we'd like have settled; but it would certainly make it a duty not to kill or steal or practise cruelty. Interestingly, such an account fits nicely with Craig's claim that God's commands ‘flow necessarily from’ his perfect moral nature. Even by his lights, there must be a fact of the matter about what a being possessing a perfect moral nature would command if there were such a being. Once again, it turns out that the actual existence of God makes no difference to the ontological foundation of morality.

In conclusion, then, neither of Craig's two theses holds up. T1 may be true, but only in a trivial way. For all we've been shown, morality has whatever foundation it requires whether or not God exists – in which case, of course, it has whatever foundation it requires if God does exist. On the other hand, T2 (and, of course, its subjunctive sibling) appears to be straightforwardly false. A Godless universe could contain loving and just persons, and if it did they would still be morally good. It might contain acts of cruelty, and if it did those acts would still be morally wrong. Neither moral value nor moral duty need be illusory in a Godless universe.Footnote ²⁶