Whether God Exists Is Irrelevant to Ethics

Abstract

The question of whether ethics is possible without God is a non-issue. While many believe that without God, morality collapses, I contend that the existence or non-existence of God has no bearing on whether ethics is possible, whether moral truths exist, or whether ethical inquiry is viable. Ethics is no more secure within a theistic framework than an atheistic one. I establish this by critically examining Divine Command Theory (DCT) and its variants, including Divine Nature Theory, demonstrating that they fail to provide truthmakers for moral statements, explain moral truths, generate moral knowledge, or serve as a practical guide for ethical decision making. If one seeks a way to justify ethical principles or resolve moral dilemmas, appealing to God does not improve the situation; supernatural explanations, including those invoking divine commands or nature, fail to meet the criteria of explanatory adequacy. I conclude by suggesting a secular approach to ethics—drawing from Ted Schick’s inference to the best action—that does not depend on God’s existence. Ultimately, if moral nihilism is a concern, God’s existence offers no solution. If ethics is possible at all, it is possible regardless of whether God exists.

One of the great tragedies of mankind is that morality has been hijacked by religion.

                                                            —Arthur C. Clarke

The question of this Special Issue, “Is an ethics without God possible?” is a non-issue—or so I shall argue. That is not to say that it is an unimportant topic of discussion; indeed, given that it is commonly thought that if God does not exist, all is permissible, the topic is highly relevant. But that also means that my conclusion is controversial: neither the existence nor the non-existence of God has any bearing on whether ethics is possible—on whether there are ethical truths, whether the study of ethics can be done, or (to put it colloquially) whether everything, anything, something, or nothing is permissible. In short, ethics’ foundations, its truths, and its applicability are no more certain in a theistic framework than in an atheistic one, and the challenges of determining whether moral facts exist or how to justify ethical principles remain unchanged regardless of God’s existence.

To establish this, I will argue that Divine Command Theory (DCT) fails to do what any good, worthwhile, or useful ethical theory must do. It fails to (1) provide truthmakers for ethical statements, (2) explain ethical truths, (3) produce ethical knowledge, or (4) serve as a practical guide for ethical decision making. Consequently, the existence (or non-existence) of God neither strengthens nor weakens our ability to “do ethics” or “have an ethic”, and therefore the worry that motivates this Special Issue of Religions (Would ethics be possible if God does not exist?) is a non-issue.1

I will conclude with a suggestion for how ethics could be done (regardless of whether God exists), but I will also make clear that, if this approach does not work, and the reader is worried about moral nihilism—the existence of God can do nothing to alleviate this worry. But, before we begin, the defining features of DCT must be articulated.

1. The Theistic Claim: Ethics Requires God

While a number of philosophers, both classic and modern, have defended DCT—such as Augustine (see Kent 2001), Scotus (see Williams 2017), Robert M. Adams (1987, 1999), and Philip L. Quinn (1978, 1979, 1992)—the basic idea behind all their theories is that morality is brought about by God. The most common version suggests that what is ethically right and wrong is dictated by the commands of God. It is because God commands X that X is morally obligatory; it is because God does not forbid Y that Y is morally permissible; and it is because God forbids Z that Z is morally wrong. God’s will is the foundation of morality.

The most famous objection to DCT, of course, is the Euthyphro problem. In the Platonic dialogue Euthyphro, Euthyphro suggests that something is pious because it is commanded by the gods, but Socrates rightly points out that the gods command pious things because they recognize them as pious. In other words, actions are not pious because they are commanded; they are commanded because they are pious. Otherwise, piety is arbitrary. Euthyphro agrees but can’t formulate another answer to Socrates’ question: he can’t articulate what the gods recognize in pious things that makes them pious.

In the same way, although God (if he exists) undoubtedly commands what is morally right, it would seem that he does so because he recognizes such things as morally right. Otherwise, morality is arbitrary. If God commanding something is the sole determining factor of what makes it good, then anything could be good—because God could command anything (like torturing puppies for fun)—and if he did, such an action would not only be morally acceptable, but morally obligatory. (Or, to put it another way: instead of God’s non-existence entailing that everything is permissible, if DCT is true, then God’s existence entails that everything could be permissible.) And if you insist that God would never command such a thing because puppy torture for mere amusement is morally wrong, then you are admitting that there is a standard of morality outside of God’s control that he must adhere to; in other words, you are admitting that DCT is false.

Indeed, that seems to reveal the two-horned dilemma, or paradox, of DCT. If God is going to be perfect, the greatest conceivable being, then he must be in full control of morality—there can be no standard of morality outside of him. That would seem to be the main motivation of theists who embrace DCT. However, if there is no standard outside of God, then any arbitrary thing could be or could have been morally permissible, praiseworthy, or forbidden; all God has to do is command or forbid it (and there is no outside standard to prevent him from commanding or forbidding anything).

Indeed, this seems to entail a kind of Lewisian paradox. Famously, David Lewis argued that time travel to the past is impossible because it would entail that it is possible for a person to kill their own grandfather and make it true that they, themselves, were both born and never born. Since logical contradictions can’t even possibly be true, time travel to the past must be impossible. In the same way, since God could hypothetically command anything—all it requires is for him to “utter certain words” (or whatever it is that the theist considers a “command” from God to be)—DCT entails that something could be both permissible and impermissible. Since God could, at one time, eternally command X and then, at another, eternally forbid it, according to DCT, X could be both obligatory and forbidden. Since that is logically impossible, it must be impossible for DCT to be true.

Now, to be fair, Robert M. Adams (1987, 1999) and William Alston (1989, 1990) have articulated a way out of this problem… of sorts. Instead of God’s commands dictating morality, what if it is God’s character or “nature” that does so? Actions are not morally wrong because God forbids them, but because they are contrary to God’s nature; they are not the kind of things that God would do. And actions are morally praiseworthy not because they are commanded by God, but because they reflect his nature; they are the kinds of things that God would do. Of course, God commands what aligns with his nature—but it is not his commands that determine morality—it is his nature that does. When Plato has Socrates ask Euthyphro for the definition of piety (in the Euthyphro), what he is seeking is knowledge of Piety itself—the Platonic Form of Piety that all pious things reflect—a standard outside of the gods which they know and can see reflected in the actions that they command. For Alston, and especially Adams, God—or more specifically, his nature—is that standard. Just like, for Plato, pious actions are pious because they reflect the Form of Piety, for Alston and Adams, moral actions are morally good because they reflect God’s nature.

Now, it would seem clear that this Divine Nature Theory (as I shall occasionally call it) solves the Lewisan “logical paradox” problem. God cannot have a logically inconsistent nature, so it’s not possible for something to be both obligatory and forbidden. It’s not clear, however, how well this revised version of DCT avoids the “arbitrariness” objection. After all, if the fact that God could have commanded different things (like murder) makes morality arbitrary, then the fact that God could have had a different nature (that was murderous) makes morality arbitrary. Of course, if God is immutable, his nature can’t change from what it already is (from being non-murderous to being murderous); but since there is no standard outside of God, God’s nature could have (always) been completely different than it is, and thus murder could have (always) been morally good.

In reply, Adams or Alston might say this hypothetical about “how God’s nature could have been” is irrelevant. God’s nature can’t only “not change”; it is what it is necessarily. Consider again Plato’s theory of Forms, and the corresponding theory that the morality of an action is due to the fact that it reflects a Form. “Just acts are just because they resemble the Form of Justice”. If I were to suggest that Plato’s theory makes morality arbitrary because the Forms could have always been different than they are, Plato would likely insist that I have a confused understanding of the Forms. “The Form of the Good is what it is out of necessity. If it were any different, it wouldn’t be the form of The Good”. In the same way, if God’s nature were different than it is—if it was murderous—then God wouldn’t be God.

Personally, I am not convinced. The only way that Forms and God’s nature can’t have been different than what they already are is if there is some outside standard that dictates what they must be; but, according to those who embrace such theories, there is nothing above the Form of the Good or God.2 Thus, they could have (always) been different than they are.

However, even if we grant that the altered Adams/Alston Divine Nature Theory can avoid the arbitrariness objection, it still fails as an ethical theory because it does not do anything that a good ethical theory must. As I shall now show, it doesn’t provide truthmakers for moral statements, explain the truth of moral statements, or produce moral knowledge, nor is it useful for making moral decisions.

2. The Truthmaker Problem

A truthmaker is something that makes a proposition true. In class, I tell my students that the proposition “The bottle is on the table” is true by pointing to the state of affairs of my bottle being on a table. “That state of affairs”, I say as I gesture towards the table with the bottle on it, “is that proposition’s truthmaker”. The notion and role of truthmakers is central to the most widely accepted theory of truth, the correspondence theory, which says that a proposition is true when it corresponds to the way the world is.

The correspondence theory, however, obviously intuitive and correct though it is, causes a problem for ethicists: where are the truthmakers for ethical statements? I can point to the referents of words in regular statements, like “bottle” and “table”, and even the physical relationship of “being on”. However, if I say, “Murder is wrong”, what in the world—literally, what piece of the world—does the word “wrong” point to? I can show you the truthmaker for “A murder occurred”, but not for “Murder is wrong”. As Hume famously put it in “A Treatise of Human Nature”, you can’t get an ought from an is. Or as George Mavrodes (1986) pointed out, moral propositions are “queer”. They are not true in the same way and for the same reason as other ordinary propositions.

Other major ethical theories are, to one degree or another, an attempt to find truthmakers for moral statements. Perhaps what makes murder wrong is (a la utilitarianism) the fact that it causes more pain than happiness, or (ala Kant) the logical fact that such actions can’t be universalized. Now, to what degree these theories are satisfactory in this regard is debatable; one might argue that states of affairs about happiness and pain are only truthmakers for statements about the amount of happiness and pain an action produces (not for moral facts); logical facts about whether an action can be universalized are only truthmakers for statements about what kinds of actions can be universalized (not moral facts). What is missing is an explanation for how such facts ground moral statements. Why does maximizing happiness make an action morally good? Why does the fact that it can’t be universalized make an action morally bad? It would seem that the best we can do is just employ an unstated assumption or axiom that they do.

In the same way, however, “God commanded X” can only serve as a truthmaker for the proposition “God commanded X”. If God is perfect, the fact that God commanded X logically entails that X is morally good—for, if God is perfect, God would only command morally good things. Likewise for that which God forbids and that which is morally wrong. But logical entailment is not truthmaking (a lawyer with perfect knowledge of the law saying that X is illegal logically entails that X is illegal, but is not what makes X illegal). God’s forbiddance is not what makes the statement “murder is wrong” true; it logically entails that it is, but it is not the statement’s truthmaker.3

The same problem remains for Divine Nature Theory. God’s nature can only serve as a truthmaker for statements about what God is like. “God is love” is true because God is loving. But the fact that God is loving is not what makes loving acts good. Again, assuming God is perfect, the fact that God is loving would logically entail that loving acts are morally good, but the theory lacks an explanation for how God’s loving nature makes the statement “loving is morally good” true.

To make the problem abundantly clear, notice how it also plagues Plato’s theory of morality. For Plato, things are pious, or just, or morally good because they resemble a non-physical abstract Form. But how and why does this resemblance generate moral facts? Plato clearly assumes it does, but absent in his works is any argument or explanation for that assumption. Their resemblance would be a truthmaker for a statement about their resemblance: “The form of justice and action X resemble each other”. The truth of that proposition might be enough to logically entail that action X is just. But that’s not what makes the proposition “action X is just” true.

Since, in Divine Nature Theory, God’s nature plays the same role as Plato’s Forms—it is through resembling God’s nature that morally good actions are morally good—Divine Nature Theory has the same problem. God’s nature can serve as a truthmaker for statements about God’s nature, and the resemblance between God’s nature and an action can serve as a truthmaker about that resemblance, but none of this provides a truthmaker for the moral statement itself. So, the “queerness” of moral statements remains.

In reply, one might argue that no further fact is required; it is just a brute fact that resembling a Form/God’s Nature is what makes something morally good. But this is not clearly the case. Take chairs. Plato would say that all chairs are chairs because they resemble the Form of Chair. But I would argue that what makes something a chair is the function it can perform: the way it can support a non-standing person. Presumably, if Forms existed, the fact that a chair resembled the Form of Chair would logically entail that it could perform that function; but the resemblance itself is not what would make it able to perform that function. Notice that chairs could still fulfill that function even if the Form of Chair didn’t exist. Indeed, notice that we who doubt that the Platonic forms exist don’t doubt the existence of chairs. (Thus, doesn’t it seem odd that we would doubt the existence of moral truths when we doubt the existence of God?)

Or, to put it another way, what generated the truthmaker problem for morality in the first place is Hume’s observation that “You can’t get an ought from an is”. You can’t derive how things ought to be from the way things are. Well, if that’s true, then facts about Forms or God’s commands/nature can’t ground moral facts because they are simply facts about the way things are. “The form of justice is like this”. “God’s nature is like that”. Even if they exist, the Forms and God can only serve as truthmakers for facts about the Forms and God.

So, to the extent that theories of morality are supposed to find truthmakers for moral statements, even the best versions of DCT fall short. But maybe moral theories aren’t supposed to supply truthmakers for moral facts; maybe they are only supposed to supply explanations for moral facts. Maybe. But as I shall now show, DCT fails to do that as well.

3. God Cannot Explain Moral Truths

Before exploring this topic, it’s important to note that providing an explanation is not the same as providing a truthmaker. The state of affairs of the bottle being on the table makes “the bottle is on the table” true, but it does not explain why the bottle is on the table. My actions are the explanation for that: I put the bottle on the table (for the purpose of explaining what truthmakers are). But even if explanations and truthmaking were the same thing, my argument would not be weakened. If they are the same, then the argument of the last section also supports the conclusion of this one, and the argument of this section also supports the conclusion of the last. But, regardless of whether explaining is the same as truthmaking, God’s existence cannot explain the truth of moral statements. To understand why, consider the work of Theodore Schick.

In his paper “Can God Explain Anything?”, Schick (2003) argues that God cannot serve as a good explanation for anything. The method for delineating good explanations from bad, and finding the best one, is well defined and understood. Colloquially, it’s called “abduction” (although, technically, “abduction” refers to the process of formulating hypotheses to compare), but more accurately, it is known as inference to the best explanation (IBE). It simply involves comparing competing explanations to objective criteria to figure out which is the best one.

What are those objective criteria? Schick (Schick and Vaughn 2024) lays them out most thoroughly in his textbook How To Think About Weird Things:

-

Testability: the hypothesis must make novel observable predictions.

-

Fruitfulness: the hypothesis must get its predictions right (and consequently open up new lines of research).

-

Scope: the hypothesis must have explanatory power (the more it explains the better, as long as it does not invoke the inexplicable or raise unanswerable questions).

-

Simplicity/Parsimony: the hypothesis that makes the fewest assumptions or requires the fewest new entities is simplest.

-

Conservative: the hypothesis must be internally logically consistent and cohere with what we already have good reason to believe (i.e., already well-established theories).

Now, when I say these criteria are “objective”, I do not mean that they are never open to interpretation; like everything else, of course they are. (Parsimony, for example, can be understood in a few different ways.) I simply mean that (a) it is usually clear when a hypothesis fits the criteria, and (b) they represent what good explanations, by definition, must be. Someone did not arbitrarily select these criteria and declare that fulfilling them makes an explanation good. Not only have the hypotheses that turned out to be true always been more adequate than their competitors, but meeting these criteria simply makes an explanation more likely to be true.

Testability and fruitfulness are obvious. If I think it’s windy and predict that I’d see swaying trees outside if it were, and then I see swaying trees outside, that’s good reason to think the hypothesis is true (even though there are other unlikely yet logically possible explanations for the swaying trees). These criteria prevent the acceptance of hypotheses that can be proven false through experimental conditions and encourage the acceptance of true hypotheses that get their predictions right.

Scope, or explanatory power, is perhaps equally obvious. By definition, being able to explain things, increasing our understanding, is what a good explanation must do—and the more a hypothesis explains, the better it can do those things. What’s more, raising unanswerable questions and invoking inexplicable entities does not increase our understanding. To borrow an example from Schick, if I want to know why a bridge collapsed, and someone told me “a gremlin did it with black magic” but could not tell me anything else about the gremlin or how black magic works, we would not take them seriously. Such an explanation explains nothing; it merely replaces the unexplained with the inexplicable.

Simplicity/parsimony, or requiring the fewest assumptions, also makes a hypothesis more likely—because the more new assumptions a theory has to make, the more chances it has to be wrong. And the less conservative it is, the less likely it is to be true. If it conflicts with already established theories, it will be less likely because all the evidence for the already established theories counts as evidence against it. And if it is internally inconsistent, we know it is false because logical contradictions cannot be true.

Now, it’s important to note that merely lacking one or even two criteria is not necessarily enough to “sink” a hypothesis; if it can make up for such a shortcoming by “dominating” the other criteria, it could still turn out to be the best explanation. For example, Einstein’s theory of relativity was not conservative when first proposed, because it conflicted with Newton’s theory of gravity; but because it turned out to be more fruitful (by, for example, more accurately predicting where certain stars near the sun would be seen during eclipses), simpler (it erased the “demon” of gravity by revealing that it is only an “apparent force”), and wider in scope (it successfully explained, for example, the perihelion of Mercury when nothing else could), it triumphed over Newton’s theory and became accepted (because it was recognized as the better, more adequate, theory).

What Schick argues is that, as a theory, “God did it” will never be able to overcome its adequacy deficit and thus can never provide the best explanation for anything. Why? Because, by its very nature (and by the admission of those who defend it), the God hypothesis is not simple; it requires the existence of an infinite entity with infinite powers; that’s about as non-simple as you can get. It is not wide-scoping; how God’s powers work and why he does what he does, again, by the admission of theists, is eternally beyond our ability to understand and comprehend.4 It is not conservative, for all well-established theories suggest that physical events have all and only physical causes, and yet God would be a non-physical being that supposedly has physical effects on the world. What’s more, many have argued that the concept of God is logically incoherent (see Drange 1998). For example, God cannot be both perfectly merciful and perfectly just, nor can he be all-powerful (and thus able to do anything) and perfectly good (and thus unable to do evil). There are literally dozens of such divine paradoxes that many argue cannot be resolved (see Martin and Monnier 2003).

It also fails the criteria of testability and fruitfulness. Now, to be fair, one might argue that God’s existence is at least testable; if it is true, one would expect there to be no gratuitous evil in the world (because a perfect being would not allow gratuitous evil to exist). But there are two reasons this doesn’t mean that the God hypothesis could overcome its adequacy deficit. First, there is gratuitous evil in the world; thus—even if the God hypothesis is testable—the God hypothesis is not fruitful (it predicts a different amount of gratuitous evil than there is). And testability alone could never overcome the simplicity/scope/conservatism deficit from which the God hypothesis already suffers.

Second, to defend theistic belief, in light of the gratuitous evil that exists in the world, theists often say that God has reasons for allowing gratuitous evil that we simply can’t comprehend (or, to put it another way, the evil that seems gratuitous can’t be evidence against God’s existence because it might not actually be gratuitous). This is known as skeptical theism—as in, “we should be skeptical of our ability to understand God”—and has been defended by important Christian philosophers like Steven Wykstra (1984, 1996). But this move doesn’t help the God hypothesis overcome its adequacy deficit. While it might be necessary to save the God hypothesis from the evidence of evil in the world, it does so by making the God hypothesis unfalsifiable (i.e., untestable); nothing could count as evidence against God’s existence, because any negative evidence could be dismissed in this way (by appealing to undetectable, incomprehensible divine reasons). Thus, in an effort to defend God’s existence, skeptical theists have completely robbed the God hypothesis of any adequacy; it’s not even testable.5

If Schick is right and God cannot explain anything, then obviously God cannot explain moral truths. Now, in response, one might argue that “Since the criteria of adequacy we have been considering are scientific criteria, they are irrelevant to whether God can explain moral facts”. But the fact that they are used in science doesn’t mean they only apply to scientific claims; they are the criteria of explanatory adequacy; thus, any attempt to explain anything must align with them.

To make my point clearer, consider the “mystery therefore magic” fallacy, a variety of appealing to ignorance that takes one’s inability to find a natural explanation for something to be a reason to conclude that there isn’t one (and thus that the real explanation is magical or supernatural).6 I’ve seen my students do this when, for example, they think magicians have real magic powers (because they can’t explain how Criss Angel did what he did on a YouTube video); I’ve seen family members do this with unlikely or unexpected medical events, like when my premature son beat the odds and survived a very early birth. “How else do you explain it?”

This move is fallacious because, when you can’t explain something, your own ignorance (the theory that there is a natural explanation that you are not smart enough to find) will always be the better explanation; it will adhere to the criteria of adequacy better. Like the God hypothesis, supernatural explanations require extra entities (and thus are not simple), invoke the inexplicable (magical divine powers), and are not conservative (conflict with established laws of physics). Conversely, your own ignorance requires no extra assumptions, can explain many other things you don’t understand, and coheres with what is well-established. Consequently, appealing to the supernatural, when you simply are unable to think of a natural explanation, will always be unjustified. The “mystery therefore magic” fallacy is a bit like the “God of the gaps” fallacy, except it extends to other kinds of supernatural explanations too.

Hopefully, my point is clear: invoking God as an explanation for moral facts just commits the mystery therefore magic fallacy. It just attempts to explain the unexplained with the inexplicable. It therefore necessarily fails to adhere to the criteria of explanatory adequacy, and thus “God did it” cannot serve as an adequate explanation for moral facts.

In objection, one might argue that I am simply assuming materialism:

“You have selected criteria that favor naturalistic assumptions and then declared supernatural explanations, like “God did it”, to be inadequate because they don’t match up to them. But someone who rejects materialism and already believes in God would find ‘God did it’ to be a satisfactory explanation for moral facts”.

But, as I tried to make clear above, the criteria of adequacy are not subjective, relative, or arbitrary; they were not selected to favor one conclusion or worldview over another. Indeed, they were not selected at all. They were identified as that which make explanations better because they make them more likely. The fewer new assumptions a hypothesis makes, the fewer chances it has to get things wrong; thus, the more likely it is. The more a hypothesis conflicts with what is already well-established, the more evidence there is against it; thus, the less likely it is. Consequently, a person “finding” an explanation to be “satisfactory” because it aligns with what they already believe does not make it a good explanation. Moreover, there is nothing intrinsically materialistic or naturalistic about the criteria of adequacy at all.

Indeed, I have argued elsewhere (Johnson 2022) that it is theoretically possible for supernatural explanations to adhere most to the criteria of adequacy and be the best among all relevant competitors. For example, if ESP existed, the hypothesis that it exists could have proven most adequate; correctly predicting experimental results and the consequent new lines of research that would have opened would have allowed its fruitfulness to overcome its initial lack of simplicity and scope. This didn’t happen because ESP is not real; but if it were real, the hypothesis that it exists could have turned out to be the best.

This, however, could raise another objection.

“If it’s theoretically possible for supernatural explanations to be the best, isn’t Schick’s thesis overstated? Isn’t it possible for ‘God did it’ to be the best explanation for something?”

Perhaps, but it’s not possible for God to be the best explanation for moral facts. Let me explain with an example.

Given its very nature, “It’s aliens” could never be the best explanation for any unidentified thing that you see in the sky. “There is a terrestrial explanation, I just can’t figure out what it is” will always be the better explanation for anything that is unidentified because the alien hypothesis is, intrinsically, not simple (it requires the existence of visiting aliens), unconservative (it conflicts with what we know about the size of the universe, how fast travel can happen, and thus how likely an alien visitation would be), and narrow in scope (because it raises unanswerable questions about how the aliens got here, how their tech works, etc.). Hypothetically, this deficit could be overcome—say, if the unidentified object became identified as a spacecraft that landed on the White House lawn out of which aliens emerged. But short of that, if we are just trying to explain something we can’t identify, nothing provides the “It’s aliens” hypothesis with the opportunity to overcome its initial inadequacy and thus “prove itself”.

Likewise, if we are trying to explain moral facts, “God did it” is never going to be the best explanation. I can imagine a situation in which “God did it” might be the best explanation for something else, like if the rapture actually happened and Jesus descended from above.7 But if we are just trying to explain moral facts, “God did it” isn’t going to cut it; our inquiry into the explanation of moral facts is not going to provide any opportunity for the “God did it” hypothesis to prove itself by being more fruitful or simpler. Like in the case of an unidentified object in the sky, it’s just trying to replace the unexplained with the inexplicable.8

So that I am not misunderstood, let me be clear: I am not saying that the criteria of adequacy are appropriate for solving moral dilemmas, or figuring out which action is morally preferable in a given situation. I’m not even sure what it would mean for one action to be simpler or more wide-scoping than another. In my opinion, it’s best to consider what all the major ethical theories would say, and choose the action that would be approved by the most. More on that later. However, solving moral dilemmas is not the same thing as explaining moral facts, and to accept that one explanation for moral facts is better than another explanation for those facts, that explanation will have to (better than its competitors) achieve what explanations are, by definition, supposed to achieve. For the reasons I have articulated, appealing to the supernatural has never done this, and when it comes to explaining moral facts, appealing to God could not. In short, since divine command theory appeals to the supernatural, it cannot satisfactorily do one of the main things that moral theories must: explain moral facts.

In reply, one might argue that this is true for classic DCT because it relies on God’s commands to determine moral facts but if, instead, as Adams/Alston suggests, God’s nature is the explanation, the same objections would not apply. Alston’s Divine Nature Theory doesn’t say “God did it”, it just says “God is it”. But, in reality, Divine Nature Theory is subject to the same criticism—and to see why, we need only return to Plato’s realm of Forms.

If someone suggested that just actions are just because they resemble the Form of Justice, it would be impossible to take them seriously. The suggestion involves extra entities (and a whole other realm in which they exist), inexplicable notions (what it means for abstractions to exist and even be “more real” than physical objects), and unanswerable questions (“How can an action resemble an immaterial object?”). Although the forms aren’t “doing” anything, the notion that they can explain moral facts fails as an explanatory hypothesis. Indeed, it seems, given their very nature, that no explanation involving the forms could meet any of the criteria of adequacy. In the same way, and for the same reasons, the notion that God’s nature explains moral facts fails as an explanation. The suggestion involves extra entities (an infinite entity with infinite powers), inexplicable notions (an immaterial being with seemingly contradictory properties), and countless unanswerable questions (what God’s nature is like, and why it is what it is).9 This last point leads to the next problem with divine command theory.

4. The Problem of Moral Knowledge

Something else a good moral theory must do is produce moral knowledge; embracing it should enable a person to distinguish good from bad. But divine command theory cannot do this. If the theory is true, to have knowledge of moral truths, one would have to have knowledge of God’s commands or nature. But, as even divine command theorists like Robert M. Adams (1987) admit, even the mere belief that God exists is a matter of faith; since faith is belief without sufficient justification, and knowledge requires justification, knowledge of God’s existence is impossible. But if we can’t even know that God exists, how much more impossible must it be to have knowledge of God’s commands or nature? And if God’s commands and nature are beyond our ability to know, yet they determine moral facts, then it is impossible to have moral knowledge.

Another way to explain this problem would be to point to the fact that there is much disagreement between religions, and even within religions, about what God’s commands actually are (or, correspondingly, what his nature is like). Centuries of disagreement and wars prove that there is no objective way to establish which (if any) religious faction has it right. Thus, clearly, knowledge of God’s commands and nature is impossible; but since such knowledge is necessary if DCT is to produce moral knowledge, DCT cannot produce moral knowledge.

Or think of scriptures, like the Bible, which (in places) has God commanding actions that everyone would agree are morally reprehensible (like the genocide of the Midianites in Numbers 31) or praising (deeming blessed) actions that everyone would agree are morally reprehensible (like the dashing of infants’ heads upon the rocks, as in Psalm 137:9). Biblical scholars agree that, in reality, God did not command (or deem “blessed”) such actions; the authors of the texts in question merely said God did in order to justify or excuse the immoral actions they favored. But the reason biblical scholars suggest this is the case is because they think that genocide and infanticide are immoral, and thus are not the kinds of things that God would command. That means they are using their own assumptions, moral intuitions, or preferred moral theory (e.g., utilitarianism) to decide what God has commanded (rather than letting God’s commands dictate what they think is immoral). Using scripture to establish what God’s commands are, therefore, is necessarily circular—for, to establish what God has actually commanded, one would have to already know what morality demands—and thus cannot produce moral knowledge.

Another version of DCT is expressed in Linda Zagzebski’s (2004) book Divine Motivation Theory, which suggests that God’s character and motivations (rather than God’s commands or nature) are the source of morality. Taking this view, moral knowledge could be obtained by understanding the virtues that God possesses. But the same kind of problems haunt Divine Motivation theory as haunt classic DCT (as well as Divine Nature Theory); in fact, they are the same kind of problems that haunt virtue theory generally: it articulates no objective way to determine what the virtues are (or what the divine motivations are) and thus provides no way to produce moral knowledge.

Virtue theory suggests that the classic approach to ethics, which was to “define the good” (as utilitarianism and deontology tried to do), was faulty and doomed to fail. Alternatively, it suggests that the goal of ethics should be to become a virtuous person. But because it rejects the definitional goal of classic ethics, it offers no objective way to identify what the virtues are. Instead, it suggests that we look to moral exemplars (individual outstanding persons who already embody the virtues) and become virtuous by emulating them. But, of course, with no objective definition of what the virtues are, which persons are included on the list of moral exemplars is completely open to interpretation (consider how Catholics consider Mother Teresa to be a moral exemplar, while atheists like Christopher Hitchens argue that she was a moral monster). As a result, virtue theory fails to produce moral knowledge; at best, one can use their own preconceived notions or intuitions about what the virtues are, or who the moral exemplars are, to try to become what one thinks a virtuous person is like, but knowledge that one is a virtuous person, or of what the virtues are, is impossible.

Similarly, the advocate of divine motivation theory cannot know what the divine is motivated to do, or what virtues the divine embodies. They can use their own preconceived notions or intuitions about what the virtues are to draw conclusions about what virtues the divine embodies, but since advocates of this view, like Zagzebski, will admit that belief in God’s existence (and attributes) requires faith, they must admit that they cannot have moral knowledge.

In reply, one might argue that knowledge of God’s existence is not necessary to know what his nature or motivation is (or would be); since, if he exists, he is a perfect being, he would simply have the nature and motivation that perfection demands. But, of course, perfection simply entails having every “good” property to the maximal degree. Since there is no non-question-begging way to establish which properties are good, we are right back to the same problem—you must use your preconceived notions of the good to determine what good properties a perfect being must have—and moral knowledge is still impossible.

Alternatively, the defender of DCT might suggest (like Michael W. Austin 2003) that a philosophical naturalist could come to see something, say, charity, as morally good through a “rational insight into the necessary character of reality” (Austin n.d.). But there are two problems with this suggestion (as a defense of DCT). First, such a revelation would not reveal that God had commanded charity; only that, if God existed, he would. Second, if moral truths can be discovered in this way, it would seem that they are not dependent upon God in any way, and thus divine command theory is false.

Another defense of DCT might go like this:

“All moral theories require faith, so no moral theory can produce moral knowledge; thus, the fact that DCT cannot produce knowledge is not a valid objection against it”.

However, there are three things to say in response.

First, if it is true that every moral theory requires faith, it is because all moral theories are based on some kind of axiom (utility, for example, is based on the axiom that maximizing happiness is the only moral good). But, if being based on axioms makes knowledge impossible, then basically all knowledge is impossible. Logic, science, and mathematics are based on axioms too, but I take it as obvious that someone raising such an objection would not say that all such knowledge is impossible. Thus, I take this line of argument to be a dead end.

Second, DCT suggests that moral knowledge is only possible if God exists; that’s what raises the question of this Special Issue. “Is an ethics without God possible?” But if moral knowledge is impossible because all moral theories require faith, then DCT is just on par with everything else; it doesn’t produce moral knowledge either. Thus, the non-existence of God poses no special threat to our ability to do ethics (which is the main claim I am defending).

Third, DCT is worse off than other theories in this regard. Even if I grant the axiom upon which DCT is based, that God’s commands/nature define morality, I still can’t have moral knowledge, because I can’t know what God’s commands/nature are. If we grant (for example) the axioms of utility about happiness, we could reason out from there to at least some moral truths and knowledge. But the same does not hold for DCT theory because you cannot reason out from its primary axiom (that God’s nature/commands define morality) to conclusions about what God’s commands or nature are (and if the divine command theorist demands that we also add, as axioms, their assumptions about what God’s commands and nature are, DCT becomes completely arbitrary).

5. DCT Is Useless for Moral Decision Making

One thing that virtue theory seems to get right is that a proper understanding of the purpose of doing ethics does not merely involve identifying moral facts; it also involves guiding people towards becoming better persons and helping them make moral decisions. But because DCT does not offer enough clarity or consensus regarding the good, it is not useful for moral decision making.

Take Augustine, for example, and his argument that the key to moral orientation is loving God (see Kent 2001). According to Augustine, one’s relationship with God serves as a moral compass, a compass that guides a person’s behavior to reflect the divine will. By prioritizing devotion to God, individuals can properly align their love for other people and things. Although it sounds nice, Augustine’s theory is not applicable in a practical way because it doesn’t offer a clear and universally accepted theory of the good or specify how individuals should resolve moral dilemmas when divine guidance is ambiguous or contested. For instance, what does love of God, and the corresponding love of others, entail about the issue of gay marriage? Does loving God involve adhering to the (supposed) divine command to prohibit such unions? Or does loving God demand we emulate God’s unconditional love and acceptance and thus support and affirm same-sex relationships? Does loving God entail that one should value justice above all else (and turn in a friend who broke the law), or mercy and forgiveness? Can a Christian voter support an ungodly candidate simply because the candidate says they align with the voter on what they take to be the most important moral issue? Without a detailed account of the good that God’s love encompasses, this approach leaves moral agents directionless in complex situations.

Or take William J. Wainwright’s (2005) observation that, according to divine command theory, atheists and agnostics cannot be moral. He considered this problematic, because obviously atheists and agnostics can be moral. Inspired by Wainwright, we might apply this kind of objection more broadly. Even among those who share a belief in God, there is widespread disagreement about what God’s commands are, what constitutes divine will, and even what God’s nature is like and what virtues he embodies. For example, while many agree on God’s omnibenevolence, there is little agreement on what this attribute entails in practical terms. Does omnibenevolence prohibit capital punishment or endorse it as just? Does it demand nonviolence or permit war in self-defense or for just causes? What’s more, any apparent agreement about God’s nature or commands among religious devotees simply stems from pre-existing moral intuitions. We think God is loving just because we already hold these values to be morally significant. Consequently, DCT’s ability to guide moral decision making is merely a function of our moral intuitions (and is thus reducible to the aforementioned Ethical Intuitionism of Michael W. Austin).

Now, I should make clear a related fact. The fact that DCT is useless for moral decision making, because it does not offer enough clarity or consensus regarding the good, does not mean that religions cannot inform (i.e., be “the basis” for) the moral systems that some persons or societies endorse. Of course they do; many people and societies embrace certain conceptions of what is good and evil based on religious teachings. And because religious systems often circumvent the need for verification, and just tell you what is (what it defines as) right and wrong, religious systems might be useful for moral decision making in a way that DCT is not. But this fact does not undermine the thesis of this section or this paper.

First, although religious systems can inform what a person or society regards as moral by making absolute statements about right and wrong, I wouldn’t say that is sufficient to make it useful for moral decision making. Any random fictional book can inform a person’s moral beliefs, but unless the morals endorsed by the book are valid (true/grounded in real virtue), I wouldn’t call it useful for moral decision making. Atlas Shrugged informs some people’s morals; I wouldn’t say that makes it useful for moral decision making. Indeed, I would consider it quite useless. Some religious traditions are no better than Atlas Shrugged, so the fact that religion informs some people’s beliefs about morality doesn’t mean that religion is useful for moral decision making.

Second, even if religions are useful for moral decision making (in the sense that they can inform moral decisions), that doesn’t mean DCT is useful for moral decision making in this way, and whether DCT is useful for moral decision making is the main issue at hand. Since DCT is not a theory of what God commanded, but only the thesis that God’s commands define morality, it cannot provide the definitive statements about morality that full religious systems attempt to provide.

What’s more, the topic of this paper, and the Special Issue in general, is whether ethics is possible without God—not whether it is possible without religion. If it were the latter, and all something needs to be able to do to “make ethics possible” is inform a person’s moral beliefs, then the question is answered before it is asked. Of course an ethics without religion is possible because there are secular ethical theories (which inform people’s moral beliefs). No, the question is whether an ethics is possible without God, and whether the assumption behind that question, that God’s existence is necessary for the existence of moral truths, is true.

And that, in turn, is why this point is irrelevant to the thesis of this paper. The main question at hand is whether God can be “the basis” for ethics in the sense that he can provide truthmakers and explanations for moral truths, not whether religions can inform moral systems. I’m questioning the assumption implied by the topic of the Special Issue: DCT generates moral knowledge. I’m not questioning whether religions can produce opinions on moral matters, because that is irrelevant to the topic of the Special Issue. Thus, pointing out that religions can be “the basis” of moral systems, in that they inform people’s ethical beliefs, does nothing to threaten my thesis.

6. Doing Ethics Without God

Given the arguments of last section, one might think that I am embracing the moral skepticism of John Leslie Mackie, especially as he articulated it in Ethics: Inventing Right and Wrong (Mackie 1977). But I am not a moral skeptic. Instead, more than anything, I am simply challenging George Mavrodes’ (aforementioned) claim that moral claims are “queer” unless God exists. No, they are queer regardless of whether God exists, because the existence of the divine does nothing to alleviate the queerness of moral claims. Knowing that God has commanded something, has a certain kind of nature, or a certain kind of motivation, would help us know what is morally right and wrong and make moral decisions. But (a) we can’t actually know what God’s commands, nature, and motivations are (thus, DCT does not produce moral knowledge and is not useful for making moral decisions), and (b) such facts about God would not produce the truthmakers or explanation for moral facts that good ethical theories must provide. Thus, even if God exists, the quest of ethics is still just as mysterious and moral facts are as “queer” as they ever have been. If ethics is impossible if God does not exist, it is still impossible even if God exists.

Fortunately, however, ethics is not impossible if God does not exist. There are, of course, plenty of secular moral theories, and even though all are subject to objection, and thus likely do not provide the moral truthmakers that ethicists seek, they are very useful for attaining moral knowledge and for moral decision making. For example, when I teach ethics in my classes, I teach my students eight ethical theories or concepts—hedonism, utilitarianism, (Kantian) deontology, rights, virtue theory, care ethics, fairness (as Rawls defines it), and justice (as in “just desserts”)—and suggest that a person can discover what action is morally best by treating each as a criterion and comparing different possible actions to see which action fits with the most criteria. Which action best promotes my well-being and the happiness of others, doesn’t treat others merely as a means to an end, respects acknowledged rights, is a virtuous “mean between extremes”, and allows me to fulfill my familial obligation to care for others, would be the kind of action that I would agree with if I didn’t know who I was in the situation, and would ensure that people get what they deserve.

I developed the method after being inspired by Theodore Schick’s (2004) paper “A Humanist Theory of Ethics: Inference to the Best Action”, which lays out a secular method for moral decision making (which mirrors inference to the best explanation). According to Schick, just as good explanations must be, by definition, fruitful, wide-scoping, simple, and conservative, it is self-evident that moral actions must be the following:

-

Just: They must treat equals equally.

-

Merciful: They must alleviate unnecessary suffering.

-

Beneficent: They must promote happiness and well-being.

-

Autonomous: They must respect individual rights.

We can and should decide, Schick argues, what is morally best by comparing different actions to these ethical criteria to determine which action adheres to the most of them to the greatest degree.

Why these criteria? Such criteria, Schick argues, are simply “contained in our concept of morality”.

Not everyone knows the difference between right and wrong. We have a label for those kind of people; we call them “criminally insane”. Anyone who has the concept of morality, however, knows that other things being equal, the best action is the one that is the most just, the most merciful, the most beneficent, or the most enabling. That’s just part of what we mean when we say that something is moral.

(p. 4)

He goes on to clarify what he means by saying that the fact that morality requires these things is self-evident; they are as self-evident as “whatever has shape has size”.

The statement, ‘Other things being equal, the most just action is the one that should be performed’ is self-evident to anyone who has the concept of morality. You can’t refute this statement by simply asserting you don’t believe it, any more than you can refute the claim that whatever has a shape has a size by simply asserting that you don’t believe it. If you believe that the statement is false, then the burden of proof is on you to provide a counterexample. If you are unable to do so—if you cannot cite a situation in which, other things being equal, the most just action is not the one that should be performed—then your assertion is irrational, for you have no reason to make it.

(p. 5).

To be clear, Schick is not appealing to Ethical Intuitionism (EI), the view espoused and defended by Michael Austin, which suggests that we can gain ethical knowledge by intuition. Schick doesn’t argue that we know that, for example, beneficence is good “by intuition”. For Schick, the way people use the word “moral” indicates that it necessarily involves beneficence, most likely because the fact that beneficence is morally praiseworthy is simply self-evident.

That said, Ethical Intuitionism (EI) would be another way to defend the criteria that Schick gives us, and Alston defends EI against the criticisms that it usually faces—that it cannot produce moral knowledge, that it cannot resolve moral conflicts, and that it “tends toward arrogance and subjectivism”. Thus, to the extent that Michael Alston is correct, we have further reason to think that (a) the non-existence of God is irrelevant to whether we can do ethics, and (b) a secular ethics is possible.10

In any event, the method I teach my students employs the same approach as Schick; it just includes more criteria, which seem, to me, to also be self-evident. Now, of course, such methods are not perfect; criteria need to be interpreted, sometimes there is not a clear winner, and the process can generate wrong answers. Sometimes, decision makers will have to arbitrarily rank different criteria or decide which duties or rights are more important than others. This, however, is no different than inference to the best explanation; but as the foundation of science, despite its imperfections, inference to the best explanation has produced more knowledge than any other method of reasoning on the planet. It follows, therefore, that despite its imperfections, inference to the best action can and does produce moral knowledge as well.11

But the fact that Schick does not include God’s commands as a criterion in his method of ethical reasoning reveals one of the main points I wish to make: DCT is not useful for moral decision making. Most of the time, one can figure out whether an action will produce the greatest good for the greatest number, or whether it respects rights, can be universalized, or treats others as a means to an end—whether it is merciful, just, or benevolent. But to incorporate DCT in Schick’s method, one would have to ascertain what God’s commands are, what God is like, or even answer questions like “What would Jesus do?” Schick recognizes that what a person thinks God commands or what God is like, or what a person thinks Jesus would do, simply reflects that person’s preconceived notions of morality. Thus, any attempt to use DCT in moral decision making can only be uselessly question-begging. It follows that, while it is possible to do ethics without God, the existence of God would not make doing ethics any easier. This is why I have concluded that the worry that motivates this volume of Religions, the worry about whether ethics is possible without God, is not worrisome at all—that the common belief that “If God doesn’t exist, then all is permissible” is false.

To put it another way: if you, dear reader, are not convinced that a secular approach to ethics is feasible, and thus are worried that we might be forced to embrace moral nihilism—that is (depending on your viewpoint, I suppose) unfortunate. But invoking God can do nothing to alleviate this worry. God’s existence makes moral facts no less “queer” and does nothing to provide what a working moral theory needs: truthmakers and explanations for moral facts, and a way to produce moral knowledge and determine how to act. “God” is no better an answer to the questions of ethics than “magic powers” is a good answer to the question of how Penn and Teller seem to catch bullets in their teeth. If moral nihilism is true, then it’s true—regardless of whether God exists.

7. Conclusions: On Fideism

In conclusion, let me make one final point. In objection to my entire approach, one might argue that there is no reason to assume (as I seem to have done) that God is comprehensible, or would create beings capable of comprehending him—of placing him in the same category as other things which provide truthmakers, serve as good explanations, make knowledge possible, etc. In other words, I have not considered the possibility that God may be paradoxical and incomprehensible and that, if he is, the categories we are capable of understanding cannot be applied to God. However, as a reply to the argument I have presented, this argument fails for two reasons.

First, if this is true, then we should all be agnostic and completely silent about what God is like, what he commands, what he forbids, and even whether he exists. If our limited human concepts cannot apply to God, and human language is incapable of describing him, then anything we say about him will be inaccurate and misleading—blasphemous, even. It can do nothing but lead us astray. Even the suggestion that God exists would be inaccurate, for “existence” is a concept that we can comprehend. Religion, religious statements, and religious beliefs, therefore, should be avoided at all costs. We should all just say “I don’t know” and leave it at that.

Secondly, far from being a reason to reject it, this argument grants my thesis. If God is an indescribable paradox, then he cannot provide truthmakers or explain moral facts (because good explanations, by definition, must be comprehensible and expand our understanding). If God is an indescribable paradox, divine command theory cannot generate moral knowledge and is not useful for moral decision making (because both require knowledge of the divine which, according to this view, is impossible).

Indeed, far from being a refutation of my argument, this view of God would make it clearer than anything else could that neither the existence nor the non-existence of God has any bearing on whether “an ethic is possible”.