Updated: Feb 12
In 2013, I published an article for the journal Think entitled “Do Souls Exist.” In 2016, it was “Does Free Will Exist.” Now, in 2019, I’m working to complete the trilogy with “Does God Exist?” In it, I will lay out the most popular arguments for and against God’s existence. I show why attempts to prove God exists fail and argue that there is good evidence that God does not exist. But evidence is one thing; hard 100% proof is another. So, this got me thinking: It is possible to prove that God doesn’t exist?
In debates about the existence of God, the question “Can you prove that God doesn’t exist?” is often asked. Now, as an attempt to provide a reason for believing God exists, this move simply commits the appeal to ignorance fallacy. The fact that you can’t prove something is false is not a reason to think it is true; likewise, not being able to prove that something doesn’t exist is not a reason to believe that it does. After all, I can’t prove that unicorns don’t exist; one could be hiding somewhere I can’t look—the universe is a big place. But that doesn’t mean unicorns exist or that it is rational to believe so. When it comes to existential matters (questions regarding what exists and doesn’t), the burden of proof is on the believer. Until evidence is provided that some thing exists, believing that thing doesn’t exist is the rational position.
A Legitimate Question
But the question “Can you prove that God doesn’t exist?” is also asked legitimately. What do I mean by that? Well….
It’s often said that “you can’t prove a negative,” as in “you can’t prove that something doesn’t exist.” And usually, this is true. Like with unicorns, the thing in question could always be somewhere you can’t look. But you can prove that something doesn’t exist by showing the very concept of it to be logically impossible. I can, for example, prove that there are no square-circles by showing the very concept of a “square-circle” to be a contradiction in terms. There cannot be a “four-sided object with no sides.” Because it contradicts itself, that string of words it not meaningful in the English language. I can know “there is a square-circle” is necessarily false just like I can know that “Bay egg jump top” is necessarily false, and for the same reason: it’s non-sense. It’s meaningless.
The reason that it might be possible to prove that God does not exist is that, upon examination, it seems that the very concept of God is logically contradictory; it’s non-sense. “There is a God” might be as much non-sense as “there is a square circle.” How so?
Although God has not always been conceived of as a perfect being, and biblical scholars agree that the God of the Bible is not perfect, God has been defined by theists as perfect since about the time that Augustine incorporated Plato’s ideas about a demiurge (a perfect being that ordered the universe according to the Platonic forms) into his Christian theology.Today, theists think that God’s perfection entails that he is (among other things) omnipotent (all-powerful), omniscience (all-knowing), omnibenevolent (all-good), omni-present (fully-everywhere), immutable (unchanging), perfectly-just, perfectly merciful, and perfectly free.
The worry, however, is that it is not logically possible for a being to have all these properties. Some of the properties, by themselves, might be incoherent. (What does it even mean for something to be “fully-everywhere”?) But others seem to contradict one another; if a being is one, that being can’t be the other.
The most common worry along this latter line is “Can God create a bolder so big that he cannot lift it?” If he can’t then he’s not all-powerful, but if he can then his power can be limited (so he’s not all powerful). A seeming paradox. A similar worry is whether God can be all-powerful but also all-good. If he is all-good then, by definition, he cannot do evil; but if he is all-powerful then, by definition, he can do anything including evil. Paradox! The very concept of a perfect being seems logically contradictory.
Atheist William Rowe argues that these paradoxes can be solved. And to do so, he doesn’t just say “God is beyond us; we can’t understand how God can have the properties he must.” For one, this just begs the question; it assumes the truth of what it is trying to prove: that a being with such properties exists. More importantly, however, Rowe understands that logical contradictions can’t be true. So, Rowe would admit, if the concept of a perfect being where logically contradictory, one could not exist. Rowe argues, however, that such a concept is not logically contradictory—at least, not for these reasons.
To understand his argument, however, we need to understand why philosophers agree that God could not do anything that is logically impossible (and yet still be omnipotent). In short, to be omnipotent is to have the power to make any proposition true. But sentences that express logical contradictions—like there is a square circle—are not really propositions. Why? Because they are contradictory and thus don’t make sense. So God can’t make them true, because they literally can’t be true—but that doesn’t mean that God is not all-powerful. (Think of it this way. God can’t make “Bay egg jump top” true, or “he-ba-ge-ba-bling-bling-blah” true either, but this doesn’t mean that he’s not all powerful. God can do anything, but those aren’t really things. Likewise, a square circle is not really a thing.)
Well, Rowe argues that wondering whether God could rob himself of an essential property—like omnipotence or omni-benevolence--is like wondering whether he can make a square circle. Of course he can’t, Rowe argues, but that doesn’t mean he is not all-powerful. Of course a being with unlimited power can’t limit his own power (by making a bolder he can’t life). That is like asking if he can make a square circle. And God can be all-powerful, without having the ability to do evil, because “There is an all-good being that can do evil” is just as logically contradictory as “there is a four sided object with no sides.”
Why a Perfect Being is Impossible
Not everyone finds Rowe’s argument convincing, however. If God is all powerful, he should be able to do anything that I can do; and since I can do evil (believe me!), God’s being all-powerful must entail he can too. Now, Rowe is right that “there is an all-good being that can do evil” is logically contradictory, just like “there is a four sided object with no sides.” But “there is an all-good being that can do evil” is what must be true if there is a being that is both all-good and all powerful. Thus the very concept of a perfect being is logically contradictory, like a square circle—or so the argument goes.
But here’s the thing: even if you don’t like that counter-argument, and think that Rowe is right—even if Rowe is right, and God being unable to rob himself of an essential property does not rob him of omnipotence—God still can’t exist by definition. Why? Because, as Rowe himself argues, if God cannot rob himself of an essential property, then he cannot rob himself of omni-benevolence. And if he can’t do that, he can do nothing but what he must do: the absolute best and perfect thing, in every circumstance. If creating the universe is best, then God must create it. Indeed, he must create the best possible world. If saving a drowning baby is morally best, he must do that. He cannot act otherwise. But (as I argued in “Does Free Will Exist”), freewill requires the ability to do otherwise.God cannot be free without it.So if God, by definition, must be both all-good and perfectly free, then God, by definition, cannot exist.
A similar problem is raised by God’s omniscience. There is some debate about whether God knowing what you will do before you do it limits your ability to do otherwise and thus your free will. (I have argued, elsewhere, that it clearly does.) But no one can argue that if you knew, infallibly, what you were going to do before you did it, that you could still do what you do freely. If you infallibly know what you are going to do before you even do it, how could you even deliberate? How could you even choose? Even if you are a compatibilist about free will, and don’t think that alternate possibilities are required for free will, you would have to agree that to freely choose something you have to be able to deliberate, decide, or choose. You can’t freely choose to do something that you already knew you were going to do. But, of course, if God is all-knowing, he infallibly knows what he will do before he does it; thus, by definition, God cannot be free.
Granted, some will argue that God does not know what he does before he does it because he is a transcendent immutable timeless being. But the concept of such a being making free choices is just as problematic; the very action of choosing (which must include a process of deliberation) is a timed process. Indeed, such a being would be unable to do most of the things that theists traditionally think a perfect being must do: love (which includes being moved by suffering), answer prayers (which includes responding to requests), be a person (which arguably requires a body), or even forgive sins.
Speaking of the latter, there is one more divine paradox worth mentioning: the logical incompatibility of perfect-mercy and perfect-justice. Justice requires someone getting what they deserve; being perfectly just, thus, is making sure everyone gets what they deserve. Mercy, by contrast, requires someone not getting what they deserve; being perfectly merciful, therefore, is to ensure that no one gets what they deserve. Clearly, then, no being can be both. One could give justice to some, and mercy to others. But it is logically impossible to ensure that everyone gets what they deserve, and that no one gets what they deserve.
Christians, of course, insists that God accomplished this feat through the crucifixion of Jesus. This, they argue, was perfectly-merciful because it allowed everyone to be forgiven, but also accomplished perfect justice by ensuring that the proper punishment for all sins was doled out. But justice requires more than just the doling out of punishment; the punishment must be inflicted upon the person who deserves it. And, as Christians willfully admit, Jesus did not deserve it; indeed, he was the person most undeserving of punishment. If a judge found someone guilty, but decided to punish the victim instead, we would not think that justice has been served. The person who does the crime, must do the time. And the same would follow, even if the judge decided to punish himself. That would be merciful; but, by definition, it would not be just.
Unless these paradoxes are solved (and I have yet to see a satisfactory answer to any of them) it does seem that one can prove that God (i.e., a perfect being) does not exist. One can’t 100% prove unicorns don’t exist because on can’t look everywhere. But one can show the concept of a perfect being to be logically incoherent.
The Response of Mythical Faith
In reply, some academics would argue that I am taking belief in God (and religious belief in general) too literally. Susan Armstrong, for example, thinks it wrongheaded to think of God as a being with properties; she instead thinks of God as an ultimate reality that all religions attempt to appreciate and worship, and emphasizes what can’t be said or known about the divine. Indeed, according to Armstrong, one of the biggest mistakes of modern religion is (what we might call) science envy—thinking that it can define, set forth, and defend the existence of God like a scientific doctrine. It cannot. Instead, she thinks religious language is largely non-literal and sees religions as a set of practices, rather than a set of doctrines.Arguing that Karen Armstrong’s God doesn’t exist (because he has contradictory properties) would be like arguing that Santa Claus doesn’t exist (because physics wouldn’t allow him to visit every home in one night). Of course he doesn’t, but it doesn’t matter because his literal existence isn’t the point. It’s the practices that surround him (including people pretending and acting like he exists) that matter.
As a reply to the arguments I presented, however, this falls short. Why? Because pretty much only Armstrong (and a handful of academics and theologians like her) embrace this kind of non-literal view of religious language and doctrines. And by pointing this out, I am not committing an “appeal to the masses” fallacy, where one thinks that a majority of people believing something is what makes it true. Why? Because Armstrong’s thesis is about the nature of religious language; she thinks it is non-literal. This cannot be true if the vast majority of religious believers use religious language literally. And that they do is undeniably the case. The vast majority of Christians believe that Jesus literally existed, that he literally performed miracles, and literally rose from the dead. Devout Jews literally believe that Moses parted the Red Sea, and Muslims literally believe that Mohamad rode a horse named Buraq up to heaven. And they all literally believe that God literally exists.This seems undeniable; indeed, I believe it is this common view of religious language that Armstrong is trying to combat with her arguments. But until she is successful, arguments which suggest that literal belief in a perfect being is logically incoherent—that a perfect being cannot literally exist—are appropriate. They are not wrongheaded; they do not commit the straw-man fallacy. These arguments are simply meeting people where they are at.
Indeed, even though Armstrong would likely answer “yes” to the question “Does God exist?”, I believe that most theists (if they knew Armstrong’s views regarding religious language) would actually consider her to be an atheist. She’s like John Caputo, who says that God doesn’t exist, but insists—that, even though God doesn’t literally exist, the concept of God is still relevant and “calls upon us, lures us, solicits us” to act a certain way. Or consider Tamar Szabó Gendler, who contrasts beliefs with aliefs—belief-like attitudes one takes that causes one to act like something is true even though one knows that it is actually, literally, false. Isn’t that, essentially, what Armstrong is doing with God? She doesn’t believe God exists. She believes God exists.
Such belief is not faith in the traditional sense—literal belief in something in the absence of evidence or despite evidence to the contrary. It’s what I call “mythical faith.” It is to treat, for example, the stories of the Bible like fans treat the stories of The Lord of the Rings: as stories that one knows are actually literally false but that contain moral lessons that affect one’s lives—that call us to live a certain kind of way.
I’m not here arguing that there is anything wrong with mythical faith.I’m just saying that people like Armstrong embracing non-literal belief cannot save the common literal belief in God from the paradoxes I articulated earlier. Yes, those arguments do not apply to the beliefs of people like Armstrong, who do not take religious language literally. But since theirs is the minority view, it is not a mistake to present arguments against the literal view, as I have done here.
David Kyle Johnson, 2020
This brief history of the concept of God is, of course, oversimplified. For the full story, see Daniel Dombrowski’s book A History of the Concept of God: A Process Approach (SUNY Press, 2018). See also Karen Armstrong’s Book A History of God (Ballantine, 1993). For more on such paradoxes see Martin, Michael and Monnier, Ricki. “The Impossibility of God”. (Prometheus Books, 2003.) Asking whether he can “make a bolder so big that he cannot lift it” is like asking whether he can make “There is a limitless being that has limits” true. Asking whether God has the power to do evil is like asking whether he can make “there is an all-good being who does evil” true. Each, Rowe argues, is a contradiction in terms.  Rowe, William. The problem of divine perfection and freedom. In Reasoned Faith, edited by E. Stump, 223-333. (Cornell University Press, 1993.)  Of course, one might try to embrace compatibilism of some sort to get out of this problem, but most theists are not compatibilists. Indeed, they are staunch libertarians; so this “way out” of this problem will not be palatable for most theists.  It important to note that, if God can’t be free, he also can’t get moral credit for what he does. As I will mention later, one can only receive moral credit for doing what one was free to do (or not do). So the very concept of omni-benevolence seems self-contradictory. Being all-good means that one is not free; being free is necessary for being all-good. So being all-good is logically impossible. “God, fatalism and temporal ontology.” Religious Studies, 45(4): 435–54 (2009). See Drange, Theodore M. “Incompatible-properties Arguments: A Survey.” In The Impossibility of God, edited by Michael Martin and Ricki Monnier. 185-197. (Prometheus Books, 2003.) Karren Armstrong, “The Case for God.” (Alfred A. Knopf, 2009.)  I should add that this is also true of religious believers in the past. Although Armstrong is right that Biblical Literalism (the idea that the Bible is literally true cover to cover) is a recent modern post-Darwin invention, it is undeniable that religious belief in general is and always has been largely literal. Christians have never taken the gospels to be as mythical as The Lord of the Rings, or to believed that “God exists” is a useful fiction. Although, for a good review of Armstrong’s book, which explains why the theistic masses find Armstrong’s kind of religious belief unsatisfying, see Ross Douthat's review for The New York Times (Oct 1, 2009), "Perpetual Revelations," at https://www.nytimes.com/2009/10/04/books/review/Douthat-t.html