Agnosticism Is Not Reasonable
Refusal to make a choice is a choice of its own. Although often presented as the intellectually humble third option between belief and atheism, Stuart Doyle argues that agnosticism presents a false middle path that is neither coherent nor practical.
If one cannot prove something to be true or false, then one should not make any claims about it either way. This seems eminently reasonable. It is a good rule of thumb, especially in science. After all, this is basically the definition of science according to the most influential theory of demarcation: Karl Popper’s falsificationism. However, when this way of thinking is exported beyond the borders of science, it becomes unreasonable—especially when it leads to agnosticism on the question of God. Despite its urbane sheen of measured rationality, agnosticism always entails a self-contradiction.

Bertrand Russell thought that claims about God are like claims about a teapot orbiting the sun between Earth and Mars—that there should be serious proof before such claims are accepted as true. What Russell did not appear to appreciate is that, while no one needs to make a decision about orbiting teapots, everyone must make a decision about God. The agnostic, just like everybody else, must choose what to do with his life. Every decision involves the loss of alternative possibilities. And one has to choose. If one puts off the decision, that is a choice.
“The agnostic does not avoid taking a side; he takes a side and calls it neutrality.”
If one kills oneself to avoid choosing, that is also a choice. It is true that we are “condemned to be free.” One must choose whom to forgive, what to protect, and when to tell the truth at a cost. There is no opting out because that is still a choice of what to do. This puts one in the inescapable position of figuring out what one ought to do. What will guide a person? It cannot be nothing because one has a mind and will act. The individual will decide.
As the philosopher Søren Kierkegaard put it:
“What I really need to get clear about is what I must do, not what I must know, except insofar as knowledge must precede every act. What matters is to find a purpose, to see what it really is that God wills that I shall do; the crucial thing is to find a truth which is truth for me, to find the idea for which I can live and die.”
So which ideas guide one’s decisions? The objective facts? Certainly not. There is—in principle—no accumulation of facts about how the world is that can tell one how the world ought to be. This "is-ought" distinction is known as Hume’s law or Hume’s guillotine: an ethical or judgmental conclusion cannot be inferred from purely descriptive factual statements. Measurable facts can only inform us about how to do what we want to do, not what we should want to do. Prominent agnostics try to figure out what they ought to do while maintaining an air of detached rationality. However, this does not hold up under examination once the veneer is stripped away.
For example, consider cognitive scientist Steven Pinker’s ultimate guiding value, described as “reason, science, humanism, and progress.” This is a typical example of what cosmopolitan agnostics endorse. On first pass, it almost sounds like it could be something that logically follows from the scientific facts of the world. But the meaning of “reason” and “science” in the slogan is that reason and science are tools for advancing humanism, and “progress” means movement toward humanism. So Pinker’s guiding value is just humanism, which is not something that can be found in the objective facts of the world.
Usually, humanism is defined as seeking well-being for humans. But “well-being” hides many assumptions. Well-being means “a good condition of existence.” But then what counts as “good”? Humanism is supposed to define what is good, but it depends on preconceptions of good. There is no “good” particle in physics. There are no rocks or plants that are good or bad. In the guiding principles of “reason, science, humanism, and progress,” reason and science are slaves of subjective judgment. Pinker’s particular judgments of good are not crazy—less disease, more curiosity, more enjoyment—but they are subjective judgments, not “science and reason.” Pinker’s ultimate guiding value is not naturalistic, and neither is yours.
As we continuously make choices in life, each of us must performatively assert some nonphysical, nonobjective guiding value that tells us what to do. That guiding principle can be God or something else. But either way, it is not anything that falls out of the objective facts of the physical world. Not only does one’s ultimate value guide one’s actions; it necessarily adjudicates rival values. That is, it must take a position on each competitor. All of the other candidate ultimate values must be either rejected or incorporated. Otherwise, one’s ultimate value cannot tell one what to do when it conflicts with other values.
Imagine a person calls his ultimate value “human flourishing,” but he remains “agnostic” about the claim that God has authority over his life. What happens when obedience to God-as-described conflicts with his best account of human flourishing? If his ultimate value cannot say “ignore God,” “defer to God,” or “reinterpret God as a poet’s metaphor,” then it is not his guide. His guide is whatever adjudicates between values.
This matters because the agnostic posture often works like this: “I accept humanism (or science, or autonomy) as my practical compass while suspending judgment on God.” But that is not suspension; that is ranking. The agnostic is already placing humanism over theism in potential conflict cases, and that is a judgment on God. The agnostic does not avoid taking a side; he takes a side and calls it neutrality.
For example, Alex O’Connor, the popular philosophy podcast host, professes to be agnostic about God. But he affirms an “emotivist” view of morality. This view is much like that of David Hume, who, upon observing the is/ought distinction, concluded that when we call actions “right” and “wrong,” we are just expressing sentiments about the actions. Hume was not merely saying that right and wrong actions make us feel emotions. His claim was that right and wrong have no meaning beyond emotional expression. On this view of morality, shared by O’Connor, right and wrong are just a matter of what people like and dislike. But this conflicts with his supposed agnosticism.
If right and wrong are just emotional expressions, then there is no higher purpose to human behavior, nothing at all we ought to do in virtue of our standing in relation to the transcendent. And this could not be the case if God—even vaguely conceived—exists. O’Connor’s ideas about morality imply a hard rejection of God as the source of morality and purpose rather than an open-minded suspension of judgment. To assert a version of morality is to reject all competing versions, including those that would follow from God’s existence. And everyone asserts a version of morality.
Some might protest that I am stretching definitions—that I am basically saying everyone has a “religion,” even when they deny it. But it is not a stretch. If there is a basis of something, then consider its necessary characteristics. It sets the top priority, excludes competing views, can demand sacrifice, and is not based on demonstrable facts. Consider the normal definition of religious values: “relating to or manifesting faithful devotion to an acknowledged ultimate reality or deity.” To serve some nonphysical something to the exclusion of all else is faithful devotion to an ultimate reality. The only difference is that it is not always acknowledged.
Some ultimate values are explicitly codified, and some are not. Some people are devoted to articulated ethical principles such as utilitarianism. Others value nothing more than being accepted by their desired social class, and some value nothing more than physical comfort. Sometimes, a person’s true guiding principle is selfish but couched in another religion that does not seem selfish. Many have noted how the so-called “woke” movement serves as a religion for those who mistakenly think of themselves as nonreligious. And Rob Henderson has noted how these beliefs are often worn as luxury goods by those whose true priority is narcissistic aggrandizement. In this way and many others, people lie to others and to themselves about what their gods are. But everyone who lives out an “ought” in the world has one god or another.
I am not directly arguing here that God exists. There are plenty of arguments for and against that claim. I think some of these arguments are rather compelling, but not so compelling that it does not take faith to believe. It would be impossible to walk through all of the arguments here, as well as to explain exactly where each of them might fall short. For the purpose of this article, I will just make the plausible claim that after the many arguments for and against God’s existence, it still takes quite a jump to assert either conclusion. But refusing to take that leap is a self-contradiction because there is no fence to sit on.
This is one way to understand the notoriously hard-to-pin-down religion of Jordan Peterson: he understands the untenability of agnosticism. He is currently the most influential proponent of the idea that in order to act, one needs values, and that one cannot get values from the facts of the natural world. He is painfully aware that this condition demands a god to sit at the top of the hierarchy of values. He wants very badly to explain how—in a functionalist sense—God exists. But he will not take a leap and assert the independent existence of God. Instead, he always defines God merely as the roles he fills in human experience. To Peterson, God is conscience, God is the call to adventure, and God is the top of the hierarchy of human values. It is evident that all of these functions exist as aspects of humanity. So by defining God as human functions, Peterson can state in a trivial sense that God exists.
But positing this thin functionalist god is an attempt to avoid both theism and atheism, and so it entails a contradiction. Functionalist definitions take away any independent existence of God as his own being. It is like defining my son as “the entity that plays in my living room.” This definition has no power to differentiate between my son and my cat because they both serve the function of playing in my living room. Just as other entities can play in my house, other values can be placed at the top by humans.
The top values are god-like in their human function but not in their ontology. Since there is no single “top of the hierarchy,” Peterson must stipulate that he is referring only to the “proper” top value. But that means he needs criteria for what counts as a good fit for the top of the hierarchy. And that means Peterson has a god-judging criterion, which is the real top of his hierarchy. God himself does not occupy the absolute top of Peterson’s hierarchy, despite Peterson defining God as the value at the top. He judges god-candidates by their ability to avoid producing miserable, tyrannical societies over time. So his real god is something nearly indistinguishable from the “human flourishing” that secular humanists like Pinker worship. This is the underlying defect people sense when they see Peterson debate atheists. He has diagnosed the contradiction inherent in scientistic agnosticism but has not left it behind.
Peterson tries to say nothing about God that cannot be inferred from human functions because human functions are empirically observable. It is a subtle way to talk about “God” without taking a leap of faith. As comforting as it may seem to have the cake of naturalism and eat it too, it is not a real option. The natural world does not tell us how things should be. To avoid self-contradiction, there has to be an intentional leap beyond what can be verified by “science and reason.” Again, Kierkegaard understood this:
“And how does God’s existence emerge from the proof? Does it follow straightway, without any breach of continuity? … When I let the proof go, the existence is there. … Must not this also be taken into account, this little moment, brief as it may be—it need not be long—for it is a leap.”
A conscious leap of faith, though not itself an act of reason, resolves the catastrophic incoherence of agnosticism. The agnostic already lives out values taken on faith but denies that he does. And his covert piety means he takes sides against rival values. If humanism or scientism or aestheticism quietly trumps God whenever they clash, then agnosticism is a costume. Life forces a ranking. The only question is whether one will confess his. One can assert the unprovable nonexistence of God, one can lie to oneself by pretending to remain neutral, or one can take a leap of faith and believe in God. None of these are exactly reasonable. Hard theism and hard atheism both assert the unprovable. But the false middle path is least reasonable of all because it entails acceptance of something unprovable, plus layers of contradiction and self-deception on top.
Since I have not actually offered any proof of God’s existence, it would be no surprise to hear a devoted agnostic such as O’Connor insist, as he has, “OK, man… I still don’t know if God exists. That is just true; I don’t know.” But the point is that there are some situations where one just does not get to “know.” Yet indecision is the least rational option. This is classically captured in the thought experiment of Buridan’s ass, where a donkey stands equidistant from two identical piles of hay. The donkey has no rational reason to choose one pile over the other—no way to “know” which one to choose. So he rationally suspends judgment and starves to death in his rational indecision. “I just don’t know” is the anthem of those who insist on making a Buridan’s ass of themselves.
“I know thy works, that thou art neither cold nor hot: I would thou wert cold or hot… Thou sayest, I am rich, and increased with goods, and have need of nothing; and knowest not that thou art wretched, and miserable, and poor, and blind, and naked.”
—Revelation 3:15–17
Stuart Doyle is a Ph.D. student in clinical psychology at the University of Kansas. He holds a B.A. in neuroscience and behavior from Columbia University and an M.S. in criminology from the University of Pennsylvania. He is a veteran of the United States Marine Corps.