Some of humanity’s greatest intellects have tried to prove the existence of God, and any atheist will have to consider these arguments and provide rebuttals, that, by his lights, are sufficient. Arguments for the existence of God fall into two broad categories: demonstrative and non-demonstrative. The former supposedly prove the existence of God with all of the rigor and formality of a mathematical proof. That is, they attempt to show that the existence of God is in some sense necessary. The necessity purportedly established in these arguments is of two types, logical necessity and metaphysical necessity. If God’s existence could be shown either logically or metaphysically necessary, this would be ideal. God’s existence would be truly indubitable, that is, beyond rational dispute. However, deep problems face any purported demonstrative argument for God’s existence.

To say that it is logically necessary that God exists is to say that the denial of God’s existence entails a contradiction. That is, “God does not exist” must entail a proposition of the form “p and not-p.” Various versions of the ontological argument attempt to show this. The problem here has always been to show how it can ever be contradictory to say of any x that x does not exist. Whatever predicates we attribute to x—existence, necessary existence, or whatever—it is not clear how the attribution of such predicates can guarantee the instantiation—in objective, extra-conceptual reality—of anything that bears such predicates.

The problem is not that “existence is not a predicate,” or even that it is only a second-order predicate. For the sake of argument, let’s assume that existence and even necessary existence are ordinary, first-order predicates like “green.” In that case, to say that “the necessarily-existent being does not exist” is indeed contradictory, just as much as to say “green grass is not green.”

But to say that there is no necessarily-existent being is not to say anything at all like “the necessarily-existent being does not exist.” Instead, what you are saying is that, limiting our discourse to extra-conceptual reality (i.e. objective, “out there” reality), there is nothing (no “x”) that has the property of instantiating the concept “necessarily existent being.” Such an assertion denies nothing in the content of the concept “necessarily existent being.” It is not even about that concept. It quantifies over extra-conceptual reality and the predicate it mentions, instantiating a concept, is something entirely different from a concept. The property of instantiating a given concept is as different from that concept as actually being green is from the idea of green.

The upshot is that to deny the existence of God, that is, to deny that the concept “God” is instantiated in extra-conceptual reality, is emphatically not a matter of simultaneously affirming and denying the conceptual content of “God.” Indeed, the atheist can agree with Anselm that “necessary existence” is a part of the concept “God.” The atheist merely holds that no actual, objective, extra-conceptual being has the property of exemplifying the concept of necessary existence. No contradiction.

A second problem with ontological arguments is that as Leibniz recognized, for it to be proven necessary that God exists it must first be shown to be possible that God exists. Leibniz saw the problem as one of showing that perfection is not a contradictory notion, and he believed that he could easily demonstrate this. However, the real problem here is the possibility that gratuitous evil exists and the consequent impossibility that an all-powerful and perfectly good being exists. If it is possible that gratuitous evil exists, then it cannot be proven possible that God exists.

What is gratuitous evil? Intuitively, a perfectly-good and all-powerful being will allow an evil e to exist only if e is a necessary condition for the realization of some good g, where g is good enough to redeem e in the long term. That is, sub specie aeternitatis, e + g is overall better than no e and no g, and e is a necessary condition for the realization of g. Speaking intuitively again, we expect a perfectly good and all-powerful being to have a morally sufficient reason for permitting any evil e, and having a morally sufficient reason for permitting an evil e means that e meets the just-stated condition. That is, e is a necessary condition for the realization of a redeeming good. An evil that is not a necessary condition for the realization of any redeeming good is a gratuitous evil, and if any such evil exists, God could have no morally sufficient reason for allowing it to exist.

It seems at least conceivable that some evils could be gratuitous, that is, they cannot serve as necessary conditions for the realization of any redeeming good. Perhaps some evils are so bad or so pointless that not even omnipotence working throughout eternity can redeem them. That is, perhaps some evil e is such that (a) no realizable good is good enough to redeem e; or perhaps (b) all goods g that are good enough to redeem e are realizable without e, i.e. e is not a necessary condition for the realization of any such good. An omnipotent being working through eternity could realize g without e.

If the idea of gratuitous evil is coherent, and if any such gratuitous evil actually exists, then an all-powerful being would have no morally sufficient reason for permitting that evil, and so cannot be perfectly good. In other words, if it is conceivable that some evils are gratuitous, and if any actual evil is in fact gratuitous, then God cannot exist. How can it be known that no actual evils are gratuitous? This is not the “problem of evil.” It is not a question of whether an atheological argument can be based on the occurrence of evil. It is a question of proving that no actual evil is gratuitous, and it is not at all clear how this could be done, especially since the worst evils might occur in some region of space/time of which we can have no knowledge. Again, God’s existence cannot be proven necessary if it cannot even be proven possible.

Well, what about arguments purporting to demonstrate the metaphysical rather than the logical necessity of God’s existence? Such an argument will not claim that “God does not exist” entails a contradiction, but it will claim that the denial contravenes a self-evident metaphysical principle. Frederick Copleston’s “argument from contingency” in his famous debate with Bertrand Russell is an example of such a purported metaphysical demonstration.

The problem with any attempted metaphysical demonstration is the supposedly self-evident principle it invokes. An instance of such a supposedly self-evident principle would be the principle of sufficient reason (PSR), the claim that nothing exists unless there is a sufficient reason for its existence. However, over the history of philosophy, claims for the alleged self-evidence of metaphysical principles have been protean and elusive. What is self-evident for one philosopher will not be for another. For instance, the PSR is not self-evident for me. If you accept the PSR, you cannot hold that there are any brute facts. However, it seems self-evident to me that there could possibly be brute facts, i.e. states of affairs that are just so with no sufficient reason for their being so.

Defenders of PSR will point to its intuitive nature and the fact that, in neither our scientific or mundane lives are we content with the “explanation” of the old bumper sticker, “shit happens.” We always want to know why it happens. However, if we deny the existence of brute facts, then we either have to say that the chain of causes that we invoke to account for any phenomenon either extends ad infinitum, or it ends with something that has no further cause or explanation, in short, a brute fact. If God is the end of our explanatory chain, then God is a brute fact. Of course, some philosophers have tried to avoid this consequence by saying that God is his own sufficient reason. However, trying to make sense of this, without returning to the concept that God is logically necessary, is notoriously difficult.

The Thomistic answer here is to say that God is Being Itself. However, critics such as Sidney Hook have noted that while it makes sense to speak of being or existence in various senses, it is hard to see how we derive a context-less concept of Being Itself, or “being qua being” from these heterogeneous meanings. The response is that God is actus purus, pure actuality, which is explained as follows:

”When Sidney Hook and company ask St. Thomas what is meant by saying of this table: ‘It has being’ or ‘it is’—the answer is that there is being predicated of the table the real act and perfection which is the basic cause of all other perfections and predicates.” (Quoted in The Quest for Being, by Sidney Hook, Dell Publishing Co., 1963, p. 153).”

Hook Comments:

On this view, Being is not a noun but a verb and modes of Being are modes of action. Metaphysics apparently is the study of action qua action.

By identifying God with the pure “act of being” Thomists indicate that God lacks all potentiality; his essence is to exist. Therefore, his existence is sufficient unto itself, needing no explanation, unlike anything that does not essentially exist. However, I consider Hook’s reply to be devastating:

“This does not escape difficulties.  It only multiplies them. It is just as unclear how we get from the action of this and the action of that as how we get from the Being of this and the Being of that to Being qua Being. The terms ‘act’ and ‘action’ are just as systematically ambiguous as the terms ‘Being’ or ‘existence.’ In many if not most usages, when we speak of ‘act’ or ‘action’ or the behavior of something it clearly presupposes the antecedent existence of some power, material, or subject matter. And when it does not clearly suppose this, it sets a problem for inquiry. Otherwise we suspect the presence of mystification. No matter how ‘pure’ the act is conceived to be, it is linked in our understanding to a preposition; it is an act of. What acts in the act of Being or existing? Certainly not possibilities, essences, or natures. The meaning of ‘death’ is not lethal; the nature of ‘fire’ burns nothing (153-154).”

In short, terms like “pure being” and “pure act” seem to be empty, mystifying abstractions. These terms represent attempts to distill from multifarious senses of being and acting a universal essence stripped of all particular associations. But removing all context means removing all content.

At a more basic level, I can see no basis whatsoever to interpret existence as some kind of act. Hook notes that it is it impossible to say of a purported act of being what supposedly is the subject of that action, i.e. what is supposed to do the acting. It is equally impossible to say what the object is supposed to be, i.e. what is acted upon. An act can only act upon something already there. An “act” that is done by nothing and to nothing does not seem to make sense.

Perhaps, though, Hook (and I) have so far misconstrued the meaning of “pure act.” Maybe an “act” in the Thomistic/Aristotelian sense does not mean an action, something done, like waving or even like exerting electrical force. Maybe it means a state of completion, fulfillment, or actualization with respect to a given potential. In God, uniquely, there is no potential whatsoever. Further, God is not a fully actualized thing (as, perhaps, a Platonic Idea or the Aristotelian Unmoved Mover would be). God is actuality itself, the realized essence of actuality. But the same old problem arises for such a construal. All sorts of tasks get completed, even at a university. Nobody, however, would attempt to reify “completeness itself.” It makes no more sense to see all sorts of potentials actualized and from these to abstract a univocal notion of a reified “actuality itself.”

Attempts to demonstrate that God’s existence is either logically or metaphysically necessary therefore seem to face deep problems. One very popular way of arguing for the existence of God is to treat God’s existence as something like a scientific hypothesis. You then invoke the principles of confirmation theory to supposedly show that theism is more probable given the evidence than naturalism. Thus, if T is theism and N is naturalism, and the evidence E is some feature or features of the universe—its existence, orderliness, or its possession of certain interesting features such as life, mind, or morality—your argue that p(T/E) > p(N/E). For this argument to work, the “relevance condition” must be met, that is it must be shown that p(E/T) > p(E/N). That is, for any evidence to confirm theism over naturalism, that evidence must be more likely given theism than naturalism. Versions of such an argument are commonly found in the writings of theistic philosophers, including Richard Swinburne, George Schlesinger, and Robin Collins.

In a sense, this is a very easy argument to make. If we are considering E to be such features as the existence of life, mind, and morality in the universe, and if T is the hypothesis that the Christian God exists, and if N is merely the denial of anything supernatural, then clearly, p(E/T) ≫ p(E/N). Given the existence of an all-powerful being that wants life, mind, or morality, the probability of getting such things is one. Given only the “hypothesis” that there is no supernatural reality, practically nothing follows about the physical universe. The only way that a naturalist could win in this situation would be to make the background probability of God nearly zero, and this would draw the charge of dogmatism or close-mindedness.

Clearly, though, such a “victory” of theism over naturalism is due to an extremely uneven playing field. Theism is loaded with lots of content, just the content needed to imply the desired results, and naturalism is given practically no content. But why should a naturalist ever agree to define terms in such a tendentious and lopsided way?

As I see it, naturalism, or—as I prefer to call it—physicalism says that the fundamental features and (if the universe is conceived as having a beginning) the primordial state of the universe are to be taken as given. That is, the universe, or rather its set of basic and original features, is taken as a brute fact, neither having nor needing any further explanation. The universe, with its distinctive laws and fundamental entities, is neither probable nor improbable. It just is. Put another way, the physicalist holds that if there is a physical “rock bottom,” a fundamental set of laws, entities, and states of the universe, our explanatory task will be done when these are reached. If and when physics finishes its task, the universe will have all the explanation it needs. Specifically, there will be no need to further invoke a creator.

Physicalism as so defined is not a competitor to theism. They are on different levels. Theism, at least in its more plausible forms, does not oppose physical explanation but attempts to incorporate it and transcend it. True, some theists in the past and even today (e.g. creationists and “intelligent design” theorists) have seen theistic hypotheses as opposed to aspects of scientific explanation. Fields such as evolution and neuroscience are still seen as battlegrounds by some theists, i.e. areas where the push for comprehensive naturalistic explanations must be opposed. However, the practice of opposing science with religion has been a losing project for several centuries, and there is no prospect whatsoever that current efforts will succeed where past ones failed.

The more promising approach is to see theism not as a competitor to science but as needed to fulfill the explanatory job that science so successfully accomplishes in its own domain. Theism argues that there are some things that need explanation but which science, in principle, cannot explain. For instance, the fact that there is a physical reality at all is not something that science can explain. Natural science explains features of the universe by appeal to other (broader, deeper, or more fundamental) features of the universe, and so must assume the existence of a physical reality. Theists might therefore argue that the existence of a physical reality needs to be explained and that theism is the only, or the best, explanation for the existence of a physical reality.

Probably the most popular theistic argument today is the fine-tuning argument (FTA). This argument claims that fundamental physical constants cannot be explained scientifically. Qua physically fundamental, they cannot be explained by science because all scientific explanation presupposes that those parameters have the values they do. The physicalist could reply that these so-called fundamental constants are perhaps not really fundamental and that it is still an open question whether deeper theories will someday account for them and explain why they have the precise values they do. Perhaps some ultimate Theory of Everything will fulfill Stephen Hawking’s desire to have a set of equations succinct enough to fit on a t-shirt, yet powerful enough to explain all physical constants. Or maybe a version of inflationary cosmology will be strongly confirmed—one that implies the existence of an infinity of universes where the basic physical parameters vary randomly with respect to each universe. In that case, it would be unsurprising that we find ourselves in one of the universes where conditions permitted our existence.

Defenders of the FTA argue that any such move would not solve the problem of fine tuning, but would only relocate it. Any deeper physical theory that explained the values of the physical constants now regarded as fundamental, would have to introduce new parameters which themselves would have to be finely tuned. That is, any more comprehensive successor to current theories would necessarily introduce its own constants and the values of these constants would have to be very precise for the theory to have the explanatory power that it would have. Further the values those new constants have would only be one selection from an apparently infinite number of other seemingly possible values.

I think that this argument is correct. Any postulated ultimate explanation will unavoidably introduce posits that themselves, qua ultimate, have no further, deeper, or more comprehensive explanation. They would be brute facts. Any ultimate physical reality will instantiate only one of innumerable possibilities, and, defenders of the FTA will say, there will still be no explanation why this was the one reality instantiated out of innumerable possibilities. Again, I see this reasoning as impeccable. However, the problem posed here is one that necessarily lacks a solution.

Ultimate physical reality is inexplicable, not because it is physical, but because any ultimate posit is necessarily inexplicable. Suppose we try to allay the problem by introducing a supernatural fine-tuner, one that chooses the values of the physical constants. Does this solve the problem?  Has everything now been explained? Of course not. The problem has just been relocated. What motivates the FTA is the intuition that we have been impossibly lucky to get just the values of the physical constants needed to have a universe hospitable to complex life forms like us. If we introduce another level of explanation—a supernatural level above the natural one—the problem just arises again there. How were we so impossibly lucky that, of all possible ultimate supernatural entities, we happened to get one that both wanted complex life forms such as us and had the power to create the necessary initial conditions?

Prima facie, no limit could be fixed on the number and kinds of ultimate supernatural realities that might have existed. “Supernatural” seems to be primarily a negative characterization, one that does not even specify a well-defined class. To say that something is a “supernatural” force, entity, occurrence, or whatever would only seem to imply that it is non-natural, i.e. something not part of the natural universe.  How many different kinds of non-natural beings might there be? The only limits on speculation here would seem to be the limits of our imagination. Humans have, in fact, conceived of quite a number of different of non-natural realities: mana, juju, the Tao, Atman-Brahman, demons, angels, gods, jinn, Plato’s Ideal Forms, Plato’s Demiurge, Heraclitus’ Logos, Plotinus’s One, and Aristotle’s Unmoved Mover. Therefore, there seems to be no way of even reasonably speculating on what percentage of possible ultimate non-natural beings would and could create beings like us. We seem to be dealing with an absolute imponderable here.

The upshot is that the question of why our universe has the constants it does has, and can have, no ultimate answer. We cannot give a finally satisfactory answer, one that leaves no brute facts and no unanswered questions. We can only relocate the question. In that case, we are free to choose which set of ultimate brute facts will satisfy us. The physicalist will say that when—if—we reach some ultimate set of physical facts, our explanatory work will be done. The physicalist will see no purpose, no explanatory advantage, in invoking a nebulous, inscrutable supernatural fine-tuner who can do no more than introduce another set of endless ultimate possibilities.

When pushed into a corner, theists will often insist that only theism can answer two questions that must be answered: (a) Why do we have this reality rather than some other that we can imagine? And (b) Why is there anything at all, i.e. why is there something instead of nothing?

The first question is just a generalized version of the question that motivates the FTA, and it is to be given the same answer: There is and can be no answer. Anything posited as ultimate, including God, will only be one of indefinitely many unrealized alternatives. The only way to escape this is for our ultimate posit to be the one, unique, logically necessary being as the ontological arguments attempted to show. But we have already seen that those arguments have deep problems.

With respect to the question “Why is there something instead of nothing?” the full, complete, and satisfactory answer is “Why not?” The assumption behind the question of why there is anything seems to be what Adolf Grünbaum called “the spontaneity of nothingness.” This would be the assumption that the normal, natural, and spontaneous state is for there to be nothing, and so the existence of something must be explained. But why make this assumption?  Why not instead assume that the normal, natural, spontaneous state is for there to be something rather than nothing?

In the final analysis, most of the arguments for theism have one thing in common: They invoke God because the physical universe is, in some sense, alleged to be inadequate. On this view, there is something vital that is missing if we only invoke the physical universe, something that can only be made right by transcending the merely physical and postulating the non-physical. The physicalist replies that the universe is enough. The universe has the resources to answer all of our legitimate “why” questions. No rational purpose is served by moving “beyond” the universe to posit some occult, supernatural entity. Such moves only land us in a quagmire of mystification. The physicalist is convinced that the physical universe holds the answers to all our answerable questions.

All arguments for the existence of God ultimately rest upon metaphysical intuitions, e.g. the intuition that perfections entails existence, or the intuition that nothing exists without a sufficient cause, or that whatever begins to exist must have a cause, or that consciousness cannot have a physical explanation, or that there must be an explanation of why this reality exists or any reality at all. However, it seems to be the nature of all such intuitions that they admit of rational doubt. As a matter of fact, many seemingly rational persons do doubt them. Where can rational argument go if someone rejects the fundamental metaphysical intuitions upon which your argument is based? Well, you might dismiss such persons as liars or cheats or self-deluded because they deny that they have such intuitions when they really do. One can take such a position, but then he or she should not be surprised if the other side is reciprocally dismissive. Really, if the debate reaches this point, it is fruitless to continue. Intuitions are a great place to begin a debate, but a terrible place to end one.  This, however, seems to be the ineluctable fate of arguments for the existence of God, to grind to an ignominious halt with an irremediable conflict of basic intuitions.