There is a particular kind of philosophical argument that feels, on first encounter, as though it ought to work. The sentences are grammatical, the vocabulary is precise, the conclusion follows from the premises with apparent inevitability, and yet something has gone wrong somewhere, and you cannot immediately put your finger on where. The ontological argument for the existence of God is the supreme example of this genre. It has been offered, refuted, revised, and re-offered for roughly nine hundred years, and it continues to attract serious defenders precisely because the sleight of hand at its centre is subtle enough to withstand casual inspection. The purpose of this essay is to make that sleight of hand visible, to trace it through both its classical and its modern forms, and to show, in terms that require no professional training in philosophy, why no combination of words, however carefully arranged, can conjure a being into existence.
The stakes are worth stating clearly at the outset. The ontological argument is not a marginal curiosity confined to medieval theology. It is routinely presented by contemporary apologists as the strongest purely rational case for theism, the one argument that works without appealing to cosmology, biology, or religious experience. If it succeeds, God’s existence follows from the bare concept of God, prior to any empirical investigation whatsoever. If it fails, and this essay will argue that it fails decisively, then an entire genre of theological reasoning collapses with it, and the apologist must retreat to ground that is, as we shall see, no more secure.
1. Anselm’s Original Move
Anselm of Canterbury formulated the argument in the Proslogion, written around 1078. His framing was devotional rather than combative; he was not producing a polemical tract against atheists but addressing God directly, attempting to demonstrate what he already believed. The prayer-like register is important because it colours the argument’s internal logic. Anselm was not trying to discover whether God exists; he was trying to show that God cannot coherently be denied. The argument, stripped to its bones, runs as follows.
We possess the concept of God as “that than which nothing greater can be conceived.” This much, Anselm insists, is uncontroversial. Even the fool who says in his heart there is no God must understand the concept he is denying, and that concept is of a maximally great being. Now consider two candidate beings: one that exists only as a concept in the mind, and one that exists both in the mind and in reality. Anselm’s answer is that the being existing in reality is greater, because existence in reality adds something to mere conceptual existence. Therefore, if the greatest conceivable being existed only as a concept, it would not be the greatest conceivable being, since we could conceive of a being just like it but with the added property of real existence, and that further being would be greater. The greatest conceivable being must therefore exist in reality, or the concept is self-contradictory, and so the argument arrives at its destination: the greatest conceivable being, which Anselm identifies as God, necessarily exists.
It is worth pausing to acknowledge what Anselm achieved, even in failure. He identified that the concept of God carries an unusual logical structure, one that invites us to ask whether existence belongs among the properties that constitute greatness. He forced philosophy to grapple seriously with the relationship between concepts and reality, between what we can think and what must be the case. These are genuine philosophical questions, and the centuries of argument his Proslogion generated were not wasted, even if the conclusion he drew was wrong. The mistake, when we locate it, is not one of carelessness but of a very deep confusion about what predicates like “existence” actually do, a confusion so deep that it took nearly seven hundred years to identify with full precision.
2. Gaunilo’s Island and the Parody Objection
Anselm did not have to wait long for a rebuttal. His contemporary Gaunilo of Marmoutiers responded with a parody that remains, after nine centuries, one of the cleanest philosophical demolitions in the literature. Gaunilo asked us to consider the concept of a perfect island, one than which no greater island can be conceived. By exactly Anselm’s logic, this island must exist in reality, because an island that existed only in the mind would be surpassed in greatness by a perfect island that also existed in reality. If the argument works for God, it works equally for islands. Since we are not obliged to believe in perfect islands on the strength of a concept, we are not obliged to believe in God on the same strength.
Anselm replied that the argument applies only to beings whose greatness is unlimited, and that an island, however perfect, is still a finite thing whose greatness has a natural ceiling. The concept of the greatest conceivable island is incoherent, he argued, because islands have intrinsic limits: the most beautiful beach, the most fertile soil, the finest climate are all bounded properties that do not admit of infinite maximisation in the way omnipotence, omniscience, and perfect goodness are said to do. A perfect island is therefore not a genuine counterpart to a perfect being, and the parody fails on those grounds.
This reply is more interesting than it might first appear, but it does not ultimately save the argument. Anselm is now relying on a distinction between bounded and unbounded greatness that he has not yet established. He is assuming that the divine properties, omnipotence, omniscience, perfect goodness, are the kind of thing that can be maximised without limit and that their joint maximisation is coherent. Both assumptions are highly contestable, as we shall see in detail below. The claim that these properties can be coherently extended to infinity is not a starting premise the argument may freely assume; it is precisely what requires demonstration, and Anselm simply asserts it.
More fundamentally, Gaunilo’s parody does not need to be watertight in every detail to do its work. Its purpose is to show that Anselm’s logical structure is defective, and it achieves this by generating an equally valid proof for something we know does not exist. The parody need not be a perfect mirror image of the original; it need only demonstrate that the form of the argument permits conclusions we are unwilling to accept. And the form does permit them, consistently. Substitute “the greatest conceivable pizza,” “the greatest conceivable library,” or “the greatest conceivable physical object” and the argument produces the same conclusion each time, with the same apparent necessity. The problem is structural, not incidental, and no restriction to “unlimited beings” can repair it without begging the very question of whether unlimited beings are coherently conceivable in the first place.
3. Existence Is Not a Predicate
The most decisive philosophical objection to Anselm was formulated, with characteristic precision, by Immanuel Kant in the Critique of Pure Reason in 1781. Kant argued that existence is not a predicate, or more precisely, not a real predicate, not a property that adds anything substantive to a concept. When I describe a hundred real thalers, Kant observed, they contain nothing more than a hundred possible thalers. The concept is exactly the same whether or not the coins exist in the world. Existence does not augment the description of a thing; it merely states that the thing described is instantiated in reality. To say “God exists” is therefore not to assign a property to God but to assert that the concept of God has an instance somewhere in the actual world. And that is precisely what the argument needs to prove, not what it may assume as a step in its own reasoning.
Consider a more mundane example to make this vivid. Suppose I describe a unicorn as a horse with a single spiral horn, the capacity to purify poisoned water, and a radiant white coat. I then add “and this unicorn actually exists.” Have I made the unicorn greater or more perfect by adding that final clause? The answer is plainly negative. I have added a fact about the world, namely that there is an instance of this description, but I have not enriched the description itself in any way. The unicorn-that-exists is not a better or more complete unicorn than the unicorn-I-merely-imagined; it is the same unicorn, and the question of whether it exists is entirely separate from what the concept contains.
Applied to Anselm, the point is lethal. When he claims that a being existing in reality is greater than one existing only in the mind, he is treating existence as a property that adds to a being’s greatness, as a perfection-making feature that can be placed on the same list as omnipotence and omniscience. But Kant’s analysis shows that existence is not that kind of property. You cannot make a concept greater by adding existence to it, any more than you can make a hundred imaginary thalers richer by declaring them real. The descriptive properties of a being, its power, its knowledge, its goodness, constitute what it would be if it existed. Whether it exists is a separate matter entirely, one that concepts alone cannot settle. Anselm’s comparative step, the move from “existing in reality is better” to “the greatest being must exist in reality,” rests on a category error, treating an existential claim as if it were a descriptive one.
Carl Sagan, who wrote with unusual clarity about exactly this kind of conceptual confusion, put it with admirable directness: “St. Anselm argued that since we can imagine a perfect being, he must exist, because he would not be perfect without the added perfection of existence. This so-called ontological argument was more or less promptly attacked on two grounds: (1) Can we imagine a completely perfect being? (2) Is it obvious that perfection is augmented by existence? To the modern ear such pious arguments seem to be about words and definitions rather than about external reality.” Sagan’s second question is the Kantian one restated in accessible terms, and his observation that the argument seems to be about “words and definitions rather than about external reality” locates the pathology with precision. The ontological argument is, at its core, a verbal manoeuvre. It rearranges concepts without making contact with the world those concepts are supposed to describe.
4. Plantinga and the Modal Upgrade
Anselm’s argument was widely regarded as refuted after Kant, and for most of the nineteenth and early twentieth centuries it occupied a place in intellectual history as an interesting failure rather than a live debate. Then, in 1974, the American philosopher Alvin Plantinga published The Nature of Necessity, which contained a reformulation of the argument using the tools of modal logic, the formal study of necessity and possibility. Plantinga’s version is considerably more sophisticated than Anselm’s, and it requires more careful handling, but its central move is ultimately the same, and it fails for closely related reasons.
Modal logic distinguishes between possible worlds, a technical term for ways the universe could have been, not literally other universes but logical possibilities about how things might be arranged. Some propositions are “possibly true,” meaning they are true in at least one possible world. Some are “necessarily true,” meaning they are true in every possible world. Mathematical truths are the standard example of the latter: two plus two equals four in every possible world, not merely in this one. The number four does not exist contingently, as tables and chairs do, but necessarily, as part of the logical structure of any coherent reality.
Plantinga defines God as a being with “maximal greatness,” which he unpacks as having maximal excellence, meaning omnipotence, omniscience, and perfect moral goodness, in every possible world. A being that had these properties in some worlds but not others would be less than maximally great, because its greatness would be contingent rather than necessary. Plantinga’s key premise is then that maximal greatness is possibly instantiated, that there is at least one possible world in which a maximally great being exists. From this, using the axioms of a particular modal logic system known as S5, he derives that a maximally great being exists in every possible world, including the actual one, and therefore that God necessarily exists.
The argument, stated more formally, proceeds as follows. First: it is possible that a maximally great being exists. Second: if a maximally great being possibly exists, then it necessarily exists, because by definition maximal greatness includes necessary existence across all possible worlds. Third: if it necessarily exists in all possible worlds, it exists in the actual world. The move from premise one to the conclusion is licensed by the S5 axiom, which holds that if something is possibly necessary, it is therefore necessary. The machinery is clean, the formal steps are valid, and the conclusion appears to follow without error.
Plantinga himself was admirably candid about what this argument does and does not achieve. He acknowledged that it does not constitute a proof that compels assent from any rational person, and he accepted that a rational person could deny premise one without obvious self-contradiction. His more modest claim was that accepting premise one is itself rational, and that someone who accepts it is rational to conclude that God exists. This is a significantly weaker claim than Anselm’s, and it has the virtue of intellectual honesty, but it also exposes the argument’s central weakness more clearly than most of its popular defenders prefer to admit.
5. The Question-Begging Possibility Premise
The decisive objection to Plantinga is that his possibility premise is not nearly as innocent as it appears. He asks us to grant that it is possible, in the technical modal sense, that a maximally great being exists. This sounds like a very modest request. Of course we can imagine such a being; of course the concept seems internally coherent at first glance; surely it is simply “possible” in the way that many things are possible before we know whether they are actual. But the technical use of “possible” in S5 modal logic is far more loaded than this intuitive sense suggests, and once that loading is understood, the premise turns out to assume precisely what the argument is supposed to prove.
In S5 modal logic, “possibly P” means “P is true in at least one possible world,” and “necessarily P” means “P is true in every possible world.” The S5 axiom in question states that if it is possible that P is necessary, then P is necessary. Plantinga’s possibility premise, properly unpacked, is therefore not the innocuous claim that “we can imagine such a being.” It is the much stronger claim that there is at least one possible world in which a maximally great being exists necessarily. But if such a being exists necessarily in any possible world, the S5 axiom immediately tells us it exists necessarily in all possible worlds, including this one. The possibility premise is therefore not a modest starting point from which the argument departs; it is the conclusion itself, wearing a slightly different grammatical costume.
To see this more clearly, observe that the argument runs in the opposite direction with equal force. Suppose we begin instead with the premise that it is possible that God does not exist, meaning there is at least one possible world in which there is no God. In that world, there is no maximally great being. But a maximally great being, by Plantinga’s own definition, would exist necessarily in every possible world if it exists in any. If there is even one possible world where God does not exist, then God is not necessary, and therefore not maximally great, and therefore does not exist in any possible world, including the actual one. The premise “possibly God does not exist” leads, by the same logical machinery, to the conclusion “God necessarily does not exist.” Both arguments have the same formal structure and the same logical validity. The question of which is sound reduces entirely to the question of which possibility premise is true, and we have no independent means of resolving that prior to any other investigation. The argument is not a proof; it is a framework that delivers whatever conclusion you smuggle in through the possibility door.
Philosophers describe this failure as question-begging, which means, in plain terms, that the conclusion is hidden inside one of the premises. Plantinga’s argument is valid in the technical sense, meaning the conclusion follows from the premises if the premises are true. The question is whether the premises are true, and the answer is that we cannot establish the truth of the possibility premise without already having established the conclusion. The argument moves in a circle, and no amount of modal sophistication changes that fundamental structural fact.
6. The Maximally Great Pizza Problem
Gaunilo’s island can be updated for the modal age, and the result is even more illuminating than the original parody. Consider the concept of a maximally great pizza, defined as a pizza with perfect taste, perfect nutritional completeness, optimal temperature, and ideal proportions, existing necessarily in every possible world in which conscious beings capable of appreciating pizza exist. By Plantinga’s logic, if it is possible that a maximally great pizza exists in this technical sense, and if a maximally great pizza would necessarily exist wherever such worlds obtain, then the S5 axiom tells us it necessarily exists in our world. The argument is structurally identical to Plantinga’s. The only difference is that we have strong independent reasons to doubt the existence of a necessarily existing pizza, whereas the existence of God is supposed to seem more antecedently plausible.
The pizza parody does more than raise a laugh, though it does that too. It reveals that the structure of the argument carries no information about whether its conclusion is true. An argument form that proves equally well whatever you insert into the subject position is not tracking truth; it is generating conclusions from grammatical templates and modal axioms. The question of whether God, or the perfect pizza, or the greatest conceivable library actually exists is a question about the world, and it cannot be settled by manipulating definitions, however formally rigorous the manipulation. That is Sagan’s point, restated: these arguments are about words, not about external reality.
Richard Dawkins made a related observation in The God Delusion when attacking a different but adjacent piece of reasoning, the Argument from Degree, and the point transfers cleanly to the ontological case: “That’s an argument? You might as well say, people vary in smelliness but we can make the comparison only by reference to a perfect maximum of conceivable smelliness. Therefore there must exist a pre-eminently peerless stinker, and we call him God. Or substitute any dimension of comparison you like, and derive an equivalently fatuous conclusion.” The substitution test is exactly what exposes the ontological argument’s emptiness. The argument works for any concept you choose to maximise, which means it is not doing the philosophical work its defenders claim. An argument that proves too much proves nothing at all.
7. Can We Even Conceive of a Maximally Great Being?
There is a prior problem that both Anselm and Plantinga tend to assume away rather than address: whether the concept of a maximally great being is coherent in the first place. The argument requires not merely that we can produce the phrase “greatest conceivable being” but that this phrase picks out a coherent concept, a consistent set of properties that could in principle be jointly instantiated. There are strong reasons to doubt that it does, and those reasons deserve careful attention before we proceed to argue about whether the concept is instantiated.
Begin with omnipotence. Philosophers since Aquinas have wrestled with the paradoxes that arise when you try to make this property precise. An omnipotent being can do anything, by definition. Can it create a stone so heavy that it cannot lift it? If the answer is yes, there is something it cannot do, namely lift the stone, and it is not omnipotent. If the answer is no, there is something it cannot do, namely create such a stone, and it is not omnipotent. This is not a cute puzzle that dissolves on close inspection; it reflects a genuine structural problem with the concept of unlimited power. Power is always power to do something, and the specification of what can be done always implies constraints on what can and cannot be achieved consistently. The attempt to remove all constraints does not produce a greater kind of power; it produces incoherence.
Omniscience generates analogous difficulties when examined carefully. Does a being that knows everything know what it is like to be uncertain? Does it know what it is like to be ignorant, to be genuinely surprised, to make a mistake? If it does possess that knowledge, then it knows something about states it cannot itself occupy, which raises deep questions about the nature of that knowledge and whether it constitutes genuine understanding or mere abstract representation. If it does not possess that knowledge, then there are truths, facts about the phenomenology of finite minds, that it does not know, and it cannot be omniscient. Does an omniscient God know what you will freely choose to do tomorrow? If it does know, in what sense is your choice genuinely free rather than already determined? The conjunction of omniscience with human free will has generated a literature large enough to fill several shelves, and there is no consensus resolution. The concept of omniscience, when pressed, does not behave like a straightforwardly coherent property.
Perfect moral goodness in conjunction with omnipotence and omniscience produces the problem of evil, which has its own vast and distinguished literature. An omnipotent being could prevent suffering without limit. An omniscient being would know about every instance of it. A perfectly good being would be motivated to prevent whatever suffering it could prevent. The world contains vast quantities of suffering, a great deal of it falling on those who have done nothing to deserve it, much of it falling on non-human animals who have no moral standing in any framework that would make their suffering instrumentally necessary to some greater good. The conjunction of these three divine properties with the observed character of the actual world is, to put it with deliberate understatement, difficult to make consistent. Defenders of theism have proposed numerous theodicies, elaborate explanations of why a maximally good being might permit the suffering we observe, but the very multiplicity and ingenuity of these theodicies is itself a signal that the conjunction is under severe pressure. You do not need multiple contradictory explanations for why something obvious is not a contradiction.
The point is not that these difficulties are demonstrably insoluble with absolute certainty. The point is that the ontological argument assumes the coherence of its subject concept before the argument begins, treating that coherence as unproblematic when it is anything but. If the concept of a maximally great being is incoherent, then the premise “a maximally great being possibly exists” is false, not because no such being happens to exist but because the description is internally contradictory, a set of properties that cannot be jointly instantiated, like a married bachelor or a four-sided triangle. And if the possibility premise is false, the entire argument collapses before it takes a single substantive step.
8. The Historical Trajectory of a Failed Argument
It is instructive to trace what has happened to the ontological argument over the nine centuries since Anselm, because the pattern it describes is philosophically diagnostic. Each time a decisive objection has been raised, the argument has not been abandoned but reformulated. Gaunilo’s parody prompted Anselm to restrict the argument to unlimited beings. Kant’s predicate objection prompted defenders to distinguish between first-order predicates and higher-order properties of properties. Plantinga’s modal reformulation prompted critics to identify the question-begging in the possibility premise, which prompted defenders to argue that the premise is intuitively reasonable even if not demonstrable. At each stage, the argument retreats to more modest ground while maintaining the claim that some version of it still works.
This is not the behaviour of an argument that is being progressively refined toward truth. It is the behaviour of a conclusion that a thinker is committed to in advance and which is being protected from refutation through strategic reformulation. The ontological argument has not accumulated independent confirmation over nine centuries; it has accumulated epicycles. Each new formulation handles the specific objection that prompted it while generating fresh vulnerabilities that require yet further reformulation. This is the epistemological signature of a hypothesis that does not correspond to reality but is too important to its defenders to be released.
Compare this trajectory with the history of successful arguments in mathematics and natural science. When Euclid demonstrated a geometric theorem, subsequent generations did not need to reformulate it repeatedly to protect it from refutation; the proof stood, and critics could be directed to its specific steps. When Darwin’s argument from natural selection was first published, it attracted powerful objections, particularly regarding the mechanism of inheritance, but those objections were resolved not by retreating to ever more modest claims but by the discovery of genetic evidence that strengthened and confirmed the original argument. Arguments that track truth tend to gather independent support over time. Arguments that track prior commitment tend to generate an ever more elaborate defensive architecture without accumulating any external confirmation.
The ontological argument’s nine-century defensive project is therefore not evidence of its resilience. It is evidence of the depth of the commitment that motivates it. Anselm was not doing philosophy in the modern sense; he was doing theology, and the distinction matters enormously. Theology characteristically begins with a conclusion and works backward to find premises that might support it. Philosophy, at its best, begins with premises and follows the argument wherever it leads, including into conclusions that are uncomfortable or unwelcome. The ontological argument has the grammatical form of philosophy and the epistemological structure of theology, and that mismatch is precisely what makes it so persistently seductive and so persistently wrong.
9. Why Apologists Keep Using It
Given that the ontological argument has been refuted in several different ways by several different thinkers operating in several different philosophical traditions, the question of why it continues to appear in apologetic literature deserves a frank and direct answer. The honest answer is that it serves a rhetorical function rather than an evidential one, and that function is to create the impression of intellectual rigour without delivering it.
An apologist who leads with the ontological argument signals that theism is a philosophically serious position, one that can be defended with the tools of formal logic rather than mere sentiment, tradition, or emotional appeal. This signal is valuable to the apologist even if the argument ultimately fails, because it shifts the conversation from the evidential question, namely “What independent reason do we have to believe in God?”, to the conceptual question of whether the concept of God is coherent. The conceptual question is genuinely interesting and genuinely difficult, and a non-specialist interlocutor can easily become lost in its technicalities. The argument functions as a distraction as much as a demonstration, and a useful one if your purpose is to avoid the evidential question rather than to answer it.
There is also a deeper psychological function worth acknowledging without condescension. For a believer who has already arrived at faith through other means, through upbringing, through religious experience, through community and culture, through a powerful sense of meaning and purpose, the ontological argument provides a post-hoc rationalisation that makes the belief feel intellectually validated. This is not a pathology unique to religious belief; it is a description of how human reasoning operates across many domains. We reliably tend to seek arguments that support conclusions we already hold, and the ontological argument, with its air of logical inevitability, is well suited to this purpose. The problem is not that believers find it convincing; the problem arises when it is presented to sceptics as a genuine demonstration, because it is demonstrably not one.
Contemporary apologists of the professional variety, the kind who publish in academic philosophy journals and who are careful about the distinction between validity and soundness, generally concede most of what has been argued in this essay. They acknowledge that the argument does not compel assent from any rational person. They accept that a rational person can reject the possibility premise without contradiction. What they maintain is that a rational person can also accept the possibility premise, and that accepting it rationally commits them to the conclusion that God exists. This is a considerably weaker claim than “God’s existence can be proven from pure reason,” and it deserves to be evaluated accordingly. The claim that a rational person can accept a premise that quietly contains the conclusion they wish to establish is a claim with very limited epistemic significance. Rational people can accept many things that turn out to be false, and the bare logical possibility of accepting a premise without immediate self-contradiction does not constitute evidence that the conclusion corresponding to it is true.
10. What the Argument Cannot Establish Even If It Worked
There is a further problem that deserves sustained attention, not because it is a logical objection to the argument’s validity but because it bears directly on the argument’s practical significance for religious belief. Suppose, for the sake of argument, that all the preceding objections were somehow answered, and the argument did succeed in establishing that a maximally great being necessarily exists. What would that being actually be, and what would follow from its existence?
The argument, as formulated by both Anselm and Plantinga, establishes a being with omnipotence, omniscience, and perfect moral goodness. It says nothing about this being’s interest in human affairs, its responsiveness to prayer, its having revealed itself through scripture, its having become incarnate in first-century Palestine, or any of the other specific claims that constitute actual religious belief as practised by billions of people. The gap between “a maximally great being necessarily exists” and “the God of Abraham, Isaac, and Jacob exists and demands that you observe the commandments” is not a small gap that further argument can easily bridge. The distance between these two claims is enormous, and no amount of modal reasoning closes it.
The ontological argument, at best, establishes something that might be called the philosopher’s God: a maximally great abstract being whose relationship to any specific religious tradition remains entirely undefined and whose properties, while impressive in the abstract, carry no implications for how one ought to live, what one ought to believe about history, or which community of practice one ought to join. To move from that abstract conclusion to the rich, specific, historically situated deity of Christianity, Islam, or Judaism requires an enormous quantity of additional argument, argument that must appeal to evidence, historical claims, and religious experience, and that must answer the problem of evil, the problem of divine hiddenness, the problem of religious diversity, and a great deal else besides.
This point is worth emphasising because the ontological argument is routinely presented in popular apologetics as if establishing the philosopher’s God were equivalent to establishing the religious one. The move is not merely logically unjustified; it is a form of argumentative bait-and-switch. The argument is deployed in a context that implies it is doing work for, say, Christian theism specifically, when in fact even a successful version of it would establish nothing more than the bare existence of some maximally great being with no known relationship to any particular tradition, scripture, or moral demand. The believer in the pew who finds the ontological argument reassuring is finding comfort in a conclusion that does not actually support the beliefs for which they sought support.
11. The Broader Lesson About Definition and Reality
The ontological argument belongs to a broader category of intellectual error that recurs across many domains: the confusion of conceptual necessity with factual necessity. A concept can have necessary internal structure without that structure implying anything about whether the concept is instantiated in the world. Triangles necessarily have three sides; this is a conceptual truth, built into the definition of “triangle.” But it does not follow from this that any triangles exist anywhere in reality. The conceptual necessity of triangularity tells us only that if something is a triangle, it has three sides; it does not tell us that anything is actually a triangle, and to determine whether anything is, we must look at the world rather than at the definition.
Applied to God, the point is exactly parallel. Even if we could establish that the concept of a maximally great being necessarily included existence as one of its properties, this would show only that if anything is a maximally great being, it exists necessarily; it would not show that anything actually is a maximally great being. The conceptual claim “all maximally great beings necessarily exist” is consistent with “there are no maximally great beings,” just as “all unicorns have exactly one horn” is consistent with “there are no unicorns.” The necessity is internal to the concept; it does not reach out into the world and compel instantiation.
This is, in a sense, the Kantian point formulated from a slightly different angle. The predicate “exists” does not add to a concept; it states that the concept is instantiated. Conceptual analysis can tell us what properties a thing would have if it existed, and how those properties relate to one another within the concept. It cannot tell us whether anything exists with those properties. To answer the question of existence, you must consult the world, not the dictionary. You must gather evidence, make observations, frame hypotheses, and subject them to the genuine possibility of disconfirmation. The ontological argument is an attempt to bypass this entire requirement, to settle what is ultimately an empirical question by examining concepts rather than reality, and that is why it fails at every stage of reformulation. The failure is not a contingent feature of Anselm’s or Plantinga’s particular formulations; it is a structural feature of any attempt to derive existence from definition alone.
This lesson has implications considerably beyond theology. The same confusion between conceptual and factual necessity appears in certain political arguments (“a truly free market would produce optimal outcomes”), in certain psychological arguments (“a truly rational agent would always choose X”), and in certain ethical arguments (“a perfectly just society would necessarily have property Y”). In each case, the move from “by definition, X has property P” to “X as described exists and has property P in the actual world” requires precisely the step that the argument omits: the empirical investigation of whether X is instantiated, and under what conditions, and to what degree. Recognising the ontological argument’s failure is therefore not merely a contribution to the philosophy of religion; it is a lesson in intellectual hygiene that applies wherever definitions are being quietly substituted for reality.
12. Responding to the Symmetry Objection
A predictable objection at this point is that the sceptic’s parallel argument, “possibly God does not exist, therefore necessarily God does not exist,” is equally question-begging against the theist, and that we are therefore in a symmetrical stand-off where neither side can claim rational advantage over the other. This objection deserves a direct and careful answer rather than a dismissal.
The objection is correct that both possibility premises beg the question in Plantinga’s modal framework, and this essay has acknowledged as much. This is precisely the point. The S5 framework converts any modal possibility claim about necessary beings into a claim about every possible world, which means the system is exquisitely sensitive to whatever you insert into the possibility slot. The framework does not adjudicate between competing possibility claims; it merely formalises and amplifies whatever intuition you bring to it. This is not a sign that the sceptic and the theist are on equal rational footing; it is a sign that the framework is simply the wrong tool for the question being asked. If a logical framework makes the answer depend entirely on which possibility premise you accept, and if we have no independent means of adjudicating between those possibility premises, then the framework has not helped us make progress on the question of God’s existence at all. It has merely formalised prior intuitions and dressed them in modal notation.
The appropriate response to this stand-off is not to conclude that both sides are equally rational in their modal intuitions but to recognise that modal arguments of this type are simply the wrong kind of argument for questions of existence. Questions about what exists are empirical questions. They are answered by evidence, by observation, by the construction and testing of hypotheses, not by modal intuitions however carefully systematised. The sceptic who abandons the modal framework and returns to the evidential question is not retreating from reason; they are insisting that the question be answered by the means appropriate to it. And on the evidential question, the situation is manifestly not symmetrical. There is no independent evidence for the existence of a maximally great being. There is, conversely, a great deal of evidence from history, from natural science, and from comparative religion that the world is entirely explicable without appeal to such a being. The burden of proof falls on the positive claim, and that claim has not been met by the ontological argument or by anything else.
For a fuller treatment of where the burden of proof lies in these debates, and why the demand for evidence is not itself a form of prejudice against theism, readers may find it useful to look at the discussion of the god of the gaps pattern, which illuminates how theistic arguments tend to fill explanatory vacuums with divine assertions rather than actually addressing evidential questions. The ontological argument is unusual in not even attempting to appeal to gaps in scientific knowledge; it operates at the purely conceptual level before empirical science enters the picture at all. But its failure at the conceptual level leaves theism in the same position as before: without evidence, and therefore without a justified claim on rational assent.
13. The Argument’s Relationship to Genuine Philosophy of Religion
It would be intellectually dishonest to pretend that the ontological argument’s history is simply one of unrelieved failure with no productive legacy. The argument has generated some of the most interesting work in the philosophy of religion, and has forced genuine advances in our understanding of modal logic, the logic of existence statements, and the relationship between conceptual analysis and metaphysical reality. Philosophers like David Lewis, Saul Kripke, and Plantinga himself made contributions to modal logic that extend far beyond the theological motivations that shaped some of their work. The argument is not without value; it is simply not a proof of God’s existence, and that distinction matters enormously.
The productive legacy of the ontological argument lies in the questions it forced rather than in the answer it attempted to provide. It compelled philosophers to ask what it means for something to exist necessarily rather than contingently, how possible worlds relate to actual reality, what logical constraints govern the concept of perfection, and whether existence can meaningfully be treated as a predicate. These are genuine philosophical questions with implications across metaphysics, logic, and the philosophy of language, and the centuries of work they generated advanced human understanding even though the original theological conclusion was unsound.
What the argument is not, and what it has never been, is evidence for the existence of God. This distinction between an argument being philosophically interesting and an argument being evidentially compelling is one that popular apologetics consistently blurs, and that blurring is not honest. The ontological argument is regularly cited in popular discussions as if Plantinga’s reformulation had settled the matter in theism’s favour. It has not settled it, and Plantinga himself, to his credit, would not claim that it has. The most that can honestly be said is that the argument raises questions about the modal status of necessary beings that have not been definitively closed in every direction, and that some philosophers who accept the possibility premise find the resulting inference compelling. This is a very different thing from the argument constituting a proof of God’s existence, and the distance between those two things is the distance between genuine philosophy and apologetic propaganda.
14. What Genuine Evidence Would Look Like
The ontological argument is, among its other features, an attempt to demonstrate God’s existence without appealing to evidence, and this reveals something important about the theistic enterprise more broadly. If God’s existence could be established purely from the concept of God, then the empirical investigation of the world would be redundant for theological purposes. The argument is designed to short-circuit the requirement for evidence by deriving existence from definition. This design is itself a symptom of the evidential difficulty facing theism. If there were abundant, clear, independently verifiable evidence for God’s existence, the ontological argument would be unnecessary, and no one would be spending their time constructing modal proofs. The fact that such proofs are offered, and offered seriously by intelligent people, is a strong indication that the evidence route has been tried and found unable to carry the weight of the conclusion.
What would genuine evidence for a maximally great being actually look like? It would look like the kind of evidence we accept for any other significant existence claim: observable, repeatable, falsifiable, and not equally explicable by more parsimonious hypotheses that do not require the being in question. A maximally great being that is omnipotent, omniscient, and perfectly good would, one might reasonably expect, leave some distinctive trace in the fabric of observable reality, a trace that was not simultaneously explicable by natural processes, psychological mechanisms, and the full range of cognitive biases to which human beings are demonstrably subject. The consistent finding across the natural sciences, from physical cosmology to evolutionary biology to neuroscience and comparative religion, is that the phenomena previously attributed to divine action are explicable without that action, and often explained better without it.
This accumulation of natural explanation is not a philosophical proof of God’s absence. It is something more practically significant: an accumulation of evidence that the hypothesis of a maximally great being is not required by the data, not called for by the phenomena, not doing explanatory work that nothing else can do. For a more detailed treatment of this point in the cosmological context, the discussion of whether theists genuinely maintain that everything came from nothing addresses the specific ways in which the cosmological argument attempts to locate a role for God at the origin of the universe, and finds the attempt no more successful than the conceptual approach examined here.
The ontological argument’s attempt to bypass the evidential requirement is therefore not a strength but a confession. It is an acknowledgement, implicit in its very structure, that the evidential case for God’s existence is insufficient to carry the weight of the conclusion unaided. The argument says: we do not need evidence, because the concept alone is sufficient. But concepts do not make things exist, and the absence of evidence, where evidence would reasonably be expected if the hypothesis were true, remains exactly what Thomas Huxley recommended treating it as: grounds for suspended judgement, for proportioned confidence, and for the intellectual courage to say “I do not know” rather than to dress an assumption in the formal clothes of a logical demonstration.
15. Anselm, Plantinga, and the Limits of Pure Reason
There is a broader philosophical point lurking behind all the specific objections examined in this essay, and it deserves to be brought into the open before we close. The ontological argument represents one of the most ambitious attempts in the history of thought to establish a substantive conclusion about reality using pure reason alone, without any appeal to observation or experience. Rationalist philosophers from Descartes to Leibniz attempted various versions of this project, seeking to derive the structure of reality from self-evident axioms by pure logical deduction. The Kantian revolution in philosophy was, in substantial part, a demonstration that this project cannot succeed: that pure reason, operating without input from experience, can generate only analytic truths, truths that are true by definition, and that synthetic truths, truths about how the world actually is, require empirical investigation.
The ontological argument is a specific instance of this broader rationalist ambition, and its failure is a specific instance of the broader Kantian lesson. The claim that a maximally great being exists is a synthetic claim; it is a claim about what is actually in the world rather than merely about what follows from a definition. To establish a synthetic claim, you need synthetic evidence, which means evidence drawn from the world rather than from the analysis of concepts. The ontological argument attempts to establish a synthetic conclusion by analytic means, and this is why it fails regardless of how the details are handled in any particular formulation. The failure is not accidental or correctable by further ingenuity; it is built into the nature of the project.
This does not mean that pure reason is without value in addressing questions about God. Logic can reveal internal contradictions in theological positions, can show that certain combinations of divine attributes are inconsistent, can demonstrate that particular theodicies fail on their own terms. These are all uses of reason that make genuine contributions without overreaching into the domain of empirical claims. The problem arises specifically when pure reason is asked to do what only evidence can do: to establish that something exists. At that point, reason must step back and defer to the investigation of the world, and no amount of modal sophistication changes that basic requirement.
Conclusion: You Cannot Think a Thing Into Being
Anselm wanted to show that God’s existence was not merely a matter of faith but a matter of logical necessity, something that any rational person, thinking clearly about the concept of God, would be compelled to accept regardless of their prior commitments or personal experience. It is a seductive aspiration, and it produced nine centuries of philosophically interesting argument. But it failed, and it failed for reasons that are now well understood and that no reformulation, however technically impressive, has managed to circumvent. Existence is not a predicate. A concept, however carefully constructed, cannot be defined into instantiation. A possibility premise that secretly contains the conclusion it is supposed to establish is question-begging rather than demonstrative. And an argument form that generates necessarily existing pizzas and peerless stinkers with equal logical facility is not tracking the structure of reality.
Plantinga’s modal version is more sophisticated than Anselm’s, and its defenders are considerably more philosophically careful than the popular apologetics that routinely cite it would suggest. Plantinga himself has been admirably honest about what the argument achieves and what it does not. But sophistication in presentation does not repair a structural defect in the underlying reasoning. The possibility premise, when properly unpacked in the S5 framework, assumes what it needs to prove. The modal machinery, powerful as it is for many logical purposes, is the wrong tool for settling questions of existence that require empirical investigation. And the retreat from “God’s existence is demonstrable from pure reason” to “it is rational to accept the possibility premise” is not a success for theism; it is a concession dressed as a holding position, and a rather thin one at that.
The deeper lesson, and the one that matters most beyond the immediate question of theology, is the one that has been implicit throughout this essay: the question of whether God exists is a question about the world, and questions about the world are answered by examining the world. The ontological argument is an attempt to answer an empirical question through conceptual analysis alone, and that method, however elegantly executed, is the wrong method for the question. You cannot think a thing into being. You cannot define it into being. You cannot modally derive it into being from a possibility premise that smuggles the conclusion through the back door. If the maximally great being exists, its existence must show up somewhere in the fabric of observable, investigable reality, and the fact that nine centuries of some of the most brilliant philosophical minds in Western history have not found it doing so is, to put it plainly, a fact that demands to be taken seriously.
For those interested in the broader question of whether any of the classical arguments for God’s existence fares better when the evidential bar is properly applied, the analysis of Pascal’s Wager and the rationality of theistic belief addresses the practical reasoning arguments that apologists tend to reach for once the purely logical ones have been exhausted. They do not fare better, for related reasons. But the verdict on the ontological argument itself is clear, and it has been clear since Kant: the argument is about words and definitions rather than about external reality, and no reformulation, however technically impressive, has changed that fundamental and final fact.