What Is Omnipotence?

The Basic Definition

Omnipotence is the property of being all-powerful. In the philosophy of religion, it is one of the core divine attributes, alongside omniscience, omnipresence, and perfect goodness. The definition sounds easy — God can do anything — but it takes about five minutes of thought to see why "anything" needs qualification. Can God make two plus two equal five? Can God create a square circle? Can God make a being more powerful than God? The answer classical theism gives to all three is no, and the reason is the same in each case: these are not things. They are contradictions dressed up as descriptions. A square circle is not a difficult thing to make; it is a nothing, a string of words that fails to refer. Asking whether God can make one is like asking whether God can blorble a fnord.

Aquinas is the one who put this in its clearest form, and his version has carried the tradition. God can do all things that are absolutely possible — that is, all things that do not involve a contradiction. A contradiction is not a possible thing, so failing to "do" a contradiction is not a failure of power. It is a failure of the question. This move looks like a dodge to some critics, and we will get to their objections. But it is the move that nearly every major philosopher of religion from the medieval period to the present has accepted as the right starting point.

The Stone Paradox

The challenge that everyone raises goes like this. Can God create a stone so heavy that God cannot lift it? If yes, then there is something God cannot do — lift the stone — and God is not omnipotent. If no, then there is something God cannot do — create the stone — and God is not omnipotent. Either way, omnipotence seems impossible.

This is a genuinely fun puzzle, and it has been around in recognizable form since the medieval period. Augustine and Aquinas were both asked versions of it. The standard response, which most philosophers now accept, is that the puzzle is a trick of language. A stone too heavy for an omnipotent being to lift is a stone that cannot be lifted by a being than whom nothing is stronger. But if the being is omnipotent, there is no such stone — the concept is self-contradictory, like "a circle that is square." The paradox works by smuggling in a contradiction and then acting surprised when the contradiction cannot be instantiated. The correct answer, on the classical view, is that God cannot create such a stone, because the thing in question is not a possible thing. And failing to create an impossibility is not a limitation on omnipotence.

There is a family of related paradoxes — can God make a being free that is not free, can God make a rule God cannot break, can God cease to be God — and the same logic applies to each. If the proposed feat involves a contradiction, omnipotence does not require the ability to perform it. Anselm made this point in his own way: "whatever implies contradiction does not fall within the scope of omnipotence." The stone paradox is the popular version of a deeper question about whether logical laws constrain God, and the mainstream answer is that logical laws are not constraints on God but features of intelligibility itself.

The Harder Questions

Setting the paradox aside, the genuinely contested questions about omnipotence are more interesting. They concern the edges of the attribute — the cases where reasonable philosophers disagree.

Can God do evil? If God is perfectly good, and goodness means God cannot act wrongly, then God's inability to do evil is a consequence of God's nature, not a weakness. Augustine argued this line: the inability to do evil is not an inability but a perfection. Most classical theists agree. But it raises a question about the definition. If omnipotence means "can do anything logically possible," and doing evil is logically possible, then a good God's inability to do evil seems to be a real limitation. Defenders respond by refining the definition: omnipotence is the power to do all things consistent with God's nature. The debate is not settled, and it turns on whether you think moral perfection is a restriction on power or a form of it.

Can God change the past? Most philosophers say no — the past, once actual, is fixed, and omnipotence does not include the power to make what happened not have happened. But some have argued that if the past is not logically necessary (it did not have to happen that way), then an omnipotent being should be able to change it. The question has practical bite in the doctrine of prayer: if you pray for something that has already happened, is that coherent? Most theists say no, but the philosophical literature keeps the question alive.

Can God act irrationally? If God is perfectly rational, can God act against reason? If not, is that a limitation? The classical answer parallels the one about evil: perfect rationality is a perfection, and the inability to act irrationally is not a weakness. But if you define omnipotence without reference to God's nature, you get the puzzle back. The pattern is consistent. The hard cases are all about whether to define omnipotence in terms of bare logical possibility or in terms of consistency with the divine nature.

Mackie and the Free Will Defense

The most consequential use of the omnipotence concept in twentieth-century philosophy came from J. L. Mackie, in his 1955 paper "Evil and Omnipotence." Mackie argued that the core attributes of God — omnipotence, omniscience, and perfect goodness — are logically incompatible with the existence of evil. If God can prevent evil and wants to, there should be no evil. But there is evil. So either God is not omnipotent, not good, or does not exist.

Mackie's argument is the problem of evil in its sharpest logical form, and the free will defense is the standard response. Alvin Plantinga showed, with formal precision, that it is possible that God could not have created a world with free creatures who always do good — because free creatures, by definition, can choose wrongly. So the existence of evil is compatible with omnipotence, as long as God's reason for permitting it (free will) is a good reason. Plantinga did not prove that this is how things are. He showed that the logical inconsistency Mackie claimed is not established. Omnipotence, on this account, does not include the power to make free creatures unfree — and again, that is not a limitation but a feature of what freedom means.

Contemporary Refinements

Modern philosophers have not stopped refining the definition. Richard Swinburne, in The Coherence of Theism, offers a version that strips away some of the classical apparatus. Swinburne defines omnipotence as the power to bring about any state of affairs that is logically possible for a being with God's nature to bring about. He keeps the logical constraint and the nature constraint, but he is more willing than the medievals to leave the edges open. The result is a definition that is slightly more flexible and slightly less tidy.

The broader lesson is that omnipotence, like the other divine attributes, is not a brute datum. It is a philosophical construction, built to capture the intuition that God is supremely powerful. Every refinement — the logical constraint, the nature constraint, the moral constraint — is a response to a real puzzle. And every puzzle reveals something about what we mean by power, by possibility, and by God. The concept is far from exhausted, and the debates about its coherence are part of what keeps the philosophy of religion a living discipline.

Sources

• Stanford Encyclopedia of Philosophy, "Omnipotence" — https://plato.stanford.edu/entries/omnipotence/
• Stanford Encyclopedia of Philosophy, "Philosophy of Religion" — https://plato.stanford.edu/entries/philosophy-religion/
• Richard Swinburne, The Coherence of Theism (Clarendon Press, 1977; 2nd ed. 2016).

Related Topics

• What Are the Attributes of God? — where omnipotence sits in the divine property cluster.
• Why Is There Suffering? — Mackie's challenge and the free will defense.
• What Is Omniscience? — the companion attribute and its own puzzles.
• Theism vs Atheism — whether the attributes are coherent enough to argue over.