The Epicurean Paradox: Is God Willing but Unable?
The Epicurean paradox is the ancient argument that God cannot be both willing and able to prevent evil, because evil exists: if He is willing but unable, He is not all-powerful; if He is able but unwilling, He is not good; and if He is both, where does evil come from? The argument is valid in form and still fails, because it uses "evil" in one sense where the question needs two. Below are its text, its history, its strongest form, the standard answers, and five objections.
Key takeaways
- The Epicurean paradox asks whether God is willing but unable, able but unwilling, or both, and if both, why evil exists.
- Epicurus probably never wrote it. Lactantius attributed it to him without a source; the oldest surviving version is in Sextus Empiricus.
- Every standard answer pays somewhere. Process theism gives up God's power, open theism the guarantee; the classical answer leaves a gap.
- Frame-Indexed Classical Theism (FICT) dissolves the paradox. Where process theism and open theism each surrender a horn, it shows the horns read "evil" in one sense where there are two: the absolute evil God must defeat, and the finite evil He permits for a season.
What is the Epicurean paradox?
The Epicurean paradox is a trilemma about God and evil: three possible relations between God and the evil in the world, each of which seems to cost Him an attribute. Its most famous wording comes from David Hume's Dialogues Concerning Natural Religion, published after his death in 1779, where the skeptic Philo presses it on his opponent Cleanthes:
"Epicurus's old questions are yet unanswered. Is he willing to prevent evil, but not able? then is he impotent. Is he able, but not willing? then is he malevolent. Is he both able and willing? whence then is evil?" (Hume, Dialogues, Part X)
Each question removes one option. A God who wants to stop evil and cannot is not almighty. A God who can stop it and will not is not good. A God who both can and will leaves the existence of evil unexplained, and evil exists.
Free — the complete book
Get Book 1: Resolving the Free Will Dilemma free when you enter your name and email.
This essay takes one piece of it. Resolving the Free Will Dilemma takes the whole dilemma from first axiom to conclusion — the complete book, free.
No spam. Unsubscribe in one click.
Put formally, the Epicurean paradox claims that four statements cannot all be true together: God is all-powerful, God is all-knowing, God is perfectly good, and evil exists. The first three are the core of the classical attributes of God, the omnipotence of God and His omniscience among them. If the paradox holds, one of them must go.
It also goes by older names, the Epicurean trilemma and the riddle of Epicurus, and it is the oldest compact statement of one of the hardest questions a believer is asked. Its first target was not God's existence so much as His care. Lactantius, the early Christian writer who preserved it, says it left many defenders of providence "almost driven against their will to admit that God takes no interest in anything." The person who types the question today usually means the same thing: if God cares, why this?
The second volume of the series takes up every form of this question, from its logic to the grief behind it.
Did Epicurus write it? Epicurus and the problem of evil
Probably not, though the question is not settled. What is often called the Epicurus problem of evil reaches us through a Christian who opposed him. In the thirteenth chapter of On the Anger of God, Lactantius sets out the argument in its fullest ancient form:
"God, he says, either wishes to take away evils, and is unable; or He is able, and is unwilling; or He is neither willing nor able, or He is both willing and able. If He is willing and is unable, He is feeble, which is not in accordance with the character of God; if He is able and unwilling, He is envious, which is equally at variance with God; if He is neither willing nor able, He is both envious and feeble, and therefore not God; if He is both willing and able, which alone is suitable to God, from what source then are evils? Or why does He not remove them?" (Lactantius, On the Anger of God, ch. 13)
Notice the fourth option, which Hume dropped: a God neither willing nor able would be no God at all.
Epicurus himself was not an atheist. He held that the gods exist, but in untroubled bliss, with no concern for human affairs. Read in that setting, the argument was an attack on divine providence, the claim that God governs and cares for the world, rather than a proof that no god exists.
Whether Epicurus ever made it is doubtful. As Wikipedia's summary of the scholarship reports, no surviving text of Epicurus contains the argument, and Lactantius names no source. The oldest preserved version of the trilemma appears in the writings of the skeptic Sextus Empiricus. The German scholar Reinhold Glei argues that it came from a non-Epicurean or anti-Epicurean academic source, and the skeptic Carneades has been proposed as its author. Scholars have not settled the question.
The modern wording travelled from Lactantius through Pierre Bayle's Historical and Critical Dictionary to Hume, whose short version became the one people quote. The name stuck to Epicurus because Lactantius put it there.
Every volume in the series begins from one such ancient puzzle and works it through to its end.
Why is the Epicurean paradox so hard to answer?
Because it looks like plain logic, and in the twentieth century philosophers said so outright. J. L. Mackie, in "Evil and Omnipotence" (1955), wrote that here "it can be shown, not that religious beliefs lack rational support, but that they are positively irrational, that several parts of the essential theological doctrine are inconsistent with one another" (quoted in the Internet Encyclopedia of Philosophy).
The four statements do not contradict one another on their face. The contradiction appears only when bridge premises are added, and the IEP sets them out. An all-powerful God would be able to prevent all evil. An all-knowing God would know about it and know how to prevent it. A perfectly good God would want to prevent it. Together these imply that if God existed there would be no evil.
The decisive step is the third: a wholly good being eliminates evil as far as it can. That is the steelman of the Epicurean paradox, and it is strong, because it sounds close to what "good" means. We blame a lifeguard who could have saved a drowning child and did not.
The Christian tradition did not dodge this. Thomas Aquinas made it the first objection to God's existence in the Summa Theologiae: "if one of two contraries be infinite, the other would be altogether destroyed. But the word 'God' means that He is infinite goodness. If, therefore, God existed, there would be no evil discoverable; but there is evil in the world. Therefore God does not exist" (I, q. 2, a. 3).
Scripture asks it too. After God announces that He will use the Babylonians to judge Judah, the prophet Habakkuk holds both premises in a single breath:
"Your eyes are too pure to look on evil; you cannot tolerate wrong. Why then do you tolerate the treacherous? Why are you silent while the wicked swallow up those more righteous than themselves?" (Habakkuk 1:13, NIV)
When the strict logical version was blunted, the debate moved to the evidential version, which says that evil makes God improbable rather than impossible. That form has its own article on why God allows suffering.
Volume One builds the method this answer rests on, and it is free to download.
How do Christians usually answer the Epicurean paradox?
Each standard answer takes one horn, or rejects the hidden premise, and each has real strength.
Lactantius's own answer. Lactantius quoted the argument in order to answer it. God, he says, "is able, therefore, to take away evils; but He does not wish to do so, and yet He is not on that account envious," because "He at the same time gives wisdom," and "unless we first know evil, we shall be unable to know good." The answer rightly denies that goodness must remove every evil. It strains by making evil a necessary condition of the good it serves, so that a world without evil could hold no wisdom and no virtue.
Augustine and Aquinas. Augustine gave the classical reply in the Enchiridion: God, "being Himself supremely good, would never permit the existence of anything evil among His works, if He were not so omnipotent and good that He can bring good even out of evil." He adds the privation account: "For what is that which we call evil but the absence of good?" (Enchiridion 11). Aquinas answers his own objection by quoting this passage, and concludes, "This is part of the infinite goodness of God, that He should allow evil to exist, and out of it produce good." The Thomist answer keeps all three attributes and explains what evil is. It leaves one thing unsaid: why a goodness that opposes evil does not require its removal now. Talk of God's permissive will names the permission without explaining it.
The free will defence. Alvin Plantinga argued that God may have a morally sufficient reason to permit evil, the creation of significantly free creatures, and that this is enough to show consistency. The IEP reports that "Even Mackie admits that Plantinga solved the problem of evil, if that problem is understood as one of inconsistency." The defence shows possibility only, and it says little about natural evil. It is treated in full, with skeptical theism, in our article on suffering, and its view of freedom in libertarian free will.
Process theism takes the first horn. Process theists accept that God is willing but unable. In the Stanford Encyclopedia's account, "A central contention of process theism is that the problem of evil is aggravated by flawed accounts of omnipotence commonly assumed by theists and their critics," and "God acts by persuasion rather than by coercion." The gain is that God is clearly not the author of evil. The cost is that a God who can only persuade cannot promise that evil will end, and the view sits uneasily with creation from nothing.
Open theism softens "able." Open theism holds that God restrains His power for love's sake. As the IEP puts it, "He cannot both respect our libertarian freedom and guarantee that we will do specific things freely." Freedom and risk are real. The cost is stated by the same entry: "If Open Theism is true, then there is no guarantee that everything will work out as God wants in the end."
So every answer either surrenders a horn or keeps all three without saying what distinguishes the evil God permits from the evil His goodness must remove.
The books weigh each of these positions at full length before answering them.
How Frame-Indexed Classical Theism answers the Epicurean paradox
Frame-Indexed Classical Theism (FICT) is a research program in analytic theology that resolves classical theological dilemmas by diagnosing them as artifacts of a single formal error: one term carried across two frames as though it meant the same thing in each. On the Epicurean paradox, process theism gives up God's power and open theism gives up His guarantee, while the classical answer keeps both without naming the distinction it needs. FICT names it.
Hume's Philo states the hidden assumption himself, in the sentence just before Epicurus's questions. He asks Cleanthes whether he will still "assert the moral attributes of the Deity, his justice, benevolence, mercy, and rectitude, to be of the same nature with these virtues in human creatures?" and then: "In what respect, then, do his benevolence and mercy resemble the benevolence and mercy of men?" The trilemma has its force only if goodness and evil mean exactly the same for God as for us, the assumption of univocity. It presses hardest on theistic personalism, the picture of God as a greater person whose goodness is ours enlarged.
Volume Two argues that "evil" carries two senses. Absolute evil is a settled, final and deliberate repudiation of the good by a will able to uphold perfect righteousness. God's nature is necessarily opposed to it, and He will finally exclude it. No fallen human being can commit it, because fallen humanity lacks the capacity. Finite evil is the evil we actually commit and suffer. It is real, culpable and grievous, and it is "not rendered 'not truly evil'" when seen from God's side. God permits it for a season, for morally sufficient reasons, under certain judgment.
Now ask the questions again. Is God willing to prevent evil? Yes: willing to defeat it absolutely, and He will. Is He able? Yes. Then whence is evil? What exists is finite evil, permitted for a time, not the absolute evil His nature cannot tolerate. His opposition and His permission fall on two different things, so nothing is affirmed and denied in the same sense. Nor in the same time: His goodness is not a state standing beside our suffering on one timeline, waiting to be reconciled with it. The Epicurean paradox is not answered by giving up a horn. It dissolves.
Habakkuk received the same answer in his own terms: "For the revelation awaits an appointed time; it speaks of the end and will not prove false. Though it linger, wait for it; it will certainly come and will not delay" (Habakkuk 2:3, NIV). Joseph saw it worked out in one life: "You intended to harm me, but God intended it for good to accomplish what is now being done, the saving of many lives" (Genesis 50:20, NIV). And the end is promised: "He will wipe every tear from their eyes. There will be no more death or mourning or crying or pain, for the old order of things has passed away" (Revelation 21:4, NIV).
The result keeps process theism's insistence that God does not author evil without trimming His power, keeps open theism's real freedom without losing the guarantee, and says what the classical answer left unsaid. The cost is univocity, and the position is refuted by any dilemma shown to turn on no term of this kind. The statement of FICT sets it out, and the companion posts on omnibenevolence and whether God created evil carry it further.
The full argument is Volume Two, Resolving the Problem of Evil.
Does the dissolution of the Epicurean paradox survive its strongest objections?
We have run thousands of simulated attacks against Frame-Indexed Classical Theism, drawn from academic and peer-reviewed sources. The five below bear directly on the Epicurean paradox, and each is met from Volume Two's own pages. They draw on the free will defence, on skeptical theism and its critics, and on the open theist William Hasker.
Can the paradox be restated entirely from God's side?
Here's one of those attacks and how the Many Beings framework would address it.
Alvin Plantinga, in God, Freedom, and Evil (1974), framed the logical problem as a question of consistency: are "God is omnipotent, omniscient and wholly good" and "evil exists" consistent? The attack builds the argument wholly in God's own terms. God is wholly good. Finite evil is evil and offends His holiness, as Volume Two itself says. And a wholly good, all-knowing, all-powerful being eliminates every evil it can, unless doing so would forfeit an outweighing good. Every term is read from God's side, and no human sense is swapped in, so the reply that the paradox mixes two frames seems to have nothing to grip. The question becomes whether an outweighing good is possible, which is Plantinga's question, not a question about frames.
The defense is that even on God's side "evil" keeps two senses. His necessary opposition falls on absolute evil, His permission falls on finite evil, and the two are not the same thing. Nothing is affirmed and denied of one object in one sense. As for the third premise, once it carries its own clause about an outweighing good, it says what Volume Two already says: God permits finite evil only for morally sufficient reasons, temporarily, and under judgment. The attack therefore confirms the answer rather than defeating it.
Must a good God prevent every evil He can?
Here's one of those attacks and how the Many Beings framework would address it.
William Rowe's evidential argument states the second horn in modern form: "An omniscient, wholly good being would prevent the occurrence of any intense suffering it could, unless it could not do so without thereby losing some greater good or permitting some evil equally bad or worse." William Hasker, in "The Necessity of Gratuitous Evil" (1992), pointed out that how grave an evil is and whether it is pointless are separate questions. The attack follows. Either the line between finite and absolute evil releases God from preventing finite evil whether or not it serves a purpose, which concedes pointless evil, or it does no work against Rowe at all.
The defense is that the distinction does a narrower job than the attack assigns it. It removes only the obligation to eradicate finite evil immediately. It never removes the need for reasons: the same passage says God permits finite evil "for a time to fulfill greater purposes," and Volume Two's central premise makes permission consistent with goodness only when it is for morally sufficient reasons. So the first reading is not the book's. On the second reading the distinction still does work, because no reason could ever license permitting absolute evil. It is what makes reasons relevant to finite evil at all. Against the claim that some evil is pointless, the book uses a different tool, the one in the next attack.
If we see no reason for an evil, is there probably none?
Here's one of those attacks and how the Many Beings framework would address it.
This is the evidential form of "whence then is evil?" Skeptical theists, following Stephen Wykstra, block the step from "I see no reason" to "there is probably none." Andrew Graham and Stephen Maitzen ("CORNEA and Closure," 2007) objected that such restrictions break a basic rule of knowledge: if we know a claim and know what it entails, we may accept what it entails. Wykstra answered in the same journal issue. The attack presses the dilemma on Volume Two. Its own conclusions about God's governance, for instance that His goodness does not require the immediate end of every finite evil, are claims about God reached from our side. Either the restriction blocks those too, or it blocks nothing.
The defense distinguishes two ways of knowing. What we know by reasoning from revelation is one mode; what appears to us directly is another. The book's conclusions about God's governance are drawn by entailment from revealed and reasoned premises, so nothing blocks them. "I see no reason for this evil" begins from how things appear to us, and then extends that appearance into a verdict about the whole of God's reasons, which no premise supplies. Entailment carries only what its premises contain. The skeptical theist's principle is near the truth, but the boundary lies between these two ways of knowing. The suffering article sets out skeptical theism in full.
Does this leave us unable to call anything evil?
Here's one of those attacks and how the Many Beings framework would address it.
William Hasker, in "All Too Skeptical Theism" (2010), argued that the usual reply to the paradox costs too much. If we cannot judge whether an evil is pointless, we cannot judge whether anything is good or evil all things considered, because that judgment depends on connections to goods beyond our sight. On Volume Two's own account every event is held within God's single governance of the world, so the restriction seems to reach every event, and moral judgment collapses into skepticism.
The defense is that our judgment that an evil is evil does not claim to place it within the whole of providence. It is a judgment of better and worse in our own finite frame, where, as Volume Two puts it, "distinctions between relatively better and worse actions are meaningful and necessary." An evil stays evil when God overrules it toward good; it is not made "not truly evil" by His purposes. What the reply withholds is only the verdict about the whole: whether the world is better or worse for this event once every connection is counted. That verdict was never ours to give, and the believer's judgment that relieving suffering is good does not need it.
If every evil has a reason, why should we fight it?
Here's one of those attacks and how the Many Beings framework would address it.
In The Triumph of God over Evil (2008), Hasker pressed a practical objection. God wants human beings to take responsibility for one another's welfare. But if I believe that God permits no evil without a sufficient reason, then any evil I fail to prevent will be one God has made sure is justified. Belief in that, he argues, weakens the motive to prevent evil, and he contends that believers often fail to see the tension.
The defense keeps human duty in the human frame. God's justified permission of an evil is a verdict from His side. It does not make the evil "no loss" in ours, where it is felt exactly as it is. Volume Two also fixes an asymmetry. Finite evil God permits without authoring it, while finite good, including the good of relieving harm, He positively enables and sustains by His grace. The one who prevents harm does what God enables. The one who fails to prevent it only lets happen what God merely permits. The duty keeps both its ground and its point.
Volume One, where these frames are first set out, is free.
Frequently asked questions
Did Epicurus actually write the Epicurean paradox?
Probably not. No surviving text of Epicurus contains the Epicurean paradox. Lactantius attributed it to him in On the Anger of God without naming a source, and the oldest preserved version appears in the skeptic Sextus Empiricus. Some scholars trace it to a non-Epicurean academic source, possibly Carneades. Hume's famous wording came through Pierre Bayle, who was quoting Lactantius, and Hume shortened it from four options to three.
Did Epicurus believe in God?
Epicurus believed that gods exist, so he was not an atheist. He held that they live in untroubled happiness and take no interest in human affairs. That is why the argument attributed to him targets providence, God's care for the world, more directly than God's existence. Lactantius saw this: the danger, he wrote, was that defenders of providence would be driven to admit "that God takes no interest in anything."
Has the Epicurean paradox been refuted?
In its strict logical form it is widely regarded as answered: even J. L. Mackie, its leading modern defender, conceded that Plantinga's free will defence shows God and evil are consistent. A defence of that kind shows only that God could have a reason, without saying what the reason is. The evidential form, that evil makes God improbable, is still debated, and our article on suffering takes it up. Volume Two argues that both forms rest on one sense of "evil" where there are two.
What is the difference between the Epicurean paradox and the problem of evil?
The problem of evil is the whole family of arguments that evil counts against God. The Epicurean paradox is its oldest compact member, a trilemma of willing and able that targets God's care. Modern versions divide into the logical problem, which claims a contradiction, and the evidential problem, which claims improbability. Related questions include whether this is the best of all possible worlds.
What does the Bible say about the Epicurean paradox?
Scripture asks the same question without denying either premise. Habakkuk says to God, "Your eyes are too pure to look on evil; you cannot tolerate wrong. Why then do you tolerate the treacherous?" (Habakkuk 1:13, NIV). God answers that the revelation "awaits an appointed time" (Habakkuk 2:3, NIV): evil is permitted for a season, and its end is certain. Joseph's words to his brothers show the same pattern in one life: what they intended for harm, "God intended it for good" (Genesis 50:20, NIV).
Where to go from here
Epicurus's questions were never unanswerable. They were asked in one sense of a word that has two, and the answers that gave up a horn paid for a mistake in the question. Habakkuk asked the same thing with both premises intact, and he was told to wait for the appointed time. The answer he received denied neither premise. It placed the evil he saw inside a time God had set, with its end already written.
Volume One, Resolving the Free Will Dilemma, lays the foundation and is yours free. The series carries the same method into evil, time and changelessness.