You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
The logical problem of evil is the most uncompromising philosophical challenge to classical theism. It does not merely claim that evil makes God's existence unlikely; it claims that the existence of evil is logically incompatible with the existence of an omnipotent, omniscient and wholly good God. If the argument succeeds, then to assert both "God exists" and "evil exists" is as self-contradictory as asserting that a single shape is at once a perfect circle and a perfect square. At least one proposition must be abandoned. This is a deductive argument: if its premises are true, its conclusion follows with logical necessity. For AQA A-Level Religious Studies (Component 1A, "Evil and suffering"), this topic supplies the framework within which every theodicy and defence must operate, so it is essential to grasp its structure with precision rather than merely repeating that "evil is a problem for believers".
The problem has ancient roots. It is conventionally traced to the Greek philosopher Epicurus (341–270 BCE), although the dilemma most often quoted survives not in Epicurus's own writings but in the work of the later Christian author Lactantius, who reports and attacks it in De Ira Dei (On the Anger of God). The so-called Epicurean paradox runs as a cascade of disjunctions:
"Is God willing to prevent evil, but not able? Then he is not omnipotent. Is he able, but not willing? Then he is malevolent. Is he both able and willing? Then whence cometh evil? Is he neither able nor willing? Then why call him God?"
The force of the dilemma is that it appears to exhaust the options. Each branch either denies one of God's perfections — power or goodness — or fails to explain why evil exists at all. The challenge for the theist is to show that there is a possibility the dilemma has overlooked, which is precisely what the later theodicies and the free will defence attempt to do. It is worth noting that the dilemma assumes a particular conception of God: a single being who is both maximally powerful and maximally good. It has no grip on a dualist who posits an evil principle alongside the good (as Manichaeism did), nor on a deity conceived as limited in power (as process theology will propose in a later lesson). The problem is therefore specifically a problem for classical monotheism, which insists on the conjunction of omnipotence and perfect goodness in one God.
The problem is generated by the conjunction of three traditional "omni" attributes, and it helps to state each precisely before seeing how they collide with the fact of suffering.
Put together, the three attributes seem to entail a world without (unnecessary) evil: an all-knowing God would know of every evil, an all-powerful God would be able to prevent it, and an all-good God would want to prevent it. The persistence of evil therefore looks like proof that at least one attribute is not fully instantiated — or that no such being exists. Mackie's contribution was to show that this intuitive worry can be tightened into a formally valid deductive argument, so that the theist cannot dismiss it as mere emotional protest. Note that omniscience plays a quieter role in Mackie's own version, which foregrounds power and goodness; but it matters for blocking the escape route "perhaps God does not know", and it becomes central to Rowe's evidential version in Lesson 2.
Key term: A deductive argument is one whose conclusion follows necessarily from its premises, so that it is impossible for the premises to be true and the conclusion false. The logical problem of evil is deductive; the evidential problem (Lesson 2) is inductive.
J. L. Mackie (1917–1981) gave the logical problem its sharpest modern statement in his 1955 article "Evil and Omnipotence", published in the journal Mind. Mackie's aim was explicit and ambitious. He argued not merely that theistic belief is unsupported by evidence but that it is positively irrational, because the believer holds a set of beliefs that are logically inconsistent with one another. He maintained that the problem of evil shows "not that religious beliefs lack rational support, but that they are positively irrational, that the several parts of the essential theological doctrine are inconsistent with one another".
The inconsistency, Mackie argued, lies in three propositions that orthodox theism affirms together:
Mackie called these three the inconsistent triad: any two can be held together, but not all three. A believer who admitted that God is not omnipotent, or that God is not wholly good, or that evil does not really exist could escape the contradiction — but each escape route abandons a doctrine central to classical theism.
Crucially, Mackie recognised that the triad is not explicitly contradictory. The bare propositions "God is omnipotent", "God is wholly good" and "evil exists" do not by themselves form the words of a formal contradiction. The contradiction emerges only when we add what Mackie called certain additional premises, or quasi-logical rules, that connect the key terms:
Given these two connecting principles, the contradiction is generated: a wholly good being would want to eliminate all evil, an omnipotent being would be able to eliminate all evil, and therefore if such a being existed, evil would not exist. Mackie insisted these rules are not arbitrary additions but unpack the very meaning of "wholly good" and "omnipotent". The strategic significance for the theist is that any successful response must show that at least one of these quasi-logical rules is false — for instance, by arguing that even an omnipotent being cannot do the logically impossible, or that a good being might tolerate some evil for the sake of a greater good it could not otherwise secure.
The argument can be set out formally:
The argument is valid: granted the premises, the conclusion cannot consistently be denied. The whole weight of the debate therefore falls on whether the premises — in particular Premises 1 and 2, which encode the quasi-logical rules — are actually true.
Mackie anticipated the standard theistic replies and divided them into "adequate solutions" and "fallacious solutions". An adequate solution would genuinely dissolve the problem — but, Mackie argued, every adequate solution does so only by tacitly giving up one of the three propositions, for instance by quietly limiting God's power or goodness. The more revealing cases are the fallacious solutions, which claim to retain all three propositions but, on inspection, smuggle in a contradiction. Mackie examined four:
Key term: The paradox of omnipotence asks whether an omnipotent being can create things it cannot subsequently control — for example, genuinely free agents, or rules it cannot then break. Mackie uses it to argue that the notion of unlimited power is itself problematic, and that, in any case, God could have made free agents who never in fact choose to sin.
Mackie's discussion of the free-will reply leads him into a deeper puzzle about the very concept of unlimited power, which is worth setting out because it shows how carefully he tried to close every theistic exit. The paradox runs: can an omnipotent being make things that it cannot afterwards control? If God can make a being whose actions are outside God's control, then there is something God cannot do — namely, control that being — so God is not omnipotent. If God cannot make such a being, then again there is something God cannot do — namely, make it — so once more God is not omnipotent. Either way, the notion of unqualified omnipotence seems to generate a contradiction, much as the old chestnut "can God make a stone too heavy for God to lift?" does.
Mackie distinguishes first-order omnipotence (unlimited power to act on things) from second-order omnipotence (unlimited power to determine what powers things shall have). He argues that a being cannot coherently possess both: a God with the second-order power to create genuinely independent free agents thereby limits its own first-order power over them. The relevance to the problem of evil is direct. The theist wants to say both that God is omnipotent and that God made free agents whose evil choices God does not control. Mackie's paradox presses the theist to admit that, in making free agents, God either retains control (in which case the agents are not really free, and God, not they, is responsible for evil) or relinquishes control (in which case God's omnipotence is qualified). This is the conceptual pressure that Plantinga's free will defence (Lesson 5) must withstand, and it explains why so much of that defence turns on the precise sense in which God can or cannot "actualise" a world of free creatures.
Much turns on how omnipotence is defined, because the theist's best reply is to deny that "there are no limits to what an omnipotent thing can do".
If omnipotence is maximal power, the door opens for the free will defence: perhaps a world of genuinely free creatures who never sin is, like a round square, not a coherent thing for even an omnipotent God to actualise. If omnipotence is absolute power, that escape is closed. Most theists adopt the maximal-power definition, which is why Lesson 5's defence becomes the decisive move against Mackie.
Students must keep the logical problem distinct from the evidential problem that follows it. The table summarises the contrast.
| Feature | Logical problem (Mackie) | Evidential problem (Rowe) |
|---|---|---|
| Form of argument | Deductive — aims at certainty | Inductive — aims at probability |
| Central claim | God and evil are logically incompatible; they cannot coexist | The amount and kinds of evil make God's existence improbable |
| What it must show | A strict contradiction in the set of theistic beliefs | That some evil is gratuitous (pointless) |
| How it is answered | By exhibiting one possibly-true proposition that removes the contradiction (the free will defence) | By arguing we cannot tell that any evil is truly gratuitous (sceptical theism) |
| Current standing | Widely thought resolved in the theist's favour | Live and actively debated |
It is widely accepted — by atheist philosophers as well as theists — that Plantinga's free will defence has shown the strict logical incompatibility cannot be demonstrated. Mackie himself, in The Miracle of Theism (1982), conceded that Plantinga's argument succeeds against the logical version, though he still regarded the existence of God as improbable. The contemporary debate has therefore largely migrated to the evidential terrain of Lesson 2.
Although Mackie gave the problem its definitive analytic statement, the modern discussion is shaped by an earlier source. David Hume's Dialogues Concerning Natural Religion (published posthumously in 1779) puts a version of the problem into the mouth of the character Philo, who revives "Epicurus's old questions" and presses them against the design argument: the order of nature does not obviously bespeak a benevolent author, given the suffering it contains. Philo argues that, even if the world's mixed character is logically compatible with a good God, it gives us no positive reason to infer one — a thought that anticipates the evidential problem more than the logical one. Hume thus supplies the bridge between Epicurus's ancient dilemma and the two distinct modern forms (logical and evidential) that this course separates. For exam purposes, it is enough to know that Hume's Dialogues is the classic early-modern treatment, and that he frames the problem evidentially — as undermining any inference from the world to a benevolent designer — rather than as a strict deductive contradiction.
The logical problem sets the agenda for everything that follows in this course, so it helps to map the responses in advance. Broadly, a theist can attack the argument at one of three points:
Each strategy targets a specific premise of Mackie's argument, which is why precise identification of the premises (above) is the foundation for the whole unit. A response that does not say which premise it rejects has not engaged the logical problem at all.
The problem draws its force from the breadth of evil to be explained. The standard division is:
The distinction is strategically important. The free will defence can in principle account for moral evil, as the unavoidable risk of genuine freedom; but natural evil has no obvious link to human choice. This is why responses to natural evil tend to appeal either to soul-making (Lesson 4), or to the free choices of non-human agents (Plantinga's speculative extension in Lesson 5), or to the necessity of a law-governed world. Any complete response to the logical problem must address both categories.
Some philosophers add a third category, metaphysical evil, a term from Leibniz denoting the imperfection and finitude inherent in any created order: because creatures are not God, they are necessarily limited, and this limitation is the precondition of the other evils. Leibniz used the idea in arguing that ours is the "best of all possible worlds" — the most perfect world an omnipotent and benevolent God could create, in which the goods could not be increased nor the evils diminished without overall loss. The phrase was famously mocked by Voltaire in Candide (1759), where the relentless catalogue of disasters that befalls Dr Pangloss and his pupil satirises the complacency of saying that whatever is, is for the best. The Leibnizian "best possible world" reply is a recognisable cousin of Mackie's third "fallacious solution" — that the universe is better with some evil in it — and faces the same objection: it appears to concede that God deliberately tolerates evil, and it strains credulity to claim that this world, with its scale of suffering, could not have been improved upon by omnipotence.
It is worth isolating the single move on which almost every theistic reply depends, because recognising it unifies the whole unit. Mackie's second quasi-logical rule says a good being eliminates evil "as far as it can". Every theodicy modifies this to: a good being eliminates evil as far as it can without forfeiting a greater good that could not be obtained otherwise. Once that qualifier is admitted, the existence of evil no longer immediately entails the absence of God, because the evil might be the necessary cost of some outweighing good (freedom, moral growth, a stable natural order).
The whole subsequent debate then becomes a contest over whether the qualifier can be honestly applied:
Seen this way, Mackie's logical problem performs a precise service: it forces the theist to commit to the greater-good qualifier, and thereby to the claim that no evil is gratuitous — a claim the evidential problem then attacks head-on. A student who grasps this hinge can move fluently between the two problems and the theodicies that answer them.
A further turn of the screw comes from divine omniscience, and especially foreknowledge. Classical theism holds that God knows all truths, including truths about the future. If so, then before creating, God already knew exactly which evils would occur — every murder, every famine, every act of cruelty — and created anyway. This appears to deepen the problem in two ways. First, it removes the excuse of ignorance: God cannot be said to have permitted evil through any failure of knowledge. Second, it raises the question of culpability: a being who knowingly sets in motion a process whose evil outcomes it foresees in full detail seems more, not less, implicated in those outcomes than one who acts in ignorance. The free-will reply must therefore show not merely that creatures' choices are free, but that God's foreknowledge of those choices does not make God their author.
Theists respond in several ways. Some, following Boethius, argue that God is timeless: God does not foreknow the future as something still to come, but sees all of time "at once" in an eternal present, so divine knowledge no more causes our choices than my seeing you act now causes your action. Others adopt the view (associated with Luis de Molina, "middle knowledge") that God knows what every free creature would freely do in any circumstance, and uses this knowledge in choosing which world to create without thereby determining the choices — a view that connects closely to Plantinga's machinery of counterfactuals of freedom in Lesson 5. For the purposes of the logical problem, the key point is that foreknowledge intensifies the demand on the free-will defence: it is not enough that creatures are free; their freedom must be shown to be compatible with God's complete prior knowledge of how they will use it. Whether that can be made out is one of the deepest questions the unit raises.
Given the consensus that Plantinga has answered the strict logical problem, a student might ask why it is still studied. There are good reasons:
In short, the logical problem is less a museum-piece than the keystone of the arch. One can hold, with the philosophical mainstream, that it has been answered, while still recognising that the answer reshapes rather than removes the difficulty: the price of escaping the logical problem is a commitment (no evil is gratuitous) that the evidential problem will test to destruction. The remaining lessons can be read as an extended examination of whether that commitment can be sustained.
Question: "The logical problem of evil proves that God does not exist." Discuss.
Introduction. Mackie's logical problem is the strongest form the problem of evil can take, because it claims not merely improbability but outright contradiction. I shall argue that, although Mackie correctly identifies a genuine tension in classical theism, the argument does not prove that God does not exist, because the free will defence shows that the alleged contradiction cannot be established. The logical problem succeeds in raising the stakes but fails in its deductive ambition.
The case for Mackie. The strength of the inconsistent triad is its rigour. If a good being eliminates evil as far as it can, and an omnipotent being can do anything logically possible, then an omnipotent and wholly good God would eliminate all evil — so the presence of any evil entails the absence of such a God. This is deductively valid; the theist cannot simply concede that evil exists and carry on, because the argument targets the very coherence of the belief-set. Mackie strengthens his case by dismantling the standard replies: the "good requires evil" move limits omnipotence, and the free-will reply, he argues, fails because God could have made free agents who always choose the good. At the level of first impression, the triad looks decisive.
The decisive objection. However, Mackie's argument depends entirely on the quasi-logical rule that "there are no limits to what an omnipotent thing can do". Once omnipotence is properly defined as the power to do anything logically possible — Aquinas's definition, accepted by almost all theists — Mackie's claim that God could make free agents who always freely choose the good becomes questionable. Plantinga's free will defence shows it is at least possible that every creature God might create would, if genuinely free, go wrong at least once (transworld depravity). If that is even possibly true, then it is possibly true that God could not actualise a world of free creatures without some evil, and the triad is not formally inconsistent after all. Crucially, the theist does not need to show that this is actual, only that it is possible — and mere possibility is enough to break a claim of strict contradiction. This is why Mackie himself later conceded the point.
Counter and reply. A defender of Mackie might respond that the free will defence accounts only for moral evil, leaving natural evil — the fawn in the forest fire, childhood cancer — untouched. This is a serious objection, and Plantinga's appeal to fallen angels to cover natural evil strikes most readers as ad hoc. Yet even here the logical argument fails on its own terms: to refute a claim of logical impossibility, the theist needs only a logically possible story, however improbable, and the fallen-angel hypothesis, whatever its implausibility, is not self-contradictory. The natural-evil objection has real force, but it is an evidential point dressed up as a logical one.
Judgement. I conclude that the logical problem does not prove that God does not exist. It establishes a genuine demand — that the theist explain how an all-good, all-powerful God could permit evil — and it rightly forces precision about omnipotence. But because the demand can be met by a merely possible defence, the deductive claim of contradiction collapses. The honest verdict is the one Mackie himself reached: the logical problem fails, and the real battleground is the evidential question of whether the sheer quantity of evil makes theism improbable. The argument is therefore better seen as the opening of the debate than as its conclusion.
Examiner-style commentary: A Mid-band response would accurately state the inconsistent triad and list Mackie's rejected solutions, but treat the conclusion as obvious and stop there. What lifts this answer to Stronger is that it engages the structure of the argument — identifying the quasi-logical rule about omnipotence as the load-bearing premise and showing exactly where the free will defence applies pressure. The Top-band features are the sustained line of reasoning that distinguishes "possible" from "actual" (the key to why a mere defence defeats a logical argument), the handling of a genuine counter (natural evil) followed by a reasoned reply, and a judgement that is argued for rather than asserted, including the self-aware point that Mackie himself conceded. The answer evaluates rather than describes, which is the heart of AO2.
Exam tip: State Mackie's triad precisely and name the quasi-logical rules — that is where the marks are. Show you understand that a defence needs only to establish logical possibility, so a single possibly-true proposition defeats a deductive argument. Never simply list types of evil; examiners reward engagement with the logical structure and an awareness that the contest has moved on to the evidential problem.