The Evidential Argument from Evil (1998)

Nicholas Tattersall


"Please understand, it is not God I do not accept, but the world he has created."

Feodor Dostoyevsky
The Brothers Karamazov


The purpose of this essay is to explain and argue against various common objections to the Argument from Evil (AE). It is necessary to first formulate the argument and explain some of its history. Second, the refutations of various criticisms to AE are generalised in such a way that the reader is able to provide a critique of many of defences and theodicies himself. As an example, these methods are applied to a theodicy developed by Richard Swinburne. Third, the objections used by Ravi Zacharias are examined. Finally, the 'defensive scepticism' argued by many philosophers of religion is discussed.

The Arguments

Alvin Plantinga has suggested that the 'problem' of evil can be taken to mean different things[1]. He differentiates between the existential problem of evil and the epistemic problem of evil. For example, a theist may wonder why there is so much apparently pointless suffering in the world. As a result of witnessing terrible anguish or misery, the believer may become suspicious, angry or bitter towards God. Nevertheless, he may never question whether or not his belief in God is factually correct. This is the pastoral or existential problem of evil. Alternatively, if the believer were to question whether the existence of apparently pointless evil were to render belief in God false or improbable, he would be dealing with a different sort of problem -- the epistemic problem of evil. Atheologians have tried to construct arguments for the nonexistence of God from the existence of evil.

There is no single argument from evil. Instead, it is found that there are numerous arguments that can be classified into two families. The members of one group conclude that the theistic God cannot exist; the members of the other group attempt to show that the existence of such a being is improbable. The former arguments are described as logical, deductive or a priori and the latter are understood as evidential, inductive or a posteriori. The distinction between and history behind the different classes is now briefly considered.

The logical argument from evil is thousands of years old. It has come down the centuries from Epicurus, and has more recently been defended by the late John Mackie[2]. Essentially, the problem is why an omniscient, omnipotent and wholly good God would allow the extreme suffering we see in the world. By simply analysing the definition of God given above and stating a few uncontroversial premises, it was thought to have been shown that such a being would have the knowledge, ability and desire to prevent the intense suffering and premature death of sentient creatures. Consequently it was thought that the theistic God was logically incompatible with universally accepted facts about the world.

Such an argument is now widely accepted as being inadequate. This is because it not true that, by definition, God is incompatible with evil. God would not exclude all evil if He had morally sufficient reason for allowing some. For example, some evil may be necessary for the realisation of a greater good, known or unknown to us. Philosophers nowadays concern themselves with evidential arguments from evil.

There are a number of evidential arguments from evil. The most common of these simply reformulates the logical argument into a probabilistic argument:


(1) Gratuitous evils probably exist


(2) Gratuitous evils are incompatible with the God of theism (omnipotent, omniscient, all-good)


(3) Therefore, the God of theism probably does not exist

Clearly, many evils can clearly lead to greater goods, for example suffering the slight pain of an injection in order to receive a life-saving drug. Advocates of AE do not generally have this kind of thing in mind when they refer to suffering or evil. What is meant by 'evil' is gratuitous or pointless evil.

How can this accusation of gratuitous evil be met? Theists usually agree with (2) but deny (1). On the theistic view, God has very good reasons for natural evils (those brought about by natural processes) and moral evils (those due to human action), and a variety of different explanations have been put forward to say what the reasons might be. These explanatory defences are known as theodicies. Popular apologists usually take the approach of trying to neutralise AE by devising these explanations. However, many professional philosophers of religion -- both theist and atheist alike -- would agree that existing theodicies are flawed. Of numerous available examples from theists, Peter van Inwagen has admitted of AE, "Examination shows there is no known way of answering this case, and there is good reason to think that no way of answering it will be forthcoming."[3] Plantinga has written, "...many of the attempts to explain why God permits evil -- theodicies as we might call them -- seem to me shallow, tepid and ultimately frivolous."[4] It will be shown that there is a common core of faults shared by most existing (and probably most future) theodicies.

So the usual formulation of AE attempts to show that evil presents good evidence for, rather than proof of, the nonexistence of the theistic God. This opens the possibility that the theist could formulate an argument against AE as follows:


(1*) The God of theism exists


(2*) Gratuitous evils are incompatible with the God of theism (omnipotent, omniscient, all-good)


(3*) Therefore there are no gratuitous evils

It is evident that the critical premise is (1*). If there were compelling evidence for the existence of God then theists would not need to refute AE. They could simply argue that it is more probable that God exists than that there are gratuitous evils. However, this could only be done with the admission that AE lowers the probability of theism to some extent. Furthermore, theists are not generally keen to acknowledge that there are any good reasons for doubting God's existence.

For atheists to establish a stronger conclusion (e.g. that God's existence is less probable than His nonexistence), AE should accompany a refutation of the traditional proofs of theism. However, this is not difficult to do as the arguments all have serious problems[5], and at the very best would manage to establish that some supernatural agents exist. It is important to note that many traditional proofs such as the cosmological and teleological arguments are compatible with polytheism, deism and finite and evil gods[6]. Moreover, a god that is evil or not all-knowing or not all-powerful would be less vulnerable to AE[7]. So any attempt to explain evil by establishing (1*) to a greater probability than (1) would require powerful arguments specific to theism that no theist has ever put forward. Such a tactic must therefore, for now, be rejected.

Undaunted by the difficulties posed by AE, theists have devised literally dozens of theodicies that attempt to explain why God chose to create a world with evils in it. For example, it has been suggested that suffering is a warning for humans to reform their sinful ways; suffering helps humans to accept God; suffering is essential for humans to learn the difference between good and evil; suffering is necessary for humans to exercise compassion. I will suggest that critical questions can be asked of many existing theodicies and that these questions can be generalised into a list. Before this is done, two things should be said. First, such a list shows the common core of faults in many theodicies; it is not intended to be exhaustive or a substitute for specific refutations. It is not the case that all of the points in the list will be relevant to the refutation of all theodicies. Second, the list is intended primarily for the simplistic theodicies given by popular apologists. It is recommended that the reader look elsewhere for a refutation of the free-will defence (FWD)[8].

Theodicy -- an overview

Why is producing a sound theodicy such a difficult task? The problem essentially comes down to the fact that things are so easy for the critic. When finding faults with attempted explanations of evil, all logical possibilities are open. Since God is omnipotent, all possible worlds with any self-consistent set of natural laws can be considered. Refutations of theodicies can often work by simply noting other possibilities open to God for achieving His goals which do not involve intense suffering.

It is often argued that because the universe operates through natural laws, God is powerless to prevent intense suffering. Furthermore, it is argued that God cannot use spectacular miracles to alter the frequencies of natural disasters, as humans would be unable to predict the outcomes of their actions and make free moral choices as a result. Since it is a greater good that humans make free moral choices, God cannot intervene. This objection fails, for many reasons.

First, great quantities of suffering could be reduced without recourse to divine intervention. For example, God could have actualised a world with natural laws that made natural disasters less probable. Humans (especially human children) could have been made less vulnerable to various diseases. Countless possibilities are open to an omnipotent being.

Second, God need not perform spectacular miracles to help the survivors of natural disasters. For example, God could perform preventative miracles that He knows will be undetectable by humans. These miracles might include alterations to the positioning of tectonic plates to stop earthquakes, or small modifications in atmospheric conditions to avoid droughts, floods or tornadoes.

Third, it is not true that the predictability of the nature would be totally broken down in the event of spectacular miracles. It is usually accepted that God intervenes in His world on a frequent basis to perform miracles. God could significantly decrease the quantity of natural evil in the world if He was more efficient in these interventions. For example, instead of performing miracles of apparently little use, such as angelic visitations, Marian manifestations or weeping statues, God could use His limited supply of miracles more efficiently to lessen the harmful effects of natural disasters.

Finally the available evidence opposes the suggestion that a greater frequency of miracles would adversely affect the human ability to make moral choices. If there is a certain threshold frequency of miracle-performing that God cannot exceed (or else human moral decision-making would be compromised), then we would expect that the frequency of miracles would be constant geographically and temporally. However, this is not the case. More miracles are supposed to have occurred in the past[9] and more alleged miracles occur in certain locations such as Lourdes. Those living in biblical times who observed countless miracles -- or the inhabitants of Lourdes a century ago -- would have witnessed many times more miracles than the current inhabitants of Bangladesh. It is a mystery, then, why God has not raised the frequency of miracles in Bangladesh to reduce the levels of intense suffering there.

The Problems with Theodicies

So, if we imagine a general theodicy that claims that the levels of intense suffering and premature death seen in the world are necessary for the some greater purpose what questions might we ask of it?


  2. Can we conceive of any logically possible world where the purpose is fulfilled and there is less intense suffering and premature death?


  3. Is the present system too ambiguous? Is it important for the theodicy that people are aware that evil and suffering are necessary for the purpose to be fulfilled?


  4. Does the theodicy explain the suffering of non-human animals, the majority of which occurred before humans existed?


  5. Does the theodicy work in all conceivable cases? To take one extreme instance of natural evil, does it explain the suffering of a one-year-old orphan who is buried and killed by a landslide and is never found or missed?


  6. Does the theodicy explain the drastic variations in suffering over time and in different places?


  7. Does the theodicy have doubtful ethical consequences? For example, should we be working against suffering according to the theodicy, or should we be allowing it as part of God's plan? Are the goods described all really worth the resulting suffering?

With these general questions now formulated, let us apply them to specific theodicies. For a first example, it is commonly argued that suffering is necessary as a warning for humans to reform their sinful ways. How does this explanation stand up to the questions posed in the list?

With regard (a) we can see that this theodicy immediately fails. There is a logically possible world that God could have actualised where people receive clear warnings to reform where there is no suffering. For example, instead of warning people by seriously injuring or killing them in natural disasters, God could warn people through dreams, by communicating the message through religious experiences, or by sending angels to inform people directly. Even if it is acknowledged that some suffering is necessary to warn people to mend their ways, we can imagine a world where there is far less suffering but people still receive enough suffering to be clearly warned.

As we move on to (b), the theodicy encounters further problems. If we must suffer as a warning from God, we would expect it to far more clear. Most people who suffer have no idea that it is part of a warning from God. Indeed, the whole question of whether evil is a warning is far from certain among theists. We would expect God to communicate His purposes to the sufferers of a natural disaster. He could easily do this through dreams, religious experiences or by writing it in holy books. For example, it could have been clearly stated in the Bible that the purpose of hurricanes is to remind people that they must commit less sin. Alternatively, the survivors of a hurricane could have dreams that inform them that the purpose of the hurricane was to help them to mend their ways.

On point (c) it should be noted that the theodicy fails to explain the suffering of non-human creatures. It seems implausible that non-human animals commit sin, as they are usually not thought to be responsible for their actions. Furthermore, almost all non-human animal suffering goes unwitnessed by humans so seems to be superfluous for warning purposes.

The theodicy runs up against problems at (d). Unwitnessed suffering of very young children does not appear to warn anyone of anything.

At point (e) we note that the theodicy is incapable of explaining the geographical variations in human suffering. For example, it is uncontroversial that an inhabitant of Bangladesh is far more likely to experience suffering (e.g. due to flooding) than a randomly chosen human elsewhere in the world. There is surely a possible world where suffering is distributed approximately evenly, giving everyone a chance to experience it first hand. Undoubtedly, learning about severe suffering and death (e.g. by TV news) may act to warn people to some extent, but it would seem reasonable that first hand experience of suffering is far more likely to effectively warn people. As a result, it seems that we can only make sense of the theodicy if we make the implausible assumption that the inhabitants of Bangladesh need far more warning than the average human.

Finally, referring to (f), it appears that the theodicy has dubious ethical consequences. When people work towards reducing human misery, their actions make the implicit assumption that such suffering serves no worthwhile purpose. Given that God is trying to warn people by causing them suffering, it is not at all clear that we should act to help people. For example, it seems that we would be directly opposing God's wishes if we give the casualty of a natural disaster some pain killing drugs before he experiences any significant suffering. To be in line with God's wishes, we should first allow the victim to receive a warning by suffering somewhat. Since this is absurd, we have good reason to doubt the plausibility of the theodicy.

Unarguably, this particular theodicy has too many problems to merit serious consideration. Of course, the example was chosen for purposes of illustration, so it is no surprise that refuting it turned out to be all too easy. It should be apparent to the reader that the same principles can be applied against numerous other explanations of evil.

As a final example, let us now consider a far more sophisticated attempt put forward by Richard Swinburne[10]. He has argued that suffering serves a number of purposes. To explain moral evils, he argues that that "The 'free-will defense,' carefully spelled out, must be a central core of theodicy, as it has been for the last two or three thousand years."[11]. As I mentioned above, critical questions can and have been asked of FWD elsewhere. Swinburne goes on to argue that there are many goods which require the existence of natural evils. For example, "Showing sympathy (as opposed to the passive state of feeling compassion), helping the suffering, and showing courage of a certain sort are like this."[12] He then argues at length that many evils can be explained in these terms. Further goods that Swinburne argues are brought about by natural disasters include the gratitude of the sufferers to their doctors, and the sufferers' fulfilled desire to be relieved of pain.

So, in evaluating Swinburne's theodicy, we come to (a). Are there any possible worlds where his criteria are fulfilled and there is less suffering? I will give several reasons for supposing this is so.

It has been noted elsewhere[13] that the exercise of compassion would be consistent with far less evil if God had given humans superior empathic abilities. There seems to be a possible world where humans with more sensitive empathic powers could read books or watch films and relate to the characters far better that we can. Note that the films and books need not be all fictional. Consequently, these humans -- if they so chose -- could experience far deeper feelings of compassion than we can. Additionally, such beings would be able to relate to the suffering reported on TV news. For example, some humans find it difficult or impossible to empathise with people suffering the effects of natural disasters thousands of miles away. Presumably, beings with greater empathic powers would be able to freely overcome these difficulties if they so chose. Since more compassion would be felt for each instance of suffering in such a world, there would be less suffering required.

Swinburne may object that he specifically avoided this objection by explicitly distinguishing sympathy from compassion. However, this distinction does not seem to be relevant. Beings capable of feeling greater levels of compassion would also be able to show more outward signs of sympathy.

We might also ask why God has not chosen to develop human courage in less harmful ways. In a world with little suffering, humans could show a great deal of courage. For example, it is not clear that the courage shown by a sports player who takes on superior opposition is qualitatively less significant than the courage shown by the victim of a natural disaster. Even if the courage shown in a natural disaster is qualitatively more significant, this is a contingent fact of human nature that God could have changed. There is surely a possible world where humans show their greatest courage, not when suffering, but when performing tasks harmless to their health such as competing in sports against superior opponents.

Even if evil is needed to help people show courage, we can wonder why the evils in our world are not better directed towards furthering this end. For example, a flood that kills a whole family may cause very little courage to be shown. The flood may strike quickly and if there are no immediate relations to the family there may be few or no mourners of the victims. If the members of the same family had been inflicted with some debilitating or fatal disease instead, there would have been far more opportunity for overcoming hardships, learning co-operation, showing sympathy and demonstrating courage, all of which, according to Swinburne, are justifying goods.

On Swinburne's view, helping the suffering is also a great good. However, we find that in this world, numerous factors reduce the extent to which we are able to help the suffering or show them any sympathy.

First, there is the problem of ignorance. Some people who live in very remote areas of the planet cannot possibly know anything about the suffering elsewhere in the world. This was certainly the case in the past before the recent invention of telecommunications. It is hard to see why God has chosen to create a world with this ignorance, resulting in fewer humans being able to feel sympathy for those who are great distances away. Moreover, there are countless possibilities open to an all-powerful God for rectifying this situation. These possibilities include humans having empathic abilities that operate over distances in some ways analogous to psychic powers, and humans having the global facts of suffering communicated to them in dreams and through religious experiences.

Second, there is the problem of geography. Some people find that the instances of the most severe suffering occur in countries other than their own. Giving money to charity is one way to help, but there is surely a possible world where people can donate money to help the most severe forms of suffering and easily give some of their time to work with the sufferers. For example, suffering due to natural disasters could have been distributed more evenly around the planet, rather than localised in certain regions. At present, some willing people in the relatively rich West find that there are insurmountable financial and geographical obstacles to overcome if they wish to frequently travel to the Third World to give direct help.

Finally, there is the problem of quantity. It might be wondered whether the amount of suffering in the world is consistent with the theodicy. How much suffering would we expect for the purposes of showing sympathy and giving help to the suffering? For example, there is a possible world where there is so little suffering that most people can sit back in comfort in the sure knowledge that a small proportion of the world's population are taking care of all the people who are suffering. But this world is not ours. There is so much untreated suffering in our world that if the number of natural disasters was cut by, say, a half we could be confident that there would still be huge amounts of suffering that would go untreated. Consequently, the quantity of suffering in our world is too excessive to be explained by the beneficial effects of the exercise of human sympathy.

Additionally, Swinburne believes that the chance to show gratitude justifies some severe suffering. Once again, we can imagine less harmful ways this good could have been realised. For example, it is not clear that the gratitude I show to someone who gives up their spare time to teach me a new skill is qualitatively less significant than the gratitude I show to a doctor who nurses me back to health from my painful injuries. Even if the latter gratitude is qualitatively more significant, this is a contingent fact of human nature and could have been arranged otherwise by an omnipotent being.

Although the relief from pain cannot be realised by any way other than experiencing pain, we can wonder why our world is not better constructed to further this good. Too many diseases effect humans in way that causes unremitting pain. We might ask why God did not avoid these terminal diseases and replace them with diseases that cause pain but regress and allow humans to fulfil their desire to experience relief from their suffering.

On (b) we find that although it is not essential that humans are aware of the reasons for natural and moral evil, it is a mystery as to why they have not been informed more clearly what the correct reasons are. Once again, the explanation for evil could have been spelt-out to humans through dreams, religious experiences, or holy texts. For example, AE could have been unequivocally settled by having the correct theodicy written in the Bible. Furthermore, on Swinburne's solution, there are no reasons for us not to have been informed what the correct theodicy is. There are several reasons for thinking that if God exists, He would have resolved the problem of evil for us. First, it would refute AE. It seems that theists of the present are forced to admit that atheists of the past had rational grounds for their beliefs. There can be no doubt that theodicies of the past were less sophisticated than those of the present -- all of them to date have failed. A person who put forward AE in defence of atheism hundreds of years ago would have been entirely justified relative to the state of religious knowledge at that time. This has serious ethical consequences for a supposedly all-good God. Second, the mystery of evil adversely affects human relations with God. It is hard to form a loving relationship with someone who is able to prevent suffering but fails to and yet gives no clear explanation. This problem faced all theists of the past. Third, the mystery of suffering has an adverse effect on missionary work. To convert people, it is important that missionaries are able to say that they have some reasonable understanding of God's will. If missionaries of the past were unable to explain suffering -- one of the most basic problems in people's lives -- then what confidence could some potential believers have had that the missionaries were working on behalf of God?

By referring to point (c) we notice that the present theodicy has no explanation of non-human animal suffering. However, Swinburne separately addresses this issue by stating other goods that are satisfied by the suffering of non-human creatures. Using William Rowe's example of a fawn trapped in a fire which slowly dies, Swinburne argues that the fawn shows great courage. Its suffering may be witnessed by other deer who learn from the incident, hence its death is not in vain. The suffering of non-human animals, according to Swinburne, satisfies these greater justifying goods.

We can immediately say that Swinburne's argument that evil is needed for non-human creatures to obtain knowledge is inadequate. There are other ways that deer could discover the dangers of fire, including God giving deer the required knowledge at birth[14]. Perhaps Swinburne's argument is that the act of learning is what is valuable, rather than the state of having knowledge. However, if fawns were created with greater intelligence, then they could learn with less suffering. For example, a highly intelligent fawn could observe a fire raging through a forest, feel the heat on its body, view the smoke and then conclude that fire should be avoided. Therefore, the ability to learn cannot explain the suffering of the fawn. This leaves Swinburne's defence of the fawn's suffering resting entirely on its demonstration of courage. Firstly it is not clear that deer show courage in the sense we understand the term. Second, Swinburne's explanation has dubious ethical consequences, as we would not normally accept that torturing a sentient being to death would be justified by the courage that the being would show in the process.

Obviously, Swinburne's theodicy does not work in every conceivable case. There appears to be no possible good that is fulfilled by (d). The child is too young to have his courage built up by a natural disaster. In any case, the landslide may strike at a rate that kills the child almost instantly. Furthermore, there are no witnesses or mourners. There seem to be no explanations open to Swinburne here.

On (e) we see that evil causes some geographical and temporal problems. We can raise all the same questions as before. Are we to suppose that the inhabitants of Bangladesh need to show more courage than most other humans? If not, we might wonder why natural evil is not spread evenly around the planet. Also, we may note that medical knowledge is improving with time. From this fact, it is clear that a given natural disaster in the past would have resulted in more suffering than the occurrence of the same natural disaster now. If we assume that people living in the past did not need to show more courage than their present-day counterparts, we may ask why the frequency or severity of natural disasters was not significantly lower in the past than at present. The fact that the there are such geographical and temporal variations in intense suffering makes nonsense of the idea that there is a 'correct' amount of suffering due to natural evil which God has allowed for the purposes of developing human sympathy and courage.

At question (f) the theodicy has doubtful ethical consequences. From the discussion directly above, the theist may reject the notion of an optimum quantity of evil for producing justifying goods. However, this causes even more problems. If we were to increase the amount of suffering in the world, then we would be increasing the total amount of sympathy and courage shown. As these are great goods, it seems that we would be furthering God's intentions. For example, if scientists found a way to alter the frequency of earthquakes, we would normally say that the scientists had a moral obligation to prevent future earthquakes, or at least lower the frequency to prevent harm to humans. According to the theodicy, we would be opposing God's intentions if we were to prevent future earthquakes. God set up the required natural laws to allow earthquakes to further the goods of sympathy and courage through human suffering. The theodicy seems to suggest that the scientists should either leave things as they are, or perhaps increase the frequency of earthquakes in order to promote higher levels of sympathy and courage. However, these conclusions are ethically dubious. If this reasoning is rejected, the theist must explain why it is good for humans to reduce the frequency of natural disasters, but not good for God to do precisely the same thing.

It is safe to conclude that Swinburne's theodicy faces serious problems.

The Moral Argument

With these criticisms of theodicies now formulated, I will move on to a more general criticism of AE. Although professional philosophers of religion would not utilise this criticism, popular apologists commonly use it. For an eloquent statement of the argument we can turn to Ravi Zacharias, in his book "Can Man Live Without God?"[15]. To explain his point Zacharias relates how he was giving a lecture at the University of Nottingham in England.


"As soon as I finished one of my lectures, a student shot up from his seat and blurted out rather angrily, "There is too much evil in this world; therefore, there cannot be a God." I asked him to remain standing and answer a few questions for me. I said, "If there is such a thing as evil, aren't you assuming there is such a thing as good?" He paused, reflected, and said, "I guess so." "If there is such a thing as good," I countered, you must affirm a moral law on the basis of which to differentiate between good and evil." ... "When you say there is evil, aren't you admitting there is good? When you accept the existence of goodness, you must affirm a moral law on the basis of which to differentiate between good and evil. But when admit to a moral law, you must posit a moral lawgiver.... For if there is no moral lawgiver, there is no moral law. If there is no moral law, there is no good. If there is no good, there is no evil. What then is your question?""[16]

Furthermore, in his earlier book, Zacharias argued that only God can provide any meaning to life or the foundations to moral absolutes[17].

To make clear what Zacharias is saying, I will restate one form of AE.


(1) Gratuitous evils probably exist


(2) Gratuitous evils are incompatible with the God of theism (omnipotent, omniscient, all-good)


(3) Therefore the God of theism probably does not exist

Zacharias holds that


(1) Gratuitous evils probably exist



(1a) God exists

But Zacharias gives little support for his thesis. His only argument seems to be that a moral law requires a moral lawgiver. However, this is not a self-evident truth, and moral philosophers have constructed numerous ethical theories that are both objective and absolute and yet posit no lawgiver. Moreover, ethical theories that are radically dependent on God fall foul of the Euthyphro dilemma and have other serious problems. If the reader has any doubts about nontheistic accounts of ethics, it is worthwhile noting the severe problems with theistic accounts[18]. Furthermore, we may accept that a moral law requires a supernatural lawgiver, but deny that this need be God. We can imagine gods that are, finite, impersonal, not omniscient or not omnipotent which would satisfy Zacharias' requirements for being capable of constructing a moral law, yet these beings would not be 'God' in the theistic sense. This may well be more justified, since AE would not count against these beings.

The most damaging blow to Zacharias' argument is that we can simply reformulate AE to overcome all of his alleged problems[19].


(1'') Gratuitous suffering and pointless premature deaths probably occur


(2'') Gratuitous suffering and pointless premature deaths are incompatible with the God of theism (omnipotent, omniscient, all-good)


(3'') Therefore the God of theism probably does not exist

This argument does not assert anything about the moral status of gratuitous suffering and pointless premature deaths if God does not exist. The argument does assume that, if God exists, He would prevent gratuitous suffering and pointless premature deaths as much as possible. But even Zacharias would admit that, if God exists, there is a moral law that we can use to judge these things as morally wrong. This formulation of the argument does not depend on any nontheistic objective moral values, even though such values are possible. If the reader has some sympathy with Zacharias' views then he can feel free to substitute the word 'evil' with the phrase 'gratuitous suffering and pointless premature deaths' wherever I use it.

This reason why Zacharias' argument fails can be explained in another way. AE is a reductio ad absurdum argument. It claims that there is an inconsistency with the theistic hypothesis and certain facts about the world. What atheism has to say about morality is irrelevant as to whether theism is contradicted or made improbable by the fact that pointless suffering probably exists.

Defensive Scepticism

Given the failure of available theodicies, many professional philosophers of religion concede that they have no answer to AE. Instead they employ what might be called defensive scepticism or the Unknown Purpose Defence (UPD). Scepticism of this sort does have the appearance of a kind of evasive damage-limitation exercise, although the theists who accept it would, of course, deny this. Some of them dispute that the existence of evil has any lessening effect on the probability that God exists.

These theists argue that, in the case of evil, appearances are deceptive. They accept that apparently gratuitous evils do exist. However, they deny that we can infer from this that there probably are some gratuitous evils. It is argued that the move from 'apparent' to 'probably actual' in this case is unjustified.

Obviously, such a defence requires some support, because we seamlessly move from 'apparent' to 'probably actual' in everyday life. The support is thought to come from God's inherent mysteriousness. God, knowing vastly more than humans, is highly likely to know things that are beyond human comprehension. Stephen John Wykstra has argued that comparing human cognitive abilities with God's is in some ways analogous to comparing the cognitive abilities of a tiny infant with those of its parents. The child often fails to understand why its parents act in certain ways, just as we fail to understand why God acts in certain ways. Furthermore, the child would not be justified in supposing that the parents have no reasons for their actions, just as we cannot suppose that God has no good reasons for evil[20]. William Alston makes the same point a different way.


"Suppose I am confronted with the activity or the productions of a master in a field which I have no expertise... .I look at a theory of quantum phenomena and fail to see any reason for the author to draw the conclusion he draws. Does that entitle me to suppose he has no sufficient reason for his conclusions? It certainly doesn't if I lack the requisite expertise."[21]

In response to defensive scepticism, several lines of argument can be taken. First, it can be argued that if there were good reasons for evil that we do not know, God would very probably reassure us that He exists and indicate to us that evil has a good explanation beyond our ken. The mystery of evil adversely affects human relations with God. It is hard to form a loving relationship with someone who is able to prevent suffering but fails to and yet gives no explanation. Using Wykstra's child-parent analogy above, we might expect that if the loving parents were forced to allow their child to suffer somewhat, they would have a moral obligation to comfort the child and do everything within their powers to explain the reasons for its suffering. In addition, the mystery of suffering has an adverse effect on missionary work. To convert people, it is important that missionaries are able to say that they have some reasonable understanding of God's actions. If missionaries are unable to explain the reasons for great evils then they make their belief system seem implausible. Moreover, there are countless ways God could have told humans the solution to the problem of evil.

Second, it can be argued that scepticism based on God's omniscience cannot be used for the case of AE but then discarded when the theist sees fit. Bruce Russell compares the view that there are reasons, beyond our ken, for God to allow apparently gratuitous evils, with the view that there are reasons, beyond our ken, for God to make the universe appear much older than it really is.


"Is the view that there is a God who, for reasons beyond our ken, allows the suffering which appears pointless to us any different epistemically from the view the view that there is a God who created the universe 100 years ago and, for reasons beyond our ken, has deceived us into thinking it is older?"[22]

From this, Russell argues that because we are justified in supposing that there are probably no morally justifying reasons beyond our ken for God to make the world appear much older than it is, we are similarly justified in supposing that there are probably no morally justifying reasons beyond our ken for God allowing great evils.

In his paper, Paul Draper wonders whether sceptical theists are as consistent as they should be:


"... it is consistent with my response to skeptical theism in this paper to grant that the skeptical theist is not mistaken that we should not be skeptical about claims that if God exists, then he would be likely to have certain goals and to perform certain actions to accomplish those goals. But many arguments for theism rely on such claims. For example, many rely on claims that God would be likely to create a world which contains order or beauty or conscious beings. Others rely on claims that God would be likely to want to reveal himself to us by speaking through prophets or performing miracles. Skeptical theists, if they are to be consistent, must treat such claims with skepticism."[23]


It can be reasonably claimed that the various evidential arguments from evil lower the probability that God exists. If the traditional theistic arguments cannot be rehabilitated and made specific to theism then disbelief in God is justified by the current objective evidence. First, the unsophisticated theodicies of the past utilised by some apologists are totally inadequate. The more ingenious attempts of Swinburne also have considerable problems. Second, the attempt of Zacharias to discredit AE rests on a double misunderstanding. AE does not need some nontheistic moral foundation to argue from (as noted above). Furthermore, nontheistic objective foundations of ethics are possible. Finally, the defensive scepticism at the cutting-edge of religious philosophy can be criticised in a number of ways.

In conclusion, although the degree to which the probability of God's existence is lowered by AE is still an open question, we can reasonably agree with Professor Russell that


"Because the hypotheses which are offered to save theism are unlikely on what we know, theism is defenseless against the evidential arguments from evil."[24][25]


[1] Alvin Plantinga (1988), "Epistemic Probability and Evil," Archivo di filosofia, volume 56. Reprinted in The Evidential Argument from Evil, (ed. Daniel Howard Snyder, Indianapolis: Indiana University Press, 1996).

[2] John Mackie (1955), "Evil and Omnipotence," Mind 64 (1955). Reprinted in The Problem of Evil, (ed. Marilyn McCord Adams and Robert Adams, New York: Oxford University Press, 1991).

[3] Peter van Inwagen (1991) "The Problem of Evil, the Problem of Air, and the Problem of Silence," Philosophical Perspectives, 5, Philosophy of Religion, ed. James E. Tomberlin. Reprinted in The Evidential Argument from Evil, (ed. Daniel Howard Snyder, Indianapolis: Indiana University Press, 1996).

[4] Alvin Plantinga (1988), "Epistemic Probability and Evil," Archivo di filosofia, volume 56. Reprinted in The Evidential Argument from Evil, (ed. Daniel Howard Snyder, Indianapolis: Indiana University Press, 1996), p. 70.

[5] Michael Martin, Atheism: a Philosophical Justification (Philadelphia: Temple University Press, 1991); John Mackie, The Miracle of Theism (Oxford: Clarendon University Press, 1981); or on the Secular Web's section on theistic arguments (URL:<>).

[6] Michael Martin, "The Gap in the Theistic Arguments," (URL:<>, 1997).

These gods are incompatible with theism, but they are compatible with atheism in the narrow sense. For example, some atheists merely disbelieve in the theistic God and hold no views on other gods.

[7] Because, unlike the God of theism, it could be plausibly be argued these beings would lack the desire to prevent evils, the knowledge of evils, and the ability to prevent evils.

[8] FWD is an attempted explanation of moral evils. Simply stated, it argues that God would be interfering with human free will if He acted to reduce moral evils. Furthermore, it is argued that human free will is a very great good. The main arguments against FWD concern whether theists have given an adequate account of human freedom that is compatible with FWD, whether there is a possible world where people always freely choose to do what is good, and whether -- given the available options -- human freedom is really worth the price. These issues are argued in much greater depth in Martin 1991; Mackie 1982; and Niclas Berggren, "Does the Free-Will Defense Constitute a Sound Theodicy?" (URL:< html>, 1996).

[9] For example, the official Roman Catholic records of miraculous occurrences at Lourdes show an indisputable decrease in the frequency of miracles with respect to time. Sceptics may feel that this is clear evidence that reports of miraculous cures are unreliable, and improvements in medical technology are gradually making miraculous explanations redundant. Here, and in the case of spectacular miracles in the Bible, the believer can only accept the reliability of miracle reports if he really believes that miracles were more frequent in the past. To dispute the claim that miracles are less frequent and spectacular nowadays relative to the past seems farfetched. Believers do not generally accept that miracles of the magnitude of, say, the resurrection of the dead and the feeding of the five thousand occur so frequently nowadays.

[10] Richard Swinburne, "Some Major Strands of Theodicy," The Evidential Argument from Evil, (ed. Daniel Howard Snyder, Indianapolis: Indiana University Press, 1996).

[11] Ibid., p. 38.

[12] Ibid., p. 40.

[13] Martin, p. 417.

[14] In fact, it is probably the case that Swinburne is incorrect here, as evolution is likely to have given deer a 'hard-wired' pre-programmed response of fear of fire.

[15] Ravi Zacharias, Can Man Live Without God (Dallas: Word Publishing, 1994).

[16] Ibid., pp. 182-184.

[17] Ravi Zacharias, AShattered Visage: The Real Face of Atheism (Grand Rapids: Baker Book House, 1993).

[18] See the Secular Web's section on the moral argument at <URL:>.

[19] This distinction is used by Ted Drange in his forthcoming book, Nonbelief and Evil: Two Atheological Arguments, (Buffalo: Prometheus Books, 1998).

[20] Stephen John Wykstra, "Rowe's Noseeum Arguments from Evil," The Evidential Argument from Evil (ed. Daniel Howard Snyder, Indianapolis: Indiana University Press, 1996).

[21] William Alston (1996), "Some Temporarily Final Thoughts on Evidential Arguments from Evil," The Evidential Argument from Evil, p.317 ed. Daniel Howard Snyder, Indiana University Press.

[22] Bruce Russell (1996), "Defenseless," The Evidential Argument from Evil, p.197 ed. Daniel Howard Snyder, Indiana University Press.

[23] Paul Draper (1996), "The Skeptical Theist," The Evidential Argument from Evil, p.188 ed. Daniel Howard Snyder, Indiana University Press.

[24] Bruce Russell (1996), "Defenseless," The Evidential Argument from Evil, p.204 ed. Daniel Howard Snyder, Indiana University Press.

[25] Kind thanks to Jeff Lowder and Ted Drange for suggestions in earlier drafts of this essay