Problem of evil
Suppose we have the following four premises:David Hume wrote, (paraphrasing Epicurus):
"Is He willing to prevent evil, but not able? Then He is impotent. Is He able, but not willing? Then He is malevolent. Is He both able and willing? Whence then is evil?"
- — Dialogues Concerning Natural Religion
We get the following contradiction. If God is omnibenevolent, then He does not want evil to exist. If God is omniscient, then He must know about all evil in the world. If God is omnipotent, then He must be capable of doing something about it. Therefore, evil should not exist. Dropping any one of those four premises would resolve the contradiction, but dropping #4 would require us to fundamentally redefine evil in some way, and dropping the other three would undermine the Christian concept of God.
It is often claimed that evil exists because God gave humans free will. According to the Bible, God's gift of free will led to the fall of Adam and Eve through their original sin. Free will is assumed to be a greater good than the evil that it causes.
The concept of omniscience is fundamentally at odds with free will, at least in any sense that would be relevant to this argument. The assertion that it is possible to be all-knowing necessarily implies some form of determinism, because without determinism, there are things (notably future events) that are unknowable. This means that omniscience is not possible if non-deterministic free will exists. This leaves only a compatibilist view of free will in which choices are made freely, but an omniscient being would know the outcome of those decisions ahead of time, as well as the new choices that result and their outcomes, ad infinitum. In this way, an omniscient creator, knowing all of the repercussions of the placement of every particle of matter or unit of energy before placing it, can be considered to be directly and intentionally responsible for every event that follows. A benevolent creator could just as easily have set events in motion in such a way that free will would never result in evil.
This argument also fails to explain why God allows natural disasters, such as hurricanes, tsunamis, and earthquakes. These events kill large numbers of people in specific geographical locations, which indicates that the concept of "evil" is not necessarily tied to what people do.
Even if we define natural disasters as not being evil, there remains the fact that they occur, and that God does not prevent them or the deaths and suffering they cause. If we replace "evil" with "suffering" in the discussion above, the problem remains: either God is unaware of people's suffering, and is therefore not omniscient; or he is unable to do anything, and is therefore not omnipotent; or he is unwilling to intervene, and is therefore not omnibenevolent.
We're also left wondering why animals suffer needlessly. Animals were suffering long before mankind, Adam and Eve, and original sin came into the picture.
There is also the question of heaven. Heaven, being a perfectly wonderful place, does not contain evil. Does this mean that inhabitants of heaven no longer retain their free will? Or does their will suddenly become perfectly good?
Apologists often claim that what appears to be harmful to humans may, in fact, be for humanity's good. How can we learn, the argument goes, without making our own mistakes?
The 'tough love' argument only works if God is limited in power. If God is omnipotent there is nothing he can not teach us gently that he can teach us harshly. If he is benevolent than he would never choose to teach us a harsh lesson when it could be taught, with exactly the same impact, gently.
Another problem with this argument is that although according to this argument, God wants us to grow as people by learning from our mistakes, according to most religious doctrine he also wants worship. Worship involves complete obedience and submission, wheras learning from mistakes requires using one's intelligence. It is contradictory to claim that God wants us to be both completely obedient and make decisions for ourselves, since complete obedience means blindly obeying authority, for example the story of Abraham and Issac.
Best Of All Possible Worlds
Many theists make the claim that although there may be some evil in the world, this is in fact the best of all possible worlds.
The problem with this assertion is that there is no way to demonstrate that this is the best of all possible worlds, simply because we cannot know what other worlds can exist. This argument has to be taken on faith rather than evidence.
One way to redefine the term 'benevolence' is to cite limited human perspective in space and time. A parent might spank a child for running into traffic, or take a child to the doctor for painful, life saving, injections. It is only in the limited, child's-eye-view that these things are malevolent.
As with the "tough love" argument, this view of God implicitly denies his omnipotence or, at least, his omnisciences. What kind of parent purposely takes his child for a surgery which he knows the child does not need or want?
Another way to redefine 'benevolence' is to argue that God may be benevolent to specific humans or to non-humans. Our entire history may exist for the positive influence it may have on aliens we have not met. We may be actors in a puppet show that makes these beings happy. After all, it is perfectly possible for benevolent humans to play comically violent video games with their delighted children.
But this argument is sophistry. To win the argument the apologist defines a God that neither we nor he would have much reason to worship. For example, if the creatures in a violent "Run and Gun" video game were to gain self awareness, would we expect them to view us as benevolent beings worthy of their love and trust as we blast them into electronic oblivion?
And, if God is not benevolent toward humans, then what differentiates him from a human sociopath or from the Devil?
The problem of evil must be taken up in the context of humanity. No other context would make a God useful to humans in any realistic way. A God that is benevolent to others at lethal expense to humans is, by definition, malevolent, or at least indifferent, toward humans. It is an unusual apologist indeed who believes in this type of God.
As with "benevolence", "evil" can be redefined. What is "evil" for humans may not be evil for God. In fact, anything that God chooses to do can be construed as "good". Using this argument, "evil" can not exist in any definable terms when applied to God.
Besides the arguments already used in the "Tough Love" response (an all powerful God would have no reason even to appear evil) here, the apologist treads dangerously close to ethical relativism. We know from information in the bible that ethical rules have changed at the will of God. Is God, then, a relativist?
If the answer is that God is following an ethical plan, then the apologist opens himself up to the Euthyphro dilemma. If the answer is that God changes as he sees fit and anything that god declares as good is good, then what is the difference between being a relativist and following a relativist God?
When an apologist tries to redefine the premises of "the problem of evil" he finds himself in a morass of relativism, but when he tries to work with the premises he finds himself unwittingly limiting the unlimited God of his religion.