Geoffrey Berg's The Six Ways of Atheism is one of the most important books about Atheism that I have ever read. Berg provides half a dozen arguments to prove that God does not exist or at least that God's existence is improbable. Although all of them may not be valid, they are all intriguing and deserve to have a hearing. My discussion of them that follows must not be treated as a substitute for reading his book, as he goes out of his way to address many possible criticisms that might be launched against his arguments.
Berg's first argument from atheism goes something like this: God is a being who is omnipotent, the creator of the universe, supremely good, and so on. But it must be statistically improbable for the creator of the universe to also possess attributes like supreme goodness which must occur in very few (if any) beings. To see this argument with more clarity, try thinking about it this way: suppose that tomorrow one person will win the lottery and one person will be struck by lightning. What are the odds that the lottery winner will be one and the same as the person who gets struck by lightning? Obviously, the odds are incredibly low that two improbable occurances will happen to the same person. Likewise, the odds that a rare (or perhaps even non-existent) attribute like supreme goodness will happen to exist in the unique being who created the universe is also improbable.
I think that Berg's argument is a valid one, but it only goes to show that the existence of God is a priori improbable. Of course, with enough a posteriori support, the existence of God might be deemed probable in spite of its a priori improbability. Fortunately Berg seems to realize this and he attempts to show that the God hypothesis can recieve no a posteriori support in another argument, which we will examine next.
Berg's second argument is the 'Comprehension Gulf' Argument. Berg asserts that since God is, by definition, immortal and infinite, that man can never recognize God and prove that God exists because man is finite and mortal. Man, being finite in both lifespan and knowledge, can never attain direct proof that something has existed for eternity. Therefore, we can never know if some being actually is God. At best we could only attain proof that some very powerful being exists who might be God, but who also might be simply a very powerful and long-lived spirit (who did not quite meet the definition of 'God').
Now this is a very interesting argument. If it is valid, it would go along very nicely with Berg's a priori proof that God is highly improbable because it effectively destroys any hope the theist might have of supporting the God hypothesis through a posteriori means. But is it valid? I don't know. If it were the case that all of humanity met some spiritual being who could stop bombs from exploding, make pink unicorns appear out of thin air, rearrange the planets of the solar system, and cause my little sister to be a little less stubborn, would we be irrational for reasoning inductively that this being was all-powerful?
The problem is analogous to another: if you have ever heard of Olber's paradox, then you will understand that, if the universe were infinite, we can predict that there would be an infinite number of stars, and thus the sky ought to glow at night because of the infinite starlight. Since the sky does not glow, we may take this as a falsification of the infinite universe theory. But what if the sky did glow? Would that mean that the universe was infinite? Not necessarily. What if the number of stars was not infinite, but just extremely large? How would we know the difference? On empirical grounds we would not know. But on logical grounds we might be able to make a decision. Richard Swinburne advocates the position that the simplest numbers are zero and infinity. Since zero represents no being at all, it follows that the simplest being we could come up with would be an infinite one. I believe that Swinburne was on to something here.
Most readers of philosophy are aware of the age-old problem of induction: inductive reasoning (or generalizing from a limited number of experiences) is not self justifiying, and it certainly is not logically necessary that what we generalize from a limited number of our experiences will hold true in all cases. For instance, if every swan that I have ever seen is white, it does not necessarily follow that all swans are white. Likewise, just because I have always observed the sun rising from the east does not necessarily mean that the sun didn't rise in the west ten million years ago or a hundred million years ago or five hundred million years ago. So why is it that inductive reasoning is valid? I believe that the answer lies in Occam's razor: the principle which states that the simplest explanation is most probably correct. Since it is simpler for the sun to rise from same direction throughout the history of sunrises than for it to rise from one direction for half of its history and another direction in another half, we can rightly say that the simplest (and therefore most probably correct) hypothesis is that the sun always rose in the east.
This piece of philosophy is completely applicable to Berg's argument. Suppose that there was a God who was only too willing to prove to us that he was omnipotent, and the God did all kinds of fantastic things like rearranging the stars, raising the dead, etc. After numerous demonstrations of power it would be simpler to suppose that, as had always happened before, the God's power would prevail over anything and everything, and therefore this God would be all-powerful.
Berg's third argument against the existence of God is the 'God has no explanatory value' argument. The argument has a very familiar feel to it. All too often I am asked, "If there is no God, why is there something rather than nothing?" I often respond by asking "Why is there a God rather than nothing?" Berg has taken this reply, which is applicable to so many theistic arguments, and actually used it as an argument against God's existence. Berg argues that if we are positing a God in order to explain questions that the God hypothesis does not really answer, then we ought to apply Occam's razor (the simplest explanation is most probably correct) and avoid postulating God in the first place. Theoretical entities should not be postulated when they explain nothing, so let us avoid them and arrive at the simpler (and more probably correct) explanation: atheism. I agree with this line of reasoning completely.
Berg's fourth argument is a brilliant variation on the age-old argument from evil. If we define God as an all-powerful and all-good creator of the universe, then it follows quite naturally that such a being would be expected to create this world in its best possible state. This world must be the best of all possible of worlds if God created it. If it is not the best of all possible worlds, then God did not create it. Berg observes that the world has improved quite a bit over the ages. After all, few of us would care to live in the Middle Ages or in the iron grip of the Roman empire. If I may bring up a quick point, I remember PBS airing a series called The 1900 House, a reality series in which a family lived in a house with conditions like those of the beginning of the twentieth century. After watching that series, I am very grateful that I was not born in that time period. And the reality series did not even fully recreate what that era was like, since many things had to be altered slightly from the way they actually were back then due to health and safety concerns. Anyway, back to the argument: Berg observes that, since the world has been improving, it follows that the world was not created in the best possible state (by definition you cannot improve something that is perfect). Therefore God does not exist.
Of course, there are some theodicies that might pose some difficulty for this argument: for instance, Richard Swinburne might counter that God did not create the world in the best possible state because he wanted to give man the responsiblity and freedom of choice to improve it (Swinburne sees these things as being good in and of themselves). Then again, such a defense seems to presuppose a libertarian or contra-causal account of free will, which I personally have chosen to reject because I view it as incoherent and unintelligible (see Daniel Dennett,Freedom Evolves, and Richard Carrier, Sense & Goodness Without God, Section III.4).
Overall, I agree with Berg's 'Best Possible World' argument, but I believe that he should make a more thorough and exhaustive attempt to defeat threatening theodicies which may, in the eyes of some, nullify his argument.
Berg's fifth argument is the 'Universal Uncertainty' argument. He argues that God must have certain knowledge of anything (if it is to be called 'God'), but that there some to be some things that God could never be certain about. How would God know if he is not simply one of many Gods who exist separately from one another and unaware of one another? However, I think a theist could suggest that perhaps God has some self-evident, a priori proof to answer these questions. Although it seems unparsimonious to postulate the existence of unknown a priori proven answers to difficult questions like the one mentioned previously, I don't see any reason to suppose it is impossible, and therefore I don't find this argument too convincing.
Berg's final argument does not feel like a single argument as much as it does a few left-over arguments slapped together to make a sixth chapter. He argues that God cannot exist because some of God's defining qualities cannot exist. Berg moves through several qualities, from his previously discussed argument against omniscience to the old argument against omnipotence (As J.L. Mackie put it: "Can God beings that he cannot control?"). He even suggests that God could not be anything more than a "grand showman" if God is not capable of giving meaning to human life (which Berg argues is impossible). It seems to me that he does not argue this very well, and does not justify it very well on logical ground. Moreover, I think Mackie's argument against omnipotetence is not very convincing: an omnipotent being could indeed make beings that he could not control. Prior to creating the uncontrollable beings the God would be omnipotent; after creating the uncontrollable beings he would not be.
Overall, Berg's book is very thought-provoking, original, and convincing. I highly recommend it to philosophers of the amateur and professional stripes.
You can purchase it here.