Several months ago, I began going through some of the basic arguments for the existence of God. I prepared a couple of them:
The Cosmological Argument for God
The Teleological Argument for God
The Entropic Argument for God
Today I am going to return to this, adding what C. S. Lewis believed was the most conclusive response to atheism (if not a proof of God) -- the Mechanical Argument.
The basic premise is this. Imagine a powerful supercomputer, with more computing resources than all the computers on the planet combined. Imagine this supercomputer has the capacity to draw inferences, made and test hypotheses, and come to conclusions which are then able to be used to create more complex and accurate hypotheses. Eventually this computer would come to a full (or fuller) understanding of the nature of reality the longer this computer was kept running.
But let us imagine, say, that because of a simple mechanical flaw in the system this computer had a basic mathematical error. Whenever this computer divided a number by two, it came up with a result one greater than it should have. So if the supercomputer attempted to divide eight by two, it would come up with five instead of four. If it tried to divide 32 by two, it would come up with 17. And so forth.
Now this mechanical error would not make reality any less predictable for the supercomputer. The computer could still make accurate predictions for the world around it, test those hypotheses, and accept or reject them based upon the data. But the view of this computer as to the very nature of reality would be fundamentally skewed. It would necessarily need to develop complex rules to explain why the world functions the way that it functions (why two divided by one is two and two divided by two is also two), but it would establish its own version of the transitive property of mathematics that would justify the results that it achieved.
In fact, over time this computer could probably get just about everything in the world predictable, rational, and be able to establish and perfect any technology that we humans have crafted. The one thing that it couldn't do would be to evaluate its own mechanical structure to see where the error is, and correct it.
Let's leave computers at this point and move on to us. Here is the problem: if our brains are nothing more than mechanical computers (however effective), then the results that our brains produce can be considered nothing more than arbitrary. We cannot think our way to atheism (particularly reductive materialistic atheism), because if the conclusion of atheism is correct then the mechanism that justifies that conclusion (our minds) are merely arbitrary mechanical constructs incapable of self-evaluation. Like the computer, if we cannot properly perform a function because of our mechanical defects, we can form a predictability around such a defect that will be effective, allow us to develop technology, and ultimately be wrong.
Our minds only have value and trustworthiness insofar as they are derived from something trustworthy. If we are the accidental happenstance resulting from eons of evolution, then we cannot trust our minds because if we are wrong we cannot know it. If, on the other hand, our minds are fashioned by God (through whatever means) and God is perfect (including perfectly reasonable) we can trust that our minds can lead us to an understanding of reality that is not absurd.
Thus using reason to understand reality is a reasonable activity for the theist and yet an absurd activity for the atheist.