5thSeptember

The Logic of Disbelief

Dawkins’ Scale

In his recent book, The God Delusion, Richard Dawkins divides religious belief into a seven-stage scale, with the first stage representing devout belief in God and the last representing total disbelief. Dawkins says that since no one can completely disprove the existence of God, no one can logically claim to be squarely in the seventh stage of disbelief. However, since Dawkins says that he finds the existence of God to be highly improbable, he claims that he fits at the far right of stage six.

My first reaction when reading this was to shake my head, because it seems laughably simple to disprove the existence of God, if by “God” one means a perfect creator of the universe. This is because, like the concept of a square circle or married bachelor, God’s various self-contradictory attributes cannot cohere. This seems to prove, in no uncertain terms, that God cannot exist.

I Was Wrong

But, does it? I thought so. And I hate being wrong, especially when I am unclear about why I am wrong. Yet, there is a problem with my logic, namely, that logic itself has been proven to be so limited that no formal system can be both consistent and complete. We have the mathematician Kurt Gödel to thank for this, um, proof.

In other words, I cannot conclude that God cannot exist, because I cannot prove it to be true in all cases, despite what seems trivially true to me. This is because language is built on top of logic and the last word on logic is that it is, as mentioned, incomplete.

Disclaimer: I understand that most people have not come to terms with Gödel’s results and that most, if not all, who read this and are even passingly familiar with Gödel’s work will be reluctant to assign relevance of those results to natural language. Nevertheless, to be clear, I am assuming here that all natural languages derive from underlying logical systems and that, in all likelihood, all logical systems derive from a single, universal grammar. This is essentially saying that all linguistic truths are mathematical truths.

Proof (Sort Of)

To show this, let us try to define God. We start by assigning God with the most general and abstract attribute available, that of perfection. Then, little by little, we expand this definition to include consequent attributes, like omniscience, omnipotence, and omnibenevolence.

We can immediately see, if we keep an open mind, that these three attributes clash with one another, since no all-knowledgeable, all-powerful, and all-loving creator would allow his creations to suffer or, for that matter, create an environment in which suffering is possible. These creations are, after all, corporeal extensions of their creator. And this is basically the root of the famous Problem of Evil, which, all by itself, has driven many theists from their once precious beliefs.

We have now ripped a mighty hole in the historically woven fabric of divinity, which seems to be more than enough to disprove God’s existence. However, remember what we are trying to do. We are trying to build a consistent and complete argument that disproves God’s existence. So, it must be constructed with a consistent and complete set of axioms or self-evident truths.

Unfortunately for us, if we plan to continue until we have completely defined God with all possible attributes, then, again, thanks to Gödel, we can expect to eventually run into an instance in which one axiom makes the sum total of all attributes internally consistent! At this point, we will have proven what we set out to disprove, namely, that God exists.

Conclusion

In effect, then, logic can prove or disprove anything we like and, at the same time, ultimately prove nothing at all. This means that logic can never be trusted to prove an allegedly irrefutable statement or argument. And in this case, specifically, it means that logic cannot be trusted to disprove God’s existence. It also follows that God’s existence cannot be proven to be impossible to disprove, but that only adds to the complexity of the problem (or any problem, for that matter).

7 Comments

Wild-Eyed Atheist Boy says 16th September @ 1:55

Have you read Hitchen’s “god is not Great” yet? Stunning. Absolutely stunning. I am into Sam Harris’ “The End of Faith” now, but “The God Delusion” is on my list.

Cheers,

WEAB

UK says 17th September @ 14:58

God Is Not Great on my list. Letter to a Christian Nation, also by Harris, is another to check out; you can read it in the bookstore.

Wild-Eyed Atheist Boy says 17th September @ 15:29

I heard “Letters” was a brief one…perhaps a hundred pages or so, more along the lines of a pamphlet. Looking forward to it, nonetheless. Enjoying Harris’ End of Faith so far. Would love to have more reading time, but between writing for my blog, arguing for “our side” on the religious forums and, well, the “rest of life…” Not enough hours in the day, I’m afraid.

Jim Lippard says 19th September @ 14:17

I’m unconvinced that Goedel’s results are applicable in the way you suggest. First, natural language is not consistent and not built upon logic in the way you suggest. Logic is a formal system designed to be unambiguous and consistent (or paraconsistent), natural language is not. Second, Goedel’s incompleteness theorem doesn’t apply to every form of logic, only to formal systems which include basic arithmetic. It doesn’t apply, for example, to propositional logic, which is provably consistent and complete. (This point is a weak one, as natural language is clearly sufficiently complex to include basic arithmetic.) Finally, Goedel’s result only says that a sufficiently powerful formal system is either inconsistent or incomplete–if it’s consistent, then there are truths expressible by the system that are not provable. It doesn’t follow that “God exists” is one of them or that “God does not exist” cannot be proved–indeed, showing that the concept of God is logically incoherent or that the existence of God entails some fact which is not the case are two ways that could be done.

I think part of your main conclusion can be drawn without reference to Goedel at all, because it’s not just logic that matters, it’s the truth of your premises. I don’t think it follows, though that we can’t trust logic. If we have true premises and sound inference rules, we can reliably infer truths, and Goedel’s theorems don’t change that. The inability to prove some truths doesn’t entail the inability to prove any truths.

UK says 19th September @ 14:50

Thanks for your feedback, Jim. I’ll think it over.

Robert-Jan Milleker says 25th October @ 1:48

A Web Site addressing some issues raised by Goedel’s Theorems:

http://www.geocities.com/robert.milleker

UK says 25th October @ 18:32

Thanks! Are you a mathematician or computer scientist?

Leave a comment.

You must login, to leave comments.