Francis Collins' new Templeton-sponsored website, BioLogos, has been getting a bit of news coverage recently. For those not familiar with the story: Collins is the Director of the Human Genome Project, and BioLogos is his attempt to persuade Christians that they don't need to be frightened of science.

I guess the guy should be applauded for trying to reign in the wackier elements of Christianity, and most of it is pretty innocuous, straightforward stuff. But there's some real intellectual howlers in there. Basically, the guy doesn't understand what science is.

At the heart of the site are some answers to big questions, and the two that really got on my goat are Question 13: What is a God-of-the-Gaps argument? and Question 4: What is the proper relationship between science and religion?

Collins (along with fellow travellers like John Polkinghorne) reckon that science provides us with "pointers to God", natural phenomena that imply the existence of a biblical God. One favourite example is fine tuning, which points out that the universe looks like it's tailor made for us, and we don't know why - and then infers that God must be behind it.

'God of the gaps' refers to the standard religious tactic that identifies a mystery and then 'explains' it by saying that it's the result of action by a magical being (i.e. a god). These kinds of arguments have a pretty sorry history - BioLogos gives planetary motion as one example of where the argument was used, only to fail miserably when science moved on. There are plenty more.

Despite their protestation, 'fine tuning' is a 'god of the gaps' argument. We don't know why the universe is the way it is. It's a mystery - for now at least. Arguing for the existence of God on this basis is exactly like arguing for gods on the grounds that that planets all move round in neat ellipses.

What's more, arguing that 'god did it' is an explicit scientific claim. BioLogos reckons that science can't disprove god because:

"God’s existence is not something that can be tested by the scientific method in the same way the existence of postulated new elementary particles are tested in supercolliders ... Rather than an empirical claim about nature or its laws, the claim that God exists is a metaphysical one, a statement about what there is, whether it be natural or supernatural."

But statements about why the universe is the way it is are empirical claims. And, despite BioLogos' misconceptions, the 'God did it' argument can be rejected on scientific grounds.

There is a misconception that science can only disprove theories, never prove them. This idea, originated in the early 20th century by Karl Popper, that you have theories that are neatly disproven by experiments, is outdated.

Modern concepts of science recognize that what actually happens is a gradual accumulation of evidence that tips the balance one way or the other. It frequently happens that a hypothesis is never actually disproved - it just becomes more and more implausible until it reaches the point that no-one is prepared to stand by it any more (it's a process called Bayesian Inference).

Now let's relate that to what Collins argues. We have a long history of claims made about the material world - whenever we see mystery, the religious claim it as a 'pointer' to god.

Time after time they have been proved wrong. It's now very clear that humans have a propensity to mistakenly give supernatural explanations for perfectly natural (though complex) things.

Each time a 'god of the gaps' argument is demolished, the Bayesian meter of improbability swings against these kinds of claims. Now God has been pushed back to the very margins of existence. And the religious are still claiming that where there is mystery, there is a 'pointer to god'.

We have to make a judgement. Is this claim plausible, given what we know about the history of these kinds of claims. Can we ever disprove god? It depends on your perspective. There will always be mysteries.

But from a Bayesian, probabalistic, scientific perspective, God is well and truly disproved.

Originally posted at: Epiphenom

Creative Commons LicenseThis work by Tom Rees is licensed under a Creative Commons Attribution-Share Alike 2.0 UK: England&Wales License.