There is a rather delicious irony floating around out there these days, and I'm a little surprised that nobody has called attention to it yet.
Sir Karl Popper was a british empiricist who did seminal work in the foundations of the philosophy of science. Among his contributions to the demarcation debate was the suggestion that the main criterion for counting something as science ought to be the criterion of falsifiability. That is, any hypothesis that is not susceptible of falsification is not a scientific hypothesis. He gave two rather interesting examples. One was Marxist economic theory, the other Freudian analysis. Marxist economic theory is not falsifiable because it is not, in principle, testable. Freudian analysis is not falsifiable because any possible falsification one can imagine for a given hypothesis can be explained away as actually being fully consistent with the theory. For example, take the Oedipal Complex. If you think that asking grown men whether they had the hots for their moms when they were kids might be a good test of the theory, think again: if they say yes, of course that confirms the theory; if they say no, then the Freudians say that they are either lying or just repressing the memory, and again the theory is confirmed.
This criterian is often applied to creationism, because it fails on both counts. Since it is a claim about a singular historical event--God's special act of creation at the beginning of time--then it is not testable even in principle, since nothing about it falls under anything like a natural law. But in addition to that, any possible falsification you can try to think of will be answered by the creationists as fully compatible with their theory. I had a friend when I was in graduate school with whom I would often go jogging, and we often talked about this issue because he was himself a creationist. One day, while we were jogging, I asked him about the fossil evidence. It's millions of years old, I said, so how could the earth be only a few thousand years old. His answer? God put those "fossils" there, and made them appear to us to be millions of years old, when in fact they are only a few thousand years old. Why on earth would God do that, I asked. To test our faith, he answered.
I have no sympathy with those who claim that creationism is science, or with those who think it ought to be taught alongside evolutionary theory in the science classroom. On the other hand, I do have sympathy for those who object to the hubristic and condescending attitude of the acolytes of scientism such as Richard Dawkins and Daniel Dennett. Dawkins, of course, has just released a new book, The God Delusion, in which he gives what amounts to a kind of just-so story about the origins and foundations of religious belief: it is a vestige of a phenotype that is in fact valuable (neonatal trust in a parental figure) but that has the effect of continuing our magical thinking well into adulthood.
Needless to say, just-so theories of this sort are not falsifiable, since they are not testable; they are merely abductions, inferences to the "best" explanation, where "best" usually means "most appealing to the researcher". Also, like the Freudian complex, any possible falsification one can think of can be explained away as a vestige of the very phenotype one is trying to falsify. So the irony here is that the so-called "God Delusion" fails to be science and, most deliciously, it fails to be science on precisely the same criterion by which creationism fails to be science. Dawkins, Dennett, and the rest are on the same intellectual level as creationists. Given the sorts of terms Dawkins uses to describe such people ("stupid", "morons", "imbeciles"), I hope he's got his falsifications ready.