But if they could be detected, then they would be natural. The problem is, if something is supernatural, it doesn't have to obey natural laws, and that means that scientists can't propose a way of testing its existence because such testing itself presumes natural law.I'm prepared to be less of a naturalist than Chris and Matt Yglesias, but less of a supernatural inclusivist than Julian Sanchez. I think that there are ways that the supernatural can play into science, I just think there are limits.
Think about intercessory prayer. The idea is that when other people secretly pray for you, you get better faster. That's a testable prediction. It's a supernatural hypothesis if it's true, but it's wholly testable. If psychics were legitimate, that would be a supernatural force which was scientific. We could predict that a true psychic could predict what card you chose more (or less) often than expected by chance, and do so repeatedly. If it happened, it would be scientific, even though it's supernatural.
The hallmark of science isn't naturalism, that's just a practical limitation which derives from the real issue. The real issue is testability. If you can generate a testable (falsifiable) prediction, it's science.
What does that say about creationism? If creationists could make predictions (not post hoc explanations) about the natural world, that would be impressive. Without specifiying anything about the creator/designer's motives and abilities, you get stuck at the problem of evil before you get to any prediction. We know that the world is imperfect, so if the creator/designer is omnipotent and omniscient, we have to explain not only the good design, but the bad design.
We all know that natural processes are flawed, so imperfection in life is to be expected. If things were too perfectly adapted, that would be the surprise. If creationists can build a model of the supernatural creator, and use it to make predictions, that's science. So far, all they offer is that other things are unlikely, which doesn't explain why it should be likely if the Creator did it.
That's why it's bad science.
Update: The Discovery Institute noticed this. Note that I don't think ID is science yet. If they would get over it and say something meaningful about the designer, we could test it and falsify it, they way we can test and falsify claims of intercessory prayer and psychic phenomena.
Most importantly, I never said "Sure, design could be “scientific,” if the design people would get their act together and make some predictions."
Would it have been so hard to quote me accurately?