(The idea and name for this post came from Steven Novella on episode 123 of the Skeptics’ Guide to the Universe)
For the past four centuries or so the human race has been conducting an experiment of sorts, though it has not been consciously instigated and so might more properly be thought of as a meta-experiment (an overarching set of many smaller experiments). In has a hypothesis, it makes predictions, and it has control groups. It is the experiment of whether science works, whether, as Carl Sagan put it, it really delivers the goods. The control groups are the attempts at knowledge that do not adhere to the scientific method.
At issue is whether methodological naturalism (that is, the rejection by necessity of supernatural explanations) can in fact reveal truthful knowledge about the universe. It is often claimed, incorrectly, that science requires faith that the universe obeys natural laws, and so is in essence the same as other forms of faith. This is wrong because “science” does not take philosophical naturalism as its premise. Science doesn’t really take anything as a premise; its simply a methodology for testing claims. It is methodologically naturalistic in that claims about the natural world are, by its very construction, that only kinds of questions that science can investigate. This is something that creationists, for example, just don’t get: It’s not that supernatural explanations are rejected out of hand from results obtained by science, it’s that the questions that can be subjected to scientific investigation preclude the supernatural because science has as one of its pillars the notion of falsifiability. Any supernatural entity that could “rig the game” as it were would cause the paralysis of science, since no experiment could be conducted in a scientific way. Since science relies on the control of variables, and the supernatural are necessarily not amenable to control, science becomes meaningless.
So what has been the result of this meta-experiment? Has there been a the expected paralysis of science that would result from a world undergoing supernatural intervention? There has been no credible evidence thus far. Science has delivered the good for hundreds of years now, and has probed deeper and further with more logically satisfying answers than any construction of the human imaginations. This isn’t to say that science has proved naturalism; there are no absolutes in science. There could always still be demons haunting the woodwork. Were science to fail categorically, we might consider that meta-evidence of the supernatural. Since the hypothesis of the meta-experiment is that the universe obeys natural laws that can be investigated scientifically, the null hypothesis would be that the universe does not obey natural laws or that those naturals laws can be violated freely by some unknown agency. But the evidence has now been stacking up for centuries: science works, so we can tentatively assume as a consequent that the universe is naturalistic (crucially, this is a result and not a premise of science).