"If you admit the supernatural into your calculations, anything goes. That is why a supernatural explanation is useless to a scientist, however pious he may be on Sundays. It provides no direction for research, suggests no testable hypotheses, and gives no reason to expect one result rather than another from any observation or experiment."