I don't either. I merely say they are never assumed by science. The default explanation will always be naturalistic, and if explanation eludes science, then the default hypothesis will be naturalistic.
There really isn't any other way to conduct science. If you don't imagine naturalistic explanations, you have nothing to test. If you assume supernatural explanations, you have no reason to test.
Arguing from structure demonstrates nothing in the absense of history. The art of science is to imagine a naturalistic history, then test each piece of the history to see if it happens without intervention. You get no proof, but you get plausibility.