Not sure, but I think empitical thought has become so important that some varient has to be used in this "reasoning" we use.
I guess one method would be along the lines of such (and coincidently, leads to empricism) that natural events have natural causes, and man can know them.
This would reason that if man can identify an occurance that happens frequently, it should be considered "natural" and thus it's origins can be studied. The workings of nature would mean that "magic (the unexplained) could not be an answer, if the answer can be understood (or assumed to be undertood) then "magic" is not the answer.
Dialetical thinking would be my "short answer."
So I think that we have come to some agreement. We cannot really be totally sure that "magic" is not the "true" answer. However, we assume that natural phenomena have natural explanations. That's what I meant when I talked about science seeking useful explanations rather than true ones. Given a sufficiently powerful magician, any observation can be said to arise from "magic". Hence this isn't a very useful explanation. However, substitute God, or some intelligent designer for the magician and the result is identical. Unless we develop some definitive test for design, the ID explanation for the development of the diversity of life is not a particularly useful one scientifically. Of course, it still makes for a terrific philosophical debate, which is where I believe students should be exposed to ID, rather than teaching ID as a scientific idea.