Magic Science
This is a short story on unexplainable phenomena.
Suppose that on a walk home, you run across a bona fide wizard from the Harry Potter books. Turns out J. K. Rowling did not write a fiction but a documentary. You now take the proposition "magic is real" as true because you observe him speak Latin-esque words, wave a stick of wood at a bush, and it catches on fire.
But hold up. What do you know for sure? At least three things: (1) A man spoke Latin-esque words, (2) The man waved a stick of wood in some pattern, and (3) A bush appeared to catch fire through no cause you can yet explain. So that's three things. You do not need to tack on a fourth thing: (4) "magic is real" when the three are sufficient to describe what happened. What does the fourth mean anyway? It refers to an aspect of your internal map of the territory, not an aspect of the territory itself.
Now, how to trace the cause of the bush catching on fire? Hold off on proposing solutions, there's a million possible hypotheses. And not to rain on the parade of "magic is real", but these three words actually wouldn't and couldn't be one of the hypotheses, because it doesn't propose to explain why the bush caught on fire. Neither would "magic did it", because it doesn't say anything about how. It's like explaining why some apples are green with "my dog made them green". Okay, but how did she make them green? Did she eat a bucket of green paint and vomit on all those apples? When did she do that? How does she find the time to do so on all apples in the world? What color were apples before they were made green in this way? Once we have a proposed mechanism, we have a hypothesis.
You ask the man to see if he knows something. Certainly looked a lot like he intended for the fire to appear. He answers "it's magic – it's fundamentally unexplainable".
Well, we'll see about that…
Fast forward a bit. As it happens, the wizarding world cancels its Statute of Secrecy. The news about the magical society spreads like wildfire. Within a mere day, some of the faculty at your local university have already set up a Department of Magic Studies. (Where do you live, dear reader? It has a university? That's the one. Would you be surprised?)
The academics start describing magic. Not explaining, just describing – getting a feel for what it can be used for, its typical and atypical traits. They catalogue how many spells have Latin words and how many have Old Welsh words, and look at Hogwarts history books to see if the Welsh and Latin spells appeared at particular periods of history, which would imply that people somewhere invented the spells at the time those languages were in use…
They start creating hypotheses like that, for the "why" of all the patterns they've found. This moves us from the descriptive stage to the explanatory stage, although the explanation is far from done. Hypotheses are just "possible explanations".
Some hypotheses will always be more plausible than others. For example, that magic comes from the flying spaghetti monsters is pretty unlikely, or that it comes from rodents secretly responsible for generating magic fields, or from the static electricity from the dust under our shoes. There's no reason to lend any weight to these explanations.
But others look more like they could actually be possibly true, even if there's no way to verify.
For example, that magic is the result of an engine we'll call the "Source of Magic" (SoM) left behind by an advanced civilization, that listens for certain humans saying certain words and makes things happen. (A character in Harry Potter and the Methods of Rationality proposes this.) Needn't be aliens—could be a human civilization that erased most traces of itself but for the SoM itself, after perhaps a magical accident. You need not believe it, but it's at least a workable hypothesis. Another possibility is that we're in a computer simulation.
The crucial thing: These hypotheses have more weight than the dust or rodents hypotheses. Thus you already have a ranking of hypotheses, and you're already doing science at this point. The next step: you look at your best hypotheses, at the top of the list, and you ask if they make any falsifiable predictions. Like, if there's a SoM that's "merely" advanced technology, then every time someone conjures water with magic, maybe there's some lake in the world where the water level goes down.
So you go out and look for evidence of that. And another crucial thing is that even negative findings are informative, because it reduces our confidence in the hypotheses that predicted them, so we raise confidence instead in the other hypotheses. Maybe after a while one will look like the by far most likely explanation.
(In the worst case, we'll have multiple observationally equivalent hypotheses, but they may still be countable on one hand, which is a huge progress away from having every imaginable explanation stand on equal ground.)
That's all an explanation is – the least falsified falsifiable proposal. And science is nothing more than a way of working – as I just showed, you can apply it even to something someone calls magic. Nothing you can see is outside the domain of science, for the same reason it isn't outside the domain of your eyes' ability to see. As long as it's describable, it's ultimately explainable, even if it takes a lot of work to make a falsifiable explanation that neither gets instantly falsified nor tacks on many suppositions. I suspect many fields of academia are still on the descriptive stage, because it really is a lot of work. The only way magic would be unexplainable was if it made you forget what you saw every time you saw magic.
What does it mean for you to be able to see something, observe its effects on the world, and for that thing to still be "fundamentally unexplainable"? It doesn't resist description, so it's impossible for it to resist explanation.