Magical thinking is only not scientific thinking because it doesn't work. As soon as sympathy manifests real effects, you're just doing new science.
A distinction I was drawing above was between expectations of consistency, even as to being
inconsistent. So one view put in this thread is that if there are world elements incompatible with what we know of the real world, the more consistent attitude would be a global suspension of expectations grounded in science.
What I pointed out is a well known category of thought -- broadly "dualism" -- which expects inconsistency. One with a dualistic attitude can accept magical elements while accepting that everything other than those elements can be explained scientifically. Perversely, they need draw no concrete boundaries around which is which.
One may point out that this seems irrational, which is accurate. A common step in games is to slip magic into the rational... ignoring that "magical thinking" in our real world is to some extent a rejection of rationalisation.
Returning to your thought, say that both dragons and birds can fly, there are a few models that can be predicated upon
A. Whatever accounts for flight in this world, it does so for both birds and dragons (this goes in the direction of your thought)
B. Whatever accounts for the flight of birds, something else accounts for the flight of dragons (here I bin Occams Razor and help myself to as much complexity as needed, which I take to be one of
@pemerton's objections; I do not deny that it's unreasonable, yet it's harder than it looks to say what exactly requires a dualistic thinker to meet "rational" standards of reasonableness?)*
C. Nothing accounts for either. My objection to this is similar to
@AlViking's i.e. that, barring dragons, if birds figure significantly in our play we may well base what we go on to say about them on our real world... and that will include any scientific knowledge we have about them. Just because dragons fly, doesn't mean that we will cease to say that birds lay eggs and instead come up with random other descriptions for them. The imaginative effort would be overwhelming! Rather, this sort of working from common experience happens again and again until we meet deliberate exceptions (so we may say that birds don't lay eggs and in doing so everyone will see that we are calling out an exception, not working from a presumed
tabula rasa!)
So this was a long winded way of saying that assuming magic to be rational -- to become scientific -- if it worked, should lead one to agree with
@pemerton. Because once A is excluded by a world containing elements that cannot be reconciled, a "magical" thinker may choose B, but a "rational" thinker ought to prefer C! And that is especially true of a fictional world whose explanations need bear only the weight of all that which comes into play, but nothing else.
*To give a real example, the Catholic church upholds that miracles have occurred and that they cannot be explained by science. God accounts for miracles. But many Catholics do not denounce science: they accept that the Universe can be explained scientifically. I suppose a Catholic would say that God accounts ultimately for physics and miracles! But my dualist need not make that move.