This whole issue is solved easily if you just think of the idea of action, rationality, or whatever as a model that we can use to understand regularities in spatial movements, which could be useful or not. And we find that it is very useful, in that modeling other people as “rational”, or as “actors” tends to let us predict how their body will move in the future. And, of course, that model is also very useful in predicting how the bodies of animals will move in the future!
We don’t need to have some sort of aprioristic reason why other people are rational. We just look inside ourselves, come up with the idea of our own action (I presume you can do this, but maybe you can’t, I don’t know), and then use that idea as a model to predict the movements of other people and animals. In that sense it is an empirical question whether the “action axiom” even “applies” to anything. (And by “applies”, again, I mean lets us understand the regularity in the succession of certain perceptions - that is lets us predict the future movements of other certain people or animals.)
Also note that Ludwig von Mises understood this. See the section in Human Action called “The Alter Ego”. He admits that the question of whether his understanding of his own action applies to anything else is an “empirical” question solved by the fact that it “works”. Nobody can deny that it works. And it should be admitted that it is impossible to say whether somebody else is “really” conscious. Who cares? The question is in vain. The point is that modeling them as conscious works!
Ask yourself this: What if somebody isn’t an actor, but just moves like one? What if somebody isn’t conscious, but his movement is exactly the same as that of everybody else? What would happen to Austrian economics? Would it cease to be “apodictically certain”? No. Saying that he moves like one is giving away the whole thing. The fact that he moves like an actor is enough to let the “action axiom” apply to him. In fact, we can’t establish anything other than the fact that other people move like they are actors! I don’t know whether you are “really” conscious, and that might be a meaningless question anyway.
Again, this whole issue is instantly solved if you just consider it like this: You look inside yourself, come up with an idea of your own volition or will, and then use that as a model to predict the movements of other people (and even animals). All of the metaphysical problems are gone, and all of your concerns are solved.
Your concerns are valid. But I think that my answer not only totally solves them, but isn’t at all out of step with Ludwig von Mises.