Hmm.
I think there's only one legitimate way to go about this. And I suspect no one will like the outcome. But just hear me out....
Typically, what most of us believe to be rights are natural rights. We look at nature, see what makes sense, and apply it. For example, it makes sense that since man is biologically able to walk, and can exercise liberty in that area, that he has a right to do so (provided, of course, he doesn't infringe on the rights of others.) What this approach does is removes arbitrariness from the equation. You aren't just picking and choosing what sounds good. You're actually observing where the lines are naturally.
When this general approach toward determining rights is extended to children (or invalids) it seems like an unavoidable conclusion is reached. And that is, that you CAN'T directly extend the approach to such individuals.
This isn't to say such beings should be considered property like rocks or insects. But to treat a child/invalid as an fully-competent adult won't work. The results of doing so speak for themselves. In a very real sense, nature itself cries out that you need to adjust things for invalids. And so I submit the following for consideration as a natural law:
Guardians of invalids have the right to restrict any of the invalid's exercise of their rights, without being lethal in doing so, until the individual is no longer an invalid.
Yes, I know, there are questions. Who is a legitimate guardian? How do you determine when the individual is no longer an invalid? Etc., etc. But those are just quibbling details that can be dealt with later. The core concept, to me, seems perfectly sound. In fact, I don't see how one could argue otherwise in the face of the natural requirements for invalids.
Of course, this leads to some curious, even unexpected conclusions. This is where the whole you're-not-gonna-like-this bit comes in. But again, I'll stand by the claim that it's a non-arbitrary, naturally-derived law. Anything else seems to be starting with a desired outcome, regardless of how natural or right, and working backwards to form an axiom supporting it. (And is that really how we want to go about this?)
I'd address some of these unpopular conclusions, but it's late, and I'm pretty sure one or two will be brought up eventually, so I'll wait until then.