Hier het antwoord van prof. Norton zelf:
Thank you for your interest in the dome.
The choice among the different futures involves more than just one bit of information. The mass can slide off in any direction. A choice has to be made among all angles from 0 to 360 degrees. A choice has to be made also over the time of spontaneous motion (T), which can be anywhere from zero time to infinity.
You wonder whether the indeterminism is a result of some unphysical, mathematical artifact. Right? Specifically, might they be discontinuities in the higher order derivatives?
The problem then is to determine how we separate proper solutions from improper ones that are merely mathematical artifacts.
The difficulty is that we need some standard or some criterion to guide us in making the separation. As I think you mentioned somewhere, experiment cannot guide us. The dome system cannot be implemented physically since quantum mechanical considerations preclude it. The dome arises within a formulation of Newtonian mechanics. I don’t see that there is any unique way to pick out the “right” formulation. Recall that Newtonian theory is a false theory physically.
The formulation I used in my paper is just that any motion is allowed if it conforms with Newton’s equations and uses the familiar entities and idealizations. In this common formulation, there is no prohibition on discontinuities in higher order derivatives. A text book example is a point mass sliding over a flat table top and coming to a sharp edge and then falling in a parabola. The table edge and the motion have discontinuities in lower order derivatives than the dome example.
Is that the “right” formulation of Newtonian mechanics. I don’t think that there is any objective criterion that helps us decide.
These last remarks are just a fragment of issues I discussed at greater length in my 2008 paper. (The original proposal was actually in a 2003 paper.)
If this example interests you, you might like to know that, if we deal with systems with infinitely many degrees of freedom, indeterminism becomes generic; and energy and momentum conservation fail. There is already a substantial literature on this topic.
My favorite example is the infinite masses and springs, which is a classical model of a one-dimensional crystal. See, for example, the appendix to
"Approximation and Idealization: Why the Difference Matters" Philosophy of Science, 79 (2012), pp. 207-232.
https://sites.pitt.edu/~jdnorton/homepage/cv.html#L2012
There are many such systems and I’ve incorporated them in my undergraduate teaching. See for example:
"Mechanical Supertasks”
https://sites.pitt.edu/~jdnorton/teachi ... nical.html
See the later part of the chapter that treats collisions among infinitely many bodies. They build to an example of an empty space that instantly fills with rapidly moving bodies.
Hope this helps.
Best
John N.