It seems to me that "determinism" is not a rigorously defined concept. It obviously involves
the order-structure of time T(what determines "before" and "after") as well as the possibility of capturing the instantaneous state of the universe at a given time t in T by an element in a certain phase-space Q.
Our notion of "determinism" will greatly depend on the order-structure of T as well as Q (for instance, its cardinality: is an infinite amount of information required to specify the state of the universe).
The popular concept of "determinism" corresponds to finite computational determinism. T is given the order structure of the natural numbers N and Q is finite. Then the state q(t) of the universe at time t can be computed via a recursive function F from the states q(t') at previous times for t' < t (more commonly the immediately preceeding state state is enough ?).
But suppose that F were not recursive but belonged to some other order of the arithmetical hierarchy (let us say Sigma^1) ? Could we still speak of "determinism" ? What if F were beyond the arithmetical hierarchy ?
What is the best way of extending our notion of "computability" to the case in which T has a dense linear order and/or in which Q has infinite cardinality ? How do we express the "determinism" paradigm of differential equations in a rigorous way ? What if the coeficients of analytic solutions are not computable ?
By "predetermination" I mean the idea that the entire evolution of the universe through time already "exists". Suppose that the law of evolution of the universe F were undefinable in first-order logic but that we had predeterminism. I call this "metaphysical predetermination".
What criteria or what experiment can we conceive of that could distinguish pure chance or free will
from metaphysical predetermination ?
I also note that for us conscious beings it seems arguable that finite computational determinism at least is false.