Reno, thanks for the reminder about 310.4. I think the most frustrating thing about studying the NEC is that there are so many rules for which no rationale is stated, even in the handbook.
I suspect the restrictions on parallel conductors have a lot to do with termination methods. Trying to connect two wildly different-sized wires into one terminal could lead to some "creative" and dangerous solutions.
Fault tolerance was probably also a concern. In your hypothetical 2/0 and #2 example, the loss of continuity in the 2/0 would leave all the current flowing in the #2, which would reduce it to glowing slag. At least if the conductors are equally sized, the overload is limited to 2:1.
I believe you when you speak from experience that unequal-sized parallel wiring produces heat at the junction, but there's got to be some other mechanism at work -- possibly higher resistance in the joint itself because of unequal pressure on the conductors.
