The self-park feature is not documented very well. Most of the wiring diagrams I have show power coming from the switch, not going to the motor first so the switch (or self-park switch) can complete the ground path. Perhaps that's how the original (no park) and 2-speed motor was wired.
However, a year ago I thought I'd finished a small portion of my restoration, when Randall noted that my self park function should work. After taking the motor apart another um-teen times, it should work now (once I have a functional car to attach it to). The self-park wire was broken inside the unit, and the wire/insulation was in need of replacing anyway. It's a PITA because that wire is soldered to a lug on the cap (that can rotate for adjustment) and the other end is soldered to the winding/brush assembly (lug #1). The wire length is only about 6", but passes through the housing, subassembly frame and gear cover. The gear cover is an illusion, and one would think it didn't need to be, until you try to assemble it.
When the E terminal is connected to ground, it connects the assembly/case to ground, allowing the self-park switch to function. The rubber mounts are not to eletrically isolate the unit/case from ground (though they do). Maybe for noise isolation ?! Both motor terminals are isolated from the case.
Now that my unit self parks, I can park it and move on to more important things... Thanks for pointing out what was not so obvious.