Rotating with respect to itself. Every particle in that space station (assuming it's rotating) is under tension, experiencing a net force and hence a net acceleration. Let a bit go, and it'll fly off in a straight line. (At that point, the particle could argue that it's at rest, and the space station is both rotating and moving linearly, because the particle *would* be in an inertial reference frame.) You don't need a reference frame to measure force, just velocity.
As for whether observing a rotating object implies that "something" must have accelerated it in the past, that's a question for the philosophers.