Yes. But if you take an existing slow ("cold") bowling ball in orbit around a mass concentration and hit it with a fast ("hot") one, it will be moving faster immediately. But now it's trajectory will take it out into a longer orbit where it slows down and spends more time moving near apogee, where it looks "cold".
So you end up with this paradoxical situation where a "hot" mass starts to interact with a "cold" one and ends up making it even colder, so it collapses.
I'm having trouble wrapping my head around how a measure with an absolute value (temperature) can relate to a measure with no absolute reference (velocity). In a purely inelastic collision, all the energy should be transfered to velocity, and thus the temperature not increase. A pure elastic collision would be the opposite. Similarly, if the ball and launcher are moving away relative to a given point, and the ball is launched towards that given point so that it is moving slower relatively, did the temperature increase or decrease? And with rotational velocity, if you double the radius the velocity halves because the total energy (and therefore temperature?) remains constant. Imagine if I used an inertialess drive to move faster - would my temperature have increased?
Temperature is a property of assemblies of particles, and in this case the “particles” are the bowling balls (dust, rocks), not necessarily atoms. Nothing in the formulation requires it to be atoms.
Of course then you get rocks sitting at a different internal temperature than the temperature of the overall cloud of rocks, and yes, over time that would reach an equilibrium. However, heat transfer in space, for large objects, is slow enough that this can be a very long time indeed.
