Listening to the recent InOurTime on the Second Law of ThermoDynamics, I was struck by something JohnGribben said.

Most of the program was dedicated to talking about the development of the idea of the second law, and of entropy, and painting the picture of the universe "running down" from a point of "high order" (energy concentrated in stores) which disperses (producing various kinds of work, including the construction of local complexity and order) leading to a universaly more disordered state.

So far, so consistent with the physics I learned at school and everything I've come across since in my PopScience reading.

But then Gribben, as Bragg put, threw a hand-grenade into this model by suggesting gravity works against or not according to this principle. Gravity creates concentrations of energy (order) by pulling stuff together.

And then I realized I don't know enough about gravity.

Imagine a closed system, consisting simply of two large, massive bodies, which exert gravitational pull on each other.

What energy is being consumed / dispersed to produce the acceleration and movement of these bodies?

OK, I suppose we're talking about something perfectly efficient (1) so PotentialEnergy is getting turned into KineticEnergy? And ConservationOfEnergy tells us that's OK.

But now I have two slightly contradictory intuitions.

nothing much can happen in ClosedSystems ie. where external energy isn't being dissipated.

if movement can stop and start AND material can get distributed more or less densely, without dissipating energy, then it seems quite a lot can happen.

Following on from the second. How important is the second-law really? It seems we could imagine information processing CellularAutomata which didn't need to increase entropy to do their thang and increase complexity.

maybe the thing here would be to try to demonstrate it :-)

By clustering stuff together, Gribben seems to imply, gravity actually works against entropy and creates more order without dissipating it elsewhere. Is this what people think? Or did I misunderstand? Or is Gribben flying a speculative kite of his own here?

(1) Maybe it can't be perfectly efficient. Even in deepest space vacuum isn't perfect. There are other particles to be collided with, so some friction. Maybe there other types of dissipation during this acceleration that I don't know about.

I think that to decrease entropy you need to lose information. Until recently it was believed that BlackHoles erase information from objects falling into them. But now the general opinion is that they actually do not completely erase information, and hence do not reduce entropy.

– ????

More :