Hacker News new | past | comments | ask | show | jobs | submit login
Gravity Might Explain Why Time Never Runs Backward (wired.com)
70 points by ghosh on Nov 5, 2014 | hide | past | favorite | 32 comments



I have read the original PRL paper, and I think that the significant issue with it is that it conflates the complexity (C5 in the paper, a measure of the clustering of the gravitational objects) with entropy. I don't think that this leap is warranted.

Complexity, as defined by the authors, is primarily a spatial dispersion measure without a corresponding measure of energy dispersion. As far as I can tell, this system should be subject to Liouville's theorem, which means that the apparent entropy decrease (the decrease in C5) that occurs when the system contracts is just hiding the entropy in the phase space of the particle velocities. So, as the particles get closer together, their velocities get farther apart.

This same kind of effect occurs in particle beams when you try to squeeze the particle beam tighter. Assuming you aren't using some kind of beam cooling (like adding cold electrons, or using stochastic cooling), every time the beam gets squeezed, the beam gets tighter but the phase spread in the particles increases. When the beam spreads out again, the phase spread usually goes back down. It can get as low as its original value, but no lower. Just like entropy (cause that's what it really is).

That's just my thoughts after a brief perusal. Feel free to point out any obvious errors in my logic.



The article is poor.

I think the research is another take on how finite reversible systems end up cycling. Their entropy has to eventually decrease because as a path through the phase space runs out of places to go it must return to the initial state to avoid getting stuck (i.e. violating reversibility).

I'm guessing that the interesting new thing here is that the researchers showed systems following gravity have this entropy-goes-up-and-down property, despite being an infinite continuous phase space where the bits could just get more and more dispersed?

Of course in the actual universe there's other forces at work and they need to be accounted for. You can't just say "gravity creates low entropy times therefore we have explained the low entropy", because things like the accelerating expansion of space have to be accounted for. You also need to account for how incredibly low entropy was near the big bang: it wasn't just a single galaxy's worth of stuff converging, which would have been sufficient and far far more likely, it was billions of times more.

Sean Carroll frequently talks about this issue in has talks and on his blog. Hopefully I didn't get the details completely wrong.


This is completely wrongheaded nonsense. Time is only an abstraction related to change. Even for unrelated events which happen in the same point in space (such as the arrival of signals from different sources), we can tell whether they happen at the same time, or else that one happens before the other. This is because we have some clock running: a moving machine, and positions of its hands on its dial are interpreted as "sooner" or "later". As to why the clock doesn't run backwards, that can easily be explained using classical mechanics.

Some time-keeping machines do in fact reverse their motion. For instance an ideal pendulum, which can serve as a clock, traces a path that perfectly reverses itself every period. Looking at just the pendulum, we cannot tell whether the pendulum has reversed direction due to the restoring force, or whether suddenly time has started flowing backwards: the motion is symmetric. This is irrelevant, because we interpret the pendulum's motion as a parametric curve. The abstract t parameter of that parametric curve marches forward.


I seem to remember from watching a Sean Carroll talk (can't remember which one now - could be this: https://www.youtube.com/watch?v=ZbPFrzliZJY - very entertaining), but he's sort of an expert in this subject, that the ever increasing entropy is a phenomenon not a law of physics, i.e. there is no underlying reason why the entropy couldn't be ever decreasing, and that possibly there was a "time" when it was the case when the universe was collapsing, which led to the big bang. If that's true, then in that period time was essentially running backwards.


"simulated the universe as a collection of 1,000 particles that interact with one another only by gravity, representing the galaxies and stars that float around the cosmos. The researchers found that regardless of starting positions and velocities, at some point the particles inevitably find themselves clustered together in a ball before dispersing again"

Wait, what? That conclusion doesn't make any sense to me.


The most interesting notion is that of using complexity, a dimensionless metric that can be measured from within, instead of entropy, which can only be estimated from within, can only be measured from without.

Since our universe doesn't have a "without", we can estimate entropy. But we can measure complexity, as it used in this article.

Intriguing. Reformulating the Laws of Thermodynamics could make for a promising avenue of research.


I fail to understand how this article (and the research that underlies it) have shown anything more sophisticated than "interacting particles interact"... Can someone help me understand the novelty?


As far as I can tell, it's not a novelty, there has been other similar research. Gravity -> certain types of interaction theories -> restrictions of entropy -> time. That connection has been condensed into "gravity causes time". Bleh.

The article is making sweeping statements, as usual.

Physics undergrad here, done some advanced courses but not an expert on this stuff


I thought the non-existence of time explained why time never runs backward.


Why wouldn't time move backwards for the entire universe? If it does, there can be no record of any future event and there is no way to determine that time is moving backwards. If time moved backwards in one spot, what about all the information connected to that spot. Why wouldn't that information rollback too?


>Instead of using entropy, the researchers describe their system with a quantity they call complexity, which they define as roughly the ratio of the distance between the two particles farthest from each other to the distance between the two particles closest to each other. When the particles are clumped together, complexity is at its lowest.

When the particles are clumped together, the entropy is highest. Gravitation pulls in the direction of entropy increase. For a system of many objects spread around some volume their "complexity" is just kind of "parallel" to entropy, not really "instead". Thus it isn't surprising that they get the first half - clumping - somewhat right. The second part - bouncing back - is pointless to discuss in the framework of their model because such important factors as, for example, space inflation (vacuum "thinning") were omitted from the model while that inflation is the key for enabling new Big Bangs in the old/inflated/cooled down Universe which thus becomes proto-Universe for the new ones.


I am pretty confused by this article. In order to simulate gravity, you have to have already decided on a direction of time, right? You'd know if time was running backwards because gravity would push things apart rather than pull them together.


Although counter-intuitive, reversing time doesn't change the rules of gravity: the basic law of gravity just says that objects accelerate in a certain way dependent on their masses and distances. Because acceleration is the second derivative of position with respect to time, it is doubly negated, and thus unaffected, by reversing time.

Reversing a planet's orbit produces another orbit just as valid under the law of gravity. And even at the scale of ordinary experience, gravity doesn't prohibit objects from rising (after all, in the canonical example of tossing a ball through a parabolic arc, there is both a rise and a fall); footage of a ball falling under the influence of gravity reverses into footage of a ball rising under the influence of gravity and vice versa, but never does footage compatible with the law of gravity reverse into footage incompatible with the law of gravity.


True enough in free fall in a vacuum. (if one could find such a perfect vacuum)

Apples don't usually leap off the ground into trees, though. Of course, that would be more of a heat to kinetic energy conversion to start that process, but I think that's more along the lines of what the original question really meant: why don't things spontaneously launch, rather than having a one way tendency to convert potential energy to kinetic energy to heat energy.

It's hard to see gravity act in isolation: trace particles eventually slow an orbit in the same way that air friction deforms an idealized parabola/elipse by dragging on a ballistic projectile; tidal forces between earth and moon alter the moon's orbit.


Sure. As you note, an apple would leap into the trees if the ground under it conspired to push it upwards. That the ground rarely does so (but the reverse, an apple falling and dispersing energy into the ground, happens frequently) is certainly an observation which merits explanation, but the explanation doesn't come from an asymmetry in the laws of gravity, as such. It's often said the explanation is found in something like the Second Law of Thermodynamics, which in turn is explained by the fact that at some moment, there was a state of extremely low entropy (the Big Bang); the "arrow of time" is then a manifestation of being to one or the other side of this particular low entropy moment. And the work discussed in the link (which I have no direct knowledge of) purports to show why some such low entropy moment should be expected to exist at all given the action of gravity over history.

But my point was simply to note that you don't have to pre-incorporate an arrow of time to make sense of the laws of gravity, in response to the poster who found this a sticking point and expressed confusion over it.


No. The thing is that gravity is a conservative force; that is, energy is conserved in a gravitational interaction (ignoring radiation of gravitational waves). Because of that, if I reverse the direction of time, gravity stays attractive, and the objects trace the exact same motion in reverse.


Energy is conserved in any iteraction. And gravity is not a force. Gravity is the curvature of space-time.


> Energy is conserved in any iteraction.

True. But gravity is a conservative field (again, neglecting radiation), meaning that the energy that goes into raising something against gravity, you can get back by lowering it to the original position again.

> And gravity is not a force. Gravity is the curvature of space-time.

If we're not in "overly pedantic argumentation" mode, then I'd say, go climb some stairs. Feel that? That's a force. But if we are in "overly pedantic argumentation" mode, then yes, gravity is the curvature of space-time in General Relativity. That may not be the final word, though. What is gravity in Loop Quantum Gravity?


Let me give you a clue. Are you in the gravitational field of the Sun? How does it compare with the gravitational field of the Earth? Is it stronger or weaker? Do you feel it? What's the trajectory of the earth in the curved space-time around the sun?

When you'd say "go climb some stairs" what you feel are the stairs. And you should also feel like a bit of an idiot, in my humble opinion.


Well, its semantics right? Gravity causes acceleration, so its a force? It can be treated mathematically as a curvature. It behaves differently in different frames, but so do electromagnetic forces.


If someone applies a force to you, you can feel it. If you're in free-fall, you can't feel anything. According to general relativity, there's no way to tell if you're stopped, moving in a straight line with constant velocity, or in free-fall in a gravitational field. They're all the same.


Actually, you only feel it if the force is applied differentially (which most forces we experience are). If your body had a volumetrically uniform electrostatic charge, which was not high enough to have charge screening, and you were acted upon by an electrostatic force, you wouldn't feel anything either. It is just that 'mass charge' is always distributed evenly along with mass, while with electrostatic forces this is a special case.

Forces can indeed be modeled as a curvature in space-time. However, there is a bit more to GR than just that, which is why it took Einstein years to go from SR to GR.


There is a distinction. A gravitational field has a center, so if you are in free-falling towards it you can determine that you are in it.

Take two objects and place them apart to float stationary( the line they make should be orthogonal to the center of the field ). Even though they started completely stationary, they will slowly start to move towards each other, as an unknown force is acting upon them. In reality they are falling towards a common center.

I guess this doesn't work on dimensionless points, so the your comment is correct as far as mathematics are concerned.


Yes I think it's semantics - different yet valid ways of looking at the same phenomenon.


Actually there are several phenomena and theories which give the arrow of time a certain direction. See:

https://en.wikipedia.org/wiki/Arrow_of_time


I know it's Wired's headline, not HN's, but let's re-write it for accuracy, shall we?

"How Gravity Might Explain Why Time Never Runs Backwards (On This Side of the Big Bang)".


Ok, we added your "might" and took out the baity "how".


Oh, wow. Cool. I wasn't actually asking for moderators to change the headline; I was just whining...


That was high-quality whining. I keep trying to tell people that the best way to complain about a title is to suggest a better one.


I wonder how much tax-payers' dollars they've spent to do this research :)


It's fairly easy to estimate. Assuming an average salary of about 50k between a grad student, a postoc, and a PI (a typical mix for a scientific publication), one only needs to ask how long it took them to do the work. Estimating about 6 months with a 3x infrastructure multiplier, I get about $225k.

This is of course giving you the benefit of the doubt that you were actually wondering, and not passive-aggressively implying the government should not fund basic scientific research. Which is understandable, because that's a ridiculous position for an educated person to articulate explicitly.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: