PDA

View Full Version : Thermodynamical thought experiment with time reversing loop?



Jarek Duda
12-17-11, 06:14 AM
There is some thought experiment constructed to help finding intuitions about the 'conflict' between CPT conservation and 2nd law of thermodynamics. Like for wormhole loops, I see it completely unrealistic - I only think it is inspiring mind exercise and may lead to some better understanding of thermodynamics (and temporal logic).
General Relativity Theory determines shape and rotation of light cones in each point of spacetime (like near black hole), but it is time (and CPT) symmetric - doesn't directly distinguish between past and future of such light cones. Without any additional reasons like entropy gradient, we could time-flip the light cones.
Another from 2nd law of thermodynamics way to distinguish past and future of such light cones is by continuity ... but let us imagine there is some loop on which GRT makes that light cones have configuration like that:
http://dl.dropbox.com/u/12405967/loop.jpg
Where the start and the beginning have the same position, but may be shifted in time. Of course it would require some really nasty singularity inside such loop - even more than in the center of black hole where spacetime is no longer a manifold.
There are rather completely no reasonable scenarios to obtain such singularity, maybe mathematics could forbid such global solution (but black holes require hypothesis: cosmic censorship hypothesis (http://en.wikipedia.org/wiki/Cosmic_censorship_hypothesis)) ...

Only for this thought experiment, let us assume for a moment that there exists such time reversing loop - and a rocket flied through it and returned back to Earth orbit.
If someone really doesn't like such loop concept, one can imagine that this rocket was transformed by CPT symmetry - it would be made of antimatter, but it would be enough for thermodynamical considerations.

Ok, let's get to the main subject - thermodynamics.
Inside this rocket, the astronaut shouldn't feel a difference - he could e.g. just break a mug ... but from our perspective it would be time reversed: pieces would get together into the mug.
Everything (like mugs) have tendency to get into higher entropic state (broken), but our things (mugs) came from past reason-result chains, so such state change (breaking) can only have e.g.: unbroken state toward past time direction and broken toward future.
In contrast, reason-result chain of mugs from the rocket came from our future time direction, so it can increase entropy only while breaking toward our past.
So it seems that 2nd law doesn't only emphasize just e.g. entropy gradient direction, but can work in both - depending on reason-result chains ... ?

This thought experiment becomes real mind feast if we allow the astronaut to land (not antimatter case) :)
Time reversed molecules are nearly the same, temperature is average energy so it also shouldn't depend on time direction - he should be able to just breath in our atmosphere (??)
His body should be in thermal equilibrium with environment - heat exchange should work normally, so I don't see a reason his time-reversed metabolism should work improperly (??)
So it would seem that he could also eat our food ... but there appears a problem - from his time perspective, it could need turning e.g. back into a chicken :)

The situation is really really strange - great mind exercise - I would gladly hear your comments, expansions ...
I think the only reasonable causality understanding here is Einstein's block universe (http://en.wikipedia.org/wiki/Eternalism_%28philosophy_of_time%29) - that like in GRT, the spacetime is already created and 'we only travel to our future there'.
So eventual time-loops are already made self-consistent, like in good SF movies (e.g. Twelve monkeys) - if one would like to kill his grandfather in the past, there would happen something that he couldn't do it.
If you disagree, how do you understand the conflict between CPT conservation and 2nd law?

How would look such contact of time-opposite natural reason-result chains? (theoretically allowed by CPT conservation)
For example some believe in cyclic universe model (http://en.wikipedia.org/wiki/Cyclic_model) - that our universe will finally collapse into nearly a point.
From the perspective of CPT conservation, such Big Collapse point would be Big Bang it reversed time - low entropy state (spatially localized) creates entropy gradient (2nd law), starts reason-result chains ... and so evolution of universe in reversed time direction ... which should finally meet with ours in some far far future.
Why against current acceleration growth, it should finally start collapsing? Because of energy conservation - gravity pulls together (1/r^2), while some 'dark energy' push it out, but its density (and so strength) should decrease with the volume (1/r^3) ...

Jarek Duda
12-17-11, 08:13 AM
Moved? Please at least explain this censorship ... e.g. that General Relativity Theory doesn't allow for such light cone configuration?
Is there a deeper reason that spacetime have to be orientable? Or maybe it needs to be additionally assumed?
This nonorientability doesn't need to be global - it can be local singularity (like for sqrt(z*(z-1)) ):
http://dl.dropbox.com/u/12405967/loop1.jpg

... and I clearly stated that I completely agree that it's completely nonrealistic - that it is only inspiring thought experiment ... it's natural for mathematicians - make some strange looking assumption and just play with it, for example lead to contradictions, understand what's wrong with it ...
Why physicists cannot just relax and play with some simple gedankenexperiment - it could be fun and really great intuition building exercise ...

origin
12-17-11, 08:54 AM
Moved? Please at least explain this censorship ... e.g. that General Relativity Theory doesn't allow for such light cone configuration?
Is there a deeper reason that spacetime have to be orientable? Or maybe it needs to be additionally assumed?
This nonorientability doesn't need to be global - it can be local singularity (like for sqrt(z*(z-1)) ):
http://dl.dropbox.com/u/12405967/loop1.jpg

... and I clearly stated that I completely agree that it's completely nonrealistic - that it is only inspiring thought experiment ... it's natural for mathematicians - make some strange looking assumption and just play with it, for example lead to contradictions, understand what's wrong with it ...
Why physicists cannot just relax and play with some simple gedankenexperiment - it could be fun and really great intuition building exercise ...

Your thread wasn't censored it was moved to the appropriate folder, you should relax.

It doesn't seemed that you have explained your idea very well (at least for me) you seem to have basically said time can go backwards with no explanation of how this occurs and then went from there.:shrug:

Jarek Duda
12-17-11, 09:28 AM
Moving to 'on the fringe' is not a censorship?
Explain ... GRT basing on local conditions (local stress-energy tensor and surrounding configuration), determines shape and rotation of light cone in given point of spacetime, like here are near black hole:
http://www.phy.syr.edu/courses/modules/LIGHTCONE/pics/blackhole.gif
It's just that GRT alone theoretically allows for really exotic configurations, like wormhole-type allowing to jump back in time (using higher dimensions) ... this example uses possibility of nonorientable configurations (like Mobius strip (http://en.wikipedia.org/wiki/M%C3%B6bius_strip)) - which while moving through a loop would lead to exchange of past and future cones ...
Maybe it leads to some deeper contradictions (on global level), but GRT doesn't seem to directly forbid it - such contradiction would need to be proven first ... maybe leading to some deeper understanding of orientability requirement (http://en.wikipedia.org/wiki/Orientability).

Personally I see both wormholes and presented configurations as completely nonrealistic, but I needed a imaginative picture for thought experiment: to reverse 2nd law of thermodynamics to see why there is no conflict with CPT conservation - that thermodynamical tendency to increase entropy also doesn't need to emphasize any time direction (is time symmetric) - if there would be a way to provide a mug from future direction (e.g. through such loop), accordingly to thermodynamics it should break in opposite direction than of 2nd law of thermodynamics.

What lead me to that thermodynamics doesn't really emphasize any time direction (what would contradict more fundamental CPT conservation), is new approach to thermodynamical modeling I was developing e.g. for my current PhD thesis: http://arxiv.org/abs/1111.2253
It shows why Brownian motion is often an approximation - makes some unjustified assumptions - only approximate maximal entropy production. If we do it right - using recent Maximal Entropy Random Walk (http://prl.aps.org/abstract/PRL/v102/i16/e160602) and expansions, it for example doesn't longer disagree with thermodynamical predictions of quantum mechanics (leads to quantum ground state stationary probability density), naturally explains Born's rules ... or e.g. shows a deeper time symmetry of thermodynamics.

Jarek Duda
12-18-11, 04:17 AM
I've just found 2002 "The orientability of spacetime" paper from Classical and Quantum Gravity Journal: http://iopscience.iop.org/0264-9381/19/17/308/
Here is arxiv version: http://arxiv.org/PS_cache/gr-qc/pdf/0202/0202031v4.pdf

It occurs that the lack of such time-reversing curves is precisely called time-orientability (there is required a distinction in semi-riemannian manifolds) and the author even propose a test:
"In a true test of time orientability on a region R through which time cannot be oriented, the clock readings would increase steadily along the path taken by the clock. Before entering R the observer sees the time values increasing on the clock. When the clock exits at a time /rho the increasing clock times would be at ever decreasing values for the observer time. The observer would still see a backwards counting clock, but only at times before /rho"

The author also refers a few other peer-reviewed ones, for example in Sorkin R D 1977 J. Physics A 10 717–725 there should be constructed a non-orientable wormhole (asymptotically flat), but it's not accessible electronically.

AlphaNumeric
12-18-11, 02:18 PM
The second law of thermodynamics (2LTD) is a statistical principle, not an absolute one. Local fluctuations can alter entropy without violating any physical principles. In fact, the Poincare recurrance theorem says for certain systems, including many physical ones, it's certain that it'll decrease by its own accord for short periods of time, if you wait long enough.

The issue, IMO, seems to be people work in terms of the macroscopic variables like temperature, pressure, volume, entropy, enthalpy etc. Of course if you reduce all the degrees of freedom of a box of gas particles to half a dozen you've projected out a huge amount. If you view everything in terms of configuration space then the system wanders around in a similar manner if you run time forwards as if you run it backwards.

Furthermore, consider what happens when the system, under the recurrence theorem, returns to be near it's initial low entropy state. It will decrease in entropy in a manner similar (but perhaps faster) than how it increases entropy in 'normal' behaviour.

Personally I think viewing the system so much through the eyes of entropy is not a good way. It's actually a rather poor measure of structure in a system.

Jarek Duda
12-18-11, 03:20 PM
The Poincare recurrence time is exponential to the number of particles, so for macroscopic systems would be much larger than the age of universe ...
While inside such time-reversed rocket ... from our time perspective, 2nd law would just go backward :)

The main purpose of this thought experiment was to show that 2nd law is not about some e.g. entropy gradient of the universe ('the arrow of time'), but mainly about reason-result chains.
If in some point of spacetime there is some ordered system (of relatively low entropy), it has tendency statistical tendency to break this order (2nd law) ... but this tendency doesn't prefer some time direction - theoretically could work backward, but our well ordered things comes always from our past reason-result chains, so 2nd law works toward our future.
Such hypothetical loop could provide us with well ordered things from future reason-result chains, so their 2nd law would work backward.

Similarly for the balls governed by Newton mechanics in your Poincare recurrence considerations - you could just reverse time - there is completely no point that some statistical 2nd law would emphasize a single time direction.


Personally I think viewing the system so much through the eyes of entropy is not a good way. It's actually a rather poor measure of structure in a system.
It's complicated ... like quantum mechanics doesn't only represent subjective knowledge, but combine it with including wave nature ... representing subjective knowledge of limited observer is only half of nature of thermodynamics.
The other other half in information theory is called Asymptotic Equipartition Property (http://en.wikipedia.org/wiki/Asymptotic_equipartition_property) - that when there are no deeper reasons for some e.g. correlations, the system will most probably choose maximizing entropy ones - among all possible configurations, the number of these ones dominates all the other (completely while N->infinity).
I've explained it better in http://arxiv.org/abs/1111.2253 (its time dependence chapter gives intuition about time symmetry of thermodynamics).

ps. Here is example of reversed 2nd law for subjective information representation (by attaching in the future):
We know that given person just couldn't miss given meeting (almost certainly will be there in that moment) ... accordingly to our knowledge, 15 minutes before the meeting probability distribution of his possible positions will be more spread (larger entropy) and so on entropy will grow while going back in time...
... but after the meeting time such entropy will grow again - this time in the proper direction.
Big Bang is kind of such a meeting - was well localized, so had relatively small entropy, so created entropy gradient and our natural reason-result chain ... (and has a twin if it was Big Bounce).