IMO too much stock is placed in what we humans think is reasonable. Why isn't it enough to say no part or phenomena of nature we have observed behaves in this manner?

You answered the question yourself, time arrow. If we agree that our physical world obey some specific laws, then second thermodynamic law (i.e. time arrow) answers the question and provides the reason for why we( and all the world) get older.

"... message might not contain any information" - Anything detectable contains some information so there will always be a paradox.

"But that is ONLY a model" - This is a favourite line among pseudoscientists. Science is only a collection of models, but we trust our lives on a daily basis to those models that have been well-tested. General relativity has passed all the tests thrown at it with flying colours.

"Is the possibility ruled out by the general relativity model?" - Technically the model WILL allow for time travel actually, but only if we include some kind of 'exotic matter' that the rest of physics rules out; in particular, matter with negative mass/energy.

"Why EXACTLYcan I send a message to be delivered in the future" - As answered by yourself and others, because time does indeed have a direction.

You know another point of view is that "a message from the future" may not be a thing that can be given any real meaning. What physical form would such a message take? A letter in your mail box from the future would violate energy conservation since the mass of the letter would have to suddenly appear from nowhere. Likewise, a radio broadcast would have to appear to come from somewhere and where would that be since the transmitter would only exist in the future? Letters could be spelled out in your tea leaves but some force would need to exist to form the letters. What would be the apparent origin of those forces? Can you offer a message from the future scenario which would make physical sense?  

I suppose that another way of asking the question might be:

"If small-scale "billiard-ball" models allow us to extrapolate either forwards or backwards from a given situation (if we're given a sufficiently complete snapshot to work from), and it doesn't seem to matter much which direction we extrapolate in, because the simplified physics seems reversible ... then in these simplified self-contained systems, a snapshot arguably contains an "imprint" of future impending events as well as past events.

... so why is it that in our larger world, the past is so much more heavily imprinted on "now" than the future is?"

I agree with the "arrow of time" answer, but it's be nice to be able to come up with something more specific.

Paul: "Why isn't it enough to say no part or phenomena of nature we have observed behaves in this manner?"

Because one of the most valuable aspects of science is that it sometimes lets us frame deep hypotheses that go beyond the simple modelling of known events, and let us discover brand new types of experience that have never before been seen or reported.

The ability to anticipate gives an organism or group a major Darwinian advantage. I think we can classify conscious anticipatory abilities into three levels:

Level One is the ability to note and remember associations, and expect those associations to hold in the future. 

Level Two is the ability to create compact models and explanations that make it easier for us to remember these rules, and to apply them more widely. But we're still basically just doing retrospective curve-fitting.

Level Three is enshrined in the scientific method - we create deeper hypotheses that are capable of predicting brand new classes of event that we've never before witnessed, and then we test and fine-tune our  predictive superpowers by sticking our necks out and making disprovable predictions that go beyond anything that's yet been seen, and (hopefully) find out whether or not we were right.

That's the whole point of having disprovable hypotheses in science, we're supposed to be constantly challenging ourselves to anticipate the discovery of  outrageously new types of effect, before they're actually been seen.

If we can only "predict" new effects after we've discovered that they are physically real, then science's provable utility to society and to the species is lessened. If we gave up on predicting the as-yet-unseen, then we'd be dropping back from Level Three to a mere Level Two, and would find it more difficult to claim that science is a "special" activity.

Stephen: The difference between a model and a theory is that a theory is based on founding principles that are (hopefully) open to scientific falsifiability, whereas models that aren't theory-based can be fudged, bent, compartmentalised, assigned a nominal range of applicability,  or retrofitted with new ad-hoc parts whenever they give the wrong answers.

So if someone is hoping to do real comparative tests to rule out one or more theories, to establish the actual rules that the universe plays by, then "only a model" can sometimes be a legitimate complaint, because models that are //only// models can find it harder to meet the "scientific falsifiability" test.

Eric:"models that aren't theory-based can be fudged, bent, compartmentalised, assigned a nominal range of applicability, or retrofitted with new ad-hoc parts whenever they give the wrong answers." Good definition, and I would then surely prefer theories to models.

But General Relativity: only a model? In the sense you mean, surely not!! It has in essence only one parameter, and that one is already known, and cannot be fudged. Also, the predictions have been successfullly tested to a considerable extent. It also meets the scientific falsifiability test to a higher degree than most other theories or models.

Answering the original question: I believe Time's arrow, meaning the second law, makes it impossible. Indeed, ``sending a message''  involves the possibility of communicating information. (otherwise, maybe it is not impossible: your father in his diary, say, notes that he saw a duck swimming in a pond on the day you were born, so you can choose to intepret this as your message.)  But if you require more of a message, then I believe the Second Law will make it impossible: somehow, it will require an irreversible process going in the wrong way, or the information that could be passed by the message could be used for some thermodynamically impossible purpose. One would require, however, a clear concept of what a message is.  

Eric: I'm not certain the "depth" of a question is a definable property. However, a well formed question must be consistent with know physical principles. The (your) reformulation of the question in terms of time reversibility of micro-physical laws ignores global boundary and initial conditions. This is why for example energy is not conserved when the letter arrives from the future. One would have to rewrite the initial conditions of the universe to create a letter to be delivered in order for the existence of letter to be consistent with know physical laws. Good luck with that. 

Thanks for interesting and insightful answers -- I am not sure much of a consensus has appeared -- I will think more.

I've been reminded that the excellent book "Times Arrow and Archimedes' Point" by Huw Price, Oxford University Press, 1996 has a lot to say about all this -- but it's not that easy reading -- but then there is "no such thing as a free lunch" ...

You might like to glance at my paper ("Agents and MAS in STaMs")  for a partially worked out approach that has something in common with Eric's answer. It's readable/downloadable here on RG. My approach is a bit too simple minded, perhaps, (which is why I asked the question) but still thought-provoking and worth developing I believe.

Thanks again

Jim

F. Leyvraz: "But General Relativity: only a model? In the sense you mean, surely not!! It has in essence only one parameter, and that one is already known, and cannot be fudged."

Einstein's original logical structure for general relativity seems to have publicly failed in about ~1960 after the publication of the Harwell group's paper, when people realised that the general principle of relativity (as outlined by Mach and Einstein) appeared to be geometrically incompatible with the idea that the theory could also include the physics of SR as a complete subset.

The community then seems to have had a bit of a crisis (which nobody involved seems to want to talk about).  Einstein had already suggested that the current "SR+GR" structure was "historical" and indefensible with hindsight, and said that it would have been better to build a general theory from the ground up without relying on any components that weren't  already conformable with the general principle of relativity from the outset. However, the "total rewrite" option seems to have been considered unacceptable in the early 1960s, and instead we decided to keep the structure that we had, but delete or downgrade any principles or other aspects of the original theory that might generate further conflicts with SR.

FL: "It also meets the scientific falsifiability test to a higher degree than most other theories or models."

It fails at the galaxy scale unless we invent dark matter to compensate, it fails at the cosmological scale unless we invent dark energy, It's believed to be incompatible with quantum mechanics, and having logically failed its founding definitions in ~1960, it's now been given internal "failure priority rules" that mean that no matter how logically inconsistent some parts of the internal structure still seem to be, we now can't use those things to claim logical inconsistency, because of catch-all rules that continuously create "soft failures" rather than "hard crashes" ("If A conflicts with B, then since A isn't allowed to conflict with B, comparing A with B in this situation is an illegal operation, making any apparent inconsistency the user's fault"). The resulting structure is theoretically inconsistent, but our operation of the structure is now //consistently// inconsistent, which I think is probably the hallmark of  a model rather than a theory.

Because of the "general" parts of the theory that were jettisoned for conflicting with SR or downgraded so that they are "allowed to fail sometimes" or be "only a guideline", GR is also no longer literally a fully "general" theory of relativity. It's maybe two-thirds of a general theory, grafted onto an SR base.

It's not uncommon to see GR people saying things like, "Oh of course, acceleration and gravitational effects aren't really interchangeable, they just look a bit similar, but they're really fundamentally different things" ... in which case, that's the theoretical basis of the general theory, as understood by Einstein, gone.

Regarding GR and experimental support, most of the "good bits" of GR (like gravitational shifts and gravitational time dilation) predate the theory and don't actually require the structure of GR1916 to calculate, they just require the basic principles that Einstein adopted for GR, not his specific implementation of those principles. For instance, Wheeler pointed out that the Gravity Probe B prediction didn't need GR, it was a back-of-the envelope calculation based on the democratic principle.

And there's also the point that, experimentally, GR was probably the most "accident-prone" major theory of the C20th, when it came to experiments that failed, or that gave the right answers where, with hindsight, they probably shouldn't have. I forget the exact quote in MTW, but it was something along the lines of "GR has been a theorist's paradise but an experimenter's hell". This isn't entirely GR's fault, but there does seem to be a certain amount of "cheerleading" about GR testing that I really don't understand - it just doesn't seem to correspond to the subject's track-record.

So from a logical standpoint, the structure of "Einstein's general theory" seems to have been criticised by its creator and then logically falsified over half a century ago, and experimentally, if we strip out the general results that don't require the GR1916 structure to calculate, we aren't left with terribly much apart from larger-scale predictions that are wrong, and a lot of fascinating "new" results about black hole horizons from the 1960s which are almost impossible to test, and which are also probably wrong because they disagree with QM.

I love the idea of a general theory of relativity, but I want one that works properly, and this one doesn't.

Paul C Colby: " The (your) reformulation of the question in terms of time reversibility of micro-physical laws ignores global boundary and initial conditions."

Yes, I agree. However, if we want a specific answer explaining time-asymmetry, which isn't just based on observation ("This is what we see"), it can be useful to start with a set of apparently "bedrock" theorems that seem to _prove_ reversibility, and then study the small print of how those proofs are arrived at, to find how they may fail.

In the case of the simple "billiard-ball physics" model, it may seem intuitive that if you start with a highly defined system (say, the balls arranged to spell a word), and then bounce the balls around, the system becomes progressively more random-looking. However, there's a theorem that says that for a closed system with constant properties, any reversible operation carried out on its contents, repeated often enough, will eventually result in the system returning to its initial state (usually after passing through a huge number of near-copies and reflections or other transformations of the initial state).

So the simple billiard-ball model has no real overall arrow of time.

Similarly, if we replace the balls with an electron-positron pair (or a set of pairs) that annihilate and send gamma-ray waves bouncing around their perfectly-reflecting container, the same theorem seems to say that eventually those gamma rays will refocus in such a way to create the original electron-antielecton pair(s).

Still no arrow of time.

However, if we gradually change the size of the box, the behaviour changes. If the box is gradually enlarging, then the original state cannot be reproduced, and with our single electron-positron pair, the tiny redshift in the gamma-rays as they bounce off the slowly-expanding box means that they no longer have enough energy to reconstitute the original particle pair. In a box with multiple pairs, mutually annihilating and re-forming, we'll initially be only one pair short, but the fact that redshifts are constantly bleeding energy out of the system means that the system can't reach an equilibrium regarding mass-energy, energy-mass reactions, and now has a bias towards mass--> energy.

So an expanding box (or an expanding universe) seems to lead to an entropic arrow of time, in which exothermic reactions are preferred over endothermic ones. Asymmetrical timeflow seems to be a feature of a universe whose most basic properties evolve rather than being static, and in our case, the “arrow of time” appears to be defined as the direction in time in which the universe can be said to show expansion rather than contraction (the expanding box).

Perhaps if we'd studied this question in the late C19th, we might have been able to tentatively suggest the existence of Hubble redshift from purely philosophical /mathematical arguments.

This still isn't a proper answer to the OP's question, but its hopefully building a platform upon which a proper answer might be possible.

Some very interesting and persuasive answers, Eric, but, as you say, not yet quite getting to my original question. I'm interested in your reaction to the following quote from Huw Price in Time's Arrow (loc cit p155): "Perhaps causal asymmetry isn't really in the world at all, but the appearance that it is is a product of our own standpoint. Perhaps it is like the warmth that we see when we look at the world through rose-tinted spectacles"? Price calls this the "perspectival approach". "On the perspectival view, causal asymmetry reflects an asymmetry in us, not an asymmetry in the external world." p161.

My own little paper on this topic (it doesn't go very far) can be read as an attempt to get a grip on the "perspectival view" from a GOFAI point of view -- which may at first sound crazy, but note --  the perspectival view has to talk about "us" (ie the mind) -- hence the need for AI -- as well as "external reality". I still think this is a promising direction to go...

Regards

Jim

@Eric: the theorem you refer to is, of course, Poincarés theorem, and it is entirely rigorous: if you start with a finite number of billard balls, say, all on the left hand side of the box in which they are contained, they will, with probability one, return infinitely often to being all on the left hand side. This seems to contradict irreversibility.

The issue is in the detailed numbers. If the billard blls are relatively small and if there are, say, 1000 billiard balls, then the mean time between the initial time and the first return to a posoition in which all 1000 billiard balls are again in the left half of the container, is about 2^1000 collision times. This, by the way, is not a guess: it follows from another theorem by Mark Kac.

Now let us think molecules. 1000 molecules is quite a small system. If you put it in a cubic micron, it is a quite dilute gas. If, on the other hand, you take collision times typical of a liquid, say a picosecond, to make the collision time as short as  possible, you have a typical recurrence time of 2^1000 ps. This is roughly 3.4*10^281 years. Compared to the age of the universe (about 1.4*10^10 years) it is fairly large. Maybe saying that a recurrence occurs every 10^271 ages of the universe brings some perspective.

So saying ``still no time arrow'' in such a case is only sensible if you wish to center your attention only on what happens on scales of 10^271 ages of the universe, or longer. It is an entirely respectable attitude, what you might call taking the long view.

If, however, you agree that things which occur on that time scale are in some sense not immediately relevant to human concerns, then a time arrow exists: once you prepare a system in a strange configuration, such as the one in which all molecules are on the left side, they will leave the left side and cover the whole box rather uniformly in short order, and recurrences to the state in which all particles return to the left hand side will not happen in ou lifetime.

An entirely similar issue arises with radiation in a cavity. Quite generally speaking, recurrence times grow exponentially with the number of degrees of freedom involved, so even for relatively small systems, recurrences rapidly become academic.

There is, however, another issue to which the answer is not easy: it is an experimental fact that I can only prepare a system in a special state and observe its later development. It is not feasiblee for me to arrange for a system to reach a special state at 15.00 by an intervention at 15.01 and to have observed it betweee 14.00 and 15.00. This is, in a sense, the true basis of irreversibility. I do not claim to understand it, but I believe that once we accept this, we will see that messaging to the past has severe difficulties. I do agree, however, that energy conservation is not an issue.

[REAL!] Mobile Phone Message Details

From: Dad!

Sent:  01/11/2015 19:28

Received:  01/11/2015 19:27

Eat your heart out Sheldon Cooper!