The Arrow of Time in Scientific American

By Sean Carroll | May 21, 2008 4:47 pm

ab230924-fa4d-9eac-5e5e8d5152c227b1_1.jpg Greetings from Paris! Just checking in to do a bit of self-promotion, from which no blog-vacation could possibly keep me. I’ve written an article in this month’s Scientific American about the arrow of time and cosmology. It’s available for free online; the given title is “Does Time Run Backward in Other Universes?”, which wasn’t my choice, but these happenings are team events.

As a teaser, here is a timeline of the history of the universe according to the standard cosmology:

  • Space is empty, featuring nothing but a tiny amount of vacuum energy and an occasional long-wavelength particle formed via fluctuations of the quantum fields that suffuse space.
  • High-intensity radiation suddenly sweeps in from across the universe, in a spherical pattern focused on a point in space. When the radiation collects at that point, a “white hole” is formed.
  • The white hole gradually grows to billions of times the mass of the sun, through accretion of additional radiation of ever decreasing temperature.
  • Other white holes begin to approach from billions of light-years away. They form a homogeneous distribution, all slowly moving toward one another.
  • The white holes begin to lose mass by ejecting gas, dust and radiation into the surrounding environment.
  • The gas and dust occasionally implode to form stars, which spread themselves into galaxies surrounding the white holes.
  • Like the white holes before them, these stars receive inwardly directed radiation. They use the energy from this radiation to convert heavy elements into lighter ones.
  • Stars disperse into gas, which gradually smooths itself out through space; matter as a whole continues to move together and grow more dense.
  • The universe becomes ever hotter and denser, eventually contracting all the way to a big crunch.

Despite appearances, this really is just the standard cosmology, not some fairy tale. I just chose to tell it from the point of view of a time coordinate that is oriented in the opposite direction from the one we usually use. Given that the laws of physics are reversible, this choice is just as legitimate as the usual one; nevertheless, one must admit that the story told this way seems rather unlikely. So why does the universe evolve this way? That’s the big mystery, of course.

  • Ethan

    Right. Because there clearly isn’t any difference in the macroscopic laws of physics with respect to time asymmetry.

    Can you even imagine how one would go about envisioning the laws of physics spontaneously reversing a supernova, for instance?

    Or, to take a simpler example, to un-fry an egg?

    It might be a useful visual tool, to play with “what would happen if we ran time backwards,” but what do we learn from this? Anything physical seems unlikely.

  • Sam

    Going to the lavatory is not going to be a pleasant experience! Shamelessly stolen from Red Dwarf: Backwards, which makes a comedy of this perspective.

  • Brian Mingus

    Sean, that was a great read. Thanks.

  • Xenophage

    Angular momentum is the strong arrow of time. Feynman’s sprinkler spins from emission but not in time-reversed absorption. The situation is not symmetric. A motion picture film can with an orthogonal port at its broad face’s center and another port tangent to its edge is an absolute direction of time detector. Fill with water then pump water center inflow to edge outflow. No problem. Reverse time – (nearly) no flow! Conservation of angular momentum creates a fluid diode with no internal obstructions.

  • Brando

    Ah, but that’s what’s so fun about computational physics – “set Time Scale to -1”

  • Tsee Lee

    What I really would like to see in any article on multiple universes is how to test the theory.

  • TheNerd

    I tried to read this earlier, but as I have ADD, I find the written format to be a very poor one for data absorption. That is always a source of disappointment for me, because I love science. I thought to myself, “if only someone else could read this and summarize it for me in an illustrative way”. You have actually done just that in your timeline above. Thank you!

  • Jo


    You failed to mention Peter Lynds’ model – something which does address the initial conditions problem.

  • Boltzmann’s Reptilian Brain

    “What I really would like to see in any article on multiple universes is how to test the theory.”

    Good point, and it’s long past time to refute people who think that it’s obvious that multiverse theories cannot be tested *in principle*. To be brief: this particular instance of the Popperian delusion is even more ludicrous than most. Just to get y’all started, read these:

    In reality, the problem with multiverse theories is to explain why we *don’t* see evidence of the other universes, why they don’t mess up our universe completely, etc etc etc. The whole “it isn’t science because other universes are not observable even in principle” industry is a load of rubbish.

  • Ronan Mehigan

    Hi Sean

    I enjoyed your Teaching Company Course on Dark Energy and Dark Matter.

    I was reading this recent article from yahoo, “Missing Matter found in Deep Space” and would love to know your opinion


  • Maurizio

    new baby universes as increasing of calculus? what is computing the multiverse? may be intelligent? why all that?

  • Jason Dick

    Boltmann’s Reptilian Brain,

    Well, that is one very specific scenario, though, that of eternal inflation. There are other potential multiverse ideas that don’t predict any such effects. With Sean’s concept, the essential way of testing whether the theory makes any sense would be to show that a galaxy (or some other feature of our universe) is more likely to arise from a big bang like our own instead of fluctuating randomly out of the vacuum. Further testing is, of course, a really messy problem, but is in principle possible.

  • Garth A Barber

    Boltzmann’s Reptilian Brain
    In reality, the problem with multiverse theories is to explain why we *don’t* see evidence of the other universes, why they don’t mess up our universe completely, etc etc etc. The whole “it isn’t science because other universes are not observable even in principle” industry is a load of rubbish.

    So, if for the sake of argument, no other universe other than our own exists, then what test would falsify the multiverse hypothesis?


  • Geoff

    So let me get this straight…. In the past week huge chunks of missing matter have been found and a supernova has been caught exploding. And you’re writing to tell us about an article you wrote?

  • Quasar9

    I can understand the reverse or ‘mirror’ universe
    like the reverse image/universe we see in a mirror.
    But I can’t understand moving backwards in time
    even in the morror one is walking backwards or forward >>> forward in time

  • Khan Muhammad

    @ Sean Carroll,

    I want to translate this article into Bengali Language. So I need your permissio. I tried to get permission to translate an article of SciAm earlier and it was not easy. Finally I translated it without permissio but failed to publish anywhere.

    I want to publish the Bengali translation in a Bengali Space Sciences Magazine name “Mohakash Barta” or “Space News”.

    So, what do you say?

  • GstuatGMU

    Another way to say it is that all particles move in only one (the one we call forward) direction in time, and this causes thermodynamics. The equations describing our microscopic laws are time-symmetric, but they apply in only the one direction. The laws of nature are not time-symmetric, because they include that constraint. Why is that? Why is alpha 1/137, and why is the Weinberg angle what it is? Because Mother Nature says so.

  • Sean

    Khan– you would have to talk directly to the people at Scientific American. I don’t have the rights to the article. You should definitely not translate articles without permission!

  • Qubit

    “Some cosmologists imagine that the universe went through a “bounce.” Before this event, space was contracting, but instead of simply crashing to a point of infinite density, new physical principles—quantum gravity, extra dimensions, string theory or other exotic phenomena—kicked in to save the day at the last minute, and the universe came out the other side into what we now perceive as the big bang. Though intriguing, bouncing cosmologies do not explain the arrow of time. Either entropy was increasing as the prior universe approached the crunch—in which case the arrow of time stretches infinitely far into the past—or the entropy was decreasing, in which case an unnatural low-entropy condition occurred in the middle of the universe’s history (at the bounce). Either way, we have again passed the buck on the question of why the entropy near what we call the big bang was small.”

    “String theory”? How’s that?! Surely you can have quantum gravity, extra dimensions and other exotic phenomena naturally, but with string theory you have to have conciseness or con-science. I really don’t think you made a mistake when you wrote, “string theory” instead of strings.

    I really enjoyed reading your article, that the first time I have enjoyed reading something about science since I read Brian Greene The Elegant Universe: Superstrings, Hidden Dimensions, and the Quest for the Ultimate Theory.

  • http://waitingonthenewm Kudzu Fire

    except for the white holes, it sounds pretty familiar

    not looking forward to the big crunch though…

  • Chris W.

    It would be interesting to flesh out this timeline with some narrative describing biological and cultural evolution, and the lives of individuals, where history and memory play crucial roles, if only to emphasize how much time-reversed scenarios conflict with our understanding of our existence. The apparent indifference of the fundamental laws of physics to such considerations is perhaps the central problem.

  • Joseph Brant

    Boltmann’s Reptilian Brain: “The whole “it isn’t science because other universes are not observable even in principle” industry is a load of rubbish.”

    I’m not sure it’s an industry, but I agree with you.

  • Peter Woit

    Reptile and Joseph,

    There is no industry claiming that “other universes are not observable even in principle”, since it is logically possible to come up with models with such observable effects.

    On the other hand, there are a lot of people who want to see specific arguments about a multiverse make some sort of standard, testable (even if only in principle) scientific prediction before they agree that these specific arguments are science. I don’t see how the arguments being sold in Scientific American lead to any possible predictions that would allow one to test them in any conventional sense of scientific testability.

  • John Merryman

    I’ve raised the point before, but; The only thing particles move through is space. Now with a sea of particles, they are creating a series of configurations, called events. Such as the earth and sun rotating and revolving relative to each other create days and years. While the physical reality goes from past to future events, these events go from being in the future to being in the past. So there are two directions of time implicit in this reality. Content going past to future, as form goes future to past. To the hands of the clock, it’s the face going counterclockwise. Our physical brain moves to the future, as our memory of past events fades into the past.

    We understand there is no objective perspective, but we assume there is such a thing as an objective past, because it is unchanging. To the extent energy is conserved, information of the past is constantly being incorporated into the present, as the energy is recycled. So the past changes because it is being erased and any knowledge of it is due to information we are consuming and adding to. Time is like a rope being woven out of strands pulled from what was previously woven.

    The result is that time has two directions, but no dimension because it is a consequence of motion, not the basis for it.

  • Matt

    Xenophage on May 21st, 2008 at 8:01 pm

    Angular momentum is the strong arrow of time. Feynman’s sprinkler spins from emission but not in time-reversed absorption. The situation is not symmetric. A motion picture film can with an orthogonal port at its broad face’s center and another port tangent to its edge is an absolute direction of time detector. Fill with water then pump water center inflow to edge outflow. No problem. Reverse time – (nearly) no flow! Conservation of angular momentum creates a fluid diode with no internal obstructions.

    I believe the point is more subtle, because simply reversing the pump is not at all equivalent to reversing time. If you were to record the Feynman sprinkler setup and run the tape backwards, nothing unphysical would happen. Like light falling into stars the situation would look very strange but no law of nature is being violated.

  • Brendon Brewer

    The origin of the 2nd law of thermodynamics is no mystery. It’s been solved for decades at least.

  • Brendon Brewer
  • Sam Cox


    A very interesting and thoughtful article! It would be, I think, interesting to conjecture on how many frames of reference in the universe it would be necessary to conceive, assume, describe and understand before we could get an accurate mental pictorial concept of the true nature of cosmological reality.

    Perhaps it would be appropriate to conceive of our “multiverse” as a progressive, eternal and infinite set of universes almost identical to our own, even including ourselves, but gradually and phylogenically developing along a single, single process time dimension, and increasing in informational complexity….

  • Boltzmann’s Reptilian Brain

    Garth Barber said: “So, if for the sake of argument, no other universe other than our own exists, then what test would falsify the multiverse hypothesis?”

    That depends on how the work I mentioned progresses. It could easily happen, for example, that the theory will predict that new universes will nucleate at a rate and collide with ours with such effect that the theory can be ruled out by observations *already* made. Or, in the specific scenario considered by Sean, it may be possible to prove that new universes never look like ours. The point, as also made by Jason D, is that it is just not true that these theories are non-falsifiable *in principle*.

    Peter Woit said: “On the other hand, there are a lot of people who want to see specific arguments about a multiverse make some sort of standard, testable (even if only in principle) scientific prediction before they agree that these specific arguments are science.

    Look at the articles I cited. Seeing the patterns in the CMB described there would of course be very difficult — but you have no complaints about that, right?

    “I don’t see how the arguments being sold in Scientific American lead to any possible predictions that would allow one to test them in any conventional sense of scientific testability.”

    Exactly what people said — on better grounds — 25 years ago about Inflation. If I had been around 25 years ago, I would have been prepared to bet serious money that nothing observable would come out of such a theory within 25 years. And I would have been wrong.

    There is nothing *in principle* unverifiable about Sean’s theory. One universe is born from another. Personally I would bet that a deeper investigation of this theory will show [sorry SC !] that it is actually not able to produce a universe like ours, except by some kind of Boltzmann-style massive fluctuation. Thinking about Boltzmann-style fluctuations has convinced most of us that it is not what happened. So at that point I would conclude that SC’s theory has been proved false. If that is not science then what is?

    In short, nothing special is going on here philosophically. We have theories that posit *currently* unobservable things, just like any speculative theory. So what?

    Anyway you might look again at Sean’s comprehensive demolition of John Horgan a couple of months back. He said it all better than I can.

  • Ed

    I don’t see how anyone wouldn’t see how much more believable the standard cosmological model is than to hypothesize an “original intelligence” which created everything.

  • pete

    @ Ed,
    Just because we are yet to understand something and at the present don`t have the technology to do that, it does`t mean that we shouldnt try….

    Quasar9 on May 22nd, 2008 at 12:31 pm

    I can understand the reverse or ‘mirror’ universe
    like the reverse image/universe we see in a mirror.
    But I can’t understand moving backwards in time
    even in the morror one is walking backwards or forward >>> forward in time

    Good point.
    Can anyone help me imagining time going backwards?? If I reverse processes, I can only visualise it (the reversed process) as a going forward…..

  • manyoso

    “You should definitely not translate articles without permission!”


    Might as well be, “Quit thinking!”

    You can’t read the article if english is your second language and you translate it first?

  • Ian

    Just got the new issue of SciAm in the mail yesterday. Flipped to Sean’s article and started to read it out loud in the car while my 16 year old was driving to school this morning, thinking she might find it interesting (and having my eyes averted from the road keeps me from thinking about all the various possibilities of increasing entropy that she, as a beginning driver, narrowly avoids).

    Unfortunately, I was unable to finish the article because, arriving at school, she asked to borrow it to share with her physics teacher today. The opening paragraphs completely grabbed her! This is the kind of stuff that gets her really excited and has her wanting to pursue physics/astronomy/cosmology in college in a couple years (She recently got the opportunity to spend the afternoon with your colleague Mike Brown at CalTech which she loved!).

    So, thanks, Sean, for putting out this “high school-accessible” article. You definitely set the hook in at least one young reader! I look forward to finishing the article for myself tonight.

  • George Musser

    Khan, I think it’s great that you’ve offered to translate the article into Bengali. Send me an email offline and I’ll help you get the necessary permission. My email address is my first initial plus last name at sciam dot com.

    manyoso, the problem isn’t the translation but the publishing! But actually I doubt there will be any problem at all.


  • John Merryman


    Good point.
    Can anyone help me imagining time going backwards?? If I reverse processes, I can only visualise it (the reversed process) as a going forward…..

    Here is a way to visualize time as going from the future into the past.
    Rather then viewing time as a directional dimension along which we move from past to future, consider just a bunch of quantum energy that simply exists, with no time dimension at all. Suppose this energy developed the tendency to clump, then break apart and scatter, then clump again.
    What this energy has done is to create a time dimension of events. Now since this dimension isn’t fundamental, but is created by the process of clumping and scattering, the quantum energy isn’t moving along it, but is creating it as each successive stage replaces the previous one. Since the only thing that exists is the energy, these stages go from being present to being past. The result is that this timeline is going from future potential, to present manifestation to past circumstance. This is one way to imagine time as going from future to past.
    Sort of like tomorrow becomes yesterday.

  • Sandy

    “Among the unnatural aspects of the universe, one stands out: time asymmetry. The microscopic laws of physics that underlie the behavior of the universe do not distinguish between past and future, yet the early universe—hot, dense, homogeneous—is completely different from today’s—cool, dilute, lumpy.”

    Is it really that surprising that equations do not fully describe reality? A multiverse could be the difference in behavior/description at different scales within this universe.

  • Ed

    Pete, I wholeheartedly, support and commend the activity of increasing our knowledge about our current and past environment (universe). That was not my point. I was pointing out that there isn’t much real difference in saying “in the beginning was void” and saying in the beginning was a tiny amount of vacuum energy.

  • Pete

    Ed, thanks for explaining it, my mistake. :)


    Sort of like tomorrow becomes yesterday.

    So, what is going to happen has already happened and therefore isnt going to happen at all? Or am I approaching this in a wrong way?
    This has twisted my brain.

  • Robert K. Vanderhoek

    Thought I just posted something, got a thank you, but, now I don’t see it.


  • John Merryman


    So, what is going to happen has already happened and therefore isnt going to happen at all? Or am I approaching this in a wrong way?
    This has twisted my brain.

    I’m just making the point that the one dimensional, linear cause and effect narrative that we call time is better explained as a consequence of motion, similar to temperature, rather than the basis of it.

    As a measure of motion, the concept of a specific dimensionless point in time is meaningless, as there would be no motion to measure, so nothing measured in time can be said to have an exact location. Reality really is fuzzy.

  • Doug Robertson


    I liked your Scientific American article, but it seems to me that you
    have neglected a significant component of the “time’s arrow” problem,
    the part that some philosophers call the “now” problem.

    Your article focused mainly on questions related to why the “beginning”
    of the universe has more or less entropy than the other end, which is an
    inherently interesting and non-trivial problem. But it misses the vital
    question of what it is that “moves” under time’s arrow. The simple
    answer is that what moves is an odd moment of time that we call “now.”

    The problem of course is that there is nothing in physics that specifies
    any particular instant of time as special in any way, and in particular,
    nothing that specifies an instant called “now” that separates a
    determined past from an indeterminate future. This might not be the
    best definition of “now,” in that it opens up a set of difficult problems
    about determinism that we don’t need just yet.

    To sidestep these questions (for now), perhaps a better definition of
    “now” is this: the moment of time that separates the time intervals that
    all of us can remember from those intervals than none of us can
    remember. At first we should probably assume that “now” is the same
    moment for all of us, at least to some (poorly specified) accuracy. We
    could define a separate “now” for each individual, but I don’t think we
    need to do that just yet. We get into serious philosophical difficulties if
    we assume that “now” for any one of us is separated by a “large” amount
    of time from that for the rest of us, where “large” means something like
    significantly greater than the light-travel-time between observers.

    You touch on this question when you ask: “Why do we remember the
    past but not the future?” You state that: “to form a reliable memory
    requires that the past be orderly, that is, have a low entropy.” This
    answer seems completely inadequate. Tomorrow may have a slightly
    higher entropy than today, but not by an amount that explains the
    radical difference in the behavior of memory over the two intervals. In
    general, the difference in entropy in my surroundings between yesterday
    and today isn’t very much different than the difference between
    today and tomorrow. But there is a radical difference in what I can
    remember about these days.

    This raises another difficulty. Clearly there are large differences in the
    change of entropy of our surroundings from day to day, as in the case,
    say, of my house burning down compared to a more normal diurnal
    interval. But I personally do not notice much difference in my memory
    abilities, nothing that correlates with the magnitude of the change in
    entropy from day to day.

    It seems to me that a complete understanding of “time’s arrow” would
    require an understanding of the “now” problem. It may even be the
    more fundamental part of the problem, in that a theoretical
    understanding of “now” might entail a full understanding of time’s arrow,
    but an understanding of the entropy differences across the universe’s
    lifetime would not appear to require an understanding of “now.”

    Some philosophers, called “nowists,” attempt to resolve this difficulty by
    assuming that “now” is the only instant of time that exists. This strikes
    me as simply defining away the problem without explaining it. And it
    has clear difficulties with special relativity. But I don’t have a
    particularly better explanation.

    Your very use of the words “past” and “future” assume the existence of
    a “now” moment, but you do not deal with this problem at all.

  • W. Arfarin

    Sean –

    Heh, cool, I obviously like your timeline with respect to large
    scale reverse cosmology
    . :-) Within our conventional
    direction of time, I think you probably wrote your version first,
    given how gaps between publication deadlines and “ink on
    paper” usually work. I promise I have not hacked
    into your ‘puter or travelled from your future after
    having read your SciAm timeline!

    A number of minutes ago I began typing a handwave about the weirdnesses of the two boundary conditions from the perspective of observers like us versus observers who see the evolution of the cosmos in reverse. This was derailed by two things, viz. conclusions that can be drawn about cosmic evolution in a Big Crunch direction, and whether reasoning about that end is really easier in either direction (forward physics or backwards physics) given a similar need for extrapolations in constucting timelines that include both boundary conditions.

    It’s going to take a long hot shower to think about some of this.

    One of the things that hit me was what the Hubble flow looks like to a reverse observer rather than a forwards observer who can simply play the “tape” backwards.

    (Also there is another plausible case that is both helpful and distracting, namely, a pair of observers who are in a universe much like ours except that both boundary conditions are strongly predicted to be hot and dense. One observer looks “forwards” and one looks “backwards”. The differences between what either observer sees after his peculiar expansion-collapse transition and what the other observer sees after her peculiar expansion-collapse transsition is worth a thought.)

    The second thing to hit me involved my immediate reaction was that the boundary condition of Big Collapse is easier to think about than either starting condition (Big Bang or “Horizon Decay”) because we only have to consider a couple oddities:

    A. changes in the metric collapse of spacetime — e.g. it may be slowing down towards the big crunch, rather than proceeding at an even, “inertial” rate. (Here’s one place where backwardsverse!observer broke my brain, since I really get stuck thinking that their metric collapse of spacetime looks to them a lot like what we in forwardsverse think of as gravitational collapse).

    B. it is not clear how the metric collapse can squish everything together to match the CMBR — we need a lot of energy to dissociate baryons (as we know from watching stellar reductions), but the energy from the metric collapse force seems insufficient, especially with respect to singularity models of the Big Collapse.

    In point (B) I wanted to predict “cosmic deflation” but my brain bent less because of mechanism (just play it backwards) than because of what could suggest it to any of our Big Crunch predictors.

    I also am tired and easily distracted by trying to add in things like possible quintessence, predictions from QM (dehadronization) and the like. You can have a lot of fun running various models backwards and forwards! Maybe everyone should do this out of habit if “Given that the laws of physics are reversible” is true.

    Penultimately, I wanted to avoid dragging everyone into long debates about the nature of reason (I’m in the “it’s an emergent property of brain structure/function/mechanics” camp), since that doesn’t obviously help with the thinking about boundary conditions, but is one of those interesting distractions. :-) This started getting very hard when thinking about how our backwards counterparts would perceive CMBR, for example.

    Finally, notch up another commenter who likes before-bedtime activities, even if he is a total outlier in your timezone-to-posting-hour distribution (UTC+1).

    Your friendly neighbourhood rodent toxin of many names.

  • W. Arfarin

    3 Meta points:

    1. Add me to the list of people who wish for a preview button!

    2. I obviously did not foresee the mangling of the link to my previous comment, but I figure you are smart and can figure it out. 😀

    3. As one way to sidestep the problems of observation and communication involving people with incompatible arrows of time, how about pretending there exists a translation system that lets you write down the observations and reasonings of a “foreign” (in the evolution of the universe sense) cosmologist acting just as his/her/it’s agent? I wish I could find a more concise and academic description than here:

    in which it is argued that some proper scientists have employed the device at least when writing hard science fiction.

    I was using it earlier without really thinking about it or acknowledging it.

    Perhaps someone will have a “conversation” with a scientist in backwards-universe and report back details? Or at least bury the use of such a reasoning device behind something more conventional and rigorous, along the lines of what you might find on arXiv or at least main postings on

    Finally: Warning: will ruin your life, or annoy you, or both. It says so itself on the front page. Apologies in advance.

  • Anchor

    Here’s a Sci Am article on the arrow of time that is interesting on quite a number of levels:

    One of the reasons why it’s interesting is because it’s 1/3 century old. Another is that it doesn’t treat it’s readers as if they are terminally afflicted with ADD.

    No offence, Sean, but I wistfully long for the days when Sci Am published elegant articles that served up some actual meat in them, “articles” that didn’t resemble a television commercial with eye-catching and monumentally irrelevant cgi artwork. Then one looks “under the hood” and finds watered-down soup that looks as if it was hammered out over a weekend rather than the hoped-for beefy stew. I spent as much time – maybe ten minutes total – admiring the opening art as it took me to read the article, which provides almost no further elaboration of the concepts that isn’t already encapsulated by the 3 so-called “infographic” boxes. Unfortunately, the article fails miserably in terms of actually learning anything new.

    Just how far can they possibly push this “simplification” gambit before anyone notices that the “articles” are basically appetizers for what readers REALLY want in a magazine subscription – you know, like a satisfying meal?

    Though they can still work up a fair sweat with biological topics, in terms of physics and astronomy articles, Sci Am has become a dreadful waste of the precious resource of paper. Looking over how the magazine has diluted the information content in it’s articles over the last 20 years or so, it’s apparent that they’ve finally settled on identifying their market: moderately bright grade-schoolers who have the attention-span of a gnat.

  • John Merryman


    The entropy problem would go away if cosmology wasn’t trying to define the universe as a sungular unit with an internal timeline that goes from beginning to end, as its external position goes from being in the future to being in the past. If it’s infinite there is universal energy exchange and all energy is conserved.

    “Now” makes much more sense if we consider that only energy in space exists and time is a consequence of measuring change. Therefore, as I previously mentioned, “now” cannot be a dimensionless point, as there would be no motion to measure.

    W. Arfarin,

    One of the things that hit me was what the Hubble flow looks like to a reverse observer rather than a forwards observer who can simply play the “tape” backwards.

    (Here’s one place where backwardsverse!observer broke my brain, since I really get stuck thinking that their metric collapse of spacetime looks to them a lot like what we in forwardsverse think of as gravitational collapse).

    B. it is not clear how the metric collapse can squish everything together to match the CMBR — we need a lot of energy to dissociate baryons (as we know from watching stellar reductions), but the energy from the metric collapse force seems insufficient, especially with respect to singularity models of the Big Collapse.

    If you think of energy expansion as the source of redshift, it is the hands of the clock of constituent energy going from one temporal unit to the next, as as gravitational structure is the face of the clock, with its units of measure going from being in the future to being in the past. That way, you have a convective cycle of expanding energy and collapsing mass, rather than a universe as a singular unit of time.


    Maybe the material they have to work with is at fault.

  • Sandy

    The apparent symmetry of time in physics could well be due to the inability of equations to fully capture and describe time. The convention is just that, a convention.

  • Brody Facoum

    I hope I am not being annoying…

    John Merryman –

    No, my thinking rathole about the Hubble flow, specifically from
    the point of view of an observer who sees *this* universe (more precisely, our Hubble volume) in the
    opposite time-like direction from us, went like this:

    1. Can a biology arise in reversed standard cosmology?

    (a) We can guess what it looks like — the (generic) luciferase
    process looks a lot like the chloroplast membrane proton gradient-ATP Synthase process in reverse!universe: photons are
    absorbed and used to phosphorylate AMP; the resulting ATP
    can do work (e.g. breaking apart polymers in ribosomes,
    synthesizing carbohydrates in mitochondria,
    bioluminescing in chloroplasts…)
    (b) Origin of life problem seems pretty damn hard, but it’s
    not exactly easy in forwards universe
    (c) Adaptation and natural selection is a bit
    confounding in reverse, especially if we look at it
    forwardsly — it seems like a spectacularly
    unsuccessful species whose population dwindles rapidly
    towards zero would be well adapted to a niche in reverse

    2. Can a biology in reverse standard cosmology result in a
    brain that reasons like our brains?

    I have nothing other than resorting to the law of large
    numbers, which as we know produces Boltzmann Brain
    problems in forwardsverse; my poor brain needs some
    pencils and paper to think about it backwards. My “yes”
    is purely intuitive.

    3. “Seeing”

    An obvious development in sensory equipment comes down to
    “sourcing” photons from biomechanical processes that emit
    them with a nonthermal distribution that can be tracked
    from point of emission. (No feedback in backwardsverse is
    an assumption to help preserve Galilean invariance and
    Lorentz covariance).

    In backwardsverse we would want to evolve strucuress which emit
    photons in such a way that the resulting work preserves
    (“encodes”) very fine grained angular distribution
    information about the photons (as well as their energies).
    “Seeing” reversed: photons stream out of phototransmissors
    in all sorts of directions with anisotropies at different
    frequencies. At point of emission we can preserve
    information about the direction of distant photon sinks.

    Using this to examine the sky we would see large areas
    which “suck” only low numbers of low frequency low
    temperature photons from our photoemissors. We would see
    quite a few pointlike and smearlike objects which would
    “suck” out larger numbers of much higher energy photons,
    and so forth. A hot blackbody sinker has a well defined
    photon sucking behaviour in terms of photon frequencies;
    many other hot objects (especially in the sky and in labs)
    do not take a full blackbody spectrum. Doppler effects
    shift these spectral lines towards higher frequencies or
    lower frequencies depending on the relative motions of the
    observer and distant photon attractor.

    Rotating and elliptical galaxies and huge stellar implosions are pretty
    uniformly distributed across the sky. When we send out
    photons to them we notice a correlation between decreasing
    angle and increasing blueshift, corresponding to a metric
    collapse of space time of approximately 71 km/s/Mpc.

    When we send out very blueshifted galactic spectra to a
    narrow angle of sky we are really transmitting photons
    into the distant past, where they will ultimately meet up
    with their appropriate photon sinks, in line with what we
    know about the second law of thermodynamics.
    Unfortunately we cannot say much about the present state
    of the sinks of these distant photons, except that they
    are probably further away from us now (and probably more
    filled with metal) than is suggested by the outgoing spectra.

    4. This is far from rigorous and hinges on forward
    reasoning informed by backwards perception (i.e., something that thinks like us but which we would probably guess was thinking backwards — “remembering the future” when it is really remembering its own past).

    This is not forbidden by an emergentist view of reasoning, but I have no idea how likely or unlikely it could be. Is it more or less likely than our own reasoning?

    This point also lends itself to metaphysical thinking which,
    frankly, annoys me.

    Lastly, with respect to your paragraph staring with “if
    you think of energy expansion as the source of

    I don’t want to do that; that’s not an approach that
    is interesting to me. I am not trying to disprove the
    concordance model AT ALL, I am trying to better understand
    it by inverting the timelike direction of observation with
    a few assumptions (reversibility of physical processes, fundamentally).

    As a result, I am (trying to be) thinking of the Hubble
    flow in standard cosmology terms (without being
    particularly rigorous), just in a timelike direction that
    leads to the Big Crunch (i.e. what we call the Big Bang).

    That is, the “clock” direction in this is with respect to
    the boundary conditions, particularly the one which is
    strongly predicted (our hot dense early universe). The
    clock metric favoured in Sean’s threads is entropy
    (specifically the log relation of microstates to
    macrostates), which increases in one direction and
    decreases in the other.

    Trying to predict the observations of a human-like
    reasoner who happens to be used to the reverse direction
    (i.e., entropy decreases over time) is hard, particularly
    since I am scratching my head about things like
    thermodynamic favourability versus forward logic, and
    whether when we flip the thermodynamic arrow we also flip
    *whether* things like electron transport chains, osmosis,
    exclusive or logic gates, and so forth *can* work, rather
    than *how* they could work.

    Particularly with respect to this rathole, I’m
    not especially keen on considering possible states outside
    the Hubble volume or beyond the BB/BR (and even the Big
    Rip (or Long Asymptotic Isolation) end is a bit fuzzy)
    without good reason.

    So, to reiterate, thinking about time reversal like this is at best an exercise that might be useful to explore the standard cosmology, and at worst is Anchor’s “dreadful waste of [some] precious resource”. Science-shattering? No. Interesting? I think so, otherwise I’d be doing something else with this. Useful? I dunno. Ask Sean Carroll. 😀

    – Your friendly neighbourhood etc.

  • Neil

    I wonder if the answer to this might not be hiding in plain sight. E=MC(Squared). A useful formulation that even Einstein had to tinker with, even if just a little. Our physics have no explanation for what happens past the “speed” of light. On the other hand our physics, certainly on our own little macro level, depends upon a formula that includes 3 elements, one of which clearly does not exist — the speed of light squared. What might energy, matter, time, or space be, or would they be, in a multiverse where one “sector” was ordered on laws of physics where the speed of light squared was the rule rather than the exception?

  • John Merryman


    My problem is the path down this particular hole started with a question about the expanding universe. Many years ago I happened to read that the sum of total expansion is very close to, if not evenly balanced by the contraction of gravity. So I’ve raised the issue of how can the entire universe be expanding, if this spatial expansion on which the assumption is based is being effectively neutralized by gravity. Now I can certainly understand why you wouldn’t want to question BBT, as it classifies one as a crackpot to do so.
    On the other hand, I’m not a scientist and in fact my main interests tend toward politics and history and I only started reading up on physics as a way to gain insight as to what motivates the larger forces in society. So when I come across bunches of people all marching in the same direction because the best and brightest told them it’s the way to go, even though the logic is incomplete, incoherent and full of patches, my impression is decidedly less then credulous. When the adherents try an enlighten me by by pointing out that it’s ordained by a higher power, in this case, mathematics and I’m just too simple to understand, I don’t necessarily assume I’m just too thick to appreciate that we are living in a new era, but tend to consider the crowd might just be falling for another version of the same old song and dance. As far as I’m concerned, equations are another form of smoke and mirrors when those promoting them cannot explain what they mean in another language. In this case, english.

    As for cosmology and time, if gravitational collapse is an inward curvature of space, than I am of the opinion the radiation and other forms of energy being ejected from gravitational wells amounts to an outward curvature of space and redshift is evidence of this cosmological constant. The result being an old fashioned convective cycle of collapsing mass that heats up and starts expanding, until it cools sufficiently to start the cycle over. As structure, mass is the face of the clock, with its units starting as future potential and ending up as past circumstance, while the outward flow of radiation/energy is the hands of the clock, the raw essence that is constantly shedding/radiating away from old forms and starting new ones. It is both of these directions of time coexisting, the outward expansion into the future and the inward collapse of the past, that forms this infinitely dimensional reality we live in.

    As for math, it is a model, not an ideal.

  • Boltzmann’s Reptilian Brain

    Anchor sneered: “No offence, Sean, but I wistfully long for the days when Sci Am published elegant articles that served up some actual meat in them, “articles” that didn’t resemble a television commercial with eye-catching and monumentally irrelevant cgi artwork.”

    Yeah, well, Anchor, “no offence”, but you are neglecting one tiny point: that article from 1975 is wrong, though at least it correctly identifies cosmology as the source of the Arrow — incredibly, there are still some people who dispute this obvious fact. Nevertheless the solution proposed is indeed wrong. Here’s a free clue: the theory advanced there sank without a trace.

    Sean’s article may seem simplistic, but that’s the point: what he is trying to draw to your attention is something which is indeed amazingly obvious, yet people —- I mean professional physicists, not just crackpots like Last Mohican —- just don’t get it. Quite recently I had a discussion with a distinguished physicist. I asked him what he thought about the old idea that the arrow might reverse if the universe began to contract. His response: “that’s ridiculous, it would require incredible correlations, etc etc etc” But he didn’t think that there was anything ridiculous about the fantastic correlations that you see if you trace the history of our universe back to the big bang! So you see that even the simplest errors in this subject live on, and Sean is doing great work in alerting everyone — not just the public — to those errors.

    ps. No, I don’t believe that the arrow will reverse if the universe contracts. The point is that the history we observe is just as bizarre as that would be, and cries out for an explanation. Meta-point: there is obviously some major gap in our understanding of the early universe. If we can’t understand this, then we are probably getting lots of other things wrong too.

  • Boltzmann’s Reptilian Brain

    By the way, I recently found this very nice quote from no less a personage than Brian Greene [Fabric of the Cosmos, page 164]

    “High entropy is the natural state of being. You should never be surprised by or feel the need to explain why any physical system is in a high-entropy state.”

    From this it follows, of course, that the natural way for the universe to be born was in a high-entropy state. But it wasn’t……..

  • James Nightshade

    This whole discussion — the arrow of time reversed, galaxies formed out of white holes, eggs spontaneously unscrambled in a way that respects chicken anatomy and the (d)evolution of life on earth — reminds me of a TV cartoon show in which some people try eating with the wrong end of their digestive tract.

  • jeff


    Most objective physicists prefer to ignore the concept of “now” because it cannot be defined without referencing your own consciousness (that doesn’t necessarily mean they’re right). There is no other other standard to determine what “now” is (what would “now” be when you’re dead?). Even Einstein did not think that a theory of the present was ever possible – relativity is a theory of possible “nows”, but not an actual “now”. And since “now” divides the past and the future, then the past and the future can also not be defined without referencing your own consciousness. Personally, I think they may be making a big mistake, but I don’t know how else they could proceed, objectively. “Now” is fundamental to time and reality. Something about “now” and consciousness are inextricably intertwined – more so than “here”, IMO.

  • jeffw


    Most objective physicists prefer to ignore the concept of “now” because it cannot be defined without referencing your own consciousness (that doesn’t necessarily mean they’re right). There is no other other standard to determine what “now” is (what would “now” be when you’re dead?). Even Einstein did not think that a theory of the present was ever possible. Relativity is a theory of possible “nows”, but not an actual “now”. And since “now” divides the past and the future, then the past and the future can also not be defined without referencing your own consciousness. Personally, I think physicists may be ignoring something important, but I don’t know how else they could proceed objectively. “Now” is fundamental to time and reality. Something about “now” and consciousness are inextricably intertwined – more so than “here”, IMO.

  • Jason

    Here’s the thing I’m not quite getting about Sean’s argument, though. He seems to be drawing no distinction between the microstates possible in the early universe and those possible today. In fact, he explicitly states that there is no such distinction, that the number of microstates available simply doesn’t change.

    I don’t understand how this can be possible. I mean, if we consider a universe with nothing but a cosmological constant (a fair approximation to the earliest epoch of our universe of which we are aware), then what lies within a cosmological constant is essentially constant with time. All descriptions of that space just don’t change as time goes on. So it stands to reason, I should think, that the number of possible microstates within a given cosmological horizon also do not change with time as this region of space expands.

    However, it is expanding. So though the possible microstates within a region are not changing, the total volume that expands from a particular region should have a number of possible microstates that increase with volume. So would not the number of possible microstates necessarily increase with the expansion of the universe? Would this not mean that inflation has at least a role to play in explaining why the entropy of our early universe was so low? I mean, by allowing a universe such as our own to come from so small a state should mean that the earliest part of our universe, though much lower in entropy than today, was also lower in its possible maximum entropy, due to the smaller volume. Surely that has to be part of the answer, unless I’m just not understanding something.

  • Jason Dick

    Oops, typo. Second paragraph, second sentence should end in:

    “then what lies within a cosmological horizon is essentially constant with time.

  • Boltzmann’s Reptilian Brain

    JD, I thnk you should look at SC’s paper with Chen. Look at the section about unitarity. Basically the size of a region has nothing to do with the number of degrees of freedom it contains. What changes is not the number of degrees of freedom, but rather the number of degrees that are excited out of their ground state.

  • Sam Cox

    James N said,

    “This whole discussion — the arrow of time reversed, galaxies formed out of white holes, eggs spontaneously unscrambled in a way that respects chicken anatomy and the (d)evolution of life on earth — reminds me of a TV cartoon show in which some people try eating with the wrong end of their digestive tract.”

    Hi, Enjoyed your comments very much! Geometric inversion is a very interesting phenomenon, and would not at all necessarily result in a universe of the ludicrous kind we would imagine from our frame of reference…observing particulate 4D event horizon surfaces at the edge of the singular.

    “galaxies formed out of white holes”…seems strange, but what are we investigating when we analyze the CBM? The existence of a “big bang” and our origins from same are experimentally valid conclusions we can draw from the universe we observe.

    When we move from culture to culture we are stunned by the incongruity of what seems “normal” to other folks.

    Eggs unscrambling seems crazy from our frame. So does doing our business and eating in reverse. An advantage of such a universe is that it is “safe”. A disadvantage is that we become more childlike.

    It took some time for us to adjust to these present conditions…we needed mentoring to survive. We might not need mentoring on the other side, in fact we wouldn’t need mentoring on the other side, but observing a 4D cross section of the universe in decreasing thermal and increasing informational entropy sure would take some “getting used to”!

  • John R Ramsden

    I thought the article was a great read, and it cleared up a few points for me while introducing some perplexing new ones (such as a distinction between virtual particles and those which fluctuate in and out of brief existence – Huh?! I thought these were one and the same, but evidently not).

    Perhaps experts critical of speculation beyond the realms of experiment (for now) should remember that the average SA reader, especially keen young future scientists, are inspired by mysteries and reassured that not everything is known, as the average college textbook would have us believe. Why bother pursuing a career in science if they can’t contribute? Isn’t that why those experts started?

    Clearly the central mystery Sean addressed was how the Universe can start in a state of low entropy when the number of microstates remains unchanged over time.

    In the context or “framing” of pre-bang evolution, which is widely accepted these days not least by Sean as he explains in the article, the only way I can make sense of the low initial entropy enigma is that the Big Bang marks a qualitative change in the types of degrees of freedom which come to the fore.

    In other words, after settling into a maximum entropy state, by some means the microstates are “folded up” in regular formation so that a low-entropy dual set of microstates of comparable cardinality “comes into focus” in a sense and takes over.

    For example, when a gas freezes into a crystalline solid, the molecules become locked into place in the lattice. But phonons and vibrations then become a new kind of dynamics in their own right, whereas in the gas their existence was only immanent (and no, I didn’t mis-spell “imminent”!)

    As a more commonplace example, consider a squad of grenadiers, or National Guards if you’re in the US, performing a drill. One common maneuvre is for alternate lines of a rectangular squad to march about and back through the complementary lines. Looks a bit messy for a while, but when the lines are re-aligned the squad can halt, all do a left or right turn, and march off as a unit at right angles to their former direction.

    I suggest the causally-connected universe can achieve an equivalent effect in the distant future, if we assume it eventually starts contracting. When the blue-shifted random background energy reaches a critical black-hole density, time in its previous measure will simply cease (in that a fictitious observer at any given point would perceive everywhere else fade away to nothing, so the local temperature would also become everywhere zero).

    At that point (God knows how, but probably by something like a Fourier transform) time and temperature get switched, as do mass and energy, and the emergent dual universe starts inflating fast even as the old contracts slowly – Everything is ass over tit, or inverted, which makes it uncommonly hard to gain a clear intuition as to how it works.

    Note that the new universe doesn’t “bud off” the old. It is inherent in the old, which still in a sense exists, like those clever pictures that show a bowl of flowers and at the same time an old guy’s face. That’s what duality means – an alternative consistent view of the same whole.

    What’s more, if we also assume this scenario is periodic, so there’s a kind of ping-pong alternating dual evolution analogous to EM radiation, then perhaps it could be verified *experimentally*. It may be that periodicity imposes some constraints on physical constants such as the Fine Structure constant.

    Apologies to experts patient enough to skim beyond the first few lines of the above if any or all of it sounds nonsensical – You may need to read between the lines, and use your imagination, to guess how I *should* have expressed it!

    As St Augustine wrote: “The new is in the old concealed. The old is in the new revealed”. In summary, I submit that to understand the beginning, it will greatly help if one understands the end ..

  • Khan Muhammad

    @ George Musser:

    Many many thanks. I am sending you a mail.

  • John Merryman

    Why suppose it takes separate universes to manifest these two directions, if they can be explained in terms of the inherent dualities of this one? Here’s a supposition; What if the philosophic basis of modern cosmolgy and physics was eastern dualism, rather than western monism. Would we be living different paradigms?

  • Plato

    Sean, it’s nice to see you are carrying on the tradition.

    Incompatible Arrows, I: Martin Amis
    Incompatible Arrows, II: Kurt Vonnegut
    Incompatible Arrows, III: Lewis Carroll
    Incompatible Arrows, IV: F. Scott Fitzgerald

    Also, there is a comparative view that one might think about in terms of the “valleys” and what genus figures are resting there, as we look at the “shape of space” and the relations you point out with regards to the Dark energy in relation to the universe? Mandelstam has pushed the boundaries of interpretation, and Jacques Distler in relation to Lee Smolin, can probably attest to that?:)

    Please note the “Not even wrong index” as to the “in the box thinking” versus the “out of the box” thinking?:) This is not a reflection on any individual character, but an assessment of what comes as a result, to a coordinated frame of reference.:)

  • Daryl McCullough

    Boltzmann’s Reptilian Brain,

    Can you say a little more about what was wrong in the 1975 article?

  • ST

    Am I the only one who finds that the debate over priors vs. dynamics is purely based on prejudice? Of course, to make progress in science we need prejudices about the unknown and hard solid work stemming from that prejudice, but to fight so vehemently about the metaphysics of it at this stage …

  • ike

    No, that seems to be sensible, ST.

    Let’s say we take our physical system and measure its variables with great accuracy. Maybe our star, the sun. Can we recover all the information about its past states? Not even slightly. Can one recover the past history of solar sunspot distribution from a single high-resolution snapshot of the sun? Not any more than one could recover the history of all storms on the Earth’s oceans from a single snapshot from today.

    Big Bang cosmology is a very special case our universe – we can indeed recover much information about the past state of our universe by looking at things like the cosmic microwave background radiation. This is very unusual, though – most information about the past states is irretrievable.

    Take the example of the history of an nitrogen atom. We know that at some point, the nitrogen atom was formed via nuclear fusion within a star – or it could have been formed due to radioactive decay of 14-Carbon in our atmosphere. If we were to isolate a single N atom, we’d have no way of reconstructing its past history. Similarly, if we put 2 N atoms side by side and track them into the future, their paths would soon diverge.

    Let’s say one of those N atoms gets swept up into a DNA molecule. It then forms the informational basis of a local entropy-defeating energy-dissipating dynamical system – the template. If we track DNA sequences forward through time, do we see “progress”? Well, what you do see is greater and greater informational complexity over time – with no particular aim at all other than protecting the informational template. Again, if we take a snapshot of all life on earth today, we can show that all life is inter-related, and that all life evolved from free-living single-celled creatures – but we lose most historical information.

    In this regard, life is the ultimate pragmatist. Whichever way the arrow of time is running, that’s the direction that life works with. Thus, if you are ever trapped in a mysterious dynamical system, and can’t tell which way time is running, just look at any living creature, and you’ll be able to figure it out. :)

  • Danny

    Of course this is a legitimate way to tell the story, given that the laws of physics are reversible. But, Universe doesn’t evolve this way. We see eternal expansion, not big Crunch. The real question would be, by my opinion – why the Universe doesn’t evolve this way?

  • Anchor

    Boltzmann’s Reptilian Brain: Hindsight is so very effective, especially fashionable with those who are expert at sneering. I’m not the one who sneered and I didn’t neglect anything worth considering. I just like to have seen Sean’s ideas presented with more detail that I haven’t already encountered numerous times elsewhere. I DO think Sean does good work and I do NOT fault him for the shortcomings of the article. I fault the editors and management of Sci Am, who generally present ALL physics articles as simplistically.

    But it seems you’ve missed my point entirely. Whether the ideas expressed in the 1975 article have withstood the test of time is utterly irrelevant. (Sci Am itself links to it via Sean’s article). One wonders what your honest “opinion” on the matter was back in 1975 when that article appeared… or had you been born yet?

    All I said was that Sci Am articles from 25+ years ago went into a level of detail which they come nowhere near approaching today. They really WERE better ARTICLES back then. I know this because I’ve subscribed to the magazine since 1964, and I’ve read every back issue extending back to 1900. The magazine peaked during the 50’s, 60’s and ’70’s. Since about the middle of the 80’s the content has slacked increasingly to the current pamphlet which delivers the published equivalent of sound bites. One wonders if the primary motivation for featuring occasional sexy (and “quickie”) articles on black holes and cosmology is to pump circulation up and draw readers to the advertisements.

    John Merryman: “Maybe the material they have to work with is at fault.”

    I suppose it’s conceivable, but I’m afraid I doubt that hypothesis. The decline in the magazine during the 80’s came in lock-step with a paper crisis and a general down-sizing free-for-all that became increasingly popular over the entire corporate sector back then. The magazine has never recovered from it.

  • John Merryman


    Gravity is the crunching of space and according to the most accurate measurements, those opposing factors are balanced, resulting in overall flat space, so the universe is doing both.
    The problem is that the assumption is the expansion is due to an original singularity, and all other galaxies are actually receding from us and redshift is due to Doppler shift. Now that redshift has been measured to reflect a cosmological constant of space itself expanding, rather then drop the singularity and actually go back to Einstein’s original theory that space expands to balance gravity in one large cycle, thus the redshift of light that travels between them is simply redshifted because it is the light that takes the route least blocked by gravity fields and thus most expanded, cosmologists have come up with this patchwork mess where the contraction of gravity is divorced from any effect on the larger space, so in order to assert galaxies are actually moving apart and explain thermal equilibrium, etc., we have dark energy, inflation and all the other patches to shoehorn an measurably infinite space into a finite space theory.

  • John Merryman


    They still have very rational articles on biology.

  • Jim Antoniadis

    John Ramsden gives a beautiful and elegant description of our universe moving in two time directions. His description fits with some thinking I have done regarding this subject over the years.
    If our universe is five dimensional this can be modelled as two four dimensional spheres each with three space dimensions and one time dimension. The time dimension would be opposite in one of these 4D spheres compared to the other.
    The presence of dark energy has been postulated as the explanation for our expanding universe and as yet there is no theoretical basis for this explanation so it has been invented as an ad hoc solution to an observational dilemma. It is also known that the cosmological constant seems finely balanced to fit the observed flatness of our observable universe and the apparent balance between the forces expanding it and those that would result in its imminent collapse.
    I believe that there could be an alternative explanation that could also explain inflation.
    If the universe is seen as being like the inside of a black hole then the only way that such a black hole could expand is if a) the gravitational constant were increasing from near zero to the current strength, b) the speed of light were decreasing from an initial velocity close to infinity to our current speed and becoming even slower in future or c) if energy were to enter the universe/black hole from an external source.

    If both a) and b) are true i.e. that ” G” is increasing and “c” is decreasing over time then eventually all matter will reside in black holes as even the gravity of an electron will have the capacity to hold in light that is travelling at a snails pace. The event horizon of all of these black holes would continue to expand until they all coalesce and once again all of the energy of the universe would be contained within a single event horizon that would usurp the original event horizon (diameter of the universe) this then would form the second half of the first 4D sphere of the 5D universe. Once this has occurred then “G” would start decreasing again to zero and “c” would start to increase to infinity again so that the event horizon/diameter would again shrink back to the Planck length type dimensions at the start of the Big Bang. At this point the process would start again. This could then be the second 4D sphere of the larger 5D universe. Its expansion and then contraction would end back at the beginning of the Big Bang of our 4D sphere.
    I know that there has been no observation that indicates massive variations in G and c as indicated by alpha although this could occur if the Planck constant were also to grow larger in inverse proportion to the diminution of the speed of light. An increasing Planck constant would also help to explain how the tiny universe at the Big Bang could still have contained as many microstates as it does at its current size.

  • Danny

    John, I do agree completely on what you have written, just one note – gravity is the geometry of spacetime, not just space. So what about the geometry of the Universe, is it open, closed or flat? The most recent measurements tell that the flat Universe is the most comprehensive explanation, given the dark energy and the observable amounts of dark and baryonic matter. Regarding recessional velocities of the galaxies, they recede from us faster and faster, and outside the Hubble sphere (which diameter is approx. 20 billions ly) they recede from us faster than light. Considering routes of light – don’t they simply follow the geodesics? Also, speed of light is constant only when strong gravitational fields are absent.

  • John Faughnan

    I love the article. I’m an avid reader of yours, so I was reasonably prepped — that probably helped.

    You did lose me at one point. I don’t have it in front of me, but I felt I was following the microstate/macrostate statistical model of entropy until you wrote that quantum mechanics requires preservation of microstates (I’m messing this up) and thus the potential microstate complexity of the initial QM fluctuation that births a child universe could not be less than that of the universe itself.

    Could you please consider a separate post on your blog in which you recap your very clear microstate/macrostate entropy discussion from the article, then reconcile that with the quantum mechanical constraint of total possible microstate preservation?

    Thanks for writing the article. I’m a fan!

    PS. For what it’s worth, I did my undergrad at caltech, albeit not in physics!

  • Sean

    Sorry that I’ve been too busy to join in, but I did at least want to thank Ian (#33) for sharing the story of his daughter, and BRB for defending my honor.

    John Faughnan, I’m not quite sure which part of the article you are referring to. Under any deterministic dynamics, specific microstates evolve into other specific microstates. At that level, to see an increase in entropy you have to coarse-grain; the entropy of a microstate is taken to be the entropy of the macrostate of which it is an element.

    The bit about the early universe was trying to explain that you might be tricked into thinking that there weren’t many possible microstates back then, because the universe was small. But that’s incorrect, at least if we believe in deterministic evolution; all configurations are part of the same space of accessible conditions. There aren’t that many microstates that look like the early universe, but that’s just restating that the entropy was low, which is the problem we are trying to explain.

  • John Merryman


    Thank you for taking the time to think through what was a hasty post with many points crammed in it.

    To further the heresy though, I do have problems with the commingling of time and space. I’ve been arguing on CV for some months (posts 24, 35, 40, 45, in this thread) that time is a consequence of motion, rather than basis for it. This would make it similar to temperature. The further question would be about the nature of space and whether it wouldn’t also qualify as a consequence of the motion defining it as well. My impression is that it is the void which constitutes the conceptual absolute. The alternative on which much of our theories are based is that the conceptual absolute is best defined as a dimensionless point. While it allows us the ability to set boundries, such as defining the entire universe as starting from such a point and space is thus created as it and the energy defining it expands from this source, as well as facilitating theoretical determinism, as all effects trace back to a singular cause, I think the void works much better as a conceptual basis. While the point constitutes a conceptual actuality, the void allows a much greater degree of potentialities. For one thing it wouldn’t be inherently deterministic, since there isn’t a singular starting point, but an infinite number. This also removes the entropy problem, since it cannot be defined as a closed set and energy lost in one place is traded around with energy losses and gains from other places.
    As for cosmology, consider the consequences of a fluctuating vacuum(as opposed to an exploding singularity); The positive and negative forces are not stable and tend to collapse, but given that the environment is fluctuating and thus not stable, this collapsing isn’t even and rotations and other anomalies develop, so that the state of near equilibrium builds up ever larger vortexes that eventually explode, or otherwise eject and radiate away energies.(1) Thus forming the sea of gravitational wells set against expanding space. Since space is infinite to begin with, this expansion cannot increase the size of the entire universe as a singular unit, so it adds to the pressure applied to those gravity wells, causing them to spin faster than their internal attraction causes. So the need for dark matter is explained, not as additional internal attraction, but external pressure.
    Since it is only the emptiest space which light travels the furtherest and this is what is expanding the most, galaxies are not actually moving apart on an absolute scale, so there is no need for the enormous amounts of additional energy required to actually drive them apart faster, which removes the need for dark energy.
    An additional point I’ve made is that if space expands, our most basic measure of it, the speed of light, should increase proportionally. Otherwise it isn’t really expanding space, but increasing distance of stable space. Which poses other problems for Big Bang Theory, since the expansion of space as a medium was originally used to explain how other galaxies could all be redshifted directly away from us, without the earth being at the center of the universe. The rising loaf of raisin bread analogy. This effect of redshift proportional to distance is accounted for if expansion is simultaneously neutralized by gravitational contraction. An analogy for this might be that running up the down escalator doesn’t cause the floors to move apart, as the extra steps are pulled back into the system.
    The further light travels, the more the effect is compounded, so that eventually a horizon line is created when the source appears to recede at the speed of light. While the source of direct visible light is no longer visible, black body radiation travels over this horizon line, explaining the bath of CMBR emanating from all directions at the approximate distance visible light ends.

    So since I see the void as the basis for motion and time as a consequence of it, space and time are not the same. Yes, the math adds up, but the same logic used to describe how the measure of time is analogous with the measure of distance, since the device of lightspeed relates to both, could also be used to argue that the measure of temperature is analogous with the measure of volume, as a identical amount of energy occupying different volumes is going to have an inversely proportional temperature, but we understand the relationship is a triangulation of measurements relating to energy levels, not that temperature and volume are synonymous.

    1. What if entire galaxies very occasionally exploded? For one thing it might explain the large voids occasionally found. If we have a universe that is infinitely old, it allows a far greater range of explanatory possibilities than trying to cram everything into 13.7 billion years.

  • Ethan

    Here’s a simple one: inelastic collisions. Totally classical physics, but inelastic collisions are not time-symmetric. Inelastic collisions are the basis for the formation of all structure in the Universe.

    Isn’t this a simple enough argument to show that the arrow of time must run the way it does, and that despite the fact that most laws of physics for 2-body interactions are time-symmetric, the end macrostate doesn’t exhibit that time symmetry?

  • Sean

    Inelastic collisions are just completely conventional examples of macroscopic time-asymmetry: start with low-entropy initial conditions, and watch the entropy increase. (Inelastic collisions produce heat, sound, and stresses, all of which increase the entropy.) They have nothing to do with why the arrow of time “must” run the way it does, since they say nothing about why the entropy was low to begin with.

  • jeffw

    What about wavefunction collapse or world-splitting (or whatever)? A particle seems more ordered than a wave (but maybe not), and i guess the larger system must be taken into account. In an Everett scenario, is the tree balanced on both sides?

  • Plato

    What was Veneziano thinking?;)

  • Plato

    And it’s Gabriele Veneziano. He is what we call a “super cosmologist” because, he thinks “outside the box.”:)

  • Ethan

    No, sorry. I didn’t meant that this explains why the entropy of the Universe was low to begin with. But as far as I understand it, it doesn’t matter what the value of entropy is or was at any point.

    What matters is that the entropy is higher now than it was in the past. All the interactions that we know of either increase entropy or conserve it. I guess what I don’t understand is what all the fuss is over. You have some initial state, like a nearly homogeneous and isotropic expanding Universe, it evolves according to the laws of physics, entropy increases, the end.

    Given that our Universe obeys the laws of physics we know, I don’t understand what question you’re actually attempting to address about the arrow of time.

  • Tom O’Bulls

    “You have some initial state, like a nearly homogeneous and isotropic expanding Universe,”


    That’s the question.

  • John Merryman

    Energy radiates across 13.7 billion lightyears. Gravitational structure coalesces from within a radius of a few hundred million lightyears. While the energy loss of radiation is diffused across much more area then the gases from which mass congeals, it is also radiating in from the same amount of area. Unless of course you assume it’s all just expanding into a void that doesn’t otherwise exist, from a point that has no boundaries. All the while measured in lightspeed that remains stable, as the very space it measures expands. It is a question.

  • John Merryman


    I wonder if the answer to this might not be hiding in plain sight. E=MC(Squared). A useful formulation that even Einstein had to tinker with, even if just a little. Our physics have no explanation for what happens past the “speed” of light. On the other hand our physics, certainly on our own little macro level, depends upon a formula that includes 3 elements, one of which clearly does not exist — the speed of light squared. What might energy, matter, time, or space be, or would they be, in a multiverse where one “sector” was ordered on laws of physics where the speed of light squared was the rule rather than the exception?

    Someone correct me if I’m wrong, but my impression of what C2 meant was it was a way to express the volume of energy, since light is a wave spreading out in all directions, it is the x axis by the y axis, not the speed of light multiplied by the figure of its own rate. I even forget if I read this somewhere, or it was just an assumption.

  • Ethan


    That’s not an interesting scientific question. “Why” do we have our Universe? As in for what reason? Not a question science pretends it can answer.

    If you meant “how” did we get to have an isotropic and homogeneous expanding Universe, well, the best mechanism we know about is inflation. Inflation sets up many of the initial conditions that we need in the big bang, including a flat, homogeneous, isotropic Universe with a spectrum of inhomogeneities, devoid of monopoles and topological defects, hot enough to bring about baryogenesis and nucleosynthesis. But that’s not an arrow of time question.

    And with an initial amount of entropy, that either stays the same or increases over time. It’s not like it’s increased all that much, either: over the course of our 14 billion years here, entropy’s increased by roughly a factor of 2. But again, what I don’t get is this notion that you can run time backwards and all the laws of physics will still be time-symmetric. They aren’t: try to run a Sun backwards. Or perhaps simpler, try to fry an egg backwards in time, or drop a glass of water onto the ground backwards. You can’t do that in this Universe. And that’s what I don’t get — since we know this, what can we ask that’s interesting about the arrow of time?

  • Daryl McCullough

    Ethan, I think you’re missing the whole point: The laws of physics are time-symmetric (at least the most common processes are). The laws of physics do not say that an egg cannot unfry or a puddle of water cannot spontaneously gather into a glass. That’s exactly the question that Sean is trying to answer: how can a time-asymmetric macroscopic rule (entropy always increases) arise from time-symmetric microscopic rules?

    The partial answer is initial conditions. If the initial state of a system has a very low entropy (compared with its highest possible entropy), then the system is overwhelmingly likely to evolve in a way that increases entropy. On the other hand, if the initial state has the maximum possible entropy, then it will evolve so that entropy decreases (slightly, for a very short period of time). But this partial answer just brings up a followup question: why was entropy low in the early universe?

    Saying that it was lower in the early universe than it is now because entropy always increases is circular thinking, because the reason entropy always increases is because it started out low.

  • Ethan


    I’m sure I’m missing the whole point. Many common processes involving the laws of physics are time-symmetric, as you state. But many processes are not time-symmetric, as well, even on microscopic scales.

    Consider any 3-body collision, for example, or neutron decay, or the process of successful recombination in the hydrogen atom. If I were to run these processes forwards in time, they would occur with a certain probability. If I were to run them backwards in time according to the same laws of physics, they occur with a different, specifically a smaller probability.

    I’m arguing two things here, I suppose. First, the microscopic rules are not all time-symmetric, and we know exactly how they aren’t time symmetric. How, therefore, could you reasonably expect that the macroscopic results would be time-symmetric? Second, the entropy was “low” (by which you mean non-maximal) in the early Universe because the physics that created the big bang created a certain amount of entropy, which happens to be low compared to the maximal value. It isn’t a philosophical question, though. The amount of entropy produced in the big bang is a consequence of the laws of physics that we have, and it turns out to be far below the maximum possible value.

    I guess I still don’t see why this is interesting.


  • Daryl McCullough

    Ethan writes: But many processes are not time-symmetric, as well, even on microscopic scales.

    No, that’s not true. There are violations of time-reversal in weak interactions, but there is no reason to believe that that violation is responsible for the second law of thermodynamics. It clearly is not.

    Consider any 3-body collision, for example, or neutron decay, or the process of successful recombination in the hydrogen atom. If I were to run these processes forwards in time, they would occur with a certain probability. If I were to run them backwards in time according to the same laws of physics, they occur with a different, specifically a smaller probability.

    No, that isn’t true.

    First, the microscopic rules are not all time-symmetric, and we know exactly how they aren’t time symmetric.

    That isn’t true, either. The microscopic rules governing ordinary interactions (which covers your examples of eggs frying and spilled water soaking into the ground) are time-symmetric.

  • BRB

    Ethan: everything that Daryl said, plus: it isn’t true that inflation explains the initial conditions. In fact, it just makes the problem worse! All explained in SC and Chen’s paper on the arxiv.

  • Jason Dick

    The bit about the early universe was trying to explain that you might be tricked into thinking that there weren’t many possible microstates back then, because the universe was small. But that’s incorrect, at least if we believe in deterministic evolution; all configurations are part of the same space of accessible conditions. There aren’t that many microstates that look like the early universe, but that’s just restating that the entropy was low, which is the problem we are trying to explain.

    Okay. I guess it just doesn’t make sense to me for a quantum theory of gravity to be unitary. If we consider, for example, that a patch inside the horizon of a region of de Sitter space, it has a finite number of degrees of freedom. But it evolves into a state that includes many now causally-disconnected patches, each of which has the exact same number of degrees of freedom. Since there are more such regions, it would appear that the total number of degrees of freedom has increased, in violation of unitarity.

    So, why should a quantum theory of gravity retain unitarity? Any such theory that retains unitarity would require, for example, that each of the patches into which a de Sitter region evolves be different in some way from the original patch, such that they individually have fewer degrees of freedom.

    And if you don’t have unitarity, it becomes really easy, it seems to me, to have inflation generate a low-entropy region: you start with a region with few degrees of freedom but high entropy, and if it evolves into a state with many degrees of freedom, it pretty much automatically ends up in a very low-entropy state. So why should quantum gravity be unitary?

  • John Merryman

    Doesn’t Inflation theory just explain how thermal equilibrium is possible within the constraints of the BBT timeline while breaking the fewest possible fundamental constants? I know this is Chinese water torture for some, but wouldn’t an infinite universe have a natural equilibrium?

    Isn’t there a basic conflict between the first and second laws of thermodynamics; All energy is conserved, yet it devolves to a thermal equilibrium. ?

    What if a complete thermal equilibrium, given the level of energy present, is just not stable. The cosmic microwave background radiation seems to be the energy closest to equilibrium, but there is a definite phase transition at 2.7k. Space doesn’t seem able to hold energy above that in a stable, uniform state, so it is constantly collapsing and expanding around equilibrium, creating a perpetual convective cycle.

  • Plato

    Sean saying it’s running backwards, but not in this universe.

    Are not similar conditions created in particle collisions?

    So are there not instances where such events within this universe could be considered “running backwards” and thusly, contributing to the dark energy? It would be by the assumption that such super symmetric states can indeed be reached and lie “under” the existing universe.

  • John Merryman


    There are two sides to the cycle quite evident in this universe. Gravity contracts energy into structure, until such a point where it breaks down, blows up, radiates back out, expanding the space occupied, as it diminishes the degree of structure. Then the process of collapse starts again. The reason there is an eternal amount of usable energy is because it is not stable at complete thermal equilibrium, so we have these tidal forces pulling it in and thus heating it up, which expands it, so it pushes back out again. No need for a singularity as uncaused cause of low entropy state.

  • Sandy

    The microscopic laws of physics are time symmetric because that is how we set up the equations. They run from time1 to time2 and then we can run them backwards from time2 to time1 and say they are the same. But, what if the nature of time isn’t captured by time1 and time2, what if time2 is inherently different than time1? Time as expressed in equations is not derived from experiment is it? Its basically a clock. It may be a simplification that misses something critical that if captured would show the source of time asymmetry. Until we can run experiments backward and forward in time, we can’t prove that the microscopic laws are time symmetric.

  • Plato


    Sometimes a picture is worth a thousand words.

    You had to know that for every “entry” there is a previous entry in thought.

  • Daryl McCullough

    Sandy, yes, it is true that we use time-symmetric equations to describe microphysics, but that is because empirically we find that time-symmetric equations most accurately describe our observations at the microscopic level.

    More specifically, you start an experiment by setting up a system in state A, you let it run for T seconds, then you check to see if the system is now in state B. Doint this repeatedly allows you to empirically determine a transition probability: The probably P(A,B,T) that the system will go from A to B in T seconds. But now, suppose we put the system in state B first, and then wait T seconds, then check to see if it is in state A. So we can do this repeatedly and compute P(B,A,T).

    Empirically, what we find is that for systems that are microscopic, involving only a very small number of particles, the two transition probabilities are the same: that is, P(A,B,T) = P(B,A,T). So at the microscopic level, anything that is possible, the reverse is also possible, and has the same probability.

    However, if instead of looking at tiny systems with only a few particles, we look at macroscopic systems involving 100000000000000000000000000 particles, then there are transitions (an egg frying, ice cubes melting in the hot sun, etc.) such that the process runs perfectly fine from state A to state B, but the reverse is never seen.

    So the problem of how to explain how microscopic reversibility gives rise to macroscopic irreversibility is not simply an artifact of the way we set up our equations.

  • John R Ramsden

    Jim Antoniadis, many thanks for your reply (#70) to my #59. I can certainly see the similarity between your intriguing model and mine in some respects.

    For now I’ll just make one point, which may have a bearing on the spheres you mentioned. Although it doesn’t detract from your ideas, I guess the moral is that in the context of cosmic evolution one must be flexible in thinking about the shape and structure of regions, and even whether these are bounded or otherwise, because suitable transformations can turn everything around in all kinds of ways. (That’s not speculative, but routine physics!)

    In my model, once the collapsing “late” universe is past the self-dual boundary (the critical density I referred to) and thus inside the black hole and approaching the Cauchy horizon inside, its rotation frame-drags it into a tremendously tightly wound spacetime “reel”, with each layer perhaps a Planck distance deep.

    The emergent dual universe (i.e. our “young” universe) is then manifested by particles and radiation travelling round and tunneling between these layers. In other words it is “intertwined” with the late universe, which thus underpins it so to speak. (To an observer *in* the dual universe, these layers appear as the miniscule “curled up” dimensions one reads about!)

    It’s the cosmic equivalent of living in a warren of limestone caves, where the stone is made up of the stacked bodies of countless creatures that once swam free, perhaps in long vanished limestone caves themselves!

    Getting to the point, some years ago, a variant Big Bang theory called the
    Ekpyrotic Model appeared. You’re probably familiar with it, and one can always pursue the references in that Wikipedia article. In summary, the idea was that a pair of parallel branes, floating in a high-dimensional space triggered the Big Bang by “colliding” and bouncing apart. I thought at the time it seemed slightly contrived. For example, why were these branes parallel in the first place?

    But the “leapfrogging duals” model I sketched is compatible (in essence) with this colliding brane scenario if one assumes that these two interacting branes are none other than the collapsing late universe, and its emergent dual (which arguably starts out parallel to the late universe, by its very nature).

    The above mentioned Wikipedia article also references related and more recent
    cyclic models, one of which involves orbifolds. As far as I can discern, these are foliated or (roughly speaking) “wrapped” objects akin to the “reel” referred to above; but the Wikipedia article on them is rather terse and formal and not very illuminating.

    You see what I mean about needing flexibility (literally!) when thinking about branes and their interactions? They needn’t be like large sheets which happen to be parallel and clash together and bounce apart. Their structure and “collision” can be a more subtle, as summarized above.

    P.S. in mentioning Fourier transforms I wasn’t simply name-dropping. Like similar transforms, it has a built-in uncertainty principle based on the product of variances.

  • Sandy

    Thank-you Daryl, that is helpful. I wonder, is saying something is reversible (the probability of it going from state A to state B is the same as it going from state B to state A) – is that the same as saying it can go either way in time?

    So the basic question is – why do we have a constraint (increasing/decreasing entropy) at macroscopic scales that we don’t have at microscopic scales? Are we using entropy as a proxy for the arrow of time? Is this warranted?

  • John Merryman


    It’s a political fact of life that you can’t question the rules and play the game. Presumably this shouldn’t apply to science, since science is all about finding what the rules are in the first place, but the irony is that instead of negating this fact, often it is re-enforced by the process, since it is assumed those ideas which have withstood questioning must be the objective truth and become the foundation of further enlightenment, which only re-enforces their status. Consider; In fifth grade geometry, did they explain that zero is the center point on the graph sheet, or did they point out that zero would actually be a blank sheet of paper? Suffice to say, the center point has become enshrined as the zero point from which everything else is defined in terms of a four dimensional coordinate system, with time represented as a dimensional projection similar to the three minimalist spatial dimensions. Since we record time as a series of events, this has instinctive logic, but formalizing it as objective reality leads to such discussions as why we can go both directions on the x, y, and z axes, so why can’t we go both directions on the t axis. Eventually this train of logic devolves into the current nonsense of multiple universes, worlds, etc, because the initial assumption that since time can be modeled as a dimension means it must be a dimension, is wrong. Actually there are natives of South America who view the past as in front of them and the future behind. That is because their point of reference is the energy, not the observer and the event occurs before the observation, so time travels toward them, with the future becoming the present, then the past. As I keep pointing out, this does represent an opposing direction of time, but using a different model is very much like speaking a different language. If your brain is not wired for it, it is just gibberish.


    Back up to what we really do know. We live on a very small spot in space, in a galaxy that is trillions of time larger then this spot, yet is itself a small spot very far from other such galaxies. We observed the light from these other galaxies to be redshifted proportional to their distance from us, although there likely are other reasons for redshift. While the history of this argument is too long to describe, accept that I’m not one who thinks redshift is best described by Doppler Shift, due to recessional velocity. As I keep pointing out, it cannot logically be argued that space itself is expanding and still have a constant speed of light to measure it against. Either space really is expanding, but lightspeed would have to increase accordingly, in which case it wouldn’t appear to expand, as any source x lightyears away would always appear x lightyears away, or it isn’t expanding space, but an increasing distance of stable space(which is what Doppler Effect is anyway) and then we get back to the reason way it is claimed space is expanding, which is that we would be at the center of the entire universe, since every other galaxy is redshifted, relative to distance, directly away from us. Which is nonsense.
    So. Is there another explanation for redshift that does make sense? We do perceive gravity as curving the measure of space around a body, yet radiation climbs directly out of this gravity well. Could it be that radiation does effectively have the opposite effect on space, since it does expand, just as mass contracts? Since it would be a far more gradual effect and effectively hidden in the vicinity of a gravity field, with no point of reference around which it curves, the only viable proof would be the redshift of light from distant galaxies. When the effect first described as proof of dark energy was first discovered by Perlmutter, et al, in 1998, it was that the supposed rate of expansion wasn’t being slowed by gravity, since the assumption was that it was all a consequence of the initial singularity and should be cooling off. On the other hand, if redshift is an optical effect on light crossing enormous expanses of space, with the effect compounded, so that earlier redshift is further redshifted the further light travels, the more the effect is multiplied and appears faster the further away it is, so that eventually the source appears to recede at the speed of light and beyond which it is no longer visible, then we have the signature of a cosmological constant that balances gravity, not the afterglow of a singularity with dark energy tacked on to explain the observation of continuing expansion. Since this is an optical effect, the actual sources are no more receding then a gravitationally lensed object moves around in space because an intervening gravity field makes it appear that way. Therefore the energy to make these objects actually move away is unnecessary.

  • Plato


    Physically, the effect can be interpreted as an object moving from the “false vacuum” (where = 0) to the more stable “true vacuum” (where = v). Gravitationally, it is similar to the more familiar case of moving from the hilltop to the valley. In the case of Higgs field, the transformation is accompanied with a “phase change”, which endows mass to some of the particles.

    Spontaneous Symmetry Breaking

    Keeping this in context of the “current state of the universe” helps me in some regard to point out the “perspective of others “who believe that the valleys can contain entropic valuations based on that “earlier time” in the universe’s expression.

    We do not disregard the phase changes.

    G -> H -> … -> SU(3) x SU(2) x U(1) -> SU(3) x U(1)

    Here, each arrow represents a symmetry breaking phase transition where matter changes form and the groups – G, H, SU(3), etc. – represent the different types of matter, specifically the symmetries that the matter exhibits and they are associated with the different fundamental forces of nature

    Unification of all the forces(including gravity) saids there is a earlier time?

    So we keep a “birds eye view” on the “nature of the universe” from this perspective? Applying “theoretical bends” to the interpretations, is just another way of exploring the potentials and realization of what has transpired from the beginning(where is that?).

    There had to “exist a place” for this cyclical nature, to join the beginning and end? Sean saids it’s not in this universe, and I am saying it is so.:)

  • Sandy

    On why the universe had such low entropy in the past…if there was a singularity, wouldn’t that be the ultimate in low entropy? All potential matter/energy united into one thing. Couldn’t you say there were no microstates, just one state?
    If there was then a big bang, all initial relationships, time and space are created…that is a lot of order as well, as the initial conditions for the whole universe are formed; all the “rules” of chemistry, physics. That too is a whole lot of order. If the question is WHY there was such low entropy in the past, that isn’t really a question that can get answered. We don’t know why anything.

  • Peter Lynds

    Dear Sean,

    Firstly, I want to say that I think your model is really interesting. I also think that you should be really commended for taking on such a difficult issue. I think your model has some problems though. Take the/a big bang. At this point, you posit that events evolve both towards the future and past with entropy increasing. This is akin to simply treating time as being both positive and negative, and presumably this is justified by the fact that the second law predicts that entropy should equally increase in both past and future directions.

    However, other than the following clearly being so problematic, there doesn’t seem to be any real reason in your model why entropy shouldn’t equally increase into the past at all times during the future directed evolution (rather than just at t=0). Moreover, and more importantly, evolving something into the past only makes sense if one has a past to work with (i.e. something that has already happened), otherwise it wouldn’t/couldn’t be the past! Without one, the temporal orientation of the model’s so-called past-directed evolution would be indistinguishable from the so-called future-directed one. Of course, at t=0, you do have a past to work with; the earlier universe. But for events to evolve into the past here, entropy would be decreasing in this direction, and it would be a no go.

    This, however, is not to say that entropy cannot increase in both time directions, as I certainly think it can. One needs a past to work with, however, and by definition, this is set and has already taken place. Moreover, the reason entropy cannot ‘simultaneously’ increase in both time directions, is because if one reverses the direction of evolution of a system with entropy increasing, entropy would be decreasing in that direction. The system already has a past which is set, and one cannot invent a new one for it.

    Forgetting the above, though, I think your model has a more basic problem. If the past was infinite and the universe had no beginning point, it would be impossible for it to evolve, not only to where we are today, but at all (forward or back). Of course, the idea of a universe having a beginning at some finite time in the past is equally problematic. Neither can be correct. Although you are obviously far from alone in this respect, I think you have to face up to this.

    Best wishes


  • Garrett Connelly

    I have a problem with the idea that the number of possible entropic microstates remains constant across the time line between nearly empty to nearly empty. It seems to me that communication of individual perceptions carries information which can effect outcomes in a blend of eddies that reflect qualitative differences.

    Yes, there is a fertile field at the extreme limit which one might refer to as quantum-gravitational microstates, even so, I cannot see why this uniform and featureless blend of perfectly organized building blocks should not be viewed as unstable zero entropy ready to burst with potential qualities expressed as conscious and democratic decisions by perceivers who act as individuals to qualitatively manipulate unfolding physics.

    Am I way off course here? It’s kinda important to get this part of things straight.


  • John R Ramsden

    Garrett Connelly wrote (#102):
    > Yes, there is a fertile field at the extreme limit which one might refer to as quantum-gravitational microstates, even so, I cannot see why this uniform and featureless blend of perfectly organized building blocks should not be viewed as unstable zero entropy ready to burst with potential qualities

    Leaving aside the (sentient?) perceivers aspect, which seems entirely irrelevant, and assuming “unstable zero entropy” is supposed to mean something like “immanent or potential low entropy”, my model (summarized in #59 and #96) does precisely that.

    Some overarching influence is needed to sweep those microstates into a configuration where the “potential qualities” you refer to can cohere into a new form of dynamics (which I argue is comparable to an early low-entropy evolutionary stage of the previous lot of microstates, and the process can be repeated).

    Although the influence must have the appearance of being fairly uniform, one need not take the word “overarching” too literally: The uniformity could be a consequence of local interactions acting the same everywhere, rather as a large flock of starlings seems to swoop and soar as a unit even though this is a consequence of each bird simply tracking its nearest (seven I gather?) neighbours.

    I think Peter Lynds is thinking along the right lines; but my impression, after a quick look at his paper, was that it doesn’t adequately address the issue of initial low entropy.

  • JTankers

    Interesting speculation, a bit out there, but interesting. Loop Quantum Gravity and have some interesting and possibly more compelling speculations.

    I do find very plausible the notion of dark energy. If space contains even any vacuum energy, that would destroy the arguments of some that Hawking Radiation might require energy to be pulled from black holes, somehow across the event horizon, to balance the energy equation. That disputed and now more frequently rejected theory was based on several other dubious conjectures also…

    Speaking of conjecture, some are claiming that the Safety arguments claiming that creation and capture of micro black holes on Earth are nothing more that conjecture.

    Any idea why some very eminent scientists such as the inventor of Chaos Theory’s Rossler Attractor are EXTREMELY concerned that the risks from the Large Hadron Collider may other that what is being told to the public. The danger might actually be a virtual certainty if some disputed assumptions are in fact as some scientist believe may be possible or likely.

    Did you know that we currently have NO credible reason to exclude possible high or inevitable risk from stable fast growing micro black holes?

    No good scientific reason to reasonably rule that possibility out. Really.
    Start this learning process at

  • Plato

    John Ramsden:I cannot see why this uniform and featureless blend of perfectly organized building blocks should not be viewed as unstable zero entropy ready to burst with potential qualities

    It’s an “alternative” to progressing the view of what is actually taking place in RHIC, and LHC, keeping in mind, what Sean is offering for perspective.

    What is self evident by way of inductive deductive inference(historically sits at the very derivations of a constitution by John Adams) is an exercise in coming to the next step in what is possible in “other universe” is not possible in this one.

    By phenomenology alone, and while there are experts here on this site, such leading experimental situations offer the continuance of this perspective, that I offer to Sean’s position?

    “A phase” in the collision process itself? Provides for, the “relativity of expression” at this level?

  • John R Ramsden

    Plato (#105), Those looked like my words. But that whole first “paragraph” was a quote from the preceding post by Garrett Connelly, to which I was replying.

  • hitesh brahmbhatt

    Does it mean that our universe and all other universes are just bubbles in the ocean of eternal unchanging energy?

    I have been fascinated by such references in ancient Indian mythology but I am not sure if there is ever a way to prove it?

    On the other hand, just the idea of things moving to higher and higher entropy for ever does not seem tenable either.

    I think I read in Brian Greene’s book that said that if the laws of physics are symmetric with regards to time that means Second Law of Thermodynamics applies in the time-reversed direction as well and that would imply that present state is always the lowest entropy state ?



  • Plato

    My apologies John R Ramsden.

    The reply still holds.

  • John R Ramsden

    No offence – easily done. I’ll start using the ‘blockquote’ syntax, which I just noticed.

    Maybe I’m being thick, but it wasn’t clear what you were trying to convey in #105.

  • Plato

    It’s a continuation of #99.

    For such “beginning and ends” to occur in our universe, they must satisfy a “relativistic interpretation “even at such quantum perspectives as denoted in the microseconds of the earlier universe? I think “this point” is important, and one must push back perspective according to the experimental data?

  • Rich Wilson

    From all the responses to Dr. Carroll article in the June 2008 issue of Sci Am. it is clear that it is well written and interesting reading. Thanks to him and perhaps his wife, Jennifer Ouellette, for making science reading fun and informative(#33). We need more such scientist to help educate us on science in our complex modern world. Thanks and keep up the good work.

    Quoting from his Sci Am article: ” Unfortunately, we do not fully understand entropy when gravity is involved” and, “In actuality, though, empty space has plenty of mocrostates-the quantum-gravitational microstates built into the fabric of space. We do not yet know what exactly these states are, any more than we know what microstates account for the entropy of a black hole”.

    Perhaps in the evolution of a black hole the intense concentration of matter and energy leads to extremely high temperatures and matter “disappears” similar to the reverse situation of the Big Bang where in the first fractions of seconds or so, the universe cools and matter first appears. In both cases it seems that without matter, gravitational energy is minimal or absence and in both cases we have a hot, dense concentrations of energy. Just as in the Big Bang energy is thought to be homogeneous and in the low entropy state, so to we might expect at least some black holes to evolve into a similar rare homogeneous microstate that reverses the entropy from “Medium” or high entropy to the state of low entropy, and poises the black hole into a Big Bang scenario and a new universe.

    While this “bounce” violates the second law of thermodynamics, it occurs under very rare and extremely different conditions than any other in the universe,and thus many not follow the same physical laws. However, it could lead to one that is similar to the state we know occurred at the time of the Big Bang.

  • Amiya Sarkar

    How do you think you’re going to deal with ‘entropy’?
    A good read though, thanks!

  • Peter Lynds

    Thanks John R Ramsden. In relation to the issue of initial low entropy, my model (if correct) simply shows that there isn’t actually an issue and that the question is a misnomer. It can be a tough one to get one’s head around though.

    Sean, I would be curious to hear how you would respond to my comments. Think of it as a friendly/collegial cosmological call out.

    Best wishes


  • Garrett Connelly

    I want/try to keep my keyboard in neutral and strive to look strictly with other’s eyes, yet, in this particular realm, is entropy made any more complicated or difficult to understand by gravity than by the emergence of sentient perceivers?

    Brian Greene says the only thing expanding as fast as the cosmos is consciousness. Albert Einstein said our perception of ourselves as individuals is something of an “optical illusion” of consciousness.

    These points of view have practical everyday implications that are difficult to bring up most of the time.

    Cordially wondering,


  • Pingback: Reality is what happens to you while you’re busy coming up with other theories.* « Communion Of Dreams()

  • M.H Khan

    Arrow of time always points towards the future as time is not just motion but also forces are part of time. One way to understand this is to consider two objects orbiting each other and imagine what would happen if time was slower faster or stopped for the orbiting pair. Both motion and forces will change increase and decrease or stop completely in the above senario with change in time.
    At quantum level time seems to symetrical because we fail to keep into account that forces are also part of time. Forces determine the arrow of time. That is why we objects fall down but rise up from the ground unless force is applied. Same applies to any other situation.

  • Nachiket pal

    i think that the way we are understanding time is wrong. there can be two reasons why time is asymmetry 1. we still dont fully understand the way time works 2. we are looking time in such a way that it seems asymmetric.

  • Dov Elyada


    Since my sophomore year some 4 decades ago, when I first met ENTROPY, I’ve been pondering the darned thing (not continuously, of course,) desperately trying to make sense of it. My efforts certainly included Boltzmann and Gibbs, but not before I came across Shannon and informational entropy did I make any significant progress. I think I can now identify the culprit impeding my understanding: it was the culturally entrenched qualitative description of entropy as a measure of “disorder.”

    The disorder explanation is widely used, especially when addressing scientists outside of physics, and the general public. In his SciAm article, Sean uses it several times, and with great stress. I claim that THE EXPLANATION OF ENTROPY AS A MEASURE OF DISORDER IS MISLEADING. But because it has become a culturally ingrained MEME that gets passed on uncriticized, even some accomplished physicists who were raised on it turn out to actually not understand entropy at all, let alone undergrads, engineers, non-physics scientists, science writers, the educated public and self-confident philosophers.

    While entropy properly applies to a SYSTEM — the totality of all system states of non-zero probability, represented by an ENSEMBLE — I think most people intuit order and disorder, as I once did, as attributes of a state. Consider, for example, a rigid cubical vessel full of an ideal, monoatomic gas in thermodynamic equilibrium. For simplicity of discussion look at the gas particle locations only. In one PARTICULAR state, call it DG1, the particles are evenly distributed throughout the volume, but randomly, unlike in a crystal. To specify DG1 completely one must list all its particle coordinates — no data compression possible. This is what one would call disorder. In another PARTICULAR state, OG1, the particles are also evenly distributed, but periodically, as in a crystal, and conforming to the vessel walls. Few parameters are needed to specify OG1 and natural language would call it an ordered state. Yet both DG1 and OG1 are states of the same system and, being fully specified, are equally probable. But to neither of them does the concept of entropy apply. It does apply to the ensemble of all such states, is related to their number in the ensemble, and is divorced from their degree of order or disorder. (If most system states happen to possess a high degree of order, then, one may argue, their number must be relatively small and so must be the system entropy. While being true, this roundabout argument has hardly any didactical value.)

    Consider now the example of the broken vs. whole egg that stars in Sean’s SciAm article. Again, for the sake of simplicity, consider a PARTICULAR whole egg named Humpty Dumpty (HD) and ignore its macroscopic mechanical motions (e.g., sitting on the wall,) and its microscopic thermal motions. Under these restrictions there exists just one distinguishable object that qualifies as a member of the set {HD}. Consequently, {HD}’s entropy is zero and we are justified in praising HD as a neat and orderly dude. Next, let HD have a great fall. The mess thus created surely deserves to be called “disgustingly disordered.” To determine the entropy of the new situation we would usually define an ensemble, a new set called “Broken Humpty Dumpty,” {BHD}, and enlist in it all distinguishable objects that qualify bona-fide as broken Humpty Dumpty. Since there is a vast number of such objects — the vast number of different ways HD might splinter and splatter and create Jackson Pollocks — the entropy of {BHD} is great and beyond the powers of all KH and KM.

    But notice that the above definition of the set {BHD} is quite arbitrary. In deterministic physics one can mean by {BHD} the actual configuration HD assumed after his historical fall on that fateful afternoon of 15 July 1648, the fall that was documented by notable eyewitness historians in the famous nursery rhyme. In that case, {BHD}, too, contains only a single member and, despite its remarkable messiness, its entropy still equals zero.

    A system may have many possible states, most may be “disordered” and some may pass as “ordered.” But regardless of whether the system accidentally or fleetingly occupies an ordered state or a disordered one, ITS ENTROPY IS THE SAME! That’s because THE ENTROPY IS RELATED TO WHAT THE SYSTEM MAY BE, NOT TO WHAT IT IS.

    (A plausible measure of state disorder may be the “algorithmic randomness”, a.k.a. Kolmogorov complexity, of the state — the size of the shortest algorithm that generates a specification of the state on a Turing machine. A theorem due to C. H. Bennett (I think; Int. J. Th. Phys. V21, N12, 1982, p.938) demonstrates that the entropy of a system is all but indistinguishable from the MEAN algorithmic randomness over all system states. This is instructive, because the mean of state-disorder taken over all system states does qualify as a system disorder measure. Nevertheless, as said above, this explanation is not at all transparent and hardly serves for clearing the entropy=disorder confusion.)

    I think it’s high time to exorcise the demon of disorder from academia and pop science alike. The wrong mental construct of entropy=state_disorder is very readily formed, in students as well as in the educated public, because people feel they understand, from everyday experience, what disorder is. But what they envision are disordered states, not disordered absract ensembles. Once the wrong construct takes root, however, uprooting it is next to impossible, especially as it has become a meme. It had played havoc with my own understanding of entropy and is surely detrimental to the understanding of all who has been exposed to it. [See Carson & Watson, Roy. Soc. Chem., 2002.] I was glad to find in Wikipedia’s “Entropy”, in the paragraph entitled Energy Dispersal, a full support of my position and some research-based references. For the sake of saving that trouble from future generations I hope that the new approach will become a trend and that Sean’s next SciAm article will contribute towards it.

  • Rich Wilson

    Thanks, Dov Elyada for your entertaining and very instructive thoughts on entropy. If nothing else, it is reassuring for me and perhaps other educated people that entropy is not as simplistic as just “order-disorder”.

    If you are so inclined would you elaborate on your statement, “Entropy of a system is…the MEAN algorithmic randomness over all system states”.

    Is “mean algorithmic randomness” quantifiable, perhaps as a measure of randomness? But isn’t entropy an energy term in a thermodynamic equation, and how does that square with “mean randomness”, which is more a statistical value?

    Thanks again for your insights.
    Rich Wilson

  • Peter Lynds

    Dear Sean,

    I take it that’s a no comment. For if you might find it interesting, this paper by John Moffat shares some features with your model (and to a lesser extent, my one too).

    Best wishes


  • Dov Elyada

    Hi, Rich.

    Please excuse the delay in my answer, I couldn’t find an earlier time.

    That entropy is indeed statistical is well known since 1877 and, by now, universally accepted. This is due to Boltzmann’s profound discovery expressed in his celebrated equation:

    S = k log(W), where S = entropy, k = Boltzmann’s constant, and W = the number of different molecular microstates, assumed equiprobable, available to the gas in the macrostate corresponding to S.

    We are not speaking here of the level of statistics that other thermodynamic state variables — temperature, pressure, etc. — are subject to. Unlike these, entropy is not a particle-average of a conserved quantity, such as energy or momentum, that survives the averaging to become a global system variable; it is related to an average of a quantity which is itself statistical in nature. Actually, no physical property of a system’s micro constituents exists the average of which can be interpreted as the system’s entropy.

    The bulk-thermodynamics facet of entropy — being a proper thermodynamic state variable and appearing in the 2nd Law and theorems derived from it — is, of course, fully consistent with the statistical view.

    Since about 1948, due to C.E. Shannon, the concept of entropy has broken through the boundaries of strict thermodynamics to become a major player in information theory, as a measure of the amount of information. At first one might think that information entropy and thermodynamic entropy are just similar or analogous, or that the term has been borrowed by one from the other and is used concurrently with different meanings. But no, they are one and the same. Entropy is the amount of information not only in a phonebook, in a movie DVD, on a communications satellite beam or in the whole of Cyberspace; it is also the amount of information in a volume of hot gas pushing on a car engine piston.

    Algorithmic randomness and its mean are indeed quantifiable. This is now a huge field and I cannot possibly do any justice to it — or to you — in a blog comment. But if I did succeed in arousing your curiosity enough to follow the lead I can assure you plenty of exciting reading and revelations.

    Yours truly
    Dov Elyada, Ph.D.
    Haifa, Israel

  • paul valletta

    In an ever expanding “de-cluttered” spacetime, the density decreases exponetially to the expansion, thus at a single point in an accelerating future-time, the “big rip” appears out of the vacuum? At this moment there is Entropic Phase “flipping”, the Universe comes to be out of the almost empty vacuum. The signal to those future beings looking backwards, must be that of a hot big-bang emerging out of a cold big crunch.

    Now interestingly, those forward looking beings may also record data that registers them to be in existence, prior to the big-bang.. if the current signal being Dark Energy?

    Is our percieved times-arrow, just like the archers arrow that has rebounded off its aimed target, and gone through a permiable target on the rebound, of area (vacuum)of less density?..the low density of a cold crunch?

  • Pingback: What Do You Say? | Cosmic Variance()

  • Pingback: Words in the Clouds | Cosmic Variance()

  • James Wan

    Sean, I have one question:
    I don’t understand why a social (or human) concept is involved in physics. What I’m wondering is the example of “broken egg”. They were given to prove that “broken egg” has higher entropy because there are countless forms of broken eggs but only one form of good egg. Here we inolves a concept that isolate “good egg” from “broken egg”, but this is from the point of view of human beings. If evaluate this from the view of nature, the “good egg” is just one of the countless forms of egg. It has not any speciall meaning to the nature than any form of “broken egg”. So when an egg is broken, why do we say that its entropy increased? For example, from the point of view of a bird, maybe the broken egg is more meaningful to him/her. So to the bird, when the egg is broken, he/she may thinks the entropy of the egg is decreased.

  • Sean

    James, there are certainly good questions to be asked about how we divide the many microstates of the world into a smaller number of macrostates. I believe that some ways of doing the division are naturally picked out by the laws of physics, but it’s still an open question.

    However, it doesn’t really matter. In terms of what we now call “broken” and “unbroken” eggs, there are certainly many more ways to re-arrange the molecules within the set of broken eggs than within the set of unbroken eggs. That’s an objective fact about the world, so you can’t just choose to insist that broken eggs have a lower entropy. You could choose some very specific form of broken eggs to define a particular kind of macrostate, and if you do it carefully there will be very few microstates that correspond to that choice; but in that case, unbroken eggs will almost never evolve into that specific form, so it will still be true that entropy almost always increases.

  • James G

    Excuse my ignorance, is entropy behaviour a classical or non classical result?

    In a small particle system ( < 1000 newtonians simulated on a computer) it seems obvious that “chaos” is the “norm” without recourse to quantum mechanics.

    At what number of particles does it become necessarry to “invoke” the existence of quantum microstates?

    Sean’s article is just one of many which fails to even address this simple question.

  • Dane C. Sorensen

    So many messy problems with the idea of reversing time. First of all and most importantly you would reverse the second law of thermodynamics. This law is not reversible as all the other laws of physics are. The Universe would not start going backward in time if it did happen to start to contract – do you get younger when you drive in reverse? Peter Lynds has done a fine job of solving the problems of the Big Bang. Also, has Lynds points out, if the Universe was created in a big bang, why was there an infinite amount of time before it decided to do so? Seems absurd. Why now? The only rational solution to Kant’s thoughts is the Universe is cyclic. Gravity is a force that can cause the Universe to return to the primordial soup once again. Lynd’s theory that singularities are not allowed, along with the second law of thermodynmics forces the Universe to Bang yet again as it has forever.

    And yes, a cylic Universe would imply that eventually in some distant cycle I will be typing these exact words under my own free will again. Cheers!

  • JCF

    Re “I will be typing these exact words under my own free will again,” one question to ponder is whether the “I” typing those words will be the same “I” typing your post, although certainly in no case aware of prior “I”s. In Nietzsche’s inconclusive consideration of “perpetual recurrence” he thought the same “I” would experience the identical same life all over again but he found the idea “horrifying.”

  • eCID

    As a latecomer to these discussions, I wonder if the ideas of Julina Barbour (“The End of Time”) have come up?

    Basically, he posits that there is no time–only NOW.

    My analogy is that of film in a motion picture projector. As per Barbour (my interpretation), only the frame actually in the projector window is “reality”.

  • eCID

    P.S. Instead of the “Arrow of Time”, does Barbour suggest THE ARROW OF ENTROPY?

    Following the motion picture projector analogy, the film spool that collects frames that have already passed the projection port are “expired QUANTUM EVENTS” (history?). The film spool that contains frames remaining to pass through the projection port are QUANTUM POSSIBILITIES. The sytem is reversible.

  • Pingback: Sean and Horacio « The Gauge Connection()


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

About Sean Carroll

Sean Carroll is a Senior Research Associate in the Department of Physics at the California Institute of Technology. His research interests include theoretical aspects of cosmology, field theory, and gravitation. His most recent book is The Particle at the End of the Universe, about the Large Hadron Collider and the search for the Higgs boson. Here are some of his favorite blog posts, home page, and email: carroll [at] .


See More

Collapse bottom bar