What Happened at the Big Bang?

I had the pleasure earlier this month of giving a plenary lecture at a meeting of the American Astronomical Society. Unfortunately, as far as I know they don’t record the lectures on video. So here, at least, are the slides I showed during my talk. I’ve been a little hesitant to put them up, since some subtleties are lost if you only have the slides and not the words that went with them, but perhaps it’s better than nothing.

My assigned topic was “What We Don’t Know About the Beginning of the Universe,” and I focused on the question of whether there could have been space and time even before the Big Bang. Short answer: sure there could have been, but we don’t actually know.

So what I did to fill my time was two things. First, I talked about different ways the universe could have existed before the Big Bang, classifying models into four possibilities (see Slide 7):

  1. Bouncing (the universe collapses to a Big Crunch, then re-expands with a Big Bang)
  2. Cyclic (a series of bounces and crunches, extending forever)
  3. Hibernating (a universe that sits quiescently for a long time, before the Bang begins)
  4. Reproducing (a background empty universe that spits off babies, each of which begins with a Bang)

I don’t claim this is a logically exhaustive set of possibilities, but most semi-popular models I know fit into one of the above categories. Given my own way of thinking about the problem, I emphasized that any decent cosmological model should try to explain why the early universe had a low entropy, and suggested that the Reproducing models did the best job.

My other goal was to talk about how thinking quantum-mechanically affects the problem. There are two questions to ask: is time emergent or fundamental, and is Hilbert space finite- or infinite-dimensional. If time is fundamental, the universe lasts forever; it doesn’t have a beginning. But if time is emergent, there may very well be a first moment. If Hilbert space is finite-dimensional it’s necessary (there are only a finite number of moments of time that can possibly emerge), while if it’s infinite-dimensional the problem is open.

Despite all that we don’t know, I remain optimistic that we are actually making progress here. I’m pretty hopeful that within my lifetime we’ll have settled on a leading theory for what happened at the very beginning of the universe.

This entry was posted in Science, Time. Bookmark the permalink.

71 Responses to What Happened at the Big Bang?

  1. Ken K says:

    Slide 16 – Time symmetric universe – I’ve seen it a few times now, but to me it seems to give special status to the parent universe at time zero.. ( even if child universes spawn in both directions of time ). This just seems wrong. Is this special status intended or am i missing something?

  2. Sean Carroll says:

    It certainly is special, in the same sense as the minimum of a parabola is a special point. But its existence doesn’t represent a fine-tuning: every single parabola will have a minimum somewhere.

  3. Jim says:

    Do these alternatives have different implications for perturbations during inflation such as in e.g. https://arxiv.org/abs/1606.07631 or https://arxiv.org/abs/1701.02544

  4. Paul Byrne says:

    Hi Sean – Thanks , I would like to have watched that lecture. May I ask a question – have you ever addressed the question of where all the energy in our universe comes from ? There is the concept of zero sum energy, that there is positive and negative energy and that at some point (the big bang) they were the same and somehow split to make the universe that we see now ? What do we know about this phenomenon or in fact if it is true ? Can infinite energy come from nothing ?

    Thanks – love your work

    Paul Byrne

  5. Sean Carroll says:

    To the best of our understanding, the total energy of the universe is exactly zero. Positive energy in matter/radiation/vacuum is canceled by negative energy in the curvature of spacetime (roughly speaking). That’s why Alan Guth calls the universe “the ultimate free lunch.”

  6. Biff says:

    Its 2017 everyone even a 6 year old can record a video. If the lecture was worthwhile at least record it with a iphone or something similar.

    Yes the slides does take something away for sure.

  7. What does ‘before the big bang’ mean?

  8. Gaehazzi says:

    How can one “remain optimistic that we are actually making progress here” and be “pretty hopeful that within [one’s] lifetime we’ll have settled on a leading theory for what happened at the very beginning of the universe” when there’s zero chance of ever putting the related theoretical conclusions to an observational test?

    Though our theories may be beautiful and elegant and what-not, they are necessarily on a reasonably short leash anchored on a realm in which we can challenge them with facts. Otherwise, we are back at pre-Galilean/pre-Baconian “science.” (Yes, Popper and falsifiability again.) Note that our Lambda-CDM cosmology and big-bang theory are rather recent developments. How dare we believe that they are conclusive and secured when their only merit is that they are the latest? What’s in the offing if quantum gravity grand unification is realized? Could we have been justified in believing in “making progress here” also prior to 1998?

  9. Arnd Rosemeier says:

    Sean, you said that the universe probably has zero energy and i understand the reasoning. But it seems that the negative energy is growing over time since its just a property of the expanding space time while “normal” energy stays the same. So how can that be?

  10. William Plettner says:

    Never wise to bet on your own lifetime.

  11. Steve Ruis says:

    I get a 404 error when trying the linked slides.

  12. Charles A. Martinson III. says:

    I have thought about this question for some time and necessarily take a slightly different point of view about the nature of Universal topology based on genetic lawfulness and the physical properties of continuity.

    Base on simple topological principles, I would seriously argue that our classical notions of Time and Space are fundamentally flawed. I would strongly suggest that Both Time and Space are emergent properties in the scalar mechanics that formulate our quantum reality and underpin initial Inflation and the structural relationship between matter and energy.

    By those principles, there is only one Universe of limited dimensions, specifically one with the reiteration of only four spatial dimensions and a constantly failing fifth dimension which manifests the properties of Time. Due to these genetic limits, these properties are patterns of infinite recursion. The process marks both the beginning and end of the inflation of the Big Bang.

  13. simon morley says:

    Please Sean, time is NOT a mystery – demonstrably and categorically !!
    IF you accept the basic premise that the only evidence we have for ‘time’ are events happening. No one has yet to provide any different evidence to my knowledge. (Event = occurrence of change)

    So, Time has two core meaning:-
    1. Its a non-specific set of [quantum] events i.e. a mass nouns (like ‘traffic’is a non-specific set of vehicles)
    2. Its the the scale or abstract framework that we use to calibrate, index and reference events.

    How is this categoric? Because every use of the word Time falls into either one or the other of these two separate meanings.

    So – time is not fundamental. The fundamental is [quantum] events…zillions and zillions of them.
    AND time is only universal in the sense that the metric system is universal – its a measuring scale in definition 2

    Is time emergent. Can an abstract noun be emergent? What is emergent is the continuum of change….the zillions and zillions of quantum events occurring.
    Asking was there time before the Big Bang is like asking was there traffic before breakfast – well only if there were vehicles on the road before breakfast, because ‘Vehicles’ is the fundamental NOT traffic. If no vehicles, then there was no traffic – the word traffic is irrelevant / redundant until the first vehicle appears.

    So – were there events before the Big Bang? If yes, then Time is relevant – if no, then the word Time is irrelevant / redundant. Simple.

    I do hope this helps

    (oh, and by the way, you cant merge two paradigms of incompatible nature – so space-time must ergo also be abstract)

  14. Torbjörn Larsson says:

    I would have thought that the entropy of the universe was handled like I think energy is which unless I am mistaken is an average energy density. In which case the local description is consistent with the global I take it, of having low (zero) energy density of an inflationary era and low entropy density of an inflationary, cold and empty, era. But then slide 4 instead compares the smooth evolution with some chaotic states. (Linde’s chaotic inflation?) And claims we can know about and measure such states.

    Where do those states come from and how is the measure problem [ https://en.wikipedia.org/wiki/Measure_problem_(cosmology) ] solved in making that claim? I don’t really see how it relates to the inflation era, which after all could be of infinite volume and indefinite time (as in one of Linde’s papers). In that case the average entropy density is that of inflation and unless I am mistaken our putatively local universe would return to that entropy density – of cold and mostly empty space – eventually. Seems to me general relativity has the same apparent violation of local conditions regarding entropy as regarding energy. I.e. we expect increasing entropy of matter configurations – unless it is exported away as here on Earth – and decreasing energy of photons as they travel freely (due to the cosmic expansion).

  15. Torbjörn Larsson says:

    @Gaehazzi: I cannot answer for Sean of course, but my impression was that he leans on the more and more popular tests that follows from bayesian learning. So you could perhaps test for internal consistency, say. In any case it seems to me even harder to make and test the extraordinary claim that this is the end of test and of accumulating knowledge. Or rather, is it not our experience – speaking of bayesian learning – that such claims always fail trivially with time?

    As for the current cosmology I always thought that the fact that it is the first self-consistent cosmology would indicate its physics should stick. (To be expanded on, for sure.) E.g. before inflationary cosmology cosmological ages – universe age, star ages, et cetera – were uncertain with a factor of ~ 100 % (a factor of two). And now we are looking at something < 0.3 % in general uncertainties if I remember correctly! (And the recent discovery of a dusty foreground that had not been cleaned away should decrease that further, to start with.)

  16. Yair says:


    If I understand correctly, the reproducing universe model is based on a de Sitter space that has a privileged direction of time determined by the distribution of dark energy (the cosmological constant), whereas the fundamental Einstein field equations are symmetric. Is there a solution to the Einstein equations that does NOT break this fundamental symmetry? Can something like the reproducing-universe model be constructed in a way that is symmetric in all four dimensions?

    I would very much like such a symmetric scenario. I’m imagining a manifold without any global time direction, within which there are pockets of time each in a random direction or from which emerge bubble universes each with its own time, not in any “direction” in respect to the original spacetime manifold. Such structures respect the underlying symmetry of the fundamental field equations. That seems aesthetically appealing.

  17. Gaehazzi says:

    @Torbjorn Larsson 2nd: Self consistence of physical theories is a necessary condition, by no means a sufficient one. A merely self consistent theory may be attacked by “the enemy” as being “a harmony of illusions” [Merton.] I hold that questions about what happened before the big bang, or about the multiverse, which not even an in-principle thought experiment can decide, are beyond physics, just as the question of how many angels can dance on the tip of a needle.

  18. JK White says:

    If we were to have a do-over of the Big Bang—under any of the models—- would we end up with the same physical laws/principles/hypotheses/conjectures we have uncovered so far. For example: Is it even possible to have a Big Bang model without a low entropy or time–or does that question even have a meaning?

  19. Charles A. Martinson III. says:

    Sean, I must apologize for my first comment here. I jumped out of your context to an unsupported assertion. I should have responded directly to your postulations.
    First, I agree with you that we don’t know much about the beginning of the Universe; however, we do know more than we might think based upon general observations so we can and do explore a range of possibilities. Like you have explained here, you only listed four categories, a few with characteristics which might relate to the history of our Universe.
    Second, I also agree with you that quantum mechanics must play an important role for understanding the genetics of emergent properties, including the question about the possible number of dimensions in Hilbert space and whether time and space are fundamental or emergent. Or more importantly, how many dimensions does our Universe manifest?
    Third, I also believe that an answer to the question of low entropy must be essential to a solution.
    So how do we get back to the inflation marked by the Big Bang? What is the initial state?
    Obviously, our Universe presents us with the opportunity for identifying and measuring differential information about comparative associations, along with lawful processes for the interaction of things. A very complex set of circumstances. I would also note, in a topological and genetic sense, the Universe is a singularity no matter how any of us might want to break it apart into bits and pieces. So, I would suggest that everything we can detect and infer through observation emerges through phased processes of evolution which must move from simplicity to greater complexity. To the point, our Universe is in a constant state of flux due to its inherent genetic instability. Due to our dependence on these emergent properties, you cannot use anything or any force within this state to examine the Universe’s initial state except in principle.
    The initial state can be conceptually reached by removing all differential information and reconstituting entropy back into its simplest form. This reversed process of evolution results in the lowest possible state of entropy. It is a topological singularity with no boundaries, no countervailing forces. It is pre-dimensional with unlimited degrees of freedom. Genetically, it is highly unstable. It is the most fundamental form of topology. It is a profound statement of unity with no differential information. It cannot be thermodynamically defined. Any kind of displacement will lead to greater complexity.
    We mostly know its current state. The real question is how does it get to our state of dynamics?
    Sean, to help resolve certain questions, I will note that evolutionary processes of emergence must be infinite in some sense. I would suggest that this form of progression precludes Time and Space as being fundamental. If this is so, then there are arguments that eliminate your options for one, two, and three. Option four is also eliminated in the form that you present it. However, processes of recursion are permitted, just not baby universes.
    So, the question for you becomes what kind of dimensional patterns and processes emerge from this state? What underpins and scales quantum processes?

  20. Mike Roberts says:

    Thank you Sean, I learned new things from the slides that I haven’t heard explained before.

  21. Charles A. Martinson III. says:

    Sean, one other point needs to be address . . . the issue of “fine tuning” and low-entropy states. I would stress that in principle a conceptual process that deconstructs emplacements of differential information within the context of physical topology only results in the simplest form of topology possible. It is not “fine tuning” in the sense of adjusting parameters to fit a desired outcome. If fact, all it does is present the simplest form of “existence,” a necessary starting point for genetic evolution, the proverbial blank sheet. The fact that the notion of “entropy” is then presented at its lowest possible state is only coincidental to this topological view of basic structure, presented as a singular unbounded state. It is void of any kind of information, except for its existence as the simplest form of topology.

  22. dmck says:

    The bottom line of every slide is cut off in my browser (Firefox on mac OSX).

  23. Richard Gaylord says:

    can you make your slideshow available as a pdf so it is easier to download rather than having to take a screenshot of each slide?

  24. marten says:

    How could a singularity come into being in the first place? As I understand it there is a quantummechanical problem. Are Polchinski’s firewalls illustrative for this problem?

  25. John B says:

    I think the real problem here is trying to determine if spacetime is an actual thing or not; or if spacetime is actually something that the early universe could have done without and still be the same. If it is an actual thing, then it would have had to have been created. If it is not an actual thing, then it wouldn’t have needed to be created. I believe that it is an actual thing, so then it would have to have been created by some kind of process at some point. I don’t think anything could have just existed forever, because it could never get to our time-frame from having to have already experienced an infinite duration of time (by the definition of infinity).

    If spacetime can be warped or dilated, then it would have to be an actual thing. Relativity itself could be indirect evidence of the existence of spacetime and the experiment that proves it. How would that be so much different that saying that the Higgs Boson exist, because of the indirect evidence? No one has ever seen or experienced a Higgs Field, but scientist all accept that it exist from finding the Higgs Boson, from just seeing two photons that shouldn’t have been there. Then no one had any previous bias about the Higgs Field like they do with with spacetime…

    Maybe, the scientific method failed us on being able to begin on problems like the Big Bang, because it doesn’t say you can use the same old experiments to prove something different as indirect evidence. Then the only way to start this problem would be to just start making assumptions and see if those assumptions give you anything which can then be proven. At some point, abductive reasoning will have to be able to find a way into being used in science before it can be completed.

    In the twin paradox, one twin actually gets younger. He/She doesn’t just appear to look or feel younger. An atomic clock used in this experiment will actually read a different time when placed side by side in the same frame of reference with the other. Time would actually be altered for one clock, but how could something be altered in this way if it doesn’t even exist?