Welcome to the Multiverse

Multiverse Mania makes the big time this week, with a cover story Welcome to the Multiverse by Brian Greene in Newsweek. While the title indicates that the Multiverse is here and part of our scientific world-view, the subtitle is a bit cagier: “The latest developments in cosmology point toward the possibility that our universe is merely one of billions.”

The article is pretty uniformly a promotional piece for multiverse mania, although buried fairly deep in the piece is something a bit more skeptical:

because the proposal is unquestionably tentative, we must approach it with healthy skepticism and invoke its explanatory framework judiciously.

Imagine that when the apple fell on Newton’s head, he wasn’t inspired to develop the law of gravity, but instead reasoned that some apples fall down, others fall up, and we observe the downward variety simply because the upward ones have long since departed for outer space. The example is facetious but the point serious: used indiscriminately, the multiverse can be a cop-out that diverts scientists from seeking deeper explanations. On the other hand, failure to consider the multiverse can place scientists on a Keplerian treadmill in which they furiously chase answers to unanswerable questions.

Which is all just to say that the multiverse falls squarely in the domain of high-risk science. There are numerous developments that could weaken the motivation for considering it, from scientists finally calculating the correct dark-energy value, or confirming a version of inflationary cosmology that only yields a single universe, or discovering that string theory no longer supports a cornucopia of possible universes. And so on.

I don’t see how we’re anywhere near finding such a version of inflation or getting rid of the string theory landscape, so the only hope of getting any evidence against the multiverse seems to be to calculate the cosmological constant. The multiverse thus looks to be pretty much impregnable and immune to any conceivable scientific challenge. A few years ago, pieces like this would hold out hope that the LHC would discover something encouraging for the multiverse, but now the LHC isn’t even mentioned. The only possible positive evidence suggested is seeing remnants of bubble collisions in the CMB, but the very likely eventuality of not seeing such a thing doesn’t count as evidence against the multiverse idea.

So, I fear Brian is right: Welcome to the Multiverse, physics is going to be stuck with it for a very long time…

This entry was posted in Multiverse Mania. Bookmark the permalink.

94 Responses to Welcome to the Multiverse

  1. ComeOn says:

    Completely agree with Bob Levine above. In academia you can go years without hearing anyone mention the word “multiverse”. But in pop science, and blog science, it’s infinitely more prominent.

  2. jg says:

    The Witten Lecture ‘String Theory and the Universe’ due at String 2012 has been given by him various times in the previous couple of years, you can watch one version online at IOP site:


  3. Giotis says:

    Yeah I’ve watched this lecture before; I wasn’t quite remembering the title though. But even if his 2012 public lecture is exactly the same to this one there would be a crucial difference; audience’s questions. Even if Witten does not want to elaborate or take sides the multiverse controversy would be certainly brought up by the audience and we may witness some interesting answers.

    Although as I said I’m not very optimistic about it. My impression so far is that Witten is more or less agnostic about the subject. Like in this lecture he is very cautious and often says that we need more clues and more work needs to be done in order to decide whether the multiverse picture is really a prediction of the theory.

    Nobody could disagree with that of course taking into account that M-theory is poorly understood. Yet again at this point it’s hard to imagine how the multiverse paradigm shift could be evaded within the context of String theory. But you never know…

  4. Truth says:

    As Brian Greene points out, the multiverse idea appears in numerous directions of research. The properties of GR, quantum mechanics, inflation, string theory, dark energy, etc, all point in the same direction and provide independent circumstantial pieces of evidence for it. Right now it is our best idea about the world, given the information available to us. It is conceivable that it is wrong, but at the moment, it is more likely to be right than wrong.
    Anyone who shows hostility to it are really just people who lack some level of imagination and want to just cling on to old archaic ideas.

  5. David Nataf says:


    How do the properties of general relativity suggest the multiverse?

  6. Peter Woit says:


    Presumably “Truth” is referring to eternal inflation models, which in his or her mind are now part of GR.


    If you’re representing Brian’s argument as it’s “more likely to be right than wrong”, I think you’re seriously misrepresenting what he has to say, which includes

    “That there are ways, long shots to be sure, to test the multiverse proposal reflects its origin in careful mathematical analysis. Nevertheless, because the proposal is unquestionably tentative, we must approach it with healthy skepticism and invoke its explanatory framework judiciously….

    But as with all rational bets, high risk comes with the potential for high reward…

    The multiverse proposal might be wrong. But it might also be the next step in this journey, unveiling a breathtaking panorama of universes populating a vast cosmic landscape. For some scientists, including me, that possibility makes the risk well worth taking.”

    Saying an idea is “high risk” means it’s more likely to be wrong, not more likely to be right. Also unlike you, I suspect that Brian’s point of view on the majority of the physics community is not that they

    “are really just people who lack some level of imagination and want to just cling on to old archaic ideas.”

    but that they have perfectly good reasons for being skeptical of a very speculative idea there is little evidence for, no matter how many articles about it appear in the popular press.

  7. Truth says:

    Peter, my first comment was my summary of Brian’s view. But from “Right now it is…” it was my own opinions.

    As to your final comment that there are “perfectly good reasons” for thinking that there is only a single universe, I have never heard it. Given that the laws of physics were evidently sufficient to create at least one universe, I dare say it could create multiple, if not infinitely many. I suspect it is much harder to show that there is only one universe than many, especially when, GR, quantum mechanics, inflation, etc, are pointing in the other way.

  8. Truth says:

    Actually, I would add that in fact you are seriously misrepresenting Brian’s view. He does not say that the multiverse idea is likely wrong, he says that there are a few possible ways to test it directly (such as bubble collisions in the CMB) and it is unlikely that we will see that direct evidence. So, according to Brian, it is high-risk to pursue a proposal for direct detection, which may not materialize. But he is certainly not saying that the multiverse idea itself is unlikely to be true, it just may be difficult to test.

  9. Peter Woit says:


    I didn’t say there are good reasons for believing there is only one universe. I said there are good reasons for being skeptical of the scientific multiverse proposal Brian is describing (and I suspect he might agree). The “high risk” in such proposals, which I think is what he is referring to, is that they may be inherently untestable in any way, and thus not science. A large majority of physicists see discussing the multiverse as an empty activity that is not science, since it is not testable, and claims to believe in either one or multiple universes are pseudoscience absent a plausible way to deal with the testability problem.

  10. Pravda says:

    “Truth” is, himself, proof of the existence of alternate universes, as he is obviously posting from one.

  11. Truth says:

    Certainly there are ways to test it, such as CMB data. So that kills your idea right there.

    Secondly, even without direct evidence of that sort, the multiverse is an extrapolation of other well tested theories. Extrapolation is not only part of science, it is THE WHOLE POINT of it, as stated very eloquently by Feynman.

    Anyone who goes around saying it is not science is being silly, and not engaging in the real issues at all.

  12. Bernhard says:


    the multiverse is a speculation, not a theory. Speculations are a part of the scientific processes but not sufficient to be consider science. A scientific theory is something more rigorous and ultimately it needs to meet the criteria of testability and it must be falsifiable. Of couse one has the right to speculate about multiverses, say time is multidimensional and pretty much whatever you want, but without a theory and solid quantitatively predictions this stays in the realm of speculations. Some may find fascinating but others will find it boring, but anyway there is ever reason to be completely skeptical about it. Could be, could not be, without solid predictions, there is little to discuss.

    The point you raise is that on the contrary, bubble collisions in the CMB are a real prediction of it and therefore it can be tested. This is certainly a scientific point you are raising. More specifically I believe you are talking about things like this paper: http://arxiv.org/abs/1012.1995

    If you are read this paper carefully you will see there are many caveats and in particular their argument rests on azimuthally symmetric temperature modulations, which as the authors observe “are not unique to bubble collisions”. The paper is at best inconclusive. So, it might be that the multiverse picture will evolve into a robust scientific theory, but until then selling the idea as scientific consensus and making noise in popular articles confusing the non-scientists is not the best idea to do this, regardless if one thinks one or 20 universes are “more likely”.

  13. Truth says:

    Just because the CMB thing is inconclusive right now, certainly does not make it unscientific. There are numerous inconclusive things in the world. I dunno…GUTs, Higgs, what killed the dinosaurs, the origin of life, you name it…it is absolute garbage to suggest that work on these topics is unscientific because these issues remain inconclusive. The WHOLE POINT of science is to work on inconclusive things. That’s what scientific RESEARCH is. We don’t know the answers today, for instance, if there is a multiverse, if a meteor killed the dinosaurs, how the pyramids were built, etc, but we can work as hard as we possibly can to try and find out. Maybe we’ll know in 1, 10, 100 years from now. That’s science.

    Sometimes the evidence fir things can be fairly direct and sometimes it is rather indirect, for instance, if it comes from extrapolating the predictions of a well tested theory. As I said before, that is the current situation for the multiverse proposal. It comes from extrapolating the predictions of GR, quantum mechanics, inflation, etc. extrapolation is what science is all about. To call it unscientific is simply wrong.

  14. Peter Woit says:


    We’ve seen this misuse of the term “test” endlessly with string theory. In the conventional meaning of the term, a “test” is something you can fail, with negative consequences for your theory. Even the most enthusiastic multiverse supporters don’t think there’s much of a chance that new CMB data will actually vindicate multiverse theory. When Planck data comes in, sometime within the next year, I predict it’s not going to contain evidence of bubble collisions, but no multiverse proponent is going to give up on the idea because of that.

    You seem to misunderstand what the word “extrapolation” means, in Feynman’s usage. He was referring to applying a tested theory to new domains, not a new theory.
    A multiverse with different laws of physics is not an extrapolation of any tested law of physics.

  15. Truth says:

    And of course, the multiverse did make a prediction. It predicted a small, but non-zero cosmological constant of the same order of magnitude that was later observed. Ok. So stop saying it does not make predictions. It did. It was observed. It is our only understanding of it. You may not accept this evidence as enough. Fine. Some people do. Some people don’t. That is science. Get it?

  16. Truth says:

    100% wrong Peter. The multiverse is an extrapolation of inflation, which combines quantum mechanics with general relativity and some scalar field source. It is a tested theory.

  17. Peter Woit says:


    The theory of inflation you are talking about involves a choice of potential for a single scalar field. If you extrapolate it way beyond what there is experimental evidence relevant to, you can get bubble universes, all with exactly the same physics as ours. That there’s an infinity of bubble universes out there with exactly our physics is possible, but probably untestable, as well as not being very interesting since it explains nothing about our own. If you want to claim that the CC and other parameters are different in every different universe, and then use anthropics to “predict” the CC or anything else, (yes, lots of people think anthropic “predictions” don’t deserve to be called “predictions”, for good reason…) you need to invoke a new theory, like string theory, not extrapolate a conventional, tested one.

  18. Henry Bolden says:

    Multiverse theory predicts the existence of a universe where multiverse theory has been proven to be incorrect. Greene’s paradox?

  19. MultipleMulti says:

    Re Witten, does anybody know why he’s working on Khovanov homology now? What do knots have to do with String theory? Do the strings form knots? Do the Feynman histories form knots?

  20. Truth says:

    Well it seems we have made some progress Peter. At least you have agreed that by extrapolating inflation we get a multitude of universes. Ok, so that much is settled. It is an extrapolation and it is part of science.

    Here your complaint is that the laws of physics are still the same though. Well, indeed there are different levels of the multiverse. Inflation forces on us a level, where there are many universes, although the laws can be the same. The universes do differ, however, in their distribution of matter. Eternal inflation generates every consistent distribution of matter. For instance, nc,uding universes with different values of the amplitude of density fluctuations.

    As to having different laws of physics, for instance with different values of the CC, we indeed need a potential with multiple minima, which is a much more severe extrapolation. But it is an extrapolation, nevertheless.

  21. Peter Woit says:


    You seem to know enough to know very well that the 10^500 different minima necessary to make the anthropic CC argument work don’t come from any tested theory, even one as little tested as inflation. The KKLT mechanism and any of its variants invoke a huge array of speculative ideas from string theory, none of which have any experimental evidence for them. KKLT is an extrapolation of our current theory only in the same sense that any complicated, untestable, speculative idea for a theory of everything is an “extrapolation”. I suspect Feynman is rolling over in his grave to have his name invoked in this way, it’s pretty clear what he would think of your claims.

  22. Truth says:

    Again, I am glad that you admitted in your previous comment that inflation implies a multiverse from extrapolation.

    In my most recent comment, I simply said that multiple minima are needed for different laws of physics. I didn’t say anything about KKLT, etc. It is a straw man argument for you to criticize KKLT as a way of attacking my quite reasonable, and largely separate, comment.

  23. vmarko says:

    I am glad that you admitted in your previous comment that inflation implies a multiverse from extrapolation.

    Khm, inflation does not imply a multiverse, at best it only suggests a multiverse. Namely, one can have inflation without multiverses. In loop quantum cosmology, for example, one doesn’t even need an inflaton field to produce inflation. And one stays in a single universe after the inflationary period. You can take a look at gr-qc/0206054 for some details.

    HTH, :-)

  24. Marty says:


    You apparently believe that inflation is correct in its current form, a form which does appear to imply “many universes.” This form may or may not ultimately be correct, but it is certainly not a “slam dunk” case. Here are a few links to talks and papers that you might find enlightening; they may help temper your confidence a bit:

    Paper by Roger Penrose (1989; behind a paywall, unfortunately)
    Penrose talks at Princeton, three total (inflation is #3; look at October 2003)
    Steinhardt talk at Perimeter Institute, 2011
    Turok talk at PI, 2011
    The Inflation Debate (Sci. Am. article on Steinhardt’s Princeton site; free)

    There is very strong observational evidence for inflation, but as you will learn if you read and watch the talks the evidence is for the much simpler picture that emerged in the early 1980s. Since then, some severe theoretical difficulties with the inflation idea (not observational problems) have become apparent, and it isn’t obvious how to fix them in spite of lots of effort. These problems have nothing to do with string theory, by the way — they are problems with inflation models themselves.

    There is essentially zero observational evidence for eternal inflation, however. It is only implied by the inflation idea if general relativity plus quantum mechanics is the right framework to use prior to the Big Bang. If you study the theoretical problems, you may agree that it doesn’t really make sense to use general relativity in a regime where it presumably breaks down, i.e., at and beyond the Planck scale.

    Regarding observational tests of bubble collisions, it is extremely optimistic to think they will ever be successful. If bubble universes are real (questionable!), almost all such collisions that would be visible today would need to have happened at very early times, and their effects would generally be restricted to an extremely small angle in the sky, far too small to be discernible, much less provide useful observational evidence (the “look elsewhere effect” …). The only real hope is that relatively recent collisions occurred and left visible traces, but those are highly unlikely.

  25. Peter Woit says:

    vmarko and Marty,

    The situation of experimental evidence for inflation is complicated, and multiverse maniacs typically start from some outrageous claim about how well established the theory is. It’s very rarely pointed out though that even if you accept the most optimistic versions of the inflation story, that actually has nothing to do with the “extrapolation” that “Truth” and others are advertising. Even eternal inflation doesn’t give you “different laws of physics” in different universes, it just gives you lots of universes with the same laws of physics. To get “different laws of physics: you need to introduce something like the hundreds of moduli fields of string compactification schemes, the exact hideous mess that has turned string theory unification into a failure.

  26. Truth says:

    Vmsrko and Marty,

    Most of what Peter says here is wrong. As I mentioned before, eternal inflation gives you different universes with different properties, e.g., different amplitude of density fluctuations, different spectral index, curvature, etc. These universes would look quite different. It is a type of multiverse.

    To get different microscopic physics, you do not need 100s of stringy moduli fields as Peter is advocating. A potential with just 2 local minima would suffice to have 2 different kinds of laws; that is enough for one to talk of a multiverse. And it does not rely on the validity of string theory at all, only quantum mechanics and GR. The only place where string theory enters the discussion is that it re-affirms this view with many, many fields and minima.

  27. Peter Woit says:


    You know very well that cooking up an inflaton potential in standard inflation theory to have extra minima is not what is being advertised. This gives no anthropic explanation of the CC for instance, which is the main “evidence” given for this scenario..

  28. Truth says:

    Peter, I have addressed all these issues clearly and correctly in my previous posts. Inflation is well tested and when extrapolated it predicts a basic type of multiverse which look different. A potential with 2 minima gives different microscopic laws and is a more crude extrapolation. While many, many minima of the string theory type is a much more severe extrapolation, which is less likely to be true, but comes from extrapolating wuantum mechanics and GR to the extreme, as it is the only known consistent picture. But even if you hate the latter, a more basic type of multiverse does come from extrapolation. That is the point I am making. If others wish to focus on other issues, that is their business.

  29. Peter Woit says:


    You just keep repeatedly abusing the term “extrapolation” (5 times in your latest comment). Repeated explanation from my why this is abuse of the English language and the memory of Richard Feynman would be pointless.

  30. Bernhard says:


    can you point out one (or more) reference (s) that corroborates your claims? I think your use of the word “extrapolation” is not entirely clear, at least to me.

  31. Pingback: Uncommon Descent | The multiverse: Just an excuse to sell books now?

  32. ScentOfViolets says:

    I suspect it is much harder to show that there is only one universe than many, especially when, GR, quantum mechanics, inflation, etc, are pointing in the other way.

    Looking at these last exchanges, it appears that “truth” is substituting some , er, nonstandard definitions of common terms for the usual ones and not telling anyone. The particular term I’m drawing attention to here is “skeptic” or “skeptical”, which “Truth” seems to think is a synonym for “disagree”. It’s not.

    For example, I’m skeptical of any sort of theorization which relies on the existence of a multiverse to make it work, but that doesn’t mean that I disagree with the notion that there are multiple universes instead of just the one. It merely means that I’m agnostic wrt the number of universes.

    Science being what it is, being a skeptic means that I don’t have anything to prove; the burden of proof in this particular instance is on “Truth”. And so far, he’s not being very convincing. Name-calling is seldom an effective way to get people to agree with you ;-)

  33. Truth says:

    Peter, above you admitted that when inflation is extrapolated you get many universes. These universes have qualitatively different properties, such as the curvature or amplitude of density fluctuatuions. So even you have conceded that the extrapolation of inflation leads to a multiverse. Here even you have used the term “extrapolated” in the same way as me, and indeed in the same way as Feynman. It is ridiculous for you to then mangle the issue by sneering at KKLT, or other such things, as though that were the basis of my argument which I have clarified repeatedly and consistently.

  34. Truth says:

    Scentofviolets, I never used the term “skeptic”, and I never did any name-calling.

  35. harryb says:

    Truth et al,

    May I quote from Feynman directly – The Character Of Physical Law, Chapter 7, Seeking New Laws. I quote at length because the original text it is well written and compact:

    “In general we look for a new law by the following process. First we guess it. Then we compute the consequences of the guess to see what would be implied if this law that we guessed is right. Then we compare the result of the computation to nature, with experiment or experience, compare it directly with observation, to see if it works. If it disagrees with experiment it is wrong. In that simple statement is the key to science. It does not make any difference how beautiful your guess is. It does not make any difference how smart you are, who made the guess, or what his name is – if it disagrees with experiment it is wrong. That is all there is to it.”
    It seems the WHOLE POINT of science – according to Feynman – is for any smart guess to agree with experiment. The Multiverse may be a smart guess – but we need an experiment that rules it in or out.
    The Multiverse sounds smart, and may well be correct. But we have to be smarter that just posing the guess – the hard work as Feynman knew – is in devising the experiment that will obviously – to many – make it more than just a good guess. Guesses are cheap – experiments to verify them are, very, hard work.

  36. Anonyrat says:

    Since inflation was proposed around 1980, and Feynman died around 1988, surely we must have some of Feynman’s opinion of the idea instead of having to rely on extrapolation.

  37. Igor Khavkine says:

    @Anonyrat: Something slightly relevant. Here’s Feynman weighing in on the somewhat larger question of the initial conditions of the universe. My interpretation: curiosity with extreme caution.

    Feynman: Take the world from another point of view (3/4)
    http://www.youtube.com/watch?v=uNOghidK2TY (from 7:10)
    Feynman: Take the world from another point of view (4/4)
    http://www.youtube.com/watch?v=mvqwm6RbxcQ (until 1:18)

  38. Tim van Beek says:

    jg wrote:

    The Witten Lecture ‘String Theory and the Universe’ due at String 2012 has been given by him various times in the previous couple of years, you can watch one version online at IOP site:


    Thanks for the tip. That is the standard history and motivation of string theory, so not very interesting to me, but probably for everyone who never heard the story before. I think it is safe to assume that Witten won’t invest much time to update the talk for Strings 2012 (no reason to).

    Nevertheless, if I go to the strings 2012 talk and get the chance to ask a question, I’ll probably ask about what he thinks a young PhD should specialize in, or what he is currently working on and why, or something like that. But I’m taking suggestions :-)

  39. Bernhard says:

    Tim van Beek,

    I would be very curious to know Witten’s answer. Years ago (2004 if I’m not mistaken) a friend of mine asked him a similar question for which he answered (I’m quoting it this from memory so it’s more or less this): “Don’t know, but something with string theory”. Would even Witten give such an advice today?

  40. Chris Austin says:

    The British Astronomer Royal Sir Martin Rees, who seems to favor a multiverse
    picture, stated on page 115 of his 1999 book “Just Six Numbers” that a
    value of the relative amplitude Q = (delta rho)/rho of the primordial
    fluctuations substantially larger than 10^{-5} would make the universe too
    violent for life. This appears to contradict Paul J. Steinhardt’s
    statement, on page 43 of the Scientific American article by linked by Marty,
    that the universe is smoother than it needs to be to support life, so that
    anthropic selection cannot explain the value of Q. Is Rees’s statement now
    discredited, or is there no consensus on the matter?

  41. Peter Woit says:

    Tim van Beek,

    There is something new since 2010, the fact that no SUSY (or anything else relevant to string theory) has shown up at the LHC. For years Witten and other string theorists have held out hope that the LHC would find something that would vindicate string theory (at least low energy SUSY). I’m curious to know at what point Witten will give up on this hope, and if that happens how it will change his view of string theory.

  42. Shantanu says:

    Peter and others, there is some discussion of multiverse in the panel debate on
    emergent spacetime at KITP conference.

  43. Peter Woit says:


    There’s not much about the multiverse there, but interesting if you want to keep up on the latest thinking about emergent gravity. I gave up after watching for a while, seemed to me there was nothing new going on in this subject, in particular nothing new about the obvious question: “emerges from what?”

  44. Nick M. says:

    Let me preface my post by saying that although there have been over ninety responses to “Welcome to the Multiverse”, it has all been a very interesting and lively read.

    So, I fear Brian is right: Welcome to the Multiverse, physics is going to be stuck with it for a very long time…

    A long time indeed, as it looks as though Stephen Hawking (along with James Hartle and Thomas Hertog) aren’t quite ready to give up on the Multiverse/String Theory paradigm anytime soon either; see version 2 of the this paper posted in arXiv on May 30th titled “Accelerated Expansion from Negative Lambda”. Incidently, I found this link via the New Scientist article “Hawking’s ‘Escher-verse’ could be a theory of everything”. (Note that you will have to be a registered user to completely read the New Scientist article.)