First Hint That the Multiverse Really Exists

Multiverse mania rolls along, with New Scientist this week running a cover story entitled Touching the Multiverse. They advertise the story by claiming that “we reveal the first hint that the multiverse really exists”.

It turns out that this is a promotional effort for the work of Raphael Bousso, with the first hint that multiverse really exists a paper of his from more than 3 years ago that purports to “predict” the observed value of the cosmological constant. This in some way improves on Weinberg’s 1987 anthropic argument, which to this day remains about the only piece of evidence backing up multiverse mania. The New Scientist article also reports on his more recent attempts to better justify the 3 year-old calculation with this decade’s buzz-word (“entropic principle”) as well as that of the last decade (“holographic”).

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

30 Responses to First Hint That the Multiverse Really Exists

  1. MyrtleParker says:

    Uhm, you insinuate that the guy went back and revisited his old paper to update it with new buzzwords when the original three year old paper’s title is, “Predicting the Cosmological Constant from the Causal Entropic Principle?”

    You positing some form of new time travel theory here?

  2. Peter Woit says:


    You’re right, Bousso had the “entropic principle” buzzword several years ago, before other people. It’s the addition of “holographic” buzzword to the story that seems to be more recent.

  3. Bee says:

    I guess one could say the first hint that the multiverse exists is that the universe exists? 😉

  4. Dave Buettner says:

    I recently bumped into a couple of theorists I know. They were jubilantly celebrating a recent victory after a long and arduous effort. Using string theory, they were able to calculate the precise number of universes in a non-infinite version of the multiverse. Their work is yet to be published because, although they believe peer review will be a breeze, an embarrassing imaginary term appeared which they have yet to account for. However they believe they can work around it by adding another dimension. Anyway, the correct number (real part) of universes is 7,477.

  5. MyrtleParker says:

    Dave, I heard something similar although the details are a bit different. I heard that the team had used string theory to determine that the number of universes in the multiverse was in fact finite. However, they could not determine the exact number. Rather different starting assumptions lead to a landscape of finite values numbering near 10^500. Definitely finite though. So there is that.

  6. Peter Woit says:

    The notion of “counting the number of universes” in string theory is completely ill-defined. There’s no reason to believe that the various constructions of “string vacua” known now are the only ones. Some of these constructions give an infinite number of vacua. The only hope of getting finiteness is to put in a cut-off at a certain scale, beyond which you expect the approximate calculation to be wrong anyway. Any numbers you get depend on the arbitrarily chosen scale.

  7. MyrtleParker says:

    Peter, I should have put tags around the post above 🙂

  8. MyrtleParker says:

    Err, should be ‘sarcasm’ tags.

  9. Peter Woit says:


    This is a subject where it’s very hard to tell what’s a joke and what is supposed to be serious science….

  10. MyrtleParker says:

    Indeed… that’s what makes it so painfully funny.

  11. neo says:

    Infinite, schminfinite. Can we even know whether the set of possible universes is a countable set?

  12. Jeffrey McGowan says:

    Physicists know what countable means, really? 🙂

  13. Amanda Gefter says:

    Hi Peter,

    While I can’t speak to the way the article was marketed, I was really writing about Bousso’s paper from last year, Complementarity in the Multiverse (arXiv:0901.4806), which argues for a duality between a global picture of the multiverse and a local causal patch measure that only considers what a single observer could see. I found this really interesting, because if it holds it circumvents the problem of unobservable universes. It would allow you to do multiverse physics without reference to anything unobservable beyond your own horizon. I thought this was an intriguing new spin on the debate about whether the multiverse can be considered science.

    Anyway, I always love to hear your opinions on these matters! Thanks for mentioning the article.


  14. Peter Woit says:

    Hi Amanda,

    Thanks for the further explanation, I should have been a lot clearer in my posting. I just don’t see how this helps in terms of testability, since it doesn’t change Bousso’s original CC calculation or make any other testable predictions as far as I can tell. So, not much of a new hint…

    I should have also noted that you start the article describing George Ellis’s skepticism about the multiverse. That there are skeptics about this, probably in a large majority, doesn’t always get mentioned. Glad to hear you enjoy the blog!

  15. Ron Macken says:

    Maybe this is too basic a question, but what is leading physicists towards a multiverse model? I’m familiar with the anthropic principle, but am not sure if there are any other logical arguments for a multiverse.

  16. Peter Woit says:


    What is leading string theorists to multiverse models is not logic, but desperation. If your theory can’t predict the value of X, you can try and explain this away by invoking a multiverse, with the value of X a random artifact of our particular universe. With string theory this gets taken to an extreme: the theory predicts nothing, but instead of giving up on it, some string theorists are trying to invoke a multiverse as an explanation for not predicting anything.

  17. neo says:

    But doesn’t chaotic inflation theory also lead to a multiverse, independently of string theory?

  18. John Baez says:

    Jeffrey writes:

    Physicists know what countable means, really?

    Yes, but different physicists use different definitions, depending on how high they can count.

  19. Peter Woit says:


    The idea of “eternal” inflation leading to a multiverse has been around for a long time, independent of string theory. It didn’t get a lot of attention though, since it was (and is) untestable. By itself, it doesn’t say anything about observable physics (except perhaps, could be used to justify the anthropic CC argument). There were initial hopes that the string theory version of the idea would be able to predict something (e.g. whether supersymmetry breaking scale is high or low), but that has not worked out.

  20. Zathras says:

    Jeffrey McGowan says:
    March 4, 2010 at 6:42 pm

    Physicists know what countable means, really?

    Well, since irrational numbers aren’t observable, strictly speaking, a Physicist would say it doesn’t matter!

  21. Michael T. says:

    @Ron. It seems that the main argument in support of the multiverse is not so much scientific but philosophical. That is it can neatly side step the fine tuning argument invoking theism.

  22. noel says:

    It still seems that the multiverse explanation is better than fine tuning. Anthropic principle always > theistic principle. So it’s not like we’re not making progress..
    What’s better, assuming that our solar system has 9 planets and that other systems might have a different number, or that our system has 9 because it’s somehow unique and special? It’s always good to expect that not everything we see around us has a particular reason for being the way it is. Accidents do happen.

    The issue is not more philosophical here than anywhere else. For example, who predicted that the standard model will be described by gauge theories? No one. We just assume a priori that this works, and it does work, and you can make predictions from there. Similarly, who can predict which of the vacua describes our universe? No one, but if we find the right one in there, it might turn out that it works, and we could make other predictions from there.

    I’m not a fan of the multiverse, I’d like all parameters to be predicted from a simple principle, but nature doesn’t give a s*** about what I like and hope. Similarly, I’d like there to be a reason why gauge theories work, in the form of a very abstract and general no-go theorem. If people can’t come up with something like that, I’m not very sure we should be expecting a selection principle for the vacua. It’s the same problem, just different names.

  23. Peter Woit says:


    Personally I think the reason gauge theories work is that they embody a very deep truth about the relationship between mathematics and physics, one that we only partially understand.

    But, that reason is not the reason gauge theories are a successful physical theory. They’re successful because if you look at some of the simplest possible gauge theories, described by a small number of parameters, you get a consistent theory that makes an infinite number of very non-trivial predictions, and every one of these that can be tested comes out exactly right, on the nose. From a small input, you get a huge non-trivial output.

    The string theory landscape is just completely different. There, the simplest examples don’t look like the real world. You have to make your constructions more and more complicated in order to evade contradiction by experiment, until you have something so complicated that you can’t predict anything. This is exactly the kind of thing you expect to happen when you pursue a wrong idea.

    Sure, we don’t know a priori whether a fundamental theory exists that is more predictive than the standard model. Based on past history, pursuing the search for such a thing seems likely to work out. But, no matter what kind of theory you want to pursue, if you’re going to do science you need some way of testing your ideas against the real world. The utter failure of the string theory landscape to meet this kind of Science 101 standard is very telling. Comparing it to the situation with gauge theory is just absurd.

  24. Giotis says:

    Well you can’t test everything. There are physical and technological limitations. It’s just the way things are. People don’t do Physics of the 19th or 20th century anymore. This doesn’t mean of course that is forbidden to construct theories for certain regimes; you must keep making progress at all fronts and push the limits of knowledge. There are other ways to evaluate a theory i.e. Its Mathematical and theoretical consistency, its explanatory power and the lack of better alternatives.

  25. Peter Woit says:


    The problem with string theory is not that you can’t test everything, it’s that you can’t test anything at all, because it predicts nothing. Arguing that you’re going to replace experiment by mathematical consistency doesn’t help here, since you don’t have a mathematically consistent theory (you don’t what string theory is non-perturbatively, can’t even write down conjecturally what the theory is that’s supposed to be producing the landscape).

    All arguments like this about the string theory landscape that I’ve seen are just attempts to excuse what is an obvious failure. This is a really bad idea.

  26. Christine says:

    push the limits of knowledge

    You can only push the limits of knowledge if nature validates back. Otherwise, it is not knowledge that you are gaining, but only the ability to construct good (?) hypotheses.

    There are other ways to evaluate a theory i.e. Its Mathematical and theoretical consistency, its explanatory power and the lack of better alternatives.

    That is a distortion of the scientific method. What you describe is a working hypothesis, not a theory. Only when the hypothesis offers predictions and/or effectively explain observed phenomena is that you have a theory which describes that phenomena. Such a theory evidently may only be valid at certain regimes and be eventually generalized or revised or an entirely new theory be formulated in its place. This is what pushing the limits of knowledge is about. And it happens exactly because of the scientific method.

    Well you can’t test everything. There are physical and technological limitations.

    We all acknowledge that. So perhaps it is time to acknowledge the difference between speculation, hypothesis and theory as well.

  27. Giotis says:

    “The problem with string theory is not that you can’t test everything, it’s that you can’t test anything at all, because it predicts nothing.”

    Peter, the problem (?) with string theory currently is not that it predicts nothing but that it predicts everything.

    This “everything” though has explanatory power and that’s important. You can’t just close your eyes to that fact.

  28. Peter Woit says:


    Predicting everything and predicting nothing are the same thing.

    A theory that predicts either nothing or everything has no explanatory power at all.

  29. neo says:


    Predicting everything correctly is definitely better than predicting nothing. I do not follow you.


  30. Bob Levine says:


    ‘Predicting everything’ in this instance means predicting *anything*. If no what what the situation is, your theory predicts that it’s one of the possibilities—if it’s compatible will anything at all—then that theory makes no predictions, because it rules nothing out.

Comments are closed.