The Multiverse, Evidence and Theology

Yes, this multiverse business is tedious, but since it is becoming mainstream physics, with colloquium talks here at Columbia devoted to it, and the Columbia University Press publishing books about it, seems to me that someone at Columbia should be commenting on these, and I don’t see anyone else doing it. Will try to make this short.

Yesterday Matthew Kleban’s talk here was entitled Testing the Multiverse. The only part that actually really was about testing the multiverse was the part describing work on bubble collisions with other universes. This has been heavily advertised in the press, see here, here, here, here, here and many others. Kleban described some of these ideas, but when it came to the experimental testing part, he just briefly acknowledged that all searches for these things have come up empty. The only prospect for the future mentioned was the polarization data to be released later this year by Planck, which would give some new things to look at, but he seemed unenthusiastic that this would realistically lead anywhere. So, as far as the “testing” goes, it has been done and the tests failed.
The rest of the talk was about various inflationary models, including Kleban’s work on “unwinding inflation” (see here, here and here). Some of these models do have testable consequences, and many do lead to “eternal inflation”, so in such models you expect to continually produce new inflated universes, although with exactly the same physics. This is being sold as “testing the multiverse”, and string theory is brought in to justify lots of possible different physics in different universes, but this is not a testable part of these scenarios. What’s being advertised is a grandiose picture of the string landscape, laws of physics determined environmentally, etc., etc., but if you actually look at the product that you’re actually buying as “testable”, you don’t get any of the cool stuff. For slides of a somewhat similar recent talk by Kleban, see here.

A while back I acquired a copy of the new book Worlds Without End: the many lives of the Multiverse, started reading it and was planning on writing a detailed review. I soon got bogged down in the first half of the book, which is a detailed intellectual history of speculation about multiple universes (so lots about relevant parts of Plato, Aristotle, Lucretius, the Stoics, Augustine, Nicholas of Cusa, Giordano Bruno, Kant and others). Finally I realized I just didn’t have the energy to serious read this material. People with other interests and/or more time on their hands may find this quite worthwhile.
The second half of the book is devoted to the question of current speculation about physics, so more up my alley, but again I found it hard to focus on this. I fear it is only mildly interesting to see what a theologian/philosopher of religion makes of the current multiverse mania, not enough so though to do more than skim the text. From this skimming, what’s in the book is a lot of retelling (sometimes introducing misunderstandings) of the hype-laden tales of the multiverse told in dozens of books and magazine articles over the past decade or so.

Rubenstein ends the section about what physicists have to say with Tegmark, seen as having reached the final endpoint of the “Ultimate Multiverse”:

So some worlds will be linear, and some will be cyclical; some will be singular, and some will be plural; some will be infinite, and some will be finite; some will branch forward, and some will branch back. Some worlds will be manufactured, and some will be simulated; some designers will be kind and some will be cruel, some capable and some all but incompetent.
And, presumably, some of the set of all possible world will have a creator-god who breathes over primordial waters, who separates the sea from dry land.
How on earth did we get back here?

I take Tegmark’s vision as empty, so a good thing to ignore, but Rubenstein sees this as an opening for theologians to get back into the mainstream cosmology business, and the rest of the book focuses on this. With the boundaries between science and religion now gone, all sorts of possibilities open up for theologians. The final part of the book begins by invoking (just like Henrich Päs, who comes at it from the mind-altering drug rather than theological angle) Nietzsche:

Nietzsche concludes the Genealogy by expanding this vision, promising “all great things bring about their own destruction through an act of self-overcoming” (3.27, emphasis added). This promise then, has me wondering. If science can be regarded as the self-overcoming of a particular form of religion, might multiverse cosmologies be something like the self-overcoming of science? Might they mark the end of the fantasy that “science” has wrested itself free from “religion”, “objectivity” free from subjectivity, and matter free from meaning? After all, we have seen each of these multiverse cosmologies open onto metaphysics and mythology not in moments of lapse or weakness, but precisely where they are scientifically most compelling.

It seems that, unlike most authors, Rubenstein actually has got the story of multiverse mania right: it’s left conventional notions of science behind and entered into the realm of theology. We do, however, disagree about whether or not this is a good thing…

Update: Bogus claims about Multiverse “predictions” are now all the rage. For the latest, see the Caltech Quantum Frontiers blog, which has Yasunri Nomura writing about Making Predictions in the Multiverse. There of course are no predictions there, just mainly a discussion of the idea that many-worlds and the eternal inflation multiverse are somehow the same, an idea I continue to find unfathomable. Nomura doesn’t mention that he actually did have a prediction from the Multiverse (and someone made a movie about it…). The prediction was for a Higgs mass of 140 GeV, but of course when you’re in the multiverse business, wrong predictions are not a problem, they’re always true somewhere.

Update: For more on the multiverse front, Edward Frenkel has a review of the Tegmark book in this Sunday’s New York Times. He does a great job of explaining the problems with the way Tegmark is trying to use mathematics. John Preskill tweets in agreement (positive and negative).

Update: Nomura, when asked about experimental evidence for the multiverse, responded that the experimental situation is

not much different from some other situations—e.g. in the big-bang theory, inflationary cosmology, and Darwinism in biology

So, the scientific evidence for the multiverse is “not much different” than the evidence for evolution? And Tegmark thinks I’m the one in league with creationists…

This entry was posted in Book Reviews, Multiverse Mania. Bookmark the permalink.

62 Responses to The Multiverse, Evidence and Theology

  1. Laymammal says:

    “This idea of the multiverse, as we currently think, is not simply a result of random imagination by theorists, but is based on several pieces of observational and theoretical evidence.” A new kind of evidence: the theoretical kind. Noumura ends the same article with the oxymoron “concrete theoretical progress.”

  2. Peter Woit says:

    I don’t think this is the problem with Nomura’s argument. If you did make a lot of real, specific theoretical progress, you could call it “concrete” if you wanted. It this progress showed that theories you do have strong evidence for were tightly linked to theories that produced a multiverse, you might call that “theoretical evidence” for a multiverse.

    The problem is that there is no theoretical progress of the sort Nomura and others want there to be. The more we learn about the properties of a conjectural “M-theory”, the more we see that just about any physics at any scale is consistent with it. We are getting “theoretical evidence” here, but it is evidence against the picture Nomura wants to sell. Similarly, “string cosmology” seems to be able to give one any inflationary model properties one wants, so again, to the extent there is “concrete theoretical progress”, it is progress towards showing that these ideas just don’t work, in the sense that they don’t tell us anything about anything.

  3. OMF says:

    I guess your blog looks like it’s found another post-string hype purpose on the internet.

  4. nasren says:

    Geore Mayer said “And even more, this makes it so hard to think seriously about anything that scratches the metaphysical discussion – e.g. like the nature of numbers/sets/mathematical laws. There are a lot of questions in this area, but once somebody asks such a questions you can find a whole lot of deep thought guys jump up and down and sing the “Pegasus exists, because we have a word for it” song.”

    You obviously like your straw men: no one has ever given that as an argument for mathematical realism. Try engaging with the discussion that exists — it is smarter than you are supposing.

  5. Brian Dennehy says:

    From Frenkel’s review of Tegmark’s book:

    ‘Conceding this point, Tegmark replaces Mathematical Universe Hypothesis with Computable Universe Hypothesis: Only “computable” mathematical structures should be allowed. But this rules out all structures that contain infinity! In fact, he admits that “our current standard model (and virtually all historically successful theories) violate the C.U.H.,” which does not bode well for the whole idea, to say the least.’

    I thought this was quite a glib comment by Frenkel. Surely the question of whether the continuum field theories of modern physics are but approximations to more basic theories (which may themselves be ‘computable’) is currently unresolved.

  6. Lee Smolin says:

    Dear SUSY,

    Just to pick up a query raised earlier in this thread, LQG does not admit a multiverse, as the spatial topology of the universe must be fixed. Either we are in the cosmological case with compact spatial sections or there is a boundary imposed, as in asymptotically flat or AdS cases. The set up of canonical quantization does not permit the spatial topology of the universe to change as indeed the action principle is only well defined with either compact or well prescribed boundary terms and conditions. In particular you cannot have non-compact critical points of an action without proper boundary conditions being satisfied.

    One thing I have wondered is how this necessary technical subtlety is dealt with in eternal inflation models.



  7. Fellow Hampshire Alum says:


    Possibly silly questions, but I’m a mathematician not a physicist. By “spatial topology” I’m assuming you mean the topology of the subspace of space-time consisting of the space dimensions? And the point is that assuming canonical quantization, the topology of the 3D (I assume) cross section at any given fixed time is always the same? The boundary would be the spatial boundary, i.e. the boundary of the observable universe?


  8. Jesper says:

    Dear Lee

    when you write that the topology in LQG is fixed – is that really the case? The configuration space of generalized connections, which is the basis for the entire LQG quantization scheme, involves all possible smooth connections on a given manifold (with a given topology) as well as a all the non-smooth generalized connections (which form the ‘bulk’ of this space connections). Couldn’t some of these (generalized) connections correspond to geometries of spaces with different topologies than the original fixed one?

    For instance, assume that we start with the trivial topology (which seems natural), then I guess there would also be points in the space of generalized connections, which actually correspond to a non-trivial topology?

  9. Narad says:

    From Frenkel:

    This does not stop [Tegmark] from entertaining various situations involving multiple copies of “you” living in those universes, such as: “When the number of yous increases, you perceive subjective randomness. When the number of yous decreases, you perceive subjective immortality.” (The real question, however, might be, What is the number of yous who can understand what this means?)

    Funny, that’s not quite how it worked in The Man Who Folded Himself.

  10. Lee Smolin says:

    Dear FHA, yes to all.

    Dear Jesper, Even if the answer was yes, which would need to be shown, time reversibility of the dynamics suggests that if such a sector could split off it could rejoin…but I doubt Peter wants to host a technical discussion of LQG dynamics.


  11. Peter Woit says:

    Lee is right. I’m glad to have his comments here, but this isn’t a great place to try and host a technical LQG discussion. It would be pretty much off topic, and beyond my abilities to sensibly moderate.

  12. Jesper says:

    Dear Lee,

    yes, I think you are right and I don’t mean to suggest that there is a multiverse setting within LQG. But I have never been convinced that there can’t be topological non-trivialities arising in LQG.

    And sorry Peter – I knew it was semi-off-topic.