The Emperor’s Last Clothes?

Bert Schellekens has posted on the arXiv an extended 87 page argument for the anthropic string theory landscape, entitled The Emperor’s Last Clothes? While most string theorists find the existence of the landscape and the corresponding inability to get any predictions out of the theory about particle physics rather discouraging, Schellekens instead sees this as an argument in its favor:

Initially, when string theory was touted as the “theory of everything” around 1984, there were hopes it would lead to exactly the opposite: a unique derivation of all the laws of physics. Evidence that quite the opposite was true started emerging almost immediately after 1984, but most people chose to ignore it. In 2003, after important additional evidence had been found, Leonard Susskind published a paper [2] entitled “The Anthropic Landscape of String Theory”, which finally started a debate that should have started fifteen years earlier. What is at stake in this debate is not only the uniqueness of our universe, but also the fate of string theory as a fundamental theory of all interactions.

In my opinion string theory gives the right answer, and the fact that it does adds to the evidence in its favour. I can say this without being accused of trying to put a positive spin on the recent developments, because I actually wrote in 1998 [3] that I hoped string theory would ultimately lead to a huge number of possible choices for the laws of physics, a point of view I have been advocating since the late eighties. I reached that conclusion after having been involved in one of the first papers [4] pointing out that the number of possibilities was humongous…

We all hope to live during a time when big things are happening in our field, and I have never doubted that this is one of those things. I have spent the last twenty years trying to convey my sense of excitement to my colleagues, but with little success. But in the last few years I have been delighted to see more evidence coming in supporting this point of view, so that the mood has started to change. I hope this is the right time to make one more attempt.

Schellekens describes in great detail the anthropic argument and the arguments for the string theory landscape. He addresses some of the counter-arguments, especially in three appendices. He doesn’t explictly deal with the main counter-argument that I’ve made repeatedly here: the anthropic landscape is not science (since it is not testable), rather it is just an elaborate excuse for the failure of the speculative idea of getting the SM out of a 10/11d string/M-theory.

Schellekens has the following comments about “string phenomenology”, noting that he worked in the area around 1987 and recently, finding not much has changed:

I have been active in this are around 1987 (which led me to the conclusions presented here) and again in the last few years, and to me the similarities are more striking than the differences. There has certainly been progress: we can obtain string solutions that are more similar to the Standard Model than twenty years ago, and we have more methods to construct them. There has been major progress in moduli stabilization and supersymmetry breaking. There is more interest in “landscape statistics”. But very little seems to have changed in the way many people view the problem we are facing. Although many of my string phenomenology colleagues claim that it was clear to them a long time ago that there are many solutions, I cannot help noticing that they still talk about their most recent “model” as if it would actually have a chance to be the Standard Model. And even nowadays one still hears the occasional expression of hope for the unknown and elusive dynamical principle that will select the vacuum. The most common way of dealing with the large vacuum degeneracy is to say “I do not care about the other 10500vacua, I only care about the one that describes our universe”. That may sound reasonable, and fact it may sound like the very definition of phenomenology, but it is actually an escape from reality.

First of all, if indeed there are 10500 vacua, it is highly unlikely that anyone will find “the Standard Model” in string theory. One should expect to find a huge number that satisfy all current experimental constraints. In addition, although we now have many techniques at our disposal to construct string theories in four dimensions, it is quite clear that we are just scratching the surface. Statistically speaking, our chances of finding even one of the expected huge number of Standard Model realizations is essentially zero. Furthermore, even if we do find one, we can only make predictions about novel phenomena if we know all the other solutions and their predictions for the same phenomena. This is a crucial change in comparison to the state of the art about ten years ago: with 1020 solutions (the largest number anyone may have expected), if one is found that agrees with all current data, the probability that there is a second one is extremely small. With 10500, the same probability is astronomical. So we should forget about the idea of finding the Standard Model and then making predictions based on it.

As for LHC predictions, Schellekens argues against the idea that it will see supersymmetry:

One could say that supersymmetry is a non-solution to a non-problem: the large weak scale hierarchy is already understood anthropically, and supersymmetry by itself does not even explain it…

With the start of the LHC just months away (at least, I hope so), this is more or less the last moment to make a prediction. Will low energy supersymmetry be found or not? I am convinced that without the strange coincidence of the gauge coupling convergence, many people (including myself) would bet against it. It just seems to have been hiding itself too well, and it creates the need for new fine-tunings that are not even anthropic (and hence more serious than the one supersymmetry is supposed to solve). But even if evidence for low energy supersymmetry emerges at the LHC, in the context of a landscape it will not be the explanation for the smallness of the weak scale. The explanation will in any case be anthropic. The landscape will undoubtedly allow a distribution of values for the weak scale, including values outside the anthropic window.

Schellekens ends up making the currently fashionable argument that it doesn’t matter that string theory doesn’t predict anything testable about particle theory, that the important thing is that it is a theory of quantum gravity:

During the last two decades there was some reason to hope that we might be able to do that [get experimental confirmation of string theory] by means of some prediction of a Standard Model feature. That hope is fading now. I am not saying that this will never happen, but I have seen too much wishful thinking to make an optimistic statement about this. Essentially, we came to that conclusion already in 1986 [4]. We are dealing with a theory of gravity. Getting information about it through the back door of particle physics is a luxury that we once had good reasons to hope for, but that may not exist. Rejecting a theory of gravity that makes no particle physics prediction may be like rejecting the theory of continental drift because it does not predict the shape of Mount Everest

He then goes on to acknowledge that we’re not going to get any experimental tests out of the quantum gravity aspect of string theory either:

One cannot count on any direct experimental check of a theory of quantum gravity, since any observable consequences it might have are extremely small, unless we are extremely lucky.

In the end, he seems to argue that the only evidence for string theory we may ever get is its consistency, something which is a very long ways from being shown. He does argue that string theory is in principle falsifiable, but the example he gives (that string theory would be wrong if coupling constants varied observably on astronomical scales) is not an uncontroversial one since other string theorists have argued that varying coupling constants would be evidence for string theory. There’s also the usual “who knows?” argument used against anyone who points out that evidence against an idea is overwhelming:

On longer timescales, it is clearly ridiculous to pretend that what we currently know will be the state of the art forever. When Darwin formulated his theory of evolution he was unaware of Mendel’s results on inheritance, and could not even have imagined DNA.

The truly peculiar thing about this is to see a scientist almost gleeful at the idea that a theory they have worked on their entire professional lives doesn’t predict anything:

To me, what is emerging looks very appealing. It fulfills and even exceeds the hopes I expressed in 1998. It is has been amazing to see this theory leading us in the right direction, sometimes even against the initial expectations of most of the people working on it. We should continue to follow its lead, and do everything in our power to strengthen its theoretical underpinnings. The emergence of a huge landscape” makes this more worthwhile then ever before.

Unfortunately, Schellekens is far from alone in this. At the FQXI web-site there’s an article about the string theory/cosmology couple Andrei Linde and Renata Kallosh entitled A Perfect Match (“How do you tie down the physics of the multiverse? With string.”) In it, Kallosh explains how “string cosmology” is now the hot topic:

These days, in fact, collaboration be-tween string people and cosmology people is all the rage.
“To give you a funny example, I had an invitation to give a talk at the Strings 2008 conference at CERN,” Kallosh says. “The way the invitation was writ-ten was, ‘Of course you are welcome to speak about any topic . . . but we would be very happy if you would give us a mini-review on string cosmology!’”

Suddenly, everyone is interested in their kind of union.

“I’m also working on other very formal, very stringy topics, which were always part of my skills,” Kallosh says. “But, at this moment, people want to know about string cosmology. I’m happily working on it . . . with Andrei’s help.”

Cosmology is now ascendant, with Kallosh arguing that it will be needed to explain what is seen at the LHC:

“Soon the LHC will start giving new information on particle physics,” she says. “But we know it will be difficult to interpret this data unless you also can digest all the data from the sky—all the observations from astrophysics and cosmology.”

The article ends with a large picture of one of the LHC detectors, captioned “Cradle of Collaboration: Will the LHC provide evidence for string theory?”

FQXI is funded by the Templeton Foundation, the goal of which is to bring science and religion together. Cormac O’Raifertaigh is at another Templeton funded event, a conference in Cambridge on From the Big Bang to the Brain: Current Issues in Science and Religion. This Wednesday will be devoted to cosmology, featuring talks on the anthropic principle, fine-tuning, God and time, and God and the Big Bang.

For another take on cosmology, this October the ENS in Paris will host a conference on Evolution and Development of the Universe. For more about this parallel universe of cosmologists who also study anthropics and the multiverse , see EvoDevoUniverse.

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

25 Responses to The Emperor’s Last Clothes?

  1. Christine says:

    Sure, it is natural to accept that any mathematically consistent theory of quantum gravity can be extremely difficult to be tested experimentally. But at the basis of any expectations like this lies the specific, unambiguous predictions of the theory, whatever the difficulty in observing them. The latter does not preclude the former, which is actually indispensable for any theory.

    The obvious point here is one of posture. The landscape does not represent different mathematical descriptions of the same physical content (something that could be manageable by, e.g., extracting the observables or invariants of the transformations between the various solutions) — a common situation in many physical theories –, but a set of 10^500 independent solutions, representing different physical realities, in which our universe is supposed to be one of them. The acceptance of this state of affairs indicates a strong epistemic attitude in the sense that one embraces a substantivalism point of view towards the landscape, in the sense of assigning an independent physical reality for each of these vaccua, but is not willing to consider the possibility that this indicates a failure that needs revision.

    Such epistemic attitude is acceptable if one proves that it does not actually break any logical reasoning or consistency, or is justifiable or even needed at some point, or under some physical grounds. I can accept such a posture in the sense that it is just part of learning, exploring a difficult subject, etc. But what I cannot easily accept is the lack of a critical reasoning on facing the consequences that result from the position taken, to the degree that one is using the scientific method as a guide and not dismissing it or deforming it in order to fit the consequences of one’s personal attraction for the theory.

    I see all this as a crisis and do not understand how can one believe that the situation is an indication of progress, except if one is really inclined to redefine what science is, which is of course, the easiest route (to go nowhere).

  2. Christine says:

    Just a question: why is the paper authored by “A.N. Schellekens” and posted by “Bert Schellekens”?

  3. Peter Woit says:


    Schellekens goes by the name “Bert”, I’m guessing the “A” is for “Albert”

  4. Christine says:


    At his old homepage, he writes:

    Research interests

    String Theory, whatever that is


    But seriously, it is quite a fair way to put it, valid for many areas…

  5. Jim Clarage says:

    I don’t see why String Theory per se is responsible for this “remarkable paradigm shift in particle physics” which Schellekens refers to in his abstract. That is it’s unclear how the shift towards anthropic reasoning follows logically, or inevitably, from the principles of string theories. Rather it is a philosophical shift, one which could have happened at any time in the history of physics.

    For example Newton could have easily adopted an anthropic approach to his inverse square law, perhaps assuming any value for the exponents in the radial and mass coordinates of the gravitating objects; and then agreeing to anthropically derive these exponents based upon e.g., what values allow and disallow stability of human bones, large tree trunks, or other mechanical properties of organisms living under the influence of Earth’s known gravitational field. Nothing in this alternative history is inconceivable, especially considering Newton himself had no trouble mentioning God and Creator in his work.

    However, instead of appealing to anthros Newton appealed to physics. He tried to pick the exponents and form of the law based upon reproducing previous physical laws, viz., Kepler’s laws. Replace Kepler’s laws with Standard Model and I think it is clear what is occuring is a philosphical shift, not a shift in physics.

  6. Chris W. says:

    Right, Jim. Ironically enough, it is a philosophical shift in a field that, for the most part, eschews and dismisses overt philosophical reflection; philosophy is just a lame attempt to get at the truth by people who lack the chops to do science. From this viewpoint, having the chops has come to be what matters, and the situation celebrated by Bert Schellekens provides unlimited scope for exercising one’s chops. That seems to be the source of the peculiar glee noted by Peter.

  7. Mitch Miller says:

    Is Schkelens view on the history of the multi-verse standard? I was under the impression that Witten and Gross had technical arguments as late as the early 2000s on how it was possible that there may only be a few string theory vacumns. I may be wrong as I am only an undergrad and have not been following this as long as everybody else.

  8. manyoso says:

    Can it be mathematically or logically demonstrated that every one of the experimental results known to the Standard Model can be produced by even one of the unique 10^500 vacua?

    Schellekens states that there are likely an astronomical amount of the 10^500 vacua that would be in agreement with the Standard Model. How does he determine this? I mean does there exist a mathematical proof that at least one of these vacua would give the Standard Model?

    I mean if we can’t find the Standard Model in the haystack that is the landscape, what reasons do String Theorists have to believe that the Standard Model is even *in* the haystack other than faith in the flexibility of String Theory?

  9. Peter Woit says:


    Schellekens has a view of the history of the string theory landscape very much oriented around his claims to have realized its importance before just about anyone else. The more standard history is that the “string vacua” known from the mid-80s on did not have stabilized moduli, so most people hoped that once one figured out how to stabilize moduli, one would end up with a small number of vacua. It was only with KKLT that exponentially large numbers of (presumably) stabilized vacua were constructed. I don’t know of any claims by Gross or Witten to have an argument pointing to a small number of vacua.


    No, it can’t now be demonstrated that you can get the the SM out of known constructions in the landscape. Many string theorists like Schellekens now argue that the efforts of theorists should be devoted to studying the landscape to see if it can be proved that the SM is there.

    Personally I don’t see the point of this, and Schellekens even gives some of the arguments why. Even if you prove the SM is somewhere in there, this doesn’t allow you to predict anything or test the theory. If you prove the SM cannot come from known string vacua constructions, there’s no reason it can’t come from still unknown ones. People keep coming up with more all the time, there’s nothing like an understanding of what “all” string vacua are like. But working on this is an extremely complex problem that could keep an exponentially large number of people busy an exponentially large amount of time, some seem to want physics to go this way.

  10. manyoso says:


    If that is true that it has not been proven that the SM is somewhere in the 10^500 “known” vacua, then it seems impossibly arrogant for Schellekens and other landscape proponents to declare anything like ‘success’ with anthropic arguments.

    How can anyone argue for anthropic origins for the Standard Model via String Theory when it can’t even be demonstrated that the Standard Model is represented in the landscape. Not even theoretically! It seems that landscape proponents are arguing that because we now have hints at an absolutely *huge* number of stringy universes… our universe just must be in there somewhere. For no other reason than blind faith.

    Just awesomely odd.

  11. Peter Woit says:


    To defend the landscapers, from what is known about the string theory landscape, it is plausible that there are “string vacua” that reproduce the SM at observable energie. They would point out that showing that such a thing is really there is work in progress.

    I don’t think the problem is that they are advertising something that is not likely to be there. The problem is that even if it is there, it’s completely useless.

  12. Richard says:

    “In my opinion string theory gives the right answer, and the fact that it does adds to the evidence in its favour.”

    Huh? It’s hard to tell for sure what was meant by this comment, but doesn’t this sound a bit circular?

  13. Esornep says:

    Penrose argues that the initial condtions of *our* universe are fine-tuned to one part in 10^(10^(123)). How are you going to get fine-tuning to that extent if you only have a miserable 10^(1500) attempts?

  14. mike says:

    “Soon the LHC will start giving new information on particle physics,” she says. “But we know it will be difficult to interpret this data unless you also can digest all the data from the sky—all the observations from astrophysics and cosmology.”

    Shouldn’t this be the other way around – do astrophysicists tell particle physicists how matter works? Maybe it’s just the experimentalist in me, but I believe the machine in front of me before I believe something we interpret from afar.

  15. F Borg says:

    Shellekens should be acknowledged for his attempt at presenting his points in a clear and accessible form. It very much centers around the old issue on what to expect of a “fundamental theory” in physics. Should such a theory e.g. fix all the 28 parameters or so of the Standard Model (as an effective low energy approximation)? If so would such a “fundamental theory” just replace one mystery by another mystery (remember Eddington and 137)? Maybe in a more encompassing theory one can find a mechanism that drives the parameters toward the region of the observed values, but it is conceivable that such a mechanism will rest on new “ad hoc” principles. Another option is to think that “the fundamental theory” will only be about the main structures and equations (“the grammar of nature”, Souriau) without singling out any unique solutions. The general assumption has been that unification will lead to “uniqueness” but it is possible that instead the space of solutions is enlarged with no recipe for picking the right one that will correspond to our observations. (Shellekens denies that e.g. “beauty” and “simplicity” will be useful guides.) I think quite a few physicists believe that we have potentially an infinite ladder of effective theories that never bottom out despite the apparently absolute scale set by Planck energy, but for “practical reason” we need only work up/down to a some level of this ladder that applies to the world (or part thereof) which we are able to observe. Against this background one may understand the optimism by those who think the stringy landscape demonstrates that there is indeed a bottoming out instead of an infinity of “onion layers”. Although the “landscape” may preclude sharp predictions (and thus “verification” in a “traditional sense”) it at least may show the basic “fabrics” (or the “template”) that reality is built of, according to the adherents. This is how I interpret the “optimism”.

  16. Christine says:

    F Borg wrote:

    Although the “landscape” may preclude sharp predictions (and thus “verification” in a “traditional sense”) it at least may show the basic “fabrics” (or the “template”) that reality is built of, according to the adherents.

    I think you have summarized all very well. But suppose that it is really our fate that our fundamental theories can never be ultimately tested (through “verification in a traditional sense”), so that what we have at the end will always be some layout, some “basic fabrics that reality is built of”. How will we know that even this is true?

  17. Chris W. says:

    Christine is exactly right. There is massive presumption in supposing an allegedly empirical assertion has been “shown” to be true, when we have no way of saying, on the basis of observation, how we would ever know that it is false. This describes the essence of string theory’s lack of testability.

    That is, we can’t say this except in a sort of lame and parasitic way, eg, that string theory would be wrong if Lorentz invariance at low energies is wrong. Of course, even on this point bets can and have been hedged, even though hardly anyone is seriously expecting the issue to arise.

  18. Pingback: String theorists betting against SUSY « A Quantum Diaries Survivor

  19. Professor R says:

    Thanks for the mention Peter!

    Today lived up to expectations – Prof Paul Shelton , a colleague of Stephen Hawking at DAMTP, Cambridge, gave a spectacular overview of today’s cosmology with a thorough review of inflation, eternal inflation, the multiverse and the landscape. Shelton went through the WMAP evidence, explaining that the evidence for some sort of inflation was very strong, but the mechanism under-determined (his word, glorious understatement -he showed a hilarious slide listing all possible flavours of inflation)…

    Shellard’s discussion of eternal inflation and the multiverse was thorough but accessible, with emphasis on the viewpoint that this may be the price we have to pay for the success of inflation (in explaining the standard BB riddles), yet emphasising also the speculative nature of the multiverse idea…

    If this weren’t enough, the high point of the day for many was ‘Meta-stories of Fine-tuning ‘ by Sir John Polkinghorne. In typical fashion, Sir John gave a succinct overview of the fine-tuning problem from a philosophical viewpoint. In particular, he focussed on a choice between the multiverse explanation and the anthropic principle from a philosophical perspective. He was clearly unimpressed with the theory of the multiverse, describing it as meta-physics and probably contrary to the principle of Occam’s razor. Calling on Lewis’s famous example of the firing squad, he suggested that it was a little excessive to suggest that the firing squad engaged in a gigantic number of shootings in order to explain a miss – suggesting that it was more likely that they simply missed by design…

    He was also unimpressed with a third possibility (from me) that an unlikely outcome – however unlikely – can simply occur without the need for an explanation. I can’t do justice to John’s persuasive arguments here, but you can get the DVD of the talk here.

    Peter Woit of NOT EVEN WRONG would be pleased to note that, along the way, Sir John gave a very terse overview of the opinion of his generation of particle physicists of string theory – not very high!.

  20. Ralph says:


    “it doesn’t matter that string theory doesn’t predict anything testable about particle theory, that the important thing is that it is a theory of quantum gravity”

    I think this idea can be read in a sensible manner, but it still doesn’t help string theory.

    Quantum gravity has purely mathematical problems: e.g., giving any completely rigorous, mathematical theory incorporating both gravity and the SM would be a huge advance (You get a Clay prize for just the ‘SM’ bit of that, more-or-less?), and physical testability is not a part of that mathematical problem.

    That changes the issue from judging string theory by physicist’s standards to judging it by mathematical standards.

    And string theory is just as badly deficient from mathematical standards as it is from physicist’s standards – the parts of string theory that are even vaguely plausible as ways of getting gravity+SM aren’t anywhere near mathematically rigorous.

  21. Bee says:

    I was very puzzled by the abstract of this paper, I haven’t noticed any paradigm shift. Possibly the majority of my friends and colleagues is just more down to earth than the average string theorist but I never had the impression anybody seriously believed one would find THE fundamental theory of everything and that was that. There’s no paradigm shifting, if anything, people who believed that are recovering their sanity.

    That’s not quite what Schellekens was aiming at with THE fundamental theory, but I hope the present confusion is only a temporary stage and eventually people will realize physics is about describing nature and not equal to philosophy. There’s enough topics people could do real work on.

    I was just checking the arxiv today and I find hep-th an increasingly depressing collection of highly specialized detail investigations that in almost all cases will remain absolutely irrelevant for actual descriptions of nature.

  22. Pingback: Rumors on SUSY « The Gauge Connection

  23. Yatima says:

    For those interested in further thinking/reading about the Anthropic Principle, Scott Aaronson (He of the Complexity Theory) has published some lecture notes on the same:

    PHYS771 Lecture 17: Fun With the Anthropic Principle

    as referenced from his recent blog entry

  24. Arun says:

    Just as it took a long while and effort to get to “N=8 supergravity might be finite” and will likely take a long time (if at all) to get to “N=8 supergravity is indeed finite”, it will likely take equivalent effort and time to get to the next step in superstring theory. Maybe the era of easy surprises is over. The anthropic ideas make such effort to seem pointless however. It will be interesting to see if the field can avoid decaying entirely.

  25. Gphillip says:

    This all reminds me of the story of the madman who had created a perfectly mad, but perfectly logical reasoning behind his paranoia. His story was so creative, logically sound and finely interwoven that even his doctors found themselves looking over their shoulders from time to time. In the end, the doctors were never able to demonstrate the error of the logic to the madman, because they couldn’t find any. All they could say was his view of reality couldn’t be correct because it was totally different from that which all sane people agreed upon.

Comments are closed.