arXiv Find: Reasons Not to Believe in the String Theory Landscape

By Sean Carroll | August 29, 2012 11:32 am

Tom Banks has long been skeptical of the popular picture of the string theory landscape — the idea that there is some extremely large (10500 or more) number of phases of string theory, representing different ways to compactify the extra dimensions, and that all these phases are dynamically connected to each other, possibly by cosmological transitions during eternal inflation. Tom’s reasons aren’t of the curmudgeonly you-kids-get-off-my-lawn sort, but arise from his views about how quantum gravity works. (He thinks different cosmological boundary conditions represent truly different quantum theories, not just different regions of one big spacetime.) Well worth considering, if only because it’s too easy to run off in the direction of conventional wisdom when you’re far away from the realm of experimental testing.

The Top 10500 Reasons Not to Believe in the Landscape

T. Banks

The String Landscape is a fantasy. We actually have a plausible landscape of minimally supersymmetric $AdS_4$ solutions of supergravity modified by an exponential superpotential. None of these solutions is accessible to world sheet perturbation theory. If they exist as models of quantum gravity, they are defined by conformal field theories, and each is an independent quantum system, which makes no transitions to any of the others. This landscape has nothing to do with CDL tunneling or eternal inflation.

A proper understanding of CDL transitions in QFT on a fixed background dS space, shows that the EI picture of this system is not justified within the approximation of low energy effective field theory. The cutoff independent physics, defined by the Euclidean functional integral over the 4-sphere admits only a finite number of instantons. Plausible extensions of these ideas to a quantum theory of gravity obeying the holographic principle explain all of the actual facts about CDL transitions in dS space, and lead to a picture radically different from eternal inflation.

Theories of Eternal Inflation (EI) have to rely too heavily on the anthropic principle to be consistent with experiment. Given the vast array of effective low energy field theories that could be produced by the conventional picture of the string landscape one is forced to conclude that the most numerous anthropically allowed theories will disagree with experiment violently.

  • Riemannium

    Personally, string theory is becoming more and more a chimera, in spite of its cool mathematical features, experimental evidence is null yet about string-like objects. I have studied it in my master degree, but, as far as I know, I can not think a good reason to believe…

  • Robert McNees

    As I have noted elsewhere, I am surprised he stopped at $10^{500}$ reasons. I love these sort of papers by Tom, as I typically learn two or three very interesting things per page.

  • Daneel Olivaw

    “…minimally supersymmetric $AdS_4$ solutions of supergravity modified by an exponential superpotential.”

  • Brian

    … Dressed up like a million-phase-space trooper.
    Wishing he was more like Sheldon Cooper.

  • Brian Too

    @4. Brian.


  • Cliff

    I thought his basic point was extremely interesting and compelling, and I think it seems valid. At very least, the idea that the landscape is some established feature of string theory should be taken with a grain of salt. It certainly shouldn’t be the presented as the only thing worth knowing about string theory, as it sometimes is.

    Overall, my personal take continues to be that there is just no way such a mathematical structure should ever exist and satisfy all these crazy properties (duality, holography…) that it needs for quantum gravity to make sense unless it is used by nature for exactly that purpose. The AdS/CFT cements string theory’s direct link to the most well-established physical frameworks, and the fact that other seemingly independent approaches (bootstraps, supergravity…) have terminated at the same answer provides satisfying additional credence. And of course there is the noticeable absence of any alternative, and the fact that the existence of another possibility would require still more “miracles” that have no reason to occur.

    I think people should at very least stop saying phrases like “we don’t know how to apply quantum mechanics to gravity”. Its taken a lot of work, but humanity has figured out how to do just that. It would be nice for more direct evidence, but there is still lots of work to do sorting through the “theoretical data” and answering well-defined mathematical questions. I think experiments will also tell us more important clues someday, but for now I find this avenue of investigation to be extremely exciting.

  • Phil

    How do you know God doesn’t exist? Maybe God does exist. Does science answer the question? If so, how? Just curious. I’m an agnostic. For me, science deals with explanations of things which can be testable. Since we only have access to the physical world of the Standard Model, etc., because we are made up of Standard Model particles, the only testable ideas and explanations are those that have the Standard Model as their basis (i.e. “natural causes”, “materialism”, etc.). But we don’t know if this aspect of reality is the only one. Hence my agnostic views.

  • Archie Pelago

    Tom needs to realise that once the words/equations ratio reaches a certain critical value, any paper on the arxiv collapses into a black hole from which no information can be extracted except on cosmological timescales.

  • Mike


    What you suggest: “we don’t know if this aspect of reality is the only one”, is not a useful way to look at the question. This approach can be used to consider and even to justify any and all conjectures, regardless of their absurdity.

    Stars made of green cheese? Well, we don’t know for sure since our belief that there aren’t any stars made of green cheese beyond what we can see and test is based on our narrow and limited scientific, physical, Standard Model, natural, material analysis and explanations, and who can ever say for sure that this “aspect of reality is the only one”?

    Magical gnomes, rocks that can speak Latin, Gods affecting the seasons? Ditto. Unfortunately, this way of looking at things isn’t helpful. I’m afraid that the scientific method and the explanations that it yields is the only useful way to approach understanding reality.

    This doesn’t mean that our current scientific explanations are complete or wholly correct. They’re not, and undoubtedly never will be. Some, in fact, may change radically. But we will not make progress by positing vague, easily variable assertions regarding other realities.

    This approach gets us no closer to better explanations of reality. It’s what humans mostly did for the better part of our time on this planet, and it was only by recently abandoning that approach that we have begun to make significant and rapidly accelerating progress.

  • Torbjörn Larsson, OM

    Thanks, an interesting read seeing Susskind et al many new papers. I suspected that Susskind’s notion that EI doesn’t care about an initial low entropy globally would get a lot of goats, and maybe this is a response to that.

    Though above my level naturally. As McNees notes, it is cram packed with goodies in any case.

    But that is also a problem. I can understand making reviews and well focused criticism. This type of long storied hypotheses building or criticism seems so much air castles to me.

    For example, I don’t see how Banks can make landscape predictions an all or nothing game. “A more serious issue [24] is the extent to which anthropic reasoning can predict more than two or three things about the model that describes our universe. If it can’t, then a landscape model in which more than two or three low energy parameters vary over the landscape, is in violent disagreement with experiment.”

    However, when he points out that these permissible and, I guess, EI inclusive coincidences amounts to finetuning he has a good point. But only if his vaunted SS model is pinning all parameters, which I doubt. Otherwise it is enough to look for predictivity, and what other theory predicts the cc?

    Air castles, too many options and too little constraint in the stories.

    A technical note:

    On p21 Banks claims in effect that habitability in our universe doesn’t need more than one flavor of neutrino. He may already have been superseded there, since earlier models of supernova explosions may have been wrong in ignoring parts of the neutrino physics:

    “It is calculated that this material will interact with the outflowing neutrinos, scattering a small percentage of them. These scattered neutrinos create a “neutrino halo” that can, in turn, interact with the outflowing neutrinos.

    But the fraction of outflowing neutrinos affected was thought to be minuscule and was often ignered [sic] during simulations. Cherry’s calculations disagree, however — his model suggests that a correction factor of 14 percent needs to be applied. In the outermost portions of the exploding supernova, this correction factor rose ten-fold. This means the neutrinos streaming from the core interacted with halo neutrinos far more often than previous theories anticipated.”

    “”Even though few neutrinos are scattered in funny directions, they can completely dominate how the neutrinos change their favors [sic!],” said George Fuller, professor of physics at UCSD, co-author of the study.”

    “Those neutrino flavor states allow the neutrinos to change protons to neutrons or neutrons to protons.” Cherry said. “What matter is produced, what kinds of atoms, elements are produced by these supernovas are changed dramatically if you change the flavor content of neutrinos.”

    What started out as an insignificant “correction factor” in a simulation could suddenly have profound effects on the elements generated by a supernova and the characteristics of the material inside a galaxy.”

    FWIW, the work is published in PRL.

  • Torbjörn Larsson, OM

    @ Phil:

    I can’t speak for Sean, but I can speak as one that is completely rejecting gods until evidence to the contrary can sway that rejection.

    The reason is very simple.

    We know that universes comes naturally out of physical laws, see Krauss et al. That makes gods neither necessary nor likely creators of universes.

    Similarly we know that physical laws comes naturally out of selection on universes or landscapes (unless Banks is correct, of course). That makes gods neither necessary nor likely creators of physical laws. And no, that isn’t circular (see Susskind especially), and even if it was in theory then observation would break that.

    And of course on the other end we know that gods are human myths, again speaking against naturalness, against necessity and for an utterly low likelihood.

    For myself, I can’t understand how one can take an agnostic position. Most of the time it seems based on making the theological claim that one “can’t know”. But one can, well beyond a 50 % likelihood situation.

    The claim that one “can’t know” is based on the idea of religious special privilege, that one can’t adjudicate the empirical claims stemming from positing gods or similar kinds of magic dualism as one can all other empirical claims. In those cases agnosticism is effectively a creationist religion, and I don’t think creationists should comment on science as they comment on religion.

  • Phil

    @ Torbjörn Larsson, OM

    “We know that universes comes naturally out of physical laws, see Krauss et al. ”

    Has this been shown to be a consequence of an experimentally validated, consistent fundamental theory of physics?

    And in what sense do the laws of physics “exist” or “result” in anything if there is no universe in the first place? I always thought the laws of physics only exist when there is a universe, field content, etc., on which to act.

    If Krauss has shown this, how come he’s not an overnight sensation the way Einstein was?

    “Similarly we know that physical laws comes naturally out of selection on universes or landscapes”

    So, you’re substituting unobservable landscapes for unobservable gods? Has the landscape been shown to be the direct consequence of an experimentally validated fundamental theory of physics? Even if, in the distant future, it is, how do we know that another fundamental theory of physics, satisfying the same experiments, which doesn’t contain a landscape, doesn’t exist?

    So it would seem as though it’s impossible to reliably observe, directly or indirectly, the existence of the landscape — just like gods.

  • Mike


    In science and in reality there are lots of things that we can’t “observe”, but we know they are basically correct because they conform to our best theories and scientific explanations. I won’t bother to begin the list — I’m sure you don’t disagree with this.

    However, people can and do disagree on the specifics. For example, some accept one form of multiverse because they say it flows naturally from a proper understanding of quantum physics; others are much less certain about ideas like the landscape multiverse because they say there remain fundamental questions about the underlying theories and explanations (like inflation) that give rise to it. However, each of these are fundamentally different than “unobservable gods”.

    In the former case, the conjectures at least flow from theories and explanations that we know are approximately true — and because they do, if we are clever enough, we stand a good chance of figuring out a way to test them someday. Even if we can’t test them satisfactorily, they will rise or fall as newer and better explanations are developed.

    In the later case, none of that is true. It’s an explanation for everything and nothing.

  • Phil

    Thanks, Mike.

  • Super_earth

    Torbjörn Larsson, OM:

    “Similarly we know that physical laws comes naturally out of selection on universes or landscapes (unless Banks is correct, of course). ”

    You mean the selection of the values of >10^500 parameters of the string theory landscape?

    What about the fundamental physical laws, i.e the (still unknown) theory of quantum gravity (hopefully superstring/brane theory is a step in the right direction), that rule that “selection of universes or landscapes” that lead to the “physical laws” that rule each universe?

    “That makes gods neither necessary nor likely creators of physical laws.”

    God(s) could still be behind the fundamental laws of nature that then rule and make new universes.

    “In those cases agnosticism is effectively a creationist religion, and I don’t think creationists should comment on science as they comment on religion”

    The position on God(s) could be either:

    • Theism: belief in an entity(God) that controls and knows everything, not just physical laws

    • Pantheism: belief that everything is God, could be stated as “the Universe/Multiverse is God”

    • Deism: belief in a God(s) that controls just physical laws, everything else works is outside God’s control.

    • Panentheism: belief where God(s) interpenetrates every part of nature and timelessly extends beyond it. Unlike Pantheism, the Universe and God(s) are different entities.

    • Atheism is the belief that all those possibilities are false.

    • Agnosticism is the statement that we don’t know which one of the previous possibilities is true

    Creationism is a subset of religious beliefs that states that the Universe/Multiverse did not exist from the infinite past (as in Eternal Inflation or Cyclic Ekpyrotic model) but began some time ago when it was created by God(s). Deism, Theism and Pantheism do not imply necessarily creationism, because God(s) could have been the “first cause” of the world for all the time since the infinite past.

    Unfortunately this term is usually associated with some groups of religious fundamentalists that deny nearly all the science in the last centuries, wanting to make people believe in literalist interpretations of religious texts (like the Book of Genesis and the Quran)like a 6000 year old Earth, a special creation of every biological species on Earth, etc. Biblical/Quranical literalists are extremists that should not be taken as representatives of all people that believe in a Creator(s).

    I see no problem with religious people commenting about science , indeed some of the greatest geniuses of history were religious (like Einstein, Lemaitre and Darwin, just to name a few).

    I think that with all the unknowns in the physical universe, the case for the existence of God(s) is far from being over and we could still make philosophical arguments about it, including all the fascinating discoveries and theories about the Universe and physical reality.

    See you later,
    Super Earth.

  • chris

    a lot of crap gets published in PRL.

  • Doug

    Hi Sean,

    Tom Banks may be correct for some aspects of string theory such as super symmetry.

    However, some aspects of string theory appear to closely resemble the theory of quantum vortices. There are well over 1,000 ArXiv papers on this subject.

    Moreover, thermophysicsts have published a book “Theory of Concentrated Vortices” which appears to work at a defferent gauge. This theory is based on mathematical dynamics which has some features in common with dynamic game theory such as pursuit-evasion.

    It may be possible that gyres such as active galactic nuclei may be another vortex manifestation at a different gauge.

  • Meh

    In physics, and science in general, if you can’t prove something through experimental evidence, then the only way to get around that is to give some sort of theoretical/mathematical proof. Since there are no complete theoretical or mathematical proofs and there is no experimental evidence; I have to think that String Theory is a load of crap until one of the criteria above can be met.

    I see Sean Carroll working on things like providing a solid framework for our understanding of the dimension of time. I personally don’t believe in singularities in nature. I think if you have a misunderstanding of root concepts like the 4th dimension, what time is and if it exists, and what it means when you stumble upon an infinite in an equation; then you are undoubtedly going to wind up with 10^500 misinterpretations.

  • http://deleted Gary

    Tom Banks has an opinion.

    Has he also a suggestion of a “violent” notion of a disproof, too?

    Isn’t that also part of the conceptual vetting process?

    Why must we ask?

  • martenvandijk

    @ Meh

    No time, no entropy.

  • Meh


    no time, no physical existence. Entropy did not exist at the beginning of our universe, it will not exist at the end of it if the popular theory is correct.

  • martenvandijk

    @ Meh

    As I understand it the popular theory implies that time was infinite at the beginning and entropy will be infinite at the end. I think that Heisenberg’s principle makes it impossible for time to be either 0 or infinite. If time cannot be infinite, singularities cannot exist.

  • Bob

    Cliff, what are bootstraps?

  • Chris Duston

    I can’t make specific comments on this, but I find one statement curious:

    “The cutoff independent physics, defined by the Euclidean functional integral over the 4-sphere admits only a finite number of instantons.”

    Since we don’t know how many diffeomorphism classes of 4-spheres there are (and the best chances are 1 and infinity I guess), shouldn’t we count each instaton on each sphere? Can someone elaborate for me a little?

  • Cliff

    Bob (#23) asks, What are bootstraps. This question has been answered in a really great way by Ron Maimon on Stack Exchange, so I’ll just link that here:

    Sorry that this is so late but I was just away for a couple weeks.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

About Sean Carroll

Sean Carroll is a Senior Research Associate in the Department of Physics at the California Institute of Technology. His research interests include theoretical aspects of cosmology, field theory, and gravitation. His most recent book is The Particle at the End of the Universe, about the Large Hadron Collider and the search for the Higgs boson. Here are some of his favorite blog posts, home page, and email: carroll [at] .


See More

Collapse bottom bar