The String Wars

I noticed recently that George Johnson will be journalist in residence and giving a talk on Friday at the KITP in Santa Barbara about “The String Wars”. Somehow I don’t really think that it’s a good thing that this is now being perceived as a “war”. Johnson is the author of an excellent biography of Murray Gell-Mann and writes for the New York Times.

For controversy on the East coast, tonight the Center for Science Writings at Stevens Institute for Technology in Hoboken will be hosting a panel discussion and debate on The End of Science?, featuring John Horgan and Michio Kaku.

This week’s New Yorker has a couple letters to the editor responding to their recent article about the string theory controversy. One points out that particle theory and quantum gravity is not all there is to theoretical physics. The second is by Lisa Randall, and mainly concerned with claiming that there is now a healthy interaction going between string theory and phenomenology, with most particle physicists eagerly awaiting the LHC.

Update: Today’s New York Times has an Op-Ed piece entitled The Universe on a String by my Columbia colleague Brian Greene, in which he responds to recent criticism of string theory. As you might guess, Brian’s piece doesn’t really convince me to change my mind (as my book and Lee Smolin’s don’t seem to have convinced him).

Brian mentions the possibility of seeing supersymmetry or extra dimensions at the LHC, and possible effects of quantum gravity in the CMB, but acknowledges that these are not definitive predictions of string theory that can be used to falsify it. He also mentions the recent attempts to apply AdS/CFT to heavy ion physics, but these don’t address the use of string theory as an idea about unification.

He deals with the landscape only by making an argument I’ve heard him make before: that just having a unified theory of gravity and particle physics would be a big accomplishment, even it this theory didn’t explain any of the things about the standard model that one would like it to explain. Besides the fact that string theory still doesn’t provide a fully consistent unified theory (since it has no non-perturbative formulation), I’ve always found this point of view problematic. If string theory can’t make any definitive predictions about particle physics, it’s very unclear that one can ever test it, which is a huge problem.

Brian does, unlike some string theorists, acknowledge that it’s possible that string theory is wrong and will have to be abandoned, in particular if “future studies reveal an insuperable barrier to making contact with experimental data”. My argument is that if string theorists accept the existence of the Landscape, such an insuperable barrier appears. He describes string theory critics as calling for research on string theory to be dropped, which really isn’t accurate. Neither Smolin nor I have ever called for this, rather our argument is that research into alternatives to string theory needs to be encouraged.

Update: The George Johnson talk is now available here. It seems that many of the string theorists at the KITP are not very happy about my book and Smolin’s, although it’s unclear if any of them have read either of the books. Amanda Peet claimed that both books have many errors (invoking the NYT review by Tom Siegfried), while Johnson repeatedly told her that it would be a good idea for her to actually read one of the books. She also kept claiming that there is “a backstory” that explains why Smolin wrote his book, but she was dissuaded from elaborating on this when someone pointed out that the talk was on video and would be on the web.

The experience of watching the talk was pretty odd, since Johnson began by connecting to my blog and discussing the fact that I was discussing his upcoming talk. I watched a lot of the talk during commercials of an episode of Numb3rs, and during this episode “Larry” the physicist was working on calculations involving branes, and playing hooky from a string theory conference.

Update: Davide Castelvecchi has put up an interview with George Johnson on his web-site.

Update: Clifford Johnson and Lubos Motl have their own takes on the KITP video.

Update: It appears that there will be a second talk by George Johnson about this, String Wars 2. After the first one, I’m having trouble figuring out why anyone at KITP thought a second one would be a good idea.

This entry was posted in Favorite Old Posts, Uncategorized. Bookmark the permalink.

97 Responses to The String Wars

  1. The End of Science? When exactly did that happen?

    Stephen Wolfram thinks it began when Newton invented the Calculus (soory, Fluxions) to do Physics, if the cosmos is actually a cellular automaton.

    The late 19th Century thought it had come, with only more decimal points to be added to measurements.

    Deterministic materialists think it came in with Quantum Mechnics.

    Aether theorists think it came with Michelson-Morley, or maybe Einstein and Poincare.

    Then there’s that pesky String Theory, which certainly blocked the rest of Physics from getting equal access to funding and tenure-track employment.

    As J. D. Bernal wrote in “The World, the Flesh, and the Devil”): “we are still too close to the birth of the universe to be certain about its death.”

  2. Timothy Clemans says:

    “Will science ever solve the riddle of the universe once and for all?”

    “This should be able to prove or disprove string theory. Personally, I feel no need to prove the theory experimentally, since I believe it can be proven using pure mathematics.” –

    Well according to Kaku he thinks that question will be answered by deducing string theory from some physical axioms so to him it may not be science that answers the question but mathematics.

  3. TheGraduate says:

    “For controversy on the East coast, tonight the Center for Science Writings at Stevens Institute for Technology in Hoboken will be hosting a panel discussion and debate on The End of Science?”

    What, no live webcasts?!

  4. King Ray says:

    I think fitting the universe with string theory is like fitting a sine wave with square waves, only it’s much worse.

  5. Aaron Bergman says:

    Then there’s that pesky String Theory, which certainly blocked the rest of Physics from getting equal access to funding and tenure-track employment.

    The “rest of Physics”? Snort.

    For controversy on the East coast, tonight the Center for Science Writings at Stevens Institute for Technology in Hoboken will be hosting a panel discussion and debate on The End of Science?, featuring John Horgan and Michio Kaku.


  6. Peter Woit says:

    OK, folks, I realize this “string controversy” coverage is getting old, and I promise to put up something on a different topic soon, but can we aim for a somewhat higher level of discourse than the above?

  7. Thomas Love says:

    Peter, The string controversy is not getting old. You fired the opening salvo in the string wars. You can’t cut (the strings) and run! I’m hoping to hear about the final demise of string theory and I expect to read it here!

  8. Gina says:


    This is a little off-topic, but I hope Peter will allow it, just this time. Over the last eight-ten weeks I made occasionally comments on several issues discussed here and in Peter Woit’s book. I did put a lot of thought into my comments, although in the blog speed I could not be always on par. My comments were mainly on issues regarding philosophy of science and the practices and ethics of debating science. Without the technical background but with some comon sense and at times help from more informed friends (e.g. my comment on 2- and 4 dimensional lattice models,) I tried sometimes to relate in a non-technical way to specific scientific matters. Following my attempt to summarize the recent landscape discussion and Peter’s rather negative comment 10 minutes later, Peter informed me that as this weblog “is largely intended for people who work professionally on the issues in math and physics that I’m interested in” and since my contributions drew a lot of criticism by other participants, a criticism that he shares, I will not be able to post here without prior monitoring. This practically means that I will not be able to comment here anymore.

    It was an interesting experiece, even if unsuccessful. best wishes everybody.

  9. woit says:

    About Gina’s comment:

    As the volume of comments has increased here, it has become more and more difficult to keep the noise level down. There are far too many people who think it is a good idea to repeatedly post comments that are off-topic, unsubstantive, uninformed, and too often a waste of everyone’s time. Dealing with this is taking up more time than I can afford, especially difficult is the phenomenon of people who sometimes post something sensible and interesting, but all too often something that isn’t. I then have to spend my time trying to decide what to do with each of their comments, and dealing with complaints from them about the ones that I deleted. I don’t have the time for this.

    Please help me by only posting comments that are on-topic, substantive and well-informed. “String theory sux” or “string theory rules” comments are not welcome. If it’s a technical issue that is the topic, I strongly encourage people with expertise to contribute and enlighten us, and equally strongly discourage people who are not informed about the issue from adding to the noise level by telling us all what they think.


    I’d like to make clear that, as far as string theory goes, I have a simple goal here, and it’s not to vanquish string theorists in battle. It’s to provide accurate information about string theory, both about those parts of the subject that are not working, and about those that are more promising. My belief is that if the physics community were operating with a much more realistic and accurate view of what parts of the string theory program are getting somewhere, and which have failed, particle theory would be a much healthier subject. One aspect of this would be greater willingness to encourage research on non-string theory formal quantum field theory research.

  10. surlygrad says:

    Re: “The Rest of Physics”

    God forbid some of us should be interested in topics (many particle systems, heterogeneous materials, etc.) that you physicists have deemed unworthy of your time and tossed to the engineers.

  11. Walt says:

    String theory sux.

  12. TheGraduate says:

    I think a lot of what is missing in the comment section of this blog in general is an honest structured discourse. For one thing people tend not to respond to what other people have said. But even so this is a minor trait common to most blogs.

    A more major problem, in my opinion, is the tendency of the discourse to careen off into minutae. It often has the form of a general discussion beginning on whether fast food is good for you and ending in a furious debate about the third decimal place in the solubility constant of FD&C Red No.40.

    Isn’t it fair to say there is a lot of ambiguity about the status of different propositions in any string theory discussion? Either a proposition is logically proven, a conjecture or substantiated by experiment. Either a proposition is marginal or it’s critical to the argument. If one can already be satisfied on whether a particular proposition of string theory is marginal or critical and it’s proof status then to me this is progress compared to how discussions usually progress on the blog.

    Maybe there is a lack of good argument standard. I think in a physics argument there can only be two objective standards, substantiated experiment and mathematial proofs. Everything else is just what you feel inside (which you are allowed to feel but it’s hard to say why what you feel is more important than what anybody else feels. And I think just because your feeling has an equation attached to it that doesn’t make it suddenly more valid.)

    (I often read the blog. I don’t always comment on everything I see but I think I might have something sensible to say on this issue.)

  13. anon says:

    Peter, I think the problem is that you have outgrown your blog tools. In my opinion, any blog that gets more than 5 comments per article is already too burdensome. What you need is a blog in which (1)each article you post can grow threads (2) commenters grade each other. Free Software for this “uberblog” already exists; viz., Slashdot.

  14. Garbarge says:

    This blog could be an interesting place, and not just Lubos’ trash can as once became, if one would be able to read the first and last comment and make any sense out of them. I dont read here that often, but I am amazed by the rapidity with which it degenerates into totally off topic discussions. In the other hand, and on the edge of toss my words into the garbage disposal too, I think the String Theory war has become a big parody/comedy with little or no contain. As long as the theater prevails over the science, I see no glean of fruitful thinking emerging from here. It is a pitty, this media has become such a powerful mean….

  15. A.J. says:

    One solution to the noise problem: Post on more technical topics! It’s relatively easy to form and offer an opinion on controversies and pop sci reporting. But it takes a bit more effort to come up with something to say about e.g. stacks or elliptic cohomology.

    Well…I guess there’s always “stacks sux” or “Hopkins is teh rul3z0r”.

  16. Thomas Love says:

    In “The Rise of the Standard Model’’, Mark Bodnarczuk has an interesting article, “Sociological Consequences of the Standard Model”, with this footnote:

    Using numerous case studies, David Hull claims that not only are infighting, mutual exploitation, and even personal vendettas typical behavior for many scientists, but that this sort of behavior actually facilitates scientific development. David Hull “Science as a process”

    So the rise of string theory is really not that different.

  17. ksh95 says:


    Find people you trust and give them admin access. This is similar to running a business, when you grow you have to add managers.


    I have a Ph.D in physics, yet I would never be arrogant enough to go to n-category cafe, start posting uninformed nonsense, jump in the middle of discussions I can’t fully understand, and then sulk when my posts get deleted. Instead, I would read every day, try to learn as much as I could, and feel lucky that I was privy to such a high level discussion.
    Any post I would make would be to thank the blog owners for making such interesting discourse public. OR, to ask for simple laymen explainations.

  18. Peter Woit says:

    Thanks for the suggestions. For a while I was interested in the idea of having a Slashdot-like self-moderated comment section. Then I read the comments associated with some of the items there about string theory. As far as I could tell, moderation was not working at all, with huge amounts of idiocy prominently displayed and substantive comments buried way down. I’m also skeptical of threaded comments, partly because they make it much harder to see what is new since one last looked, partly because they encourage people to get on off-topic threads, instead of encouraging a discussion of a single topic.

  19. John A says:


    Can I say with the benefit of experience, that you cannot choose what level your commentators write at. As Aaron Bergman has amply demonstrated, sometimes people just simply don’t want to express themselves in a rational way, but simply want to disrupt, flame or just be an asshole, hoping to drag down the discussion and turn away interested posters and shut down proper debate.

    That said, my suggestions to you would be to shorten your posts to one or two key points, and invite your visitors to make discoveries/comments/reviews for themselves.

    In my view you should take the initiative, talking about short topics in your book, or within other books and looking at alternatives to string theory in the same critical light.

    You are not using categories on WordPress which is a mistake. This means that when a topic or topics covers a particular subject, and those topics disappear from the front page, then it disappears from your active discussions as well. Try to make good references both to your previous posts and to interesting comments in your newest postings.

    In my view, its not the noise that’s increasing, its the signal getting a little weaker.

  20. Hal says:

    To add something missing from ksh95’s comment I would like to wish Gina well since she did, after all, give her best wishes to everyone.

  21. Aaron Bergman says:

    As Aaron Bergman has amply demonstrated, sometimes people just simply don’t want to express themselves in a rational way, but simply want to disrupt, flame or just be an asshole, hoping to drag down the discussion and turn away interested posters and shut down proper debate.

    So that’s why I’m here.

    I’ve been trying to figure it out myself….

  22. A.J. says:

    John A,

    What are you talking about? Aaron’s consistently been one of the most reasonable and well-informed commenters here.

  23. Peter Woit says:


    Please stop with the attempt to start a flame-war. I can understand why Aaron had the reaction he did to the things he quoted, just wish he’d ignore some things, make more substantive comments on others. There’s a lot of less than completely intelligent stuff posted here in the comment section, even after a fair number of deletions. People are encouraged to ignore such stuff and not feel that it is necessary to respond.

  24. Kent G. Budge says:

    “is largely intended for people who work professionally on the issues in math and physics that I’m interested in”

    When I was a new astronomy graduate student at Caltech, I sometimes wondered if I ought to switch over to theoretical physics, since I believed I had much more aptitude for mathematical analysis than for observational and experimental technique.

    However, during my first year, I had the poor judgement to ask a dumb question at a physics department seminar. I wasn’t trying to start a flame war and I wasn’t being deliberately ignorant; I was a young graduate student struggling to understand some of the basics of quantum field theory. However, the speaker thought it appropriate to take five minutes to respond to my question by berating my stupidity and wondering out loud how I ever got to Caltech.

    At the time, I was humiliated and deeply hurt. However, in retrospect, I think the speaker may have done me a favor, by disabusing me of the notion that the physics department at Caltech saw itself as a teaching institution, and by steering me away from what would probably have been a poor career choice, given my dislike for feeling stupid and for getting in arguments.

    Perhaps you are doing me a similar favor here. S’long.

  25. woit says:


    I’m not berating anyone for stupidity or questioning their credentials. If someone who doesn’t know what is going on asks an uninformed question in a seminar, it’s perfectly appropriate for the speaker to say that he or she doesn’t want to take everyone’s time by answering it, especially if the answer is not something that can be given in a sentence or two. It’s not appropriate to do this in an insulting way.

    Similarly, here I also need to deal with uninformed comments, sometimes by telling the commenter that this is not an appropriate place for the kind of comments they are making. I hope that I do so in a polite way.

  26. ak says:

    anyway, I point out that a seminar circles by definition around a well-formulated subject or theme, it is by definition designed to fit persons with similar ‘background’, knowledge base and interest. At least as I understood the internet blogosphere is of a fundamental different character and this does not exclude excpetions as ‘the n-category cafe’ as it was mentioned. Undenied the fact that any blog owner might design his or her blog in a given way or directed to a chosen public and might do so by explicit forms of moderating, at least from my point of view this was not the initial aim of the phenomenon ‘blog’. That certain characteristics of this initially ‘liberal’ form of electronic discussion also carry over to scientific forms cannot be regarded as a pure ‘noise phenomenon’. The more informal and liberal form makes it possible to share information flow between possibly poorly overlapping subjects and standpoints and people with very different knowledge backgrounds. As far as physics and mathematics as classically highly impenetrable subjects for non-experts share responsibility to communicate with the ‘outer world’, the scientific blogosphere possibly could realize this with relatively low levels of formal and financial effort. To be more concrete: as it ssemed to me Gina, fo instance, did have a scientific background, which while it aparently did not quite overlap with particle physics (chemistry?) does not quite exclude her form any form of rational reasoning.
    The scientific blogosphere could well be aware of even much more radically differing forms of scientific reasoning having potential and justified interest in informal discussions with the mathematical or physical world, I mention philosophers, sociologists as already discussed, not to mention cultural scientitists or even artists, baring viewpoints which could ‘possibly’ be of interest or benefit for physicists or mathematicians themselves and who up to now share mostly very deformed and wrong views of scientific knowledge in the mathematical or physical sphere, I think the most ‘distant’ views shared in this blog came from engineers, so I do in fact think the situation could be much ‘worse’ (or better, what you prefer).

  27. Peter Woit says:


    A “blog” is just a software tool, and people are using it with many different purposes. Many scientists have blogs designed to try and explain and promote science to the widest possible public. Many people would like this blog to be a forum for the discussion of their and other unconventional ideas about physics. Those are fine things to do, I’m just not interested in doing them. One reason is purely a lack of time: maintaining this blog is already taking up too much of my time and I’m trying to figure out how to change that.

    Sorry to be a bit obnoxious about this, but one of the whole points of the blog technology is to allow people to try and create an information source and discussion forum of whatever kind they want. I’m trying to make one for people who share some of the same interests as me, with the highest possible level of substantive material. Doing this requires deleting comments from people who want to use this to discuss something I’m not interested in, as well as comments that are uninformed. One obvious thing to point out to people who don’t like this is that the technology is free, you’re welcome to create your own blog, and there you can do exactly what you want.

  28. Jud says:

    I thought the suggestion about allowing some trusted folks to perform moderation was not an awful idea, though of course I don’t know what your supply of such folks may be.

  29. Tony Smith says:

    Brian Greene’s 20 October 2006 New York Times Op-Ed says:
    “… at high energies, the electromagnetic and weak nuclear forces seamlessly combine … at even higher energies the strong nuclear force would also meld …
    For decades, however, the force of gravity stubbornly resisted joining the fold. …
    Time and again, attempts to merge the two theories resulted in ill-defined mathematics

    Such was the case until the mid-1980’s, when a new approach, string theory, burst onto the stage. Difficult and complex calculations … gave compelling evidence that this new approach … unified gravity and quantum mechanics …”.

    There is a very significant omission in Brian Greene’s recital of historical “context”:
    In fact,
    the first approach that successfully “unified gravity and quantum mechanics” was supergravity, in which gravity and other forces were unified by the use of Lie superalgebras and description of gravity by mechanisms of the MacDowell-Mansouri type, using either the anti-deSitter group Spin(2,3) = Sp(2) or the conformal group Spin(2,4) = SU(2,2).

    Interest in supergravity waned in the mid-1980s when it seemed difficult to get it to contain the Standard Model groups and it also seemed that it might not have the cancellations needed for finiteness (although the difficult calculations have yet to be done).

    Then, as Brian Greene said, in the mid-1980s conventional superstring theory became dominant in the high-energy theoretical physics game, and it maintains its dominance even now.

    As Brian Greene said, the basic idea of conventional supertstring theory is to change the view of “matter’s fundamental constituents” from “point-like dots of virtually no size” to “minuscule, vibrating, string-like filaments”.

    The competing basic idea of supergravity, ignored in Brian Greene’s discussion of history, is to change the view of forces from being based on the generators of Lie algebras to being based on the generators of Lie superalgebras.

    While the conventional superstring theory approach focuses on a new view of matter as filaments rather than points,
    the supergravity approach focuses on a new view of the setting for force generators.

    Even if changing the setting from Lie algebras to Lie superalgebras might not have been totally successful,
    there are some other natural settings for the force generators (one of which is to look at them as root vectors, which is what I like to do)
    I feel that such generalized/modified-supergravity approaches should be investigated as alternatives to conventional superstring theory,
    since supergravity (despite Brian Greene’s omission of it from his recital of historical “context”) was actually the first theory that successfully “unified gravity and quantum mechanics”.

    Tony Smith

  30. Gumbi says:

    In some ways the dispute reminds me of the old debate about the Many-Worlds Interpretation of QM. The MWI is just like textbook QM except there is no wave function collapse. Instead, proponents argue that in a universe with no collapse, macroscopic observers would see the illusion of collapse, which is consistent with what we observe.

    This may be good philosophy but the problem from the scientific perspective is that the MWI is too perfect. It reproduces the predictions of regular QM so well that no experiment can distinguish them.

    We are left with philosophical arguments – MWI proponents advocating its somewhat more parsimonious axiomatic basis, while opponents object to its version of the “landscape”. All those parallel worlds, all those alternate histories and variations on ourselves, all as real as we are. It’s a lot to accept.

    Although the analogy is not exact, we might end in a similar situation with regard to string theory. If the string theory landscape is not falsifiable, then we can’t prove the string theory landscape is real; but we also can’t prove that the idea is wrong. In the end it may be a matter of philosophical taste as to which theoretical approach produces the most attractive and elegant model of the universe.

  31. Peter Shor says:

    I’m going to stick my neck out and say you’re being too harsh on Brian Greene in at least one respect. I think that having a unified theory of gravity and quantum mechanics would be a big step forward (not that string theory is actually one yet).

    Why do I say so? Because I suspect that all consistent unified theories of gravity and quantum mechanics are going to share a lot of properties, and if we could figure out what some of these properties are, then it might help us to figure out the right theory of quantum gravity. I believe that it is quite possible that any consistent theory of quantum gravity is going to lack one of the things we think are absolutely essential, for example Lorentz invariance, unitarity, or locality. (All three of these seem to be crucial in formulating the black hole information loss paradox.) One possible reason we haven’t come up with any reasonable theories of quantum gravity might be that we’re looking in the wrong place. So if we could formulate a non-perturbative theory of string theory (even in 10 or 11 non-compactified dimensions), it might teach us something something about all theories of quantum gravity.

    I’m sure that one of the string theorists reading this will object and say that string theory has all these properties, and is also a consistent theory of quantum gravity. But my understanding is that we only “know” that perturbative string theory is unitary, and that this theory lives in the limit where the curvature of space-time doesn’t change. The black hole information loss paradox is perfectly consistent with unitarity in this limit.

    In fact, this is where I first lost some of my trust in string theorists. Shortly after quantum fault tolerance was discovered, I was talking to a famous string theorist, and trying to explain to him Kitaev’s result about error correction with anyons, which I was very excited about at the time. This result led Preskill (I believe) to speculate that the laws of nature might be non-unitary at bottom, but that this non-unitarity is masked by an inherent error-correction process, which makes everything look unitary at any scale significantly larger than the Planck length. He was very dismissive, and said that it had been proved that any non-unitarity at small scales would also show up at large scales. Since this result was exactly contrary to the one I was trying to explain to him, I asked where this proof was, and how it worked. He couldn’t tell me exactly who had proved it, or whether it was written down, and he muttered something with the words “at least in the generic case …” in it. At this point, I gave up.

    I still think that Kitaev’s result shows that it’s possible that the true theory of quantum gravity is not unitary. Unfortunately, I have come to realize that it gives very little guidance as to where to look for a non-unitary but self-correcting theory of quantum gravity, so it seems unlikely that anybody will come across the correct theory of quantum gravity starting from this direction.

  32. Garbage says:

    To Peter Shor,

    and why not.

    I agree, unitarity is not such a sacred concept. Also, axiomatic constraint may be very well tested.

  33. Peter Woit says:

    Peter Shor,

    My comment just reflects my often expressed worry about quantum gravity: what if there are lots and lots of them, and there is no way to ever get any experimental evidence about which one corresponds to reality? If you believe the most optimistic claims of string theorists these days, they have not one consistent theory of quantum gravity, but 10^500, and nothing at all about these theories is experimentally testable. This seems to me to be highly problematic.

    It’s certainly true that you may learn something generic about quantum gravity by thinking about all these models. But, I’m really a particle physicist, and you’re not going to learn anything about particle theory by thinking about them. Where string theory has failed is not as a theory of quantum gravity, but as a theory of particle physics. If string theorists would just admit that, and formal particle theory was no longer so heavily dominated by string theory, we’d be a lot better off.

  34. anonymous says:

    I have never heard a string theorist say string theory gives rise to 10^500 theories of quantum gravity. Many of them think there is one theory with 10^500 solutions. It is a very big difference.

  35. Garbarge says:

    “Many of them think there is one theory with 10^500 solutions.”

    Well, it might be 5 or so after all 😉

    Peter, I think the main purpose of ST is to provide a theory of quantum gravity, which happens to include all of particle physics [Something inevitable, if you believe “Georgi”]
    String theory has some achievements in that area. Like the entropy calculations/predictions for extreme BH, although it fails in more reasonable circunstances. Also one could say the ADS/CFT correspondence opens the bridge to understand gravity as a field theory as well. I agree ST does sort of what Peter Shor says, although still falls short to get the M-picture. I personally think the “background independence” issue is a big *problem* not yet fully addressed/understood. I asked once Ed Witten about this, his answer was:

    “…another point is that though string theory was discovered historically in the context of a fixed spacetime background, it turns out that the string describes fluctuations in the geometry, though we don’t understand very fully how this happens.”

  36. woit says:


    I’m well aware of what the string theory ideology is, with its “one unique string theory”, which no one actually knows the definition of. The fact of the matter is that current known definitions of string theory depend on a choice of “background”, and many string theorists believe there are 10^500 or more different such choices.

  37. Chris W. says:

    I have never heard a string theorist say string theory gives rise to 10^500 theories of quantum gravity. Many of them think there is one theory with 10^500 solutions. It is a very big difference.

    Given that any one of those solutions is potentially an effective theory for describing the entire observed universe, and a failure of any one of them in fulfilling that goal merely sends us off on a wild goose chase through the other 10^500 looking for a solution that does work, one can say that we have 10^500 theories, for all intents and purposes.

    We would be in somewhat the same predicament in atomic physics (non-relativistic quantum theory, shall we say) if it allowed for millions of solutions (stable or meta-stable configurations), we only had only example of such a system to test the theory against, and we couldn’t independently observe the configuration to prevent the theorist from saying “no, wait, it’s not a helium atom in state X,” or “it’s actually a boron atom in state Y [or an iron atom in state Z, etc, etc]”. Now think of the options afforded us by a vast array of vacua, not to mention braneworld scenarios, and on and on.

    Arguably, the central lesson of investigations into quantum gravity is that the issues are inherently cosmological in character. This enormously raises the stakes and and increases the risks of running off the rails epistemologically and methodologically. I can see why a sober particle physicist would just as soon avoid the entire territory while attempting to understand the existing problems of the Standard Model. The question is, is this really possible? What could a satisfying solution of these problems look like without involving gravity, or the dynamical structure of spacetime, in an essential way?

  38. Aaron Bergman says:

    The fact of the matter is that current known definitions of string theory depend on a choice of “background”, and many string theorists believe there are 10^500 or more different such choices.

    If you want to be a stickler about it, none of the examples with all moduli stabilized have a description as a perturbative string theory at all because of the presence of Ramond-Ramond fluxes. However, if you believe that string theory has supergravity as a low energy limit, you can tunnel between vacua and see that they cannot be considered separately.

    On the other hand, Tom Banks might argue that none of these things will change the asymptotics of the solution and that different choices of asymptotics should be considered different theories.

    Frankly, I’m not sure why anyone should care whether string theory should be thought of as a framework (like QFT) or as one theory with zillions of (not-so-rigorously constructed) vacua. What matters are the results of the calculations that you do.

    If you believe Steve Carlip, there are plenty of inequivalent quantizations of 2+1D quantum gravity, so perhaps the same holds 3+1 also. I still think that string theory is interested because, at least with AdS/CFT, I believe that it really is a theory of quantum gravity. As such, even if it’s not the right theory of quantum gravity, we can learn things about how the usual problems associated to quantum gravity might be solved. String theory has succeeded towards that end in some respects (black hole entropy, for example) but not entirely in others. (The black hole entropy calculation does not ‘fail’ in far-from-extremal situations as in give the wrong answer, in case there was some misunderstanding; it is just not doable yet.)

    For Peter Shor, we know that AdS/CFT is unitary, and there the metric is fully dynamical. This argument is essentially what convinced Stephen Hawking as I remember it.

  39. Moshe says:

    Peter Shor, in my mind there are some indications in string theory that black hole physics does not require modifying QM, but there is nothing I know of that amounts to a watertight proof.
    I am curious though about something related, in particle physics circles it is common belief that QM is “rigid” in the sense that there are no consistent “small” modifications of QM (e.g adding small non-linear terms to Schrodinger equation), example that is cited often is Weinberg’s attempt to modify QM.
    On the other hand it is my impression that in other communities the sitaution is different. So, out of curiosity, are you aware of any model that modifies QM and reduces to it for observable physics? Do you feel like something like that ought to exist?



  40. nontrad says:

    Aaron B,

    Thanks for bringing attention to 2+1 QG. In the above posts discussing the posibility of several QG’s I immediately thought of 2+1, which is an odd place / ‘laboratory’ in more ways then one.

    Again, thanks for bringing attention to this point.



  41. anon. says:

    For Peter Shor: one of the arguments that is sometimes referred to in the context of claims that modifications of unitary QM are dangerous is a paper by Banks, Peskin, and Susskind, motivated by the black hole information problem:

    They argue that modifying QM to try to get around the black hole information problem leads either to acausal signal propagation or to violations of energy-momentum conservation. They further argue that these effects are not suppressed by powers of the Planck scale. Their argument applies to a class of modified QM theories with a “superscattering matrix.”

    I’m not expert in these things and I don’t know what the space of all reasonable modifications of QM looks like, so I don’t know whether there is any danger for the sort of thing you have in mind. I am fairly confident, though, that this is the paper that has devolved into the sort of “folk theorem” you mention that “any non-unitarity at small scales would also show up at large scales.”

  42. Tony Smith says:

    Peter Shor said “… Kitaev’s result about error correction with anyons … led Preskill (I believe) to speculate that the laws of nature might be non-unitary at bottom, but that this non-unitarity is masked by an inherent error-correction process, which makes everything look unitary at any scale significantly larger than the Planck length. …”.

    John Preskill said in his paper Fault-Tolerant Quantum Computation, quant-ph/9712048:
    “… In Kitaev’s spin models, we might imagine that localized processes that destroy quantum information are quite common. Yet were we to follow the evolution of the system with coarser resolution, tracking only the information encoded in the charges of distantly separated quasiparticles, we would observe unitary evolution to remarkable accuracy; we would detect no glimmer of the turmoil beneath the surface. …”.

    Stephen Adler said, in his book Quaternionic Quantum Mechanics and Quantum Fields ((Oxford 1995):
    “… If the multiplication is associative, as in the complex and quaternionic cases, we can remove parentheses in … Schroedinger equation dynamics … this … fails in the octonionic case, and hence one cannot follow the standard procedure to get a unitary dynamics. …
    [so there is a]… failure of unitarity in octonionic quantum mechanics…”.

    Would the Preskill idea of non-unitarity at Planck energies be consistent with an octonionic-non-unitary model at Planck energies that becomes quaternionic-unitary by “freezing out” a preferred quaternionic subspace at sub-Planck energies ?

    Tony Smith

  43. Garbarge says:

    hey anon,

    you might like to read this,

    along with the other references I pointed out before.
    There is no real *danger*, unitarity isnt such a sacred principle. One can conserve probabilities and charges (including H) without it. This has been known for a while in the theory of open systems (Lindblad formalism) . The nice thing about these papers is that the modified Schroedinger dynamics is derived from quantum gravity ideas, and applies independently of any other environmental decoherence effect. Also that the black hole paradox is rendered *practically* unobservable

    and provides a conceptually satisfatory solution to the problem of time,

    There are indeed attempts to experimentally detect this type of effects,

    nice reading… 🙂

    bed time…

  44. Peter Shor says:


    Many thanks for the reference. I’ve looked at it now, and I’m convinced you’re right: this is probably the paper that turned into the folk theorem. As for AdS/CFT, I’m not going to be completely convinced by this until somebody can explain exactly why the black hole information paradox goes away in AdS/CFT. I’ve discussed this with a couple of string theorists, and they admitted they didn’t understand it completely themselves.

    Peter Shor

  45. Lee Smolin says:

    Dear Peter Shor,

    If it really is the case that fault tolerant quantum computing can provide a counterexample to the Banks-Peskin-Susskind argument that is very worth working through carefully and writing up.

    As to small testible modifications from quantum theory, it has been known for a long time that one can add terms to the schroedinger equation that are non-linear in the wavefunction Psi, which violate unitarity but not conservation of probability and then use precision experiments like the Lamb shift or neutron diffraction to bound their coefficients. (I wrote such a paper in the mid 80s but it was already an old idea then.)

    As to consistent quantum theories of gravity which show that general principles like Lorentz invariance must be modified, let me recommend the recent Freidel-Livine work on 2+1 gravity with matter, which shows that Poincare invariance is necessarily quantum deformed as a result of including quantum gravity. (btw, not for the first time, I have no idea what the ambiguity is in 2+1 quantum gravity some people keep refering to. Carlip wrote about different appraoches to 2+1 gravity but my understanding is that with the same matter content they all give the same theory.)



  46. Aaron Bergman says:

    Let me quote from Carlip:

    Perhaps the most important lesson of (2+1)-dimensional quantum gravity is that general relativity can, in fact, be quantized. While additional ingredients—strings, for instance—may have their own attractions, they are evidently not necessary for the existence of quantum gravity. More than an “existence theorem,” though, the (2+1)- dimensional models also provide a “nonuniqueness theorem”: many approaches to the quantum theory are possible, and they are not all equivalent. This is perhaps a bit of a disappointment, since many in this field had hoped that once we found a self-consistent quantum theory of gravity, the consistency conditions might be strin- gent enough to make that theory unique. In retrospect, though, we should not be so surprised: quantum gravity is presumably more fundamental than classical general relativity, and it is not so strange to learn that more than one quantum theory can have the same classical limit.

  47. Lee Smolin says:

    Carlip concludes however (p 46).

    “A more general problem is to understand which of the approaches described here are equivalent. In particular, it is not obvious how much of the difference among various methods of quantization can be attributed to operator ordering ambiguities, and how much reflects a deeper inequivalence.”

    Here is why I suspect that it is the former. Let us consider the case of no matter, to which his comments mostly refer. Given a compact two topology and a value of Lambda, there is a fixed finite dimensional physical phase space for 2+1 GR. Given that the physical phase space is finite dimensional, there are only two ways that one can get inequivalent quantizations. First, from different choices of subalgebras of the physical phase space algebra to be quantized. Second, from different representations of those algebras. Both of these kinds of ambiguities are well understood in conventional quantum mechanics. One cannot expect the situation here to be better than in ordinary QM, but if it is no worse than it is misleading to say that different approaches lead to different theories.

  48. Hmm says:

    Peter Shor–

    Talk to better string theorists; the understanding of the information paradox is now relatively standard. The process of black hole formation and evaporation in AdS space can be desribed in the unitary boundary CFT. Furthermore, for the case of eternal black holes in AdS, there is a sharp form of the information paradox and its resolution as discussed by Maldacena in 2001, hep-th/0106112; this was the paper that convinced Hawking to capitulate. The sharp form of the paradox in this case is this: correlation functions for (say) a scalar field in the BH background naively fall exponentially to zero–whereas a unitary description would say that it can’t get smaller than exp(-S). Of course the dual CFT tells you precisely this. There is a separate question of how to understand the breakdown of Hawkings original argument. Maldacena (and subsequently Hawking) suggested that this arises from a sum over all bulk geometries, including ones where the BH doesn’t form, but this is still controversial. This does not change the fact that AdS/CFT provides the correct answer-no information loss.

    With all due respect to you and your great accomplishments in quantum computing, your questions and comments on quantum gravity are somewhat naive. This also goes for a number of your “the world in a quantum computer!” colleagues too. This makes your insinuations that the string theorists don’t know what they are talking about a mite annoying. Perhaps in the future you should consult some of the standard literature. Of course actually being an expert in some subject and actually knowing what one is talking about is completely at odds with the ethos of the blogosphere, especially here…

  49. woit says:


    My apologies. Your last comment ended up in multiple copies in the spam queue (presumably because it had multiple links, something this mysterious software is very suspicious of). I was trying to delete most copies, keep one, managed to irretrieveably delete them all. Please repost if you still have a copy.


Comments are closed.