Carroll and Johnson on Bloggingheads

There’s an interesting exchange today between Sean Carroll and George Johnson on They cover a range of topics, including the controversy over string theory and the role of blogs. Johnson describes his time last fall at the KITP, telling about drinking very expensive Scotch with Steve Shenker, and many discussions with the string theorists there who were concerned that they were getting bad publicity and didn’t know what to do about it. Some were concerned that Lubos Motl was not exactly representing them well.

Sean gives a fairly standard defense of the landscape (“maybe it really is the way the world works”), with no discussion of the main problem with the landscape, that it shows no signs of making falsifiable predictions that would make it legitimate science. Some other points of the discussion include speculation about “what if Feynman and Gell-Mann had blogs?”, and Sean’s analogy of the string/LQG debate with mainstream/”heterodox” academic economics.

: Sean has his own posting about this.

This entry was posted in Uncategorized. Bookmark the permalink.

81 Responses to Carroll and Johnson on Bloggingheads

  1. Chris Oakley says:

    Well – it’s over an hour, but I saw it through. Once again, we discover that Sean does not care whether the String Theory Landscape is ugly or not, he just cares about whether it is right.

    With all due respect to Sean, this is not the sort of view that a scientist should take. “Right” and “wrong” imply certainties, a luxury never afforded to the scientist. There are only good explanations and bad explanations. Invokes the Anthropic Landscape being a bad explanation, one should, at that point, or, maybe even before, be seeking something better, and if this involves discarding twenty years of work, then so be it.

  2. SnarkFest says:

    A mildly fanciful elaboration:

    “Maybe it really is the way the world works. If so, we’ll never be able to test this entrancing conjecture. We’ll have to content ourselves with asserting it, and reveling in the irrefutability of the assertion. It will be the greatest achievement of the human mind since G. W. F. Hegel proved that the solar system contains seven planets.”

  3. mr.M. says:

    I’m glad this debate happened and i wish there will be more in future. It’s more interesting if debaters don’t agree and they debate the issue.

  4. Chris Oakley says:

    “…It will be the greatest achievement of the human mind since G. W. F. Hegel proved that the solar system contains seven planets.”

    Well … just get them to de-list Uranus and Neptune, just like they have already done for Pluto, and he will have been right!
    The S.T. equivalent, I suppose, would be Susskind (or whoever) altering the definition of the term “science” to allow the Anthropic Landscape to be respectable.

  5. AGeek says:

    Two related questions come to mind:

    1) If some string theorists are concerned that someone is not representing them well, why don’t they represent themselves? Who’s out there blogging, besides Motl, Clifford and Distler?

    2) How come Carroll is seen taking the pro-string side of the debate so often? Not only is he not a string theorist, it is questionable if he can even be considered a physicist. Is there really nobody more suited than him for the “job”?

  6. Peter Woit says:


    Sean is definitely a physicist, although definitely not a string theorist. It is a somewhat weird situation that he’s taking a leading role defending the theory.

    One of Sean’s main arguments for string theory is that it has been victorious “in the marketplace of ideas”, i.e. it has achieved its dominant position in particle theory by convincing theorists it is the way to go. My take on why more sensible string theorists than Lubos haven’t tried to mount a serious defense of string theory against its critics is that they know that they would not fare well in open debate. String theory has not been doing well in the “marketplace of ideas” recently, and that’s not because Lee Smolin and I are marketing geniuses. I’ve been told that one prominent string theorist has said that he wanted to write a book refuting Lee’s book and mine, but couldn’t bring himself to, since he would have to directly refer to us by name. For the moment, I think string theorists have made the tactical decision that they won’t do well in an open marketplace of ideas, that entering it would just give serious criticism more exposure. If you were promoting an idea like the anthropic landscape, wouldn’t you do what you could to stay away from forums where you would have to answer your critics?

  7. AGeek says:

    Carroll’s academic degrees are all in astronomy

    Based on that, and on what I’ve seen of his work, I wouldn’t call him a physicist.

    His talk of “marketplace of ideas” has been criticized by others as trivially flawed: in academia, you don’t have producers and consumers, with the latter only concerned with choosing the best product offered by the former; everybody is a producer. Imagine ten Apples all trying to sell iPods and iPhones to each other… the metaphore may reveal Carroll’s level of understanding of economics, but little else.

    If I were promoting an idea like the anthropic landscape, I guess I’d concentrate on the people who really count for the bottom line. That would be funding agencies and workers in more or less related fields, especially cosmology. I would probably not worry too much about the general public unless there were a real outcry about wasted funding threatening to turn into political action.

  8. DB says:

    Carroll is very slick. Very PR savvy. Very mediagenic. I think you’ll see more of him as “the acceptable public face” of string theory. Motl was a PR disaster for string theorists – and a boon to critics. My bet (but what do I know anyway) is that Sean would like to carve out a niche for himself in the lucrative area of popular science presentation. A guy as smooth as Sean will be able to turn on a dime and dump the string theorists whenever he wants.

  9. Peter Woit says:


    “lucrative area of popular science presentation”

    I know for a fact how much Sean made in one of his recent media appearances, since I was promised a check of the same size. After taxes, the sum is in the high two figures. I doubt bloggingheads pays even that well.

    As far as I can tell there are extremely few physicists or cosmologists making enough money from “popular science presentation” to support even an extremely modest lifestyle. My guess is that you could count them on one hand, with fingers to spare. I doubt Sean is doing this because it’s “lucrative”, because it really isn’t, except in very unusual cases (basically I think you need to have a successful TV show). People’s motivations are complicated, and I think every scientist who makes it part of their mission to get ideas about science before the public has some mixture of an altruistic desire to enlighten others and a less-altruistic desire to get attention for themselves. Rarely does money have much to do with it at all, it seems to me.

  10. DB says:

    Fair point. It doesn’t usually pay well, although a bestselling book can be a real moneyspinner (e.g. Brief History of Time), especially if tied into a successful TV show (e.g. Sagan’s Cosmos).

    Anyhow, Sean’s May rallying call in New Scientist
    is having the requisite effect today:

    If I’m not mistaken, dark energy is Sean’s area, so it will be interesting to read his take on this.

  11. Bee says:

    @AGeek above: Reg. the ‘marketplace of ideas’, you might find this interesting.

  12. Bee says:

    sorry, off-topic, but you just have to see this paper

    Search for Future Influence from L.H.C
    Authors: Holger B. Nielsen, Masao Ninomiya

    “When the Higgs particle shall be produced, we shall retest if there could be influence from the future so that, for instance, the potential production of a large number of Higgs particles in a certain time development would cause a pre-arrangement so that the large number of Higgs productions, should be avoided.”

    what an ingenious idea! Peter, one shouldn’t worry, I think ST just has a backward causation: Every potential observability of stringy effects just causes a pre-arrangement such that the observation is avoided. Thus, no observation is the correct prediction. Best,


  13. Peter Woit says:

    Thanks for mentioning the Nielsen-Ninomiya paper. I noticed it, was thinking about writing something about it since it is so bizarre, but couldn’t think of anything sensible to say…

  14. gunpowder&noodles says:

    Maybe the NN paper is some kind of elaborate joke? Write on a piece of paper, “THE LHC WILL HAVE ITS FUNDING CUT OFF” and see what happens to Higgs observations……

  15. Eric Mayes says:

    I’m pretty sure this paper has absolutely no connection to string theory. Isn’t it a little unfair to give lay readers the impression that it does?

  16. Peter Woit says:


    My, my, you string theorists are mighty defensive these days.

    I made no connection of this to string theory at all. But now that you mention it, one could note that this paper at least proposes a falsifiable test of their idea at the LHC, which is more that string theory is able to do…

  17. Eric Mayes says:

    Bee made the connection with string theory. However, there are several falsifiable tests of string theory at LHC, namely supersymmetry and large extra dimensions. You may claim that a discovery of either of these does not prove string theory, however either of these would be very strong evidence in it’s favor.

  18. Peter Woit says:


    You seem to have a non-standard definition of “falsifiable”. Supersymmetry and large extra dimensions are not falsifiable predictions. Their status is the usual one of string theory “tests”: if we see them, that’s great for string theory, if we don’t, no problem. Heads we win, tails doesn’t count…

  19. Eric Mayes says:

    Sure, it’s falsifiable. All we need is a big enough collider. Now, it may be that LHC isn’t big enough, but this a technological liimitation not a fundamental one. However, it is clearly possible to find evidence in favor of string theory at LHC, even if LHC will not be able to rule it out.

  20. Peter Woit says:


    “there are several falsifiable tests of string theory at LHC”

    “Sure, it’s falsifiable. All we need is a big enough collider. Now, it may be that LHC isn’t big enough”

    Funny the way these string theorists keep changing their story when you point out to them that they’re wrong…

  21. Eric Mayes says:

    Do you deny that it’s possible, in principle, to find evidence in support of string theory at LHC?

  22. Peter Woit says:


    Sure, although I don’t think it’s likely. But it’s really a tedious waste of time dealing with the comments that you keep posting here which make completely false claims (e.g. “there are several falsifiable tests of string theory at LHC”), which you never acknowledge are untrue, instead continually returning here to make the same false claims again and again.

  23. Eric Mayes says:

    If you don’t think it’s likely that supersymmetry will be found at LHC, then you are in for a rude awakening. Supersymmetry, by the way, is the smoking gun of string theory.

  24. Chris Oakley says:

    Sure, Eric. Superstrings being – let me guess – the language in which God wrote the world – ?

    Just for the benefit of lay readers – Supersymmetry is, and always has been in conflict with experiment. No pair of known particles are classifiable as supersymmetric partners of each other (fermions and bosons of adjacent spin would need to have the same mass). So supersymmetry is wrong. It does not agree with experiment. The talk is therefore about broken supersymmetry, which means a supersymmetric theory that has been doctored according to a well-defined procedure known, through abuse of the word “spontaneous”, as spontaneous symmetry breaking. However, the ways in which the broken supersymmetric theories can be constructed are legion, and however little evidence appears for supersymmetry in LHC and similar experiments, it is always possible to claim that supersymmetry is still there, but is broken at a higher energy.

  25. super says:


    It seems to me a little dishonest to say that SUSY is “doctored” because it must be broken in nature. The gauge symmety of the standard model is also broken, which doesn’t mean that it’s wrong.

  26. Eric Mayes says:

    Yes, we know supersymmetry must be broken at low energies. However, there are very good reasons for the superpartners to have 1 TeV scale masses, just in the range that will be probed by LHC. Now, it’s possible that the superpartner masses will be much greater than this, but in this case we are giving up naturalness and introducing fine-tuning into the Higgs potential. One could make the case that the reason for some people to introduce the anthropic landscape is to justify unnaturalness and fine-tuning.

  27. Thomas Larsson says:

    “However, there are several falsifiable tests of string theory at LHC, namely supersymmetry and large extra dimensions. ”

    Supersymmetry will not be discovered at the LHC.

    Proof: String theory predicts supersymmetry (Witten 1984-2002). String theory predictions are always wrong. Hence supersymmetry does not exist, and will in particular not be found at the LHC. QED.

    Lubos Motl will lose his experimental-susy-by-2006 bet.

  28. Eric Mayes says:

    I thought string theory predictions were ‘not even wrong’, so doesn’t this invalidate the proof of your theorem?

  29. ori says:

    Regardless of whether string theory is right or wrong, it is consistent mathematically, and does decribe on the purely formal level unified quantum gravity. I think that any theoretical physicist interested in these ideas must know string theory, even if he does not believe it. It has proved useful in calculating highly non trivial objects in gauge theories, with lots of possible uses to real QCD (see Alday-Maldacena for instance) and needless to say, many results in mathematics.

  30. lyme says:

    Nielsen, who has made superb contributions to HEP in the past, has been saying that kind of crazy things for a long time. Somehow, scalar particles seem to trigger a kind of mystic madness in him.

    Or maybe he’s just joking, which would be equally crazy. I’ve seen him speak at the ICHEP conference in 1996, saying the same kind of nonsense. Cracking the same joke time and again for eleven years would certainly not be a sign of mental health…

  31. woit says:


    The line “string theory is mathematically consistent” is stated as a fact in its favor quite often, but maybe you can tell me precisely what theory it is that is mathematically consistent. Is it

    1. the string perturbation expansion?

    according to d’Hoker/Phong, in flat space this is well-defined through two-loops, quite possibly it is well-defined at any order, and quite possibly you can extend this to other backgrounds. But all evidence is that it’s a divergent series. You can’t sum it, and you can’t cut it off at fixed order without violating unitarity.

    2. non-perturbative string theory?

    last I heard, no one knew what that was. In some special cases, via making the AdS/CFT conjecture a definition, I suppose you could define it in terms of a QFT, but then it’s not so clear why you shouldn’t just be studying QFT. In this context, string theory is just a conjectural approximate calculational method for certain QFTs, something which is quite interesting, but not necessarily fundamental, and not necessarily of interest to all particle theorists.

  32. Who says:

    this just out. Can anyone enlighten me as to its possible significance?
    Conformal Gravity Challenges String Theory
    Philip D. Mannheim
    8 pages. Proceedings write-up of talk presented at PASCOS-07, Imperial College London, July 2007
    (Submitted on 16 Jul 2007)

    “The cosmological constant problem and the compatibility of gravity with quantum mechanics are the two most pressing problems in all of gravitational theory. While string theory nicely addresses the latter, it has so far failed to provide any compelling solution to the former. On the other hand, while conformal gravity nicely addresses the cosmological constant problem (by naturally quenching the amount by which the cosmological constant gravitates rather than by quenching the cosmological constant itself), the fourth order derivative conformal theory has long been thought to possess a ghost when quantized. However, it has recently been shown by Bender and Mannheim that not only do theories based on fourth order derivative equations of motion not have ghosts, they actually never had any to begin with, with the apparent presence of ghosts being due entirely to treating operators which were not Hermitian on the real axis as though they were. When this is taken care of via an underlying PT symmetry that such theories are found to possess, there are then no ghosts at all and the S-matrix is fully unitary. Conformal gravity is thus advanced as a fully consistent four-dimensional alternative to ten-dimensional string theory.

  33. Peter Woit says:


    Please, this has nothing to do with the posting, and this is not a general discussion forum for new ideas about quantum gravity. The idea of “conformal gravity” has been around for a very long time, with lots of people claiming to have ideas about how to deal with the problems caused by its fourth order action. This is one more, I have no idea whether it works. You should use another forum to discuss this issue.

  34. Bee says:

    Hi There,

    I’ve put some remarks on the bizarre paper here. In case it wasn’t clear, my above comment about string theory was meant as a joke.


    you write Do you deny that it’s possible, in principle, to find evidence in support of string theory at LHC?. It seems to me you are kind of missing the point. The question is not whether there is ‘in principle’ a ‘possibility’ to find something or the other, but whether there is a prediction it has to be found, or otherwise the theory the prediction was made with is falsified. The problem is if the LHC doesn’t find anything like SUSY or extra dimensions, you can still argue ST has not been outruled.

    – B.

  35. ori says:


    it is string theory as a whole on all of its thousands of papers with independent calculations which have never contradicted each other. Surely, the whole object is not yet well defined mathematically and non perturbatively, but this does not stop people who are physics-minded to do many non trivial calculations and reach fascinating conciusions which have tought us many lessons in mathematics and physics (and were confirmed in many independent ways).

  36. Peter Woit says:


    “string theory is consistent mathematically”

    “the whole object is not yet well defined mathematically and non

    It’s pretty misleading to go around telling people that you have something mathematically consistent when you don’t have a definition of what it is…

  37. Eric Mayes says:

    Science is about research and discovery. Sure, it’s true that we can’t guarantee that we’ll find anything, but this is the case with any theory. This is what we like to call ‘research’.

  38. anonymous says:


    It is most definitely *not* the case with any theory. Einsteins theory of general relativity made specific concrete falsifiable predictions. If those predictions did not pan out then the theory would have been proved wrong.

    String theory is not like that. String theorists have not offered any falsifiable prediction. Unless and until string theory can offer up a real prediction you can not lump it in with ‘any theory’, but of course you already know all of this.

    Eric, why do you insist on tip toeing around the elephant in the room? It has been pointed out time and time again and you are obviously intelligent enough to see it, so why do you ignore it and use word games to prance around it like it isn’t right in front of your face?

  39. Santo D'Agostino says:


    The terms “theory” and “prediction” are sometimes used in non-traditional ways, and that has made some discussions of string “theory” unproductive. I note that recently there has been some movement towards using these terms in their traditional sense. For example, David Gross has said that we don’t know what string theory is, and Barton Zweibach has said (in his textbook) that string “theory” makes no “sharp predictions”. Traditionally, there is only one kind of scientific prediction, so the phrase “sharp prediction” is not a return to standard usage, but at least it acknowledges that there is a problem.

    Is there a string “theory” calculation for the specific mass of a particle that has not yet been observed? Or some other specific quantity that has never been measured? If so, then we have a prediction. If not, then let’s just admit this. And if there is no prediction, then it’s difficult to see how anything measured in any experiment could provide support for string “theory”.

    I have no problem with any individual person pursuing research in string “theory”. It has a number of features that seem to promise great riches. However, “we don’t know what it is yet” precludes it being called a scientific theory. A possible scientific theory in formation, maybe, but not a theory.

    All the best,

  40. anon2 says:


    Suppose Einstein had instead invented a scalar-tensor theory. This theory can be made to look arbitrarily close to GR by making the scalar heavier and heavier. It will predict some fifth force which, if seen, would confirm the theory. However, the theory can never be ruled out, you just get stronger and stronger upper bounds on the scalar mass. Would you consider this theory “not even wrong” and accuse its advocates of promoting religious nonesense in the guise of physics?

  41. anonymous says:

    “Would you consider this theory “not even wrong” and accuse its advocates of promoting religious nonesense in the guise of physics?”

    Yes, I would say they were “not even wrong.”

    OTOH, I wouldn’t say they were promoting religious nonsense. String theory might well be a useful mathematical tool, but it is not a scientific theory. It offers the observer no predictions, no scaffolding on which to better understand the real physical world.

    String theory is just speculation(s). It is a proposal for some 10^500 possible scenarios that -might- be happening at scales and magnitudes completely beyond our ability to penetrate. It does not offer one iota of explanation for the large mysteries of physics that have been discovered during string theories incubation.

    When the special theory and quantum mechanics were being developed theorists had some shocking mysteries they were struggling to understand. The constancy of the speed of light… The mysterious nature of the photon… The theories developed during this time helped explain the what and why of these experimental phenomena.

    The same can not be said of string theory. With string theory, if some new and shocking experimental finding or observation comes along the string theorists run along not long after and shout, “We can fit that in string theory too!”

    Glashow’s comment hurts the sensibilities of string theorists so much because it cuts so close to the truth.

  42. Peter Woit says:


    When I accuse landscape proponents of engaging in pseudo-science, it is because they are promoting something very different than what you describe, something that makes no predictions of any kind (or to the minimal extent that it does, some are already falsified), not predictions that depend on a parameter.

    Sure, it’s conventional scientific activity to speculate about what happens when you extend tested theories by adding one or more parameters, and trying to put bounds on these. If there were a unique string theory, with just an undetermined scale, extracting predictions from it and putting bounds on the scale would be a standard kind of scientific activity. But that’s not what string theory is like. Instead you have exponentially large numbers of possibilities, any of which still contains an undetermined scale (or scales). This is not conventional science, and resembles in no way a one-parameter scalar-tensor theory.

  43. Eric Mayes says:

    Does quantum field theory by itself offer any falsifable predictions? The answer is no, it does not. To get any predictions, you must input into the theory information such as gauge groups, particle masses, gauge couplings, all of which are not fixed by the theory itself. Does this mean that QFT is not science? Absolutely not. The same is true of string theory.

    On the other hand, there are theoretical and mathematical criteria that may be used to falsify various candidate theories of quantum gravity. To date, string theory is the only such theory which passes these test, which is the reason it is studied.

  44. King Ray says:

    “The value of a principle is the number of things it will explain.”—Ralph Waldo Emerson

  45. woit says:

    King Ray,

    String theory promises to explain everything, but so far explains nothing. So, it achieves the highest possible quotient of promised value/actual value (infinity/zero). Perhaps this is why string theorists like to describe it as the “most promising” principle for unifying physics.

    Then there’s the landscape, which explains everything with the explanation that it can’t be explained…

  46. Santo D'Agostino says:


    You said:

    “Does quantum field theory by itself offer any falsifable predictions? The answer is no, it does not. To get any predictions, you must input into the theory information such as gauge groups, particle masses, gauge couplings, all of which are not fixed by the theory itself. Does this mean that QFT is not science? Absolutely not. The same is true of string theory.”

    I agree with you, except for the last sentence. Would you please describe a calculation from string theory that goes through a process analogous to the one you describe for QFT and results in a (specific numerical) prediction?

    If, at the very least, it is possible using string theory to calculate quantities that have already been measured, and whose values are well established, then string theory might qualify as a scientific theory (there are other criteria as well). If not even that is possible, then one might say it is a theory in development, and so work on it would certainly qualify as scientific activity, but how could you call it a theory?

    All the best,

  47. anon. says:

    Santo, you are right. In quantum field theory, the few inputs to the standard model can easily be found from observed data, based on experiments. But for string theory, there are many vital inputs that can’t be found from experiments, because they describe features of the 6 compactified dimensions which (are supposed to) exist at the unobservably tiny Planck scale. Hence string theory is a failure.

  48. King Ray says:

    Peter, you have hit the nail on the head once again.

    As Feynman said, if string theory is so great then they should be able to calculate the mass of the electron. Forget about predicting superpartners, just show us a calculation for the mass of the electron!

    I think the string theorists are dodging honest work; they are so divorced from reality that they don’t even know when they are lost. They seem to be divergent rather than convergent thinkers. To get something done you have to limit the possibilities you’re going to explore.

    I would start by eliminating the extra dimensions, and stop spending so much time in 2 and 3 dimensions, which in a lot of cases seems to be just avoiding the difficulties in 4 dimensions. Also, if you can’t handle 4 dimensions, how are you going to handle 11?

    If there is no evidence for something, such as extra dimensions, then why should you believe in it with such vigor?

  49. ori says:

    such a situation occured many times in history. I am sure that if you lived in the era of Dirac you would go around and say that whatever Dirac does does not make any sense since this strange delta function object is not well defined, and you can not say that you do something well defined when you don’t have axiomatic understanding of it…

    This is why physics is physics and math is math, and it is good that it is this way. The sole fact that people have been able to compute so many non trivial things (mutually consistent) without a rigorous definition of the theory clearly implies that there is such a definition, and it will be found in the future.

    Your objection have been raised many times in the history, and was never relevant. Whenever something in physics works, even if it seems ad-hoc, it must make sense in some more well defined frame work. The examples are too many to list, but I am sure you understand this point very well.

    I also wish to emphasize that all string theorist are very interested in finding a more natural and more axiomatic framework to unite whatever they were doing in the last 25 years – this is not something which is being ignored. The evidences for the existance of such a formulation is overwhelming, there is no doubt that the theory is mathematically consistent (in similar way that Dirac’s delta functions turned out to be not an ad hoc object with no clear rules for the game) as I said.

  50. Peter Orland says:


    Criticizing string theory is like criticizing the Dirac delta function? Are you kidding?

Comments are closed.