Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

A critical look at strings - Carlo Rovelli

  1. Aug 6, 2011 #1

    tom.stoer

    User Avatar
    Science Advisor

    http://arxiv.org/abs/1108.0868
    A critical look at strings
    Authors: Carlo Rovelli
    (Submitted on 3 Aug 2011)
    Abstract: This is an invited contribution to the Special Issue of "Foundations of Physics" titled "Forty Years Of String Theory: Reflecting On the Foundations". I have been asked to assess string theory as an outsider, and to compare it with the theory, methods, and expectations in my own field.
     
  2. jcsd
  3. Aug 6, 2011 #2

    Physics Monkey

    User Avatar
    Science Advisor
    Homework Helper

    There is a lot to discuss here, and I really appreciate the attempt to cross the (unnecessary) boundaries between these two communities. Although there were a number of interesting points made, I will begin with my criticisms. I focus only on the quantum gravity aspects.

    1. How can any review about string theory not mention the word "duality"? The review certainly mentions some examples of duality, but no explicit attention is given to this. This suggests to me a major category of results in string theory that string theorists believe are extremely important. Are these results judged not important or not relevant?

    2. String theory contains GR as a special case. There is no mention of this fact as far as I can tell, or at least no mention of the way this classic result flows from the world sheet description. I think you can't call GR background independent and not give the same status to string theory. String theory contains and string theorists study all the interesting non-perturbative solutions mentioned near the bottom of page 3 on the left. Of course there is much to learn, but isn't string theory already as good as GR?

    3. Related to 2, string theory definitely contains classical GR, but as far as I know this continues to be uncertain in loop quantum gravity. I know we discussed this issue relatively recently, but has any progress been made since then? Rovelli and others may believe strongly that the classical limit is right (or maybe not), but in string theory it pops out in a completely unambiguous way, so surely this dichotomy should have been mentioned? If the pros think we're almost there, then I'm fine with that, but not mentioning this issue while simultaneously insisting on the calling ads-cft a conjecture seems unfair to me.

    4. Related to the end of 3, ads-cft may technically be a conjecture, but in my opinion this is disingenuous. Yes, the duality is not formally proven and some people have objections, but there is an enormous amount of evidence in its favor. Rovelli's review explicitly refers to ads-cft as not "addressing the real problem". I think that is an extremely unfair characterization. Ads-cft provides a complete non-perturbative definition of quantum gravity in certain asymptotic spacetimes with the right classical limit.

    5. I think the review was right to accept the UV finiteness of strings, but I would have liked to have seen more about this issue and the related issue of "quantum geometry". What about T duality? What about topology change? What about singularity resolution?

    I'd be happy to change my mind if I missed something in my reading, but right now I feel like a lot of important stuff was left out.
     
  4. Aug 6, 2011 #3

    marcus

    User Avatar
    Science Advisor
    Gold Member
    Dearly Missed

    The critical look explicitly refers to string dualities and certainly uses the word :biggrin: Look on page 4, first full paragraph.

    I think it is misleading to term this brief 7-page critical look a "review". The review would, I guess, be the whole special issue of Foundations of Physics journal, which will be devoted to a retrospective "forty years of string theory".

    Gerard 't Hooft took charge of compiling and editing this special issue of Foundations, and he gathered contributions from many people---mostly string theorists. I think he probably invited Carlo Rovelli specifically to give a critical look from an outsider's perspective. Certainly not what one usually considers to be a full review article!

    Regarding this 7-page contribution by Rovelli, the question is not 'how complete is the review?' but 'are the criticisms well posed?'

    But as for dualities, strong positive mention is explicitly made of them. You may have missed it in a hasty reading.
     
    Last edited: Aug 6, 2011
  5. Aug 6, 2011 #4

    Physics Monkey

    User Avatar
    Science Advisor
    Homework Helper

    Fine, don't call it a review. Call it an assessment. I just needed a noun. Drawing attention to my nomenclature is an irrelevant semantic dodge since Rovelli does review many aspects of the string framework. Unless you or someone else wants to suggest that calling it or not calling it a review is actually an important issue?

    Note that I didn't ask for completeness, I suggested that the assessment was missing major crucial pieces of the physics. There is a long way between missing major idea clusters (which I hope we agree would lead to less enlightened comparison) and missing details which no 7 page work could be expected to contain.

    And you're right, I made a mistake. He does use the word duality once (in a different tense than I had searched for after not seeing it), but nothing is said about it! Or almost nothing, except that the relations are "beautiful" and difficult to evaluate the significance of. If missing that is hasty reading then so be it. Furthermore, the statements at the beginning of that same paragraph betray a certain prejudice about the way string theory is supposed to work. Maybe there are no "basic degrees of freedom", this is after all part of the message of dualities.
     
  6. Aug 6, 2011 #5

    marcus

    User Avatar
    Science Advisor
    Gold Member
    Dearly Missed

    It seems to me that he took pains to be fair to the string approach---and more than that: kind, friendly, collegial. He had a lot of positive appreciation and praise for stringy achievements.

    The question as I see it is whether this very brief critical outside assessment could contribute some value or usefulness to the broad review of string foundations and progress that Gerard 't Hooft and the other editors are preparing.

    You may have seen other preprints of articles (by string theorists) contributed to this collection. What can the perspective of someone outside the field, working on a fundamentally different approach, add? What issues can he put in contrast that might not get such focused attention in the other papers?

    Some of the invited contributions to that special issue of Foundations of Physics.
    Giddings: http://arxiv.org/abs/1105.6359
    Rovelli: http://arxiv.org/abs/1108.0868
    Gubser: http://arxiv.org/abs/1103.3636

    Other articles written for the issue, which may be included.
    Balasubramanian: http://arxiv.org/abs/1107.2897
    Rickles: http://arxiv.org/abs/1004.4491
     
    Last edited: Aug 6, 2011
  7. Aug 6, 2011 #6

    martinbn

    User Avatar
    Science Advisor

    How does it contain GR as a special case? I have always been curious to see that.
     
  8. Aug 6, 2011 #7
    This appears a very fair overview to me. Rovelli indicates that he will be assessing String Theory according to precise bounds at the beginning (vis a vis LQG), so dualities are not discussed because they are rather the (powerful) mathematical motivation for the theory, something which LQG doesn't (tediously) constantly appeal to.

    Rather sneakily he does suggest that string theory might only fulfill the aims of its beginnings with Veneziano, in helping to explain hadron physics via improved QCD calculations

    In any case he has the very strong case that low energy SUSY particles haven't been found yet, and ST conveniently can get around that and any other finding at the LHC.

    Maybe both String Theory and LQG will end up as a branch of "toy" mathematics.
     
  9. Aug 6, 2011 #8

    tom.stoer

    User Avatar
    Science Advisor

    The weak point in the article could be that Rovelli compares strings with LQG. The restriction to quantum gravity and the focus on one single theory of quantum gravity could very well miss the essential nature of string theory.

    Let's assume for a moment that we have a rather nice effective theory for strong interactions, e.g. chiral perturbation theory. Now let's assume that for some reason we have not yet identified the SU(3) color-gauge theory for strong interactions, but a lot of other gauge theories (not relevant for strong interactions). Now let's compare gauge theory with chiral perturbation theory: our judgement would be that chiral perturbation theory wins agains gauge theory regarding the description of strong interaction ... what a misjudgment.

    String theory could very well be a framework for constructing consistent theories. Perhaps one should compare string theory with spin networks; here "spin networks" would mean "the set of all spin network theories constructed from arbitrary Lie groups, quantum groups, affine Lie groups, ..."

    Rearding many details and weak points of string theory a share Rovelli's opinion
     
  10. Aug 6, 2011 #9
    But he clearly states that the ONLY reason he agreed to write the article was to compare ST to LQG, so he is not arguing against the fact that ST might be applicable to the Standard Model of particles better than LQG, but rather that ST is fundamentally not convincingly physically motivated, perhaps because it is trying to do too much.

    Anyway, I think its a brave article in these times, and well worth a read.
     
  11. Aug 7, 2011 #10

    tom.stoer

    User Avatar
    Science Advisor

    Yes!
     
  12. Aug 7, 2011 #11

    Ben Niehoff

    User Avatar
    Science Advisor
    Gold Member

    This is a standard textbook result. One obtains the Einstein equations for the background metric by conformal symmetry on the world sheet (i.e., from the vanishing of beta functions for the worldsheet CFT). It should be discussed in Becker, Becker, Schwarz, and most any other string theory textbook.
     
  13. Aug 7, 2011 #12

    Fra

    User Avatar

    What I like about most Rovelli's writings is that he has a reasonably clear and honest writing style, which also means it's easy to see when you disagree.

    Knowing about Rovelli's logic and views since before, when I read the paper it provides no constructive new views on ST, to me it rather just confirms what really is Rovelli's stance.

    I have put forward my objections to Rovelli's reasoning before, and don't want to repeat myself, since what rovelli writes is stuff I've read before.

    But shortly, if I may be so bold to critique Rovellis thinking, I again must say that some of Rovelli's critique against ST seems (to me) be rooted in his own understanding of measurement theory, inferences and what a theory is. (Needless to say I disagree with his viewe). The reason I think that some parts are unfair critique is not because I like or advocate ST, it's because some of that critique can be closely applied to my own personal ideas. While I see that I'm noone compare to Rovelli I feel quite confident that something just isn't right about his application of "measurement theory" on the "observer invariants".

    The reason for difficulty of finding/understanding an observer invariant measurement theory is IMO much deeper that what Rovelli's stance seems to allow. IMHO, clearly the "background" comes with hte observer. Rovelli's way of doing away with the observer simply does not make sense to me, no matter how stupid that may make me look.

    I fully agree. But in a general sense (et give or take ST). Rovelli doesn't seem to consider this an option. This is one aspect of all the things I mentioned above. IMHO. degrees of freedom (or I prefer to say complexions because the continuum is not an a priori), are INFERRED, and inference requires a context(background; however not a static one). This is why there is a deep problem to even define the notion of DOF in a proper inference (which is to me what should take place in a generalizedr measurement theory).

    Some of this critique applies to ST as well, but it apparently applies even stronger to Rovelli's/LQG logic.

    /Fredrik
     
  14. Aug 7, 2011 #13

    martinbn

    User Avatar
    Science Advisor

    That is the answer I was given before, but what I found in the textbooks was: First part of the book covers the theory on flat ten dimensional space-time. When one wants to develop the theory on a different manifold one has to work with a Ricci flat ten dimensional manifold. Surely that is not the same as contains GR as a spacial case or classical limit.
     
  15. Aug 7, 2011 #14

    Ben Niehoff

    User Avatar
    Science Advisor
    Gold Member

    A Ricci-flat manifold is an Einstein manifold, isn't it? And it's not just that one must work in a Ricci-flat background...it's also that one obtains an action for the background fields, and that action is the Einstein-Hilbert action coupled to additional matter fields.

    To obtain the Einstein-Hilbert action coupled to matter fields in 4 dimensions, all you have to do is compactify somehow. There are lots of ways to do that.

    See also Clifford Johnson's book, D-Branes. One can use stacks of D-branes to create all sorts of background geometries, including the analogues of Schwarzschild, Kerr, Reissner-Nordstrom, etc., in all dimensions from 2 to 11.
     
  16. Aug 7, 2011 #15

    Fra

    User Avatar

    This is how Rovelli objects to the quest for a full B/I independent string theory:
    There are plenty of fair objections to ST, but the above seems to be a central objection of Rovelli, but I have to say that I can actually see a good logic behind the struggle that goes on in ST. And it can be partly understood in a way that has nothing to do with ST specifically, but rather that would be common to any framework that might aim at considering evolving theories (which is at least ONE possible route in ST, but there are others).

    IMO, the background unavoidably comes once you Pick a real observer. This is the observer doing ALL inferences. The observing system is also physical hosting and encoding a theory. In the general case the background is more than JUST spacetime, but the argument still holds for spacetime only if we insist on that B/I definition.

    The "indirect" adressing Rovelli objects to IMO, can be understood like this: The only way to measure or assess one theory, is from the point of view of another theory. Ie. The only way to infer background independence of a certain class of theories, you need another theory. In other words, it must actually be true that background independence paradoxally IS background dependent. The exceptions with GR - is since it's a CLASSICAL theory. I think it's a mistake to extrapolate this non-inference (from a realist theory) to the quantum domain. So the problems seems to actually BE how to understand a populating of interacting theories (backgrounds in the sub-problem). The quest for fundamental timeless eternal DOF's that are observer invariant seems to be a extrapolation from classical physics, where we do not have to worry about how one observer in detail infers(measures) something, and how they compare the result, because in the classical world this is much more trivial.

    The asymptotic correspondence of a boundary can be understood as how an observer at infinity would "infer" or assess the physics that's inside the bound. The only objection I have to this is not that of Rovelli but rather that the asymptotics does not seem to qualify as a general realistic observer. Asymptotics work for special cases, such as for assessing microscopic black holes Maybe, but it seems to be a special case only, because a finite observer can never access this asymptotic information in reality - and thus not make any inferences. But rovelli's critique seems to be quite different, he seems to address not such details but the general idea of two different theories describing each other. I think it does make sense in a very general sense (which you can appreciate without ST), it seems he does not?

    I think maybe some of this roots also in the obsession with falsifiability of theories. But it could very well be that the understanding of what a theory is, is needed. Again, I know alot of people has written that this is ST making excuses for this and that, but i've given this this good thought outside ST, and there are actually good rational thinking in this. This is interesting because it does touches not only the foundations of physics, but also of science.

    /Fredrik
     
  17. Aug 8, 2011 #16

    martinbn

    User Avatar
    Science Advisor

    Yes, but there is a difference between having a theory, which reduces to GR and being able to beging to build the theory only if you take a fixed Ricci flat manifold as a background.
     
  18. Aug 8, 2011 #17

    Ben Niehoff

    User Avatar
    Science Advisor
    Gold Member

    This is becoming kind of a tangential discussion, but anyway...

    Perhaps you have misunderstood. The whole point is that string theory demands GR as a consistency condition. Kind of like a Ward identity. In order for any relativistic quantum theory of 1-dimensional extended objects to be consistent, it must be on an Einstein background in 10 (with SUSY) or 26 (with bosonic modes only) dimensions.

    I'm not sure where you got the qualifier "fixed" from. The background metric G is a coherent state of string modes. The background itself is built out of a field of virtual closed strings being exchanged.

    In QED one solves the hydrogen atom by assuming a classical background (i.e., a 1/r potential) and looking at small perturbations from it (i.e., we solve the Dirac equation in the given background). Does this mean that QED only works for fixed backgrounds? No, if we were to look at the full interacting theory we should be able to find bound states of protons and electrons and the 1/r potential would come out as a low-energy solution to the theory rather than a background we put in by hand. But that is a very difficult calculation to do, as all our perturbative techniques are designed to attack scattering problems that do not result in bound states.

    The situation is similar in string theory. We know that in the full interacting theory at high energy, the "background" one would get is not even describable by traditional geometry. In some regimes it may be some kind of non-commutative geometry. In other regimes, who knows. This is a very difficult problem. But we do know that at low energies, the theory gives us things that look like geometry, and those geometries are required to solve the Einstein equations (with some matter fields). So the easiest thing to do is find a classical background, and then look at what happens with strings propagating on that background.
     
  19. Aug 9, 2011 #18

    martinbn

    User Avatar
    Science Advisor

    Yes, I am sorry about that.

    That is how I understood it. The consistency condition has to be imposed before we have that theory. When I hear that ST contains GR, I expect to see the theory and from it as a classical limit to derive GT, but I didn't see that. What i found was, we want to quantise strings in not necessarily Minkowski spacetime, and the only spacetimes we can do it on is Ricci flat. Perhaps I need look more into it.

    Most likely I didn't understand something. But I got it out of the textbooks.

    That is different. I have no problem with this, making simplifying assumptions when solving specific problems.

    This sounds like what I would like to see, but the books gave me a different impression.
     
  20. Aug 9, 2011 #19
    Relativists (used to?) have trouble with this picture because it wasn't clear why flat metrics should be preferential --- many took/take background independence to mean that all metrics should be considered equally. The particle physics tradition (of which string theory must be regarded as being born from) has always been about perturbation analysis of some sort --- thus the focus on vacuums and excitation on top.

    The thing is, from LQG I actually think the relativists might have finally found something "special" about flat backgrounds which distinguishes them (in the quantum regime) from others --- the so-call Ditt invariance; i.e. when you "refine" the combinatorial structure in LQG flat metrics are fixed points of this procedure, and reflects the fact that LQG works by "gluing" flat simplices together, a la Regge.

    It's worth (for relativists) to think deeply about the string view point. As various string theorists point out, the fact that GR is in string theory is a "trivial" consequence --- the whole edifice is designed to give weak-field GR! One day, we might even understand that, like gauge theories, string theory is an essentially unique (i.e. self-consistent) extension of QFT to include gravitons, and so necessarily reproduces the correct low energy picture, but one which is not incompatible with a more "covariant" point of view. For example, high energy scattering in QCD is perfectly reasonably calculated by considering perturbations to an empty vacuum, but one also reasonably concludes that in other regimes it is sensible to use a lattice method.
     
  21. Aug 9, 2011 #20
    Originally - but more recently one has seen glimpses of non-geometrical regimes and begins to understand "space-time geometry beyond space-time". For example, near singularities or more generally in regions of large curvature where notions of classical geometry break down. Certainly string theory is much more generally defined than in terms of smooth manifolds etc.

    I guess this is a fairy reasonable point of view: as formulated today, string theory is more akin to lagrangian QCD, and LQG (ie: whatever formulation of the many approaches that were proposed that will ultimately make sense) is more akin to lattice QCD.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook