Sometime around now is the tenth anniversary of my first foray into the business of public criticism of string theory. I wrote something up over the end-of-year holiday in 2000, and circulated it by e-mail to a list of prominent theorists (some of whom I knew, some I didn’t), asking for advice. The main motivation was that it seemed to me that it had become clear that string theory had failed as an idea about unification and while this was increasingly well understood in the particle theory community, the news had not gotten out to the wider world. Instead, a fairly active campaign to promote string theory to the public continued unabated. This was a rather peculiar situation, one that I felt someone should do something about, and I was curious what my correspondents thought of it. Most responded with quite interesting comments on their views on the matter, and one of them put me in touch with an editor at Physics Today. After some back and forth with Physics Today it became clear that they were unlikely to publish anything on the matter, especially from me, so in February I posted what I had written to the arXiv as String Theory: An Evaluation.
Hard as it is to imagine, back in those days there were no physics blogs. Perhaps the closest thing was Usenet newsgroups, especially sci.physics.research, where John Baez and others had taken on the thankless job of moderating discussions which often addressed issues about string theory. The archive of these discussions is here, and some discussion of my arXiv piece broke out there, appropriately in a thread about Lie algebra cohomology. For my first posting joining that discussion, see here. This led to my first encounters with the surprising phenomena of Jacques Distler and Lubos Motl.
Scientifically, not much has changed in ten years, but the public perception of string theory has changed a lot and become much more realistic. The next decade will undoubtedly be dominated by the effects of whatever we learn over the next few years at the LHC, although I don’t think this is likely to affect proponents of string theory unification very much. Many of the remaining defenders of this idea are by now pretty well dug in and make it clear that “never give up” is their policy, even if it involves abandoning all hope of understanding this universe and putting faith in the existence of others.
One outcome of this that I never expected ten years ago is that I now have a book that has been published in Czech, with the title Dokonce ani ne spatne. I can’t read a word of it, which doesn’t matter except that I’m intrigued to see that Martin Schnabl has written an afterword and wonder what he has to say. The publisher just sent me a few copies, but I can’t think of anyone I know who reads Czech to give them to. Other than Lubos, of course, but I suspect he wouldn’t appreciate the gesture…
I don’t speak any Chekh but Google Translate helped me with an interview that Martin Shnabl gave to newspapers:
http://translate.google.com/translate?u=http://www.lidovky.cz/nejde-o-moc-ale-o-dobrou-vedu-dp8-/ln_noviny.asp%3Fc%3DA101130_000112_ln_noviny_sko%26klic%3D240160%26mes%3D101130_0&hl=en&langpair=auto|en&tbb=1&ie=windows-1250
He says that string theory is getting money in the United States because it is the best science that is available and it has no competition. Everyone else is just waving his hands and no one has yet challenged string theory as a science, Shnabl concludes.
Were it not for the generally high profile of theoretical physics, I think the last ten years of controversy would probably be described as an academic storm in a teacup by most. However, given the central important of theoretical physics in the pantheon of the sciences, I think this storm was one well worth having.
Given the intersection of science, (academic) culture, society and new technology(by which I mean blogs), I would also say that the last ten years of the String Theory debate will be a topic of some historical interest.
It is not about power, but about good science
November 30, 2010
No-one invalidated string theory yet, says Martin Schnabl from The Institute for Physics of Czech Academy of Sciences.
LN (the newspaper): Are top faculty appointments and generous grants really won mostly by the string theorists?
That’s true in US. Although I see the origin somewhere else than Peter Woit. According to him, the Stringmen have been usurping power and they are now holding onto their position no matter what. I just think they are doing the best science. But the situation in Europe is different. I have also won a large grant for string theory research myself but I was only second one in our field within Europe.
LN: What’s behind the string domination in US?
String theorists are getting grants because they are bright, hard-working and creative. They would have been successful even if they focused on another research subject. In reality the grant awards are based more on results and individual track record rather than hopes and promises. The scientists decide for themselves what they want to work on.
LN: But why are the physicists so often choosing string theory? Is there really no other alternative?
Fundamental physics can be approached in two ways. Going from bottom up, we find that there are abundant possible extensions of our current particle physics model. Scientist are testing them by performing experiments in accelerators, for example at CERN. It is an expensive but a very solid way of doing things. The string theory represents the opposite approach, from top down. It is the only one that explains gravitation within quantum framework. Since 30s when the scientists first started thinking about this, they had not come up with any other theory that could handle this. The alternatives, even if proposed with serious intentions, are still just little more than handwawing. String theory is in agreement with all experiments so far. Even though the String theory is pretty complicated and we don’t fully understand it yet, it seems to describe and unify all laws of physics. There is however a large number of possible solutions and each one of them can correspond to a somewhat different physics. That’s why it’s very hard to invalidate this theory
BTW Peter, have you seen this recent small article in the pop-ph arxiv (http://arxiv.org/abs/1012.5417)? Your book is presented as a representative example of contemporary criticism towards String theory.
the abstract mentions “sinergy”.
is that the finnish rock band ?
I would like to know how did you manage to write the book in a language you can not read.
Thanks Giotis, hadn’t seen that.
Peter,
There’s a translator, who presumably has done a good job…
milkshake,
Thanks. I like the argument for string theory that “it’s in agreement with all experiments so far.” There are lots of competitors though which can say the same thing, for instance my unified theory which is much simpler. It goes: “I dunno…”
I think there are some unintended funny bits in that interview:
… It is [accelerator experiments] an expensive but a very solid way of doing things. The string theory represents the opposite approach…
The verbatim translation gets even better: “It is an expensive but honest way of doing things. The string theory represents the opposite approach”
Peter and others, happy new year. anyhow what do you think of the fact that we do not have any experimental evidence for physics beyond standard model for more than 30 years? what do you think will be
the future of particle physics if LHC sees nothing?
Given the fact that we have not found anything are we on a completely wrong track (wrt supersymmetry, technicolor etc)?
Being the one who started the spr thread on Lie algebra cohomology, let me recall how the discussion started.
Anybody can criticize string theory on the grounds that it makes no hard predictions, and that all weak predictions (extra-dimensions, susy, etc.) disagree with experiments. Such critique, albeit well motivated, is not very constructive. In contrast, my point was more original and based on mathematical discoveries (by myself and mathematicians) that at the time were quite novel: the Virasoro-like extensions (an element in the Lie algebra cohomology, hence the title) of the spacetime diffeomorphism algebra and its off-shell representation theory. The Lie algebras that appear in string theory are either one-dimensional (Gelfand-Kirillov dimension = 1) or classical (no extensions, representations not of lowest-energy type), and this can not be good enough for a reality which is both four-dimensional and quantum.
At the time I harbored the idea that physicists might be interested in new mathematics, especially with such obviously close ties to the symmetry principles of both GR and QM. Ah, the naivité of the youth.
Easy: scan -> OCR -> Google translate
The “problem” with string theory (in my very humble opinion) is the fact that it is so very successful mathematically.
It is a framework, a way of thinking (that uses “physics words” as B Schroer would say) that effectively unifies almost all of mathematics. You want a short cut to very large chunks of 19th and 20th century mathematics? Read Wittens (and Vafa’s) papers!
How could that be? How can we explain that? I would say that this is the reason why die hard string theorists would never give it up.
Happy 2011!
Peter Woit: If you ever need help translating something of interest from Czech, just drop me a message. I am an occasional reader of your blog already for many years and I will be glad to help, if it is needed.
Anyway, I haven’t read a “popular science” book in ages, so this seems like a good place to start. I may even do a review for our magazine.
It is not surprising to me, that it is published in Czech, as the anti-string-world-dominance feelings were always quite strong among many physicists around here. I am not implying that the community here is anyhow against string theory (after all, I think we all value having Martin Schnabl here), only that plurality of opinions is (also historically) highly valued and the view advocated by some, that the string theory is the only truth in the universe, is simply unacceptable.
Greetings,
Jan
Hi,
I made a vow of not contributing to blogs but you make it so hard. Its the throw away comments that insult the hard work of so many. How can you say, “Scientifically nothing has changed”, in ten years! Are you really saying the intellectual endevours of so many for so long mean nothing. That is just not true. I am made happy by the excellent work produced by my colleagues over the last ten years. The huge number of insights and ideas. I cannot fathom a psyche that fails to see the beauty and wishes to see fault. There isn’t a Hollywood Eureka moment; but that isn’t science. There is the hard work of many and the results they produce. I am thankful and so should you, since exactly where would you be without string theory?
(How do you spell endevour?)
@Jan
you wrote: “I think we all value having Martin Schnabl here”
What about Lubos Motl?
David,
Obviously the “scientifically nothing has changed” remark refers to what the posting was about: the argument I gave ten years ago that string theory unification was a failed idea. It’s still a failed idea, for the same reasons as ten years ago. Actually, I was trying to be rather kind and not bring up the one thing that has changed, the promotion of landscape pseudo-science.
Outside of the unification failure, “string theory” is by now such a huge subject that of course there are plenty of interesting things learned over the past decade to point to.
And yes, I am saying that a lot of hard intellectual work by a lot of smart people turned out to be wasted as they worked their way into a dead end. But that’s how science progresses, most of what one spends one’s time trying to do doesn’t work. The problem arises when people refuse to admit that an idea they put a lot of effort into has failed…
There is one common theme that keeps repeating itself in this discussion as exemplified by this statement…
…besides Peter’s point that string unification has failed. Of course there will always be some kind of progress in a scientific community consisting of at least 500 highly qualified scientists, progress that will be appreciated and occasionally celebrated. The question is if there are any results that are of interest to anyone outside of the community.
If you take any obscure method and calculate a scattering amplitude of a process that can be observed at a collider, this is a result that is of potential interest to experimentalists, regardless of the involved theory. But if all of the results are about answering questions that no one outside the community would ever have asked, and no one outside the community can do anything with the question and its answer, this will be of course observed as “no progress” from the outside.
In this situation there no point in lamenting the lacking interest in the progress of one’s favorite topic.
Time says: “But if all of the results are about answering questions that no one outside the community would ever have asked, and no one outside the community can do anything with the question and its answer, this will be of course observed as “no progress” from the outside. ”
This is a reasonable querry from an interested outsider. The point however is that unification at the Planck scale is far away from our experimentally accessible scales, and we need quite a good control on the theory in order to make connections with low energy experiments. The “progress” that David talks about in string theory is (mostly) towards that goal: the goal of understanding the theory well enough. This is the necessary prerequisite to see whether it can say something about things of interest to “outsiders”. By definition, such progress is of not direct interest to an outsider, especially if only the result at the end of the tunnel is considered as worthwhile.
Except for some universal statements (like those regarding black holes, and AdS/CFT inspired consistency checks) the detailed tests of string theory are very likely to require such an understanding. The reason to study string theory is that the universal things seem to work(!), which is way more than what other attempts at quantum gravity can claim.
Peter talks as if the experimental inaccessibility of string theory is an obvious fact. This is FAR from the case. Especially in light of the successes of string theory in producing the universal aspects of low energy physics, it is very worth investigating this matter in detail.
But I do think that David should have resisted the bait. Peter thrives on these… 🙂
Somebody,
The problem with your argument is that all progress towards better understanding string theory in the past ten years has just given added force to the arguments for why string theory unification doesn’t work. Most notably, 10 years ago one could have claimed “when we figure out how to stabilize moduli”, then we will be able to make predictions. Instead what happened was that the moduli stabilization mechanism found gave us the landscape and a radically non-predictive unification scenario. Yes, this is progress towards understanding string theory, but it leads to understanding more clearly why string theory unification can’t work.
As for claims about “the successes of string theory in producing the universal aspects of low energy physics”, that’s just ridiculous.
Hi all,
I will resist the general discussion but in reply to Tim Van Beek, I couldn’t quite figure out if you know that indeed progress in QCD scattering amplitudes is happening as a result of a stringy understanding. My colleagues at Queen Mary are amongst those that provide calculational tools for people interested in QCD amplitudes (I believe they are incorporated into the state of the art monte carlos at CERN). So indeed your example is completely poigniant to the discussion but perhaps not in the way you imagine. Or perhaps I’ve misunderstood you.
I will now gracefully bow out and now the holliday is over go back to my day job.
Very fast damn Peter. You haven’t wasted time. Peter, I hate you. You’re just looking for money. I’m going to attack your blog. You hate physics, You hate the science.
Peter Woit said
“the moduli stabilization mechanism found gave us the landscape and a radically non-predictive unification scenario. Yes, this is progress towards understanding string theory, but it leads to understanding more clearly why string theory unification can’t work.”
In fundamental physics, we are trying to learn about reality. We have a working model with all sorts of unexplained numbers. We may hope that all those numbers are uniquely determined for reasons we don’t currently understand. But that is just a hope.
Meanwhile, we have string theory, with its landscape of solutions. Judging just by experimental adequacy – and not the apriori hope that all those numbers are uniquely determined – every single string vacuum which perfectly matches experiment is a candidate for the correct description of reality. This is true whether there’s only one such vacuum, or whether there are infinitely many of them.
So what is the situation in string theory? There are many promising classes of vacua, but not even one vacuum has been *proven* to fully reproduce the standard model. This is a symptom of slow progress on topics like moduli dynamics, though such progress is occurring (e.g. see 1101.0108 at the arxiv today). Eventually we *will* know whether some of those promising vacua really do match experiment, or whether none of them do.
The moment that a string vacuum is found which is both mathematically tractable and fully consistent with experiment, then string theory will finally have delivered a candidate description of the world. If it produces a hundred or a million such vacua, they will all be possibilities – and most likely they will fall into classes which do make distinct predictions about higher energies, making them subject to future falsification. Finally, if it can be shown that *no* string vacuum can match experiment, then string theory really will have been falsified. Only then would you be able to say that string theory can’t be the truth. And given the number of phenomenologically promising classes of vacua, it appears far more likely that string theory will instead provide the “standard models” of the future.
Mitchell Porter,
“Finally, if it can be shown that *no* string vacuum can match experiment, then string theory really will have been falsified. Only then would you be able to say that string theory can’t be the truth.”
Well, since there’s no way to computationally ever do that, there’s no way to ever show that string theory is wrong. Your conclusion about a theory that can’t be shown to be wrong is
“string theory will instead provide the “standard models” of the future.”
My guess is that the idea of using a theory that can’t be shown to be wrong as the “standard model” of the future is one that few physicists are willing to go along with.
Peter,
Let’s distinguish between falsifying string theory as a whole, falsifying a class of string models, and falsifying a specific string model.
So far as I can tell, the set of string models for which moduli dynamics are fully understood, and the set of string models that are of phenomenological interest, are completely disjoint. All of the latter, like the heterotic MSSM or the intersecting IIA braneworlds, still pose unsolved mathematical problems which prevent the calculation of standard-model masses and coupling constants.
Now let us suppose that with time, after sufficient mathematical progress, the models of phenomenological interest become fully tractable. This is when we can really talk about falsifying, not just specific models, but even whole classes of models, by *proving* that an empirically necessary conjunction of properties does not occur anywhere in the class.
When you say that “there’s no way to computationally ever do that” (for the whole of string theory), you must be thinking of arguments like Denef & Douglas’s, that the Bousso-Polchinski model of the cosmological constant would be NP-hard to test. But as Denef himself suggests, such complexity-theoretic results might be more like “no-go theorems”, telling you how *not* to proceed.
In computer science, two things which can speed up search are (1) structure in the search space (2) search for a special property, and string phenomenology possesses both of these. Looking for the cosmological constant a la Bousso-Polchinski – a finely-tuned near-cancellation of numerous fluxes – must be just about the least efficient way to falsify string theory. The qualitative features of the standard model already suffice to rule out the vast majority of string vacua, and subtle facts like Koide’s relation connecting the masses of the charged leptons will also be highly constraining.
When I say that string theory will supply the standard models of the future, I mean only that I expect some string vacua to pass all these tests. The standard model won’t be “string theory, in which we believe because it contains gravity”, it will be “string theory on background X, because it predicts the correct masses and coupling constants”.
Peter says: “As for claims about “the successes of string theory in producing the universal aspects of low energy physics”, that’s just ridiculous.”
Alright, I’ll bite. Here are some aspects of how string theory captures general features of low energy physics. While these will not impress you, I will keep an open mind that some of your readers are not as opinionated as you are.
All phases of string theory contain black holes. Whenever they are under computational control, the Bekenstein-Hawking entropy is reproducible microscopically. In the limit of small temperature of the black hole, even dynamical phenomena like Hawking decay rates and their greybody corrections can be produced from string computations. As Sen likes to emphasize, if there was a mismatch in *any* of the phases of string theory, then it would immediately have been game-over for string theory as a whole. Many spacetime singularities are naturally resolved in string theory. Chiral fermions, multiple generations, non-abelian gauge symmetry, hierarchies in Yukawa couplings etc. are all essentially automatic in any string compactification. Each of these things is a puzzle from a purely low energy perspective. Things like AdS/CFT and the qualitative matches with RHIC-like experiments is another generic test. Finally, all of these things are coming from a quantum theory that containes gravity – usually, attempts in this direction result in instantly-dead-at-birth theories. (Another one of the things that is lost on the layman is how remarkably gauge theories and gravity seem to mesh together to make AdS/CFT and string theory consistent, but this is more theoretical, so its not too relevant for this discussion.)
So yes, while you might argue that string theory hasn’t made precise quantitative predictions yet, generic aspects of string theory do capture low energy physics. Its not “just ridiculous”, sorry.
An over-riding theme in the things I mentioned above is that “whenever we understand string theory”, things seem to work. This is the reason why it is interesting for some of us to work on understanding the theory fully.
Somebody,
I see, the “successes of string theory” in low-energy physics that you are hyping have really nothing to do with the topic at hand: the failure of string theory unification. Black hole entropy tells us nothing about observable low-energy physics, and AdS/CFT is a completely irrelevant issue.
The one sentence you write that is relevant:
“Chiral fermions, multiple generations, non-abelian gauge symmetry, hierarchies in Yukawa couplings etc. are all essentially automatic in any string compactification.”
is nonsense. You can find string theory “vacua” violating any of those, and if low-energy physics were different, I’m sure you’d be telling us that something else was “essentially automatic”. For instance, in a non-chiral world, we’d probably be hearing a lot about how N=2 supersymmetry was crucial in string theory, so non-chiral couplings were “essentially automatic”.
It’s interesting that you don’t mention the most famous argument of this kind: the “string theory predicts low-energy supersymmetry” argument. I notice that over the past ten years as the Tevatron has seen no evidence of supersymmetry, and as the arrival of relevant LHC results approaches, string theorists have started to back away from this argument. My prediction is that over the next few years they’ll be running away from it, and claiming they never believed it. Unless of course, evidence for supersymmetry appears, in which case we’ll hear something very different…
Mitchell,
You’re creating for yourself a hypothetical world in which all the problems on which no progress has been made in 26 years (for very good reasons) have been solved, and then arguing on the basis of that. This is just wishful thinking, not science.
“Black hole entropy tells us nothing about observable low-energy physics,… ”
My low energy world contains gravity as well, Peter. And not just the standard model of particle physics. Since general relativity is part of low energy physics, we are forced to confront black holes. Its not an option.
About the rest, the claim was that string theory contains the ingredients of particle physics. That statement is true and non-trivial, notwithstanding your outbursts.
Somebody,
We’ve gone from:
“the successes of string theory in producing the universal aspects of low energy physics”
to
“the claim was that string theory contains the ingredients of particle physics”
which is rather different. Sure, string theory contains “the ingredients of particle physics”, along with 10^500 other things. The problem is that it has failed to tell us anything about these ingredients.
There is only one particle physics in the real world. So the word “universal” doesn’t make much sense in my original statement. I am pretty sure that you understood that what I meant was “generic”.
If you want to be cynical … 🙂
I think it should be pointed out, lest someone get the wrong idea from the above arguments, that it is in fact possible to find string vacua which are very close to the observed Standard Model. This includes all of the features of the Standard Model such as chiral fermions, multiple generations, hieararchical Yukawa couplings, etc. as Somebody has pointed out. This by itself is reason enough to believe that string theory is on the right track.
What has not been accomplished up to this point is to uniquely fix the parameters of these models in order to determine the actual values of gauge and Yukawa couplings. The study of the mechanisms for doing this has been a very active area of research over the last decade and much has been accomplished. I fully expect to see a string model constructed in the next 10 years or so where this can be done. As to why our universe contains the three-generation Standard Model rather than some other gauge symmetry or matter content is a question that will take some time to answer.
Somebody said:
“Chiral fermions, multiple generations, non-abelian gauge symmetry, hierarchies in Yukawa couplings etc. are all essentially automatic in any string compactification.”
Such a statement is absurd of course and I find it strange that you could say such thing since as it seems you know quite a lot about the subject. Maybe the above nice properties are present in the models you know but this is exactly because they have been engineered to produce them. In any case I don’t understand the point you are trying to make with similar statements.
In the current situation I think there are 5 options that someone could choose from regardless of any LHC findings:
1)Dismiss String theory as a framework for unification altogether.
2)Adopt the multiverse idea, eternal inflation, anthropic reasoning etc..
3)Dismiss the current concept of low energy effective vacua and wait until there is a better understanding of the theory hoping that this will eventually help us understand how our world emerges naturally from the theory.
4)Accept the validity of the current concept of low energy effective vacua and wait until a vacuum selection principle is found which will hopefully select our vacuum.
5)Do nothing of the above and continue to be skeptic.
David Berman said:
This blog is mainly about string theory as a theory of unification (of the standard model and gravitation), and that’s what my remark was about. If one counts the application of “stringy understanding” to QCD as a success of string theory is a matter of discretion, depending on one’s understanding of what “string theory” comprises. In the sense of this blog’s understanding of string theory as unification, it is not.
To pick another not too elementary example: One can use white noise calculus in order to give QFT an axiomatic foundation using the Osterwalder-Schrader axioms. One can also use white noise calculus to try to understand stochastic partial differential equations and turbulence in Navier-Stokes equations.
These are completely different physical theories, but of course one is free to claim that progress in QFT has led to a better understanding of turbulence in classical physics (and if you think both topics through and solve all problems, you could even win two millenia problem prices at once, think about that!).
Giotis, the general context in which my statement is to be understood is the original Calabi-Yau compactification papers of Candelas, Horowitz, Strominger and Witten. As I repeatedly emphasized, the question was not what is possible[1], but what seems to arise essentially “automatically”. This is admittedly a not very precise notion. For example, inherent in this claim there are some assumptions: like we demand four non-compact dimensions. But there exist large classes of string theory models where my statement is true, is the point. Its hard to define a measure on the space of theories, so it is not very meaningful to be more precise.
[1] There is a quote attributed to Coleman that string theory is not a theory of everything, but a theory of anything. (This is one of the things that bothers Peter and probably you.) This is in fact less true about string theory than quantum field theory, because UV completion puts constraints. The hope of the last decade (which seems completely unreasonable in hindsight) was that there will be a unique vacuum that will emerge. Instead it turns out that UV completion is constraining, but still we do have a large choice: in the industry-lingo, this is what distinguishes between the “swampland” and the “landscape”. But a theory can be predictive without being unique. In fact, quantum field theory does just that, i.e, standard model, a specific quantum field theory is predictive. So string theory might turn out to be predictive if we could only wield it well.
I don’t think your 5 bullet points cover all that is possible. Keep in mind that a theory doesn’t have to predict the values of the couplings to make it a predictive theory. The choice of the vacuum might be anthropic or accidental or initial condition or whatever you want to call it, but after a finite number of experiments things can be predictive. The standard model contains 19 undetermined couplings, but (ideally) after 19 experiments to calibrate those numbers, the theory is predictive. String theory could work the same way: the choice of vacuum would determine the couplings, and indirectly we will be determining the vacuum by doing the calibration runs. We simply don’t know if this situation is possible in string theory, which is why pronouncements irritate the practitioners. There are some subtleties here, like how much purely low energy experimenst will be able to fix the vacuum: I like the way Mitchell describes the situation in the last para of his post at 10:51.
Hope it was not too useless. I don’t have the time to write a longer message.
Somebody,
If there is no underlying principle which will point to a specific vacuum you could always change the vacuum to fit the experimental data by manipulating your model. That’s the whole point of the criticism.
Well, there is no underlying principle that uniquely picks out the standard model from the plethora of possible quantum field theories. But thats hardly a problem. As I said, uniqueness is not a necessary requirement for predictivity. You are thinking of computing the coupling constants, I am talking about fixing them experimentally. Please read my post again.
The point is whether you can fix a vacuum/model by doing a finite number of experiments. If you can, then all experiments after that are predictions. “Manipulating” a model as you call it, is what went into (for example) the construction of the standard model in the last century. One way to look at the question is whether this scenario can be realized in string theory. There is the possibility that string theory might provide the same kind of arena for UV complete model building that QFT provided for theories without gravity. Only time and hard work will tell. One trouble is that model-building in string theory as it stands is sketchy because we don’t understand the theory well enough, so we have no good organizing principles (like gauge symmetry in QFT).
Somebody,
The Calabi-Yau compactifications chosen for study in 1985 were chosen very specifically to have the right general characteristics to have some hope of giving the (supersymmetric) standard model at low energies. You can’t point to this as providing “generic” features of string theory at low energies.
It’s quite understandable that there was a flurry of excitement in 1985 when people started looking at these models. On the other hand, from the beginning there were also very obvious reasons to be skeptical. Not a single specific feature of the standard model was predicted (gauge groups, couplings, matter representations). As time went on, it soon became clear that there was so much freedom in how you chose your “string vacuum” that you could probably get just about anything. The only hard thing to understand is why 26 years later, people are still claiming this is a viable idea about unification.
As for the “string theory more predictive than qft” argument, it’s a waste of time to keep pointing out why that’s wrong. If someone wants to keep arguing that a theory that predicts nothing is more predictive than the most successful theory we have ever had, there’s not much point in going there…
Hi Somebody (and others with similar intentions …),
isn`t it a bit pointless for You wanting to explain and discuss things in this blog? Your efforts will never change the positions of anybody here…
You have certainly more important things to do 😉 …
Happy New Year
Hahaha! Its my holiday good deed. I like to think that I am talking to the lurkers who never post on this blog, but who might be more reasonable. Otherwise, it is easy to flip out and go on a shooting spree when faced with these relentless misrepresentations. 🙂
Cheers!
He he,
I sometimes silently lurk around 😉 …
Had to laugh about the joke linked to; most of these abstruse goose comics are fun.
I was just asking myself if the game they are playing here would lose it`s fascination if they could not evoke any response?
Anyway, enjoy Your holiday then 🙂
Cheers
A question, If you say that you aren’t an expert in string theory, what the hell are you criticising? You don’t know of what you are talking about. You aren’t a physicist, any longer. Maybe you’ve got in a Ph.D. in theoretical physics in a prestigious university long time ago, but now you don’t want to learn theoretical physics. why don’t you try to learn the hard and necessary math and then to try understanding string theory before to criticise the theory?
Because the only aim for you is to argue with the true physicists for laughing of them and your blog is only about sociology. Your book is for earning a lot of money with this dishonest behaviour. Because of this I hate you and I’m going to attack your blog. You are an enemy of science you are an enemy of the hard work.
Peter, probably you already know this. But Rutgers also archives their
HET seminars.
http://www.physics.rutgers.edu/het/video/het-video.html
there are two talks by Witten on this
Somebdoy says: (January 5, 2011 at 6:21 am)
“But a theory can be predictive without being unique. In fact, quantum field theory does just that, i.e, standard model, a specific quantum field theory is predictive.”
That’s an excellent argument. In fact, it works so well that it also applies to not-necessarily renormalizable QFTs. It then puts gravity-as-a-QFT on the table as a theory of quantum gravity. That plus the standard model does an exceptionally good job of accounting for all known physics.
Obviously unknown physics cannot be accounted for, since we do not know what it is. It does sting a little bit that unknown physics currently includes the very interesting questions of the fate of evaporating black holes, the mass of the Higgs field, and the possible additional fields and interactions relevant in the early universe. But that can only be helped by looking for the answers in nature, rather than on paper.
Suddenly the rationale for using string theory at all, instead of plain old QFT, has evaporated…
I think probably the best way to view perturbative string theory is that it is really an effective theory which is valid up to some scale M_{string} < M_{Planck}. So, in this sense it isn't a complete unification theory, at least at the perturbative level, as Peter likes to say. However, I do believe it does take us a little closer to unification than do gauge theories/QFT by themselves since now gravity can be brought into the picture. It just doesn't take us close enough to say, uniquely predict the Standard Model, if that's even possible. What is needed is to find the theory which is valid at the Planck-scale, and to which perturbative string theory is an approximation. This sometimes goes by the name of "non-perturbative formulation" or "M-Theory".
Khavkine says: “In fact, it works so well that it also applies to not-necessarily renormalizable QFTs. It then puts gravity-as-a-QFT on the table as a theory of quantum gravity… Suddenly the rationale for using string theory at all, instead of plain old QFT, has evaporated”
No, you will need an infinite number of experiments to fix all couplings if the theory is non-renormalizable. Thats not predictive, even in the best case scenario. UV completeness is key. Note that we are talking about all energies at once, not just low energies.
Here is one scenario: if we had all the control on string theory that we wanted, then we will be able to construct vacua which reproduce everything we see at low energies: standard model and black hole puzzles and cosmological constant and what not. The question that we don’t know the answer to is, (1) is there any vacuum that allows this? (2) is there more than one vacuum which allow this? If the answer to (1) is no, string theory is ruled out. If the answer to (2) is yes, then it will not be within the realm of science to distinguish between these degenerate (at low energy) vacua without doing high energy experiments.
My strong suspicion is that if we fully understood string theory, it will actually enable us to relate low energy and high energy in more interesting ways, so the last setup may not even be realized. But even if it is, we will still have a candidate for a (universality class of) complete descriptions of Nature.
For the prupose of full disclosure, I should add that the “string phenomenology” as it is practised today is not really in line with this philosophy. It tries to construct vacua using more short-sighted approaches, because of our inadequacy in understanding non-perturbative string theory. But since it seems quite likely that the answer to question (1) is “yes”, I think this might indeed be enough to settle that once and for all. But I suspect that answering (2) in any sort of useful way is going to require a real understanding of string theory, instead of the various bits and pieces that we currently refer to by that name.
Somebody wrote: “No, you will need an infinite number of experiments to fix all couplings if the theory is non-renormalizable. Thats not predictive, even in the best case scenario. UV completeness is key. Note that we are talking about all energies at once, not just low energies.”
Yes, thank you for pointing out the obvious difference between renormalizable and non-renormalizable theories. However, despite this difference, non-renormalizable theories are not any less predictive. In fact, once I pick any set of values for the infinite set of coupling constants needed to define my QFT, it is as predictive as desired and at all energies to boot. Whether this fixed theory is correct is, like any other hypothesis, to be tested by experiment. Do you still disagree?
Now, even if you agree with the above, you might take issue with exactly how the coupling constants were chosen. For the sake of argument, let me set all of them to zero, except the finitely many that have been directly probed by experiments to date; those I set to the measured values. I didn’t have to use zeros, I could have used the outcomes of an imagined infinite sequence of coin flips. The point is that there is no known preferred mechanism for making that choice. And even if string theory is a candidate for such a mechanism, it is not known to be preferred to, for example, the two methods I just made up on the spot.
The other comments you’ve made rely on the hypothesis that the fundamental model we need will be a string theory. Now that is the hypothesis that IMHO severely lacks evidence.
Khavkine: “In fact, once I pick any set of values for the infinite set of coupling constants needed to define my QFT, it is as predictive as desired and at all energies to boot.”
Hmm? How does one fix these coulings experimentally is the question. Couplings are fixed by EXPERIMENT, not by “picking” whatever value one wants. If you have a theory that had infinite number of couplings, one needs an infinite number of experiments to determine them, so the theory is not predictive. Simple as that. PS: Anomalies would rule out many theories, but this is even before that.
“Now, even if you agree with the above, you might take issue with exactly how the coupling constants were chosen. For the sake of argument, let me set all of them to zero, except the finitely many that have been directly probed by experiments to date; those I set to the measured values. I didn’t have to use zeros, I could have used the outcomes of an imagined infinite sequence of coin flips. The point is that there is no known preferred mechanism for making that choice.”
No, I don’t agree with the above. I agree that your setup is operationally okay at low energies, but is meaningless as a full theory. I expect that questions that we could ask at any energy should have reasonable answers. When all energies are allowed, your set up is *meaningless*.
Your claim is tantamount to arguing that the world is not predictive at a fundamental level. That was a valid argument at any stage of science when there was something we didn’t understand. But so far that argument has been wrong.
Dear Somebody, I do not disagree that you are arguing from conventional wisdom. However, conventional wisdom needs to be shaken up and re-examined once in a while. With that in mind, perhaps you can be a bit more specific with your counter arguments when it comes to predictivity and meaningfulness at arbitrary energies.
I described two specific (though possibly appearing convoluted to some eyes) QFT models. What is it that these models cannot predict? A specific example would be nice. And what kind of question gets a meaningless answer at arbitrarily high energies? I cannot think of any such examples myself. After all, a fully renormalized QFT model yields n-point functions that are finite where expected and defined for all energies and all n. Moreover, any quantitative question posed comes down to the knowledge of n-point functions. So I’m puzzled by where your examples could come from.
“Your claim is tantamount to arguing that the world is not predictive at a fundamental level. That was a valid argument at any stage of science when there was something we didn’t understand. But so far that argument has been wrong.”
I’m sorry, I really do not see where you drew that conclusion from. I’m also completely stumped by the second sentence above. There are certainly many things in science we do not understand. One does not even need to stretch to meaning of the word “understand” to extremes. The particle/field content and their interactions that were relevant in the early universe is an example of something we do not understand. Though, still, I’m not sure how these factoids pertain to the discussion.
Here is where your basic fallacy is (in my view):
“In fact, once I pick any set of values for the infinite set of coupling constants needed to define my QFT, it is as predictive as desired and at all energies to boot. Whether this fixed theory is correct is, like any other hypothesis, to be tested by experiment. Do you still disagree?”
One doesn’t fix the theory by devination and then check it against experient. The fixing itself done VIA experiment. And when you have infinite couplings there is no operational way to do it.
When you are truncating the theory to a finite number of operators the situation *qualitatively* changes. What I am saying is that the truncation can arise from either a UV complete theory or a non-renormalizable theory, but only one of them is really a theory (the UV complete one).
In any event, the situation regarding non-renormalizable theories being unpredictive is not something that I would refer to by the romantic term, “conventional wisdom”. I would say that it is an automatic consequence of what non-renormalizability means. So if you still don’t agree, maybe it is better that we agree to disagree.
Best wishes!
As usual, a great blog by Peter, giving space for fascinating comments from all sides. (I am speaking as a complete outsider to mathematical physics, and am grateful for this accessible window looking in). BTW: if Peter were more diplomatic and less provocative, I doubt the resulting discussions would be half as interesting!