Why I am REALLY disappointed about string theory

[suprised]

Well, just .. read...above!

 

[MTd2]

Are there any paper where perturbative finitiness of 11d SUGRA is proposed?

 

[suprised]

Are there any paper where perturbative finitiness of 11d SUGRA is proposed?

Well this has been a hot topic over the last couple of years, and hep-th has a lot of papers on that, see for example:

http://arxiv.org/pdf/1005.2703
http://arxiv.org/pdf/0808.1446
http://www.worldscinet.com/ijmpd/18/1814/S0218271809015874.html

Actually these deal with N=8 sugra in d=4, which is very closely related (by torus compactification).

 

[MTd2]

Actually these deal with N=8 sugra in d=4, which is very closely related (by torus compactification).

I try keep track of the papers about N=8 4d sugra and I don`t remember reading about this kind of conclusion. How can you be prove that one being renomalizable means that the other is also renormalizable? Type IIA string is renormalizable but I don`t see anyone saying that because of this 11d sugra is.

 

[tom.stoer]

Type IIA string is renormalizable but I don`t see anyone saying that because of this 11d sugra is.

How do you prove renormalizable of IIA string theory?

 

[Haelfix]

can you give us a hint why 11d SUGRA fails to be asymptotically safe? simple power counting isn't sufficient (as we now from 4d GR)
my sentence "What is its action, its path integral, its Hamiltonian? I think this consistent quantization is not known; otherwise the while program would have been succeeded - but as we know it's still work in progress. " to which yozu responded was not about SUGRA, but M-theory. You stated that 11d SUGRA is the classicsllimit of M-theory. In order to understand that you would have to write down quantized version the 11d M-theory and then explain its classical limit. How does this quantized version look like?

Its hard for me to disprove a negative (almost all nonrenormalizable field theories fail to be asymptotically safe, and only a few examples are known for sure eg the Wilson-Fischer fixed point etc). Nearly always, the existence of such points are highly nongeneric and subtle.

In this case, what people know is that there are extended objects in the nonperturbative spectrum of the supergravity theory (roughly acting like soliton states). We know what they are, and can write them down. Now b/c some of them act loosely like esoteric version of black holes, you instantly run into many of the subtleties that black holes entail including the information loss paradox. The consistency of the theory is very much threatened, and you have to find a way to unitarize it somehow.

Now how do I know that supergravity is the classical limit of Mtheory or has anything at all to do with it? Well, I'm going to cop out and direct you to a textbook b/c the whole chain of reasoning is lengthy and I won't be able to handwave my way through it. Suffice it to say, its not obvious at all and really does require some work.

Try chapter 10 of quantum fields and strings: a course for mathematicians, Volume 2. Its pretty explicit.

 

[suprised]

I try keep track of the papers about N=8 4d sugra and I don`t remember reading about this kind of conclusion. How can you be prove that one being renomalizable means that the other is also renormalizable?

A compactification concerns the IR properties and not the UV ones; by going up in energy (>> the scale of the compact space) the effect of the compactifciation becomes negligible. So the UV properties of the theories should be the same.

 

[mitchell porter]

Try chapter 10 of quantum fields and strings: a course for mathematicians, Volume 2. Its pretty explicit.

For people outside the USA who can't access this via http://books.google.com/books?id=TQIsyvw1KnsC" (including a whole course on gauge theory, delivered by Edward Witten).

Haelfix, do you mean lecture 10 by Eric D'Hoker? (D'Hoker's 10th lecture is listed as "Lecture 15" on that page, but if you open the file you'll see it's numbered as part "X".) It discusses the realization of susy in string theory and some properties of supergravity, but it doesn't talk about those extended objects.

 

[Haelfix]

Yes that's right, otoh I like that series of lectures b/c that's where I came across much of the material on the dimensional reduction of d=11 supergravity to d=10 for the first time and I happen to like the notation.

Try Ortin's book called Gravity and Strings for a full chapter on the extended objects as well as the entirety of above material but done with atrocious notation (you have been warned)..

 

[MTd2]

A compactification concerns the IR properties and not the UV ones; by going up in energy (>> the scale of the compact space) the effect of the compactifciation becomes negligible. So the UV properties of the theories should be the same.

But degrees of freedom are suppressed. The theory in 4d is not written with those in mind.

 

[MTd2]

In this case, what people know is that there are extended objects in the nonperturbative spectrum of the supergravity theory (roughly acting like soliton states). We know what they are, and can write them down.

This is not a barrier to asymptotic safety. Even pure GR has a 2d surface at the asymptotic limit. 11d SUGRA maybe would have those on its asymptotic limit.

 

[MTd2]

BTW, LM answered me this:
"
Dear Daniel,

nope, nope. The compactification obviously doesn't preserve finiteness of field theories. Higher-dimensional field theories of the same kind are always more divergent - because of the higher-dimensional momentum integration. The whole counting has to be redone.

The maximal d=4 SUGRA is most likely perturbatively finite - because of the KLT relations and other things. The 11D SUGRA is not finite.

Cheers
LM"

 

[tom.stoer]

So it's basically power counting.

 

[suprised]

Ups, you are right, my bad. I got carried away into sloppy thinking. The argument fails because of KK modes arising the compactification. Power counting alone gives only superficial estimates of divergences, because amplitudes can be better behaved due to symmetries (or other, "miraculous" cancellations).

In particular, I had even forgotten that there are arguments that the N=8 supergravity in d=4 cannot be obtained by _any_ decoupling limit of 10d strings, so may be in fact separate from string/M-theory and thus be part of the swampland. see http://www-spires.dur.ac.uk/cgi-bin/spiface/hep/www?eprint=arXiv:0704.0777

 

[tom.stoer]

I would like to come back to an idea I had a couple of days ago: why not drawing a much more detailed map of the theory space?

• It should list all theories
• It should provide a notion for the relevant degrees of freedom (open/closed strings, M2/M5 branes, matrices - which one?,...) for each theory
• It should provide a small set of fundamental equations for each theory (action, ...)
• It should list all relations (the self-duality is missing in this picture)
• Each relation should be qualified like "S" with the appropriate couplings, "low energy", "classical limit" - with an explanation what that means, ...
• It should distinguish between different theories; it should especially not list one central M-theory as we know that this is stll a hope; instead it should show M2/M5-, matrix-theory etc. ...
• it should list what prevents a certain theory from being the mother-theory

I guess there isn't such a map - otherwise it would have been posted here. Is there a chance to construct it here in this forum? e.g. as a simple table?

 

[atyy]

http://arxiv.org/abs/1007.4001

"As a consequence of the last property, we consider such QFTs to be definitions of models of quantum gravity, with fixed asymptotic background. The idea that AdS/CFT defines a duality between two independently defined theories, is probably without merit."

"String perturbation theory is a non-convergent asymptotic expansion."

 

[mitchell porter]

I would like to come back to an idea I had a couple of days ago: why not drawing a much more detailed map of the theory space? ...
I guess there isn't such a map - otherwise it would have been posted here. Is there a chance to construct it here in this forum? e.g. as a simple table?

It won't be simple because the dualities get more and more complicated in lower dimensions. A full map of theory space would be an atlas. And it would be great for someone (or a dozen someones) to create an atlas of string theory, but it would have to be done by people working in the field, and it would have to be updated every few years, like one of those Particle Data Group publications.

What we can do here is work just to understand the basic dualities in 9, 10, and 11 dimensions which connect all the theories. Very briefly, M-theory compactified on S^1 is IIA, M-theory on T^2 is IIB, M-theory on S^1/Z_2 is heterotic E8xE8, M-theory on T^2/Z_2 is heterotic SO(32). (Simplest possible diagram of this, http://www.sukidog.com/jpierre/strings/duality.htm" .)

Many of the details are in Ashoke Sen's http://arxiv.org/abs/hep-th/9802051" . There must be a more orderly exposition somewhere, at the level of detail that Tom wants, but I haven't seen it.

 

[Lt_Dax]

Here is my view. Feel free to squash it if there's reason to.

Isn't the main criticism of String Theory nothing at all to do with falsifiability, use of abstract mathematical trickery, or usefulness?

(It is clear that an unfalsifiable idea may become falsifiable at technology improves. Weird mathematics is not a sign of overstretching ourselves, as was pointed out with regards to hilbert spaces in QM, for example. And nobody doubts the usefulness of string theory and AdS/CFT in pure mathematics for example).

Isn't the primary criticism of string theory that it introduces features uncalled for? What do I mean, well we usually formulate a quantum field theory with 4 space time dimensions and point particles because that is exactly what we observe in nature. Forgive me for being "old-fashioned", but usually we develop a scientific theory by making observations first, then drawing conclusions. We chastise literalist creationists for doing things the opposite way around.

Of course, sometimes theory matures so fast that it gets a little ahead of experiment. String theorists often cite Dirac's theory and the prediction of the positron as proof that it's acceptable to jump the gun. I disagree. While it is clear that Dirac was inspired and a genius to bring special relativity together with quantum mechanics, he didn't introduce anything exotic. No strings, bells, whistles, extra dimensions or any other peculiar entities. His theory assumes point particles and four dimensions, since there was no experimental evidence that anything more than this exists (and there still isn't, as far as I know).

It's also tempting to claim that Planck introduced, seemingly as a random fudge, the quantum in order to explain the unexplainable. "See, there you have it, that's what string theorists do!" they claim. Except that Planck did what he did because there was no other way to explain the data. He knew it felt like a fudge and he felt bad doing it, but it worked well because there was no alternative, and experiment quickly supported the idea (in fact, the experimental proof for the existence of the photon already existed, it was just languishing without explanation).

Then there is the claim that Einstein developed special and general relativity "just by thinking about it", divorced from experiment. String theorists seem to use this idea (which is highly questionable anyway) to validate the claim that they're actually "discovering" new physics all the time. Ed Witten is particularly fond of making this claim, suggesting that by writing equations down we are actually discovering things. But you don't discover things with theory, you explain things with theory. You discover things with experiment. Surely experiment (observation) is the arbiter of truth.

In any case the comparison with Einstein's approach is flawed. Special relativity follows from the requirement that electromagnetism produces invariant results in different reference frames coupled with the constancy of the speed of light, an experimental observation. General relativity follows from these things but obviously not straightforwardly otherwise it wouldn't have taken him another ten years, but general relativity is as devoid of exotic concepts as it can be. Curved space time is a necessity, and was quickly confirmed experimentally, just as Planck's hypothesis was.

Another claim I've noticed that string theorists make is that it does in fact originate from experiment, because it was proposed to explain the strong force (another attempt to parallel the introduction of the quantum, for example). However, this hypothesis was rejected in favour of QCD which (unsurprisingly in my view) in based on point particles and four dimensions, nothing uncalled for and nothing unobserved. So what is the continuing motivation for studying string theory as a candidate for explaining nature (as opposed to just mathematical interest)?

It may finally be tempting to say that unification is the motivation, but although it looks compelling, there is not yet any proof that all the forces are unified. Making the coupling strengths match at the appropriate energy scale required altering the standard model so I remain unconvinced. Physics may or may not be unified. It is my firm belief that experiment will tell us whether or not all of the forces are unified together, and how they are (we don't even have a grand unified theory minus gravity yet!).

Isn't it also clear that Einstein, if we wish to follow his career with interest (which string theorists understandably like to), was far more successful in his early years. What distinguished his early career? In his younger years, he studied experimental results, unsolved problems and he reaped large results from listening to nature. In his later life, he tried to dictate to nature, and he made little progress. The rest of the physics community viewed him as an old timer who'd lost touch. And yet the string theorist community choose to imitate the later half of his career. The fact that Einstein tried to find a unified theory seems to provide some sort of validation for thousands of physicists to engage in a global effort to do so, when in my view his failure actually teaches the opposite lesson.

Isn't this the key, then: experiment is king. Theory is (or should be) subservient, in my view. The predictability of theory is surely meaningless unless the theory has its origins in observation in the first place; an experiment sandwich with theory in the middle. Why do I say that? Because without some form of experimental constraint both in the formulation and confirmation of a theory, there are literally an infinite number of ways to solve the same problem (isn't that the fundamental reason why there are so many string vacua?) There are probably an infinite number of (unconstrained) uses for a paper clip.

In my view, because I view the origin of a theory as just as crucial as its later validation, string theory doesn't even fit the usual definition of an empirical theory, which originates from experimental observations, by definition. The standard model fits this description well, and contains within its parts some of the most spectacularly verified predictions ever produced. Isn't it rather unfortunate that the Standard "Model" is a true theory, but String "Theory" is actually just a model? (Isn't that also a source of enormous confusion when discussing it?).

It's worth noting that although it's acceptable for a (real) theory to get ahead of experiment for a little while, this uncomfortable no-man's land does not typically last long. Never has it lasted for 30 years! Dirac's theory would have been discarded long before 30 years had passed. Before anyone claims that our technology has become so limited that we've reached an impasse, consider that there is currently an enormous list of unsolved problems in physics that are far more deserving of our attention. If our technology (ability to make experimental observations) is so limited/limiting, then how did it produce such an enormous list of unsolved observations?

Far fewer unsolved problems led to two huge revisions in our understanding of physics in the early 20th century, so we could probably learn a lot about the universe just by taking the approach of the young Einstein, to sit down and try to explain them by developing or revising a theory. We may even learn some things about high energy physics. If string theory is pure mathematics then put it in the mathematics department, surely. This is not a trivial issue; it actually takes up lots of PhDs and funding which the funding councils assume is used to advance physics, not mathematics.

Am I old-fashioned, behind the times and hopelessly naive? I'm not being belligerent, I genuinely want to know why I have the view I do. Did the entire physics community decide in the 1980s that we could change the way we discover new science, and I haven't realized the new "sophisticated" way of doing things? Or is the claim of sophistication and beauty just a cover to play in a sand box?

 

[mitchell porter]

Reality has fermions, gauge fields, and gravity. String theory has all those things. It also offers a framework in which masses, coupling constants, and symmetry groups have a deeper explanation. It would be crazy to ignore it.

 

[tom.stoer]

We had all these issues during our discussion; so I would say that everything has been discussed carefully, but I understand that some arguments are hidden in 400 posts so it is clear that you come up with some of them again.

I'll try to answer rather quickly and gope you will study the other posts, too.

The problem of falsifiability (in practice, not in principle - we discussed this difference) is not specific for string theory but applies to all theories including quantum gravity. Therefore either you accept this paradigm shift (that your guidelines are more mathematical then experimental) or you have to stop doing physics at all.

Yes, it's true that string theory introduces more new and abstract theoretical concepts then every other theory we had so far (but that was true even for quantum mechanics). The difference is again testability, not additional assumptions. Once proved experimentally nobody will care about.

A theory does not necessarily originate from experimental observations only. This was a famous discussion between Einstein and Heisenberg.I thin with QM it became clear that the basic entities are not necessarily directly accessible via experiment.

Regarding the 30 years: the Higgs mechanism has been discussed and used in the sixties to explain gauge invariant mass generation. Still the Higgs remains unobserved and is probably only a low energy effective degree of freedom, i.e. the GSW theory may fail to me UV complete.

Last sentence regarding what string theory "is". String theory seems to be more a framework than a single theory which allows one to consider all theories including quantum gravity consistently; the main difference to gauge theories (which do the same for theories w/o gravity) is that in string theory the different theories are solutions (SU(N) + SUSY + ... theories emerge as solutions) instead of isolated theories.

I agree with most of your skepticism (this was the reason why I started this thread) but I have to admit that there are rather convincing arguments that string theory is a unique framework which has a lot of potential - I haven't seen any other approach which has the same richness.

 

[Lt_Dax]

Thanks for your views Tom. You're right it is difficult to find these points in 400 post, so I especially appreciate your response all the more. I've been trying to develop a more mature viewpoint, so this thread has been useful to read (if not in its entirety).

 

[marcus]

...
The problem of falsifiability (in practice, not in principle - we discussed this difference) is not specific for string theory but applies to all theories including quantum gravity. Therefore either you accept this paradigm shift (that your guidelines are more mathematical then experimental) or you have to stop doing physics at all...

... I've been trying to develop a more mature viewpoint, so this thread has been useful to read (if not in its entirety).

Tom, I don't think this is as fairly balanced as many of your posts, so I will suggest a different viewpoint on testing. Loop has become a coherent theory of quantum gravity (QG) which stands to be falsified by observation of the cosmic microwave background (CMB) by proposed next generation missions such as B-Pol.

There are certainly details of the combined theory (canonical+path integral+cosmology) that still have to be worked out. But there has been a remarkable convergence and the prediction of a cosmic bounce is robust. Repeatedly, under varying assumptions, Loop applied to cosmology replaces the singularity with a bounce and a natural inflation episode.
This has consequences for practical B-mode polarization maps of the CMB. The B-Pol mission has been proposed for the 2015-2025 timeframe. If it gets funded (a big "if") it could effectively falsify Loop.

So that QG theory predicts new phenomena and bets its life on the prediction. This is the customary behavior of scientific theories that we expect since Bacon set out the empirical philosophy 400 years ago.

It's not granted yet that we discard the Baconian paradigm or "stop doing physics".

======================

A Loop-related conjecture by Smolin that had stood since 1993 was falsified this year by the discovery of a neutron star with mass > 1.9 solar. Smolin had conjectured that the parameters of the standard particle model were at a local optimum for black hole production and had derived from this the prediction that no neutron star would be seen with mass > 1.6 solar. This conjecture was not a prediction of QG theory, but it assumed that classic black hole singularities were replaced by bounces. That is still possible, but the optimality conjecture has been falsified.

======================

QG theory itself, and QG-related conjectures, can be tested.

=======================

A relevant philosophical point is that fundamental physical theory is inferential. As Bohr told us, physics is not about how nature "is". It is about how it responds to measurement. What we can detect, measure, infer, and check by further measurement.

As Newton said "Hypotheses non fingo". We do not pretend that little wiggling strings exist. Or that spin networks exist--with nodes of quantum volume and links of quantum area--flickering into and out of existence. Or that little triangles of quantum triangulation exist. So theory-testing does not require a big "magnifying glass" to see these human imaginings.
There is no fundamental ontology.

These imagined things represent nothing but ways of calculating and relating measurements to other measurements.

Theories can be roughly graded according to how fundamental since from a deeper theory one can sometimes derive a more coarse-grain theory.

So in order to test comparatively fundamental theories such as Loop (I think everybody here realizes) instead of building an impossibly powerful "magnifying glass" what one needs to do is use the theory to predict new phenomena (which distinguish one theory from another) and look for the phenomena.

A theory must be predictive about the universe we actually live in, including the ancient light which we observe, or it is useless. And so it must be possible to test.

Apart from that, it can be as mathematical as anyone could wish there is no contradiction between empirical testing and mathematical elegance as guides.

That's my contribution to help balance the view you expressed. Hope it does not repeat too much from the previous 400 posts of this great thread!

===some references===
B-Pol mission proposal:
http://www.b-pol.org/index.php
Bee Hossenfelder's post:
http://www.math.columbia.edu/~woit/wordpress/?p=3262&cpage=1#comment-67988
Recent paper by Julien Grain et al:
http://arxiv.org/abs/1011.1811
Observing the Big Bounce with Tensor Modes in the Cosmic Microwave Background: Phenomenology and Fundamental LQC Parameters
Julien Grain, A. Barrau, T. Cailleteau, J. Mielczarek
12 pages, 5 figures
(Submitted on 8 Nov 2010)
“Cosmological models where the standard Big Bang is replaced by a bounce have been studied for decades. The situation has however dramatically changed in the last years for two reasons. First, because new ways to probe the early Universe have emerged, in particular thanks to the Cosmic Microwave Background (CMB). Second, because some well grounded theories — especially Loop Quantum Cosmology — unambiguously predict a bounce, at least for homogeneous models. In this article, we investigate into the details the phenomenological parameters that could be constrained or measured by next-generation B-mode CMB experiments. We point out that an important observational window could be opened. We then show that those constraints can be converted into very meaningful limits on the fundamental Loop Quantum Cosmology (LQC) parameters. This establishes the early universe as an invaluable quantum gravity laboratory.”

 

[atyy]

LQC is not a theory of QG until it can be connected with LQG.

When it is connected with LQG in its present form, it will make no predictions that can falsify LQG, because (i) canonical LQG is not defined - no hamiltonian constraint (ii) Rovellian LQG in its latest form is probably divergent.

 

[marcus]

Gravity/gaugefield/Higgs unification bid by Lisi-Smolin-Speziale

LQC is not a theory of QG until it can be connected with LQG..

Connected.

http://arxiv.org/abs/1010.0502
Local spinfoam expansion in loop quantum cosmology

http://arxiv.org/abs/1010.1258
Big Bounce in Dipole Cosmology

Your objection about Hamiltonian does not make sense to me, since the current leading formulation of LQG does not employ a Hamiltonian.
We will just have to see if "Rovellian LQG", the current formulation of the theory, is divergent--and how that plays out. Are you making a prediction?

 

[marcus]

This thread is about string theory---it would be nice to keep it that way.

I only want to point out that the customary apology for string: that other programs (in QG) have equivalent problems does not hold water.

It should not be made. And it is not a very good apology in the first place.

There is no need to "change the paradigm" of empirical science simply to give one elderly overgrown late-20th century research program a break.

One doesn't have to talk about Loop to make that point, either. You can for example ask Matilde Marcolli, the hot noncommutative babe at Caltech (this is a joke, I am not being a male chauvenist or something, she is a first rate mathematician: Oberwolfach grade.)

Anyway it is a general point. We do not change the paradigm of empirical science for light causes.

 

[Careful]

Just out of curiosity, I thought these non-commutative approaches were not quantized yet. So, it wouldn't be fair to compare it with something like string theory...

This thread is about string theory---it would be nice to keep it that way.

I only want to point out that the customary apology for string: that other programs (in QG) have equivalent problems does not hold water.

It should not be made. And it is not a very good apology in the first place.

There is no need to "change the paradigm" of empirical science simply to give one elderly overgrown late-20th century research program a break.

One doesn't have to talk about Loop to make that point, either. You can for example ask Matilda Marcolli, the hot noncommutative babe at Caltech (this is a joke, I am not being a male chauvenist or something, she is a first rate mathematician: Oberwolfach grade.)

Anyway it is a general point. We do not change the paradigm of empirical science for light causes.

 

[marcus]

Just out of curiosity, I thought these non-commutative approaches were not quantized yet. So, it wouldn't be fair to compare it with something like string theory...

Careful! Nice to see you after a 2 year absence! I saw you thanked in the acknowledgments section of a QG paper recently for helpful discussions with the authors---glad to see that indication of your continued interest and activity in QG.
I forget what type it was, Noncommutative, GFT, AsymSafe, maybe even spinfoams--it's gone out of my head.

Anyway I was not talking merely about specific rivals for some specific research goal but about the bid to abandon the empirical science paradigm. If you asked Marcolli I don't think she would say there is any pressing need to do that.

She recently organized a 2010 workshop at Oberwolfach on "Spinfoam+Noncommutative Geometry". Krajewski participated. You may know him. Indeed you may know more about that workshop than I do.

And Marcolli recently posted a "Spinfoam+NCG" paper.
http://arxiv.org/abs/1005.1057

So there is QG stuff brewing with NCG. Also NC field theory is considered one of the new "paths" to quantum gravity. There was a conference in Denmark on that exact theme funded by Barrett's QGQG branch of the ESF (Euro. Sci. Foundn.) I forget who organized it. Richard Nest probably.

Nest has collaborated on some "LQG+NCG papers". It is not a fully formed research gambit but some kind of initiative is taking shape. Just another reason not to be scared into abandoning the 400-year old Baconian tradition.

I'll get some links to introduce people to Marcolli, Krajewski, Nest etc, in case they are unfamiliar.

 

[marcus]

Here Matilde's Caltech page
http://www.its.caltech.edu/~matilde/
with a nice picture of her and a description of her research interests.
Here's a sample paper:
http://arxiv.org/abs/1005.1057
Spin Foams and Noncommutative Geometry
Domenic Denicola (Caltech), Matilde Marcolli (Caltech), Ahmad Zainy al-Yasry (ICTP)
48 pages, 30 figures
(Submitted on 6 May 2010)
"We extend the formalism of embedded spin networks and spin foams ... We end by sketching a possible approach to combining the spin network and spin foam formalism with matter within the framework of spectral triples in noncommutative geometry."

Here's some pictures from the 2010 Oberwolfach workshop she co-organized.
http://owpdb.mfo.de/show_workshop?id=783
There is a shot of Nest and Krajewski there, just so folks here can attach faces to the names. My intuitive feeling is that what these people are doing is potentially significant because NCG has already in a sense reproduced the Standard Model of particle physics at least in a rudimentary way, and spinfoam LQG has a pretty good model of evolving quantum geometry.

The point is all these people's theories can make predictions testable with available technical means--given the funding.
Spinfoam has been merged with Loop Cosmology (and appears to predict a bounce) so observational astrophysics is available to test it.
NCG is capable of generating predictions about particle masses, respectable despite some initial reckless shooting from the hip , and NCG is on track to supply matter to Spinfoams. Another possible merger.

This is just in answer to Careful's question. The main thing is this should be a string discussion thread, and I'm generally not interested in participating (I've kept out almost totally until now). But the point should be made that the apologetic argument that other approaches have equivalent problems, and that we need to relax empirical standards, should not be made.

String has to stand scrutiny on its own in the usual light.

 

[arivero]

It won't be simple because the dualities get more and more complicated in lower dimensions. A full map of theory space would be an atlas. And it would be great for someone (or a dozen someones) to create an atlas of string theory, but it would have to be done by people working in the field, and it would have to be updated every few years, like one of those Particle Data Group publications.

What we can do here is work just to understand the basic dualities in 9, 10, and 11 dimensions which connect all the theories. Very briefly, M-theory compactified on S^1 is IIA, M-theory on T^2 is IIB, M-theory on S^1/Z_2 is heterotic E8xE8, M-theory on T^2/Z_2 is heterotic SO(32). (Simplest possible diagram of this, http://www.sukidog.com/jpierre/strings/duality.htm" .)

Many of the details are in Ashoke Sen's http://arxiv.org/abs/hep-th/9802051" . There must be a more orderly exposition somewhere, at the level of detail that Tom wants, but I haven't seen it.

Lets assume that such atlas exists, very much as a new version of Slansky's group atlas for unified theories. In Slansky, if a given group happens to be the GUT group, it will be easy to give meaning to the subgroup descendants of such theory, and probably also to the nonexistence of bigger groups above the GUT one; it is a pretty trivial relationship, a directed tree or network in any case.

But with the "string atlas", we have the network of dualities, going up and down and around. If we locate a specific theory (and remember I have my guess ) as the real one, what are we supposed to do of all the network of dualities relating such theories with a lot of other ones. Should all the theories of the network be observed? Or perhaps should we look for a theory with a minimal quantity of dualities?

 

[atyy]

The main thing is this should be a string discussion thread, and I'm generally not interested in participating (I've kept out almost totally until now). But the point should be made that the apologetic argument that other approaches have equivalent problems, and that we need to relax empirical standards, should not be made.

Other approaches have similar problems, no question about that. No one is saying that physics is not about experiments any more. Only that the lack of immediate testability should not prevent one from exploring. In this spirit "But let us finish by a quote from Erwin Schroedinger, ... , ' ... or else, one might seriously worry that just where we forbid further questions there could be still quite a bit worth knowing about.'" http://arxiv.org/abs/1005.2471