Why I am REALLY disappointed about string theory

[suprised]

About counting black hole states, I still consider that to be somewhat semi-classical and speculative as it's "results" arrived at my extrapolating things from different domains to some QG domain. ST has made some success there, to connect to semiclassical approximation results but I don't know how the entire notion and view of entropy, and states will be once we have a proper theory. Most treatments of that, make extrapolations of things into doubtful context - where we in fact lack experimental confirmation of methods.

The whole point of the extrapolations done in these computations is that they are exact in the full quantum theory. This goes far beyond semiclassical analysis!

One thing is clear that string theory works extremely well and makes sense as a _physical model_. It is based on physical principles and produces, for example, scattering amplitudes that can be measured in principle, so it is _not_ just a mathematical game. Whether it actually describes nature is a different question.

I would view it analogous to "gauge theory". The detailed study of the various incarnations of gauge theories (susy eg) is certainly a subject in both physics and mathematics, and has provided many important insights into the real world gauge theories like QCD.
"String theory" is simply the natural generalization of gauge theory when including gravity; and as said before, it is intimately tied to it due to dualities. One should consider both, gauge and string theory, as part of one package.

 

[Fra]

In despite of my critique, I want to note that I appreciate the discussion and this thread. We all want the same thing in the end.

The whole point of the extrapolations done in these computations is that they are exact in the full quantum theory. This goes far beyond semiclassical analysis!

Just to make sure I understand - What do you mean by "full quantum theory"?

Correct me if I am wrong, but part of ST reasoning or premises, contains the premise that the mathematical formalism of QM as the framework for a measurement theory is correct, and need no revision. Ie. unitarity, timless hilbert spaces, linear operators etc as the correct abstraction for measurements. Right?

This is part of the "extrapolation" I refer to, and to which I object. But this isn't unique to string theory, it's indeed inherited from normal QFT. But then again, we don't know to what extent QFT makes sense beyond the experimentally verified domains. Ie. where the extrapolation is valid.

When it comes to things like very massive or extended systems (cosmological scale), we do not have a clean exprimental verification of wether the structure QM holds. So extrapolating things we know are right at laboratory scale, to QG scale is I think at least speculative. I personally am of the opinion that some of the rigid framework of QM can't make sense in the "full theory". Instead current formalism may be a special case where the observer is non-dynamical, and it studies a small subsystem (sufficiently small so that it does not deform the overall environment). This is why I think that extrapolating this to the general case is to formulate myself diplomatic at least not obvious! There is no experimental confirmation of the QM structure in this domain.

One thing is clear that string theory works extremely well and makes sense as a _physical model_. It is based on physical principles and produces, for example, scattering amplitudes that can be measured in principle, so it is _not_ just a mathematical game. Whether it actually describes nature is a different question.

Don't you rather have a whole set of possible predictons? ie. that for each (consistent) choice of background (ie observer in my assoication) you get different predictions?

Here I like to add one thing: I previously "defended" this possibility and I would like to have a balanced discussion. So from my perspective I don't object to the fact that there is no manifest B/I (I see reasons for why this is not possible), I object to the fact that you've constructed a set of possibilities without navigation scheme.

/Fredrik

 

[suprised]

Just to make sure I understand - What do you mean by "full quantum theory"?

The extact quantum theory including all perturbative and non-perturbative corrections to the classical one.

Correct me if I am wrong, but part of ST reasoning or premises, contains the premise that the mathematical formalism of QM as the framework for a measurement theory is correct, and need no revision. Ie. unitarity, timless hilbert spaces, linear operators etc as the correct abstraction for measurements. Right?

Of course. If you like to challenge the basic axioms of quantum theory.. uh oh...good luck. Well unless there is a good reason for doing this, we rather prefer to go on with something constructive, and get somewhere with that.

There is no experimental confirmation of the QM structure in this domain.

Indeed very much so. And what shell we conclude from this - stop research?

Don't you rather have a whole set of possible predictons? ie. that for each (consistent) choice of background (ie observer in my assoication) you get different predictions?

I don't see what you mean. It is rather the other way around - "different" backgrounds can give rise to the same predictions. That is what underlies duality.

, I object to the fact that you've constructed a set of possibilities without navigation scheme.

I am not sure to understand what you mean - you object to, or agree with, the statement that string theory constructs a set of possibilities without navigation scheme?

I myself do agree with it.

 

[Fra]

If you like to challenge the basic axioms of quantum theory.. uh oh...good luck. Well unless there is a good reason for doing this, we rather prefer to go on with something constructive, and get somewhere with that.

Thanks, I'll probably need it. Of course I do see good reasons for it, otherwise I wouldn't suggest it. The details are separate discussion.

Indeed very much so. And what shell we conclude from this - stop research?

No, of course not. My response to this will get me into WHY I think QM structure needs to change. It may be an interesting discussion, where I've argued in some other threads already, so I won't comment further in here.

I don't see what you mean. It is rather the other way around - "different" backgrounds can give rise to the same predictions. That is what underlies duality.

Again please correct me if I'm wrong (I'm certainly no string expert) but there are dualities that connects SOME choices of background. This is of course a good thing, I agree!

But the full claim that there is some kind of picture, where EVERY background is equivalent (predictionwise) to EVERY othre background is I think not true because then there would be not landscape problem? That full equivalence is probably more like the string theorist dream or "vision"? But we aren't yet there, right?

I am not sure to understand what you mean - you object to, or agree with, the statement that string theory constructs a set of possibilities without navigation scheme?

Both. I agree that is the way You do it. But I object to that it makes sense.

Here I think my objection differs to the usual critique against the lack of manifest background independence. I hold the opinon, that any measurement, and any predction needs a context to be formulated. A physical prediction that lacks any reference, is to me not possible.

So I do not per see object to the fact that you need a background, to reference a prediction. It's not more strange han the fact that you need an observer/measurement device (a real observer) to make a (real) observation or measurement.

What I find strange, and which I think is a symptom of the way ST is constructed, is that fact that you have "deducted" a gigantic set of possibilities, to the point where the progress stalls because there seems to be no competitive way to sort or scan this set in a rational way. I think this should not happen.

I guess from your response above that your idea is that the landscape is really somehow a gauge, and that you will eventually see that the big set of options is rather just the one and the same, so that you do get unique predictions?

I personally find that idea doubtful, but I understand the idea.

Another idea is that the landscape is "somehow" real, BUT that it's not really that large, AND that maybe there is a way to navigate in it. I just find it strange that if the construction of the landscape was "physical" then there should have been some kind of probability measure on it, where a given observer would then drift in the landscape like an evolution. But I'm not aware of anything in this direction from ST. There are these people who refer to anthropic principles, but that is more like an excuse, it does not solve the real problem of the making a choice, or explain the paradox of how it can be rational to generate such a landscap without navigation.

/Fredrik

 

[suprised]

But the full claim that there is some kind of picture, where EVERY background is equivalent (predictionwise) to EVERY othre background is I think not true because then there would be not landscape problem? That full equivalence is probably more like the string theorist dream or "vision"? But we aren't yet there, right?

Not at all and no one wants to be there! It would imply to saying eg that all the solutions of the Maxwell equations and GR would be equivalent and give the same physical answer!

Here I think my objection differs to the usual critique against the lack of manifest background independence. I hold the opinon, that any measurement, and any predction needs a context to be formulated. A physical prediction that lacks any reference, is to me not possible.

So I do not per see object to the fact that you need a background, to reference a prediction. It's not more strange han the fact that you need an observer/measurement device (a real observer) to make a (real) observation or measurement.

True… we do live (approx) in Minkowski space and that’s a flat backround of GR
and certainly in order to make measurements in our world we’d like to have a space-time around...

What I find strange, and which I think is a symptom of the way ST is constructed, is that fact that you have "deducted" a gigantic set of possibilities, to the point where the progress stalls because there seems to be no competitive way to sort or scan this set in a rational way. I think this should not happen.

How would you like to sort eg, the solutions of the Einstein equations of GR?
That’d be an “easy” sub-part of the problem.

I guess from your response above that your idea is that the landscape is really somehow a gauge, and that you will eventually see that the big set of options is rather just the one and the same, so that you do get unique predictions?

No no.. unique predictions are not at all wanted.

I sense some confusion about landscape and background independence. I don’t have the time right now, but may like expand on this a bit later. In the mean time, I recommend: http://arxiv.org/pdf/0807.3249v3 for some food for thoughts.

 

[Fra]

Not at all and no one wants to be there! It would imply to saying eg that all the solutions of the Maxwell equations and GR would be equivalent and give the same physical answer!

Mmm, that wasn't what I meant. I guess by equivalence I don't mean that all backgrounds give the same "numbers", I mean that they are different but "consistent", just like different SR observer get different numbers, but they are related in a "consistent way" - but the symmetry transformations.

Anyway, this was just ONE possibility I saw (I wouldn't share it). But I'll get to it more later... I'll try to skim the paper you suggested and see if it answers any questions about the String logic. Thanks for the reference!

How would you like to sort eg, the solutions of the Einstein equations of GR? That’d be an “easy” sub-part of the problem.

It's not that easy as Einsteins equation and GR as it stands (ie classical GR) is simply not cast in an appropriate form. They way I envision it, the corresponding information contained in "Einsteins equation" should be coded in an observer view. Anyway, all possibili solutions - in a given perspective - should come with a "probability" and be countable. I would expect gravity and Einsteins equation to be an emergent entropic phenomomen in this reconsturction. GR is not a starting point for me. But this this is as much a "vision" as string theory. I certainly have no solution on the table. But each will have to judge the plausability of the approach.

/Fredrik

 

[tom.stoer]

I understand what you are saying and I am with you in the sense that string theory has a lot of physical guiding principles - except for missing hard experimental facts. The latter one is due to the fact that string theory tries to reproduce the SM at reachable low energies and also tries to produce new results at unreachable higher energies. But somehow this seems to be natural for any approach towards a completion of SM + gravity and does apply in some sense it to LQG, NCG, ... as well.

Unfortunately the picture is a little bit more complex than that: there are two different experimental facts:
a) know facts string theory is not able to post-dict
b) new (and possibly experimentally inaccessable) facts string theory is not able to predict (again this applies to LQG, NCG, ... as well).

So there is a paradigm shift we see in modern theoretical physics: stronger focus on mathematics, less stress on phenomenology. This is a rather strange situation, but I see no way out.

Anyway - what I am expecting from string theory (seen as a framework) is more than just something like the framework of "gauge theory + SUSY". There should be some kind of uniqueness, some hint like "the swampland", "solutions / vacua of the landscape instead of theories", something like a definition of string theory.

Having a non-unique framework w/o any predictability in practice (not in principle) does not lead anywhere. So string theory either must be able make physical predictions which are unique to string theory (= which are not just something that could be derived from xMSSM w/o strings) and which can be tested in practice - or string theory must be able to go one step further than just providing a collection of losely coupled frameworks which contain both gauge symmetry and gravity.

As I do not expect a unique new low-energy prediction (post-diction would already be nice) from string theory, I guess it's more the conceptual side that should be stressed.

Coming back to my example regarding the early days of quantum mechanics. I still think that the situations are comparable, but on a different level. String theory is a broader framework addressing different issues, but we already know a lot of magic numbers and structures to be addressed within this framework: U(1)*SU(2)*SU(3), 6 flavors / 3 generations, Higgs particle (?) Weinberg angle, Yukawa couplings, 4 dimensions, ...

It would be phantastic if string theory could produce a results like
- a list or category of allowed low-energy theories (*)
- one theory very closed to the SM
- a list or category of forbidden low-energy theories (~ selection rules)
- the statement that the theories (*) are not just frameworks or "theories" but "solutions"

We have quite a lot of these results in our hands - just like Einstein, Planck, Bohr etc. had the photoelectric effect, black body radiation, the hydrogen spectrum, selection rules, half-integer spin (has been introduced by Pauli on purely phenomenological reasons and has been critizised by Sommerfeld :-)

Therefore the next logical steps seem to be
- understand more details (e.g. come even closer to the SM, proof of perturbative finiteness, ...)
- make predictions which can be addressed experimentally in principle or even in practice
- provide a unique definition of string theory = a guiding principle, set of axioms or rules

In QM we had exactly that:
- spectra, fine structure, Zeeman and Stark effect, ...
- formulations developed by Heisenberg, Schrödinger, Dirac and last but not least Feynman

In string theory there is a lot of progress regarding model building: D-branes, F-theory, xMSSM-like theories. There is also some progress regarding new predictions (unfortunately not testable in practice afaik). And there are even different concepts which one could compare with the Schrödinger and the Heisenberg picture, Feynman's formulation etc. - of course on a different level. What is missing is a guiding principle, a sound proof of dualities, an axiomatic derivation of the different concepts, formulations, sectors, ... This is exactly what string theorists promised to be provided by M-theory!

So I end up with the same conclusions as David Gross:

"What is the fundamental formulation of string theory? or: What is string theory?
This is a strange question since we clearly know what string theory is to the extent that we can construct the theory and calculate some of its properties. However our construction of the theory has proceeded in an ad hoc fashion, often producing, for apparently mysterious reasons, structures that appear miraculous. It is evident that we are far from fully understanding the deep symmetries and physical principles that must underlie these theories."

 

[Haelfix]

Tom, everything you say is more or less correct. But I mean at one point you do have to throw your hands up in the air.

Quantum gravity is something like 15 orders of magnitude away from anything we might be able to test. Making some sort of testable and practical prediction, is a running issue the entire field faces, and everyone is accutely aware of it. Even if you could figure out some observable that was possible to measure, you would likely end up facing a maxwell demon type situation. Eg the scale difference is so huge, that invariably any practical approximation you might make at the Planck scale, gets amplified and distorted many times over.

So absent a theoretical breakthrough (eg a vacuum selection method), it seems like the only hope is to get lucky with astrophysics and cosmology. For instance the observation of a cosmic string (in the stringy sense) would be a relatively clear and smoking gun (although of course it would take years for astronomers to be sure).

So at this time, it seems like only mathematical consistency has any hope of guiding us to a theory of qg, and I insist that st is the best candidate currently out there under those guidelines. It is really the only candidate that takes the minimal amount of physical inputs and outputs something that smells roughly correct. There is no alteration of the Rules of Quantum Mechanics or special/general relativity, and no steps are avoided, there is only a simple statement (strings are extended objects) and many of the miracles of modern physics seem to fall out of that.

I think perhaps its just one of those things you actually have to see being done to appreciate, as it sounds like snake oil when its being explained.

 

[ensabah6]

So at this time, it seems like only mathematical consistency has any hope of guiding us to a theory of qg, and I insist that st is the best candidate currently out there under those guidelines. .

More so than LQG?

 

[Fra]

(The paper surprise referred to is 87 pages, I'll try to get time to skim some of it during the next days and get back on that.)

Tom, I guess I don't disagree with anything particular your wrote.
I just have some minor comments.

So there is a paradigm shift we see in modern theoretical physics: stronger focus on mathematics, less stress on phenomenology. This is a rather strange situation, but I see no way out.

I still think it's important to distinguish between not only physics and mathematics, but also between mathematical modelling and mathematics itself.

You're right about the paradigm shift, and it's in particular why I chose to focus stronly on the guiding principles, logic of reasoning and methodology, as this is the only thing you can assess in different programs. The "mathematics itself" is not really something to comment on, beyond consistency issues and level of stringency. To discuss CHOICES of axioms is not a mathematical problem. To discuss "unification" of mathematical branches is also indeed interesting, which some mathematicians do. I find that interesting, but it's somehow not what theoretical physics should be about IMO.

So My focus is really on inference models! In a sense this is reall not physics! It's actually about learning models. This is how I secretly think of many things, and it certainly have applications outside of theoretical physics. This is best IMO seen as ET Jaynes puts it - an extension to logic. The normal language for this is probability theory, usually in some bayesian form or so. There are also other inference rules, such as entropic reasoning.

If you listened to Ariel Caticha's talk on perimeter not too long ago about nature of laws, he points out during the first explanation that "he is not doing physics, he is doing inference".

This is quite different from mathematics. Inference models can be tested without physics. It can be tested in different ways.

One of my conjectures which I share with Ariel and those that work in this direction, is that the laws of physics are really just "rules of inference". And that physical processes, and observations, backreactions from environment etc, can all be described abstractly in terms of inference processing, where each subsystems constantly tries to infer and learn and take control of it's environment.

It's in this light, I critique some of the current researhc programs. From this point of view, I simply find their methodology to not be quite rational. (ie. from the point of view of inference! not phenomenology).

After all, physicists do make inference on nature, this is what we do when we model and construct experiments. I put this on par with physical processes, where one atom makes inference about a neighbouting atom for example.

All this is, in line with the paradigm shift you mention. However it's not just "mathematics" - it's inference, or mathematical models for inference or learning. An inference model is judged not by truthness of theorems or consistency, it's judged on it's fitness. A good inference models, simply makes efficient and good inferences.

Coming back to my example regarding the early days of quantum mechanics. I still think that the situations are comparable, but on a different level. String theory is a broader framework addressing different issues, but we already know a lot of magic numbers and structures to be addressed within this framework: U(1)*SU(2)*SU(3), 6 flavors / 3 generations, Higgs particle (?) Weinberg angle, Yukawa couplings, 4 dimensions, ...

I agree that there are these "magic things" but you refer more or less to the structure of the forces, particle families etc. I too think this calls for some unification, but just because string theory is one of the few "mainstream-candidates" doesn't mean I think it's even close to the only option.

/Fredrik

 

[Fra]

So at this time, it seems like only mathematical consistency has any hope of guiding us to a theory of qg, and I insist that st is the best candidate currently out there under those guidelines.

I know you're very knowledgeable as you typically come with great posts, but I still think a good alternativ to "mathematical consistency" as a guide, is to instead too at our methodology, and in particular formalise it and study inference models. This is also I think a very good guide, as it focuses on "rational behaviour". That can in fact be seen as "consistent reasoning", which is not the same as "consistent results", because it may in fact be that inconsistent states DRIVES development.

I think it IS rational to FIRST try minimal extensions to what we know (ie keep QM intact etc) but OTOH long time has passed with many unsolved problems, and there are also good reasons to doubt the validity of extrapolation of current frameworks. So it is rational to try this first. But at some point, I think it's irrational to not start to question our first conjectures and consider alternatives.

/Fredrik

 

[suprised]

More so than LQG?

Can't you realize that ST plays in a different league?

Otherwise I mostly agree to what has been said esp. by tom, Haelfix, finbar.

 

[Fra]

I realize that I just don't have time to comment on this paper in one session the nearest time as it's too long, so I'll try to skim it in parts and comment.

I recommend: http://arxiv.org/pdf/0807.3249v3 for some food for thoughts.

A first comment (without having started reading) on the abstract on what seems to be the main point

We are in the middle of a remarkable paradigm shift in particle physics, a shift of opinion
that occurred so slowly that some even try to deny that they changed their minds at all.
It concerns a very basic question: can we expect to derive the laws of particle physics
from a fundamental theory? The Standard Model of particle physics as well as the 1984
string theory revolution provided ample food for thought about this. The reason this was
ignored for so long can be traced back to an old fallacy: a misguided idea about our own
importance.

IMO, taking it litteraly, I do not think it's possible to _derive_ (ie. deduce) the laws of particle physics from a static fundamental theory.

But I do think it's possible to makes a very could guess and infer the best expecations of the laws of physics as per some inference scheme, and that this working may in fact reflect not only how science works, but how nature itself works. Ie, the uncertainty in the inference is not an "approximation" it is in fact corresponding to real physics uncertaintes. And we know these things exists in general (just think QM).

For me, from the inference perspective, it's important to distinguish between deduction and induction or abduction. All of them quality as inferences, but deduction is a certain, fault free idealized form of it that doesn't correspond to most real life situations, which are almost without except inference under conditions of uncertainty, where moreover there is no objective independent way of MEASURE the uncertainty.

(The remaning part of the paper, I will see if they expand on htis, but I see conceptual associations to what I suggest here and the meaning of the landscape, and observers. It's just my prior experience that string theorists really does not (genereally at least) see it like this - this is why I think this could be interesting to discuss)

/Fredrik

 

[tom.stoer]

I agree that there are these "magic things" but you refer more or less to the structure of the forces, particle families etc. I too think this calls for some unification, but just because string theory is one of the few "mainstream-candidates" doesn't mean I think it's even close to the only option. /Fredrik

I only wanted to stress that both today and in the early days of qm there are / were this magic things, so the situations are comparable.

 

[tom.stoer]

More so than LQG?

Of course!

LQG is not aiming to unify gravity with anything else. LQG just says that the very nature of gravity is different both mathematically and physically (just read Rovelli's book or Smolin's and Ashtekar's summaries) that one must first understand how to quantize gravity in order to be prepared to couple it to other forces. LQG does not deny the need for unification, but it postpones it (discoveries like algebraic structures similar to particles emerging from quantum-deformed / framed loops are not mainstream).

 

[tom.stoer]

Tom, everything you say is more or less correct.

Thanks for the agreement.

Quantum gravity is something like 15 orders of magnitude away from anything we might be able to test. ... it seems like the only hope is to get lucky with astrophysics and cosmology.

yes

So at this time, it seems like only mathematical consistency has any hope of guiding us to a theory of qg, ...

yes

... and I insist that st is the best candidate currently out there under those guidelines.

yes and no;
it's not the best candidate w.r.t. these guidelines, it's the only candidate which truly aims for unification ...

It is really the only candidate that takes the minimal amount of physical inputs and outputs something that smells roughly correct.

... and/or the only candidate that mainstream physicists are able to understand; I still try to understand what NCG means and I always come to the conclusion that it is some magic trick; Connes is too clever for me.

Regarding qg, there are a couple of viable approaches (LQG, asymptotic safety, ...), regarding unification I only know ST and NCG.

 

[Haelfix]

"More so than LQG?"

My opinion of LQG is not even wrong. I am not a gravitational theorist, and I have zero experience with it.

I looked at it briefly several years ago in my spare time when I was a graduate student, but it ceased to interest me when it became clear that it involved a lot more than standard Dirac quantization and that the models seemed to change every six months.

It is in that regard that I claim that it is highly nonminimal. You have to learn a new type of quantum mechanics in order to begin understanding it, and its totally unclear if it makes contact with general or special relativity (this is the issue of the semiclassical limit and lorentz invariance respectively). Heck the new version of LQG is not even about the Einstein Hilbert action, so again I have no idea what they're really trying to do.

And much more importantly, it makes absolutely no contact with anything that I work on in my own research (unlike string theory, which often provides a toolset that is valuable well outside the field of quantum gravity).

 

[suprised]

Well I guess that LQG is represented here at this forum very well, IMHO much more than it deserves for its merits and prospects, so perhaps we can try to stick to the title of the thread at least here.

 

[tom.stoer]

As I said: LQG in its current fashion is of no interest for unification. NCG may be interesting, but not under the topic "string theory" (which is rather nice because it may be the first theory which cannot be called string theory :-)

=> I agree with suprised.

 

[ensabah6]

Can't you realize that ST plays in a different league?

Otherwise I mostly agree to what has been said esp. by tom, Haelfix, finbar.

Of course!

LQG is not aiming to unify gravity with anything else. LQG just says that the very nature of gravity is different both mathematically and physically (just read Rovelli's book or Smolin's and Ashtekar's summaries) that one must first understand how to quantize gravity in order to be prepared to couple it to other forces. LQG does not deny the need for unification, but it postpones it (discoveries like algebraic structures similar to particles emerging from quantum-deformed / framed loops are not mainstream).

"More so than LQG?"

My opinion of LQG is not even wrong. I am not a gravitational theorist, and I have zero experience with it.

I looked at it briefly several years ago in my spare time when I was a graduate student, but it ceased to interest me when it became clear that it involved a lot more than standard Dirac quantization and that the models seemed to change every six months.

It is in that regard that I claim that it is highly nonminimal. You have to learn a new type of quantum mechanics in order to begin understanding it, and its totally unclear if it makes contact with general or special relativity (this is the issue of the semiclassical limit and lorentz invariance respectively). Heck the new version of LQG is not even about the Einstein Hilbert action, so again I have no idea what they're really trying to do.

And much more importantly, it makes absolutely no contact with anything that I work on in my own research (unlike string theory, which often provides a toolset that is valuable well outside the field of quantum gravity).

Suppose that future evidence coming from the LHC, SuperCDMS, Fermi Gamma-ray Space Telescope, Super-Kamiokande etc., is empirically consistent only with the theory that nature is only 4D and SUSY is unrealized in nature, and GUT such as SU(5) and SO(10) are falsified. Do you think string theory is, under the additional, experimentally observed guidelines of no-SUSY, 4D, no SO(10) unification, still the "st is the best candidate currently out there under those guidelines"

 

[ensabah6]

As I said: LQG in its current fashion is of no interest for unification. NCG may be interesting, but not under the topic "string theory" (which is rather nice because it may be the first theory which cannot be called string theory :-)

=> I agree with suprised.

Is ST a more promising QG than LQG ?

 

[suprised]

Suppose that future evidence coming from the LHC, SuperCDMS, Fermi Gamma-ray Space Telescope, etc., is empirically consistent only with the theory that nature is only 4D and SUSY is unrealized in nature. Do you think string theory is, in this scenario, still the "st is the best candidate currently out there under those guidelines"

Try to read what was written above, about the meaning of higher dimensions and the role of SUSY in string theory. I had commented on this in the past several times as well. It makes no sense to repeat it all the time, again and again, esp. if poeple don't want to read and/or understand.

The answer to your question is yes, as I don't see a good reason, as of today, to think otherwise.

Is ST a more promising QG than LQG ?

LQG hasn't even convincingly demonstrated that it describes gravity in 4d in the first place; nor that it's quantization makes any sense; nor that it is a single theory with a clear definition (rather it is a collection of attempts whose starting points change all the time and, after decades of research, didn't get anywhere near to what has been promised).

So again, please let's keep at least this thread clean from this fog, ok?

We could open other thread about what is wrong with LQG, why so few people believe in its merits, why there is so much hype involved promoting it, and so on. Though I suspect you wouldn't like to read it.

 

[tom.stoer]

I do not agree with suprised's statements regarding LQG, but I will not respond here in this thread.

The situation with strings and the LHC is the following:

A) even if the LHC does not produce a single mini black hole in the TeV range, even if there are no large extradimensions, and even if the LHC does not find a single SUSY particle, nearly nothing will happen to the basic principles of string theory. String theory can still be right or wrong. Our interpretation of the situation will be different; the position of string theory regarding fashion, acceptance, support, money, ... will be affected, but not its fundamental principles.

B) if the LHC finds SUSY particles fitting perfectly to some (x)MSSM the same applies; of course this will support string theory in some sense, but it does not affect its basic principles. Verifying SUSY does not help in identifying new underlying principles or axioms of string theory, it does not help in developing an off-shell formalism, pushing string field theory forward, establishing a background independent framework, identifying a vacuum selection principle, ...

Regardless what the LHC finds, the situation regarding basic principles will stay the same.

 

[suprised]

I can't answer all what I wanted but a few comments on those:

(LQG just says) that the very nature of gravity is different both mathematically and physically (just read Rovelli's book or Smolin's and Ashtekar's summaries) that one must first understand how to quantize gravity in order to be prepared to couple it to other forces. LQG does not deny the need for unification, but it postpones it (discoveries like algebraic structures similar to particles emerging from quantum-deformed / framed loops are not mainstream).

Let's see whether they will ever succeed. If string theory is right, then this program should fail, because there matter IS needed for consistency. This is a nice dividing line between those approaches. If matter is not needed, even worse for them, because how could they ever hope to get constraints on particle physics then?



Fully agreed on:

Regardless what the LHC finds, the situation regarding basic principles will stay the same.

But not quite on:

A..NCG may be interesting, but not under the topic "string theory" (which is rather nice because it may be the first theory which cannot be called string theory :-)

Why cannot? They present a cute way to parametrize (part of) the standard model. You want to imply that the standard model is not contained in string theory?

More specifically: as written before, the dualities discovered during the last 15 years have shown that there are in general many different ways to geometrically represent a given theory; there is no absolute meaning of a given background geometry, be it continuous or discrete. Those authors have discovered still another way, by using a non-commutative structure. There is no a priori reason why that would not be contained in string theory.

 

[tom.stoer]

It's rather simple: I am referring to a statement "whatever ... it may be ... it may look like, we will still call it string theory" I know from Smolin's book ... I am not sure so I have to check ...

What I am sying is this: NCG is not related mathematically to string theory.

If you have the QCD beta function which contains 11/3 and if I give you an 11-faced geometrical object with 3 yellow faces, that does not automatically mean that the QCD beta function is somehow related to this 11-faced object; it does not mean that this object appears somewhere in QCD; nor does it mean that probability theory is dual to QCD.

So even if both ST (for which it is not proven) and NCG converge in some appropriate limit to the SM that does not automatically mean that NCG is related to ST mathematically. What you are saying here is rather dangerous - and I have the feeling that this applies to the web of dualities as well: the fact that certain theories or formulations match in some appropriate limit does not automatically mean that they really match. It is dangerous to think about theories like coordinate patches for which a small smooth overlapping region is sufficient. Low-energy effective theories (chiral perturbation theory, nonrelativistic quark model) are in some appropriate limit related to each other. But they are not identical; they can only be identified via QCD; w/o QCD the essential unifying structure is missing.

For ST this means that the web of approximate dualities is perhaps too weak to be called a theory. It is perhaps not sufficient to stay with this web of approximate dualities w/o being able to identify the underlying unifying structure. The initial idea and program of M-theory is still waiting for completion!

 

[suprised]

It's rather simple: I am referring to a statement "whatever ... it may be ... it may look like, we will still call it string theory" I know from Smolin's book ... I am not sure so I have to check ...

You don't need to look up, it doesn't matter what he said. I'd even refine it: everything what makes sense should be contained in the big blob of consistent theories, that one calls string theory. In particular the standard model, no matter in what parametrization! On the other hand, not Lisi's model, for example.

NCG is not related mathematically to string theory.

Ehm? Didn't you see papers on compactifications on non-commutative spaces? Eg http://arXiv.org/abs/hep-th/9711162
Non-commutativity is in fact a basic feature of any theory with open strings. Configurations of D-branes can generically be viewed as some kind of non-commutative manifolds, so called D-manifolds. And background "B"-fields automatically lead to non-commutativity as well.

If you have the QCD beta function which contains 11/3 and if I give you an 11-faced geometrical object with 3 yellow faces, that does not automatically mean that the QCD beta function is somehow related to this 11-faced object; it does not mean that this object appears somewhere in QCD; nor does it mean that Probability theory is dual to QCD.

Indeed so, but that's not claimed. Their model seems _nothing else_ than the (though incomplete, unquantized) standard model, isn't it? So why you insist it can't have a realization in terms of string theory?

Their only point, IMHO, is that this structure seems natural in their language, while not obviously so in string theory. But other choices would also be possible. I don't know what to conclude from this.

Let's present a Gedanken-teaser. Recall that the standard model gauge group, SU(3)xSU(2)xU(1), can be viewed as the exceptional group E4, and forms a member of the natural sequence E8,E7,E6,E5=SO(10),E4. So one could say: oh I "derive" the gauge group of the standard model incl its matter reps by postulating that the fundamental theory is based on E4! Sounds great, as exceptional structures fascinate everybody... but what does it buy? Another choice would also be possible.

This just boils down whether there is a mathematical principle that is free of any choices, that leads to the standard model. Surely everyone would dream about monster groups, leech lattes, division algebras...but so far nothing has ever be found! And there are good reasons to believe that it does not exist - for some arguments, read the article on the landscape I cited above.

 

[Haelfix]

Regarding NCG and string theory. Isn't there some sort of theorem that was found in the early days, whereby every low energy realization of NCG was consistent only if the high energy theory was also part of string theory (or a limit thereof).

I seem to recall a lecture about this, but I didn't understand it so I have now promptly forgotten.

 

[tom.stoer]

I do not know enough details regarding NCG a la Connes, but all what I have seen so far has nothing to do with strings; if this impression is wrong please forget about it; it's due to my limited understanding of NCG.

But be careful and please take my 11/3-example seriously. Yet another example: just because LQG and ST both use complex numbers, they need not be related.

As I don't think that you responded to the most important part of my post, so I'll re-post it here:

Even if both ST (for which it is not proven) and NCG (a la Connes) converge in some appropriate limit to the SM that does not automatically mean that NCG is related to ST mathematically. What you are saying here is rather dangerous - and I have the feeling that this applies to the web of dualities as well: the fact that certain theories or formulations match in some appropriate limit does not automatically mean that they match exactly. It is dangerous to think about theories like coordinate patches for which a small smooth overlapping region is sufficient. Low-energy effective theories (chiral perturbation theory, nonrelativistic quark model) are in some appropriate limit related to each other. But they are not identical; they can only be identified via QCD and w/o QCD the essential unifying structure is missing. For ST this means that the web of approximate dualities is perhaps too weak to be called a theory. It is perhaps not sufficient to stay with this web of approximate dualities w/o being able to identify the underlying unifying structure. The initial idea and program of M-theory is still waiting for completion!

 

[atyy]

There's NCFT which is a part of string theory, and NCG which is not a part of string theory (or at least not known to be).

 

[suprised]

Ill answer tom tomorrow, here just a remark on:

Regarding NCG and string theory. Isn't there some sort of theorem that was found in the early days, whereby every low energy realization of NCG was consistent only if the high energy theory was also part of string theory (or a limit thereof).
.

The NCG model of Connes et al focuses on the standard model coupled to gravity, at the classical level. Its virtue is in the derivation of these couplings from simple-to-state principles. It says nothing whatsoever on quantum gravity, and all problems with quantizing gravity will be there. In other words, the UV completion is not addressed, and either one assumes a string embedding, which means to be on the safe side, or one bets on a miracle by which this theory turns out to be consistent (for which there is not the slightest indication).

And I can't grasp why this model is listed here in other threads under quantum gravity...but never mind...

 

[Haelfix]

Hi Surprised,

Actually Connes' model wasn't what I had in mind. I was some statement that I heard in a colloquim regarding decoupling limits in string theory. Evidently the claim was something like a NC Yang Mills theory was only consistent if it had some sort of UV limit that also included extended states. The authors arrived at the result by carefuly analyzing the decoupling limits of nc backgrounds in string theory (which is probably where I lost it)

Something like that anyway, no matter.

 

[atyy]

Evidently the claim was something like a NC Yang Mills theory was only consistent if it had some sort of UV limit that also included extended states.

Are you thinking of something like Steinacker's work? He talks about the relationship to string theory a bit on p 18 of http://arxiv.org/abs/1003.4134

 

[Fra]

One point (still haven't read the entire paper properlt).

An overall impression I have is that he is closing his eyes for his own reasoning in certain respect.

At one point comments about the possible idea of (smolins) evolving law in the context of the ST landscape.

The answer to questions like these requires a proper understanding of the thorny issue of probabilities, and may well be beyond the edge of our current knowledge

I fully agree, but the point is that this is unavoidable anyway. He uses this implicitly all over the place in his inference.''

If different laws of physics are possible, we { or anything else of comparable intelligence { might not exist in some of those dierent circumstances.

Already here is is using precursors of inductive and probabilistic reasoning.

About the string landscape he talkes about his past conclusiosn

Observing that the number of possibilities was huge, in the approximation used at the time, was a relatively simple matter. But it was not obvious what to conclude from that.

In several places he still uses the notion which he admits is thorny.

It's clear that he is inconsistent in the sense that he seems to thinkg that probability is thorny when it comes to evolution, but uses it without problems in his mathematical sense.

My suggestion is merely that we need to be aware of our own reasoning, and try to reason consistently. I suggest here that the "thorny probability" issues has to be solved. He seems to reject evolution of law on this basis, yet he uses similar reasoning himself.

He also violates the use of observation, he sats he "observes" the probabilities in hte landscape.

When we try to analyse the meaning of probability, we also come to the meaning of physical law? something which accoriding to hime as I see it is a forcing constraint (except of course there is a whole landscape of such POSSIBLE constraints) (*)

We certainly must aim to understand this. This in fact has nothing to do with stirng theory, but the fact that he touches on it, and manages to avoid digging into the problem is interesting and I think part of the core of SOME of the issues with string theory.

I think these things (the constructing principles and logic) are the most interesting parts that at this point - with the lack of experimental contant - are most WORTH discussing.

edit: more later...
(*) I could argue thta from the inference perspective this is irrational. Physical laws are better thought of as "expectations of the future" that constrains and guides the observers ACTION and interactions with the environmnt. In this way, several of the above things gets easier to understand. This indeed also gives us in a certain sense several different "laws" BUT the major difference is that they are interacting, and there is an evolution as a result of their negotiation. This is the projection I mentioned earler. But for some reason there is no hint of this association in the Schellekens paper. I think the rigid reasoning and the quest for deductive schemes, prevents him from seeing the beauty of a possible inductive scheme, which is related to evolution.

/Fredrik

 

[Fra]

I've skimmed Schellekens paper a couple of times, and while I agree on some things, some of his overall reasoning is objectionable, and I trying to figure how to make the point clear.

My point is quite the opposite: to show that even the greatest minds of their time can make the obvious anthropocentric mistake of thinking that what we see around us is all there is.

Again here is a difference between deduction and induction; absolute truth and rational expectations. A distinction I don't think Schellekens makes; he only sees truth and deduction.

Can we deduce that all we see is all there is and will ever be, or that we will ever see? Of course not, that would indeed be a fallacy.

This seems to be his conclusion.

I think the question he poses is wrong. The question is; is it possibly RATIONAL to base your actions on what you do see? I think the answer is yes. This how a rational player in a game acts. The players information about the rest of the world, largely rules his actions.

The question is not, to answer to eternal truth, what is or or not, or what we will or will not ever see. The rational question any information processing agent asks is: what is my next step ~ what is my optimally rational action?

For example a decision maker, makes a decistion based on available information and acts rationally from that. This can be rational, even if the final result proves to be wrong. There was long time ago a discussion about the notion of "false information" in the QM interpretaton, and the story is similar. A systems ACTION, given an initial state does not need a measure of false or true.

Similarly, we do not need to KNOW in some objective way, wether all we KNOW or see is all we will EVER know or so, or wether this will change in the future. All we need to do, is to evaluate and execute the next step. That's the only rational action.

This gets philosophical and it seems really hard to get this point through, but I'm trying to make a point here, about evolving model building in general, and suggest that it applies to the inference we physicists do, and that it also applies to any physical inferences.

Note that I don't object to lack of perfect uniqeness per see, I agree with part of what he says - that it's impossible to deduce from an obvious axiom system - the laws of physics in detail. The objection is that he seems to focus on the wrong question. He seems to focus on the ultimate state of law, rather than the process of inferring the law! Now that's to me as irrational.

The flawed focus is IMO what gives the gigantic landscape of possibilities. But I do not think these possibilities are physical, or relevant in the way they are devised.

Edit: implicit in my objection and reasoning here is also a different view of what "law" means, although I didn't write it explicitly. The law is rather a "rule" or "guide" for construction an action. A "tool" for interactions. This tool is itself evolving. This is quite different from the realist view of law as "forcing mathematical constraints".

/Fredrik

 

[suprised]

..
just because LQG and ST both use complex numbers, they need not be related…
..
Even if both ST (for which it is not proven) and NCG (a la Connes) converge in some appropriate limit to the SM that does not automatically mean that NCG is related to ST mathematically.

Of course, I completely agree. But what I was contesting was not this, but a statement that their model cannot be called string theory.

Actually I re-read what you wrote and found that you said it weaker than what I had in mind: "… it *may* be the first theory which cannot be called string theory :-)"
I won't argue against this ;-)

I also looked again at their paper, in order to find out to what extent it is an approximation to the standard model. In fact, it is an effective theory up to some scale L, and contains an arbitrary function f which becomes important at energy scales comparable to L. So without specifying f the theory is not fully defined and its consistency and the issue of UV completion is left open; as is the question whether there is some string construction into which it embeds and which would determine f.

This is not to play this nice construction down, rather to put it into perspective with string theory, which is the topic here. I would view its status as similar to what is called "local model building" in string phaenomenology. Here one also postulates a background geometry which reproduces the standard models as well as possible, as an effective field theory (containing corrections involving functions f that are not important at low energies). Again, this description breaks down at a scale above which the full theory becomes important.

In the string construction, one knows what the full theory is, and in principle one can compute the functions f.

For ST this means that the web of approximate dualities is perhaps too weak to be called a theory. It is perhaps not sufficient to stay with this web of approximate dualities w/o being able to identify the underlying unifying structure. The initial idea and program of M-theory is still waiting for completion!

Perfectly agreed. That's why I prefer to view it as a framework that generalizes QFT.