Why I am REALLY disappointed about string theory

[suprised]

I have two comments for the time being:

1) Whether SUSY is needed for consistency of strings. I don't believe, nor I hope so. Our world is not SUSY at low energies, we don't see a tachyon, so strings better are able to cope, as a matter of consistency, with non-susy backgrounds. And if they could cope with that, why should be SUSY then need to be restored at the weak scale and not, say, right at the Planck scale?

Indeed SUSY is a powerful symmetry principle that helps to prevent tachyons, and protect the cc etc; generically models without SUSY are threatened by tachyonic instabilities. And of course SUSY is very handy for heaving treatable models, but that's not an argument why Nature should care about this. AFAIK there is no theorem that says, non-susy strings imply tachyons (and as said above, there better be no such theorem). Thus it is unclear whether SUSY is a technical convenience for toy model building, or really a deep principle of nature.

The situation is in fact more complicated, as there exist metastable vacua where SUSY is "broken temporarily" and the true ground state is SUSY. AFAIK no definite conclusions can be drawn here, all I want to say is that the scenario of having unbroken SUSY at low energies may have been quite a substantial blind ally.

2) Geometrical vs non-geometrical compactifications. With geometrical I meant "classical geometry" involving manifolds, field configurations like vector bundles, etc. In short, all what comprises the good old supergravity school of thinking. Certainly this has been very useful and fruitful, but neverless captures only the boundary of the string parameter space.

With non-geometrical I meant "stringy geometry". I make my life easy and define this by simply saying that's it is a kind of generalized geometry that takes stringy features properly into account (eg by identifying classical geometries that are related by dualities). It is eg well-known that the notion of D-branes wrapping sub-manifolds needs to be replaced by abstract mathematical constructs like derived categories of coherent sheaves, when we move away from the boundary of parameter space). This would be the proper language to describe strings in the bulk of their parameter space. And generically this can NOT be mapped back, by dualities, so some classical geometry.

Again, all what I want to say is that focusing on the language of classical geometry, can be a major blind ally since it excludes the _main part_ of the string parameter space.

 

[suprised]

Since you didn't specify the current state of thinking, may I ask your opinion about it? For example, would a majority of string theorists disagree with: string theory is a rich theory, with a landscape of solutions where to 0th order anything goes, where susy is not essential or generic and where higher dimensional geometry is not essential or generic?

Well that depends to whom you talk to. I believe the majority of colleagues would say that SUSY and naive extra dimensional geometry are essential, not only technically. I raise these points here as a Devil's advocate since Tom asked for potential blind allys.

 

[suprised]

Anyway, my issue with Gepner models is they seem to have issues generating correct family structures in the standard model, which is to be contrasted with some of the other vacua that seem to pick out 3 generations uniquely. Further it is unclear which way the generalization goes. I distinctly recall a theory seminar where it was shown that Gepner models typically reproduce isolated points in the moduli space arising from usual CY compactifications. Consequently it was perhaps the case that the nongeometric vacua were subsets of the geometric ones..

edit: for non string theory cognescenti.. Gepner models naively seem to generalize world sheet coordinates. Instead of scalar fields, we are thinking about more abstract mathematical objects like Ising models or conformal minimal models and things like that. They are rather weird in that you have states with fractional charges floating around the place. However, surprisingly you can prove the equivalency of these models with the more familiar ones (eg ones with the usual boson and fermion degrees of freedoms) by analyzing how objects behave in the target space. Here the actual dimensionality of the critical string is completely obscured, although other physical criteria (like recuperating supersymmetry) becomes manifest. The surprising thing here is that what started out as an apparent generalization from the worldsheet point of view, actually becomes equivalent or even perhaps weaker looking at the target space.

Gepner models illustrate my points. Some of these models have a direct relation to string compactifications in CY spaces (in the deep quantum regime, ie, where the CYs are very small and string effects are important).

The prime example is the quintic, described by a Gepner model with Landau-Ginzburg superpotential

W = Sum_(i=1)^5 (x_i)^5

W=0 is nothing but the equation of the quintic CY:
5 coordinates minus the equation W=0, minus rescaling gives 3 (complex) coordinates, so this yields indeed a six real-dimensional CY. So here we can make a nice map between 2d CFT and space-time compactification manifold.

The following Gepner model is from the 2d CFT point of view on the same footing as the model before:

W = Sum_(i=1)^9 (x_i)^3

However, 9 coordinates minus the equation W=0, minus rescaling gives 7 (complex) coordinates... so this looks naively like a 14 dimensional manifold... as such certainly not useable as compactification manifold! But this model defines a perfectly valid string vacuum.

So we see here that the 2d, "non-geometrical" formulation gives rise to more string vacua than most would have thought of when naively thinking about compactification manifolds!

 

[atyy]

Since the extra parameters are not necessarily spacetime dimensions, are there any examples where these parameters give rise to 4D spacetime and the rest being non-geometrical?

 

[fzero]

Since the extra parameters are not necessarily spacetime dimensions, are there any examples where these parameters give rise to 4D spacetime and the rest being non-geometrical?

Yes, the restrictions on matter in the superstring come from requiring worldsheet conformal invariance. The anomaly is proportional to the total central charge, c, of the worldsheet theory. One first computes the central charge of the ghosts required to fix worldsheet diffeomorphism and Weyl invariance. This is the b,c system and has central charge -26. If we don't include worldsheet SUSY, then the worldsheet matter must have c=26. Since a free boson CFT has central charge 1, this leads to the result that D=26 for the bosonic string.

If we include SUSY, we find another ghost system \beta,\gamma that fixes local worldsheet SUSY. This system contributes another central charge +11, leaving -15. A free fermion has central charge 1/2, so 15 = 10(1+1/2) gives D=10 for the usual superstring.

If we want only a 4D system, we will be left with central charge -15+4(1+1/2) = -9, so our not-necessarily-geometric "internal" CFT must have central charge c=9. In the geometric compactifications, this is supplied by 6 more boson-fermion pairs, but in general could be supplied by any CFT we can stitch together.

 

[suprised]

Since the extra parameters are not necessarily spacetime dimensions, are there any examples where these parameters give rise to 4D spacetime and the rest being non-geometrical?

Zillions of examples...

 

[arivero]

1) Whether SUSY is needed for consistency of strings. I don't believe, nor I hope so. Our world is not SUSY at low energies,

This is the blindness -the wrong turn- I try to fight in the last years: our world IS susy at low energies, and because of it we confused the pion with the muon in the fifties.

It was a very prepostereous thing to say, so five minutes after proposing it (basically a couple of publications by John H. Schwarz in 1971, following the discovery of the Ramond string), everyone, including Schwarz, forgot about it. But with three generations, the degrees of freedom match. It is susy, it is the qcd string, they were right from the start, and the only point today is why the non-chiral interactions get their gauge bosons massless, but not the partners. If we find the gauginos -and only them- the question will be settled.

 

[tom.stoer]

oops :-[

Could you please write down how to "pair" the known particles?

 

[atyy]

Yes, the restrictions on matter in the superstring come from requiring worldsheet conformal invariance. The anomaly is proportional to the total central charge, c, of the worldsheet theory. One first computes the central charge of the ghosts required to fix worldsheet diffeomorphism and Weyl invariance. This is the b,c system and has central charge -26. If we don't include worldsheet SUSY, then the worldsheet matter must have c=26. Since a free boson CFT has central charge 1, this leads to the result that D=26 for the bosonic string.

If we include SUSY, we find another ghost system \beta,\gamma that fixes local worldsheet SUSY. This system contributes another central charge +11, leaving -15. A free fermion has central charge 1/2, so 15 = 10(1+1/2) gives D=10 for the usual superstring.

If we want only a 4D system, we will be left with central charge -15+4(1+1/2) = -9, so our not-necessarily-geometric "internal" CFT must have central charge c=9. In the geometric compactifications, this is supplied by 6 more boson-fermion pairs, but in general could be supplied by any CFT we can stitch together.

Zillions of examples...

So we don't obviously need Calabi-Yau compactifications?

 

[arivero]

oops :-[

Could you please write down how to "pair" the known particles?

Sigh . Guys, just look at the data.

It is about taking seriously the ideas of http://dx.doi.org/10.1016/0370-2693(71)90028-1" ): the fermion in the dual model is susy to gluonic strings. So now all you need is to terminate the gluonic string. Regretly in 1971 there were only three states available to terminate the string: u, d, and s. Now we have the full history, and the experimental data tell us that we can terminate the gluonic string with five and only five different states: u, d, s, c, b.

So just count, please, just do the SU(5) global flavour game, and count. How many states do you get of charge +1? six, by terminating with particle and antiparticle. How many of charge +2/3? six of each colour, by terminating with an antiparticle at each end of the string. How many of -1/3? six. How many +1/3, -1, -2/3? Same: six, six, six. And how many neutrals? of course, twelve: the other half of the 24 of SU(5).

BONUS: Does it means that string theory, given as input the 3-2-1 gauge theory of the SM, predicts three generations? No exactly; only if we require that the neutral leptons must be produced too. If we only look at the quark sector, then any pairing of 2^{p} "up quarks" with 2^{p+1} -1 "down quarks" will produce equal number, 2^p (2^{p+1} -1) of up and down combinations, and p=1 is just the simplest case. Numerically minded people will notice that p=4 amounts to 496, but a theory with 16 light "down" quarks, 31 light "up" quarks and a total of 248 generations seems not to be the object that Nature has offered us.

EDIT: other references using "fermion-meson": http://dx.doi.org/10.1016/0550-3213(74)90529-X Nuclear Physics B Volume 74, Issue 2, 25 May 1974, Pages 321-342 L. Brink and D. B. Fairlie; http://www.slac.stanford.edu/spires/find/hep/www?j=NUCIA,A11,749 Nuovo Cim.A11:749-773, 1972 by Edward Corrigan and David I. Olive. Modernly, they are some works, in the framework of SQCD and also in Holography, that work with "mesinos", in the sense of susy partners of mesons. But note that phenomenologists call also "mesino" to the combination of squark and quark.

 

[fzero]

So we don't obviously need Calabi-Yau compactifications?

I wouldn't say don't need, for a couple of reasons. For one, CY compactifications are a class of c=9 theories. However, it could be that a nongeometric model gives physics that is close to reality. As an example, there are some models with 3 generations in http://arxiv.org/abs/1009.1320 though they also seem to find massless fractionally charged states that could be a problem for phenomenology.

However, it is known that many of these nongeometric theories are equivalent to CY compactifications at special points in moduli space. For some evidence of this, one can look at Witten's http://arxiv.org/abs/hep-th/9304026, which relates some of them (so called N=2 minimal models) to Landau-Ginzburg theories. These LG theories are themselves known to be a phase of CY sigma models http://arxiv.org/abs/hep-th/9301042

I don't believe that the state of knowledge about the equivalence between nongeometric and geometric models is developed completely, but I think it's strong enough that it wouldn't make sense to drop CY models. If anything, the equivalence itself should be studied further, since it might teach us more things about the space of c=9 models.

 

[tom.stoer]

I would like to come back to suprised's list regarding possibly wrong turns.

1. - That geometric compactification of a higher dimensional theory is a good way to think about the string parameter space
2. - That perturbative quantum and supergravity approximations are a good way to understand string theory
3. - That strings predict susy, or have an intrinsic relation to it (in space-time)
4. - That strings need to compactify first on a CY space and then susy is further broken. That's basically a toy model but tends to be confused with the real thing
5. - That there should be a selection principle somehow favoring "our" vacuum
6. - That a landscape of vacua would be a disaster
7. - That there exists a unique underlying theory
8. - That things like electron mass should be computable from first principles

Let's look at this list again: there is a deep connection between some topics; that's why I was mentioning background independence. I would like to comment on this once more.

String theory walked - for a rather long time - on the trail of particle physics and quantum field theory. Of course there was a graviton, but after recognizing this particle one immediately focussed on QFT-like reasoning (background, strings on top of this background, perturbative quantization, ...). I would say that the first few topics are essentially due to this perception of string theory.

Looking at the field today most researchers are convinced that non-perturbative approaches are required. Thousands of backgrounds / vacua have been identified, but still they are mostly perceived as reasonable backgrounds on which standard particle- or QFT-like theories can be formulated. This is OK for model building an phenomenology (it is not only OK but of course heavily required in order to achieve a closer relation to reality).

But using intuition to find such backgrounds and doing "ordinary physics" on top of these backgrounds does not help in order to understand the relation between these backgrounds and to identify the "unique" and deeper origin of these backgrounds, which I would call the underlying theory.

I think another wrong turn - perhaps the most serious one - would be to turn a bug (the missing unique underlying theory) into a feature (we do not need a unique underlying theory). It would be same as looking at the periodic system and stating that happily there is no underlying theory required as we have a collection of relations between different chemical elements.

I think we do not need to look for a selection principle ("why is it iron instead of copper?"), we do not need to condemn the landscape ("iron, copper, mercury, oxigen, ... is too much; we need a single solution"), we do not need to look for a way to calculate the mass of the electron ("how do we calculate the mass of the mercury atom in a theory which does not explain why there is a mercury atom?"). All what we have to do is to understand what string theory really is. My impression is that we still do not know, we are scratching at the surface, we see some "effective models", not more (and not less).

So 1. - 4. may have been wrong turns - but were overcome somehow over the last years. 5., 6. and 8. are perhaps wrong turns which are in the spotlight today. 7. is not a wrong turn but the essential driving force of progress in physics. I would not abandon it w/o having a worthy successor.

I am still with David Gross (and others - like Weinberg I guess) who asked exactly these questions:

• 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. It is hoped that the recent efforts to construct covariant second quantized string field theories will shed light on this crucial question.
  
• We still do not understand what string theory is.
  We do not have a formulation of the dynamical principle behind ST. All we have is a vast array of dual formulations, most of which are defined by methods for constructing consistent semiclassical (perturbative) expansions about a given background (classical solution).
  
• What is the fundamental formulation of string theory?

Denying the relevance of these questions is - in my opinion - the "wrongest turn ever".

 

[suprised]

Nicely said, Tom.

Though I think I should explain what I meant with 7) "there exists a unique underlying theory".
Much could be said here. For the time being, let me provocative and say the following:

Strings seem to be the natural generalization of gauge theory, actually closely related to it by dualities, such as AdS/CFT; in the latter context, strings are indeed reconstructed from gauge theory. So let's view strings as analogous to gauge theory; and then re-ask the same question: "what is the underlying unique theory of gauge theory" ?

Clearly this is a not very fruitful question to ask, because it presupposes something which does not exist, at least in the sense of the question. All there is with gauge theory, are various degreses of freedom that are exposed depending on the energy scale (gluons, quarks, mesons...)

As for strings, the situation is unclear but it may be similar - there may be no further "unique underlying theory". All there might be is the complicated web of perturbative approximations related by dualities, but there is no regime where "universal, more fundamental" degrees of freedom would be liberated.

The real question is whether there is an encompassing, "off-shell" mother theory which would contain all the known theories as "critical points", and describe transitions between them, etc. This may, or may not exist (analogous to gauge theory). So this question is a potential blind ally as well!

 

[suprised]

So we don't obviously need Calabi-Yau compactifications?

They are just special examples of vacua, their main advantage is being relatively well under technical control. That's why there has been so much focus on them, unfortunately thereby creating the impression that they would be somehow essential. But there are zillions of other constructions (generalized geometries with fluxes, non-geometric vacua, brane backgrounds, non-perturbative F-theory vacua, M-Theory vacua,... ).

Of course, many of such vacua are equivalent via dualities, and this shows, again, that there is no objective, unambiguous meaning of a compactification geometry.

 

[Fra]

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.

This is the most serious concern I've always had.

Exactly becuase, string theory seems to be a framework or research program - rather than a unique mature theory, the logic of reasoning used is even more important; because this is what defines the program.

/Fredrik

 

[tom.stoer]

... let's view strings as analogous to gauge theory; and then re-ask the same question: "what is the underlying unique theory of gauge theory" ?

Clearly this is a not very fruitful question to ask, because it presupposes something which does not exist, at least in the sense of the question. All there is with gauge theory, are various degreses of freedom that are exposed depending on the energy scale (gluons, quarks, mesons...)

As for strings, the situation is unclear but it may be similar - there may be no further "unique underlying theory". All there might be is the complicated web of perturbative approximations related by dualities, but there is no regime where "universal, more fundamental" degrees of freedom would be liberated.

The real question is whether there is an encompassing, "off-shell" mother theory which would contain all the known theories as "critical points", and describe transitions between them, etc. This may, or may not exist (analogous to gauge theory). So this question is a potential blind ally as well!

I agree to this view - at least currently string theory seems to be a framework for constructing and defining theories; this framework is capable of producing ordinary (SUSY) gauge theory plus gravity (which is not possible in the framework of gauge theory alone).

But already in gauge theory we asked the question "why the standad model? why exactly U(1)*SU(2)*SU(3)"? Or "why gauge bosons, why not spin 5/2 particles, ...?"

I agree that these questions (translated to the string theory language) could be dead ends. But I bet that going into these directions we will learn a lot - even if they are dead ends.

 

[arivero]

BONUS: Does it means that string theory, given as input the 3-2-1 gauge theory of the SM, predicts three generations? No exactly; only if we require that the neutral leptons must be produced too. If we only look at the quark sector, then any pairing of 2^{p} "up quarks" with 2^{p+1} -1 "down quarks" will produce equal number, 2^p (2^{p+1} -1) of up and down combinations, and p=1 is just the simplest case. Numerically minded people will notice that p=4 amounts to 496, but a theory with 16 light "down" quarks, 31 light "up" quarks and a total of 248 generations seems not to be the object that Nature has offered us.

Allow me a correction to this remark: Of course, the quark sector condition works for any integers q and 2 q -1, with q an even number, not necessarily a power of two. But that the powers of two are an interesing subset was noted by Peter Crawley in other thread time ago and I am kind of obsessed with this, because it could constitute the way to reconnect with usual string models, via the above p=4 case.

 

[Fra]

Clearly this is a not very fruitful question to ask, because it presupposes something which does not exist, at least in the sense of the question. All there is with gauge theory, are various degreses of freedom that are exposed depending on the energy scale (gluons, quarks, mesons...)

Unique theory is a strong phrase, and I do not expect that either in the meaning of eternal objective theory.

But I think a fruitful and necessariy question seems to require an understanding of these "various degrees of freedom" and how and why they are related by means of gauge symmetries in the context of a measurement theory.

I expect that state spaces and theories are to be described as the result of an interaction history. This includes also inferred "gauge symmetries". That are like inferred evolving constrainst that constrain the action of the observer. It's interesting that these symmetries are "energy dependen" as you say, but one can also see them as generally observer dependent. All this is quite interesting and seems to lead to an intrinsic measurement theory that involved emergent constraints (gauge symmetries).

This MAY suggest a general framework for inference (this is exactly why there is no external FIXED unique description, since it keeps evolving)

The question I ask is: could string theory be that framework? If we can understand ratianally that the action of quantized strings in classical backgrounds somehow corresponds to such "gauge choices" that are furthermore scale dependent (so as to give rise to a range of dualities) then I think that would be extremely beautiful and powerful.

Ie. that vision is nice. But is really string theory this theory of theory that I think a lot of people that do not today enjoy string can appreciate?

For example, has any string theorist ever tried to justfiy the basic string action, from a pure inferencial perspective? Ie. that the string action can be understood as an optimal action on the set of possible changes constriaied by historically inferred constraints? (we are conceptual analogues of gauge symmetry)

I think that the focus and hope of string theory is to actually BE the "theory of theory" that some hopes for.

What traits would one ask for such type of theory, and what is the purpose of such theory? descriptive or as an interaction tool?

/Fredrik

 

[suprised]

Fra, I really can't answer your questions, I barely understand them.

But I comment on this:

For example, has any string theorist ever tried to justfiy the basic string action, from a pure inferencial perspective?

There is no such thing like a basic string action. There are various actions, with different symmetries (like heterotic string world sheet, like type II string world sheet, like open type I world sheet...). They are all different, and each one refers to some particular perturbative approximation centered at a different regime. Moreover, for F-theory or M-theory such "world-sheet" actions are not known or may not exist; as we have discussed earlier, there are quantum theories which are strongly coupled and no lagrangian or action description of them exists.

So the string world-sheet perspective (Polyakov action and generalizations), while very useful in many situations (eg see the above discussion about CFT and internal degrees of freedom), is hardly fundamental. Trying to find a deeper meaning of it had been another of many blind ally's.

That's one of the most important conceptional riddles: does a "fundamental" action that would universally describe strings in every corner of the parameter space exist at all? I don't know but I doubt it.

 

[Fra]

I know what I asked is fuzzy, but thanks for trying to answer.

There is no such thing like a basic string action. There are various actions, with different symmetries (like heterotic string world sheet, like type II string world sheet, like open type I world sheet...). They are all different, and each one refers to some particular perturbative approximation centered at a different regime. Moreover, for F-theory or M-theory such "world-sheet" actions are not known or may not exist; as we have discussed earlier, there are quantum theories which are strongly coupled and no lagrangian or action description of them exists.

So the string world-sheet perspective (Polyakov action and generalizations), while very useful in many situations (eg see the above discussion about CFT and internal degrees of freedom), is hardly fundamental. Trying to find a deeper meaning of it had been another of many blind ally's.

Yes there are different string actions dependong on what string theory you consider, but that doesn't avoid my question:

Since you might have figured from my strange comments that I'm slowly working on an inference perspective to physics, and in this context, one can talk about actions as a way to measure the information divergence of possible futures relative to present. The idea is to define expected change not as dynamics realtive to external time, but with respect to a observer dependent entropic flow. IE to understand the concepts w/o referencing mechanical or geometrical visualisations.

As far as I know (even though yes there are different string actions) the actions is understood at least originally simply from the CLASSICAL ACTION you would expect from a litterally oscillating string. Then this is put in a background and you quantize etc.

The reason what I keep asking this because I sincerely think that there IS a deeper way to understand strings (or a way to at least connect string theory to something else). But this would require a deeper understanding of string actions and background beyond the classical geometric "picture" it started out as.

Maybe this is included in the open issue you already defined, but the basic string itself and the string action is a good starting point.

That's one of the most important conceptional riddles: does a "fundamental" action that would universally describe strings in every corner of the parameter space exist at all? I don't know but I doubt it.

I don't think so either it wasn't what I meant.

I meant that you can only "measure" one theory with respect to another one; by including a renormalized version of the first in the second one in a holographic sense.

But maybe we can in this way understand how theories interact. If I understand you, you also seek a way to understand how say transitions between different theories work, right?

What I am suggesting, and that does connect to the question I asked about the meaning of string actions, is that instead of thinkg in terms of a gigantic state space where you have transitions between theories, maybe the better way is to think of the "transitions" in terms of INTERACTING theories, that are negotiating.

Ie. the transitions are then simply internal revision in the light of new information. There is a good change to connect then the understanding of a string (seen as a simple measure on it's environment) to the foundations of measurement theory.

This means that the "background of the string" is defined by the interaction context (ie. neighbouring strings). But the difference is that, this "background space" only exists from the point of view of the string itself.

Ie if we thinkg of a string as an observer! then the string can "as far as it cna infer" conlude that it lives in this background space, and thus the rational action of the string (defined in the way I SEEK in the original question) is then merely doing a random walk in this effective background.

Transitions from different string theories would then (maybe?) correspong to the string observer remapping it's internal structure, so that giving instnatly "consistent" expectations, it becomes more stable.

What comes to my mind first is to tro "reproduce" or connect the ordinary string actions to some probabilistic measure based on permutations of string configurations - assuming ou can count it, maybe starting with discrete strings?

If such a deeper understanding of the string, and the string action as observers resp rational actions, I think it would be a major boost and it would help solve many questions. It would also force a new way of thinking about this.

Totally relased from the simple "geometrical pictures" you also mention you want to loose.

So the question is, what do we replace that with? I propose the inferentical perspective, but the connection to string seems to be in sight, but yet I'm not sure of anyone works in this direction.

Edit: Thinking in the direction is this http://math.ucr.edu/home/baez/nth_quantization.html. This is related to probabilities of probabilities which in turn related to renormalization of theories.

Could be generate string from something else, that does not come with the ad hoc or classical pictures to it? Something purely inferential?

/Fredrik

 

[marcus]

FOR CONTINUITY since we're on a new page, it may help to carry over some essential posts. This of Suprised was seminal:

I guess there were many potentially wrong turns - at least in the sense of bias towards certain ways of thinking about string theory. Here a partial list of traditional ideas/beliefs/claims that have their merits but that potentially did great damage by providing misleading intuition:
...

Tom's most recent long one was:

I would like to come back to suprised's list regarding possibly wrong turns...

===quote post #554 ===
I would like to come back to suprised's list regarding possibly wrong turns.

1. - That geometric compactification of a higher dimensional theory is a good way to think about the string parameter space
2. - That perturbative quantum and supergravity approximations are a good way to understand string theory
3. - That strings predict susy, or have an intrinsic relation to it (in space-time)
4. - That strings need to compactify first on a CY space and then susy is further broken. That's basically a toy model but tends to be confused with the real thing
5. - That there should be a selection principle somehow favoring "our" vacuum
6. - That a landscape of vacua would be a disaster
7. - That there exists a unique underlying theory
8. - That things like electron mass should be computable from first principles

Let's look at this list again: there is a deep connection between some topics; that's why I was mentioning background independence. I would like to comment on this once more.

String theory walked - for a rather long time - on the trail of particle physics and quantum field theory. Of course there was a graviton, but after recognizing this particle one immediately focussed on QFT-like reasoning (background, strings on top of this background, perturbative quantization, ...). I would say that the first few topics are essentially due to this perception of string theory.

Looking at the field today most researchers are convinced that non-perturbative approaches are required. Thousands of backgrounds / vacua have been identified, but still they are mostly perceived as reasonable backgrounds on which standard particle- or QFT-like theories can be formulated. This is OK for model building an phenomenology (it is not only OK but of course heavily required in order to achieve a closer relation to reality).

But using intuition to find such backgrounds and doing "ordinary physics" on top of these backgrounds does not help in order to understand the relation between these backgrounds and to identify the "unique" and deeper origin of these backgrounds, which I would call the underlying theory.

I think another wrong turn - perhaps the most serious one - would be to turn a bug (the missing unique underlying theory) into a feature (we do not need a unique underlying theory). It would be same as looking at the periodic system and stating that happily there is no underlying theory required as we have a collection of relations between different chemical elements.

I think we do not need to look for a selection principle ("why is it iron instead of copper?"), we do not need to condemn the landscape ("iron, copper, mercury, oxigen, ... is too much; we need a single solution"), we do not need to look for a way to calculate the mass of the electron ("how do we calculate the mass of the mercury atom in a theory which does not explain why there is a mercury atom?"). All what we have to do is to understand what string theory really is. My impression is that we still do not know, we are scratching at the surface, we see some "effective models", not more (and not less).

So 1. - 4. may have been wrong turns - but were overcome somehow over the last years. 5., 6. and 8. are perhaps wrong turns which are in the spotlight today. 7. is not a wrong turn but the essential driving force of progress in physics. I would not abandon it w/o having a worthy successor.

I am still with David Gross (and others - like Weinberg I guess) who asked exactly these questions:

• 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. It is hoped that the recent efforts to construct covariant second quantized string field theories will shed light on this crucial question.
  
• We still do not understand what string theory is.
  We do not have a formulation of the dynamical principle behind ST. All we have is a vast array of dual formulations, most of which are defined by methods for constructing consistent semiclassical (perturbative) expansions about a given background (classical solution).
  
• What is the fundamental formulation of string theory?

Denying the relevance of these questions is - in my opinion - the "wrongest turn ever".
==endquote from Tom's post #554 ==

===quote Suprised reply, post #555===
Nicely said, Tom.

Though I think I should explain what I meant with 7) "there exists a unique underlying theory".
Much could be said here. For the time being, let me provocative and say the following:

Strings seem to be the natural generalization of gauge theory, actually closely related to it by dualities, such as AdS/CFT; in the latter context, strings are indeed reconstructed from gauge theory. So let's view strings as analogous to gauge theory; and then re-ask the same question: "what is the underlying unique theory of gauge theory" ?

Clearly this is a not very fruitful question to ask, because it presupposes something which does not exist, at least in the sense of the question. All there is with gauge theory, are various degreses of freedom that are exposed depending on the energy scale (gluons, quarks, mesons...)

As for strings, the situation is unclear but it may be similar - there may be no further "unique underlying theory". All there might be is the complicated web of perturbative approximations related by dualities, but there is no regime where "universal, more fundamental" degrees of freedom would be liberated.

The real question is whether there is an encompassing, "off-shell" mother theory which would contain all the known theories as "critical points", and describe transitions between them, etc. This may, or may not exist (analogous to gauge theory). So this question is a potential blind ally as well!
==endquote Suprised==

==quote Atyy==
So we don't obviously need Calabi-Yau compactifications?
==endquote==

==Suprised reply to Atyy, post #556==
They are just special examples of vacua, their main advantage is being relatively well under technical control. That's why there has been so much focus on them, unfortunately thereby creating the impression that they would be somehow essential. But there are zillions of other constructions (generalized geometries with fluxes, non-geometric vacua, brane backgrounds, non-perturbative F-theory vacua, M-Theory vacua,... ).

Of course, many of such vacua are equivalent via dualities, and this shows, again, that there is no objective, unambiguous meaning of a compactification geometry.
==endquote==

 

[marcus]

This discussion is plowing deep and turning up the soil in a potentially fertile way. I reflect on on the title of the thread: "Why I am REALLY disappointed about string theory."

What has just now come up, interestingly, are not faults/limitations of theory (as I see it) but deficiencies of "program management". As I hear it, the leadership (funding committees, conference organizers) may have allowed too many "wrong turns"---so that creative talent was wasted on "blind alleys".

So a kind of meta-question would be does Tom's question matter: "Does it matter why experts are disappointed about the string program?"

Or if "disappointed" is too specific, be more general and say experts show a loss of interest, loss of energy, tendency to go off into borderline areas or spend more time in other fields, loss of focus on the hard core problems---some or all these things.

If loss of focus by the best people matters to you, and if it is real, then that looks like a program management problem. Is the string leadership listening enough to what David Gross says, or for that matter, what some people in this thread are saying? Just a thought.

 

[tom.stoer]

Marcus, thanks for the summary.

... not faults/limitations of theory ... but deficiencies of "program management". As I hear it, the leadership ... may have allowed too many "wrong turns"

Hindsight is always wiser; I was listening to a talk of a great QCD guy 20 years ago. His reply to my question how to find the best way to proceed was "how shall I do the calculation if I don't know the result?" Unfortunately this approach is not available in string theory :-(

Is the string leadership listening enough to what David Gross says, or for that matter, what some people in this thread are saying? Just a thought.

They should definately listen to Gross. The problem may simply be to identify a blind spot. In order to achieve that new questions and perspectives are required.

String theory is (in my opinion) in a situation like the strong interaction with all its hadron multiplets but w/o the fundamental representation = w/o quarks. Nice relationships, but no fundamental building block.

My guess is that strings, branes, dimensions and spacetime are only "effective" descriptions valid in certain regimes.

 

[Haelfix]

My guess is that strings, branes, dimensions and spacetime are only "effective" descriptions valid in certain regimes.

That is definitely the state of the art right now. The very real possibility (which I believe Surprised has hinted at) is that this might *always* be the case. It might be that's simply how nature has made her mathematics! Actually, it might be the case for low energy QCD as well. There might simply not be a simple analytic result that humans can package up in a simple way and pretend like it covers the entire energy range perfectly.

Certainly, most of the discoveries about dualities as well as insight into the nonperturbative physics in the last 15 years has followed this road.

Then again, there are so many very intricate mathematical relations and surprises going on within String theory, that I think the original belief that there is some as yet unknown 'super structure' that controls it all is not entirely without merit either.

 

[tom.stoer]

Now that we have consensus (OK, not really, Suprised will not agree) the interesting question is how to identify the underlying theory from which all these effective string models do emerge.

 

[atyy]

Now that we have consensus (OK, not really, Suprised will not agree) the interesting question is how to identify the underlying theory from which all these effective string models do emerge.

As a layman, the main line so far seems to have been that AdS/CFT is the sector in which this underlying theory exists, so let's study it better. The main results in recent years seem to have been about integrability and the ABJM case. In here, there is also the hope that twistors may be a reformulation of the gauge theory which will generalize - Arkani-Hamed even talks about emergent unitarity.

The other line, which is a minority, but very pretty, is the West/Damour, Henneaux, Kleinschmidt, Nicolai work on E10,E11.

I remember Mitchell Porter some time ago pointed to http://arxiv.org/abs/1008.1763 as yet another line trying to formulate the underlying theory.

Naturally, I don't know the relationship between these, or if there are in fact other more important approaches, would be interested to hear from the pros.

 

[Physics Monkey]

It seems to me that holographic duality suggests that there is no simple metatheory. Of course, many of those terms are undefined so who knows if it means anything. I don't usually think of there being some kind of metaframework for gauge theory beyond the basic structures inherent in any quantum field theory, but the existence of a string metatheory along with holographic duality would seem to imply that there is such a metaframework for gauge theory. That would be cool but also surprising in my opinion.

 

[atyy]

But could one hope for non-perturbative definitions of other sectors of the theory in the same spirit?

 

[Calrid]

Now that we have consensus (OK, not really, Suprised will not agree) the interesting question is how to identify the underlying theory from which all these effective string models do emerge.

How about trying an experiment?

Ok just kidding.

That is probably the most salient reason to be disappointed though. No evidence. Disagreements about meta questions are after all just philosophical objections atm.

Nice try but no cigar is the best thing we can say atm.

Peturbative or non peturbative, back ground dependant or not, one thing science is dependent on is discernible reality.

 

[Fra]

If we take a theory to - rather than be some objective description of outcomes of all possible measurements - be one observers inferred expectations of possible measurements it can do - than it seems plausible that two interacting observers is the same thing as two interacting theories, and in addition to that that there is no objective meta theory of how the theories interact. All there can be, is a holographic connection between theories. And that the theories that we do see in nature are somehow the result of some evolutionary selection, just like one can imagine all kinds of crazy by physicall consistent orgnisms on earth, yet the organisms we do see are many but constrained.

There can't be an *inferrable* fixed super meta space of theories. If it exists, it's only in the sense of structural realism.

So my projection of string theory, I think surprised hunch that there may not exists unique timeless eternal mother theory makes perfect sense.

But that doesn't mean it can't exists an evolving meta theory that solves our problems. This evolving meta theory then IS the same thing as what we usually call effectiv theories. I mean it could be that all there is are effective theories. But what is wrongwith that? I see nothing wrong with that. On the contrary; the search beyond effective teories is the search for realism! I was hoping that after a couple of scientific revolusions we was done with that ;) But I was wrong.

Seens as inference, this is just the same thing as acknowledging that there is no ultimate eternal truth. Ie. from the point of view of LEARNING, its' wrong to FOCUS on some ultimate truth. Doing this may in fact inhibit progress. The focus should I think be on learning, without bias of some ultimate truth.

It's the description of this process, I seek. This is exactly what interacting theories is about. So I definitely defend som of these weird things of ST, MY question is merely where the methodology of string research is optimal. Ie. is future string theory the ultimate theory of theory, or do we need to rethink the entire business from scratch?

If I understand this summary right...

Loosely speaking? Many people here except surprised, at least hopes that there will be found some unique mother theory (in order to ST ot make sense)? Is that fair?

/Fredrik