Why I am REALLY disappointed about string theory

[marcus]

I'm glad you said "similar" Atyy An equilateral triangle that is 2 inches on a side is similar to an equilateral triangle that is one mile on each side. But I would not call them "equivalent".

Where is Lieutenant Dax? Methinks she was well-spoken and made interesting points. I hope she decides to rejoin us.

 

[mitchell porter]

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

If you look at compactifications of M-theory on a torus, the more compact dimensions there are, the bigger the U-duality group. Since for a torus with dimension n $$\geq$$ 6, the U-duality group is just a form of E_n, and since it is speculated that the master symmetry of M-theory is E₁₀ or E₁₁, and also that the dualities may be derived from M-brane worldvolume symmetries... this growth of U-duality seems to be a migration from explicit string/brane symmetries you could see on a fixed background, to implicit symmetries which require a change of variables and/or background to be demonstrated. So you might say that any phenomenologically relevant solution of string theory is going to be low-dimensional, therefore, it will have a large U-duality group, because most of the symmetries will have become "implicit". (Though remember we might also be on a braneworld in a high-dimensional bulk space.)

However, these results apply to solutions with a large or even maximal number of supersymmetries, and something else about the real world is that you expect supersymmetry to be broken, probably completely. In terms of the explicit+implicit framework I just gave, this might mean that the total number of "remaining explicit symmetries + dualities" is small. Actually I'm just confused about the relationship between supersymmetry and the dualities. Solutions with maximal supersymmetry have been useful in understanding the nonperturbative dualities, but that's because they're easier to analyse. I simply don't know whether breaking supersymmetry also means reducing the "total amount of non-susy M symmetry", whatever that means, or whether the "susy and non-susy parts of M symmetry" can be broken independently.

If you have a preferred GUT model, you could also just look for string models which contain it, and find out afterwards how they look from the fundamental perspective.

 

[suprised]

As if precisely these points haven't been spelled out here N times... there seems again a confusion about the meaning of dualities. In fact there are two main meanings that need to be separated.

In the strict sense, duality refers to different descriptions or parametrizations of the very same theory. Since there is just one theory, there is no issue of "observing or not the others".

In a wider sense, duality refers to different descriptions or parametrizations of the same theory, but related to different parameter values of the same theory. Again, observing "other" theories is an ill-defined question.

 

[arivero]

In the strict sense, duality refers to different descriptions or parametrizations of the very same theory. Since there is just one theory, there is no issue of "observing or not the others".

You are right! But it is very easy to forget the point, as I did, because each of these different desctiptions have a different set of massless states. In a duality, the fundamental states of a "parametrization" are mapped nonperturbatibely to massive, excited or solitonic states of the other, and reciprocally.

I can imagine one or two such maps in our particle spectrum, particularly for the top, which is massive while all their cousins are massless, and for the neutrinos, which have a light Dirac mass and a heavy "majorana" mass for the see-saw. But this should allow for a couple of different "parametrizations", not a whole web of them.

 

[arivero]

Addendum, let me put an extra bit in the stack of experimental evidence: it is well known (or well neglected) that in the experimental spectrum there are 84 almost massless fermionic states. They should be protected by some symmetry. Opening at random the Slansky report, I can see 84 in SU(4) (with triality!), SU(6), SO(9),... and I could also look for 42 (hattip Douglas Adams) or 21. So it does not seem a big clue. But the source of the 2-brane of M-theory is the antisymmetric tensor of 84 components, the complement of the 11D graviton (44+84=128) in the N=1 sugra fundamental multiplet. Thus I'd say that the M-theory brane is a candidate to protect the Yukawa couplings of the fermions, in some yet unknown parametrisation of a yet unknown compactification.

 

[Lt_Dax]

Where is Lieutenant Dax? Methinks she was well-spoken and made interesting points. I hope she decides to rejoin us.

She made some interesting points, but is still too inexperienced to really get into the thick of it (some of the mathematical terminology others use makes it tough for me to extract the arguments people are making).

One thing I did pick up on from a response to my original post: I'm still skeptical that it is possible to actually learn things about the universe without "interrogating it".

It is probably possible to learn new physics by developing mathematics in ways which are consistent with the rest of mathematics (which is more a definition of mathematical proof than physical proof), but I can't avoid the feeling that the parts of mathematics which tell us real things about the universe actually originate from observation in the first place. For example, we can develop lots of new physics starting from the assumption that space is Euclidean, but someone made that assumption from observing the behaviour of lines and triangles, real objects (of course we often have to revise the assumption as our knowledge becomes more sophisticated).

String theory could be described like this (having the lofty goal of being self consistent, consistent with the rest of mathematics, and in principle at least, predictive), but my problem is that its fundamental premise hasn't been established. It is probably possible to "predict" the entire standard model by using any fundamental object as a starting point. The string idea still relies on the assumption that the only alternative to a point object is a quantized oscillating string (the wave "paradigm"). Is this a failure of imagination?

So even if string theory can predict, say, the entire known particle spectrum, or anything else we already know about, I'm unmoved by it. Does what I'm saying make sense?

Some string people say that string theory solves the problem of unification, but that assumes that unification is a problem (it might not be). When they say that string theory "predicts gravity", I assume they mean the same thing - that gravity must be part of a completely unified theory which satisfies our own biases about what a unified model would look like. Even if we could develop a quantum theory of gravity, there could be ten other fundamental forces we don't know about. Will string theory predict those? Can we unlock deep secrets about the universe with a pen and paper? I'm highly skeptical.

 

[marcus]

Lieutenant, one thing we could do is start a new thread, with copies of several of your posts, called "The Dax Discussions" and let people reply to you in that thread, and comment on the issues you raise.

That would allow Tom, Arivero, Mitchell, Surprised, and the others to stay focused on the ground they've been plowing so far.

I'm not PF staff, so I can't make a judgement call and move posts and split off a piece of a thread. But my personal inclination is to let them have uninterrupted technical discussion and reply to you in a separate thread. I think your opening posts are engaging.

Maybe what I'll do as a private "on spec" initiative is start such a thread, then if you don't like that you can simply not participate and the thread will die, if Tom does not like that and wants to keep Dax discussions here, he simply has to say and we can abandon the split-off thread.
I'll wait a few hours and see if there is any immediate reaction to that idea.

 

[Lt_Dax]

I'd support that idea marcus. If you start a new thread, I'll participate when I can. It doesn't matter if the thread dies, it's worth it just to find out if anyone else is interested in discussing my points.

 

[mitchell porter]

Speaking of new threads, I'm going to start one for Alejandro Rivero's idea in comment #425. I don't think the number of degrees of freedom in 11 dimensions is much of a clue for phenomenology, because moving to lower dimensions creates so many new states and relationships. But it would be a good exercise for interested parties to really think this through, and the technicalities might interfere with the discussion here.

edit: https://www.physicsforums.com/showthread.php?p=2983996"

 

[friend]

One of the complaints about string theory is that we really don't know what the theory is, what it's basic premises are. Let me offer a suggestion.

Could the basic premise of String Theory/ M-Theory be that, "the concepts of the Lagrangian and the Action are applicable to sets as well as points"? Then those sets might come in the form of open and closed strings and higher dimensional branes.

Just like scalars and vectors and matrices are generalized by tensors, perhaps branes are generalizations of strings and points. It would be an interesting mathematical study to prove that the Feynman path integral also applies to "sets" as well as to point particles, whatever that means. Perhaps it also applies to sets of discrete point as well as to continuous lines and branes. Or has this already be proposed?

 

[tom.stoer]

This is true from the perspective of "first quantization" where the action of world lines of point particles / of world sheets of strings is promoted to the path integral.
But we know that in quantum field theory we have to describe nature not via first quantized point particles but via "second quantized" fields.

I am confused b/c there is still no understanding of string field theory which would do the same with strings.

 

[arivero]

Other question is if the string theory we are dissapointed about, is the same string theory that people is studying. Check for instance this list of recent seminars.

http://www.talks.cam.ac.uk/show/archive/18332

 

[tom.stoer]

Can you give me a hint what exactly I should look for?

 

[arivero]

Nothing particular, just the feeling, arguable I expect, that the topics studied there doesn't seem to fit with the topics usually discussed in this subforum.

 

[elivil]

@tom.stoer

I've been following this thread on an on-and-off basis for some time but it's really hard to navigate in the space of >400 posts. In your post #333 you suggested writing a short summary of identified problems and questions. I think it would be indeed great to have such a summary and it would help young physics students to cut through the hype from both sides if there were a technical summary dealing with actual physical questions that are investigated at the moment.

From my part, I can add that there is a substantial interest in AdS/CFT correspondence as applied to different problems (including outside of particle physics). At my university recently there was a short introductory course into AdS/CFT (or, to be more precise, better to call it "gauge/gravity duality" since after Maldacena's first paper other instances of similar dualities to AdS/CFT have been identified) and this course was well attended by the members of the condensed matter groups and even some senior undergrads (have to 'fess up:). So that offshoot of string theory is growing strongly and with results from LHC (q-g plasma etc) will have at least some hope of its predictions being experimentally verified.

 

[Careful]

Careful! Nice to see you after a 2 year absence! I saw you thanked in the acknowledgments section of a QG paper recently for helpful discussions with the authors---glad to see that indication of your continued interest and activity in QG.

Well, I have been busy... :-) Ah, if you have a good idea how to solve a not so trivial problem and you are not interested in publishing it yourself (because you think the general idea fails in a deeper way), then you just give it away in case the author happens to be a nice, intelligent and open minded person. That's how you make friends. I am not interested in publications, but in solving the problem (and I happen to disagree with almost everyone on rather substantial things). So, if I would publish in the (near) future, you know the content of the paper

 

[Careful]

Here Matilde's Caltech page
http://www.its.caltech.edu/~matilde/
with a nice picture of her and a description of her research interests.
Here's a sample paper:
http://arxiv.org/abs/1005.1057
Spin Foams and Noncommutative Geometry
Domenic Denicola (Caltech), Matilde Marcolli (Caltech), Ahmad Zainy al-Yasry (ICTP)
48 pages, 30 figures
(Submitted on 6 May 2010)
".

Ok, to actually answer the content of your message. My understanding here is that some kind of landscape problem is unavoidable in *any* approach to quantum gravity. NCG isn't quantized yet, so God knows how many free parameters, new particles, symmetries will be necessary to make the whole thing consistent. It is always like that with unification ... the number of possibilities goes up and you will have to figure out new types of boundary conditions on your theory restricting the number of solutions/kind of physics drastically. So, in that sense, string theorists are correct beyond reasonable doubt that any approach will face a landscape problem. This will weaken 'predictivity' of your theory: the same thing happened already with for example general relativity. It is never said in this way, but you could easily hold the point of view that for example the precession of mercury is *not* a prediction of GR. It is just a possiblity ... if I were to turn on some gravitational waves so strong that they would precisely knock mercury out of its regular orbit (at some point in space and time), that would be a false prediction of GR! The same goes with quantum mechanics, we really never know the state of the system and have to make lot's of assumptions about (a) decoupling of the system under study from the evironment (b) the precise form of the wavepackages and so on... There are so many possible choices that eventually everything can be fit. The question of course is, are these assumptions 'natural' ? What do we mean with that? For example is it natural to assume in GR that post Newtonian corrections are the good thing to calculate? Isn't it just psychology because we think Newton must be valid everywhere in the universe (which it isn't because of the 'dark matter' puzzle) to some high degree of accuracy? From the point of GR, this doesn't *explain* Newton unless you somehow find a criterion why the universe must almost be spatially flat.

Unification will make the possible worlds even crazier and string theory is so far the only candidate which did manage to even adress this issue. I think we will have to learn to live with theories where we cannot really predict the future anymore unless we fix lots of boundary conditions. If you want to have a theory of boundary values, fine! That's the next step. But Connes and co did not *predict* the standard model either, they almost have put it in by hand.

 

[rdjesch]

I'm quoting Leonard Suskind from his book.
"Elegant theories have more beauty depending on the lowest number of defining equations. Therefore String theory is the most beautiful theory because it has exactly zero defining equations."

 

[tom.stoer]

wow!

 

[rdjesch]

I guess Mr Susskind may have been joking in his book "The Cosmic Landscape."

Quoted correctly "A beautiful theory is one with a few elegant defining equations. By these standards, String theory is the most beautiful. String theory still does not have a single defining equation."
I think I understand a little about the too many landscapes problem but the book was very sparse on explanations or mathematics.
But seriously, Is there some way of using string theory to predict something?
How does it work?
I'm trying to decide if I should devote more energy into understanding this.

Please help.
Where would you start?

 

[suprised]

Please help.
Where would you start?

A good idea would actually be to start reading this thread, rather than posting meaningless stuff!

 

[tom.stoer]

@tom.stoer

I've been following this thread on an on-and-off basis for some time but it's really hard to navigate in the space of >400 posts. In your post #333 you suggested writing a short summary of identified problems and questions. I think it would be indeed great to have such a summary and it would help young physics students to cut through the hype from both sides if there were a technical summary dealing with actual physical questions that are investigated at the moment.

I think the only thing we can do here is to write a summary regarding the discussion and the conclusions in this thread. For a full review regarding string theory (current status, research directions, open issues, ...) we have to find independent review articles.

 

[Careful]

I guess Mr Susskind may have been joking in his book "The Cosmic Landscape."

Quoted correctly "A beautiful theory is one with a few elegant defining equations. By these standards, String theory is the most beautiful. String theory still does not have a single defining equation."
I think I understand a little about the too many landscapes problem but the book was very sparse on explanations or mathematics.
But seriously, Is there some way of using string theory to predict something?
How does it work?
I'm trying to decide if I should devote more energy into understanding this.

Please help.
Where would you start?

Of course, Susskind was joking: a theory without defining equations simply does not exist, hence string theory doesn't exist - whether the theoretical vacuum is a beautiful thing or not, I leave that to the philosophers I think if you read my post about the more modern meaning of what it means to ''predict something'' you should get a clearer idea. The discussion is not whether there will be a landscape or not, just how big the f*cking thing has to be.

 

[lumidek]

The problem with the theory is that it thinks the universe does calculus every time. In genetics there are only four base units and that creates all living things. What if there is something simple that creates all of matter from energy?

Hasn't someone already told you that it's not your problem with string theory but rather your problem with all of physics? It was Isaac Newton who first figured out that the Universe does calculus every time. To show that it was the case, he had to single-handedly invent or discover the calculus, too. ;-)

 

[rdjesch]

A good idea would actually be to start reading this thread, rather than posting meaningless stuff!

Thank you, and I do mean that sincerely. I guess somebody let me out of my cage. I don't know what came over me. For sure I'm a mega junior compared to most of you on this list. I apologize. There must be something done about this mess and this thread seemed to have some potential to do it.

Meaningless... Hmmmm ... do you actually read most of these posts?
Some out of context quotes may surprise you because of how much you can ignore.

Just trying to focus on the headliner of this thread:
"Why I am REALLY disappointed about string theory"

Have you contributed today?

 

[suprised]

I guess I have contributed my fair share. The first part of my remark was about your question: "But seriously, Is there some way of using string theory to predict something?"
This had been discussed at length here; and not just once! We can't repeat this again and again simply because people don't want to spend some effort in reading. Though I admit that a forum is not a suitable medium to confer this, as information gets incoherently presented, diluted and mixed with desinformation.

The second part was aimed at your Susskind quote; for what was this good for?

 

[MTd2]

The second part was aimed at your Susskind quote; for what was this good for?

To show that string theorists are getting the increasing impression by many people outside this field of being paranoids, having delusions of grandeur and being cranks.

 

[suprised]

Sigh... I was hoping to have this thread concentrate on science and avoid sociology. Seems impossible.

 

[Greg Bernhardt]

first call to get this thread back on track

 

[tom.stoer]

As I wrote in #333:

I am afraid I can't do more than indicate what the central problems and questions are which have been identified throughout the discussion (to be honest, I don't think that we found out something new; we only collected facts and questions well-known to the experts). ... It could make sense to write a short summary and conclude this thread instead of reiterate and spin in circles.

Anybody there to summarize from a string theory perspective what the essential conclusions are?

 

[mitchell porter]

There's a great little introduction to M-branes on the arxiv today http://arxiv.org/abs/1012.0459" .

There are two main perspectives we can take on extended objects such as the membrane and fivebrane. We can look at them as solutions of 11-dimensional supergravity (these solutions will also have near horizon limits) and look at the field theories on their worldvolumes. This is at the heart of the AdS/CFT correspondence. The degrees of freedom on the worldvolume are goldstone modes from broken symmetries, including supersymmetries. Requiring that the Bosonic and Fermionic degrees of freedom match to give a supersymmetric worldvolume theory puts very strong constrains on the allowed extended objects, importantly the maximal dimension this can occur in is 11. Here the 8 scalars from broken translations in the directions transverse to the brane match with 8 Fermions from the broken supersymmetry. A fivebrane thus has only 5 scalars but will still have 8 Fermions if it preserves half the supersymmetry. The three additional Bosonic degrees of freedom come from broken gauge symmetries of the three-form C. This leads to a 2-form with anti-self-dual field strength on the fivebrane worldvolume. This makes the fivebrane worldvolume theory difficult to formulate. (Chapter 3)

 

[mitchell porter]

Another addition to this old discussion about the fundamental formulation of string theory:

I have been studying the AdS/CFT correspondence, and I am only now realizing how central it is to answering that question! d=4 N=4 super-Yang-Mills theory is believed to be exactly equivalent to Type IIB string theory on AdS₅ x S⁵. The fifth, "AdS" dimension; the five further compact dimensions; string states and brane states in this ten-dimensional space - they're all entirely constructible from operators in the four-dimensional theory, which lives on the boundary of the AdS space. So if you want to understand Type IIB string theory - at least on an AdS background, which I admit is an unusual space - study this four-dimensional super-gauge theory!

As for M theory, the three-dimensional http://arxiv.org/abs/0806.1218" is a recent review of one aspect of ABJM (when it talks about Type IIA on AdS₄ x CP³ as the dual theory, that's the same as the M-theory description, but with the eleventh dimension removed from the S⁷ so it becomes CP³), and it makes it clear how much struggle and hard work is involved in extracting each increment of additional insight from these theories, even though we can write down the whole equation.

 

[Calrid]

I personally have nothing against string theory even though of course technically it is not a scientific theory or even a very good hypothesis yet as it makes no measurable predictions.

The only thing that disappoints me is the fact that it seems to be very esoteric so only String Theorists can understand it, and hence it seems to be defended with a sort of religious zealotry, it is not as open as one would like, perhaps that is something that will change with evidence when theory meets experiment. Would be fine if it was dark/energy matter theory, that has some sort of evidence although nothing concrete. But I fear strings is putting the cart before the horse, and M-theory the car before the horse, and I don't think that is good science if it remains untestable, for well the foreseeable future at least.

As long as Scientists understand where philosophy ends and science begins though then I am fine with the whole thing. It should really be in maths departments if it is not applied (and again it can be if not to a ToE or gravitation as yet). I think mostly it is good science - if it tackles applications to science such as quantum chromodynamics issues at least - in a model and or environmental predictions or materials science.

It's a promising idea but as yet that is all it is, the disappointment probably comes from expecting too much from too little if you see what I mean.

Anyway this is mostly a popcorn post. I am by no means an expert on this subject. I just find the "theory" interesting.

 

[tom.stoer]

@mitchell porter: thanks for these interesting ideas. It could very well be that this is a direction where progress can and should be made. What you are saying - better: what I understand from it - would mean that:
- there may be a full, non-perturbative formulation on a certain class of backgrounds
- there may be a duality between strings and SUSY-gauge-theories
- there is still no selection principle for this special background (or class of backgrounds)

If this is correct, then the following questions apply:
- why a certain class of backgrounds / topological sector? unfair questions: nobody asks this for QED or GR
- why strings and not only SUSY-gauge-theory?

Perhaps string theory is only a different perspective to look at the same thing

 

[Physics Monkey]

The only thing that disappoints me is the fact that it seems to be very esoteric so only String Theorists can understand it, and hence it seems to be defended with a sort of religious zealotry, it is not as open as one would like, perhaps that is something that will change with evidence when theory meets experiment.

Have you ever tried looking at Barton Zwiebach's book "A First Course in String Theory"? Granted, this book is a lot about the classical mechanics of relativistic strings and membranes, but it really does give some nice intuition for those subjects. One can learn a surprising amount about certain solutions of string theory just using classical mechanics (SUSY helps make this true).

And speaking personally, I've found that string theorists are actually among the friendliest and most open groups in physics. The people I know are generally a pretty laid back group, far from zealots, and I suspect its partially because a lack of direct experimental contact forced them to adopt a more conciliatory stance. Of course, my experience could be limited.