- #1
mitchell porter
Gold Member
- 1,422
- 657
http://arxiv.org/abs/1211.0004
Reflections and Impressionistic Portrait at the Conference "Frontiers Beyond the Standard Model," FTPI, Oct. 2012
Authors: M. Shifman
Abstract: Comments on the results presented at the Conference "Frontiers Beyond the Standard Model," FTPI, Oct. 2012. This summary traces a historical perspective.
As a brief personal view of fundamental physics from the 1970s to now, this talk strikes a nice balance. I disagree with a few details of what he says about string theory, but his opinion is still something I can constructively engage with. And otherwise, there's a lot to learn from and think about here.
Shifman is expert regarding the SM and especially QCD, so there are the usual insights to be gleaned whenever "someone who was there" (when it was discovered) talks about the SM. (Shifman didn't invent the SM but he co-discovered some of its important properties.)
He speaks about supersymmetry, as a theory of the real world, with detachment but also with sympathy and understanding, so there is some objectivity in his remarks. In this regard, I think his talk is a good counterpoint to the discussions we get from inside the supersymmetric research program these days. The Higgs is a little too heavy for what had been the dominant paradigm in MSSM research, and so there's a lot of turmoil and creativity going on. People are thinking about new regions of MSSM parameter space, and extensions of the MSSM like NMSSM. A good current talk or review will discuss the ins and outs of all those arguments.
Shifman positions himself as a sympathetic outsider: he understands the logic of what is going on; the theorists of supersymmetric phenomenology certainly have reasons to be interested in that approach; but he also clearly understands that reality may be work in some very different way. This is also the "stance" that I cultivate, something like "50% orthodoxy, 50% pluralistic heterodoxy", where the percentages indicate how much attention each gets; so it's no surprise that I like Shifman's attitude.
When he gets to string theory, he says (part 8) that 1980s string theory was too ambitious, "promising to give answers to each and every question that arises in the realm of fundamental physics" and "by now, the “theory-of-everything-doers” are in disarray, and a less formal branch of string theory is in crisis"; but also that "string theory, as a qualitative extension of field theory" is rich and a source of insights and we can use it.
This is where I disagree - about the status of the mainstream TOE approach in string theory. The only "crisis" here is an excess of possibilities. Critics of string theory often charge that it could be consistent with anything; but it's not as if we have a known string vacuum that gives the standard model, complete with correct parameter values, that we could then abandon in favor of another, if new physics shows up. String theory has only produced qualitative approximations to the standard model, we're still not at all good at calculating e.g. the masses and mixings in a particular string vacuum, so string theorists still have plenty of solid and respectable work to do.
I think it's worth really thinking about the epicycle analogy for a moment. Epicycles were extra circular motions, wheels within wheels, that could be added to the geocentric model of the solar system, finetuning it to match the details of observed planetary motion. Later came ellipses, Kepler, Galileo, Newton. The "epicycle criticism" of string theory today is that it's an apparatus where you can plug in a new set of epicycles for any conceivable set of observations, so no matter what data we get, there will be a string model, and therefore string theory as such isn't falsifiable, it's "not science".
String theory as it is now is definitely not at this level, at least with respect to the "engineering at will" of specific SM parameter values. But you can use it to engineer supersymmetric GUTs "at will", at a qualitative level. You can get an N=1 SU(5) theory, or an N=2 SO(10) theory... In that sense, the diversity of options lives up to the epicycle criticism. There are still lots of technical constraints on what can and can't be done, all sorts of interesting details about how GUTs work within string theory, but the bottom line is that there are a zillion models to consider, and you can go looking for the sort of model you want: an E6 GUT, with certain discrete symmetries, etc.
But is this bad or a crisis? It might be disappointing to people who first experienced the freedom of model-building in QFT, then witnessed all the miracles of string theory, and who thereby began to hope that, not just realistic-looking GUTs, but a unique vacuum, might materialize as the prediction of string theory. But this was never anything more than induction on miracles: miracles keep happening, therefore maybe a miracle that is convenient to me will happen. It hasn't happened, and the phenomenological situation in string theory therefore resembles the "phenomenological situation in field theory", except that there is a potential for much better predictivity, since the arbitrary real-valued parameters of a QFT are replaced by the low-energy behavior of a specific string vacuum, which in principle are predictable to arbitrary precision. Individual string models are extremely predictive (or they will be once we solve a lot of hard math problems... analogies with QCD are possible here).
Returning to the "qualitative" level where the epicycle analogy has some accuracy to it - where we can go looking, not for a version of the SM where the proton-to-electron mass ratio is 1836, but for a GUT with a desired symmetry group and other desired features - the criticism doesn't have much force, if you understand model-building in particle physics as incremental and proceeding on many fronts at once. Think again of celestial mechanics. If you don't have Newton's laws yet, but you do have the idea of epicycles, is it a bad thing to try to fit the data with epicycles? For that matter, the epicyclic models probably played a role in the discovery of the truth - I haven't studied the history, but it stands to reason.
You don't just stare at the data and have the correct causal model pop into your conscious mind. You work with wrong models and incomplete models, you discover new properties of the data that way, and you come up with your own wrong and incomplete models that others then use in the same way. Even if the phenomenological mainstream in string theory is wrong, it represents a very vigorous and sophisticated engagement with the data, using a mathematical apparatus that is also sophisticated and has good physical pedigree, and so one would have to think that it has a good chance of helping to inspire the correct ideas, even if it is wrong.
And it remains true that some version of the mainstream program may be the truth. I certainly don't have a refutation of the idea that the real world is a heterotic MSSM. When I read older string literature, I keep discovering nuggets of information that must be lore among older researchers, and which must provide positive encouragement for the overall program. For example, "Topics in string phenomenology", dating from 1993, mentions that the standard structure of susy-breaking in the MSSM, should arise generically in a string vacuum with certain properties. It would be really interesting to know how that fact looks now - is it affected by the recent turmoil in supersymmetric field-theoretic phenomenology, or is it untouched by some of the new alternatives under consideration since the observation of the Higgs?
Either way, that's an example of how string phenomenology evolves in response to experiment. It's not like, "oh, the numbers have changed, let's put in a new set of epicycles"; there are much higher-level facts about how string theory works and what it can achieve to be discovered and evaluated. On that level I think string phenomenology is fine. It's hard, it's slow, there are too many options for anyone person to understand, but in that respect it's like the application of QFT to the real world.
Most of this post ended up being about strings, but most of Shifman's talk is not about strings. But this is an example of why it's a stimulating talk. His opinions are highly informed, so even in opposing them, you'll have to dig deep and you should learn something, so I recommend it.
Reflections and Impressionistic Portrait at the Conference "Frontiers Beyond the Standard Model," FTPI, Oct. 2012
Authors: M. Shifman
Abstract: Comments on the results presented at the Conference "Frontiers Beyond the Standard Model," FTPI, Oct. 2012. This summary traces a historical perspective.
As a brief personal view of fundamental physics from the 1970s to now, this talk strikes a nice balance. I disagree with a few details of what he says about string theory, but his opinion is still something I can constructively engage with. And otherwise, there's a lot to learn from and think about here.
Shifman is expert regarding the SM and especially QCD, so there are the usual insights to be gleaned whenever "someone who was there" (when it was discovered) talks about the SM. (Shifman didn't invent the SM but he co-discovered some of its important properties.)
He speaks about supersymmetry, as a theory of the real world, with detachment but also with sympathy and understanding, so there is some objectivity in his remarks. In this regard, I think his talk is a good counterpoint to the discussions we get from inside the supersymmetric research program these days. The Higgs is a little too heavy for what had been the dominant paradigm in MSSM research, and so there's a lot of turmoil and creativity going on. People are thinking about new regions of MSSM parameter space, and extensions of the MSSM like NMSSM. A good current talk or review will discuss the ins and outs of all those arguments.
Shifman positions himself as a sympathetic outsider: he understands the logic of what is going on; the theorists of supersymmetric phenomenology certainly have reasons to be interested in that approach; but he also clearly understands that reality may be work in some very different way. This is also the "stance" that I cultivate, something like "50% orthodoxy, 50% pluralistic heterodoxy", where the percentages indicate how much attention each gets; so it's no surprise that I like Shifman's attitude.
When he gets to string theory, he says (part 8) that 1980s string theory was too ambitious, "promising to give answers to each and every question that arises in the realm of fundamental physics" and "by now, the “theory-of-everything-doers” are in disarray, and a less formal branch of string theory is in crisis"; but also that "string theory, as a qualitative extension of field theory" is rich and a source of insights and we can use it.
This is where I disagree - about the status of the mainstream TOE approach in string theory. The only "crisis" here is an excess of possibilities. Critics of string theory often charge that it could be consistent with anything; but it's not as if we have a known string vacuum that gives the standard model, complete with correct parameter values, that we could then abandon in favor of another, if new physics shows up. String theory has only produced qualitative approximations to the standard model, we're still not at all good at calculating e.g. the masses and mixings in a particular string vacuum, so string theorists still have plenty of solid and respectable work to do.
I think it's worth really thinking about the epicycle analogy for a moment. Epicycles were extra circular motions, wheels within wheels, that could be added to the geocentric model of the solar system, finetuning it to match the details of observed planetary motion. Later came ellipses, Kepler, Galileo, Newton. The "epicycle criticism" of string theory today is that it's an apparatus where you can plug in a new set of epicycles for any conceivable set of observations, so no matter what data we get, there will be a string model, and therefore string theory as such isn't falsifiable, it's "not science".
String theory as it is now is definitely not at this level, at least with respect to the "engineering at will" of specific SM parameter values. But you can use it to engineer supersymmetric GUTs "at will", at a qualitative level. You can get an N=1 SU(5) theory, or an N=2 SO(10) theory... In that sense, the diversity of options lives up to the epicycle criticism. There are still lots of technical constraints on what can and can't be done, all sorts of interesting details about how GUTs work within string theory, but the bottom line is that there are a zillion models to consider, and you can go looking for the sort of model you want: an E6 GUT, with certain discrete symmetries, etc.
But is this bad or a crisis? It might be disappointing to people who first experienced the freedom of model-building in QFT, then witnessed all the miracles of string theory, and who thereby began to hope that, not just realistic-looking GUTs, but a unique vacuum, might materialize as the prediction of string theory. But this was never anything more than induction on miracles: miracles keep happening, therefore maybe a miracle that is convenient to me will happen. It hasn't happened, and the phenomenological situation in string theory therefore resembles the "phenomenological situation in field theory", except that there is a potential for much better predictivity, since the arbitrary real-valued parameters of a QFT are replaced by the low-energy behavior of a specific string vacuum, which in principle are predictable to arbitrary precision. Individual string models are extremely predictive (or they will be once we solve a lot of hard math problems... analogies with QCD are possible here).
Returning to the "qualitative" level where the epicycle analogy has some accuracy to it - where we can go looking, not for a version of the SM where the proton-to-electron mass ratio is 1836, but for a GUT with a desired symmetry group and other desired features - the criticism doesn't have much force, if you understand model-building in particle physics as incremental and proceeding on many fronts at once. Think again of celestial mechanics. If you don't have Newton's laws yet, but you do have the idea of epicycles, is it a bad thing to try to fit the data with epicycles? For that matter, the epicyclic models probably played a role in the discovery of the truth - I haven't studied the history, but it stands to reason.
You don't just stare at the data and have the correct causal model pop into your conscious mind. You work with wrong models and incomplete models, you discover new properties of the data that way, and you come up with your own wrong and incomplete models that others then use in the same way. Even if the phenomenological mainstream in string theory is wrong, it represents a very vigorous and sophisticated engagement with the data, using a mathematical apparatus that is also sophisticated and has good physical pedigree, and so one would have to think that it has a good chance of helping to inspire the correct ideas, even if it is wrong.
And it remains true that some version of the mainstream program may be the truth. I certainly don't have a refutation of the idea that the real world is a heterotic MSSM. When I read older string literature, I keep discovering nuggets of information that must be lore among older researchers, and which must provide positive encouragement for the overall program. For example, "Topics in string phenomenology", dating from 1993, mentions that the standard structure of susy-breaking in the MSSM, should arise generically in a string vacuum with certain properties. It would be really interesting to know how that fact looks now - is it affected by the recent turmoil in supersymmetric field-theoretic phenomenology, or is it untouched by some of the new alternatives under consideration since the observation of the Higgs?
Either way, that's an example of how string phenomenology evolves in response to experiment. It's not like, "oh, the numbers have changed, let's put in a new set of epicycles"; there are much higher-level facts about how string theory works and what it can achieve to be discovered and evaluated. On that level I think string phenomenology is fine. It's hard, it's slow, there are too many options for anyone person to understand, but in that respect it's like the application of QFT to the real world.
Most of this post ended up being about strings, but most of Shifman's talk is not about strings. But this is an example of why it's a stimulating talk. His opinions are highly informed, so even in opposing them, you'll have to dig deep and you should learn something, so I recommend it.