Today’s “Science Saturday” on Bloggingheads features me and Sabine Hossenfelder, supposedly talking about What’s wrong with string theory. Actually, we both agreed that we were pretty tired of that topic, so tried to discuss some more interesting related issues we both have an interest in. Here’s a clip from the full thing, I promise to not start regularly embedding video in this blog:
I hope this thing came out all right. It was recorded a couple weeks ago, in a process involving no trouble on my end, but heroic efforts on Sabine’s. While Sabine had to set the whole thing up on her end, and ended up crouched in an attic since it was the only place she could get a connection of good enough quality, I just sat at my office chair and someone from Bloggingheads took care of everything. Unfortunately we couldn’t see each other while talking. I can’t really bear the thought of watching myself on video, so I guess I’ll never see exactly how this turned out, but I’m glad to see that the turtles on the bookcase behind me made it into the frame.
Sabine has her own posting about this here, and the full thing is here.
I just made the heroic effort of indeed watching the whole thing again! It is actually quite good, we managed to cover a lot of interesting topics. I was as well somewhat confused by the title the bloggingheaders chose.
Peter,
This is a great video blog but I have a few comments regarding “the market place of ideas”. First, any marketplace of ideas as I see it (like an economic marketplace) is an open market – open to exchanges by any and all individuals like the free market that we currently barter in for goods and services. This is almost universally seen as the reason free markets are successful – they are “open” to exchanges by many without much intervention on control (although we understand that some intervention may be necessary). To exclude the public from the string theory debate seems counter this market philosphy, as it is the public that seems decides our fate in terms of economics and political process. The beauty of our system in this and other free countries is that regardless of how bad a product is, or how limited the intelligence of a political candidate, the system manages to correct itself through some magical feedback loop (bad product –> unhappy customer –> product is not purchased and company goes bankrupt). This system is probably what keeps us a float even when the most limited intelligences occupy positions of power in our government or economic system.
This being said, I believe that the public should (and always will – we have no choice in a free market) be involved in the string theory debate. The last time we may have had a true vacuum in science was when a certain oligarchy taught and enforced without debate to a select few ideas of about an earth-centered universe. We have fortunately moved away from this, but like all dynamical systems that we rely on, cycles are the inevitable result and we are perhaps in a down cycle in terms of the “idea marketplace”. The solution to getting out of this down cycle, as the government well knows, is to stimulate the economy, and for ideas it’s the same. We stimulate the “idea economy” by involving the public, and earnestly hope as we do every election year that the right idea or candidate floats to the top of the pile over some time. As part of that process, we have to tolerate the ideas we don’t like and because that feedback loop that finds a happy medium for all of us, in the end things seem to get better for everybody (GDP in this country is still pretty good).
Therefore, I don’t think there is anything to be feared about a public debate – the lack of understanding of QFT or string theory is irrelevant. Ideas that survive the market place rise to the top, ideas that don’t sink to the bottom. In the end, the public becomes more educated (which is more than they might get in public schools in some cases) and it’s better to have a somewhat educated public voting on science funding than a non-educated public doing the same. In fact, it seems like there is little input from the public on science funding and the more interested they get in the debate, the more likely they are to let congress know about it. Vacuum’s of knowledge never kept theories alive nor did they stimulate growth. Only open debate – not just academic debate (which as you both mentioned doesn’t seem to be happening anyway) but open public debate, can bring us to an understanding of the truth.
Hi Mike,
I have written two long posts on the topic that you touch: We have only ourselves to judge on each other and The Marketplace of Ideas that I recommend you read. What you miss is the following:
a) Especially in theoretical physics, there is potentially a long time in which there is no evidence (in form of observation or experiment) to judge on the promise of a research program.
b) During this time then when judgement by Nature is still absent, judgment is left to experts in peer review – which is thus essential for ideas to survive and grow to adulthood. The public opinion should play no role in this, this isn’t the way science works. The public opinion is relevant to the question what fields are generally of importance for our societies and how money should be invested to be as beneficial as possible, but the public never ‘votes’ on research funding on the specialist level. There is a reason why funding agencies have an elaborate referee process.
There is a priori no problem with leading these discussions publicly, as long as this does not influence the truth finding process. But the danger of such influence when people worry about their funding is very obviously there. It should also give us a lot to think that apparently in the case of string theory it was not possible to lead this discussion without exerting public pressure (or possibly it wasn’t tried hard enough?). This is my opinion shows that the judgement process on our own ‘Marketplace of Ideas’ works very insufficiently.
Best,
B.
Mike,
I think I, like Sabine, have mixed feelings about this whole “marketplace of ideas” business. What I’ve seen in particle theory over the past 20 years has been a massive failure in how that particular marketplace is supposed to work, and that was one of the main concerns of my book and Lee Smolin’s.
There are different marketplaces at work here, an internal marketplace inside string theory, where different proposals within string theory battle it out, a marketplace within particle theory, where the traditional lines are phenomenologist vs. string theorists (although with many trying to cross those lines to gain marketplace advantage) and the marketplace within physics departments, where decisions are make about who to hire. This is a relatively closed system, and I don’t think the public has much of a role to play there, since good decisions require significant technical expertise.
There’s also a public marketplace of scientific ideas, and string theory has traditionally been extremely successful there, due to the impressive public relations efforts of some of its proponents. This has changed the last few years, with the public seeing a much more balanced range of views, and string theory losing market share. String theorists are naturally upset about this now, while they were relatively pleased with how things were going back when string theory was being heavily promoted, and its problems pretty much ignored. When my book and Smolin’s first came out, one frequent claim made was that the books should be ignored, since string theory had already triumphed in the marketplace of ideas. The books were very much an effort to engage in this marketplace, and I think did so quite successfully.
The relation between these two marketplaces is a complicated one, and that’s something Sabine was repeatedly asking me about. One of her questions was why I chose to enter the public marketplace, as opposed to sticking to the professional marketplace. One answer is that the book was intended to address both, aimed at both professional physicists and mathematicians (who generally know very little about what the state of string theory is), as well as the public. Both marketplaces needed a countervailing force to the promotional effort that has been backing string theory.
The public marketplace affects the professional one, as ideas that get public attention get more funding from various sources and more support from university administrations. I wouldn’t want this public attention to determine the fate of a scientific research program, but some influence is unavoidable. In the past string theory was probably better funded and supported because of the promotion to the public. The professional marketplace affects the public one too, largely through the popular science press, which observes the professional marketplace, tries to evaluate what is new and how things are going in a science, and then transmits some understanding of this to the public. Part of the most serious problem for string theorists now is that a sizable and vocal contingent amongst them has taken to promoting a research program (landscape anthropics), which can be readily explained both to their colleagues and to the public (anthropic papers tend to have few equations…), but most people recognize pseudo-science when they see if. If you’re not a string theory ideolog, this stuff looks extremely dubious.
I am watching as I write this, and I was surprised to hear you agree with her statement that it is bad to make arguments about String Theory to the general public, because it polarizes people. What does that mean — that you regret writing the book? That laymen should not know about Physics controversies? That it is bad to express polarizing opinions? C’mon, you ought to defend yourself better than that.
Peter,
I also think that both your book and Lee Smolin’s did well in countering the string theory promotion that went on so long in the public – they rose to a high level of public opinion with no tangible evidence to show. Both books did a good job in countering string theory claims of TOE and I think the public market place was a better venue to have this discussion than professional journals because there appeared to be a bias in the journals towards “the new trend” without giving it much scrutiny. It’s always good to get a public sanity check once in a while – it’s not that I think the public will help comment immediately on formulas or theories, but just that they act as a sounding board and “ethical referee” check for professional scientists who engage in a debate where the cards are clearly stacked on one side. It’s kind of like a jury of one’s peers – they aren’t supposed to be experts in forensics or other specialties because a bias may exist – but they still are presented with highly technical facts in which they use to make a decision. In the end, it leads to a rather unbiased opinion (the fact that lawyers trust the average person on the street to make a decision about medical evidence is a fascinating concept). Perhaps this is the advantage of science in the public – just to keep what we know as science and the scientific method going and to call pseudoscience what it is. That being said, I am definitely not in favor of landscape anthropics which appears to be epicycles all over again (but with more advanced mathematics). I think the sooner these concepts of “anything is possible” are examined in the public view, the faster they will go away.
Hi Roger,
I did not say it is bad to make arguments about string theory (or any other scientific topic for that matter) to the public. I said it is of great concern this was necessary, since doing so has predictable drawbacks eg, as you say, in that it polarizes people. Uncertainty is an elementary ingredient of science, but there seems to be no place for this in public discussions which usually forces people to take either side, leaving out detailed nuances. This is not a good environment in which to discuss whether the present research directions are pursued in a balanced way, and if not what the reason is for this. Instead, it will prompt a food fight. Best,
B.
Roger,
I’m obviously not opposed to communicating with the general public about string theory and its problems, since I’ve done a lot of it. I think that exchange was more about how it was a shame that discussion of the problems of string theory within the theoretical physics community was often suppressed (string theorist referees stopped publication of an earlier, less popular, version of my book by a university press), with the result that it ended up taking place in public forums not necessarily known for civility or a high standard of informed professional discourse.
Hi Bee,
I definitely don’t advocate that the public take over the technical referee process – there’s obvious reasons why this wouldn’t work, but it appears at times (like when Peter submitted an earlier book to the University press that got trounced by string theorists bent on keeping their shrine sacred) that professional peer-reviewers stop acting professionally when they let their personal opinions get involved of what should be an unbiased observation of facts. During these times it seems many who feel the peer review process is biased seek the public domain and in the case of string theory this rightly seemed to correct the problem. I don’t believe this is because the public decided string theory was bad, but I believe it because the embarrasment of potentially being proven wrong in the public domain about a theory is much worse than anything that happens in the peer review process and therefore has the effect of correcting peer-review so it becomes less biased and more effective.
This also affects funding to a certain extent as you mentioned in one of your posts about the funding for string theory starting to dry up. What is popular does get funded and I agree with you that this is a real problem. Perhaps some funding should be set aside for “fringe theories” so that it’s not always “right or wrong” to fund something. This is already taking place, like Naval research funding cold fusion well into the last decade, or NASA funding the blacklight propulsion project a few years ago. These funding sources take it upon themselves to ensure the research gets done even if it’s unpopular. I sometimes take issue with funding some of these projects but as in a democratic system, I believe we have to tolerate some things we may not like so we can meet common ground and create a better environment for research overall.
I agree with your ideas about what happens in the market – it’s not a “pure process” and I don’t think the market place is the ideal place to bring about change in science, but sometimes it may be necessary if the peer-review process isn’t working as it should (which then hopefully gets corrected). In mathematic terms, the relationship between science and the public seems to be non-linearly dependent.
Best regards,
Mike
What is the problem with “public forums not necessarily known for civility”? You mean that if you denigrate String Theory, then Lubos Motl will write nasty things about you on his blog? He is apt to do that whether you write for the general public or not. You don’t need to give excuses for writing for the general public.
The reason why only the public debate has gotten any traction, and why the attitude of the particle/gravity community to string theory is the same before and after the books, is simple: there were no new scientific ideas in either book that was not already well known.
Woit and Smolin did a good job of counterbalancing the Kaku style nonsense that was being promoted to the public as string theory. Many string theorists (eg: myself) would have liked nothing better. But the relative merits and difficulties of string theory were well known to string theorists before the books, and the books did nothing to affect that scientific judgement.
Also, discussions even about well-known things usually get caught in ruts – the typical example is that of the landscape. The oft-repeated claim that “having many vacua which look like ours at low energies kills predictability” is simply an incorrect one, and no sensible discussion about the landscape is possible when this sort of factual errors are not acknowledged. What could potentially kill predictability is the difficulty of constructing vacua, and it is not easy to find the motivtion to debate against (for example) the comment thread on Schellekens’ article below.
I really don’t understand the problem some people have with the Multiverse. To me the idea seems both plausible and very satisfying. If the Earth were the only planet that existed, the philosophical problem of coincidences would be exceedingly difficult to deal with. But the minute you realize that there are many planets the problem simply disappears, and I don’t see anything at all upsetting about the way it disappears, rather it seems to me that it disappears in the best possible way.
Likewise, the Multiverse hypothesis makes the (very serious!) problem of cosmic coincidences disappear in the best possible way. The complaints about it all strike me as the equivalent of being unhappy with the Multiplanet hypothesis because it means that you will never be able to derive the distance from the Earth to the Sun from first principles. Well no, in actual reality you can’t, and that’s just the way it is. And isn’t it satisfying to discover why you don’t have to?
(@ John Brock)
… and you can argue for anything in a Multi-Strawman post.
The thing that you must keep in mind is that it is in fact possible that the anthropic landscape could be the correct description of reality. In this case, then it would be utterly impossible to make unique predictions, although this is not the case within the confines of specific vacua. This just may be a fundamental limitation of reality, just as it’s impossible to simultaneously measure position and momentum. So, you can criticize string theory because of this, but if this is the true reality then we have no choice but to accept it. Is it the true reality? Nobody can really say at the moment, and so more work must be done.
somebody,
Schellekens is the one making the claims about predictivity that you say are in error. Instead of arguing with my often ill-informed commenters, you should e-mail him and demand that he retract his paper from the arXiv since it is wrong.
I don’t think it’s just the public debate about string theory that has gotten traction, so has the professional debate within the physics community. I find that the way physicists in general think about string theory is quite different now than it was 5 years ago. Probably one of the main reasons for that is not my book and Smolin’s, but the public relations efforts of some string theorists themselves. Lubos and Lenny Susskind have done more to discredit the subject than I ever could. The one place where there hasn’t been anywhere near as much of a serious debate as there should be is within the string theory community itself. People there would be well advised to start acknowledging the problems they are facing and engaging in serious discussion about what to do about them.
John Brock and Eric,
For about the 100th time, the problem with the multiverse is not that it is logically impossible or can’t be used to solve a philosophical problem. The problem is that it’s not an idea that can be tested and so isn’t science.
Woit,
I’m sure this is probably the 1000th time, because it is such an obvious point, but although I’ve done some graduate work in physics (decades ago) I’m nowhere near the level of you guys, and I am just trying to understand.
My problem is that when you say it “isn’t science,” it sounds a lot like you are asserting that this is a *problem* with String Theory, and I just don’t understand what the problem is.
Let me put it this way, suppose that tomorrow I were to resolve all the technical issues with String Theory, and I handed you a well defined and consistent theory which demonstrably included our own universe, but which also included a huge number of alternative universes. What would you do with this theory? Would you reject it, because it isn’t testable, and insist we continue looking for a theory that could be tested, even though such a theory might not exist? That’s what seems to be implied by equating “not testable” with “not science.”
And why would you even want to keep on looking? The point I was trying to make in my post that if a theory happens to yield a Multiverse that should be seen as a bonus, not a negative, since it does after all resolve what is otherwise a difficult philosophical problem. You don’t seem to agree that a Multiverse is a good thing to come out of a theory, and that is basically what I’d like to understand.
“Schellekens is the one making the claims about predictivity that you say are in error.”
No Peter, his argument is perfectly clear. He wants to go anthropic because he doesn’t know how to systematically construct vacua. That is his only real argument. Take away the difficulty to construct vacua, and he has no argument.
But your claim is stronger – that string theory is INHERENTLY non-predictive because of the landscape. What I am trying to say is that if we had infinite powers of model-building there is absolutely no philosophical difference between QFT and string theory. So it is a logically defensible stand to try to get string theory under better control so we can build models.
Anyway, I have some marginal disagreements with Eric as well. The only way in which I can make sense of his claim is by the statement that there might be many ways to UV complete our low energy physics. I don’t see why this is bad for making predictions. Because at any energy scale, all I can even in principle hope for, is to identify equivalence classes of theories which could potentially differ from each other at higher scales. We will be able to resolve the different models as we probe higher and higher energies. Again, the thing we really need is control on string theory.
“Lubos and Lenny Susskind have done more to discredit the subject than I ever could”
I really doubt that, since, Lubos keeps giving clear and patient expositions of the central tenets of string theory on his blog, and Lenny has done string theory a big service by drawing the inescapable discretuum of vacua to the center of the debate. Even if one doesnt like that idea, it doesnt help to ignore it, as mother told us.
somebody,
Schellekens (see the paragraph I quoted starting “First of all”) is very explicitly claiming lack of predictivity, based on what is currently known of string vacua constructions. Sure, you are welcome to argue that the problem is that we just don’t understand these well enough, and that when we do the problem will go away. I think he would respond to this the way I would: you’re engaging in wishful thinking.
John Brock,
I just don’t care whether a proposed fundamental theory includes a multiverse or not. If you present me with a theory that can be tested, it passes the tests, and it also predicts a multiverse, that’s evidence for a multiverse. If you present me with a theory that can’t be tested, even in principle, you’re just not doing science, but something else that I don’t care about. The problem with the string theory multiverse is that its role is precisely to make the theory untestable. You can imagine plenty of multiverse theories that are still testable, the problem here is that this one just isn’t, and all the going on about anthropics, the distance from the earth to the sun, etc. is just blowing smoke to avoid acknowledging that there’s a speculative idea here which simply has failed.
Robert,
Have you talked to many non-string theorist physicists and asked them what they think of either Susskind’s promotion of anthropics, of Lubos’s behavior?
Somebody,
My point is that even if there is a specific model in which it’s possible to reproduce everything we know at low energy, it will not generally be accepted or recognized as unique or even interesting. This is due exactly to the fact that many within the string theory community have the same view as Schellekens, who doesn’t think it’s important to try to construct models where gauge couplings and particle masses can be calculated. This is the real danger of the anthropic argument, that people don’t even think it’s worthwhile to find solutions which agree with our world simply because we cannot say it is a unique solution.
OK, I feel I’m getting closer. I had thought that you considered all theories with Multiverses to be in principle untestable, but apparently not. I may already be over my head, but I do appreciate your making the effort! A couple of further questions:
1) Is there any way you can explain what distinguishes a testable Multiverse theory from a non-testable one? I don’t see any obvious way to create two categories; to me it would seem that they would all have to be equally untestable experimentally.
2) How do you deal with the possibility that a theory, like String Theory, might be both untestable and true? Or do you feel that String Theory is just wrong for reasons that have nothing to do with testability, in which case why does the issue of testability come up at all? If the Multiverse is being invoked simply to cover up other failings of String Theory, why not just put it that way, rather than bringing up testability?
3) Do you really not care at all whether a theory implies a Multiverse?!? To me it seems like such an obvious Good Thing, because of the philosophical considerations. All else being equal, an incomplete theory that seems to imply a Multiverse would seem to have a leg up over one that does not, because the Multiverse is such a compelling resolution to the cosmic coincidences problem.
Robert, do you really think someone who has written a book like “L’équation Bogdanov” should be taken seriously? That he writes clearly about string theory is no surprise (it would be quite unfortunate if it were otherwise); he was hardworking and lucky enough to be involved in some developments in string theory, in one of the most specialised, almost carefree academic environments in the world (Rutgers, Harvard), from a relatively early period of his physics education. He had all his time to understand the technical details of much of what he has been writing about, when it comes to string theory.
I am a physics graduate student, and if I took Motl’s rants seriously (with or without the string theory), I would never have thought about spending time studying this subject, or interacting with string theorists (I’m quite frankly relieved that he’s no longer part of the mainstream community). Sadly, he has given a bad image of string theorists (and consequently has made the subject itself less “friendly”), for sure! I have so far met a number of string theorists, and thankfully almost all of them are very different to Motl’s blog-personality. I know of at least 5 other physics graduate students (all in HEP, and all in Europe!) who would agree with me on these terms.
John,
1) The multiverse aspect of a theory may be testable via statistical predictions, and, more importantly, the parts of a theory with a multiverse that just involve our universe may or may not be testable.
2) If string theory is inherently not testable, then the question of whether it is true or not is just not science. The idea that the we live in a simulation performed by higher beings is also not testable, and could be true. You’re welcome to think about this and believe it or not believe it, but you’re not welcome to claim that this has anything to do with science. As for string theory, my specific claim is that the idea that a 10/11d string/M-theory can provide a successful unified theory is simply a failed idea. One way speculative ideas fail is that they turn out to be vacuous, and thus explain nothing. That’s what has happened here, with the multiverse being invoked to avoid admitting failure. It’s an interesting and important exercise to examine exactly what goes wrong with the string theory unification idea, but the bottom line for any scientific theory is whether it provides a testable explanation of some phenomenon. String theory unification fails this test.
3) No, I honestly don’t care. I spent a lot of my youth studying philosophy, and at some point lost interest in the idea that philosophy was going to provide by itself some insight into the nature of the universe. Philosophy can be quite useful as a tool for clearly thinking things through, but that’s not at all how I see it being used in this multiverse story.
Eric says:
“My point is that even if there is a specific model in which it’s possible to reproduce everything we know at low energy, it will not generally be accepted or recognized as unique or even interesting. This is due exactly to the fact that many within the string theory community have the same view as Schellekens, who doesn’t think it’s important to try to construct models where gauge couplings and particle masses can be calculated.”
My point is that uniqueness is not a problem if we know how to build models in string theory. Uniqueness is not how we EVER built models. The virtue of stringy models is not that they are unique, but that they are UV complete. There could be many UV complete models, and to pick between them we need experiments – just like we needed experiments to determine standard model as the QFT at low energies. But this is still a perfectly predictive framework. Predicting gauge couplings and particle masses has never happened in the histoty of physics, but still we have had predictive theories.
There is a subtle but all too important distinction between the anthropism that Peter and Schellekens see in string theory. The former is based on the belief (see his previous comment) that there is an inherent impossibility of constructing predictive models, the latter is based on the fact that constructing models is very hard. My claim is that the effort to understand string theory non-perturbatively is the best way we can gain control of model building. Peter (invoking Schellekens) calls this wishful thinking. I must admit that he could be right, because the problem is in fact very hard. For that the only answer I have is that one must try!
In any event, the way it is often presented on this blog – that even if everything went well for string theory, STILL it would not cut it – that is simply wrong. If we only could wield string theory with sufficient ease, we would have a perfectly predictive framework. Whether we should work on solving this problem, or whether we should give up because it is too hard, thats where only history can be the arbiter. My suspicion is that a problem of this level of difficulty will have to be solved one way or another in any theory that purports to be a fundamental theory.
Why should we assume that the rejection by (Cambridge?) University Press was necessarily a bad thing, or unjustified, or unprofessional? We don’t know what the earlier version of the book said, we don’t know much of what the referee reports said, we simply assume it was a representative case of big bad string theorists using their power unjustifiably? Surely it could simply be a very typical University Press decision. Academic publishers, for quite good reasons, don’t deal with works which are primarily criticisms of other scientists and their research. CUP publishes research, monographs, reviews, textbooks, they don’t publish controversialists or polemics.
That rejection also didn’t in any way force Peter to rewrite the book into a popularizing kitchen-sink attack (which may or may not be an accurate description…)
It simply isn’t necessary or productive to write and publish academic books that are primarily critical of some trend in theoretical/mathematical physics: if and when that trend gets mined out, people will get bored and do something else. Researchers’ time and effort are never well served by deliberate negativity in publications, whether academic or popularizing.
Thomas,
1. The book as published is quite close to what was considered by Cambridge University Press, what was added were mainly some summaries at beginnings of chapter, and suggestions for further reading. No, I didn’t rewrite it into a “popularizing kitchen-sink attack”, which it isn’t. Have you actually read it? The fact that the book ended up getting published by a non-academic publisher didn’t change the actual book, but did make sure the book got more attention and a much wider distribution.
As for whether it deserved to be published by CUP, people can read the book and make up their own minds. I’ve been extremely gratified to hear from a large number of mathematicians and physicists that they learned a lot from the book and had a high opinion of it.
2. The referee reports from non-string theorists were quite positive and endorsed publication, the two from string theorists quite negative. One of the string theorist ones said that they would only give one example of what was wrong with the book because they couldn’t believe any referee would favor publishing the book. The example he or she gave was a perfectly accurate sentence that he or she changed to make it inaccurate. None of the referees or Cambridge editors ever suggested that the problem with the book was that it was just a “criticism of other scientists and their research”. Again, have you read it? There is nothing in the book personally critical of any scientist, or even of any scientists as a group. The only significant discussion of any particular scientist is about Witten, and there the discussion is extremely positive.
3. The book was rejected later on by Princeton University Press because of a referee report from Lubos Motl. For this story, see
http://www.math.columbia.edu/~woit/wordpress/?p=438
I have thought about it for many years but I still don’t get it… Can someone explain to me how an “anthropic argument” is different to a normal “scientific argument”? (Please don’t delete this comment, I can assure you I am a well-meaning honest academic who’d just like to understand this point, I’m not a heckler). For instance, in a previous comment, Jim Clarage wrote:
“For example Newton could have easily adopted an anthropic approach to his inverse square law, perhaps assuming any value for the exponents in the radial and mass coordinates of the gravitating objects; and then agreeing to anthropically derive these exponents based upon e.g., what values allow and disallow stability of human bones, large tree trunks, or other mechanical properties of organisms living under the influence of Earth’s known gravitational field.
…
However, instead of appealing to anthros Newton appealed to physics. He tried to pick the exponents and form of the law based upon reproducing previous physical laws, viz., Kepler’s laws.”
[italics added]
Could someone help me out here? Surely it amounts to exactly the same thing… what is the difference here?
(a) Newton makes the ansatz F = G m_1^a m_2^b r^c for a Law of Gravitation, in terms of some unknown exponents a,b,c,d. He then performs a rigorous mathematical tour de force and shows that large tree trunks can only exist if a=b=1 and c=-2.
(b) Newton makes the ansatz F = G m_1^a m_2^b r^c for a Law of Gravitation, in terms of some unknown exponents a,b,c,d. He then performs a rigorous mathematical tour de force and shows that elliptic orbits (whose existence is gleamed from previous astronomical observations by Brahe, Kepler, etc) can only exist if a=b=1 and c=-2.
Why on earth is (a) called an anthropic argument and discredited as “pseudo-science”, while (b) is considered as valid science? They are the same thing, no? What is the difference between observing the existence of large tree trunks with binoculars and observing the existence of elliptic orbits with a telescope?
“I’ve been extremely gratified to hear from a large number of mathematicians and physicists that they learned a lot from the book and had a high opinion of it.”
Indeed, I considered buying the book (I read it in a bookshop) simply for the clarity of its mathematical presentation of the history of the ideas coming from witten, segal, atiyah, etc. Believe me, as a mathematician I find the usual description of these topics written by conventional string theorists very difficult to understand.
Johnson,
I’m glad you liked the book. But please, the discussion of anthropics is getting off-topic and I don’t want this comment section to turn into yet again another endless discussion of this issue.
Okay… I will wait until this issue comes up again.
Somebody said,
“In any event, the way it is often presented on this blog – that even if everything went well for string theory, STILL it would not cut it – that is simply wrong. If we only could wield string theory with sufficient ease, we would have a perfectly predictive framework.”
I’m just a layman struggling to understand the ideas presented here, but I’d really like to know what evidence you have that forms the basis for this belief? How do you *know* that String Theory would “cut it” if only we had enough control over how to construct vacua? What guarantee do you have that our Universe is correctly modeled (not at just the currently known low energy results) by one of these sufficiently well constructed vacua?
Even more, what guarantee can you give that if only we knew how to construct the vacua better that the Standard Model would fall out at even low energies? To me, it seems that the only answer I’ve ever heard from String Theorists is that the math is just sooo elegant that, “It must be so!” This is not particularly convincing for a theory that hasn’t even been demonstrated to be consistent yet.
I just don’t understand the confidence you guys voice when top guys like Witten admit that you don’t even know what String Theory *is* yet. How can you be sure that it’ll end up describing our Universe at any energy level when so little is known of String Theory?
John Brock said,
“How do you deal with the possibility that a theory, like String Theory, might be both untestable and true?”
How on earth would you ever know that it was true if you couldn’t test it? More to the point, there are countless theories one could come up with that ‘might be true’, but are untestable. We’re all living in a VR made by higher beings… The Spaghetti Monster did it…
“Do you really not care at all whether a theory implies a Multiverse?!? To me it seems like such an obvious Good Thing, because of the philosophical considerations.”
How so? I don’t understand how some vacuous idea about a ‘Multiverse’ can be useful to explain anything? It sounds like you are arguing for ignorance. Just suppose for a minute that the Multiverse was true: we’re all living in just one of the countless number of Universes in existence. Ok, what can you tell me about the Multiverse itself? What are it’s physical principles and dynamics? Since we’re already supposing a Multiverse, it is only fair to suppose this Multiverse’s physical dymamics are also inscrutable (at least at first). We form theories with new physical constants that seem to have just popped out of nowhere… Now should I start supposing a Multi-Multiverse?? Perhaps the Multiverse concept is only appealing because it is so void of content that people can ascribe anything they wish to it.
“Even more, what guarantee can you give that if only we knew how to construct the vacua better that the Standard Model would fall out at even low energies?”
I cannot give you any guarantees whatsoever. But neither can I give you any guarantee that Yang-Mills has a mass gap, nor that black holes radiate, nor that the gravitational coupling doesn’t change with time. But I still believe all these things are “true” in the contingent sense that scientists use that word. The point is that I see no reason why my quality control should be any ore “string”ent with string theory.
Anyway the reason why I think the standard model can arise from string theory is pretty simple, and I have a suspicion that it will not convince you: string theory permits many kinds of standard model like models, and there doesn’t seem to be any reason why standard model should be particularly impossible. That is it. Proving it one way or another, I think, is going to be HARDER than actually constructing one.
Just to draw a caricature to drive home a point: if you ask me why I don’t believe in telepathy, the answer is again only that it seems pretty unlikely. I have been held to court and asked whether I can PROVE it doesn’t exist. This business is all about judgement calls and thats something people don’t seem to get. Its not about worrying about exceptions even before the get-go.
” To me, it seems that the only answer I’ve ever heard from String Theorists is that the math is just sooo elegant that, “It must be so!” ”
You asked a string theorist why he believes the standard model could arise from string theory, and the answer you got was that because the math is elegant? You should certainly talk to better string theorists.
Somebody says,
“Anyway the reason why I think the standard model can arise from string theory is pretty simple, and I have a suspicion that it will not convince you: string theory permits many kinds of standard model like models, and there doesn’t seem to be any reason why standard model should be particularly impossible. That is it.”
Ok, let’s presume that String Theory provides a sufficiently malleable framework wherein we’ll eventually learn to construct vauca to accommodate the Standard Model. Given this, what reason is there now to believe that this will elucidate or provide predictive power for future experiments beyond the Standard Model?
manyoso,
Peter seemed not to want to continue on this topic, so I wasn’t going to press him, although I wasn’t content with his answers. But I’ll take a shot with you.
“How on earth would you ever know that it was true if you couldn’t test it? More to the point, there are countless theories one could come up with that ‘might be true’, but are untestable. We’re all living in a VR made by higher beings… The Spaghetti Monster did it…”
The thing is, Virtual Reality and the Spaghetti Monster are obvious and uninteresting cheats. They are strawmen. Yes, you can trivially come up with any number of theories like that, and who cares. But can you come with any number of mathematical theories which are well defined and consistent, in some way mathematically elegant and unique, and demonstrably include, as one of a huge number of possible solutions, a solution which describes our actual universe? How many of those can you come up with in the next five minutes???
Now here is the point: my understanding, as one of the ill-informed laymen who are part of the intended audience of Peter’s book, is that such a theory is what the String Theorists are still hoping to find. This seems like an entirely reasonable judgment call to me! I do understand they haven’t found it yet, despite decades of searching. I understand there are many possible reasons why they may never find it (that it doesn’t exist being only one of the possibilities!). But if they did find such a theory, wouldn’t you have to agree that it would be enormously more compelling as a candidate for the Theory Of Everything than the Spaghetti Monster, even if the theory included a Multiverse which made testing impossible?
That’s it in a nutshell. The Spaghetti Monster will never be a compelling TOE. A well defined and consistent String Theory which actually described our universe (along with many others) would be, even if you couldn’t test it. If you believe the universe is describable by mathematics, and I hand you a mathematical theory which can be shown to describe the universe, doesn’t that have to count for something? Unlike Spaghetti Monsters, such theories are not a dime a dozen. So far we don’t even have one.
John,
I find the idea that there is a mathematically simple and compelling TOE that doesn’t predict anything is about as plausible as the FSM. You’re making up mythical creatures that there is no evidence for and debating their properties.
But is that what the String Theorists are looking for?
If they believe they can find such a thing and you don’t then this whole thing is just a difference of opinion at a level where there is no point talking to the lay reader such as me. What I’m trying to figure out is whether their is some problem with String Theory which can be understood at my level. I understand that you don’t think they can come up with the sort of well defined theory that I am talking about. But do you understand why it seems to me that if such a theory were found it would have to be given a lot of weight, predictions or no?
The thing is, presumably the String Theorists are more or less as smart as you are, and if they think they can find such a theory, and you think they can’t, and the argument is on technical issues, then there is no way I can take a side. The only level I can deal with involves questions such a Multiverses and testability, and so far I just don’t see that String Theory has any problems that can be expressed at that level.
manyoso:
“Ok, let’s presume that String Theory provides a sufficiently malleable framework wherein we’ll eventually learn to construct vauca to accommodate the Standard Model. Given this, what reason is there now to believe that this will elucidate or provide predictive power for future experiments beyond the Standard Model?”
You could ask the exact same question with QFT : “yeah, sure, you can build models, but why should we believe it is going to be predictive?”
Model-building is something which requires ingenuity and intuition, and what we are seeking is a paradigm to do that in. QFT provided that for non-gravitational theories.
If you don’t see the answer from that comment, see my other responses on this thread. This fallacy of “predictivity implies uniqueness” seems to be too deeply ingrained in everyone’s mind, but right now I don’t have the time to go through it again.
John Brock,
“But if they did find such a theory, wouldn’t you have to agree that it would be enormously more compelling as a candidate for the Theory Of Everything than the Spaghetti Monster, even if the theory included a Multiverse which made testing impossible?”
You are misstating the situation I believe. If in the future String Theorists discover that the theory is consistent and provides a way to calculate the Standard Model, then they’ll have found a theory that is testable. If this same theory goes on to predict multiple universes or other bizarre and fantastical physical outcomes so be it. Even if we can not find a way to test all of the predictions, the theory would still provide others. I’ll bet that the entire community of physical scientists would celebrate such a discovery.
However, the current situation looks nothing like that from what I can discern. What exists is a mathematical framework that has not been proven consistent and has not been shown useful to calculate any known physical result. It has not been shown a super set of the Standard Model.
All we have is this hugely malleable framework and a landscape of possibilities describing perhaps ~10^500 different universes. Currently, the String Theorists take it on faith that our Universe is some needle described by this huge haystack. And why do they? Because they’ve found some other needles which if you squint your eyes and turn your head kind of resemble the needle we’re looking for. That and because the framework has an elegant way of including gravity.
The guys that argue for the anthropic landscape throw up there hands at ever having to find the actual needle in this landscape. They don’t even see a particularly compelling reason or need to find it?! They are ready to argue that String Theory is a success even without finding the part of it that makes contact with our known world. This is the point where I think the exercise becomes completely unscientific.
Again I’m just a layman myself, but this is the situation from what I understand. Upshot? If String Theory wants to predict fantastical things that is fine with me, but largely irrelevant. The test for any physical theory is whether it can be used to model known experimental measurement.
Hi John,
I rarely butt in here, but this one is like a thorn in my side. It just seems to me that you would expect a theory which “demonstrably include[s] a solution which describes our actual universe” to be testable, whether it includes a multiverse or not, especially if it is “mathematically elegant and unique”. Such a theory would have to go beyond the Standard Model and have something to say about, for example, neutrino masses, the hierarchy problem, dark matter and dark energy, the number of Higgs bosons and their mass, particle masses in general etc.
If it can’t do that, it’s just another pretty model, a curiosity. Worse, I think we’re really talking about little more than a unicorn, since, as far as I can tell, these conditions—an elegant and unique TOE description of our actual universe that provides nothing more than the SM—are mutually exclusive.
It’s not the presence or absence of a multiverse in the theory, in and of itself, that’s the deal breaker. And I think this is what Peter is saying: he’s multiverse agnostic. Just because a theory spawns a multiverse doesn’t—or shouldn’t—make it intrinsically untestable. If it also includes a beautiful, unique, and complete description of our universe, it must have something new, unique and testable to say about it. The confirmation of such a feature, or features, of our universe would just, at that point, lend credibility to the theory’s multiverse.
The problem is that in a kind of logical « tour de force » this has been spun around backwards and upside down in, uh, certain circles, so that the inherent untestablity of a theory gets blamed on the multiverse it spawns.
Best,
Gilbert
For those such as Gilbert and manyoso who have made claims that string theory does not contain anything that looks like the Standard Model, I suggest you do a search of the literature. You will find that there are several good models, both heterotic and Type II, which resemble the SM closely at low energies.
John,
Claiming that string theory unification should be believed if it is completely unpredictive is an extreme minority viewpoint among string theorists. You can find such people, but most string theorists know that if string predicts nothing it will have failed. Some have given up on string theory as providing predictions about particle physics, but think it worth studying for other reasons (e.g. as a model for quantum gravity, as an approximation scheme for strongly coupled gauge theories, etc.). Others are still trying to get predictions. But virtually none are out there telling their non-string theorist colleagues that string theory can be a successful unified theory that doesn’t predict anything. They are well aware that their colleagues would laugh in their face if they did this, and never take them seriously again.
Eric, please read again. I never made that claim. And I don’t see where Gilbert made it either.
Hi Eric,
I think you’re reading things into my post that aren’t there, as I certainly make no such claim. I mean, it’s just simply not there, unless your browser is seriously mangling the text somehow.
And while I wouldn’t want to speak for manyoso, I don’t see that he (she?) does either. In case you missed it, “…they’ve found some other needles which if you squint your eyes and turn your head kind of resemble the needle we’re looking for” lines up better with “there are several good models… which resemble the SM closely at low energies” than “does not contain anything that looks like the Standard Model.” Seems to me, the difference being a (rather large) difference of opinion as to how close the models line up to the SM.
Methinks perhaps you’re making too many assumptions here.
Best,
Gilbert
Manyoso, you said
“If in the future String Theorists discover that the theory is consistent and provides a way to calculate the Standard Model, then they’ll have found a theory that is testable.”
I suppose you mean to calculate the SM parameters since you acknowledge that there exists models which look like the SM with three generations. Essentially, this boils down to stabilizing the moduli which parameterize the size and shape of the compactified manifold, upon which all couplings depend. For a given MSSM-like vacua with all moduli stabilized via flux or other nonperturbative effects, all couplings are determined and everything is predicted.
Eric, Gilbert, Manyoso,
This discussion has become both off-topic and completely unenlightening. No more about this.
Peter,
I think that Sabine’s claim, it would have been better to keep the string theory discussion within the physics community leads to the questions, whether you agree and if so, why you saw the need to write your book still. In other words, I would think Sabine’s position would have required a stronger response from your side. In would be interesting to hear more along these lines now.
Mikael,
I think I already responded to this at Sabine’s blog, but here’s some more. There were two distinct problems at the time I wrote my book (2002). One was an unwillingness of string theorists to acknowledge to their colleagues how bad the situation of the theory was, and to engage in thinking about what to do about this. When Sabine says that it would have been better to have this discussion among physicists, I don’t think she’s saying that string theorists were willing to do this, but I somehow insisted on not engaging with them and went to the public. We both agree that there is a discussion that needs to take place among physicists, and that this was not happening. I tried to get an article published in Physics Today in 2001, this was rejected. The book I wrote was aimed at a scientific, university press audience. String theorists stopped university presses from publishing the book.
To this day, string theorists are still stone-walling, refusing to admit what the problems are and that something needs to be done about them. All that has happened since 2002 is the arrival on the scene of the anthropic landscape pseudo-science mania, which really has not made string theory look good. At this point, string theorist’s plan seems to be to just do nothing until LHC results arrive, and hope that will change things. We’ll see how this works out. In the meantime, their colleagues are noticing this and I’ve heard sometimes explicitly deciding not to hire string theorists until they see what the LHC says. The kind of discussion that should take place hasn’t happened, and that’s a shame.
The other problem was that the discussions of string theory available to the public (and to mathematicians, who are a significant audience for the book) were pretty uniformly overhyped and misleading. I thought outsiders deserved a more honest explanation of what was going on. The book was written to be as accessible to as many people as possible, while giving a serious discussion of the subject. I think this was worth doing. I’m not at all sorry that the book ended up being published in a way and at a time (with Smolin’s) that got a lot more public attention than I ever expected. I’d have preferred to have the book published two years earlier by CUP, where maybe it would have been ignored, maybe it would have generated some more serious discussion (although I fear this was not very likely).
The oft-repeated claim that “having many vacua which look like ours at low energies kills predictability” is simply an incorrect one, and no sensible discussion about the landscape is possible when this sort of factual errors are not acknowledged.
Hi Somebody,
Could you possibly briefly explain exactly how predictability is possible in the case where there are many vacua that look like ours at low energies? (Assuming we pretend for the moment that actually constructing those vacua, as you are concerned with, somehow turned out to be easy.) If we find that we can identify a whole bunch of string vacua which produce the standard model at observable energies, but split off into a variety of higher-energy behaviors past that depending on which vacua you choose to study– I am unable to understand how to proceed in this case. What do you do next? How is the prediction made?
What I am trying to say is that if we had infinite powers of model-building there is absolutely no philosophical difference between QFT and string theory.
So let me see if I understand what you’re saying here. You’re suggesting something like that the general QFT toolbox allows us to construct many different models; each of these models is basically a “vacua”, or analogous to one. We can thus imagine there is a “landscape” of QFT “vacua”, each describing a different QFT which one might construct; and by nature there are many different QFTs which reproduce the standard model at “low energies” but do different strange things at higher energies (i.e. SU(3)xSU(2)xU(1) is one, SU(5) is another). Do I understand your argument correctly?
But surely this is not how people actually use QFT? Nobody I’ve ever heard of actually treats the space of usable QFTs as something to work with directly. Instead we pick one QFT “vacua” (say, SU(3)xSU(2)xU(1)) and stick with it. We say QFT is predictively successful not because “QFT” produced useful predictions, but because a QFT produced useful predictions. Moreover the problem you describe– that there are many different QFTs with different high-energy behavior and we have to pick one– it’s not that this is a standing problem with QFT which we have simply chosen to excuse, it seems more that this is a problem in QFT which can be addressed because the different QFTs are highly distinguishable and some QFTs (say, ones which are simpler) are going to be inherently considered preferable over others. We can recover predictability despite a multiplicity of choices, because some of the choices are better than the others. I have never heard of any analogous selection principle being offered for string vacua. Am I missing something?
From my perspective as an outsider, it looks like faced with the problem of there being more than one potential QFT which produces the Standard Model at low energies, physicists seem to have coped just fine with picking the QFT to use going forward; they even seem to be coping with this problem in extending the Standard Model, we have things like the MSSM and the minimal dark matter model (chosen as I understand based on the simpler-is-better rule). But what would physicists do when faced with more than one potential string solution that fits experiment at low energies? You say what you need for predictivity is a vacua-construction method; if you had one, what would you do with it? I’m not trying to be snarky or difficult, I’m just trying to understand what the procedure here would be.
Peter,
I think it was good (or at least necessary) you wrote the book.
You know, what really strikes me is that even a guy like David Gross seems to have become a little more skeptical about string theory in recent years. You’d think that somebody like him as an independent thinker would give a damn about such books and still he seems to have been influenced by this ongoing discussion.
Coin and Somebody,
Claiming that the situation of QFT (which makes a huge number of non-trivial, verifiable predictions of high accuracy), and string theoy (which predicts absolutely nothing) are the same is just sophistry. QFT would be in the same situation as string theory if simple QFTs didn’t work, so you had to keep adding more fields and complexity, just to evade falsification, ending up with a complicated mess, unusable for predictions. That’s where string theory is now, it’s not where QFT is.