Here’s some interesting Sunday reading: Frank Close wrote a very nice article for Prospect Magazine on the business of testing scientific theories in Physics. Ideas about multiverses and also string theory are the main subjects under consideration. I recommend it. My own thoughts on the matter? Well, I think most of you know them. Here are three key points:- (1) Many important ideas in physics started out as purely mathematical digressions inspired by physics… You can’t find those ideas and make them work without exploring where they lead for a while, perhaps long before you even know how to test them. So we need that component of physics research as much as we need some of the other aspects, like designing and performing new experiments, etc. (2) You never know where a good idea will ultimately find its applications… It is often not where we think it might be initially. (3) Of course, testability (confronting an idea with experiment/observation) is key to the enterprise of doing science, there is no doubt in my mind. I do not think we need to start considering whether testability is something we can abandon or not. That’s clearly silly. We just need to be careful about rushing in to declare something testable or not testable before it has had a chance to develop into something useful. Unfortunately, everyone has a different take on just when it is time to make that declaration… and that’s what causes all the shouting and political arguments that generate a lot of heat and precious little light.
Enjoy!
-cvj
Ion Vasile Vancea liked this on Facebook.
Mitchell,
He really should have been talking about the version of the theory that derives the standard model and allows us to calculate something we know about in our universe.
Hi Mitchell,
I think that’s what he means. He means the ten dimensional theory, uncompactified and so forth, since then it has the wrong dimensions, too much supersymmetry, the wrong gauge theory, and so on and so forth….
-cvj
I am somewhat bewildered to see David Bailin saying that heterotic E8xE8 string theory is obviously false! Does he mean the uncompactified theory?
Depends on how you define “QFT”. For the working definition “a quantum theory that is aware of localization” I would count a violation of locality/causality as in indication that the assumptions of QFT are wrong.
Well for me the cosmological constant failure makes it clear that QFT is incomplete, at least.
Separating the issue of scientific judgment, which can be done only by the scientists (I think is equally silly when philosophers, untrained in the relevant technical issues, make judgment on these issues nonetheless). But, when it comes to history, philosophy and sociology of science (how it is organized, how are decisions made, why is it so successful, etc., etc.), we are sadly, by and large, equally naive. I guess the time for meaningful cross-disciplinary contributions is over.
Well, they (we) are surely not the best in describing how science is done (both for the reasons I described, being untrained and fundamentally biased, and as an empirical fact, evidenced by the insistence on sticking to the Popperian myth despite its glaring flaws).
🙂 see my remarks to Moshe on this very point about scientific communities being the best hope for working out how to do science….
Permit me to not lose any sleep awaiting for their enlightenment, and be happily surprised should it actually happen. See, the founders of meta-mathematics were actual, practicing mathematicians. Perhaps all too naïvely, I can’t see anyone but actual, practicing (mathematical) physicists figuring out the rigors of “whereof and how we can speak clearly, and whereof we should shut up.” 🙂
For my part, I find that Close’s article, while well written and worthy, still seems to perpetuate the pitting of empirical (direct experimental) observation/measurement against sheer mathematical elegance. In doing so (as does most if not all of layman-accessible literature), it omits logical consistency—the factual sine qua non of all (sane) attempts to understand Nature. And, if sanity alone is not enough of a motivation, recall that 100+ years ago, Bohr’s original beef with Rutherford’s planetary model was the unexplained stability of electron orbits (commented upon in Rutherford’s original 1911 paper), not the theoretical derivation of the experimentally observed Balmer formula; that came as a bonus. Let’s learn from the past century’s worth of predictions from logical consistency.
What can be observed must fit together in a coherent and logically consistent way: Nature is one, and is not inconsistent (malicious/frivolous… w/apologies to Einstein 🙂 ).
Gabriel Cozzella liked this on Facebook.
Tristan Hubsch liked this on Facebook.
I agree, but see what I said about editors.
The best part is that science wins!!
“it is the nuanced story that will never be told in a popular level magazine article”—but should. Nay; must.
Tristan … Nicely put. I think that there are some interesting philosophers of physics thinking about this, such as Halvorson at Princeton, but there is a lot of work to do. The main problem, as I’ve said to him, is that they’re still stuck at about 1973 in terms of what has happened in theoretical physics, and there are so many things to incorporate to enrich their discourse if it is to feed back to the physics community and get noticed and be of use. We shall see what they do with it.
Sabine: You can easily Google for Carl’s talks online, and find all sorts of references. For example, here’s one: follow the slides that, at some point, Carl shows the experiment in question, and the relevant papers,
» <http://hatano-lab.iis.u-tokyo.ac.jp/hatano/NonHermite/Talks_files/1211_02_bender.pdf>.
Further, note that my point about Resurgence and Alien Calculus, was that these tools are currently being used in order to make QFT mathematically rigorous, yielding all sorts of extremely interesting results along the way. This is not a “minor detail” that “physicists can relegate to purists” or some such thing: if you’re going to be serious about the whole falsifiability-à-la-Popper-thing, then this is an extremely interesting path to follow for several reasons, including historical ones.
Hope this helps.
Indeed, (the increasingly mis-named) STRING theory is undeniably a theoretical (axiomatic, to some extent) system (what I understand you call “toolbox” and Moshe calls “method”) for building models, rather than a concrete model. Just as “classical mechanics” is too, which is a limiting corner of “quantum mechanics,” which is a limiting corner of “quantum field theory,” etc., each of the latter constructs in this chain extending the former; “string theory” is but one such extension in the chain of ever broader theoretical systems. The careful distinction between concrete models within any one of these theoretical systems and the theoretical system itself is something we ourselves do not study rigorously.
Mathematicians have meta-mathematics: very rigorous and definitive, and certainly useful in thinking clearly about how one thinks mathematically (e.g., Gödel’s 2nd theorem as a response to the Russell-Whitehead “Principia Mathematica” efforts). Physics does not have meta-physics of the same ilk, but (I believe) needs it as much — if not more so.
Cribbing from Robert L. Forward’s “Indistinguishable from Magic” [Baen, 1995, ISBN 0671876864; also (shameless plug alert), look for ISBN 1107097487 in a few weeks]: some perfectly scientific (to a naïve popperite’s chagrin) predictions cannot possibly be tested for reasons that have nothing to do with the theory that predicts them. Think of (1) the requisite resources are unreachable (e.g., collider of a parsec radius), (2) the requisite procedures are prohibited by ethical/moral… reasons (e.g., cloning, bionic/cyber-organic modifications…), (3) the requisite fundamental science is known but the engineering lacks (e.g., paint the ceiling with neutronium to create a weightless room), (4) profoundly new concept/methodology is required (e.g., to experimentally determine the _upper_ limit of proton lifetime). And, that’s not even mentioning the manifet impossibility of double-blind experiments in astrophysics and cosmology—which are none less scientific for it.
Of course, thought experiments suffer none of these limitations, and when properly executed are as useful. (ABJ and GIM and other theorists predicted the 4th quark years before the experimentalists were utterly surprised by J/psi; to name one of last century’s examples—and lessons.)
Ha ha ha! Let us declare it so and carry on with doing science!
Thanks Clifford, so now aplying the Popperian strategy to itself, did we just falsify it once and for all, via that counter-example? That *would* be nice.
And yes, obviously it is more complicated than the Y-in-box discussion, but I hold that a step in the spirit of what I wrote to Elliot is still a key part of the enterprise.
Moshe , I agree wholeheartedly. Good example. (QFT + cosmological constant)
And, while I am poking holes in the Popperian story: QFT either does not know how to calculate the cosmological constant, or gives a ridiculous value for it. Did this fact, known for decades, falsify QFT in anyone’s mind? Clearly the real story of how things work in the real world is slightly more complicated that the falsification myth.
This is true, but one comment though. In the real world you don’t shake boxes but do a complex operation whose interpretation depends on many layers of theory. This is one of the problematic aspects of naive views of testability, the impossibility of cleanly separating out the empirical from the theoretical (relatedly: the underdetermination of model from data). None of this undermines the necessity of empirical verification, just that this is not a simple black and white issue. As I said, life is messy and simple stories don’t do it justice.
Elliot – you stopped before you were done. The next step is to write down a model of Y. That model then suggests properties of Y-in-the-box that you can test, by shaking the box in a particular way. The model which predicts the most phenomena (you can also build in some postdicting success too if you want, just not too much) correctly is the model to keep, I’d say. It’s your current theory of Y, until a better model comes along. That is the role of testability. It has served us well for a long time. I see no compelling reason to leave that step out in the science we are doing now…. But yes, (see my thoughts in the post) sometimes it might take a long time to getting to that testability step: but we still need it at some point, otherwise, a beautifully choreographed interpretive dance about Y would be just as useful for doing physics.
Herbi – indeed, that’s what I said in the closing of my covering remarks in the main post.
Daniel – Indeed! Thanks.
The problem with this argument is that you can use it to justify all kinds of speculation, and the question arises where to draw the line. The simplest answer is: “to each his/her own”. It only becomes a problem with regard to funding, I believe.
Can you give me a reference?
Sabine: than, in that case, you should really pay close attention to this foray into the Complex domain. It links very diverse corners of physics and math, in particular PT-symmetric Quantum Mechanics, which has been showing very interesting experimental results lately. I had a chance to talk to to Carl [Bender] maybe a month ago, and some of the experimental developments he mentioned were quite compelling.
Daniel: Yes, I gather the string theory – math connection is going nicely. But I’m a physicist and I care mostly about the theory’s potential to describe observations.
Sabine, Clifford: Just as a side comment, it’s also worth noting that, lately, there have been quite interesting and different fronts of development that, in some sense, were derived from broad stringy topics. For example, extensions into the Complex plane, which now have transformed into Resurgence and Alien Calculus; novel categorical approaches to knot theory (and particular invariants), tying quite neatly with previous work done in n-categories; etc. The point being that there are many interesting areas ‘sprouting’ here and there, even if a bit under the radar as of now.
OK since I am not a practicing physicist, let me proceed by an analogy. Suppose we have a box X with object Y completed sealed inside. We can do “some” experiments on the box. We can weigh it for example to get an upper bound on the weight of the box plus the object. But we cannot detemine the actual weight. We can shake the box to determine if the object attached or freely moving. We can listen to the sound. We can determine if it is magnetic. We can learn a great deal but we cannot “know” what the object is ontologically. We can however exclude a lot of possibilities. I think testability is not necessary a vehicle for truth, but a vehicle for narrowing potential answers. When looked at in this light, it is not a true/not true issue but one of getting closer to it. I consider that scientifc progress even if it is decoupled from physical “tests” and is only based on mathematcal consistency and elegence.
HI Sabine. This is a topic that is dear to my heart, and I’ve spoken about it a lot in the past. I see it as one of the most important aspects of the field that has come along (gradually) in the last several years, and represents an important maturing of the field. There is a very diverse set of topics within the subject (or inspired by examples that first arose in the subject) that (along with topics like applications to condensed matter and nuclear topics) are all vital explorations of what string/M-theory really is, and what it can teach us about quantum field theory, spacetime, etc. It was clear to me (and I imagine, others) a very long time ago that it was very premature to have the entire field all working on trying to squeeze the theory into one simple (‘theory of everything’) role, and that we needed to diversify and explore it in many contexts (especially connecting with other types of experimental physics) in order to really get to grips with what we’ve got, and what the theory can and can’t do. The benefits are that we (1) Get insights and useful tools for all those different corners of exploration, and (2) We strengthen the program of developing the subject for its application to the (naive, in my view) “theory of everything” quest. So in my view, even though it might _seem_ like all these people are not following the original string theory narrative of seeing what it can do for understanding physics beyond the standard model, and far beyond the standard model, I think that they all are part of the program whether they know it or not…. In other words, all that work will lead us back there eventually, stronger and better equipped. It may or may not be string theory as we think of it now that we come back to these issues with, or we may have found something entirely better suited to the job, but either way, it is by this healthy diversification and exploration that we will find our way, not by just sitting collectively working on the same narrow ideas. [As I was saying to/with moshe above, it is the nuanced story that will never be told in a popular level magazine article, for all the reasons I said.]
Sergio Fierro liked this on Facebook.
Sabine Hossenfelder liked this on Facebook.
Mathias Völlinger liked this on Facebook.
David Hurn liked this on Facebook.
Virgil Guevara liked this on Facebook.
And while I am here, let me also ask you a question. I have the vague impression that there are not so many people left working on string theory as ‘the theory of everything’ and instead most are now doing AdS/CFT and extensions thereof (dS, time-dependent, etc), dualities in general and applications. Do you share this impression?
I agree with Moshe that physicists are not in the position (and don’t have the interest either) to analyze the practice of science. Neither, for that matter, are philosophers. To my knowledge there isn’t a historical precedent in which sociologists corrected some trend in science. But then you would expect this to become relevant only with large communities and the increase of external pressures.
Physicists are of course biased when it comes to judging their own theories. That is a priori not a problem. The problem is that they’re not educated to become aware of and account for their biases. The bigger problem though is that the academic system today creates other problems. And btw, just to be clear on this, my issue isn’t with string theory, I have a general worry about academia. My pet peeve is that it’s extremely hard to change topics after PhD. It’s like people grow up (excuse the analogy) in some religion and then can’t get out of it. This is clearly very bad for objectivity because everybody has a huge incentive to stay loyal to whatever it is they are working on.
Interesting, thanks, I haven’t read Close’s article. It’s basically what Dawid discussed in his book http://backreaction.blogspot.com/2015/05/book-review-string-theory-and.html
??? liked this on Facebook.
As for your question, I don’t know of any example, but I am not all that knowledgable about this.
Clifford: I think the subject of the “method” could be an interesting discussion, or at least more interesting than it usually is. I think that it is just whatever you have to do to increase your confidence (Bayesian probability, if you want) in your results, so it is fluid and context-dependent, but life is messy…Sometimes it involves empirical checks, or consistency checks, or whatever else works. Analyzing what has worked at different circumstances is an interesting intellectual pursuit, I think.
But it is good that you spelled it out with specifics, David. Thanks.
Clifford, Thanks, I see what you mean. I did not appreciate it fully when I first read it.
David – Exactly! This is part of Moshe’s point (2), in a way.
To give perspective to the question”Is superstring theory falsifiable?”, I think the question “Is quantum field theory falsifiable?” should also be addressed. Of course, we can and have falsified particular QFTs, QED for example. Equally, particular superstring theories are obviously false, heterotic E8xE8 for example. Indeed, so far as I know, the string theories corresponding to all vacua so far written down are also false. But of course this this does not falsify string theory, any more than the falsity of QED falsifies QFT. Further, it has been argued that some string theories are N=4 QFTs, which I think ties together the two questions very nicely.
I am curious though as to whether you have a historical precedent in mind for the “deep study of the subject” aspect. Can you point to a time where the physics community was all adrift and people from outside the community came and , after careful study, pointed the way? My own feeling is that there are no such clear examples, and in all cases, it was people from the community itself, one way or another, that got the ball rolling. I don’t happen to believe that we live in some particularly special time in history, so this is why I think we will do what we’ve always done eventually…. Finding our own way.
Indeed, I should have not used the word “method” lest it be construed as meaning some rigid dogmatic way of proceeding. That was not my intention.
My own view on how things progress does not involve any rigid a priori “method”, but there are certainly decisions to be made, continually, on the personal level and the community level. The manner in which those decisions are made, and any regularity that could be found in that, is fascinating. Deep (even quantitative!) study of the subject can also, conceivably, improve the way those decisions are made and therefore their outcome. I don’t think the practitioners are in any position to do that analysis, and I don’t think any simple stories involving “methods” really penetrate into this subject.
I agree with you on that. I am not one of those physicists who pours disrespectful scorn and contempt on those who reflect upon our field professionally… (Historians, philosophers, etc)… And I agree with you about individual physicists not being the best judges of their work… So that is why I said “scientists… as a group”. I am a firm believer in the overall process eventually producing great lasting knowledge and real unexpected value… Not an individual scientist with extraordinarily great wisdom. I don’t believe there are any, in the really long run, for the reasons you hinted at…. Eventually everyones own work gets in the way of them seeing the big picture clearly…. But the process itself….. It is powerful. So scientists as a group using the scientific method properly are the best deciders of what’s good science. With enough time to shake out the wrong turns…. That’s my point.
Yes, I agree with that, I am referring to the debate as a whole and not this one article which is rather good.
As for point (1): I agree with you that the analysis of the practice of science, rather than normative discussion of what science should do, is more useful. But physicists are equipped to do neither by training. Physicists as a whole are of course very smart, and perceptive, and tend to be handsome and humble, and have many other superb qualities. What they don’t have is the relevant professional training to analyze the practice of science. Not only the relevant knowledge (which they are more likely to have in their subfield of science, but how many physicists know about 18th century biology?), but most crucially the tools of analysis. Then again, they are also deeply emotionally attached to the subject…So, you could see why I prefer the grain of salt in reading their analysis of their own activity. Especially when (surprise, surprise) it ends up that they adopt the viewpoint that paints them in the most heroic light.
Moshe – Well, you and I know the people who were quoted, and they are good people, and Frank writes very well and knows the issues too. Remember that at the end of the day, a magazines content is controlled by editors who want a certain content, and a certain emphasis and a certain spin on what they consider the big issue. My guess is that many times it comes down to writing it with the focus they want, or you don’t get the piece showing up at all. I see this all the time with excellent well informed writers who, at the end of the day, are not the editors and just have to do their best. There is no shortage of excellent writers willing to write on the real contemporary issues…. What we need is better informed editors to allow that nuanced more up to date stuff to appear. That is a harder row to hoe.
As for (2): Some professional physicists quoted in the article must know about the evolution of string theory in the last 15 years, and the new viewpoint that entails. They may not think much about it, which is fine, but debating the pros and cons of the subject circa 1984 is even less useful than debating its current status.
And on point 1… I tend to look to practicing scientists as the ones who really, as a group, carve out what science really is and isn’t. I think there is value in people such as historians of science, or philosophers, etc, looking in and making what they will of what we are doing and what it all means for the enterprise, but it is mostly post-hoc, and not frequently that useful for the scientists carrying on with the business of getting on with doing the science.
Moshe – Indeed, you and I have emphasized that view in point number 2 on this blog many times over the years. Popular level articles tend not to care much about string theory as a powerful toolbox, presumably because they are aimed at audiences who have been bombarded only with discussions about theories of everything and the holy grail of physics, and the like…. Pragmatic discussions of what is calculable in various fields of physics and useful tools for doing so are just not as sexy. Occasionally you do get the toolbox discussion, but then it is most couched as a separate issue on its own (the “shocking news – string theory may be useful for something!” Type articles….)…. But indeed my hope for many years is for more to see that the two issues are intimately connected.
I like the piece, it is more thoughtful than most in this genre. Two points I’d love to see addressed more in this type of discussion: 1. The demarcation of science is a huge subject, and it was not by any means finished in 1934, any more so than any subject of similar depth in say, Physics. In fact, I believe you’d find very few proponents of unmodified Popper in current philosophy departments, and for good reasons I think. 2. Many of us have come to think about string theory as a method rather than a model. This viewpoint certainly has it’s weaknesses, but it is sufficiently common as to merit at least addressing.
Frank Close liked this on Facebook.
RT @BlackPhysicists: On Testability… http://t.co/GhXa9swyRq via @Asymptotia
On Testability… http://t.co/GhXa9swyRq via @Asymptotia