Friday, May 03, 2019

Graham Farmelo’s interview of Edward Witten. Transcript.

[I’ve meant for some while to try an automatic transcription software, and Graham Farmelo’s interview of Edward Witten (mentioned by Peter Woit) seemed a good occasion. I used an app called “Trint” which seems to work okay. But both the software and I have trouble with Farmelo’s British accent and with Witten’s mumbling. I have marked the places that I didn’t understand with [xxx]. Please leave me a comment in case you can figure out what’s being said. Also notify me of any blunders that I might have missed. Thanks!]

GF [00:00:06] A mind of the brilliance of Edward Witten’s comes along in mathematical physics about once every 50 years if we’re lucky. Since the late 1970s he’s been preeminent among the physicists who are trying to understand the underlying order of the universe. Or, as you might say, trying to discover the most fundamental equations of physics. More than that, by studying the mathematical qualities of nature, Witten became remarkably influential in pure mathematics. The only physicist ever to have won the coveted Fields Medal which has much the same stature in mathematics as a Nobel Prize has in physics.

GF [00:00:46] My name is Graham Farmelo, author of  “The universe speaks in numbers.” Witten is a central figure in my book and he’s been helpful to me. Though he’s a reluctant interviewee so I was pleased when he agreed to talk with me last August about some aspects of his career and the relationship between mathematics and physics. He was in a relaxed mood sitting on a sofa in his office at the Institute for Advanced Study in Princeton wearing his tennis clothes. As usual, he speaks quietly so you’ll have to listen hard.

GF [00:01:20] He uses quite a few technical terms too. But if you’re not familiar with them I suggest that you just let them wash over you. The key thing is to get a sense of Witten’s thinking about the big picture. He is worth it.

GF [00:01:32] He gives us several illuminating insights into how he became interested in state-of-the-art mathematics while remaining a physicist to his fingertips. I began by asking him if he’d always been interested in mathematics and physics.

EW [00:01:47] When I was a kid I was very interested in astronomy. It was the period of the space race and everybody was interested in space. Then, when I was a little older, I was exposed to calculus by my father. And for a while I was very interested in math.

GF [00:02:02] You said for a while, so did that lapse?

EW [00:02:04] Yes, it did lapse for a few years, and the reason it lapsed, I think, was that after being exposed to calculus at the age of eleven it actually was quite a while before I was shown anything that was really more advanced. So I wasn't really aware that there was much more interesting more advanced math. Probably not the only reason, but certainly one reason that my interest lapsed.

GF [00:02:22] Yeah. Were you ever interested in any other subjects? I mean because you know you came on to study history and things like that. Did that really interest you comparably to math and physics?

EW [00:02:31] I guess there was a period when I imagined doing journalism or history or something but at about the age of 21 or 22 I realized that I wasn't going to work out well in my case.

GF [00:02:42] After studying modern languages he worked on George McGovern’s ill fated presidential campaign and even studied economics for one semester before he finally turned to physics.

GF [00:02:53] Apparently he showed up at Princeton University wanting to do a Ph.D. in theoretical physics and they wisely took him on after he made short work of some preliminary exams. Boy did he learn quickly. One of the instructors tasked with teaching him in the lab told me that within three weeks Witten’s questions on the experiments went from basic to brilliant to Nobel level. As a postdoc at Harvard, Witten became acquainted with several of the theorist pioneers of this model including Steven Weinberg, Shelly Glashow, Howard Georgi, and Sydney Coleman, who helped interest the young Witten in the mathematics of these new theories.

EW [00:03:33] The physicists I learned from most during those years were definitely Weinberg, Glashow, Georgi, and Coleman. And they were completely different. So Georgi and Glashow were doing model building, basically weak interaction model building, elaborations on the Standard Model. I found it fascinating but it was a little bit hard to find an entree there. If the world had been a little bit different, I might have made my career doing things like they were doing.

GF [00:04:01] Wow. This was the first time I’d heard Witten say that he was at first expecting to be like most other theorists and take his inspiration from the results of experiments building so-called models of the real world. What, I wondered, led him to change direction and become so mathematical.

EW [00:04:19] Let me provide a little background for listeners. Up to and including the time I was a graduate student, for 20, 25 years, there had been constant waves of new discoveries in elementary particle physics: strange particles, muons, hadronic resonances, parity violations, CP violation, scaling and deep inelastic scattering, the Charm particle, and I’m forgetting a whole bunch. But that’s enough to give you the idea. So that was over a period of over 20 years. So even after a lot of the big discoveries that was one every three years. Now, if experimental surprises and discoveries had continued like that, which at the time I think is what would have happened because it had been going on for a quarter century, then I would have expected to be involved in model building, or grappling with it, like colleagues such as Georgi and Glashow all were doing. Most notably, however, it turned out that this period of constant surprise and turmoil was ending just while I was a graduate and therefore later on I had no successful directions.

GF [00:05:20] Do you remember being disappointed by that in any sense?

EW [00:05:23] Of course I was, you never stop being disappointed.

GF [00:05:27] Oh dear, oh it’s a hard life.

GF [00:05:31] You were disappointed by the drawing up so to speak of the.

EW [00:05:33] There have been important experimental discoveries since then. But the pace has not been quite the same. Although they’ve been very important they’ve been a little bit more abstract in what they teach us and definitely they’ve offered fewer opportunities for model building than was the case in the 60s and 70s. I’d like to just tell you a word or two about my interaction with the other physicists. There was Steve Weinberg and what I remember best from Weinberg. He was one of the pioneers of a subject called current algebra which was an important part of understanding the nuclear force. But he obviously thought most other physicists didn't understand it properly and I was one of those. So whenever current algebra was mentioned at a seminar or a discussion meeting he would always give a short little speech explaining his understanding of it. In my case after hearing those speeches the eight to 10 times [laughter] what Steve was telling us.

EW [00:06:28] Then there was Sidney Coleman. First of all Sidney was the only one who was interested in strong coupling behavior of quantum field theories which is what I’d become interested in as a graduate student with encouragement from my advisor David Gross. So, he was really the only one I could interact with about that. Others regarded strong coupling as a black box. So, maybe for your listeners, I should explain that if you’re a student in physics they teach you what to do when quantum effects are small, but no one tells you what to do when quantum effects are big, there’s no general answer. It’s a smorgasbord of different methods that work for different problems and a lot of problems that are intractable. So, I'd become interested in that as a student but I was mostly beating my head against a brick wall because it is usually intractable, and Sydney was the only one of the professors at Harvard interested in such matters. So, apart from interacting with him about that, also he exposed me to a number of mathematical topics I wouldn’t have known about otherwise but that eventually were important in my work which most physicists didn’t know about. And certainly I didn’t know about.

GF [00:07:27] Yeah, can I ask were you consciously interested in advanced pure math at that time?

EW [00:07:32] Definitely not

GF [00:07:32] You were not?

EW [00:07:32] No, most definitely not. I got dragged into math gradually because you see the standard model had been discovered so the problems in physics were not exactly the same as they had been before. But there were new problems that were opened up by the standard model. For one thing there is new math that came into understanding the standard model. Just when I was finishing graduate school more or less Polyakov and others introduced the Yang-Mills instanton which has proved to be important in understanding physics. It’s also had a lot of mathematical applications.

GF [00:08:02] You can think of instantons as fleeting events that occur in space and time on the subatomic scale. These events are predicted by the theories of the subatomic world known as gauge theories. A key moment in this story is Witten’s first meeting with the great mathematician Michael Atiyah at the Massachusetts Institute of Technology. They will become the leaders of the trend towards a more mathematical approach to our understanding of the world.

EW [00:08:32] So Polyakov and others had discovered the Yang-Mills instanton and it was important in physics and proved to have many other applications. And then Atiyah was one of the mathematicians who discovered amazing mathematical methods that could be used to solve the instanton equations. So he was lecturing about that when he visited in Cambridge. I think in the spring of 1977, but I could be off by a few months, and I was extremely interested. And so we talked about it a lot. I probably made more of an effort to understand the math involved than most of the other physicists did. Anyway this interaction surely led to my learning all kinds of math I’d never heard of before, complex manifolds, sheaf cohomology groups.

GF [00:09:16] This was news to you at that time.

EW [00:09:18] Definitely. So I might tell you at an even more basic level the Atiyah-Singer index theorem had been news to me a few months earlier when I heard about it from Sidney Coleman.

GF [00:09:28] The index theorem first proved by Michael Atiyah and his friend Isidore Singer connects two branches of mathematics that had seemed unconnected. Calculus, that’s the mathematics of changing quantities, and topology about the properties of objects that don’t change when they’re stretched, twisted, or deformed in some way, topology is now central to our understanding of fundamental physics.

EW [00:09:51] Like other physics graduate students of the period, I had no inkling of any 20th century math, really. So, I’d never heard of the names Atiyah and Singer or of the concept of the index or if the index theorem until Albert Schwarz showed that it was relevant to understanding instantons. And even then that paper didn’t make an immediate splash. If Coleman hadn’t pointed it out, I’m not sure how long it would have been before I knew about it. And then there was progress in understanding instanton equations by Atiyah among others. The first actually was Richard Ward, Penrose’s doctoral student. So, I got interested in that but I was interested in a sense in a narrow way which is what good would it be in physics. And I learned the math or some of the math that the teacher was using. But I was a little skeptical about the applicability for physics and I wasn’t really wrong because the original program of Polyakov didn’t quite work out. The details of the Instanton equations that were beautifully elucidated by the mathematicians were not in practice that helpful for things you can actually do as a physicist. So, to sort of summarize what happened in the long run, Atiyah’s work and that of his colleagues made me learn a lot of math I’d never heard of before which turned out to be very important later but not per se for the original reasons.

GF [00:11:10] When did you start to become convinced that math was really going to be interesting?

EW [00:11:14] Well that gradually happend in the 1980s I guess. So, for example one early episode which was in 1981 or two I was trying to understand the properties of what's called the vacuum the quantum ground state in supersymmetric field theories and it really had some behavior that was hard to explain using standard physics ideas and since I couldn't understand it I kept looking at simpler and simpler models and they all had the same puzzle. So finally I got to what seemed like the simplest possible model which you could ask the question and it still had a puzzling behavior. But at a certain point, I think when I was in a swimming pool in Aspen Colorado, I remembered Raoul Bott and actually Atiyah also had given some lectures to physicists a couple of years earlier in Cargesse, and they had tried to explain something called Morse theory to us. I’m sure there are like me many other physicists that have never heard of Morse theory or are familiar with any of the questions it addresses or.

GF [00:12:11] Would you like to say what Morse theory is roughly speaking?

EW [00:12:14] Well if you’ve got a rubber ball floating in space it’s got a lowest point, where the elevation is lowest, it’s got a highest point where the elevation is highest. So it’s got a maximum and a minimum. If you have a more complicated surface like for example a rubber inner tube, it’ll have saddle points of height function as well as a maximum and minimum. And Morse theory relates the maxima and minima and the saddle points of a function such as height function to the topology of a surface or topological manifold on which the function is defined.

GF [00:12:48] You see that paper by Maxwell on that what he spoke about see in 1870.

EW [00:12:52] I’ve not read that.

GF [00:12:53] Oh I’ll show it to you later. It’s “On Hills and Dales,” gave it in Liverpool, very thinly attended talk, erm, anyway.

EW [00:13:01] So was he in fact describing the two dimensional version of Morse theory.

GF [00:13:04] I can’t go into detail but the historians of Morse theory, they often refer to that. At a public meeting incidentally in Liverpool. 

EW [00:13:13] Actually now you mentioned it, I heard the title of the Hills and Dales talk by Maxwell that had something to do with the beginnings of topology. And topology was just barely beginning in roughly that period.

GF [00:13:23] But this was useful in physics. Your Aspen swimming pool revelation...

EW [00:13:28] Well, it shed a little bit of light on the vacuum state in super symmetric quantum theories. So anyway I developed that further so you know at first that seemed exceptional but eventually there were too many of these exceptions to completely ignore.

GF [00:13:42] Am I right in saying, not to put into your mouth, but it was the advent of String Theory post Michael Greene and John Schwartz where these things started going front and center, is that fair?

EW [00:13:50] After... Following the first super string revolution as people call it which came to fruition in 1984 with the work of Greene and Schwarz on the anomalies after that the sort of math that Atiyah and others had used for the instaton equation was suddenly actually useful. Because to understand string theory, complex manifolds and index theory sheaf cohomology groups, all those funny things were actually useful in doing basic things like constructing models of the elementary particles in string theory. I should give a slightly better explanation. In physics there are the forces that we see for the elementary particles that means basically everything except gravity. Then there's gravity which is so weak that we only see it for macroscopic masses like the earth or the sun. Now we describe gravity by Einstein's theory and then we describe the rest of it by quantum field theory. It's difficult to combine the two together. Before 1984 you couldn't even make a halfway reasonable models for elementary particles that included all the forces together with gravity. The advance that Greene and Schwarz made with anomaly cancellation in 1984 made that possible. But to make such models you needed to use a lot of the math that physicists had not used previously but which was introduced by Atiyah and others when they solved the instanton equations and you had to use complex manifolds, sheaf cohomology groups and things that were totally alien to the education of a physics graduate student back in the days when I'd been a student. So those things were useful even at a basic level in making a model of the elementary particles with gravity. And if you wanted to understand it more deeply you ended up using still more maths. After string theory was developed enough that you could use it in an interesting way to make models of particle physics it was clear that a lot of previously unfamiliar math was important. I speak loosely when I say previously unfamiliar because obviously it was familiar to some people. First of all the to the mathematicians. Secondly in some areas like Penrose had used some of it in his Twister theory. But broadly speaking unfamiliar to most physicists.

GF [00:15:46] So we actually went very well in physics very very important for mathematicians in mathematics a very important physicist they're working harmoniously alongside each other. You go back to Leibnitz who used to talk about the pre established harmony between math and physics. That was one of Einstein's favorite phrases. Is there something you regard as a fact of life or is it something you would regard as possibly can be explained one day will never be explained. Do you have any comment at all on that relationship.

EW [00:16:09] Well, the intimate tie between math and physics seems to be a fact of life. I can't imagine what it would mean to explain it. The world only seems to be based on theories that involve interesting math and a lot of interesting math is at least partly inspired by the role that it plays in physics. Not all of course.

GF [00:16:25] But does it inspire you when you see a piece of math that's very relevant to physics and vice versa when you're helping mathematicians. Does that motivate you in some way to think you're on the right track.

EW [00:16:35] Well when something turns out to be beautiful that does encourage you believe that it's on the right track.

GF [00:16:39] Classic Dirac. But he took it as he put it to almost a religion. But I sense you are  a little bit more skeptical, if that's the right word or hard nosed about it I don't know.

EW [00:16:51] Having discovered the Dirac equation, Dirac was entitled to commit its use to extremes, to put it that way.

GF [00:16:58] Witten has long been a leading pioneer of the string framework which seeks to give a unified account of all the fundamental forces based on quantum mechanics and special relativity. It describes the basic entities of nature in terms of tiny pieces of string.

GF [00:17:14] Go back to string theory. Do you see that as one among several candidates or the preeminent candidate or what? I mean what do you see the status of that framework in the landscape of mathematical physics.

EW [00:17:24] Id say that string slash M theory is the only really interesting direction we have for going beyond the established framework of physics by which I mean quantum field theory at the quantum level and classical general relativity at the macroscopic scale. So where where we've made progress that's been in the string slash M theory framework where a lot of interesting things have been discovered. I'd say that there's a lot of interesting things we don't understand at all.

EW [00:17:48] But you’ve never been tempted down the other route. The other options are not.

EW [00:17:52] I’m not even sure what you would mean by other routes.

GF [00:17:54] Loop quantum gravity?

EW [00:17:56] Those are just words. There aren’t any other routes.

GF [00:17:58] Okay, all right, fair enough.

GF [00:18:01]  So there we have it. The preternaturally cautious Witten says that if we want to discover a unified theory of all the fundamental forces, string theory is the only interesting way forward that’s arisen.

GF [00:18:17] Where we are now strikes me as being quite an unusual time in particle physics because so many of us were looking forward to the Large Hadron Collider, huge energy available ,and finding the Higgs boson and maybe supersymmetry. And yet it seems that we have gotten the Higgs particle just as we were hoping and expecting. But nothing else that’s really stimulating. What are your views on where we are now?

EW [00:18:39] My generation grew up with a belief very very strong belief which by the way was drummed into us by Steven Weinberg and by others. That when physics reached the energy scale at which you can understand the weak interactions. You would not only discover the mechanism of electroweak symmetry breaking but you’d learn what fixes its energy scale as been relatively low compared to the scale of gravity. That’s what ultimately makes gravity so weak in ordinary terms. So, it came as a big surprise that we reached the energy scale to study the W and the Z and even the Higgs particle without finding a bigger mechanism behind it. That’s an extremely shocking development in the context of the thinking that I grew up with.

EW [00:19:22] There is another shock which also occurred during that 40 year period which possibly should be comparative. This is the discovery that the acceleration of the expansion of the universe. For decades physicists assumed that because of the gravitational attraction of matter the expansion of the universe with be slowing down and tried to measure it. It turned out that the expansion is actually speeding up. We don't know this for sure it seems quite likely that the results from the effects of Einstein's cosmological constant which is incredibly small but non-zero. The two things the very very small but non-zero cosmological constant and the scale of weak interactions the scale of elementary particle masses which in human terms can seem like a lot of energies. But it's very small compared to other energies in physics. The two puzzles are analogous and they're both extremely bothersome. These two puzzles although primarily the one about gravity which was discovered first are perhaps the main motivation for discussions of a cosmic landscape of vacua. Which is an idea that used to make me extremely uncomfortable and unhappy. I guess because of the challenge it poses to trying to understand the universe and the possibly unfortunate implications for our distant descendants tens of billions of years from now. I guess I ultimately made my peace with it recognizing that the universe hadn't been created for our convenience.

GF [00:20:43] So  you come to terms with it.

EW [00:20:45] I've come to terms with the landscape idea and the sense of not being upset about it. As I was for many years.

GF [00:20:49] Really upset?

EW [00:20:50] I still would prefer to have a different explanation but it doesn't upset me personally to the extent it used to.

GF [00:20:56] So just to conclude what would you say the principal challenge is all down to people looking at fundamental physics.

EW [00:21:01] I think it's quite possible that new observations either in astronomy or accelerators will turn up new and more down to earth challenges. But with what we have now and also with my own personal inclinations it's hard to avoid answering new terms of cosmic challenges. I actually believe that string slash M theory is on the right track toward a more deeper explanation. But at a very fundamental level it's not well understood. And I'm not even confident that we have a good concept of what sort of thing is missing or where to find it. The reason I'm not is that in hindsight it's clear that a view we might have given in the 1980s was what was missing was too narrow. Instead of discovering what we thought was missing instead we broadened the picture in the 90s in unexpected directions. And having lived through that I feel it might happen again.

EW [00:21:49] To give you a slightly less cosmic answer if you ask me where I think is the most likely direction for another major theoretical upheaval like happened in the 80s and then again in the 90s. I've come to believe that the whole it from qbit stuff, the relation between geometry and entanglement, is the most interesting direction.

GF [00:22:12] It from bit that was a phrase coined by the late American theoretician John Wheeler who guessed that the stuff of nature the "it" might ultimately be built from the bits of information. Perhaps the theory of information is showing us the best way forward in fundamental physics. Witten is usually wary of making strong pronouncements about the future of his subjects. So I was struck by his interest in this line of inquiry, now extremely popular.

EW [00:22:39] I feel that if in my active career there will be another real upheaval that's where it's most likely to be coming [xxx]

EW [00:22:47] I had a sense both in the early 80s and in the early 90s. I had a sense a couple of years in advance of the big upheavals where they were most likely to come from it and those two times did turn out to be right. Then for a long long time I had no idea where another upheaval might come from. By the last few years I've become convinced that it's most likely to be the it from qbit stuff of which I have not been a pioneer now. But I was not one of the first to reach the conclusion or a suspicion that I'm telling you right now. But anyway it's the view I've come to.

GF [00:23:20] There's a famous book about night thoughts of a quantum physics. are there night thoughts of a string theorists is where you have a wonderful theory that's developing you know unable to test it. Does that ever bother you.

EW [00:23:31] Of course it bothers us but we have to live with our existential condition. But let's backtrack 34 years. So in the early 80s there were a lot of hints that something important was happening in string theory but once Greene and Schwartz discovered the anomaly cancellation and it became possible to make models of elementary particle physics unified with gravity. From then I thought the direction was clear. But some senior physicists rejected it completely on the grounds that it would supposedly be untestable. Or even have cracked it would be too hard to understand. My view at the time was that when we reached the energies of the W, Z and the Higgs particle we'd get all kinds of fantastic new clues.

EW [00:24:11] So. I found it very very surprising that any colleagues would be so convinced that you wouldn't be able to get important clues that would shed light on the validity of a fundamental new theory that might in fact be valid. Now if you analyze that 34 years later I'm tempted to say we were both a little bit wrong. So the scale of clues that I thought would materialize from accelerators has not come. In fact the most important clue possibly is that we've confirmed the standard model without getting what we fully expected would come with him. And as I told you earlier that might be a clue concerning the landscape. I think the flaw in the thinking of the critics though is that while it's a shame that the period of incredible turmoil and constant experiment and discovery that existed until roughly when I started graduate school hasn't continued. I think that the progress which has been made in physics since 1984 is much greater than it would have been if the naysayers had been heeded and string theory hadn't been done in that period.

GF [00:25:11] And it's had this bonus of benefiting mathematics as well.

EW [00:25:14] Mathematics and by now even in other areas of physics because for example new ideas about black hole of thermodynamics have influenced areas of condensed metaphysics* even in the study of quantum phase transitions, quantum chaos and really other areas.

GF [00:25:31] Well let's hope we all live to see some revolutionary triumph that was completely unexpected that's the best one of all. Edward thank you very much indeed.

EW [00:25:38] Sure thing.

GF [00:25:43] I’m always struck by the precision with which Edward expresses himself and by his avoidance of fuzzy philosophical talk. He's plainly fascinated by the closeness of the relationship between fundamental physics and pure mathematics. He isn't prepared to go further to say that their relationship is a fact of life. Yet no one has done more to demonstrate that not only is mathematics unreasonably effective in physics physics is unreasonably effective in mathematics.

GF [00:26:15] This Witten said makes sense only if our modern theories are on the right track. One last point. Amazingly Witten is sometimes underestimated by physicists who characterize him as a mathematician, someone who has only a passing interest in physics. This is quite wrong. When I talk with a great theoretician Steven Weinberg he told me of his awe at Witten's physical intuition and elsewhere said that Witten's got more mathematical muscles in his head than I like to think about. You can find out more about Witten and his work in my book "The universe speaks in numbers.


* Condensed matter physics. I am sure he says condensed matter physics. But really I think condensed metaphysics fits better.


  1. Not really important but the final EW comment actually appears to be by GF, a continuation of the previous paragraph.

    1. That's right of course, I fixed that. Thanks for pointing out.

  2. Well...that last comment is labeled EW and it’s certainly not...

  3. This doesn't need to be published.
    I didn't listen to all of the interview, but to react to your request about [xxx] in interview transcription (in some places there are a few words from what was already transcribed, I hope this makes some sort of sense):
    12.14: rubber inner tube? (kind of makes sense)
    12.53: thinly attended talk, erm, anyway
    13.04: but that - historians of Morse theory - they
    13.13: that hills and dales talk
    13.42: started going front and centre
    13.50: more maths (transcription has months, there are probably other places where similar incorrect transcription has occurred)
    And two more where I couldn't completely fill in the gaps:
    15.46: in his ---
    16.51: commit to his ---

    1. ksd,

      Thanks so much for this.

      12:14 The "rubber inner tube" doesn't really make sense to me.

      12:53 I changed that.

      13:04 Yes! I'd never have figured that out.

      13:13 What's hills and dales? Doesn't make any sense to me, help!

      13:42 Okay. Is that a thing people say? Never heard this before.

      13:50 I changed this

      15:46 I don't know what this refers to, sorry, can you be more specific?

      16:51 There still seems to be a word missing. What I hear is "was very entitled to commit to his strings, to put it that way" but that doesn't make any sense.

    2. Regarding "Hills and Dales", I believe this must have been the title of the talk, at least he wrote a paper with that title. I added a reference for this.

    3. A "rubber inner tube" is a donut-shaped object that used to go inside an automobile tire. You may have to be fairly old to remember them ... they make tires differently now. Try googling it on "Google images".

    4. Peter,

      Thanks for clearing this up. I had no idea this is an actual word, sorry!

  4. First [xxx] (in EW [00:09:51]): "Richard Ward, Penrose's doctoral student".

    Here are a few mistakes that aren't tagged with [xxx]:

    In the first paragraph, GF says "More than that..." not "Or on that..."

    EW [00:04:19]: He says "even though I forgot [or left out?] a lot of the big discoveries", not "even after I'd ordered the big discoveries". Also in the same paragraph he says "which at the time I assumed would happen because it had been going on for a quarter century", not "which at the time I think what happened was even going on for a quarter century" (!).

    EW [00:05:33]: Replace "General Steve Weinberg" with "There was Steve Weinberg" (and put a full stop between "physicists" and "There"). Also replace "discussion meaning" with "discussion meeting".

    EW [00:06:28]: Get rid of the first full stop (the sentence doesn't end there) and replace "my advisor David Gross said" with "my advisor David Gross, so". Replace "they teach you what to do when quantum effects were small but no one tells you what to do when you don't have extra big there's no general answer" with "they teach you what to do when quantum effects are small, but no one tells you what to do when quantum effects are big, there's no general answer".

    EW [00:07:32]: Replace "there is no math" with "there was new math".

    GF [00:08:02]: The first word ("applications") belongs to the end of the previous paragraph and is spoken by EW not GF. Replace "Witness" with "Witten's".

    EW [00:08:32]: Replace "could go on for a few months" with "but I could be off by a few months". Replace "cosmology" with "cohomology".

    That's all I have for now (stopped listening at around the 10 minute mark). I will probably come back to this later...

    1. Hi Jesse,

      Thanks so much. I had meanwhile made several of these changes already, but added the rest you point out. Esp the "doctoral student" I couldn't figure out!

      If you come back, please make sure to reload the page.

    2. Condensed metaphysics! Please don't change it!! 😄

    3. Haha, I hadn't seen that. I actually thought you were saying the whole interview is "condensed metaphysics".

  5. Thank you for the transcription. I much prefer reading then listening to a talk or video.

  6. Hello Sabine,
    I note on occasion you use a very dry/drol sense of humor, so when you said "I used an app called “Trint” which seems to works okay." in your opening paragraph I didn't know if you were just drawing attention to transcribing software making mistakes or if you meant to say 'seems to work okay' instead of 'seems to works okay'

    1. CFT,

      Sorry, typo, I fixed that. Thanks for pointing out.

  7. I gave a quick listen to some xxx you have marked and I got this:
    but not [xxx] for the original reasons
    but not per se for the original reasons

    it still had a puzzling [xxx].
    it still had a puzzling behaviour.

    [xxx] say what morse
    Would you like to say what morse

    like for example a [xxx]
    this could be - rubber interdune - but I am not sure about the second word.


  8. Replies
    1. Shippey,

      You can listen to it on this website. Or use the link in the first sentence of my blogpost.

  9. from the transcript

    EW [00:17:48] But you’ve never been tempted down the other route. The other options are not.

    EW [00:17:52] I’m not even sure what you would mean by other routes.

    GF [00:17:54] Loop quantum gravity?

    EW [00:17:56] Those are just words. There aren’t any other routes.


    do QG researchers typically only focus on strings and completely ignore LQG?

    any idea why Witten is so anti-LQG?

    1. I don't know Witten, so, I am sorry, I do not know what his misgivings are about LQG. A lot of string theorists seem to think that LQG doesn't have a graviton propagator and hold that against the theory. What I find somewhat more perplexing though is that Witten doesn't seem to be aware of asymptotically safe gravity, or at least doesn't consider it to be viable for whatever reason.

    2. what if nature isn't super symmetric or is only 4 dimensional, it seems Witten is putting all his eggs on an unproven basket.

  10. Thanks for transcribing this!

    At 13:50 "sheaf cosmology groups" has got to be "sheaf cohomology groups", because the latter exist and the former don't.

    1. Hi John,

      Thanks for pointing out. I had fixed this in a few places, but must have missed it in other. The "cohomology" generally came out as cosmology. And sometimes it was sheep cosmology :o) Oh, and Witten usually was a "witness".

    2. sheaf cohomology was the most important chunk. Here some tiny pieces:
      - But nothing else that’s really [xxx]. -> … really stimulating
      - which possibly should be [xxx]. -> … should be comparative.

    3. Interesting, that Witten was shocked that no supersymmetry was found at the LHC. I was so deeply impressed at the time by Alain Connes' derivation of the Standard Model, which doesn't require supersymmetry, that I wrote in Sept. 2010:
      "What will not be found: Supersymmetric particles (and I would bet a fortune one that)."

    4. Yes. I'm shocked Witten was shocked.

      We've known all along the Standard Model does not require supersymmetry. What has Connes to do with that?

    5. I would have been shocked if supersymmetry had been discovered, since I used string theory to prove that supersymmetry will not be discovered at the LHC.

    6. Connes said:
      "This prediction (mass of the Higgs particle) was based on the hypothesis of the 'big desert', namely that there will be no new physics up to the unification
      scale, besides the Standard Model coupled to gravity. Thus it was a bit like trying to see a fly in a cup of tea by looking at the earth from another planetary system. But very strangely the model also predicted the correct mass of the top quark, and a surprising number of mechanisms such as the Higgs and the see saw mechanisms."
      Moreover, he wrote:
      "The noncommutative approach predicts all the fermionic and bosonic spectrum of the standard model, and the correct representations. One can also take as a prediction that there are no other particles to be discovered, except for the three scalar fields: the Higgs field, the singlet field and the dilaton field."

    7. Except that his prediction for the Higgs mass was wrong. In any case, we have all known since the invention of supersymmetry that the standard model does not require supersymmetry.

    8. His prediction was actually right, he only mistakenly neglected a term.
      The important point of his derivation is that the spectral action applies to the GUT scale and it contains no supersymmetry. He then "runs it down" to the SM energy scale. This is all extremely minimalistic and not the kind of Rube Goldberg machine supersymmetric people prefer.

    9. "Except that his prediction for the Higgs mass was wrong."
      That was already corrected here:

    10. Sure, experiment falsified an earlier (2007!) noncommutative Standard Model, but progress has been made in more recent years, pointing to physics beyond the Standard Model while staying compatible with experiment. It is a small group of mathematical/theoretical physicists working on this, but definitely worth taking a look at if you ask me ;-)

      Thanks for the transcript!

    11. If you predict a measurement after it's been made it's not a prediction.

    12. But it just corrected an earlier calculation. The math was already there, he didn't change the formalism after the discovery of the Higgs! Look, Sabine, if you find any fault with Connes' math, let us know. But otherwise, the intellectually honest position would be to remain in silence. I'm fed up with the same rubbish argument (i.e. 'he corrected the prediction after the measurement'), especially comming from people who aren't even acquainted with the relevant math (operator algebra, spin manifolds, spectral triples, etc).

    13. Of course it's not. But the interesting thing is that the scalar field they used to correct the wrong initial prediction was already there in the original papers, i.e., *before* the measurement, they simply ignored its contribution for the wrong reasons. Anyway, it's not my intent to enter into the semantics of what is and what is not a correct prediction, but just to point out that the story is more interesting than a mere wrong prediction.

    14. I read several of Connes' papers. I will not claim that I understood them, but I understood enough to know that he did not just correct a mistake. He did, as a matter of fact, correct the prediction after the measurement, regardless of how much you dislike that.

    15. Yes, but what he did is not the usual "make the mass bigger" that susy advocates do whenever experiment shows no superpartners. He did, as a matter of fact, correct the prediction after the measurement but by taking into account elements that were already predicted by his model and which he neglected in his initial calculation, regardless of how much you dislike that.

    16. I'd like to be precise at this point: in 2012 Ali Chamseddine and Alain Connes did not "correct the prediction after the measurement", but merely found a natural extension of the Standard Model formulated in the same formalism of noncommutative geometry that "restores the consistency of the noncommutative geometric model with the low Higgs mass" (last phrase in the abstract of their arXiv:1208.1030, and confirmed by others)

    17. It's not even a retrodiction, because there's a new parameter that has to be adjusted (value depends on the unification scale) to get 125 GeV - the constant n, introduced on page 3, also see figure 2 on page 9. What changed is that such a Higgs mass was actually inconsistent with their simpler model ("invalidating the positivity of the coupling at unification which is an essential prediction of the spectral action", page 2), but is consistent with the expanded model.

    18. aleazk,

      You are wrong in thinking that I dislike that. What I dislike are false statements, like the ones you have made. I suggest you stop doing it.

    19. Walter,

      Well, if you want to say that they argued it's no longer a prediction, fine with me. In any case, the statement that Connes' prediction for the Higgs mas was right is incorrect. The only reason I can think of that Markus would make such a statement is that he was hoping no one here would know any better.

      Fwiw, the only correct prediction of the Higgs mass that I know of is that by Wetterich et al, which no one seems to pay any attention to. Why is that, I ask you? Is it that asymptotically safe gravity isn't pretty enough because it doesn't unify three forces into one unified force?

    20. "The only reason I can think of that Markus would make such a statement is that he was hoping no one here would know any better."
      No, I am just trying to rephrase what Connes said:

    21. "I am just trying to rephrase what Connes said"

      Okay, thanks. That certainly explains it.

    22. @Mitchell Thanks for that hint. That figure 2 looks more like a "Higgs mass landscape" than a Higgs mass prediction :-)
      @Sabine I agree, no prediction here it now seems to me.

    23. It's indeed not the best way to frame it as a "prediction" or "retrodiction" or whatever of a *single value* for the mass. What the model gives is a certain range of possible mass values that are compatible with the model. But the model is tight enough that this range is bounded from both sides and, therefore, there are values which are incompatible with the model and that would falsify it if observed. And, indeed, in their first model, the range of compatible values was 160 to 180 GeV. That model was, of course, falsified by the observation of the value being 125 GeV. Now, they came back to their calculations and realized that they ignored the full contribution of a scalar field that was already predicted by the spectral action (one must realize that the spectral action and some other theorems, severely constraint the field content of the model, one cannot add just whatever field one desires.) With the full contribution of this field (which involves a minimal extension of the SM) now being taken into account, the range of mass values (for the Higgs field) compatible with the model now contains the observed value.

    24. "Is it that asymptotically safe gravity isn't pretty enough because it doesn't unify three forces into one unified force?"

      As you have argued before, the quest for unification per se is not completely justified in terms of *directly* looking for the answer of a well posed problem (which is defined as a contradiction between different parts of our current knowledge.) But, one could argue that unification (of all forces, gravity included) can give clues to the resolution of an actual problem, quantuam gravity, since it would imply that at that energy scale things are much more subtle than the mere quantization of gravity alone. Now, you could say, why overcomplicating it in that way? We don't know if unification is correct, you are just adding an additional, ill-motivated, and potentially superfluous thing to the original problem. That would be true if the proposed unification involves currently unobserved features (sure, it may be falsifiable, but we could add pretty much any thing to the problem if we follow that logic.) And this is indeed the case with many of the popular unifications. But, what makes Connes' model interesting, is that it's pretty much a fully geometrical *reformulation* of the SM, without any additional and controversial hypothesis (with the exception, perhaps, of that additional field, but it's a minimal extension, i.e., no new spacetime dimensions or an arbitrary amount of new unobserved particles.) That is, the "unification" is practically already here, with the currently observed data. That's really a very interesting discovery that is not getting the deserved attention. And, unlike what you tacitly suggest in your comparisom with asymptotically safe gravity, Connes' model is actually a marginal research area, with only Connes himself, and collaborators, and a small group in the Netherlands working on it (as far as I know.)

    25. "... the only correct prediction of the Higgs mass that I know of is that by Wetterich et al, which no one seems to pay any attention to."
      At least I did so early on and was pretty impressed by their prediction. But as in the case of the noncommutative standard model, no low energy supersymmetry is required. So why was Witten so shocked? Would there be a problem if supersymmetry was broken at very high energies instead?

    26. MarkusM,

      No, it would not be a problem for string theory if supersymmetry was broken at much higher energies. I explained this here.

      It it a problem for gauge-coupling unification, if that's something you believe in.

      Yes, why was Witten shocked? I can only guess that he actually believed in naturalness arguments. Someone should tell him to read my book ;o)

  11. @11:50 ...But at a certain point, I think when I was in a swimming pool in Aspen Colorado, I remembered [xxx]. and actually Atiyah...

    xxx is Raoul Bott a Hungarian-American mathematician who worked with Atiyah.

  12. Around 15 minutes I think he's saying "Secondly in some areas like Penrose had used some of it in his twistor theory."

    1. yes, now I also can hear it.
      In “Road to Reality” there is the chapter “33.9 Twistor sheaf cohomology” - it is about gluing of patches …
      It is unbelievable, that only with the right (predictive) model we can see/hear what someone or nature is telling us all the time.

    2. Witten developed what has been called the twistor-string mini-revolution. Twistor theory is CP^3 with the isometry group SU(2,2). The projective twistor space CP^3 ~ SU(2,2)/SO(4,1)×U(1) with is isotopy group SO(4,1). This connects with the Maldecena AdS/CFT and M-theory.

      If you want to read a really great paper by Witten that does not lean on string theory that heavily I suggest Three Dimensional Gravity Revisited, or a title of that sort. It is almost poetry in my opinion on how Witten links results of analysis and topology with physical theory.

      Theoretical physics has a lot of structure developed, which runs from twistor theory to supersymmetry and string to unification based on group theory to attempts to make forms of general relativity quantizable in a direct setting --- LQG or dynamic triangles etc. We have no general anchor for this. There is a sort of crisis here, which in one sense is a wonderful thing to be facing.

    3. Lawrence,

      This Witten paper is indeed impressive how he juggles with the various mathematical structures. It would take me some time to understand most of the details and to grasp what you already can sense as almost poetry.
      It is impressive, but on the other hand, it is also very restrictive since the Einstein-Hilbert action lives in D=2+1 and needs Λ<0. We live in D=3+1 with Λ>0. Also, the CFT in AdS/CFT is restricted to D=2.
      (And the conformal algebra in D = 2 has an infinite-dimensional extension, the Virasoro algebra, unlike in D=3 or D=4 where it has finitely many generators.)

      One fascinating application of CFT are continuous/(second order) phase transitions, where scale invariance is actually taking place. Already 1970 Polyakov conjectured not only scale but more general conformal invariance [1]. Recently [2, 3] the additional symmetry in the special conformal transformation was used to calculate critical exponents in D=3.
      Renormalization group and lattice methods work together to explore the ‘“kink” on the boundary of the region allowed by the constraints of crossing symmetry and unitarity.’ [3].

      The transition to Bose-Einstein condensate (BEC) and the superfluid transition of helium belong to the same universality class. Both are in O(N) with N=2. Phases in statistical field theory are characterized by the symmetry G of the free energy and the symmetry H of the ground state. The free energy, aka Wilsonian effective action represents the struggle between energy E and entropy S in F=E-TS (F=log Z).
      I find it fascination that a second order phase transition washes away all memory of the underlaying details, whether it is He or rubidium-87 atoms. Kind of a major information loss.
      (I should have added here also the link where Sabine talks in her book about decoupling, effective field theories, i.e. the renormalisation group)

      You said “There is a sort of crisis here, which in one sense is a wonderful thing to be facing.”
      Yes, indeed, it is and I guess we have to rethink sentences like “information is never lost” not only at the horizon of forming and evaporating black holes, a topological transition.

      ceterum censeo:
      Our universe is not evolving exclusive unitarily – so far, we ignored the measurement problem.
      And (observer independent triggered) measurements are the source for random flukes all the time.

      The variable “time” in QM is just used to calculate probability amplitudes and this of course must be unitary.
      And we already know how Fermions and Bosons evolve on a curved non-quantized spacetime.
      This separation of QM “time” immediately makes it plausible why virtual particles are, well, virtual, way off mass-shell and entanglement simply is non-local.
      This all has to do with why it is that cyclic imaginary time is connected to temperature, besides being just an analytic continuation …
      Whether this separated QM “time” also has some connection with the second “time” variable in AdS, i.e. SO(2,2) would in principle be an interesting question, but I for now (and for reasons of limited time ;-) just want to understand the world we are living in.

      [1] A. M. Polyakov, “Conformal symmetry of critical fluctuations" (pdf)

  13. At 12:14:

    "...example a [xxx], it’ll have saddle..." definitely sounds like "rubber inner tube" to my English ears.

    In the earlier sentence he talks of a rubber ball, so I guess he's contrasting a rubber ball with a rubber inner tube (presumably it's a topology thing about a torus versus a sphere).

  14. Bee,

    Witten states he was 11 when he learned calculus. Is this common for physics professors and how old were you when you learned it?

  15. This is what I make out:

    12:14 rubber inner tube

    13:42 not to put words into your mouth

    16:51 Having discovered the Dirac equation Dirac was entitled to commit its use to extremes, let's put it that way.

  16. Thanks for this report from the enemy, i.e. those day dreaming, 'head-the-clouds' string theorists with their infinity of unfalsifiable possibilities. :-)

    EW [00:16:35] ...doesn't hurt you believe, sounds like ...does encourage you to believe.

    EW [00:25:14] ...have influenced areas of condensed metaphysics, sounds more like ...have influenced areas of condensed matter physics, to my ear at least.

    But who knows, maybe Ted is suggesting that mathematics has somehow bolstered the assertions of the monotheisticly inclined dispensations, ...or something. :-)

  17. I liked the 'areas of condensed metaphysics' at 00:25:14,
    guess it means more vonventional condensec matter physics

  18. Without listening:
    EW [00:17:24] ... scale. So where [where] we've made progress that's been in the string slash M theory framework ...
    (probably doubled)

  19. GF [00:15:46] ... Is there something you regard as a fact of life or is it something you would regard as possibly can be explained one day [word missing here?] will never be explained. ...

  20. I have read a fair number of Witten's papers. They are long, but they are very approachable. He presents these advanced concepts in a clear and concise way.

    He points to something I have been pondering. He states:

    EW [00:21:01] ... I actually believe that string slash M theory is on the right track toward a more deeper explanation. But at a very fundamental level it's not well understood. And I'm not even confident that we have a good concept of what sort of thing is missing or where to find it.

    The connection between string types, with S. T and other transformations, is on the boundary of somethings else. The most general M-theory is missing. Will that clear things up or will we end up with a far more ponderous and never ending or closing set of problems? Who knows.

  21. In other words :

    EW : " I know nothing, I come from Barcelona "

  22. Is Edward Witten a hero that hurts science?

  23. Why even bother with Witten and co ? He may have an IQ of 190 or so, but since he's been barking up the wrong tree and will continue doing so till his final breath, what he says is a complete waste of time.

  24. I noticed there are few more xxx remaining, so I checked those out. I got:

    some of it [xxx] > some of it in (his) Twistor theory. - can not make out if there is "his" in between, and name of theory found through hearing+Googling. :)

    The last one is really tough ... Tried to reduce speed of the audio by 10, 20, 30 and 50 percent and listen to it, but could not make sense.

  25. GF 2:53 Sydney <— Sidney
    EW 6:28 Sydney <— Sidney
    GF 13:42 Greene <— Green & Schwartz <— Schwarz
    EW 13:50 Greene <— Green (2 occurrences)
    EW 23:31 Greene <— Green & Schwartz <— Schwarz

  26. Hi Sabine, !
    Just getting to this (been busy)
    Thanks for the transcript. I didn't have much trouble figuring out the'xxx's. Nice program anyway.
    Witten, what a wonderful mind.
    - always a pleasure to to
    If you'll permit me; a brief remark to the ,'peanut gallery'.
    @ Dennis, - regardless frames
    of reference (age, generation,
    How about this;
    YOU go ahead and win
    a Fields Award.
    ... and,
    we'll let you piss
    on any tree you want.
    - you don't even have to bark.
    --. smartass.

    Also, @ Adrian
    - indeed, a topological
    dimensional reference. (difference between a sphere and torus) basic, you can twist a torus any way you want and the surface equations stay the same.
    ( not to get all 'Callabi-Yau'
    on ya ;-)


  27. also, 'hills and dales',
    - part of 'first generation'
    topology (c. 1870 ?)
    - really cool.
    Thanks again, Sabine.
    Ich hoffe du hast einen
    Guten Tag.
    - Love Your Work.

  28. Could it be that the 126 GeV prediction of Shaposhnikov & Wetterich for the Higgs mass (computed with a 173 GeV top quark mass)* gets more attention with the last measurements of a 171 GeV top quark mass reported by CMS & ATLAS from LHC data at 13 TeV

    *Shaposhnikov & Wetterich mention a former pretty correct Higgs mass prediction (with different hypothesis) by Frogatt & Nielsen



Comment moderation on this blog is turned on.
Submitted comments will only appear after manual approval, which can take up to 24 hours.
Comments posted as "Unknown" go straight to junk. You may have to click on the orange-white blogger icon next to your name to change to a different account.