Interview with Kip Thorne — Part 2

0:00 It's the 17th of July, a quarter to two, and I'm talking to Kip Thor. Now let me say Kip Thor. I'm busy trying to get some of the chronological details done. And I remember that you mentioned that at some point in the late 60s to early 70s that you had maybe said in print that you felt that the radiation reaction problem was basically settled and that the quadruple formula was correct and that I think Peter Havis and Morgan Ehlers, you mentioned, at least perhaps Peter Havis, objected to this and that then some of the debate sort of ensued from that. I was sort of wondering when that... Well, there was a meeting in Paris, France. Probably, it was a matter of weeks after Werner Israel had given his first lecture on the No Hair Theorem. So it was, I'm trying to pinpoint it, it has to have been summer of 67. 68 was Invasion of Czechoslovakia. and so it was probably June 1967 there is a proceedings of that meeting probably entitled Ones et Radiation Gravitational that I probably have a copy of I think that was the one. Anyway, at this meeting I did discuss gravitational radiation reaction as more or less a solved problem. and I had some interchange with Peter Havosh

2:30 who was somewhat critical about that kind of a claim but I don't remember much more than that except just a general impression that Peter was put out by the cavalier attitude that I had about the issue. And I'm not 100% sure I've got the date right. There was such a meeting in 67. It's possible that there was another meeting of this sort in 71, roughly, and it may have been there. If I work at it, I can probably pinpoint it for sure. But I would guess it was the 67 meeting. If I go, if you want, turn that off for a minute. That was my principal interaction with him on it from the very, or maybe my first interaction with him and principal one. I think Josh Goldberg was also at that meeting. Peter Bergman was probably there. And there was a moderate amount of discussion. But again, I'll have to check whether it was the 71 meeting or the 67 meeting. Or even 72. But it was maybe at that point that I was first sensitized to the fact that they really wanted a much higher level of rigor than I thought was necessary, and that, or they, particularly Peter, and that, and it was only an impression I had, not anything that was very explicitly stated, because Peter is certainly a gentleman, but an impression I had that he was put out, that he was just coming through the veneer of his gentlemanly way of dealing with me. Now, given the dating question, 67 to 71, what were the main, presumably, maybe there was recent work which made you feel that the issue had... Well, I, by then, I think, had read the paper,

5:00 Or was it to Perez or the paper that dated way back to 64 or so? I guess that's Perez. I by then had read that and was pretty satisfied with it. I had done work of my own by then on... It's published in Physical Review Letters. on gravitational waves from pulsating neutron stars in which I adapted tools that Johnny Wheeler pointed out to me that date back to the theory of alpha decay of atomic nuclei whereby one can build a traveling wave description of something decaying, an alpha particle leaking out of an atomic nucleus on the one hand or the emission of a gravitational wave on the other hand by this pulsating star, one can build that description from standing wave modes and one can see the radiation reaction in effect occurring. So I had published that approach to what anybody else had done in this particular field, but it strongly convinced me. And if this was 71, then Bill Burke had invented the radiation reaction potential that was normally attributed to him and me. He invented this radiation reaction potential but had cold feet about it, and I kept pushing him, and I was the one that publicized it because he had the cold feet, and somehow in the end, my name got associated with it as well as his. It really was his idea. And so that had already been done, and in thinking about that, and I had derived it in the context, again, of neutron star pulsations, but it had given a much more general derivation as well,

7:30 the same tools in this paper in 69. The question is whether this interaction with Peter was before or after that particular piece of work. That work was done basically in 68, probably, with Bill. So it may or may not have preceded this discussion with Peter, but those were the things that I had as a basis for it. There's a fix. I forgot I had to do that. Those were the early pieces of work that convinced me that everything was pretty clean and alright. Let me see, this might be only applicable if it was 71, but was Chandra, Seclar, and Esposito, and that sort of work, influential, or was that above and beyond work? When I was working out my derivation of the radiation reaction potential, I was on leave from Caltech to the University of Chicago, and let me think, I arrived at Caltech in That must have been 68, that was the period January 68, January 68 to July of 1968. I was at the University of Chicago as a guest of Chandra, and I and Bill Burke had previously, so it's probably as even in the fall of 67 had come up with a radiation reaction potential and I was, but had cold feet on it I was working it out for myself in the context of stellar pulsations perturbations of neutron stars and then as I said

10:00 used the same techniques in the Reggie Wheeler gauge to derive it quite generally for weak field slow motion sources and all of that I was doing while I was a guest of Chandra at the University of Chicago in parallel he was working in the early stages of working on this stuff with Esposito and so we had a lot of discussion and the idea that there might be a radiation reaction potential basically I passed on to him from from having gotten it from Bill Burke. And he and Esposito ultimately showed that they could also describe it in that language. Their work, aside from that, was totally independent, pursued on a very independent path from mine, a very independent gauge, but, as I say, with extensive discussion between us during those six months I was at Chicago. Just look here. There are two different conferences in Paris that I was involved in in that era. These are the publication reviews. There are my publications 26 and 53. The dates of the publications are 69 and 74. Uh, and I would guess that this was, uh, the one whose publication date was 69, which, uh, in which case it almost, it probably precedes, well, it may precede the radiation reaction potential. Let me just see if I have copies of those two papers in here. I have them at home. I might have them in here. This preprint, orange jade preprint from August 1967, shows the date of the meeting as being in Paris, June 23, 1967. And at that point, I had done the first piece of work that I had described on using the techniques from alpha particle decay to describe radiation reaction and pulsating stars.

12:30 and I would have guessed that was when Peter and I had the interaction but it seems to me a little more plausible now in retrospect that it was after the second of these meetings which is the one called Onze Radiation Gravitational is the publication that publication is in 74 You'd have to look at the volume itself to see when the meeting occurred, but I would guess now the meeting was in 71 or 72. But I described a lot of other gravitational wave work at that meeting, including, I believe, the radiation reaction potential. Let's just see what all is in here. No, maybe not. So, the radiation reaction potential by that time was fairly old. And so, this does seem to me awfully late. So, my best guess is that it really was the 67 meeting where we had this interaction. But at that time, all I had in hand was this alpha-particle decay argument. Let me just see. I seem not to even, as yet at that point, have, I have the method but I don't have numerical results.

15:00 I can't be sure. I can't be sure which meeting it was. But anyway, I have printed out for you the list of all the other relevant publications that I was involved in in that era. The one with Bill Burke, where we jointly presented the radiation reaction potential at a conference. The conference was held in 1969, I believe, and the publication was in 1970. This conference was in Cincinnati, Ohio. That's publication 33. The publication in which I derived, using Reggie Wheeler Gauge, the radiation reaction potential, which is publication 29. it's buried in this paper on pulsations of relativistic stars but if you look in detail at the derivation of the radiation reaction potential it's done in the form valid for any slow motion weak field system and that was the way I convinced myself that that way of describing radiation reaction was correct and then there's my paper in Physic Review Letters in Publication 21 which was where I derived where I presented the argument based on alpha particle decay and did it also in the context of now having numerical solutions for the normal modes, for the complex normal modes of a pulsating neutron star in particular equates of state so that this was then improving on the thing that I had presented at the conference in 67 where I didn't yet have the numerical results. And then publication 27, which is the detailed paper on the numerical results for stellar pulsations. You mentioned that when Bill Burke first proposed the radiation action potential that you were uncertain about whether...

17:30 Well, he was very uncertain. He was uncertain about communities. It looked to me like a beautiful idea. And as I recall, he was not jumping up and down with enthusiasm nearly as much as he should. his worry was what everybody else's worry was about was a non-linear effect centering in and invalidating it and I just couldn't see any way that you could get into trouble maybe I was naive but it seemed to me quite adequate to be totally convincing to me for slow motion weak field systems I at the time was already working had been working for a while on the stellar pulsation problem so that was the natural place for a natural gauge for me to use for seeing if I could reproduce his result which he had derived by probably in the Donder gauge or Lorentz gauge but I just don't remember He never published it himself independently for several years afterwards until he had to his own satisfaction resolved his worries about non-linear effects. He has two beautiful papers in Physical Review D about radiation reaction. The first one does radiation reaction in simple model problems involving, for example, a pulsating rubber ball, a spherically pulsating rubber ball that's sitting in a fluid and produces sound waves in the fluid. And so he did problems like this where you didn't have gauge problems or any non-linearity questions. and to perfect the method, convince himself the method was fine, and to understand radiation reaction in general. And I have always pointed to that as, to my mind, the very nicest discussion of the general issue of radiation reaction and of the higher order derivatives and of things like the Dirac radiation reaction force

20:00 and not genuinely giving rise to pre-acceleration or problems because solutions that exhibit pre-acceleration or runaway solutions, these things violate the approximations that go into the scheme. And he shows this very, very clearly, that the derivations of these kinds of radiation reaction equations are inherently slow motion derivations and any fast motion solution is invalid since it doesn't satisfy the assumptions that go into the derivation of the equation of motion. So he did all that for himself in other contexts where it was simpler and then ultimately convinced himself that it also worked in general relativity, though I had been convinced almost from the outset. Okay. And his second paper extends this stuff to general relativity. But I've already been pushing this idea that came from him for several, two or three years, before he finally publishes his definitive paper on it. That's interesting. Now, you said, for instance, in passing, that while at such and such a point, the radiation reaction potential idea was already old hat, which I suppose means either that it had become something which people were using a lot, or something that people were perhaps ignoring. I was just wondering, who were the audiences? I mean, on the one hand, you had people who obviously weren't happy with these arguments, but who would be the audience of people who... I don't remember the context in which I said it was old hat. Well, you were wondering if you would have been putting this forward at the meeting in the early 70s. Oh, okay. And so you were saying, would I be announcing this then? Yeah, well, it was certainly old hat in that Burke had had the result, I had had the result, and I had been talking about the result for several years. The extent to which the potential had been accepted by the community, I didn't pay an awful lot of attention to. I mean, basically, I was fully convinced that it was right, that was a valid way to deal with radiation reaction, a very simple way to deal with it.

22:30 And that was basically enough for me. And Chandra came along and agreed once he had completed his work with Esposito, though he had a different form of the radiation reaction potential. And I think it was probably Bonnie Miller working, who was a student of his, who showed clearly the relationship between his version and mine. that they were equivalent modulo approximations that were made in the analysis. But by 1971, I think at least Chandra and I were fully agreed on this, and people around me were using it a bit in calculations. it was I was canonizing it in preliminary versions of MTW though as Cliff Will and Martin Walker pointed out the derivation in MTW was flawed so it was old hat in those senses that I'd been beating the drum long enough that I was sort of sick and tired of it It was just a, as far as I was concerned, it was a useful tool. Certainly, I'm sure, as far as Peter Havosh and people from that community were concerned, it was very suspect, I presume. But I never really discussed these things with them to speak of. Bill Burke did go to the Ereche Summer School, which was several years later, and interacted with them on it. or not. No, but I went and got his email address. I'm trying to do at least right. It probably would be worth talking with him about it and getting his take on what I've been telling you and also his take on the conference at Ariche or at Como. It was at the Enrico Fermi Summer School where that conference that Ehlers organized. Oh, okay.

25:00 But I guess I regard all this as just sort of a little pocket in the long history of the controversy of the field. And as I say, already at that stage when I was doing this work and was looking at literature, I did feel that Paras had basically resolved the issue, that what I was doing was providing new viewpoints on an issue that was more or less in hand. I don't think Chandra probably would agree with that description, but that, I think, was my feeling in that era. New viewpoints and new computational tools. But I'm not 100% sure when it was I became aware of Peres' work and adopted that view. It could have been early, it could have been later. I'm not certain. You'd have to look at when I start citing Peres. Yeah, that's a good idea. Let me see. Well, So, well, this sort of gets bigger, but sometime around 1970 or so, you sort of became aware of the feeling amongst certain people that they weren't happy with the state of affairs at that point. And I know that essentially, in one form or another, People continued arguing about it up until the mid-80s, I guess, at GR10 in 1983. Cooper Scott mentioned that was where he had a showdown with various people when he was presenting his argument about the struts and so on. But then by the next GR conference, apparently, they were more in agreement amongst themselves. So, I mean, do you have anything that you can communicate about the sort of intervening period? Did things just sort of rumble away, or did there were occasional outbursts? Cooperstock and Arnold Rosenblum kept the pot stirred in the late 70s, early 80s, but

27:30 I don't think anybody else was actively arguing that there were problems. Arnold was fairly close to Jürgen Ehlers, and Jürgen, I had an impression, was fairly supportive of Arnold's concerns and was still not satisfied in that era at all that things were adequately at hand. So it was my impression also that Jürgen thought that the answer was probably right, and satisfied with the levels of rigor. But Jürgen was certainly not terribly active at keeping the pot boiling in the way that Arnold and Fred Cooperstock were. The principal defender of the faith during that period was probably Cliff Will or Cliff and Martin Walker. And I, by and large, didn't pay a lot of attention to it. of attention to the controversy at any point through the whole thing in the sense that I felt I understood what was going on to my level of satisfaction and other issues were of greater interest. The one thing that was enlightening to me personally was my work with Shondor Kovach on the Bremstrahlung problem where because we were dealing with a fast motion problem we had to develop fast motion techniques it was then that I really became aware of how extremely difficult the fast motion techniques are and how fraught with perils they were and how hard it was even with fast motion techniques even just to get the leading order radiation reaction. So at that point, I was not surprised that Havas and Hu and others had had terrible difficulties and had been getting wrong answers back in the late 50s, early 60s, whenever it was.

30:00 I mean, and Kovach and I spent several years, enormous amounts of time and effort and sweat getting that right for the remstrelin problem. And I find it remarkable that Havosh and company had the fortitude to attempt to do it by the fast motion technique. and they got as far as they did. I'm not surprised that they got a wrong result having gone through the Bremschlund problem that way. The Bremschlund problem, at least we had to some extent, is some guidance, a perturbation theory analysis that Phil Peters had done with small objects scattering off of Schwarzschild black holes. So he basically used regiweiler techniques. And we were trying to do the same problem for arbitrary mass ratios. So we had some answers to help as guidance that helped us sort things out in the long run. And Havosh and company didn't have anything except what was regarded by most everybody in those days as very unreliable, simple-minded arguments to get the standard result. Well, Rosenblum also focused on, well, he focused on kind of black hole scattering too with his fast motion things. Is there a technical reason why that's easier to handle in the fast motion case than, say, the binary, or why the fast motion is more appropriate for that? The only reason that I used fast motion for that was I wanted to do a fast motion problem. And that is, in the Bremschland problem, I wanted these objects to have speeds that were arbitrary, ranging from slow motion to trans-relativistic to ultra-relativistic. And so there was no choice. I still don't know any other way to do it. And why, I presume, Rosenblum used those kinds of techniques because he had been a student, I believe, of Peter Havish's, and with Peter he was skeptical of the slow motion techniques.

32:30 Yes, I guess so. I was just wondering if there was a reason why both of you ended up doing kind of a scanning. I think that's completely independent. Well, having worked on a similar problem with somewhat similar techniques, did you have any feeling for why, I guess Rosenblum was claiming a different answer? Rosenblum never published the details of his calculations. They were very long, very complicated, and I think he never had the energy, or whatever anyway, to try to put them into a form that was sufficiently concise for publication. So I don't think anybody except maybe Peter Havosh ever really, I don't know whether Peter did. Anyway, conceivably Peter went through the details, conceivably Juergen Ehlers went through the details, but I don't know whether that's the case. Certainly, almost certainly nobody else did because they weren't publicly available. Well, Ehlers did mention to me, recalling Rosenblum sitting through the night and working through these endless things and feeling that he didn't, his personal opinion was that he didn't think Rosenblum was going to be able to carry it through to the end just because it was so far in the same country. Oh, still on this subject of fast person versus slow motion, one thing that confused me greatly when I was starting to look at all this was the question of if you wanted to categorize it in terms of these other things kind of where DeMore fitted in because I guess his approach to the problem of motion is basically post-Minkowski but eventually he Well, it's really post-Newtonian. His approach to gravitational radiation is heavily post-Minkowskian, but that comes later. He's getting deeply into gravitational radiation around the time that he came here to Caltech for six months, nine months, which is early 80s, but probably after Lezouche. He was here, and then the following year, Blanchet was here.

35:00 And they developed their beautiful, basically uniformly, I don't know what the right words are, well-behaved Post-Minkowski approximation and algorithm for solving the Post-Minkowski equations to compute the radiation in far zone and near zone to all orders. They developed that largely when they were visiting here. Previous to that, although post-Nutonkowsky may have been used, I don't really remember details, but may have been used in part as a foundation for equations of motion, their analyses, or Damur's analyses, Damur-Duriel, were fundamentally post-Newtonian. But they developed techniques of their own for dealing with the infinities that arise and for dealing with the possibility that the source has strong internal gravity. These techniques of analytic continuation, the complex plane and so forth. it's my impression that this was all done quite independently of the other of other people's work Damur reads literature heavily but I don't think he was awfully strongly influenced it's my impression he was not awfully strongly influenced by the literature or by other people's approaches he wanted to do it himself in a way that he was fully satisfied with and it would enable him to go up to as high in orders as he chose, basically. And he devised techniques that are probably the best techniques around, or almost certainly the best techniques around for doing our radiation reaction problem for coalescing binaries. Relying on many ideas that came from literature, but generating a number of new ideas of his own. along the way. I don't think he was particularly involved in the controversies along the way. After all, he was beginning this work about the time he finished his PhD and he was really a very independent person

37:30 already, even as a graduate student. Yeah, I asked him, in fact, if any of the area would be the person who was like that, Yeah, I think that's right. It's fair. So, the approach that you mentioned, that he and Blanchet were working on when they were here, that's essentially the same approach that they... What they did here, largely while they were here, was develop the technique to solve the relaxed Einstein equations in a post-Minkowski expansion. This wasn't really the equations of motion. It was assuming that you had some specified motion, which may or may not satisfy the gauge conditions. When you impose the gauge conditions, then you're forcing yourself to get the full solution to the full relativity equations. Let's specify a motion which may or may not satisfy be a correct equation of motion, and now you want to compute the solution to the nonlinear equations that say box H is equal to source. and how do you do that, including all the non-linear terms, do that in a uniformly valid post-Minkowski algorithm that can go to all orders. So they have a pair of papers by the two of them that appear in transactions of the Royal Society of London, or at least the first one's in transactions, the second one may be in proceedings of the Royal Society of London, that is basically the work that they did while they were here. But that becomes one of the foundations for their current approach to things, but it's just one. The other is the issue of how do you actually then get the equations of motion themselves. And that was separate work where most of the foundations for that,

40:00 been done by D'Amour and D'Ariel and are discussed in the Lezouche summer school of 1982 or whatever that year was. I need to go. We can talk some more. Well, just to continue to drag down from where we were, so you can tell I found it a little bit confusing where the post-Munkowski and post-Newtonian parts are in DeMuro. So, would it be truer to say maybe that, so what they do is, as you say, taking the relaxed equations, they approach the radiation problem with the post-Munkowski equation, but then when they have to come back and say what motions are allowed, then they're restricting to a post-Newtonian. Well, they did anyway in the stuff at Les Zouche, preceding Les Zouche of 1982. Even now, the issue of how do you get the equations of motion is not algorithmic. And we had a little discussion with Blanchet about that at Penn State. And so I'm not sure. I'm not clear on how they approach it. get up to these higher orders, clearly they are putting in the radiation reaction in one manner or another. But I suspect it's fundamentally post-Newtonian with the radiation reaction introduced via input that comes from the post-Minkowski analysis. Okay, that's interesting. That sort of clear things up for me. I've been, in my letters, I've been trying to prod Havas to give me his view on the demoral work, and he keeps promising to do so. But I'm interested precisely because I know that one thing that Havis and Bondi always said was that it wasn't clear to them that the equations of motion would allow motions which would produce radiation reaction. So if Havis has doubts about Demorti, which I think he does, it may well reside in that point, I don't know. So I'm just curious to try and get some set on my own mind at least.

42:30 The thing to look at is for the way that things are formulated conceptually not all the details but conceptual formulation and methods is Damur's piece in the Les Housh Proceedings but then keep in mind that he has not really started going to higher order than post 2.5 yet, so he has only the leading order radiation reaction in at that point, and he has not yet really formulated this uniformly valid post-Minkowski approach, which is the key, which is one key to going to higher orders. Yeah, I have been looking through the lasers thing. I've just gone through it and found it, but I guess I've already found it's an interesting thing. You know, De Moor is very good on the history. Yeah. Or at least he's very good at identifying obscure pieces of history. That's right. How accurate he is, I don't know, but certainly in his physics he's extremely accurate, and I would expect he would be in his history. He does seem to be, one of the really obscure pieces he came up with there is that Laplace did a radiation reaction catheter, which I'd vaguely heard about before, and then this inspired me to go and get Laplace's folks we're a bit heavy I think even worse than Ampego but anyway yes I was probably not very interesting let me see let me jump to a completely different topic so that I can at least touch on that one I was going to ask you about the Russian side of things you know here I haven't really much detail, and in any case, it certainly seems that they never had any dispute amongst themselves about this issue. But of course, I've looked at Landau, Nish, and Falk. One thing I was going to ask is, in Russia, did you have, the Soviet Union, I should say, did you have kind of people who specialize in relativity much? I mean, Falk obviously did a lot of work in it, but I know he did a lot of... There is a community of people who specialize in relativity, at least in the period that I was in touch with Russia in the field,

45:00 which was basically 67 onward or 68 onward. I rather doubt that that was the case in the 30s and 40s, and even probably into the 50s, Relativity was ideologically suspect during that period, and it would not have been very safe politically for somebody to specialize in relativity. But I don't know for sure. I only know that there were these problems, ideological problems with relativity in the party. But by 68, or even by early 60s, focusing very heavily on relativity. But those people are largely obscure in the sense that they don't have much reputation in the West, perhaps because relativity during that period was a real backwater. But Zelmanov, for example, who was the mentor for Novikov, Zelmanov was a professor at Moscow State University, I think he was associated with the Sternberg Astronomical Institute. He spent some years developing three plus one splits, which he called the theory of chronometric invariance. And he had considerable influence on Novikov. He clearly was a rather clever man. He was very heavily steeped in this subject, as far as I know. I don't know of work in other areas. There's a man named Mitskevich. Zelmanov was already, by the mid-60s, he must have been.