Interview with Stuart Shapiro (contd.)

0:00 There's also this interesting result that was in mathematics, and the... I recall that the reason why it drew our attention immediately was the earliest work that we had done in the early 90s that I showed you the Apgé supplement paper where in Newtonian physics, in an ellipsoidal approximation, we looked at what happens to the stability of a star when put in a binary equilibrium. In particular, we calculated a critical adiabatic index for radial collapse, which is four-thirds for an isolated spherical star. That is, if the star has a gamma-low equation of state, where gamma is less than four-thirds, it undergoes catastrophic collapse, it has to have a stiffer equation of state, gamma greater than 4, that critical gamma actually decreases as a star moves in a tidal field of a companion. Meaning that one has to have an extremely soft equational state in order to collapse. and in particular and more precisely any star which is stable in isolation that has a gamma greater than four-thirds will always be stable in a tidal field of a companion to collapse. And in fact, there are stars which would be unstable in isolation, but when

2:30 put in the tidal field of a companion, they would actually be stabilized And as Jim Nelson, Matthews, and the company reported just the opposite finding from their numerical simulations, a couple years later, that's what stunned us. and certainly our interest generating our own post-Newtonian and fully relativistic calculations. Come on in. So that's where our second wave of interest got triggered. That was an ending contradiction. As I said, I think that APJ supplement be the earliest discussion in the literature about the ability of the tidal field of a binary companion to stabilize a star of inspiration in the collapse. You may find earlier that if you find earlier discussions, that would be interesting. At least the first I was aware from the actual data of software. Well, so, that led to responses to what some methods were by Don Lye and others that come back, and then you and others here did work with the co-rotating case for relativistic I talked to Thomas Stomberg yesterday. He was saying that initially you had begun working on that code before Wilson's attack. Yes. Yes, we had actually started to develop that equilibrium calculation before we knew Wilson's resolve. and yet it was a rather non-trivial problem to solve the equilibrium equations in full GR in 3D so it took quite a while for that to get fully online part of the problem was part of

5:00 all problems in 3D is to take efficient use existing computer-generated hardware to do parallel computing, one needs very special software really to manage an integration in parallel efficiently in 3D. Some of that software was provided through our Grand Challenge Alliance, where professional computer scientists helped construct a package which essentially manages and distributes the calculation among many processes we needed to do that in order to get sufficient large grid, sufficient coverage of our computational domain numerically to have accurate results. And that's of course one of the concerns we have with Jim's results. Did he have adequate resources, adequate coverage, a sufficient number of zonies to do that calculation. If we don't know, we won't know for sure until the last week do we take this calculation, which is something we're considering doing. That's part of the general approach? Yes, well, as I mentioned yesterday, we are launching a more intense investigation investigation into the binary neutron star problem. We want to do it in full general activity, hyperdynamically. We've done it in Newtonian physics. We've done the equilibrium problem in MGR. We've done the ellipsoidal post-Newtonian approximation. Now we want in the full problem, time dependent, light generally done, and we hope to avoid some of the pitfalls and hopefully confirm the result that we've obtained already with our equilibrium stabilization that E-Stars are stable to collapse. And of course that's only one question, there are many other questions that we're wants to ask, as soon as that one is resolved, what does the n-spiral look like over the

7:30 gravitational waveforms? Can we see signature of the transition between slow n-spiral to rapid plunge? Does that transition tell us something about the size, the radius of the neutron star and if we know the mass of the neutron star by looking at the waveforms far away and then determine something about the size of the neutron star by looking at the rapid plunging coalescence waveform to the first something about the equation of state of nuclear matter these are hosts of questions other people have raised them But we're now ready to begin looking at this computation. We have some tricks up our sleeve to approach this problem a little slowly and carefully perhaps that has been done before, where we can gain confidence in our calculations at our stages. Based on the experience that you've done with the creatures? Based on, yes, I mean, we've done all these other calculations. We know what to look for. We've identified concerned quantities. We also have ways of dealing with this problem in a hierarchical way. there are simplifications one can make which will allow for example focus on the gravitational field equations simplifying the hydrodynamics while we look at what happens to the gravitational wave forms and the gravitational field equations. And when they are under control, we can then replace the hydrodynamical simplifications by the full set of hydrodynamical equations. And we've already taken advantage of these simplifications in some of our equilibrium work. For example, there are two timescales in this problem of interest.

10:00 of a star about its companion, and the other is the in-spiral time. In-spiral is triggered by the radiation of gravitational waves and is slow compared to the orbital time. So there's a small dimensionless ratio, and that's T orbit over T in-spiral. one can imagine a procedure in which one more or less expands in that small parameter and that's a simplification that is justified to the extent that that parameter is small and we think we know how to take advantage of that small parameter in the first calculation that parameter ceases to become small when the stars really get on top of the innermost state of circular orbit and so we will ultimately have to resort to a full-fled approximation-free calculation, but we can work our way to that part, gaining confidence in various sectors of our numerical calculation curves. We thought Would you be taking a step on the way similar to Wilson's technique of defending the form of flatness and perhaps taking the range out of the from the binary to China. So, be appreciative of what I would be getting. Um, in our first go at it, probably not. We'll probably use conformal flatness. Um, because I think what we first may want to do is repeat Rosen's calculation decisively. we assumed conformal flatness in our equilibrium calculation where we got a different result than Wilson. So it would be interesting to see if yet another, perhaps an improved Wilson-like calculation gave a different answer from the one he gave. I'm not convinced that conformal flatness

12:30 has anything to do with his result. the basic assumption that he may be there, but with other numerical reasons and numerical assumptions. I suspect that was true. But I'm willing to test that in stages. First I'll make that assumption and see what happens. But I remind you, in our equilibrium case, We made that assumption and we got a different answer. So I don't think that's the root cause at the time. Although talking to Grant Matthews the other day, he was saying, I don't know if I'm practicing the details of this work, but he said that they went back and put the co-rotating case in that code and that particular case found that they didn't, But they got results, they should be able to do it, and that he thought there was something significant in regard to the non-COVID data case. The non-COVID data case? Well, there is that possibility. The only calculation that I'm familiar with that deals with the non-COVID data case is the post-Metorian calculation that I did with Lombardi and Razio, which is a paper that's in press, where we were able to deal with co- and e-rotation equally. One did not have to assume co-rotation. And one found results that were basically in agreement. And Kipp, in his recent paper, which is analytic, reports a similar result, that tidal fields stabilize irrespective of the co-rotation, the degree of co-rotation. So, I don't know that that is a crucial feature, but it may be. It certainly may be. All of these point to the need to do new numerical calculations. assumptions, perhaps several different groups, because this is a tricky problem.

15:00 And each group to make progress has had to assume one more non-negligible assumptions. And we have to begin relaxing them, but in order to do that, we need more work online. There does seem to be a considerable degree of interest in this set of different models Well, the interest is triggered in two directions. One is it's a problem that is considered by the gravity wave experimentalists to be one of the most promising sources for wave detection. and these detectors are being built and we need to understand theoretically. And secondly there's a pure theoretical interest in this problem. As I mentioned yesterday, binary problems, a two-body problem is the outstanding unsolved problem in classical general relativity. And that is a challenge right there. What is the solution? Perhaps there's another motivation that we know there are binary neutron stars. From the earliest days of Hulse and Taylor, we discovered the first known binary neutron star. To the present time, where there are at least four of them in our own galaxy, they do exist. And we know for sure they do in spiral at the rate predicted by general relativity for gravitational radiation dissipation. All we do not know is what happens when they start to get close. So this is a real event that happens in our real universe. We do not know the answer, and that is a challenge. Just to return to one other thing that we discussed briefly yesterday, the reaction of some of the experimentalists to those of the methods, So what you were saying was, I was one of concern that here was something very unexpected that I had counted. Right. I recounted the fact that Lee Holloway, who we met, Professor of Physics in Illinois, who was on the Virgo collaboration, came running to my office, arriving home from an Aspen workshop,

17:30 not having heard Jim Wilson's story, alarmed that much of his effort would be going down the tubes because he has been grooming his effort with Virgo to the search for binal neutron stars. And here he was told, well, we really will not have two neutron stars at least at the very last stage, they won't afford to collapse the black holes. Now, quite frankly, even if that is the case, there is a, it is no great loss from the point of being an experimentalist. Experimentalists will have two binary, will have a binary black hole system, and that will be very interesting. Maybe even more interesting. The only loss will be, I think, theorists who would like colliding neutron stars to provide gamma ray bursts, which are observed. And if the two neutron stars collapsed in black holes prior to merger, they will not generate gamma rays. So they're not gamma ray bursts. But for the Grenadation Award astronomers, I don't think there's much to lose if two neutron stars happen to collapse in black holes. Although some information on equations of state and radii will be lost. But I don't think that happens. And I explained to the only why I felt that long, I think it was pretty short. The, um, I suppose from the, um, the experience point of view, um, it really, it was difficult not to take because of the Japanese work at the base family. Well, that is, it's, unless you redo the calculation or analyze pieces of it or have worked on this tricky problem, you have no recourse. In some sense, the numerical work is so complicated these days

20:00 that it's like experimental work. It's like an experimentalist reporting on their findings. Unless you want to acquire the apparatus, build a new experiment, you have to take the results at face value. Now, no one has really assembled that machine, exactly as Jim Wilson has done, to completely rule out what he was talking about, but we have analyzed it in pieces and have done very in a significant way to rule it out. that is the difficulty of modern, large-scale computational physics in general and relativity in particular. It can only be done with a huge investment and a large resources, computational, personnel, time, and when a result is reported, it is not easily ruled out or confirmed. It takes another effort in order to do that. The reputation becomes highly not true. Yes, exactly. That's where I pointed out yesterday the analogy of the supernova problem, that the supernova problem has been like this for 20 years, 30 years. Trying to explain how one can get a supernova explosion on the one hand and a neutron star remnant on the other has been a puzzle. and when one group reports a result, positive or negative, it's very difficult to confirm that with another calculation because the calculations themselves are not trivial. And those calculations, for the most part, have been one-dimensional, two-dimensional, only now three-dimensional. And typically, they're Newtonian. We're talking now about a three-dimensional, fully relativistic calculation. That's the complexity and suggests, following a supernormal analogy, that we're going to need many different code builders over the years building different codes and

22:30 approaching this problem at different angles to be confident with the result. It's interesting, actually, that I've already spoken to a two people working in the neutron star, why did your problem begin working in the neutron star? Well, they have a lot in common. Compact objects, three-dimensional, hydrodynamics, GR is important. Most people who work on the neutron star problem have an easier time than these would be the microphysics. That is, the stars are cold, typically not hot to begin with. Hot supernova have radiative transport to be concerned about. The radiation is the retrinos. That's coupled into the hydrodynamics. That's a very complicated problem. And they have very complicated equation of state variation as the horizontal line collapse. And the equation of state is a hot nuclear equation of state, which is more complicated than a cold equation of state. People who have looked at the neutron star problem today have dealt with the simplest cold nuclear equation of state, and when such a cold neutron star is free from neutrinos. And yet, there's controversy. When these neutron stars smash into each other ultimately and their kinetic energy is converted in part to thermal energy so they become hot and then they begin to radiate neutrinos. All the complexity of the supernova problem is revisited in the binary neutron star. No one has gotten to that stage yet, except in our Newtonian calculations where we have seen this start to get in place. it is like the supernova time, in many ways, only maybe harder. I wonder if that, when I was looking at the, previously looking at the radiation action problem in GR, the analytics problem going on in a few decades, one thing that people often come to me was the difference in sort of cultural outlook between people from different backgrounds. Well, so-and-so takes a mathematical view and is always satisfied by a certain high level of rigor, whereas I'm going into this system.

25:00 I was wondering if, within the numerical field, if you have any kind of similar phenomenon? Jim Wilson, for example, is a computational genius in the sense that he is able to translate his intuitive feelings about computational requirements to code very quickly. And he is able to get code up and running dealing with the problem much more quickly than most people. part because he is somewhat more relaxed about all of the modern slogans and requirements which insist on second-order convergence of codes, which insist on a testing of code accuracy by doubling, redoubling, quadrupling grids and watching how the answers converge. It's not at all that Jim is not interested in accuracy and reliability, but he is not overburdened by that either. and he can often be out in the front laying out the essentials of a problem early and quick by not being so overburdened by these other concerns. And there are other people who will not run a code, certainly not communicate the results of a code unless the code passes Numerical analysis tests. So I think there is a whole range of variation in the numerical field as there is in the

27:30 I think both approaches are useful. If we wait all the time to cross our T's and dot all of our I's, we will never barrel through and learn what's going on. If on the other hand, we don't eventually cross our T's and dot our I's, we might be Do, um, is that more of a question, do you think that this sort of style is more, more, probably more a question of individual temperament or do you see any evidence that it reflects a sort of background and presence in people coming from an experience of supernova and No, I think it's more temperament, but even that is molded somewhat by background. We have some people in our field of numerical activity who are trained partly in computer science. And we have others who are trained in astrophysics, and still others in hydrodynamics, and still others who are trained doing very practical problems for, if you like, industrial application in large scale. Each of these people bring a certain style and assumed sheet of criteria or validity with them. And so, backgrounds certainly play a role. It has two personalities. There's one issue I thought might be interesting to address, because if you're going to discuss gravity-wave astronomy, and you go outside the gravity-wave community, you're likely to get a certain point, another point of view. In particular, you're going to have to confront, or you will confront, the notion shared by many people outside, relatively, some astrophysicists, any astrophysicists, people in other areas of science, physics,

30:00 whose work competes with gravitational wave astronomy for funding and that attitude is one of some hostility on the table. Hostility is a product of the first competition for funds but also it has some scientific basis in the sense that well, those forms you're looking for are exceedingly small in amplitude. The sensitivities you require are exceedingly high, and the likelihood that you'll ever reach that is so small, why spend all the money now? And then on top of that, you have the argument that, well, you don't even know exactly what you're looking for. You have no idea what the event rate will be. we'll be building a very expensive apparatus that we're unable to see for any reasonable rate an event. And why are we doing this? That is a prevalent attitude outside of Corpus River. Not in all sectors, but I've heard it from now. And it's even led to difficulty in getting things like LIGO started and then maintaining these online programs. So one question everyone addresses in the field is, well, what's the answer? What's the response to that? You know, I've thought about it. And I do agree on the one hand that the probability of a detection in the very near future is not extraordinarily high. On the other hand, the bonus, the knowledge gain, the significance of the detection is very high if we make one. So the product of the low probability times the significance I consider to be, you know, fairly reasonable for a go. And, you know, that on the one hand. My more thoughtful reply is to again look for analogy.

32:30 For me, the analogy is the solar neutrino experiment. Here was an experiment that was designed in the 60s to look at, to literally count the few neutrinos that were being generated in the core of the sun in order to confirm the whole nuclear fusion scenario that was constructed for stellar evolutionary theory. Ray Davis, Davis Experiment, Corman Experiment in South Dakota would launch that whole field. And I remember listening to talks by Davis when I was a grad student in the 70s, even before that in the 60s. I participated in some elementary theoretical calculations with people even who did some the neutrino expected rates, McCall, John McCall, of course, has participated theoretically in his problem for many, many years, had done calculations, and if I looked at that experiment over time, I find that that experiment really did not gain credibility for many years, if if not a couple of decades, in the physics community at large. The results were reported by Davis of the neutrino deficiency. And these results had very little impact among physicists who believed that the astrophysics was so crude for those theoretical calculations that who's surprised that the neutrino count would be so long. Surely the calculations are off. And the difficulty in capturing those few neutrinos in this huge tank of tetrachlorine at the base of a gold mine was so unlikely that, you know, who can take it seriously? You're looking for too few events amidst theoretical calculations that are, uh, too complicated to believe have been done correctly.

35:00 But finally, after years and years of perfecting the experiment and perfecting the calculations, perhaps, uh, end of the 1970s, beginning of the 1980s, when neutrinos were thought to possibly have finite mass a whole new group of physicists became interested in this problem the experiment began to be taken fairly seriously and in the late 1980s and now in the 1990s whole new neutrino experiments have been designed and it is now considered one of the most important ongoing experiments in particle astrophysics if you like It is an enterprise that is contributing to not only astrophysics, its original intent, but the fundamental theory of particle physics. It's addressing certain properties of weak interaction and properties of neutrinos, neutrino oscillations, neutrino finite masses, if they have it, whole sets of new questions. But it took 30-odd years or longer for this experiment to rise from an epoch of total skepticism to one of fundamental significance. I believe that LIGO and the gravitational wave program might follow a similar pattern. That it is time to launch this, and it has been launched. and there will be a period of great skepticism but it has to be launched sometimes, at some point and there will have to be a 10 to 20 to 30 year period over which this experiment will have to run be improved trigger theoretical calculations and improved calculations before it may take grasp but in the end And I say, go for it, for that reason, not because I am counting on an early triumph. I would love to have that, but even if there is not, in the end, I think it will be extremely important. So that's my analogy, and that's my answer to whether we go forward or not.

37:30 Yeah, obviously this is an important area. Regarding that, you were saying that one of the arguments perhaps against LIGO is that the theory is too unsophisticated at this point. The theory is unsophisticated and the experiment is so delicate that unless we're lucky to get something going off in our backyard, it's a little iffy. and so those who say well let's do LIGO because it opens up the new window to astronomy there was radio astronomy there was x-ray astronomy now there's gravity wave astronomy we're on slightly thinner ice when they make that argument because the other detections and the other sources were on a solid footing But I think the analogy is not with the onset of radio spectrum and x-ray spectrum, it's with the solar neutrino spectrum. That's the analogy. I'm very a physicist with doubt the significance of that experiment, and yet few are still more enthusiastic or supportive when it was first conceived. I know the skepticism in the 70's. I was there and listening, and I was not taking it very seriously. Do you think the, in the ultimately successful battle process, to get Lago Fungi to get the ball, Do you think theory played a positive role in the system? Was theory something that proponents of the experiment could want to be to say, well, we have these arguments? Oh, yes, sure. That's what propelled that experiment. The numbers were discrepancy. The observations disagreed with theory. I don't think there were any really good explanations for the discrepancy until particle physics came through with other options later. I don't think the issue is resolved, but I think that new theoretical ideas had to come

40:00 on board to make that experiment, but to drive that experiment more. In the case of Lido, for instance, was theory in a position to play a positive role in the arguments for beginning the Lido? Oh, yes. Sure. Sure. I mean, it's theory that tells us about black holes and it's theory that tells us that it's strong gravitational fields are probed by black hole collisions, and that we are not probing strong gravitational fields directly by any other means. And we certainly are not verifying Einstein's theory in the strong field limit by any other way. We have all these linearized predictions with contoning verifications experimentally, but we have a really low horizon, and this will do better. That's what theory is, that's why theory is propelling us all. And even though, I suppose, the argument could be made that theoretical predictions, for instance, and source frequencies such as they were were on the one hand not certain and on the other hand not hugely optimistic even if that theory was still able to provide sufficient justification. Oh sure. In other words, canonical theory tells us that there will be sources and event rates will be adequate and we will be able to detect them at what levels Vigo hosts achieve. not hoping for a miraculous overturning of conventional wisdom. There are people who can't argue that if they don't see anything conventional wisdom when we overturn, we should be seeing things. But, so, we probably have to take off. I've got to move on to a couple Thanks.

42:30 Thank you.