Interview with Cliff Will

0:00 Okay, I think it's actually working. So I'll just say quickly that it's the 2nd of March at 11.30 in the morning, and I'm speaking with Cliff Whip. So I guess I outlined sort of what I was interested in, which I also ran out of, so I don't know what I'm thinking about in the time. um so i suppose well well so i suppose i'll start off with something that i mentioned harry collins was interested in this and he was struck by the fact that some of the ligo people were saying well we should do some theory testing and we want to know what kind of theory testing we can do and you know maybe we can invalidate this theory or that theory and he got the impression that they were saying well you know that theory isn't a very good theory but if we could invalidate we would at least be saying something. So he got interested in the question of, well, what is a bad theory and what is a good theory, and why is it that you can have a bad theory that it's worthwhile to go in and validate with an experiment and so on. So I was going to ask you if you have any ideas about bad theories and good theories. Are there such things as bad theories, and what would that constitute? It's a bit of a continuum. There are degrees of badness. I mean, there are theories that are just so bad that they are either poorly motivated or they don't even agree with sort of basic Newtonian physics, even though frequently the authors of these theories don't recognize that fact. You have to battle it and so on. But there are theories, and Rosen's theory is one example, that's in a certain sense not a bad theory in that it's well-posed, it's based on an action principle. It's not very pretty, but it has the capability of making full-scale calculations of anything you want and also agrees with not only Newtonian gravity but post-Newtonian gravity in the solar system. And so that's sort of an example of a reasonably good theory. I wouldn't say it's a bad theory. Whereas the Yilmaz theory crazy. It's not even worth mentioning, although Yilmaz and his collaborators continue to write lots of papers in journals like New Oval, which meant though, which is the only place they can get them published. but I mean, in your email, I guess you mentioned

2:30 this issue of Rosen's theory versus scalar tensor gravity. I mean, why people discuss the fact that the binary plus arc of Rosen's theory is not, you know, Franz Zicke theory was never sort of discussed. But that actually comes down to quite a practical reason, namely, the binary pulsar does not put a decent amount of scalar tensor gravity. It's not as good as the solar system, just because of the bad luck that the two neutron stars are virtually identical. And so the main effect of Franz Zicke theory type of radiation is just suppressed. So in fact, a student and I went through the exercise and found that the best limit you could put on omega The coupling constant was about 100, and now solar system tests. Even then we're at 500, and now they're up to 3,000 just from the solar system. Whereas in Rosen's theory, it turns out that even though it's pretty close to GR in the solar system, as soon as you have neutron stars with strong self-gravity, those nonlinear aspects just totally make the theory completely different from general relativity. and so you get a qualitatively different prediction that then is easy to kill with binary pulsar but you couldn't kill it with the solar system because you could always they could agree with any experiment so that's sort of why there was some focus on this particular example of the theory it was a reasonable theory but you could really show that it was sort of representative of a class of theories that in a weak field limit or virtually identical to GR but in the Strongfield regime, are quite different. It's qualitatively different. So from two different points of view, on the one hand, these were theories that you hadn't been able to test before. And on the other hand, they were the particular type of theory that you could discriminate with this particular test. Right. Well, just to go back, you mentioned poorly motivated theories. what would be an example of a poor motivation for a theory do you think? Well, the motivation I mean, the individual motivations of various inventors I'm not sure that's that relevant I mean, some people in the past have tried to generate their own theories because they don't like black holes or they don't like curved space-time so they want something that's more like a field theory type theory

5:00 I mean the sort of basic thing that a theory really needs to have is some sort of a complete theory has to be complete in a certain sense in that it has to have enough of a structure that it can evolve gravitational it can do not only gravity but also determine how matter responds to gravity in kind of a self-consist in a way. You can't just postulate that particles move on geodesics. You have to show that an atom, which is charged particles and electromagnetic fields, you have to be able to calculate the behavior of all of that stuff in your gravitational field. These days, good motivation also means that a theory should be based on an action principle, just because essentially all successful theories in physics are those based on an action, and there are no examples of successful theories that aren't. And so in some ways that's a strong starting point, although the direct evidence that you have to have an action principle is experimental evidence is pretty weak and indirect, but it just seems to be one way you're sort of guaranteed of getting to the next step of being able to make real predictions and having things be reasonably consistent and so on. And then clearly you have to agree with sort of basic stuff like Newtonian gravity. I mean, there's a theory that's being promoted by a guy named Philip Mannheim at the University of Connecticut that sort of is an action-based theory and has all the bells and whistles, but has very strange behavior at the Newtonian limit. And I've tried to argue with them in various reports and letters and such that the theory just doesn't agree with basic solar system dynamics at the post-returnian level. They have some ways they try to argue out of this, but there's just something that seems to be drastically wrong with this kind of theory. That's something I was going to ask about since I guess I was wondering if that happens a lot and if it's a very turbulent lengthy process, if you're analyzing say, for the purpose of theory testing, and you disagree with the progenitors of the theory over the results that it gets?

7:30 Well, I mean, it can be, and it rarely reaches a conclusion. The only time I know, and I don't get involved in this all that much. I mean, you know, I don't sort of grab theories of the literature and analyze them. It's kind of a hopeless and not terribly rewarding task. But my experience has been that, you know, it takes a long time because people who propose it always try to wiggle out of it. But there are only two cases that I know of where it has actually come to a conclusion whereby the person said, yes, I agree this theory is wrong. And one was Rosen himself, because when we did this work on the binary pulsar and showed that Rosen's theory disagreed with the observations. In fact, I was giving a talk in Haifa shortly after that and he had this lecture and said to Rose's theory it's wrong and at the end of the lecture, Rose said, yep, I agree with you, it's wrong but I have a new theory. It is a rather different theory which he then went on to argue. It's on nice properties and agrees with all the experiments. The other example is this notorious theory by John Moffat of Toronto of this non-symmetric gravitation theory. After years and years and years of discussion on various points and disagreement over what things it did predict or didn't predict and what bounds were or were not placed by various observations, finally Thibaut D'Amour and collaborators showed that it had a real sort of basic flaw at a basic field theory level. In some sense, it predicted ghost modes and gravitational waves unquantizable, you know, kind of a fatal flaw if you ever wanted to think of this as a theory, at least equivalent to GR. And in the end, Moffat conceded on that. It took several years of, you know, arguing back and forth and trying to wiggle out of this problem, trying to find a way out of it. He finally conceded with a bottle of wine, and I was there when the wine was open and shared, so that was, that doesn't happen. It happens occasionally. I was going to ask that, in fact, the prediction that, the kind of anti-damping prediction for binders and gravitating those from binders that Rosen's theory, that was something that you calculated as being a result of his theory,

10:00 and then subsequently he's agreed that that was. Yeah. Now, see, in a case like that, it really depends on your point of view. Some people would have argued that just having anti-damping, that a bad theory right off the bat and you would just throw it away without further ado so you know my attitude is slightly more phenomenological than that I'm willing to say that it looks strange to me but it's compared with observations and of course there the comparison is easy because we see damping and not anti-damping and so it really is wiped out but some people would just say on theoretical grounds it's that theory's dead yeah yeah I guess so I think you mentioned that in the say well but we should just leave it there did Rosen I don't know where I got this impression I had the impression that Rosen might have actually tried to come up with some elaborate justification for the result I know that a student of his there were a few papers for a time trying to argue trying to combine this idea trying somehow to use the absorber model, Wheeler and Feynman to circumvent this problem so his argument was that because of the absorber idea you really need to use half retarded plus half advanced green swung so you don't get any damping at all so you've got around the empty damping if you put in this idea but of course the observations So he's wrong even then, because there's no way in the theory to get actual damping of the orbit. So I was familiar with that paper, so I wonder if that was related, I knew it was the same time period. And curiously enough... But it was before the actual observations for the damping we were in, so there was a period between when I wrote this paper that showed this theory and others like it anti-damping predictions and then he had these arguments and then of course the observations came in to clearly show damping at that point so there was kind of a period in which three or four years so when you initially made the calculations about rosen's theory and perhaps others was that

12:30 in anticipation of the money closer yeah shortly after the discovery it was this so it was discovered in september 1974 and sort of for the next two years first i think it was doug early who pointed out that we sort of did the calculation in scalar tensor theory and pointed out there'd be dipole gravitational radiation in scalar tensor and brand sticky theory and so i then looked at sort of a whole bunch of theories about six or seven of the various alternative theories that were around and went through the details to find out what they would predict and including also extended his calculation to sort of the brand sticky modifications of quadruple radiation as and found that, in fact, in all cases except Brans-Dickey theory, the damping was anti-damping, which is, I think there are probably some good reasons for that, just because all these theories tend to introduce things like vector and tensor gravitational fields in addition to the metric, and these fields tend not to have basically positive definite energy density when you do some formal calculation of the stress energy, in these theories, they're just not positive or definite in general, unlike GR or, say, the scalar tensor case. Just as a quick aside, am I right in thinking that people have argued that if the gravitation theory doesn't have positive or definite energy, that it should be incompatible with quantum mechanics? So that would be an argument for simply not even thinking about these theories further, but that's a theorist's argument. Since we can't even quantize GR yet, to use that as a criterion to rule out other theories seems to be a value of theoretical arrogance even though it could be right but that's just I think more phenomenologically than many others I remember some people telling me that when the binary was first discovered that maybe Demure said he wrote a paper with Ruffin saying that well it probably won't be able to see effects from gravitational radiation so were you sort of just inclined to take a more optimistic view at that point or did the experimentalists were there more optimistic or just in case it wasn't that people said it wouldn't be seen but people predicted that it would be anywhere between 10 and 15 years before it could be seen before the build up would be

15:00 enough to be able to detect it that's true of course they did do it very quickly and So the first observation was within five years of the original discovery. But basically no one anticipated the fact that the pulsar would be so stable and that using tonal clocks they could really keep phase with the pulsar over a long period of time and really just track its phase the way they do in a tonal clock. And that just gives you, so if you can be within one cycle, and so basically an integer thing you're measuring cycle by cycle over five years, the accuracy for a 59 millisecond pulsar is just so enormous that they were able to do it in five and not 15. So they just beat the predictions. But it was clear what was going to happen, at least to those of us who interacted with people like Joe Taylor and stuff. It was clear what happened someday. And so we all sort of jumped on it and started calculating away. I used to show this slide in the early days when I gave a talk on binary pulsar. In fact, the list is in my book, too. These papers that came out within the first six months of the discard. The theorists were jumping on the bandwagon. So was Rosen's theory, this is probably a difficult question to answer, but since nowadays it means someone like me will only go over Rosen's theory kind of for historical reasons. Was it kind of a prominent theory at the time? Was it one in which you would have had any reason to do any calculations in other than the fact that, well, now there's a possibility of testing the theory? Probably, certainly not to the degree that scalar tensor theory was prominent, if you can use that word. But it was, I mean, it was out there in the sense that obviously Rose and the name attached to it was a person you sort of took seriously, I mean, he was a serious person, famous person. And it was also, and again, just from the point of view of the subgroup of us who were

17:30 thinking about theory testing, it was an interesting example of a theory that was, you know, had a lot of the basic properties of general relativity. It was a metric theory and it had Lagrangian blah, blah, blah. But it was still very different because it had these two metrics, two tensors, a flat background metric that you composed non-dynamically plus a physical metric. So it had sort of interesting properties from that point of view, sort of conceptual properties that made it different. So, I guess, and again, this is something that I'll have to at some point ask an experimentalist, I guess, but that's why it seemed, since that's kind of, in a general sense, the kind of impression in the hand, it sort of struck me, looking back over the paper, say, Weinsberg and Taylor are one of those papers, that, you know, here's a result which I think we nowadays mostly put in terms of, well, here's the first evidence of the existence of gravitational waves. And yet the paper spends a large part of its time disproving Rosen's biometric theory. So I why theory testing is so important that, you know, it's kind of ranked right up there with, say, discoveries of new phenomena. I don't know. I mean, you know, I guess you'd have to ask Joel or Jill what they were thinking at that time. I mean, it may have been a form of modesty in some sense. I mean, because in some sense they had performed, you know, the first real detection of gravitational waves, however indirect, but it was really a real thing. But they were a little too modest to actually come out and say, we've done this, but they could certainly say, you know, and it was still an impressive statement, that we have ruled out this theory and maybe many others and verified general relativity in this sort of more in this way where you know you sort of have well-defined rules of what to do and we've done this but then to come up with this grand statement we have detected may have been a little more than they wanted to do at that precise point also I mean you know the original detection of the damping was It was good to only about 10 percent, and it was still pretty, and it would even, also

20:00 I think, you know, there was a while before it even settled down to agree with, precisely with the GR value. It was like, you know, for a time it fluctuated, sometimes it was a one and a half signal away from what GR would predict and so on, so it may also just, on a technical ground, it's They might have wanted to be a bit more cautious before they said we have verified GR in detected gravitational waves just because it wasn't quite there yet for a while. There was this whole business of a correction. You have to correct the data due to this galactic differential rotation effect, which is about a 1% effect. and that became important as the error bars went down below 1%. Again, the observation then started to move away from GR. And so we had to carefully take this effect into account and then we brought it smack on. But anyway, they can probably characterize their thinking at that time better than I can, but that's my sense. Well, but that seems interesting. And I suppose in that sense, is the fact that the theoretical predictions between the two, between, say, Dr. and Rosen's theory, predicted that qualitative differences, to say, between anti-damping and anti-damping and attractive. If you may be saying about Rosen's theory, that was totally valid and disputable. So you didn't have to worry about your error worse too much. So I guess I was interested in asking, in general terms, since I've just somehow gotten myself onto the subject, kind of the origins of the apparatus of theory testing from the theoretical point of view, and the origins of the parametrized post-entennial formalism what the initial motivations were was it pretty much solely connected with the idea of theory testing or was it more of the theoretical concerns? I mean, the origins of course go back to Eddington to stick parameters in the Schwarzschild metric and then to say that if we put these parameters in then we can make the deflection of light in terms of these parameters so that's a way to test theories. Now we had no alternatives to think about in those days although alternatives existed

22:30 like the Nordstrom theory and some of these others but I don't, you know, I don't think It was really until sort of later on that people started actually sort of compiling all the other theories and calculating these values in other theories. So there was kind of a disconnection between the formalism, as Eddington wrote it down and Robertson and others used, and sort of the universe of theories that were floating around in the literature. It was only really until, I guess, in some sense, me, that people started saying, okay, let's take some of these theories and actually calculate their PPM parameters and make tables of the list and see how they compare and ask what experiments to rule out what and what different experiments could you do to rule out this or that and what kinds of theories, what classes of theories would be ruled out. I mean, even Nordvet, who sort of originated the more, added more parameters from the original two of Eddington. Right. Ken only worked out the parameters in Sticky Theory. Never sort of went beyond that. So it was really me and some of the other students in Kip's group who started thinking about kind of systematizing the whole deal. Right. And was that pretty much inspired by the theory testing experiments? Yeah, I mean, it was inspired by the idea that there were, you know, one could think about a bunch of different experiments coming down the line. I mean, in 1968, when I entered Caltech, people at JPL were planning the Mariner 6-7 tracking and worrying about the Shapiro time delay and how good a job they could do and what accuracy and so people were really thinking about solar, starting to really think seriously about solar system experiments. You know, business of Brands Dickey theory and solar of lateness was still controversial in 68, and some people wondered about that. So, yeah, so it was just really finding a way to test theories in a way that didn't put general relativity sort of on a pedestal. Just put it in the hopper along with the others and kind of level the playing field.

25:00 just a bunch of parameters, and you measured the values, and then you asked which theory in a table gave the best fit. So I was curious, was it kind of intended that way, that as sort of a tool for the experimenter, that the experimenter would have to be a table of values that they could apply to give a problem? Well, there was that, and also, I think the experimenters liked the whole approach because that gave them predictions and things kind of in their language that they could appreciate. They could say, when I'm doing this experiment I'm measuring this effect and this parameter. It was a very handy way to connect between theorists and experimentists. And I guess to touch back then on this issue of when the theorists disagree over the prediction of the theory, which I guess, in the case of radiation damping, was really true of GR because there would have been all this wonderful form of controversy and so on. So I suppose that explains why, for instance, Weisberg and Taylor, they said, well, we know I'm going to just leave that to the theorists. In the case of alternative theories, if it occasionally happened that people disagreed over what the prediction of the theory was, is it your impression that the experimentless had a way of their own notions about what results to believe or who to believe about these things? I'm just sort of wondering if, you know, in Weisberg and Taylor, I guess they're reluctant, they seem perfectly happy to make a prediction between theories, and they definitely don't want to get too much involved in deciding which particular version of a prediction from general relativity is correct. So I was curious if it was something that was a problem in the less prominent theories that they had to say, well, okay, we think we could rule out that one except that someone so says that maybe the result is different from what another person says.

27:30 Was that ever a reason why you were involved in the debate with another theorist? Well, I mean, I, by and large, I'm not sure I sort of fall in with... Yeah, the notion's gotten a bit wooly. I guess what I mean is, you've said occasionally, not very often, you've been involved in debating with a theorist. And does that arise more or less from the need to do an experiment, or is it more from theoretical principles? Yeah, but it's more from theoretical ideas, and it's more between the theorists. either wait till the dust settles or they tend partly because they don't feel quick to make these judgments and they quite often these things just turn on fairly technical points this thing with Moffat's theory for example I mean you could make lots of predictions with the theory and then you could make pictures of the binary pulsar this that and the other and put limits on the various parameters in that theory whether they were enough to make him throw in the towel or not. I mean, that was one side of the issue. Experimentalists were quite happy to take a prediction that was made and to say, well, okay, we can put a bound on Moffitt's theory of such and such with this such and such experiment. But there wasn't a lot of disagreement, I don't think, about those kinds of issues. A lot. Something, but not a lot. But it's this thing that finally killed the theory. I mean, that was really a very technical discussion. I was not involved in it directly. I was involved in refereeing the papers back and forth. for me, Thiebaud and Stan Dezer and another collaborator would write paper and Loppa would respond and there'd be back and forth, back and forth, and five different responses, and then Fisrael would just dump it on my lap and say, here, you decide. So that was, so I had to get into it in some detail to follow it all. It was a pretty technical question, but I don't think the experimenters really care one way or the other about it. I mean, And they're happy to see it be dead. But a lot of, I'd be asked a lot by experimenters, what about, what's the status of Moffitt's theory?

30:00 Because he gave a lot of co-occurring lectures, so people certainly heard about it, and it was in the physical review. He published a lot of papers, a lot of students working on it. So they certainly were interested in the theory to the degree that if it made some prediction that they could maybe test, they'd they'd like to hear about that. But somehow they only wanted to know when it was agreed, what the prediction was. If it was a solid prediction, then they'd be interested in testing it. I guess that's some question to hang. You mentioned, of course, then, that there's a certain point where if a theory still has its proponents but it's been widely discredited and they're limited in where they can publish and so on. Is that something that's liable to happen to a theory only after it's been invalidated somewhere by experiment or can it just happen because the theorists don't credit it anymore? It's probably not quite so clean cut. I mean, I don't really have a clean answer to it. I mean, partly because so many, there are a lot of theories out there that just really We haven't gone through this sort of cycle of being compared with experiments systematically. Because there's this basic flaw in the theory that still hasn't been resolved. And so somebody like me is not going to go out and go through the effort of making a bunch of calculations because I think there's something basically wrong with the theory at a more fundamental level. It's like this Mannheim theory, this formal, wild theory, or whatever it's called. I mean, it just doesn't stop Tim from making lots of calculations and predictions, although mostly in the cosmological realm, because he uses this to change, basically change the law of gravity. Large distances is a way to get around having dark matter. But, in my

32:30 opinion, he still hasn't solved the basic problem of post-Newtonian gravity, where I think there's a big unresolved issue. So, but he publishes in APJ, in usual places, and such. Occasionally I get the paper to write free, but it's a funny phenomenon in some of these things, you know. I objected to one of the first papers that was submitted, but ultimately it got published anyway, or they, you know, they managed to get it in. And now, of course, if they sent me a paper 12 in the series, I mean, it's hard to know back in paper one, I still was not satisfied with the answer, but in paper twelve they're now doing it on to some other topic and maybe the calculation is just right. But it's always a bad, puts you in a bad position. But I don't get this very often because I think a lot of these people don't, they tell the editors not to send it to me because I'll just be too mean or something. Well, I was curious about that side of it, because even in my experience as a referee, I was asked to referee a paper of Rosen's for Aft-J, and really the only thing that I think they asked Kip, and Kip gave it to me all the time, the only thing they really wanted to know was, is this the product worked out in the biometric theory or in GR? I presume the implication was that it was in the biometric theory, they didn't want to know about it, but otherwise it was okay with that. so it was interesting to see I guess that was a case where the journal knew what they wanted what theory was involved there's some point at which theories become unacceptable for major journals but it's sort of murky as to where they it's murky Oh, so I was going to ask, I don't know if I've gotten a good impression of this from your book, and I guess there may be another one since then, but do you have an estimate of how many theories you've, let's say, gone as far as to work out on the post-natalian parameters? It's actually not a huge number, and strangely enough, many of the theories were, not many, but some of them were just, you know, straw-to-man theories generated by people in Kip's group during this period. I mean, there's the, there's the Leitman-Lee-Nee theory, and I mean, once we sort of knew the

35:00 rules of the game and what it took to generate a theory that, you know, you could guarantee to have, at the lowest level, all the right properties, then it didn't become such a hard thing to do. I mean, you could generate some nice theories. And so various people sort of in the group at that time produced some model theories of this sort. But the real difficulty is if you go into the literature and look at some of these theories, some of them are just so so incomplete and they just don't have enough you know they they write down some equations and they get a metric that looks like the source of metric and they say okay this is a viable theory but then you want to ask you know how do you get what's the post newtonian limit of this theory uh and how do you calculate the motion of something like the moon where you have to really take into account the fact that it's a finite fluid body and works with its own grab self-gravity if you want to get the northern effect right it's internal gravity all these theories having a hope of being able to calculate this thing because they don't have enough equations right they just there's not enough machinery to do the job so there's just a lot of theories like that these days in fact the story is really is quite different actually than it was say 20 years ago these days of course the most interesting alternative theories are theories that are coming from the other direction, down from string theories because there, there's a real sense that the generic theory that will come out of superstrings is not going to be GR it's going to be some scalar tensor theory and maybe not even a metric theory there'll be some violations of the equivalence principle albeit rather weak coming from these extra interactions that come from dilettantes and all these crazy fields that come out of strings But the trouble is there, there just isn't, there's no literature yet, I mean people, string theories just don't yet know, except in some very ad hoc models. There are no firm predictions yet of what string theory really predicts. But we'd really like to know what these things predict, because there are some ideas from improved experiments that could, it's like a space equivalence principle test, they could really make a huge leap from current bounds and, you know, maybe actually say something

37:30 depending on the details but string theory is not yet in a shape although people I talk to who work in string say that they may in the next 10 years be able to really do some of these things so would that involve probably as it were taking a limiting case of string theory the theory of gravity which a particular version of string theory would predict and then you can take that proposed theory of gravity and make sure that it seems quite clear the theory of gravity that string theory predicts is scalar tensor theory but the trouble is there are two big unknowns one is the mass associated with the scalar field could be very heavy long scale maybe who knows and so then the theory would truly be equivalent to GR because they wouldn't affect anything. No one scales. On the other hand, no one really knows. The mass could be very small or zero. It could be light enough that you could have long-range scalar effects. It actually turns out that generic scalar tensor theory is a bit of an embarrassment that no one talks about it. It's kind of a scandal. It's like the blue dress. I mean, it's really a scandal for the theory that the scalar tensor theory that pure string theory predicts has an omega of minus one, which is totally unviable. So they go through these arm-waving things and say, well, it requires a mass, or there's some symmetry breaking, or it's this, that, and the other. But they're confident it'll work out in the end, but no one really knows. but to me that in some ways is the much more interesting direction for theory testing to come trying to get some some kind of predictions or some range of predictions out of these quantum unification type theories that could be tested I was wondering so you mentioned that on the one hand there are ideas for new experiments more accurate experiments and on the other hand then there's this possible new source of alternative theories, as you say, coming from a different direction and are these kind of two arms of the potential enterprise connected in any way? Is it just that the experimentalists are interested in the idea of more detailed experiments just for its own sake

40:00 and then hopefully the theorists will come up with something they can test or are the experimentalists thinking, well, hopefully somewhere along the line there will be a need for this type of experiment so we should think about it. There's some of both. I mean, there's some links. I mean, obviously, many experimentalists would like to do better experiments in some of these areas, if they think they can and have the technology. But on the other hand, they need motivation. I mean, because anything that you do in space, for example, is expensive. Or even on the ground, if you're really talking about pushing, you know, bounds on, say, null experiments, those are not cheap either. So the experimenters would like motivation to do these better tests, so they asked theorists about this. And so people like Thibault-Denor and others sort of look at what's going on on the theory side and see what might go along. So there have been a lot of ideas from the theory end that say that if the sort of low of string theory had such and such properties, then you could get violations of the equivalence principle as large as 10 minus 17, which could then be tested by one of these space experiments, which hopes to go to 10 minus 18. Other people would say, well, it requires a certain amount of fine-tuning, and I don't think that model is what motivated, but that's just the theorists arguing. But it's not inconceivable that you could get some violations that would be tested if string theory had certain properties. Another thing is actually also that's been pointed out by Thibault, Damore, and Ken Nordvitt, that, again, if string theory had such and such properties, then it turns out that the effective Bransdicke omega today could... doesn't have to be infinite. I mean, it could be as small as 10 to the minus 4, 10 to the minus 5. It turns out that in this sort of class of models, which Thiebaud says is a very general class, but other string theorists say it's not all that general, so again, there's some disagreement. But within some class of models, if you look at standard cosmology, gr is kind of an attractor. Omega generically grows in any of these theories, but it doesn't grow all the way to infinity.

42:30 it goes to a value that, depending on the initial conditions, and you put in inflation and put in standard early universe cosmology that we all know and love, you end up with omegas that could be as small as 10 to the minus 5. Well, those are interesting values because we're now approaching a few times 10 to the 3, and an experiment like the gyroscope experiment could reach omega of 10 to the 5, not measuring the frame-dragging, but measuring the geodetic procession due to space curvature. So there are some experiments that could actually start to test predictions that at least some theorists think are reasonable possibilities from some of these string theories. So again, there's still lots of room to wiggle, but those kinds of ideas then motivate the experimentalists to put some effort into some new tests because at least they see some sort It's not just adding another decimal place to some bound. They see a possible goal. And even if they end up with a null result, at least they push the boundaries in a theory space in a way that is sort of non-trivial. It could have been there and it wasn't, rather than saying, well, it couldn't possibly be there, whatever this theory predicts is at the Planck scale. They're actually, even Ed Witten, for example, who certainly knows string theories as well as anyone else, there could be effects in these regimes where some of these experiments are being discussed. So he, as a theorist, says you should definitely be doing these experiments because there might be something. It's not out of the question. So the experiment would probably have particular theorists with an expertise in this area, such as yourself, or Thelon Moore, or Ken Norbert, who they would go to saying, well, if you're thinking we could go farther in the experiment, what kind of things would we say if we didn't go to such a point I have a couple questions of mine for now but first of all I guess I'll quickly say about theory testing with LIGO is that a topic that you're interested in or is that I've been thinking about it a fair amount there are a couple of results I've done, but other people have been thinking about things too. One is that you can test

45:00 scalar tensor gravity with LIGO, and if you're extremely lucky, and maybe not too extremely lucky, you could put a bound on the theory that's better than solar system bounds, and certainly better than binary pulsar bounds. And that's by looking at, again, it's dipole radiation and how it affects the inspiro of two compact bodies. But for it to be feasible, you have to have a mixed system of a neutron star and a black hole. You get two neutron stars, they always tend to be 1.4 solar masses, and so it's suppressed by the symmetry of black holes. Of course, black holes in Bransicki theory are identical to black holes in GR, because basically the scalar field gets radiated away. And so you get no tests at all. Everything is identical in the two theories. But if you have a mixed pair, then you get dipole radiation. And that changes the damping in spiral enough that by mass filtering you can put a bound on the theory. And then recently I also pointed out that if the graviton is massive, and that brings up a theorist versus experimentalist question, if you have a massive graviton then And the speed of propagation of gravitational waves depends on its weight length. And so, because the frequency during an in-spiral is evolving, then the speed of the emitted radiation will differ. And so, the apparent waveform at the source will be different from the waveform you detect at the receiver. It basically will be scrunched in time because of the distortion produced by the propagation speed difference. Of course, the effect will be bigger the further away it is. worked out the kinds of bounds you could place on the Graviton mass using LIGO and using LISA as space-based. You can actually do better, mainly because you can see it further away. And the bounds are better than you can get from solar system bounds on a Yukawa modification to the Uniter's Square Law. But, of course, when I speak about this, people in the audience, especially if they're particle theorists, then I would say, well, but the graviton can't be massive because Veltman and Van Damme proved that you would get four-fifths of perihelion advance as soon as you put it in massive gravity. You'd get sort of this qualitative shift from GR whenever you put it in mass. Well, again, that's sort of a field theorist's viewpoint, and I'm thinking more phenomenologically.

47:30 Let's suppose you could have a theory that was close to general relativity in all other regards of wavelength-dependent propagation speed here's what you can do I don't necessarily abide by no go theorems that come from some narrow viewpoint all right so that would be an example of the test but not with a particular theory alternative theory in mind right basically all I assume that the theory is general relativity plus some correction that involves the propagation but I Although, actually, Matt Fisser upstairs in the particle physics group has actually been thinking a little bit about how you could modify a GR to introduce a massive graph. It is actually a very non-trivial problem. And so we've talked vaguely about seeing if we could think harder about this and try to about at least some candidate alternative. The, I guess, that's a question that Harry would probably ask. What he, the conversation that he was hearing between, I guess, some other people, from what he told me, was, well, what kind of theory testing can we do if we haven't got a detection? and if all we can do is kind of put limits on the flux of gravitational waves. So is that something that you've given any thought to or heard? No, partly because I don't see what, without a detection, I'm not sure what you would do because there's so many astrophysical insurgents. I mean, in the absence of a detection simply means your source estimates are wrong. I mean, you'd have to really go a long way in sensitivity, you know, and still be seeing nothing to start to think about whether that means that the theory of gravitational waves is wrong and maybe that either GR is wrong or something, because the event rate estimates are so uncertain. Even if gamma and reverse are caused by inspiring neutron stars, which gives you kind of a lower bound on the rate of inspirals in the universe, but even that is sufficiently uncertain that you'd have to go a long way in seeing nothing before you'd really start to question what's going on.

50:00 So, to me, non-detection is non-information, apart from maybe giving some astrophysical information. But once you detect, I think, you know, the other thing you can think about doing with an array of interferometers is to think about polarization, measuring the various points and seeing if there's just the quadruple or trying to rule out a scalar mode or whatever, or see a scalar mode. Yeah, because I noticed that right back to the early 70s, I was reading some paper, Kips, and I was talking about theory testing at that time, different polarizations, because a number of different theories that you discussed also have quite different polarization predictions. Right, I mean, all the theories that I worked out, where I worked out the gravitational wave damping predictions, can also work on the polarization modes, and they all predict more than, in fact, the most you get is six with, in the generic case, if you're just measuring the, sort of the electric components of the Riemann tensor, in fact, all, virtually all theories predict all six. So, have you had a lot of interest from gravitation wave experimenters in theory testing, or are they more focused on other scientists? Well, I mean, most of them, of course, are focused on getting the physical, right? But, no, but I think my sense is that they're certainly very pleased to have these things, you know, these possible things to do. I'm not sure they give much more thought to it at this stage, but I suspect that once some detections are made then people will want to put in waveform templates that include these other alternative theory terms and then do some fitting and see what you can do relatively easy thing to do at that time I get the impression but I don't know for sure but maybe you have some idea that the kind of community of way detection is really very distinct from the community doing high precision gravity experiments. A lot of the gravitational wave people have actually come in and have come from other fields like particle. Not so many of them were previously working gravity. Obviously

52:30 there's no statistical notion. I was wondering if the two communities were somewhat different I can think of examples and counterexamples, so it's hard to know statistically. I mean, like the people in Colorado, whom you previously thought of as being interested in experimental tests like Jim Fowler and Peter Bender and all those people are now very actively involved in either LIGO or LISA type things for gravitational wave detection. Whereas people sort of like the JPL crowd and the Harvard crowd who deal with spacecraft tracking and such testing GR are not terribly involved in gravitational waves except for those people doing spacecraft doctor tracking for doing some upper bounds but they're not involved in the lab as far as I can tell they don't have a strong I'm not sure much of a case can be made either way it's just kind of yes and no Well then, since we were discussing in some way the question of how the experimentalist's interest in doing theory testing and the theorist's interest in having theories be tested play off each other, I was curious to know what kind of role that played. my impression is and that seemed to me to be confirmed by you for instance reading your book there was a real renaissance of theory testing I guess the late 60s or the 70s and so I was curious to know that was mostly the result of experimentalists to do new tests, or the theorists deciding to start coming up with alternative theories, or whether the two sort of interact with each other's problem? I think the two really interacted.

55:00 They interacted in many ways, I mean, in the sense that, you know, a lot of experimentalists sort of learn BPN formalism, that someone sort of really interested in looking for new things that they could test, or were interested in what they would learn if they pushed a test to a certain higher degree of precision and sort of asked the theorists about theirs and thought there was that kind of interplay. But also, there were new tests done that had not been thought of before, before, for example, we had this more general framework. I know people in this business are looking for preferred frame effects. by geophysical measurements, a number of groups that actually did experiments that they would not have thought of before. There was a group using gravimeters at UC San Diego. When I pointed out, if you looked at earth tides in a certain way, you could put a bound on one of the PPM rounders. So they went out and had these superconducting gravimeters that were very precise. So they did the experiments and placed a nice upper bound. Better than, I mean, I placed some bounds myself reading the literature and learning what was known, but they could improve the bounds by a factor of ten or more by actually explicitly taking data and analyzing it carefully. So the idea of having new experiments was an important thing. The Nord-Ged effect in lunar laser ranging is a clear example. and even earlier the Shapiro time delay predated my entry into the field but there were certain ideas that came along out of theory that motivated experiments so was that probably were they probably I guess I'm wondering if there's a distinction between new effects that theories come up with that are possible that are amenable to being searched for by experimentalists, and new alternative theories. In other words, were there many cases or any cases of new interesting possible effects arising out of work on alternative

57:30 of theories accepts, well, I think you did give an example of saying that that came out of the alternative, the framework of parameterized close-leaf funding, but did any of the alternative theories necessarily suggest? Probably not, in the sense that... I mean, you can do all these analyses within the PPM framework and you didn't even have to have an example theory, you just had to frame it. I mean, the stuff on the various effects related to preferred frame terms in the PPM framework that I worked out and Ken and I did, we had no particular theory or example of theories at that time. In fact, the first paper I wrote, I had to sort of cook up a kind of totally bogus ad hoc theory to illustrate that you could have a non-zero value for one of these parameters. And in itself is not a real theory. I just sort of play it around, kind of a toy thing. And so in some ways, generating some theories that had these effects came after developing the general framework and working out what the predicted effects would be. So those were examples. And then, I suppose, then once I started to look in the literature and actually could find a few other example theories that had these effects, but they weren't sort of there, you know, a priori. So in many ways, many of these new experiments came out of the PPN framework rather than out of specific theories that said, here, I'm predicting this effect. So is there a connection between the growth in theory testing and theoretical frameworks like PPM? Did that have an effect on the production of alternative theories? You certainly said it did within Kibbs Group, but I was wondering, was it the case that theorists had always played around with alternative theories that had just sort of sat there? No, I doubt if it had much of an effect. There were always alternative theories around in literature. And let's just put aside some of the really junk and crazy theories, but sort of reasonably serious theories.

1:00:00 And partly because they were motivated for their own serious reasons. I mean, there's a whole class of theories based on something called Poincaré, a way to try to connect, you know, add new symmetries to this meaning of Heidel's group and other groups in Germany and such, and generating some theories of this sort for years. And so they come at it from a fundamental level, field theory, group theory level, they want some kind of structure, and it's not even, and so then the theories are very similar to GR in some way. Or theories, people have always been interested in theories that have either a non-symmetric connection or some spinner structure. And so there's a whole industry of theories like that. And so they've always been around and they haven't been changed as far as I can tell, one way or the other by this theory testing program. Partly, especially in the case of these theories with torsion, that's the torsion theories. At some level, they make no predictions that are different from general relativity for ordinary experiments because the torsion always gets wiped out for macroscopic bodies. So if they make a prediction, any difference, it's only either for quantum mechanically spinning bodies or other regimes that aren't accessible to testing. So, you know, they're out there and there are all these theories around and they're, in principle, different from GR, but as far as we can tell, they don't make any different predictions in anywhere we can access. So, but there's certainly a continuing industry in that area. So would it be fair to say that the theorists have their own reasons, or some theorists, but anyway, have their own reasons for coming up with alternative theories for arising out of issues as a principle within theory, and that on the whole these alternative theories have, by and large, a little bit of an impact on experimentalists, but what's interesting for the experimentalists are new effects, such as some of the ones you mentioned, and that perhaps the Paramatron's post-Newtonian framework, to give an example, gives them a set of quantities, a set of measurable quantities, that they can apply their tools

1:02:30 Yeah, I mean, I would agree with that, and it's also, these kinds of frameworks are sufficiently phenomenological that they can actually understand them, too. You find that even in particle physics, I mean, there's a particle phenomenology where you, I mean, you take QCD, which is this horribly complicated thing, but you can then work out things where you calculate quantities or calculate certain quantities in terms of other quantities. there's this phenomenological theory that comes out of a more fundamental theory. But it's the phenomenology that the experimentalists interact with, not the underlying basic theory. So in that sense, the PBN framework plays that kind of a role. It's similar to particle phenomenology. And because in the case of the particle form of countries and so on, you had this, what seems almost like, cultural passion within relativity but some people were saying, well, these people are being too phenomenological in their approach and insufficiently rigorous. Was any of that criticism ever directed at things like PPN frameworks and so on? Not that I know. It's a different level altogether. I think it just came down to the issue of radiation that seemed to make a big difference because I think from the point of view of the mathematical relativists, it was the first place where the real causal structure of space-time became important. Space-time, I mean, PPN framework doesn't really involve space-time all that much. It's just like it's Newtonian theory, but with some corrections, but it's all basically retardation, doesn't affect anything. So it's really pretty straightforward. It doesn't really probe anything about space-time that makes it uniquely space-time, But as soon as you got into radiation, radiation damping, then their ears perked up and they got involved in the issues, not that they were trivial issues, they were important issues. How do you solve a problem like this when you now have to worry about the light cone structure and retarding the fields, radiation and such, and how the radiation backs back. But all that's absent in the PPM frameworks, and I've never heard anyone, like, no Juergen Ehlers ever came to me and said,

1:05:00 you know, I disagree with the theoretical foundations of the PPM framework. He always loved the PPM framework. So since I'm on the topic, maybe I'll finish by getting some of the sort of biographical details right. So you began as a student at Caltech in 68? I was 68, entered Caltech. And did you, at what stage did you start working on PPN stuff? Was that something that was really in the air when you started as a student? It wasn't so much in the air. I mean, at that time, I know Kip was already interested in sort of thinking about experimental gravity, because he'd been talking to people at JPL about the upcoming space missions, and the fourth test of general relativity, as it was called then, the Shapiro delay, and, of course, things like quadruple moment of the sun were still described. So it was, you know, it was a topic he was interested in. And at that point, it was just a question of which student to assign to work on these things. You had other students like Richard Price and Jim Hipser, and they were all working on things related to black holes or perturbations of realistic stars, and things that really have to use with gravitational waves, because at the same time, people were thinking about Weber's experiments and possible detection of gravitational waves that everyone thought might be imminent. And so basically, as it just happened, the spring of that year, in Kipps' relativity course, we had to do a term project. And so I did a project of basically re-deriving the Einstein and Feldhoffman post-Newtonian equations of motion, but using Chandra's fluid, starting with Chandra's fluid equations finding fluid balls and integrating and reducing it to a, sort of a, quote, point mass form. And so basically, essentially that summer, Kip said, well, why don't you think about testing, do some reading and think about testing theories. And so by the end of the summer, we had written just that we wrote a little commentary paper

1:07:30 from some journal that no longer exists about the potential for testing alternative theories. But it was, I think, during that summer that I discovered Ken Norvett's papers and some earlier paper by Ralph Beyerlein, which did a sort of a half PPN framework. He threw in some parameters, but not all of them, in some sense. And so, then I realized, I saw there were things about Ken's way of doing things that I didn't like, and so I thought it would be better using Chandra's fluid equations. So, I took off. I mean, that was done in two years. I hold the record for, I think, one of the fastest PhDs in Caltech history. So I did my thesis defense in sort of the middle of March of my third year there. And I didn't have a master's degree when I came in. I was just a regular bachelor's degree person. But it was just, you know, everything was just right. I was there at the right time, hit the right thing. I mean, it was pure luck. And so then, I guess you had the work sort of done by the time these experiments were were coming to fruition, sort of, like the, or, I mean, the, the, the, the time that you sort of graduated, someone coincided with one of the experiments that were being considered or starting at the time that you began as a graduate student, were they sort of finishing coming to fruition then, or did that take somewhat longer? Let's see. I have to look at the various experiments. For example, there were a number of experiments going on in a period between 1969 and 1972 or so doing long baseline interferometry for blade deflection. And so they were sort of happening all the time and papers were republished. Some of the early tracking of experiments involving Mariner and such, I think some of those results came out in like 1970, so sort of midway, but some of these other things

1:10:00 like new experiments that sort of came out of the PPN framework, I mean it took a while for them to sort of happen in a, I mean, in summer 71, I stayed at Caltech another year as an instructor, but so between my graduating in that year, in summer 71, I spent a summer in Montana State, Ken and I sort of decided to unify the two frameworks, because we used different rotations and different ways of doing things, so we decided to agree on a unified approach, and so we wrote a couple papers where we presented sort of the official a canonical PPN framework and also looked at things like some additional tests of preferred frame effects. It was sort of that paper that then got some people involved in looking for some of these preferred frame tests that were done within a few years after that. So once you had the kind of canonical formation, it was at that stage that it really had most Most of its impact and inspiring new experiments. But also, you know, it also helped that, again, the summer after that, in 72, I lectured at the Verena Summer School and then wrote up the lecture. So there were then some places where all this stuff could be written in monograph form or there was a long review article that people could use and refer to. Right. You mentioned, I guess it's sort of that thing, you mentioned that while you were at Caltech and once you sort of worked on this for a while it sort of became easy to make up your own theories and that this was sort of the industry in the group for a while so you've kind of had experience that's not too, the generating theories is the attraction just intellectual curiosity or is it kind of a way to explore particular issues, theoretical issues that come up, or what were the motivations in that case? It's just that you could do it fairly easily then. Partly, yeah, I mean, it was some of both. I mean, it would be curiosity in the sense that it was kind of fun to do. You certainly learn a bit more about both GR and just general things, doing these calculations. but we also especially at that time wanted to explore some issues

1:12:30 having it seems that if you take general relativity and you can extend it to scalar tensor gravity but if you try to go beyond with some other kinds of things like a vector field you run into some crazy theories although Ken Norbert and Ron Hellings did produce a theory with a sort of vector field added to the metric so we sort of asked a tensor field, and what about these theories where you have sort of these non-dynamical ingredients like a flat background metric or theories where you, you know, people would invent theories in which you simply define the metric once and for all to be diagonal with a, you know, g0,0 to be of some form and a spatial metric to be of some form. So in the old days, people would write these things down and then propose some field equations for the various fields, like e to the phi and then e to the minus phi, sort of a quasi-Schwarzman-like metric, and then cook up some field equations for that. Well, the question is, can you really turn such a theory into a well-defined theory? Is there a way formally to make this a good theory in the usual sense? Whereas previously, you would have rejected it as being crazy. It proposes a preferred frame in the universe because it says that in the universe there's a frame diagonal form, whereas ordinarily it wouldn't be diagonal, it would have all components. But can you define this mathematically and make this a rigorously defined theory? And so it turns out you can if you define various so-called prior geometric variables, things that aren't dynamical, that you can still define geometrically, they're just not dynamic. Cosmic time function that establishes the bird frame and a flat background metric that you've just defined by demanding that it's Riemann tensor banish everywhere. So that's, you know, you would do things. So that was nice to be able to generate theories that had nice, rigorously, and geometrically defined properties, and then you could calculate what they predicted. And so this is what people like Victor Nene did, and David Lee and Alan Lightman. So they produced a few of these kinds of theories. Strangely enough, I mean, right now, there's an issue that I haven't worked on, but I'm thinking hard about having it with the gyroscope experiment, because partly as a result of all this theory, this PPM framework stuff that we've been doing, there's now all the

1:15:00 evidence that shows that the effect of dragging of initial frames, PPM parameters that enter into that formula, are now known far more accurately than the gyroscope experiment will ever measure. For years, this has been a big criticism of the experiment. We're spending $600 million on this experiment to test something that's already been tested much more accurately, even though indirectly, by other means. And I've been involved in lots of committees to evaluate it and to recommend it, you know, so I've been involved in all these discussions about whether it's a worthwhile thing or not. And currently, I'm chairing a science advisory committee for NASA for the experiment to help them to make sure that whatever science comes out of it is good science and that they get answers that are believed. But one issue that I've said to myself that I want to look at again is whether or not you can invent a theory of gravity that agrees with all of these existing constraints for frame-dragging from GR. Standard metric theory, any metric theory that agrees with the PPN structure can't. It's just guaranteed. If you have such a theory, then the same parameters that appear in some of these other tests will appear in frame-dragging, and you're dead. So is there a way to prove something more general that might, you know, tell you that the gyroscope experiment is measuring something that's not totally been measured before. It's actually a very hard question. I still don't know if it can be done. It'll probably be a non-metric theory. It'll probably involve intermediate-range fields, fields with different eucalypt-type potentials, and maybe several. So you have to fine-tune things so that on some scales where certain experiments work, You measure one thing, but then on the scale of the gyrus experiment, that effect's been suppressed because it's got a, it's exponentially damped, and then you have another one that's, it'll be, I think it'll be tremendously ugly. And it still may not be possible, but it's something I'll have in the back of my mind to try. Yeah, it's something, it's something to channel. And so are non-metric theories still generally regarded as viable, at least in some instances? Well, they're not regarded as terribly viable, just because all the various constraints from that class of null experience is so tight.

1:17:30 But on the other hand, as everyone in string theory says, string theory does predict non-metric violations. Because the scalar fields that come out interact directly with matter, either with electronics or various parts of the standard model sector. So they'll generically produce violations of quality of free fall. again, they could be negligible because the scalar field is suppressed by a short range cutoff or they could just be weak much weaker than experiments but they're generically should be there they're dead but people are strongly motivated to push tests like tests of the equivalent of the weak equivalence principle further because in some sense that's our only hope to look for, to see any such effects that might But the kinds of non-metric theories that we sort of thought about in the 70s that used to be in the literature, those kinds are pretty dead. Things that we have like two metrics and one metric couples to electrodynamics and one couples to ordinary matter. I mean, you pretty much kill those by experiments. So the theory that you're thinking about now with the intermediate range forces is an example of one where an alternative theory that's in a way motivated by experimental needs appears. Yeah. Suppose we had a theory at this time. Yeah, I mean, that's right. So that's coming up from time to time. It'll be kind of after the fact. It's not a theory that you want to test, but it's basically a straw man theory that you set up to see what the potential of a certain experiment might be. But it's so late in the game because it's actually very hard to do, and no one has come up with such a theory beforehand. But in the end, everyone hopes that the experiment will yield a result of the degrees of general relativity. But it would be nice to know what it would mean, what it would imply if it gave a result that didn't agree. One thing it could apply was that there's something wrong with the experiment, and that, of course, is what everyone would think. But it would be nice to have at least something, some statement you could make and say, well, if you really believed that this was a real result, that they measured something, this is what it would imply.

1:20:00 And you could actually go, how horrible would that be? So with the whole paraphernalia of the advisory committees and so on that exist for these various experiments, Do you ever find yourself kind of at loggerheads with the experimentalists or with the people who are, with the NSF people who are wanting the committees to be set out for a visit? Well, my main experience on this has been with Gravity Pro B. And there, it was the The one committee that I was on was set up by Dan Golden to basically give a thumbs up or thumbs down, a final thumbs up or thumbs down on the mission. He had decided that he was either going to kill it and close it down, or commit to it, and commit to it all the way to the end. Because for years it had suffered under previous administrators who every year would cut its budget to zero, and then Francis Everett would go to Congress and get Congressmen to write the money back into the budget. It was always, there were years of game playing with it because no one really wanted to commit all the funds to make it go because it was very expensive, astrophysicists hated it, blah, blah, blah. But Golden decided this was crazy to do this. We should either kill it outright now or say we're just going to do it no matter what. So we had this committee to review it. And it was very controversial because as many people told me, one of the great successes basically destroyed the motivation for the gyroscope experiment. In the 1960s, when it first came along, measuring frame-diagging was an important thing, because it had never been measured before. We didn't know the parameters that went into it, but because of the PPM framework, the two parameters that go in, gamma and alpha-1, now through these new one-th gamma, of course, from the bending of light, which got better and better, and alpha-1 through some of these geophysical and other tests frame effects over time became known just so much better and so it basically destroyed the motivation for doing the experiment. Also, as people like Ken Norvath point out, frame dragging is implicit in a bunch of other effects like the Norvath effect. It's absence. If you sort of turned off the frame dragging term, then lunar laser range wouldn't agree with what's observed, similarly in the binary pulse.

1:22:30 So, you know, because it took 30 years to get the experiment to go, and now pushing 40 years, sort of time passed it by. Right. And so the argument was made that we shouldn't be spending all this money on this experiment, because it's not going to tell us anything new. And even if it told us, and if it told us that if it gave an answer in disagreement with GR, no one would believe it, because we now believe GR so strongly, your natural assumption would be that the experiment was flawed or something went wrong or there's a systematic effect that wasn't taken into account. So it was very, I don't want to say very controversial, I mean, in the way people weren't angry, there was great disagreement on this particular panel, as there had been on other panels, about what to do. But ultimately it was decided that for many complicated reasons, including simply the fact that the effect hadn't been measured directly before. Sure, you could make all these arguments about the TPM framework. Maybe you could cook up a theory that could give a different answer. No one's done it heretofore, but maybe it's possible. But still, no one's actually measured this precession of gyroscopes. So in that sense, it was important. Of course, obviously, you can't deny the fact that half the money had already been spent. and so to kill a project after $350 million it would have been spent seems like a waste of money and technology so in the end they recommended to Golden that we should go forward and he is committed they've gotten the full funding every year to go forward the launch date's set the spacecraft is built they're putting the stuff into the spacecraft now cooling it down October 2000 is the launch date and they're on track to do it so I have to give him credit for biting the bullet, and I don't think it was a very popular decision. But strangely enough, that decision has really improved the situation on the astrophysics part of NASA, because basically the managers of those programs and the astrophysicists who deal with those programs have seen, okay, fine, we're going to spend the final $250 million. Instead of arguing over that money and bitching over it, it's committed, let's take the money that's left and make the best use of it. And so it's decreased a lot of the acrimony that continually was going around in those various branches of NASA that had to deal with gravity probing.

1:25:00 So I think it was a smart decision on this part. So I guess there were two possible answers, or two answers, maybe there would be two possible answers to the problem that the PPM formalism, as you say, helped somehow take away a large part the motivation for the thing, for the experiment, and one is to sort of, you know, ignore the theory to some extent and say, well, look, we still have to actually measure the effect directly, never mind, the theory now gives us a way to calculate it from other results, and the other is, well, okay, is to go back to the theory and say, well, okay, why don't we, we could still step outside PPN and still come up with an alternative theory, an alternative non-metric theory perhaps that would still predict a different result that we could test Right, quite frequently I would find myself arguing both sides of those things simultaneously the trouble is from things like this I would be arguing that the PPM framework says all this and says the experiment's worthless but then at the same time I was I mean I supported it, I felt that we should do it so I would sort of go back to the ultimate in the phenomenological viewpoint and say that even if we know all these ppm parameters we should still measure this now strangely it turns out people have someone like Ken Norvitt argues that he thinks gravity probably is important but not because of frame dragging but because of measuring gamma by this geodetic recession because if you can measure gamma at a part of two to five then you're testing scalar tensor gravity to that position and then you're getting into this you know testing strength So to him, that's the more important scientific motivation. Okay, well, thank you very much. Okay, thanks. Very interesting. I don't know if there'll be a need to do this, but I should ask you if Harry Collins wanted to listen to the transcript of the tape, would you have any objections? I don't think I said anything that inflammatory. Okay. Just don't show it to Ken Starr, that's all I can ask. Well, I've been...