310 | Marc Kamionkowski on Dark Energy and Cosmic Anomalies

0:00:00.9 Sean Carroll: Hello everyone, and welcome to the Mindscape podcast. I'm your host, Sean Carroll. I'm a working theoretical cosmologist, among other job descriptions. So recently there's been some news in cosmology that may or may not turn out to be a big deal. This is often how it is in science. You get a result, but of course, by the very nature of having gotten a new result, it's a hard result to get, otherwise he would have gotten it earlier. So the first indications that something interesting might be happening are typically faint. And you're not sure whether they're on the right track or not. But there's a couple of different things that have indicated that perhaps there are kinks in the armor of the standard cosmological model, the so called Lambda-CDM model. Lambda for cosmological constant, CDM for cold dark matter. Not something that throws away the whole Big Bang scenario or anything like that, but specific details might need to be tweaked. This is something that I could have done a solo episode about, but the data and exactly what the data are telling us really, really matter here. So I thought it would be better to have a true expert on the podcast.

0:01:11.0 SC: So we're happy to welcome Marc Kamionkowski, who is my colleague at Johns Hopkins. And someone I've known for a long time. We've written papers together, including suggesting the idea of dark electromagnetism in addition to dark matter out there in the universe. We don't talk about that in this podcast. Instead we're talking about these accumulating possible anomalies in cosmology. Most recently, there's a survey called the Dark Energy Spectroscopic Instrument, DESI, that has suggested that perhaps the density of dark energy is changing with time. Which is not what you would expect if it was just a cosmological constant. If it were a dynamical field, you might expect something like that. And there was a hint a year ago that that was true. Very recently, the hint has become stronger. And there is another instrument called the Dark Energy Survey DES, as opposed to DESI, for the Dark Energy Spectroscopic Instrument, that has less firm results, but also pointing in the same direction, that dark energy might be evolving with time. These are both amazing surveys. Interestingly, they both look at galaxies, out there in the universe, and they look at the distribution of galaxies and how they're evolving with time and things like that.

0:02:32.3 SC: They're both ground-based cameras that replaced previous cameras. The Dark Energy Spectroscopic Instrument, DESI, replaced the camera at Kitt Peak in Arizona. And the Dark Energy Survey replaced a camera in Chile, the Victor Blanco Telescope. Anyway, these hints that dark energy might be changing with time are still tentative. It's not completely clear yet. And indeed, at face value, it would be remarkable if they were really true because of the specific way in which the dark energy is evolving with time. So we're going to get into that. But of course, I have to take advantage of this to also talk about other cosmological anomalies. The Hubble Tension, which we talked about with Adam Riess some time ago. Marc turns out to be one of the world's experts in thinking about models to explain the Hubble Tension. Marc was in on the ground floor in thinking about the cosmic microwave background as a cosmological probe, and also was the author of some interesting ideas about what dark energy could be back in the day. So he's really the best person to talk to about what the microwave background tells us, what these galaxy surveys tell us, and what the theoretical implications are of all this stuff.

0:03:45.5 SC: I would say that right now, I'm still on the fence about whether there really truly is something dramatic going on, but it's absolutely a legitimate possibility. Sadly, we're still gonna have to wait for even better data to come in. That's how science goes sometimes. But if you listen to this episode, you'll be well prepared to understand what's happening when that data does come in. So let's go.

[music]

0:04:25.3 SC: Marc Kamionkowski. Welcome to the Mindscape podcast.

0:04:28.3 Marc Kamionkowski: Hello. Pleasure to be here. Nice to be talking to you on this beautiful Wednesday morning.

0:04:33.2 SC: I know you're back in sunny Baltimore. I'm here in Santa Fe. But, yeah. It's a reasonably nice day today. A little cooler than yesterday, but probably more oxygen. Santa Fe is at very high altitude. Yes, gets me. We are here because there's been a couple of, more than a couple of anomalies, challenges, puzzles, whatever you wanna call them with respect to the standard cosmological model, which is nowadays known as Lambda-CDM. So we're gonna talk about that. But let's first explain what is the standard cosmological model and why do we believe it? Give us a medium-sized intro to where we are before we have any anomalies.

0:05:15.1 MK: Okay. Medium-sized intro to where we are before we have any anomalies. So we live in a universe that we have been observing for centuries, but I would say over the past 100 years in particular, our understanding of the universe, which is everything that we know, it's one physical system, has evolved tremendously. And it sort of started just under 100 years ago really, with Hubble's discovery that the universe was expanding. So everybody knows that the Earth spins around the sun. And the sun is the center of the solar system. Most people know that the sun is one of about 10 million stars in our galaxy, the Milky Way. And the sun spins around the center of the Milky Way for the same reasons the Earth spins around the Sun.

0:06:11.7 SC: So I think you said 10 million?

0:06:13.6 MK: 10 billion. Sorry.

[laughter]

0:06:16.0 MK: Good to know you're paying attention.

0:06:18.9 MK: Team up the galaxy. Yeah, 10 billion stars. And so, the sun spins around the center of the Milky Way for the same reason the Earth spins around the Sun. And that's because all of the stars in the Milky Way generate a very strong gravitational field. And you might then wonder whether our galaxy is part of some larger structure, whether our galaxy is one of 10 to 10 billion galaxies that spin around each other. But it turns out that the hierarchy ends there. And our galaxy, it turns out, is one of tens of billions of galaxies that are more or less the same that we know about. But the galaxies don't spin around from each other. It turns out that every galaxy is moving away from every other galaxy. And this is what Hubble discovered almost 100 years ago. And the relative, the speed at which any two galaxies are moving away from each other is proportional to their distance. And so, the interpretation of this is that the entire universe is expanding. This was discovered by Hubble. And it turned out that it was kind of convenient, because Einstein had discovered general relativity 12 years before that.

0:07:29.3 MK: And several people who were studying general relativity realized that, equations of general relativity allowed for such a universe that was filled with a bunch of stuff where everything was expanding, everything was moving away from everything else. So that was sort of the birth of the standard cosmological model. And since then, we've discovered a bunch of other things. Perhaps the next big breakthrough was sort of in the mid '60s. There was a discovery of something that we now call the cosmic microwave background. Basically, the idea is that if everything is moving away from everything else today, if we were to make a movie of that expansion, then run it backwards, at some earlier time, everything in the universe would be on top of everything else. So although the universe is a fairly low density place now, if everything's moving away from everything else at some time in the past, which we call it the Big Bang, the density of the universe would have been very high. Anybody who puts lots of air in tires and drives them around knows that when densities get high, the pressures get high, the temperatures get high. So the early universe, we have good reason to believe, was very hot.

0:08:40.4 MK: And if you look at a fireplace where there was a fire that is now out, the embers still glow for some amount of time afterwards, even though there's no fire, you can still see residual heat. And in 1965, we discovered this residual heat, the cosmic microwave background. So it turns out that we discovered another relic from this Big Bang that consistent with this picture of an expanding universe that Hubble sort of gave us a hundred years ago. Is this good so far?

0:09:10.7 SC: This is great. Yeah, I love it.

0:09:12.3 MK: Okay. Just checking. [laughter] And then, so that was 1965. So that was 60 years ago. And since then, we've learned even more about our universe. So we've been able to study the distribution of galaxies in the universe. And we find that the universe, on the very largest scales, is very, very smooth. So it's like a pond on a clear day, on a calm day. But if you look very carefully, there are some fluctuations. There are some small amplitude ripples as if there was some light wind. We've also been able to look at this cosmic microwave background very, very precisely, very carefully, and we've been able to see, that the temperature of this glow, this afterglow of the Big Bang, is not precisely the same everywhere. It's pretty close. It's close. The temperature is the same to one part in 100,000. But if you actually look really, really carefully, there are small fluctuation. And we believe, we have very good reason to believe that these small fluctuations that we see in the cosmic microwave background, for then the seeds for the larger amplitude fluctuations see in the galaxy distributed verse today. We believe that those small fluctuations were amplified by gravitational gravitational forces.

0:10:37.6 MK: So we have all these very, very detailed measurements of the cosmic microwave background of the distribution of galaxies. And we have a model, that allows us to relate the distribution of galaxies in the universe today, to the distribution of the cosmic microwave background that we see, the afterglow from the Big Bang. And in order for our model to account for the features that we see both in the cosmic microwave background and in galaxies, we need to have, we need in these models, in addition to the ordinary stuff that you and I and everything in the solar system are made of, which we call baryonic matter with jargon for ordinary atomic stuff. In addition to the baryons, we also know that there has to be a lot of dark matter about five times as much mass in dark matter as in baryons. We don't know what dark matter is, but the models require that the dark matter is required in order for the models to work. And then there's also something called the cosmological constant that was inferred in the late 1990s. But we now also understand from the models that we have for these fluctuations that it has to be there.

0:11:47.4 MK: And again, the cosmological constant is something we don't really know what it is, but some sense, it's some energy density that pervades all of space. So we have this great model explains the origin of the universe, the expansion of the universe. We have some ideas about why it's expanding, although those are not fully formed yet, I would say.

0:12:09.9 SC: You mean what started it in some sense?

0:12:11.7 MK: Yeah. What set it in motion?

0:12:13.0 SC: The good news for you, is that I have a future upcoming podcast about what happened near the Big Bang, so you don't have to worry about that. [laughter]

0:12:19.6 MK: Oh, really? Near the Big Bang? What about before the Big Bang?

0:12:22.4 SC: Oh, yeah. That's gonna be there.

0:12:24.5 MK: Yeah, that should be fun. Okay, so we have this great model that explains all this wealth of observations we have of the galaxy distribution. This is millions and millions of galaxies that we've been able to map, and the temperature of the cosmic microwave background, we've been able to measure it about a million different points in the sky. So there's a lot of data. It's not just a hand, wavy, squiggly approximate model. It's not like about 3,000 miles from New York to Los Angeles. It's 3,118.632. And it's a really good model. And we're really proud of ourselves.

0:13:02.8 SC: I think we should be. Let me pause though, for a second because something sneaked in there that I think is really interesting. A lot of people, I'm sure that you get emails from people who have explained away dark matter without being professional scientists, et cetera. And of course, they always concentrate on the rotation curves of spiral galaxies. So this idea that the amount, the rate at which stars and gas are rotating around the centers of spirals, depends on how much mass there is, et cetera, et cetera. Ordinarily, Vera Rubin and her collaborators prove this. We attribute that to dark matter, but it could be something else. But you didn't even mention spiral galaxies. You went right to the microwave background.

0:13:42.9 MK: Yeah, that's a good point. So I think I did that because I was trying to give you a capsule summary of...

0:13:51.6 SC: No, I like it.

0:13:53.4 MK: The model for the universe. But yes. So the measurements of the cosmic microwave background and large scale distribution of galaxies that I told you about that required the existence of dark matter. Those happened about 25, started happening about 25 years ago. But you are correct that even 20 years before that, around 1970, Vera Rubin and her collaborators and a few other people started to realize, that most of the matter in the galaxy has to be dark. And so, we actually had reason to believe, we had good reasons to believe that there would be dark matter in the universe even before these large scale structure cosmic microwave backgrounds that I told you about. So in some sense it wasn't a surprise when that happened, but it was a confirmation and it gave us much more confidence that the anomalies that we were seeing with galactic rotation curves were actually real and due to some new form of math.

0:14:53.0 SC: But the reason why I like to emphasize it is because, it does kind of highlight a difference in how the professionals think about this, than how we perhaps talk about it to the broader public. We tend to be historically quasi accurate and we wanna give the early people credit. So we talk about spiral galaxies. But the real reason we are confident that there's something like dark matter is much more, something like some combination of the microwave background radiation, large scale structure, things like that. And so, accounting for spiral galaxies doesn't actually get you out of the need for dark matter.

0:15:28.1 MK: Accounting for spiral galaxy does not get you out of the need for dark matter? Yes, that's right. That's right. [laughter] I'm trying to parse what you said. If there were indeed, if somebody had some other explanation for the galactic rotation curve that did not involve dark matter, we would still have reason to believe that dark matter exists as of observations of cosmic microwave background and galaxy disk.

0:15:52.6 SC: Sorry to hector you on that, but it is the internet that we're talking to here and there are people out there who have ideas. And we love them, and we support their efforts, but we wanna be clear about why we believe these things.

0:16:03.3 MK: Yeah, it's actually, it's a good point that you make, and I think it's something that we are becoming, we've always known, but appreciate more in cosmology with time. And that is that, when we do cosmology, it's sort of like archaeology or physical anthropology or paleontology. Yes, paleontology. [laughter] Missing the word there. With paleontology, what you do is you find bones somewhere, and they have these funny looking shapes, but you look at them and it's sort of like a puzzle. And you sort of try to put the pieces of the puzzle together, consistent with what you know about bones of animals that exist. So it's a puzzle, but it's also informed by, your solution to that puzzle is informed by other solutions to similar puzzles you have. And we do the same thing in cosmology. It's very similar. It's not an experimental science, we don't like in paleontology, you don't build a dinosaur, although some people are trying yet, you don't build the dinosaur. We can't alter the system. We just have observations. There are things that we find with telescopes. And so, we try to construct a model that's consistent with the observations and consistent with what we know about the laws of physics.

0:17:30.2 MK: And so, if we have a model for galactic rotation curves that involve something other than dark matter, that's a perfectly legitimate thing to try. But then, you have to ask, is that solution going to be consistent with other things that I think? And now with cosmology, we try to make as many different observations we can, try to study as many different systems as we can in detail. And in some cases, there are things we can try in the laboratory. But basically, in order to actually have confidence in conclusions that we make, in order to increase our confidence, we wanna have different measurement and different observation from different systems and different types of observational techniques, that then all match and give you...

0:18:16.8 SC: Speaking of which, this dark energy business, this cosmological constant business, where did we figure out that?

0:18:25.0 MK: So the cosmological constant, the story is that Einstein had realized that there might be a cosmological constant, the Einstein equations, and called it viscous blunder. Whether that's true or not, I don't know. I actually saw the notebook, the page, you know Diana Buchwald, Einstein papers project?

0:18:48.0 SC: Yeah.

0:18:49.5 MK: So Diana once showed me the actual notebook pages, pages in Einstein's Notebook, where he was doing the calculation that led him to think of the cosmological constant. And it was kind of interesting what she told me. She said, it's the only case that they have in all of his papers, the only example in all of his papers where he was actually doing a numerical calculation.

[laughter]

0:19:11.2 SC: Flagging in the numbers?

0:19:12.5 MK: Yeah, he actually had a graph, and those graph papers, he was trying to calculate the area under the curve.

0:19:19.0 SC: Oh, wow.

0:19:19.5 MK: He did it by counting the boxes. So anyway, the cosmological constant sort of existed as a possible theoretical addition to the basic theory of general relativity for over 100 years. But the observational evidence that I think actually exists, came about in the late 1990s. And you can look in the literature, even before the late 1990s, there were people who were sort of speculating that various cosmological observations were better fit with the non-zero cosmological constant. But the real smoking gun was measurements made by two independent groups, the Supernova Cosmology Project and the, God I don't remember the name.

0:20:08.7 SC: High-Z Supernova Team.

0:20:10.3 MK: Yes, the High-Z Supernova Team. They sound like the same thing, yeah.

0:20:14.7 SC: And so, we talked earlier about how every galaxy in the universe is flying apart from every other galaxy. And if you think about a ball that I throw in the air, if I throw a ball in the air, it goes up, but then experiences the gravitational attraction to the Earth. And so, even though I throw it up initially with some large velocity, the velocity slows, eventually goes to zero, becomes negative, and then it falls back down. Now, if I had a really, really good arm and I could throw the baseball at a velocity bigger than 11 kilometers per second, I don't know what that is in miles. If I could throw a ball with a velocity greater than 11 kilometers per second, it would actually escape the gravitational field of the Earth. Would actually, instead of going up and then flying back, the ball will actually fly away from the Earth and continue flying away from the Earth forever. But since gravity is a long range interaction, a long range attractive interaction, even though that baseball was flying away from the Earth and would continue to fly away from Earth forever, the speed at which it does so, would be continually decreased.

0:21:30.6 MK: So ordinary gravity, ordinary Newton's gravity, suggests that if two galaxies are flying apart from each other, the relative speed at which they fly apart from each other should be decreasing with time. And that is in fact, what was in the standard cosmological model based on Einstein's general relativity, with no cosmological constant until the late 1990s. And then what happened is, that the High-Z Supernova Team and the Supernova Cosmology Project independently actually measured how fast galaxies were moving away from each other. And what they found, is that the speed at which galaxies are moving away from each other is actually increasing with time rather than decreasing. And this was Science magazine's breakthrough of the year in 1998. It was completely and utterly shocking to everybody in physics. We knew that general relativity, allowed for the possibility of a non-zero cosmological constant. But everybody just assumed that it would be zero, because the actual value is something like 0.000 with 120 zeros. The actual value is extremely, extremely small. And physicists don't like extremely small numbers. We like one, we like Pi, we like 2.3. [laughter] We don't like extremely small or extremely large numbers. So everybody was very, very shocked.

0:23:04.6 MK: I remember being very, very skeptical. People tried to explain it away. They tried to suggest that maybe the supernovae themselves were evolving with time. They speculated that maybe light was being absorbed by the more distant supernovae, thus making them look fainter. And the people in both projects did a really good job, checking all of these things and dispelling all of these possibilities, ruling out all these possibilities. And then, I became really, really convinced when the cosmic microwave background experiments came out in the early 2000s. And from a completely different type of measurement, different type of observation, they also inferred that there had to be a non-zero value of the cosmological constant. So, did I answer your question?

0:23:51.9 SC: You did. And thus Lambda-CDM, the lambda for the cosmological constant, CDM for cold dark matter. That is the standard target fiducial cosmological model?

0:24:03.5 MK: That is our standard cosmological model. I don't like the name.

[laughter]

0:24:08.3 SC: It's not the sexiest name, but maybe we'll overturn it. So that's okay. We can come up with a better name. I guess, the one other piece of cosmological measurement that I wanted to get on the table was the idea of a baryon acoustic oscillation. I think that's probably the trickiest thing for the person on the street to wrap their brains around, but apparently very, very important to modern cosmology.

0:24:31.4 MK: Yeah, this is the hardest thing to explain, but I'll try. So we know from our observations of the cosmic microwave background that the early universe was very, very smooth. So I said it was sort of like the surface of a pond on a very calm day. But suppose, I threw a pebble into that pond. There would be a splash, but then there would be a wave that propagates out from the place where, a circular wave that propagates out from where the pebble landed in the pond. And so, that wave expands with time, and it's moving at some velocity. And at early times, if I were to take a snapshot just a few seconds afterwards, the circle would be small, that wave would be small, and at later times, that circular wave would be larger. Now, I told you that although the early universe was very, very smooth, it was not perfectly smooth. And so, there were sound waves propagating the early universe. The early universe consisted of this fluid. All the baryons that make up the galaxy and you and me, the sun, all the other stars, all those baryons would have made up a fluid in the early universe.

0:25:49.1 MK: If I have a disturbance in the early universe, if I were to throw a pebble into the early universe, there would be a wave that propagates out at the speed of sound. Now, although we don't see an individual such wave, what we do see, if I have a pebble in a pond, and I throw it in the pond, there will be that circle which is the wave propagating out, but there will also still be some bubbling right at the center.

0:26:16.3 SC: And so, there will actually be a correlation in the surface height of the water, at the center and at the wave. So the surface height far away from the wave is zero. The surface height inside the wave is pretty small. But there will be an increase in the surface height at just this right distance. And so, when we look at the galaxy distribute, we don't see any individual wave, but we can measure the probability to find one galaxy at some distance from some other galaxy. And if you look at that probability to find one galaxy at some distance from some other galaxy, that probability decreases as you go to larger and larger radii, the excess probability. But then, it turns out that there's a bump somewhere around 100 megaparsecs. And that bump is actually essentially a consequence of these sound waves in the early universe.

0:27:16.3 SC: Okay. So it's, roughly speaking, metaphorically, God threw pebbles at the smooth pond of the early universe and ripples went out. And there's gonna be sort of a natural correlation length between the different galaxies that we see today, because of just the timescales at which everything happened?

0:27:32.7 MK: That is correct.

0:27:34.2 SC: Good. And that's the baryon acoustic oscillation, BAO?

0:27:38.2 MK: Yep. It's pretty remarkable when you see it in the data.

0:27:41.2 SC: Yeah. [laughter]

0:27:43.1 MK: This is another example of the whole, everything hanging together. We had the cosmic, we have the expansion of the universe, we have the cosmic microwave background. We have these cosmic microwave background fluctuations. And in the, I remember when I was a postdoc and assistant Professor in the mid 1990s, people sort of understood that you should also see a bump in the galaxy distribution. But I remember thinking that, there's no way we'd ever be able to see this. It's an interesting theoretical idea, but you'd need to measure the positions of, God knows how many millions of galaxies to actually ever see this. And even so, all kinds of complicated things happen between the Big Bang and now. But it turns out that the model actually works. And we'd actually do have surveys of millions and millions of galaxies, with very well measured position. And we see this bump in the galaxy distribution. And it's not at all subtle now, with current measurements yes, really in your face.

[laughter]

0:28:56.9 MK: We're actually, you look at the distribution of galaxies on distant scales of hundreds of millions of light years, and you see the imprint of this physics that we have in our models to describe the early universe.

0:29:14.9 SC: Well, and it's important to emphasize that with both the temperature fluctuations in the microwave background, and with the baryon acoustic oscillations, you are measuring, in principle, all these parameters. Like the density of dark matter, the density of the cosmological constant, et cetera. But you're not measuring them directly. You're saying that, I have a model for everything at once, and I'm gonna make all sorts of predictions, and those predictions will depend on the parameters. I'm gonna measure some things and then ask which values of the parameters give me the best fit.

0:29:44.8 MK: That is correct.

0:29:46.4 SC: So on the one hand, very, very impressive that the standard model, works so well. On the other hand, there's a lot of moving parts. If something doesn't start fitting, then it's not gonna be perfectly obvious where to look.

0:30:00.3 MK: Yes, that is correct. So another thing I think that's been surprising about our understanding, our evolution of the understanding of the universe, is that the universe has turned out, for reasons we probably don't fully understand, to be a much simpler system than anyone might have surmised. So you're right. We measure the distribution of bazillions of galaxies, we measure the distribution of millions, the temperature of the cosmic microwave background over millions of points in the sky. It's a really complicated system because galaxies have gas, they have stars, they have interaction between outflows from stars. There are supernovae that blow up and then pollute the intergalactic medium with heavier elements. There's gravity, there's this cosmological constant, there's dark matter. It turns out to be a very, very complicated, seemingly very, very complicated system. But it turns out that it's much... When we look at it carefully, the system as a whole turns out to be much simpler than anyone might have. Our understanding of the origin and evolution of the universe is actually, I think on much more precisely parameterized and specified by the model, than is our understanding of a solar system.

0:31:22.4 SC: Yeah.

0:31:23.0 MK: Even though, we live in the solar system, have visited parts of it, and it's much simpler physics in principle, gravity and Kepler's law. So it turns out that you're right. We have all this data, we have all these parameters to turn. It seems like it would be a really complicated system. It would be hard to have any confidence, have much confidence in any individual parameter, given that you're trying to simultaneously fit for these other parameters. But it turns out that a model with five parameters, can account for it all. And that's been one of the things that's been so surprising and impressive. That one model can explain what we'll see in the galaxy distribution and the cosmic microwave background. And that's why we were so proud of ourselves.

0:32:15.9 SC: We should be proud, but we should also be interested to see if there's anything weird going on. In terms of weird things that could go on, there's theorists favorite ideas. Then there's what the experimenters actually come back and tell us about. What is it, very quickly I think, what are some of the main alternatives in terms of perhaps the physics of dark matter and dark energy, that we're trying to test when we do the cosmic experiments?

0:32:44.9 MK: Oh, okay. That's a good question. So one thing that we do to test the cosmological constant. So as we said, the cosmological constant is a very, very strange thing from the point of view of our understanding of fundamental physics. And so, one thing that you can wonder is whether the cosmological constant is really constant. So it sort of says that there's some mysterious energy pervading all of space. But you can ask, is that changing with time, as the universe expands, or is it really, really constant? And so, if your cosmological constant is not really constant, and people have been using the word dark energy in place of cosmological constant, because if the cosmological constant isn't constant then, not a good name for it. So you can ask, whether the dark energy density is constant in time or evolving in time. There's been a major effort over the past 25 years to try to address this question, try to figure out whether the energy density is changing with time or not. And that is sort of done with the same types of measurements that we use to determine the expansion rate and to determine dark matter density, et cetera.

0:34:03.4 MK: We have models for how the galaxy distribution, the cosmic microwave background should look. Those models have incorporated into them as one ingredient, dark energy. In the simplest model, is the dark energy, the only parameter that we use to describe the dark energy, it's density, which we assume to be constant. But you can also see what happens if you have a model where the dark energy density evolves with time. And so, we have parameters now that we could measure for fit from the model, fit from the data with the model, to figure out or see if the dark energy density is evolving with time.

0:34:40.3 SC: With dark matter, it seems to harder to have sort of physically plausible modifications, but people still do play around with it?

0:34:48.8 MK: Yeah, that's actually in some ways a bigger industry. We have no idea what dark matter is. The models work very well, if we make the simplest assumption that dark matter interacts with itself and with everything else only gravitationally. So in other words, dark matter particles don't scatter from themselves. They don't scatter from the ordinary stuff. But that's an assumption. And again, you can construct more complicated models where dark matter has some type of interaction with itself or some type of interaction with ordinary matter. And then you can describe those interactions in terms of parameters that you can then try to fit from the data, from the galaxy distribution cosmic graph background. But with dark matter, it's a little. We've got a few more possibilities. The dark matter is not only out there in the universe, it's presumably also in the Milky Way and in the solar system and presumably passing through us here on Earth every single day. So one of the prevailing ideas for dark matter is that it's an elementary particle that has a mass of roughly 100 times the proton. And it turns out that the dark matter density locally is roughly half a proton mass per cc. And so, what that means, is that every time you buy a liter of milk at the store, in addition to your recommended daily allowance of calcium and vitamin D, you are also getting one dark matter particle.

[laughter]

0:36:26.4 SC: If it's axions, you're getting a lot of dark matter particles?

0:36:29.3 MK: Yeah, it could be, yeah. The question is, are you buying it by weight or by... So as I said, the canonical idea for dark matter is that it interacts with nothing else except gravitationally, so you don't have to worry about it if it winds up in your milk. But if it does have some very weak interaction with ordinary matter, then we can construct laboratory detectors to try to see the effects of interactions of these very rare dark matter particles with ordinary matter.

0:37:04.1 SC: That's a fairly big industry, and so far, we have seen zero. [laughter] But just to emphasize, there's plenty of room for very, very sensible, viable dark matter candidates that we would not have seen yet.

0:37:19.7 MK: Yes, that is correct. Well, it's actually, it's interesting that, I think the people first started to think about elementary particle, dark matter seriously about 40, 45 years ago. So in the late 1980s, people started to get serious about actually looking for these dark matter particles in the lab. And so, we've been looking for dark matter particles in the lab for 40 years. And during that time, 40 years ago, we had predictions, or very, very elegant, attractive predictions for what the dark matter should be. And many of those models, have been ruled out because we haven't seen them. So in some sense, we don't know what dark matter is, in some sense a shot in the dark, but we have actually had over the past few decades, a number of really intriguing and interesting and promising theoretical models for dark matter. And it's interesting that we've been able to rule those out. It's dark matter. This is an experimental science. We're not just casting, flailing about completely in the dark.

0:38:39.9 SC: That's nice to hear. But it brings us back into the fact that we do have puzzles that we need to deal with. I guess chronologically, the first puzzle that I personally took seriously that is still lingering, is the Hubble tension. In fact, we had our mutual colleague Adam Riess on the podcast talking about it a couple years ago. So update from a couple years ago. Is it still there? Are we still worried about the Hubble tension? What is it?

0:39:06.3 MK: Hubble tension is big problem. So discussed, we have these models, we fit for a bunch of parameters to try to explain the measurements in the cosmic microwave background and galaxy surveys. And one of the parameters is the Hubble constant, which is the rate at which galaxies are moving apart from each other. So essentially measures the speeds at which galaxies are moving apart from each other. And so, the galaxy distribution, the cosmic microwave background, we don't actually see the universe expanding, but this expansion rate is a parameter in the models that we describe the distributions of the galaxy in the cosmic microwave background. But alternatively, you can try to measure the Hubble constant directly, just like Hubble did 100 years ago. So you can look at some galaxies that are not too far away, and you see those galaxies moving away from us, from our galaxy. And if you can also figure out the distance, then that gives you the Hubble constant. So the Hubble constant is the ratio of the velocity at which galaxies are moving away from us to their distance. In principle, it's straightforward. So measuring the velocity at which galaxies are moving away from each other, from us is actually fairly easy.

0:40:27.3 MK: And the reason is, the galaxies emit light. And some of that light is either absorbed or emitted by various atomic transitions. And so there are spectral lines. There are lines in the spectrum, the frequency spectrum of the light that we see. And if the galaxy that's emitting this light is moving away from us, then those lines are Doppler shifted to different treaties or longer wavelength. So the same effect is when an ambulance is moving away from you, it sounds lower pitched than it does when it's towards you. So we measure these Doppler shifts. We can figure out the velocities very, very well. The distances are surprisingly difficult. And the reason is that, when we look at a galaxy on the sky, it has some angular size. And if we knew what the physical size was, then we could infer the distance. Or if we see a galaxy, it has some brightness that we can measure very, very precisely. And if we knew exactly how luminous the galaxy was, then we could figure out exactly how far away it was. If I give you a standard flashlight and you shine it at me, I can figure out how far away you are, because I know how bright the flashlight is, and then I can, how luminous the flashlight is, and I can measure how bright it is.

0:41:47.1 MK: But galaxies don't all have the same luminosity. And so, we can't infer the distance just by looking at the luminosity. It turns out though, that there are things called supernovae. And in fact, a very specific type of supernova type Ia. A type Ia supernova, is a white dwarf, an exploding white dwarf. So what happens is when a star uses up all of its nuclear fuel, it evolves to a state where it's a gravitationally bound star in which there's no nuclear fuel being burned. And the star is held up from gravitational collapse by quantum pressure, quantum electron degeneracy pressure. But there's a limit as to how massive such a star could be before the gravitational forces overcome, this electron degeneracy pressure. And so, if I have a white dwarf that's in a binary with some other star and it's accreting matter from that other star, as soon as that white dwarf exceeds this limit, which is called the Chandrasekhar limit, it explodes. And since that happens at a very specific type of mass, we believe that all of these have, all these supernovae are exactly the same. So there's a good theoretical reason to believe all type Ia supernovae have the same luminosity.

0:43:11.6 MK: And it's also been measured empirically. You can look at a bunch of supernovae the same galaxy, and they do have the same brightness. And so, these supernovae are what we call standard candles. They're objects that have a very well determined luminosity. And so, if we observe how bright they are, we can actually figure out the distance to the supernova and therefore to the galaxy that hosts it. So there has been a project that's been going on for 15-ish years called the SH0ES Collaboration, S-H-0-E-S. And there's also been another collaboration called the Carnegie-Chicago Hubble Project, CCHP. And they've been measuring the Hubble constant in this way. They've been looking at supernovae and distant galaxies and measuring the brightnesses of the supernovae and the velocities at which the galaxies are moving. And when they do these measurements, especially the SH0ES Collaboration, they find that the expansion rate is about 10% larger than the expansion rate you infer from the cosmic microwave background in galaxy search and our models to account for them. And we call it the Hubble tension because when this was first noticed, it was a discrepancy, but it wasn't clear whether it was statistically significant or not.

0:44:31.4 MK: It also wasn't clear whether it was maybe some misunderstanding of how supernovae work or how the observations work, or perhaps some problem with the calibration of the distances and brightnesses. But things have evolved with time. The measurements have become better, there have been more of them, and the error bars have shrunk. And many of the systematic effects that people were concerned about 10 years ago have been shown to be of no concern or not the source of the discrepancy. And if anything, the Hubble tension became much more serious just a few years ago with the launch of JWST.

0:45:10.4 MK: So until I would say, three years ago, it was reasonable to be skeptical about the brightnesses of the supernovae. And the reason is that the supernovae brightnesses are calibrated to things called Cepheid variables, which are stars, variable stars. They're stars that become brighter and dimmer over time scales of weeks to months. And Cepheid variables are also standard Kant. And so, we have Cepheid variables in the nearby in the Milky Way, and we can measure their distances very well. And then there's some Cepheid variables in nearby galaxy, that also host supernovae. There are about 40 such galaxies that host Cepheid variables that we've observed very well and supernovae.

0:46:00.3 MK: So the supernova distances are calibrated to the Cepheid variables. And the Cepheid variables then, are what we're calibrating against. So when you look at a Cepheid variable in one of these nearby galaxies, if you look at it with the Hubble Space Telescope, which was the best instrument that we had for doing these measurements, the angular resolution of the Hubble Space Telescope is great, but not perfect. And in many cases, when you were looking at one of these Cepheid variables, it'd be light from nearby stars that would sort of spill over onto the Cepheid variable limit. So this was something that you might be reasonably concerned about. Can we actually separate the light from the Cepheid variable, from the light from nearby stars well enough to actually tell how bright that Cepheid variable is? But now, we have the James Webb Space Telescope that launched what, three years ago now. And the James Webb Space Telescope has much better angular resolution than the Hubble Space Telescope does. And so you can go back, and look at some of the Cepheid variables and do the measurement with this better telescope. And when you do that, there's no issue of crowding.

0:47:09.9 MK: The Cepheid variables are very, very well separated from the nearby stars, and you can make the measurement. And for the 16 such supernovae, or sorry, 16 such Cepheid hosts for which they've done this measurement, are the sample of 45-ish HST set hosts. The measurements are spot on from what was inferred from HST. So this crowding issue is no longer a concern. And so, the Hubble tension is more serious now than it was three years ago because of this.

0:47:43.8 SC: Okay. So unless there's some huge mistake that we are really missing, and the experiment has obviously been very good, there's a mismatch between the sort of direct local measurements of the expansion rate, and the preferred expansion rate we need to fit the data of the CMB and the wire models. So at a very broad strokes, without getting into model building or anything like that. But, how would we solve this? Could we just do. What are we looking for? Are we looking for something that makes the universe slow down at later times or speeds it up at later times, or what is the target here?

0:48:22.4 MK: So the target here is the baryon acoustic oscillation in the cosmic microwave background. So we talked about how you see this peak in the correlation function for galaxies at 100 megaparsecs, but you also see this peak in the cosmic microwave background. But you don't measure the physical size of the peak, you measure the angular size. Same thing with a pond. If you're viewing a pond from some distance and you see the ripples going out, you see the waves going out from where you throw the pebble, you can't tell how big those waves are unless you know how far away you are viewing and viewing them from. So we measured the angular size of this sound horizon variant acoustic oscillation very, very precisely, one part in the 10,000 with cosmic microwave background measurements. And the angular size is the physical size divided by the distance to the cosmic microwave background. That angular size is determined from cosmological models, that have as their parameters the Hubble constant, the dark matter density and the baryon density.

0:49:33.0 SC: And the dark energy density?

0:49:36.0 MK: And the dark energy density. And how it evolves with, whether it evolves with time. So the two solutions that people have, can be sort of classified into late time solutions and early time solutions. The late time solutions, we sort of change that angle, the model predictions for that angle, by changing the distance to the surface of last scatter. And that would happen if we sum, if the expansion rate in the recent past was somehow different than in the standard cosmological models. Those tend to not work because we can also get a measurement of the Hubble constant, from the a baryon acoustic oscillation in the galaxy distribution. And that sort of agrees with what we get from the cosmic microwave background. So these late time solutions people thought about early on, but they don't really work. So the other possibility, are early time solutions where we somehow change the physical size of the sound horizon in the early universe and ways to do that. And in fact, we have to decrease the physical size of the horizon in the universe to account for the Hubble tension.

0:50:41.2 MK: And so, one idea that people have spent a lot of time thinking about is called early dark energy, where we have something that's sort of like a cosmological constant, but has a much larger magnitude and it's around only in the early universe for the first half million years of the universe and then somehow decays away. And people spend a lot of time thinking about these from about 2018 until now. And I'd say initially, we thought it was a promising idea.

0:51:16.1 SC: You were one of these people, by the way?

0:51:17.8 MK: I was one of these people.

[laughter]

0:51:20.0 MK: In 2021, there was this very interesting result from one of the cosmic microwave background collaborations, the ACT Atacama Cosmology Telescope Collaboration. They published a paper in late 2021, where they said that the new measurements were actually more consistent with the early dark energy model than the standard Lambda-CDM. And that was very exciting. Got all excited. And that got people even more revved up about early dark energy models. There was more, a bigger another round of model building, but then also more scrutiny from experiments. And I think that was four years ago, three and a half years ago. And over the past three and a half years we've had more measurements from the cosmic microwave background, from galaxy surveys, more data, more scrutiny of the data. And I think the pendulum is swinging back to Lambda-CDM away from early dark energy.

0:52:13.7 SC: So it sounds like there's two options. Late time solutions and early time solutions. And neither one of them work.

0:52:20.0 MK: That is, I would say, fair summary of the situation.

[laughter]

0:52:24.1 SC: So it's very different, just again for the sort of non-experts here. In 1998 when people claimed, when the two teams claimed that the universe is accelerating and that was an anomaly, et cetera, but there was instantly a theory that explained it, and everyone could say, "Oh okay, so we found this thing?" Here we have an anomaly and my impression is, it's still not clear what could explain it.

0:52:46.3 MK: That is correct. Early dark energy was a really, a very plausible idea until just a few years ago. And I would not say that it's ruled out, because early dark energy is not a model. It's sort of a class of models or an idea that can go into models. But what's happened is, with new data that is increasingly more consistent with Lambda-CDM, the wiggle room for constructing early dark energy models has decreased and it becomes harder and harder to find some model that actually works. But your summary. Yes, is probably the first approximation correct. The simplest late time models don't work. The simplest early time models don't work. But it's sort of a... Yeah, I think if you had Adam on your podcast, what Adam would say if he were here, is that the evidence for the Hubble tension is now much stronger than the evidence they had for accelerated expansion in the late '90s. But people are much more reluctant to accept this because we don't have a model to explain it. So interestingly enough, if it was the other way around, if the cosmic microwave background was giving us a Hubble constant that was 10% bigger than that from supernovae, then we would just change our dark energy model.

0:54:15.8 MK: So we'd say it was time evolving dark energy. Given that the Hubble constant from supernovae is larger than that infer from CMB. The same solution could also be tried, but it would require a dark energy density that increases with time rather than decreases. And decreasing with time is okay because the energy can go somewhere else. But in order to have a dark energy density that increases with time, is sort of equivalent to having energy just appear out of the vacuum, which is not something that we like. I think you probably know this better than I, in the general relativity community, it violates the strong energy.

0:55:00.9 SC: It violates the weak energy.

0:55:01.5 MK: There we go. See?

0:55:03.6 SC: Yeah. I think I learned this from you. But I did like apparently I didn't learn it from you.

[laughter]

0:55:09.8 SC: I'm forgetting myself. But it violates. But your point is right, it's just sort of so magical and scary to have energy appear out of nothing.

0:55:18.3 MK: Yeah, I think that's real. Weak energy condition is general relativity for creating energy out of the vacuum, which makes us physicists uncomfortable. But the cosmological constant also made us uncomfortable.

0:55:34.0 SC: Yeah, to be fair, it's more than uncomfortable. When you try to construct a model in which it happens, other things tend to go disastrously. Our discomfort is not purely emotional and vibes-based. [laughter]

0:55:42.0 MK: Yeah.

0:55:44.5 SC: Okay. I guess, remembering now and I know that we're running long so let me know if I'm abusing your kindness here. But there's another tension even before we get to the variable dark energy stuff. There's the S8 tension that cosmologists worry about and is not sunk into the popular imagination yet. Should we worry about that?

0:56:06.0 MK: Yeah. The S8 tension is a little more subtle and I sort of have less confidence in it. And whether it's a tension or not seems to bounce around a lot more depending on who you ask, and which data set. And one of the conclusions from the new results that we've seen from the DESI collaboration and the Dark Energy Survey collaboration just last week, is that with new data, the S8 tension is going away. So the S8 tension is a discrepancy between the amplitude of fluctuations in galaxy surveys on small distance scales, compared with that expected from the cosmic microwave, the models that best fit the cosmic microwave background. And it's a strange tension because it sort of depends on which data set you look at and in some cases, how you analyze the data set. And there's some measurements that seem to indicate that there's a discrepancy, but then other measurements that indicate that it's not a discrepancy. And it also involves measurements or observations of the galaxy distribution on smaller scales, where the theory becomes more complicated and we have less confidence. So a lot of people in cosmology worry about the S8 tension. There's probably less consensus about whether it's there or not, and I think it might be going up.

0:57:33.1 SC: Well, that's another reason why the Hubble tension hasn't quite been as completely accepted as the accelerating universe. Because when you find a tension like that, maybe you find other tensions or other signals that kind of go in the same direction, but rather than building up, we're having other things sort of come and go and waffle around and nothing quite definitive yet.

0:57:55.5 MK: Yeah, it's a perfectly reasonable view. And I think 10 years ago, most people would think that as we analyze more CMB data, as we understand the analyses better, and as we understand the supernova and the analyses better, we'll find small errors in one or both that sort of have accrued and together give you some consistency. But that has not happened.

0:58:29.3 SC: Not yet happened, yeah. Okay. You already referred to two new results which have sadly, very similar names. DES and DESI. And they're both attempts to measure dark energy, and they're both hinting that it is not the cosmological constant. So if true, I'm very... I have sort of public statements that I'm skeptical that something like that would come to pass, but it would be a big deal if it were true.

0:58:58.6 MK: I agree. I am also skeptical. I think sometimes when I'm skeptical, I have to try to dial it back.

0:59:09.5 SC: [laughter] Yeah.

0:59:10.5 MK: Because, I think often what you find when you study history of science, is that discoveries are made not just by, major fundamental discoveries are made not just by people who've made really good measurements and are really good at the analysis, but they're people who have an open mind and will be accepting of the possibility that this might be big.

0:59:36.5 SC: We're not as young as we used to be.

0:59:38.0 MK: Yeah. If you're... I think most of us have this attitude that when there's something strange in the data, there must have been something that goes wrong, went wrong, or we missed it. The Hubble Tension. I think many people have that attitude. They're obviously missing something in the supernovae, complicated systems, they haven't modeled it correctly. The standard cosmological model is fine. And if you always have that attitude, you'll never discover anything.

1:00:01.3 SC: Yeah. So what is the new result?

1:00:04.7 MK: So the new result is coming from DESI and then there's sort of consistent information coming from DES and from various supernova measurements, that show that the dark energy density evolving with time. And in particular, they show that it is or has in the recent past been increasing with time. So it's a complicated result. It's not hugely statistically significant, but a little more statistically significant than it was a year ago. But it suggests that the dark energy density was smaller early times, became larger with time, and then fairly recently started to decrease in energy again. So it's a very unusual result. So I would say it's unusual in several ways. The first is that, if the dark energy density was evolving in time, that is instant Nobel Prize. And we've been looking for this. So we shouldn't say, can't be right. But as we just discussed earlier, the preferred fit suggests that the dark energy density was increasing with time, which I just learned violates the Weak Energy Condition, which I already knew is creating energy out of a vacuum, which is I'm supposed to keep an open mind to. But it's really very, very, very strange. The point of view of theoretical.

1:01:36.3 SC: One has priors, that's okay. And your priors are never zero, but they're bigger on some possibilities than others.

1:01:44.2 MK: Yeah. Another way of saying, it's like sort of higher order and discovery space. Learning that the dark energy density evolved in time is spectacular enough. But then learning that it increases with time, that's even beyond that. So I think the bar for that is even higher than it would be just for a dark evolution with.

1:02:06.8 SC: What are these experiments? What are they measuring?

1:02:10.3 MK: So the principal experiment for this is the DESI collaboration, which stands for Dark Energy Spectroscopic Instrument I think.

1:02:18.0 SC: The DESI. That's right, I looked it up.

1:02:22.7 MK: So this is a really spectacular project where they measure the redshifts and therefore the distances to millions of galaxies over a huge volume of the universe. And with these measurements, they can measure the baryon acoustic oscillation feature at a variety of different redshift or distance spin. So they can measure the angular size of this bump in the clustering, the galaxy clustering. They can measure the angular size at a variety of different distances. And in that way, they can figure out the expansion rate as a function of redshift or as a function of time. And so, they can actually see the expansion rate changing with time and distance. So, some of the issues are that they're splitting all of their galaxy survey into a bunch of different distance spins. And so, they have bazillions of galaxies, but they have six or seven different distances. And so, on each spin it's a bazillion divided by six or seven. And then, the other things that you might be concerned about is that, the universe is actually evolving with time. And maybe there's something about the properties of the galaxies that they're looking at, that are evolving with time.

1:03:45.6 MK: And you can read the papers, they've got hundreds of pages. They spent a huge amount of effort checking for all of these obvious things that you would check for. And none of these obvious things that you would check for has shown up. But with a result like this, that's so unusual, you really require a higher level degree of scrutiny. The way I look at it... The other thing I should say is that they're also supernova measurements that are sort of like those. But they look at supernovae out to larger distances. So they're interested not so much in the expansion rate today, but how it evolves with time. So they're doing sort of complementary measurements. They're sort of doing the same thing that DESI is trying to do with the baryon acoustic oscillation, but in a slightly different way. And then there's the dark energy survey, which doesn't have distances quite as well, but they have tons and tons of galaxies. And their measurements are sort of consistent as well.

1:04:48.2 SC: Consistent with the DESI results of the time dependent dark?

1:04:52.0 MK: Yeah, they can't really determine the time evolution of the dark energy quite as well. But there are other places where their observations overlap with DESI's observation. And in places where they overlap, there's consistent. So I think the way I look at it, is that DESI has shown that these galaxy surveys can be extremely powerful. They can work. And the other thing that's important to notice that DESI is not the last such project. We've got the Rubin Observatory, that's gonna start taking data any day now. And they're gonna do analogous things over...

1:05:29.9 SC: That's a big ground-based telescope.

1:05:31.8 MK: Ground-based telescope. There's then the European Space Agency last year launched Euclid, which is a space-based. They're gonna do a space-based galaxy survey. And there's then NASA's Roman Space Telescope would also be launching soon. And that's gonna also do a huge galaxy survey from space. And all these projects have some overlap, but they also have complementarities. They check different things. They will be affected by different types of systematic artifacts. They have different ways of observing the same galaxy populations, and they also have access to slightly different galaxy populations. And also two weeks ago, NASA launched a project called SPHEREx, which is gonna have some galaxy mapping capabilities. And so, the way that I look at it, is that these projects can work. They do work. The level of precision that we're getting from them is absolutely stunning and was unimaginable just even 10 years ago. And the other thing is the DESI results are new and we've always found with new telescopes, projects in cosmology and astronomy, when you build a new telescope to make new observations, you're learning about the universe and the telescope at the same time.

[laughter]

1:06:51.1 MK: And so, I think it's gonna be really interesting, important and interesting for us to really look at the telescope and the detectors and the analysis pipelines simultaneously with our scrutiny of the cosmological implications. And I think that in the process, we'll understand better what's going on. And even if it's not time evolving dark energy, it's definitely gonna feed into our ability to do these measurements even better in the future.

1:07:23.1 SC: I'm sure some people are gonna want to know why we need to build so many different telescopes. Why can't JWST do this? But these are experiments designed for different purposes.

1:07:33.3 MK: Yeah. So JWST is an absolutely phenomenal instrument. I actually got to see it in the high bay at Lockheed Martin in December of 2019. And I was looking at and watching the videos of how it was gonna unfold and unpack, and I was thinking, there's no way it's gonna work. [laughter] Like, oh my God, it was crazy. As a...

1:07:58.0 SC: You can't fix it. You can't go up, and repair it?

1:08:02.2 MK: The fact that it worked, and actually worked better than they anticipate in many ways, it's absolutely spectacular. And the images and the things we're finding with it, absolutely amazing. But the thing is, JWST is a narrow field of view. It's really good for looking at a very small number of objects at very large distances or very faint objects. But if we're trying to do cosmology, we'll wanna map the distribution of galaxies over as large a volume as we can, so over as much of the sky as we can. So it's a different type of telescope.

1:08:33.2 SC: And one of the things about the DESI results, is it brings home at least my very, very casual looking at the papers, brings home the fact that it really does depend on your model that you think you're testing. When you come across and say, "Here's our result." 'Cause if you just fit to, there's a constant dark energy, et cetera, you get one result. If you say, "Well, I'm gonna let it vary linearly with time." You get a different result. If I'm gonna let many things happen, you get a different result. Is it possible that there's different levels of confidence in the dark energy used to be increasing result and the dark energy is somehow changing result, or do they go hand in hand?

1:09:16.9 MK: I'm still trying to understand that. So yes, what we do is we construct models and then we fit for the parameters in those models. And one of the things I'm trying to understand is that if you take the DESI results and you model them with the standard Lambda-CDM model, my understanding is that actually gives you a pretty good fit. And I don't know whether it's... If there was something that was desperately wrong with Lambda-CDM, then when you tried to fit it with Lambda-CDM, you would get a result that was not, that you would not get a good result. But my understanding is that they do get a good result with Lambda-CDM. But then, if they expand the model parameter, say, so instead of Lambda-CDM, they have this time evolving dark energy density.

1:10:07.3 SC: Yeah.

1:10:07.6 MK: So this is a model that now has two additional parameters. It has the time evolution, and then they have a second parameter which is the time evolution of the time evolution. And when you do that, that model seems to provide a better fit than Lambda-CDM. But I have not yet really understood whether it implies that Lambda-CDM is not a good.

1:10:27.6 SC: Right.

1:10:28.1 MK: And my understanding is also that if they just try to fit a model where you have dark energy, that is evolving with time, but has just one parameter model, it's just evolving with time. You don't have any evolving of the evolving. That that also gives you a fit that is consistent with the standard Lambda-CDM.

1:10:48.9 SC: Okay. Is there any relationship between this result and the Hubble tension?

1:10:56.3 MK: Yeah, that is a great question. As I said earlier, we're always looking for consistency. There is no obvious way in which this connects with a Hubble tension.

1:11:08.1 SC: Okay.

1:11:08.8 MK: I think it would be much more exciting if there was.

1:11:12.4 SC: Yeah.

1:11:14.0 MK: But it turns out that when you change or expand the model parameter space this particular way, it does not change the Hubble constants inferred from the measurement. It does change the upper limit to the neutrino mass, which is sort of something that has been, I think one of the most exciting things from these results that people have not been paying a whole lot of attention to.

1:11:36.3 SC: What is that? Where does that come from? So what does a neutrino have to do with this? You haven't even mentioned neutrinos yet?

1:11:41.5 MK: I haven't even mentioned neutrino. Yeah. No one's been mentioning it. So in the standard model of elementary particle physics, there are three different types of neutrino, electron, muon and tau neutrino. And in the standard model, when it was constructed in the early 1970s, the neutrinos were thought to be massless. And in the standard model, they are massless. They have don't weigh anything. But then, about 20 something years ago, it was discovered, that neutrinos actually have a small non-zero mass. And so, we know now the neutrino masses are not zero, but we don't know what they are. They're very, very small. So we know that they're bigger than zero, but they're smaller than some upper limit. Those upper limits come from a variety of accelerator experiments and laboratory experiments and beta decay experiment. But it turns out that neutrinos, the standard cosmological model predicts that there should be neutrinos running around the universe just like there's light and baryons. And if the neutrinos have a mass, then they would actually contribute something to the cosmological energy density and affect our cosmological models. And the measurements that we have now in cosmology are so precise that the fact that neutrino masses are non-zero, actually has to be taken into account.

1:13:03.5 MK: And in fact, with our cosmological measurements, we now have upper limits to neutrino masses which are complementary and in some ways better than those that we have from laboratory experiments. And one of the things that was really interesting about DESI, is that it improves the sensitivity to a non-zero neutrino mass over what we had before.

1:13:26.0 SC: So to be clear, are we saying that they have detected, if we didn't know that neutrinos had mass, would this tell us that they did?

1:13:33.7 MK: No, but they have... So what they have are upper limits that are starting to distinguish between the two different neutrino mass scenarios. So there are three neutrino masses. We've got good reason to believe that all three of them are non-zero. We know that two of the masses have some small mass splitting, and then another pair of masses has a larger mass splitting, but we don't know whether how those masses are assigned to the electron muon or tau. And so, we don't know whether there's two lighter states and one heavier state or two heavier states and one lighter state. And the DESI results are starting to say that the inverted hierarchy, the system with two heavier masses is ruled out. And this is kind of super useful, kind of gone under the radar in terms of popular press coverage of the DESI results, but it's a really, really impressive accomplishment and could be very important for elementary particle physics.

1:14:39.0 SC: So it's a tradition. Late in the podcast, we always get to let our hair down and explore wilder ideas. So you've already said that the straightforward fitting the data, implies this dark energy increasing for a while in density and then decreasing. That's already a very, very wild idea. Are there wilder ideas that could fit the data? So I'll just tell the audience, back in the day, when we were younger than we are now, you were involved in a couple of papers establishing the idea of the big grip, as a possible future for the universe. With the dark energy density just going crazy upward in the future. And friends of mine and I, wrote papers saying, first, so the technical language we use for increasing energy density is W less than minus 1. W is a little Paramus. And if W is less than minus one for dark energy, then the density goes up. So I wrote a paper saying, "Can W be less than minus one? And we argued probably not." But then we wrote a follow up paper saying, "Could you be tricked into thinking that W is less than minus 1. If gravity were changing it's strength over cosmological time?" And we said, maybe, but it doesn't look very pretty. Are people exploring ideas like that, gravity changing it's strength?

1:16:01.3 MK: I don't know in the sense that I haven't seen much. So in the dark energy literature, they're sort of like the early simplest type dark energy models, which we called, I think you called them quintessence?

1:16:16.2 SC: No, that was called...

1:16:18.1 MK: All right, you had quintessence in the rest of the whole... I did have that. I helped popularize.

1:16:21.1 SC: Yeah.

1:16:23.7 MK: So those quintessence. But then there was sort of a wave of alternative gravity models for dark energy. I have not seen many of these alternative gravity models showing up in connection with new DESI result. But I don't know if I haven't seen them 'cause they're not there. I just haven't noticed.

1:16:41.9 SC: They haven't had that long. We like to think that people take more than a week to write a good paper. [laughter]

1:16:45.7 MK: Well, the DESI results were also around last year.

1:16:48.0 SC: That's true.

1:16:49.8 MK: So I don't know. The crazy idea that I like to think about, is oscillating dark energy. So there's sort of... We know that there's a cosmological constant now, we have very good reason to believe that there was a period that we call an inflation, in the very early universe which was powered by a non-zero cosmological constant with a very large magnitude that then decayed away. And the early dark energy models also surmise that there's a period of cosmological constant domination, half a million years after the Big Bang that then dies away. And so, people have, over the years considered the possibility that every few logarithmic times in the history of the universe, for some reason, there's a cosmological constant shows up for a little while, and then disappears again.

1:17:47.3 SC: A cascade of dark energies at different times?

1:17:49.9 MK: Yes. So there were papers where you would just have essentially just one quintessence model, but it had, instead of rolling down a smooth hill, it rolled down a bumpy hill that would do that. And then there was also this idea called the string Axiverse, which was quite popular about 15 years ago. And the basic idea there is that, in string theory, there are in addition to the fundamental fields responsible for electron and muon and quarks and photons, there are many, many, many more fundamental fields. And there could be hundreds of them, that they call axion field. And it is conceivable that in these scenarios, you could have different axion fields sort of randomly becoming dynamically important at different periods of the universe. So that's the thing I kind of like to entertain. And it kind of fits in to some extent with these DESI results because these scenarios suggest there should be some type of cosmological constant domination that then decays away. In these scenarios, whatever we think is a cosmological constant now, will then decay in the near future, several billion years from now, and the universe will then proceed as if there's no cosmological constant once again.

1:19:05.0 MK: And so, with DESI, you look at the, I told you expand, the dark energy density is increasing, but now it's decreasing in time. So it kind of fits in with that scenario. The only thing that doesn't fit in, is the increase density, which we can't fit or has not yet been explained well.

1:19:21.6 SC: But the idea, this is an important one that we should mention. String theorists never liked the idea of a positive cosmological constant that's hard to fit into string theory, but zero or negative, they could make their peace with. And if the dark energy is evolving and if it's decreasing right now, that is back on the table, we could have a big crunch in the future. We could have a negative vacuum energy at the end of the day.

1:19:48.1 MK: Yep, that is definitely on the table. I don't think though, I don't think that those allow for the increased intensity.

1:19:55.6 SC: No one, no sensible person allows for that, which is obviously this motivates people to really get that right.

1:20:03.4 MK: Yeah.

1:20:04.5 SC: And the final thing then, I will give you a chance to wax eloquent on birefringence. 'Cause there's been a couple of hints that maybe there is something funny going on with the polarization of light from the CMB. That's the last anomaly that I'll lay in front of you.

1:20:24.8 MK: Okay. So the cosmic birefringence. I learned about from a 1998 paper by Sean Carroll and collaborators, is that right?

1:20:34.5 SC: Yeah, maybe.

1:20:35.5 MK: So your listeners should know that you wrote this spectacular paper in the late 1990s, where you pointed out that, there may be some physical models in which light that has a right circular polarization could travel at a slightly different velocity than light with a left circular polarization. And if so, a light wave that was linearly polarized would have a linear polarization that rotated with time as it propagated. And that is called cosmic birefringence. And I wrote a paper a few years after that or the year after that, that showed how you could test the scenario by looking at the cosmic microwave background. The cosmic microwave background, we're looking at light that's been propagating for 14 billion years. And so, if there's any subtle effect, it would have more time to accrue in the cosmic wave background than anything else. And people have been trying to make these measurements with cosmic microwave background experiments since then. And there has been some hints in the data, that the rotation, that the linear polarization actually does get rotated by 0.3 degrees over 14 billion years. I think it's very exciting, very interesting. It's a very, very difficult thing to measure from the data though.

1:21:55.3 MK: And the primary reason, is that it's hard to calibrate the linear polarization. So they can measure differences in linear polarization very well. So if I give you two rays of light that are side by side, and ask you what's the difference in the linear polarization? You can measure that very well, but the absolute linear polarization is harder to get.

1:22:19.2 SC: And that's once again because telescopes are complicated things.

1:22:22.1 MK: Yes, because telescopes are complex. It's not, it's a fairly... I don't know, it's just easier to measure the separation often between two points that are nearby, than it is to measure the separation of two points that are really far away.

1:22:35.7 SC: Okay, that's fair enough.

1:22:37.1 MK: So it's the same thing with polarization.

1:22:39.1 SC: But did you notice that act? The Atacama Cosmology Telescope also has a tiny little detection of birefringence.

1:22:47.7 MK: No, I had not noticed that yet.

1:22:49.8 SC: So the result that you're talking about was from Planck. You had this beautiful all sky thing and it's marginally statistically significant. And like you say, it's very difficult, so people didn't get too excited. But the Atacama Cosmology Telescope, which is a ground-based thing, they had a recent data release where they said, everything fits Lambda-CDM perfectly well, but there is a two point something sigma detection of birefringence.

1:23:16.1 MK: So I hadn't looked at those 'cause I was studying the DESI papers really, really carefully in preparation for this podcast. I haven't had a chance to dig down deep in those papers yet.

1:23:27.3 SC: Well yeah, I don't know.

1:23:29.4 MK: That's interesting.

1:23:30.4 SC: It's interesting. Yes.

1:23:32.2 MK: I'll have to take a look.

1:23:34.1 SC: What are your feelings? What are your... This is where we close up. Your final thoughts 20 years from now, what do you think we'll have landed on? Most probably.

1:23:43.8 MK: Most probably. I think 20 years from now, we'll know much more, with much more certainty whether the Hubble tension is real. I'm guessing that 20 years from now, there'll be new ideas from elementary particle theory and theoretical physics, that make a phantom energy much more palatable.

1:24:14.9 SC: Phantom energy is the increasing density.

1:24:16.9 MK: Yes, WS. Increasing energy density more palatable to us? And I'm guessing that we will see dark energy evolution.

1:24:27.1 SC: Okay. All right.

1:24:28.5 MK: Bold.

1:24:29.4 SC: 'Cause 20 years, I picked that because we might both still be around.

[laughter]

1:24:33.5 MK: Yeah.

1:24:35.2 SC: 50 years, it's easy to make crazy predictions.

1:24:37.3 MK: Well, not only will we still be around, but this podcast will still be around. You'll be able to turn it on and say, look.

1:24:42.2 SC: Much less likely. It might have to have a revivification if that's the case. [laughter] All right. Well, that's a lot to think about. It's kind of good because, for a while there, it was possible to believe that cosmologists had figured it all out. That we had a theory that fit the data too well. But now we're in a more normal science area, where there are anomalies and we got to bang our head against them. It feels good.

1:25:09.3 MK: Yep. Yeah. I think the frustrating thing is, as you know, I spent a lot of time working on early dark energy, and then I go and give talks. I was invited to give talks all the time. And at the end, people say, "Well, what is it? Is it early dark energy?" I'll let them see then I said, well, the measurements will be done in the next few years, and if it's really dark energy, we'll know. But now, so what's going on? [laughter] I don't know what to say. I have no idea what's going on.

1:25:37.1 SC: It's good for the young people out there. There's still a room for a really good idea.

1:25:41.0 MK: Yep.

1:25:41.9 SC: All right, get to work. All right. Marc Kamionkowski, thanks so much for being on the Mindscape podcast.

1:25:46.3 MK: Thank you very much for inviting me. It's been an honor and a joy to be here.

[music]