There is so much we don’t know about our universe. But our curiosity about the unknown shouldn’t blind us to the incredible progress we have made in cosmology over the last century. We know the universe is big, expanding, and accelerating. Modern cosmologists are using unprecedentedly precise datasets to uncover more details about the evolution and structure of galaxies and the distribution and nature of dark matter. Priya Natarajan is a cosmologist working at the interface of data, theory, and simulation. We talk about the state of modern cosmology, and how tools like gravitational lensing are providing us with detailed views of what’s happening in the distant universe.
Support Mindscape on Patreon.
Priya Natarajan received her Ph.D. in astrophysics from the University of Cambridge. She is currently professor of astronomy at Yale University, the Sophie and Tycho Brahe Professor at the Niels Bohr Institute of the University of Copenhagen, and an honorary professor for life at the University of Delhi, India. She is an Affiliate at the Black Hole Initiative at Harvard University and an Associate Member of the Center for Computational Astrophysics at the Flatiron Institute in New York. She is a frequent contributor to the New York Review of Books and other publications. Among her awards are a Guggenheim Fellowship, the India Abroad Foundation’s “Face of the Future” Award, and an India Empire NRI award for Achievement in the Sciences. She is the author of Mapping the Heavens: The Radical Scientific Ideas That Reveal the Cosmos.
- Web site
- Yale web page
- Google Scholar publications
- Articles at the New York Review of Books
- Wikipedia
[accordion clicktoclose=”true”][accordion-item tag=”p” state=closed title=”Click to Show Episode Transcript”]Click above to close.
0:00:00.0 Sean Carroll: Hello, everyone, and welcome to the Mindscape podcast. I’m your host, Sean Carroll. And as I’ve often said on podcasts and elsewhere, cosmology is a great science because the universe is so simple. It has ingredients in it, right? Our universe has electrons and atoms, and photons and neutrinos, dark matter, dark energy, but it’s a fairly finite list. It’s not like there’s a hundred different ingredients that really matter for the evolution of the universe, and even better, when you look on sufficiently large scales, the universe looks smooth. If you kind of squint and average out over a scale of, I don’t know, 100 million light years across, it’s more or less the same amount of stuff in every part of the universe, the same number of galaxies and stars and so forth, and that helps us.
0:00:46.3 SC: We can put together a picture of what the universe is doing, how old it is, etcetera. But then, of course, if you squint a little bit closer, if you zoom in on what the universe is doing, all that simplicity kind of evaporates, and you see a very rich, complicated universe. There are stars and planets, galaxies, clusters of galaxies, filaments of galaxies, and what have you. And then, the particular way that these structures both exist today and have evolved over time depends on the list of ingredients and the expansion history of universe and so forth.
0:01:21.0 SC: So the bread and butter of modern cosmology is not measuring the expansion rate or the acceleration rate. Those are important things, we do them and they matter, but most of observational cosmology and theoretical analytic cosmology is interested in the structures that are there in the universe. That’s what we’re going to be talking about today, with today’s guest is Priya Natarajan, who is an astrophysicist at Yale, and the author of a book called Mapping the Heavens: The Radical Scientific Ideas That Revealed the Cosmos, which traces both through the history of mapping the heavens, but also the modern version of it. And that’s what we’re talking about, mapping where the galaxies are, how they get distributed, and in particular, Priya is an expert on gravitational lensing as a technique to study structure in the universe.
0:02:11.3 SC: Einstein predicted a long time ago that gravitational fields will bend light and we can use the amount of bending on cosmological scales to figure out exactly how much stuff there is in the universe, whether or not we can see it. This is obviously hugely important for figuring out properties of dark matter, and more than that. So we get into some details on modern cosmology, and another thing that this is a wonderful discussion for is the fact that it really involves so many of the different moving parts of modern science. More generally, there are, of course, observers collecting data, there are theorists proposing models, but in modern cosmology, it also really matters that we have simulations, that we have large-scale computer programs that are trying to bridge the gap between the theoretical ideas and the observations of the data.
0:03:02.2 SC: And finally, not to neglect them, we have human beings. We have the fact that human beings respond differently to different scientific results and are pre-disposed in different ways to different kinds of scientific theory. Priya and I talk about that, the existence of anomalies; despite the fact we have a really, really successful cosmological model, no model is perfectly successful, so where are the gaps, where are the weak points, which might lead us to a better model in the future? This will bring you up to date on what cosmologists are thinking about right now, and what to look for as we learn more stuff in the future. Let’s go.
[music]
0:03:57.1 SC: Priya Natarajan, welcome to the Mindscape podcast.
0:04:00.8 Priya Natarajan: Delighted to be here.
0:04:02.1 SC: We’re going to talk about cosmology. And I’m going to take for granted, I’m going to assume the audience knows that the universe is big, that it’s expanding, that it’s full of galaxies that are mostly smoothly distributed, and that the interesting part comes when we say, “But it’s not perfectly smoothly distributed.” As a working cosmologist at the interface of data and theory, explain to us how you think about the large-scale structure of the universe. What are the basic facts that the person out there should have in mind while we have this conversation?
0:04:34.0 PN: I think you sort of outlined the basic picture we need to have. You can just think of the universe as expanding and with its entire contents, not just expanding, but the expansion is accelerating and that it is filled with a bizarre kind of matter, dark matter, which is nothing that we are familiar with. Nothing on the periodic table. Some likely exotic particle made in the early universe, and that much of the action in the universe starts kind of late, like relevant to us, that is the assembly of ordinary atoms happens a little late. And I’m particularly interested in this sort of aggregation and clumping of matter, sort of a departure from the simple kind of linear, homogeneous kind of distribution.
0:05:31.7 PN: Early on, we believe the universe was more or less homogenous, but that over time gravity, which is the most powerful organizing force in the universe, collects, aggregates mass and matter. Both dark and baryonic ordinary matter fuel gravity, because we believe this dark matter is actually a particle. And so, then they aggregate, they co-evolve, in fact, the dark matter is sort of a cocoon in which ordinary matter falls in. You can really think of it as like a cradle and a cocoon. And gas, basically… You know, for us, all of matter essentially in the universe is kind of hydrogen, really. We’re really talking about hydrogen. It’s all gas. Gas falls in, cools, condenses, form stars, and then, you know, stars evolve.
0:06:27.7 PN: And I think there are many intriguing mysteries that have to do with the first set of stars that formed. So, the onset. How did it all get started? We kind of know how the universe itself got started, from a very hot, dense state, but I think there are many more open questions as the universe expanded, cooled down, and then these dark matter cocoons lit up for the first time. You have the properties of the first stars.
0:07:00.3 SC: Well, actually, and that makes me… That leads me to a question, because I think that this is something that is often glossed over in the details, but we have ordinary matter. Which is, like you say, everything on the periodic table, everything we’ve ever found here on Earth. Dark matter is something new, something different. And we’ll talk about why we think that’s there in a second, but there’s more dark matter than ordinary matter, but also… Let’s just get straight what form the ordinary matter is in. You said mostly hydrogen. And so, what fraction of it is stars and planets versus what fraction of it is just out there as gas, between the stars and planet?
0:07:39.7 PN: Yeah, so much of this hydrogen is actually captured in stars and planets, and some of it, a small portion of it, is kind of smeared everywhere as material in between galaxies in space. And dark matter is distributed similarly, in the sense that there’s dark matter is there everywhere in the universe, very lightly smeared. And then there are places, pockets, where structure forms. And what I mean by structure is galaxies, the visible part sort of forms, and in those regions you have a concentration of dark matter. And light really traces dark matter. Which is why the regions where you have an aggregation of dark matter, you also tend to have an aggregation of ordinary matter.
0:08:34.7 SC: Can I ask you? Because back in my day when I was in graduate school, before we had really mapped out the cosmic microwave background radiation, which you should tell us about, but just to set the stage, we didn’t know how structure grew. There was this theory that there were tiny fluctuations in the early universe that grew over time, but there was a competing theory that there were seeds, like cosmic strings or something else, that sort of stirred up the universe all along. And I get the feeling, in fact, I know for sure, but the audience doesn’t know, that we’ve more or less discarded this cosmic string theory in favor of the primordial theory. Is that right?
0:09:11.6 PN: Yeah. Yeah, absolutely. And I think the compelling evidence for this idea of dark matter, of cold dark matter, very sluggish material particles constituting the universe. I should note here, sluggish… Everything is relative to the speed of light, right?
0:09:30.0 SC: Yeah.
0:09:31.9 PN: These are super slow compared to the speed of light. The theory in which all of the structure that we see is formed from the gravitational amplification of these little inhomogeneities, so you have these little piles. It’s almost like little piles of sand, and then because you already have a little excess in one region, you tend to attract more because of gravity and you amplify and you grow masses in clumps. And this theory has been ratified, and I would say… It’s called the standard cold dark matter paradigm. It’s been ratified, and probably the most powerful cosmological probe has been the cosmic microwave background radiation. This is the relic radiation, hot radiation from the big bang, the universe was at a very dense, hot, violent beginning. And so, this is radiation that comes to us when the universe was about 400,000 years old, and this is the last direct signal that we get from farthest back in the universe.
0:10:42.6 PN: And I think what’s remarkable about this signal is not only does it telegraph properties of the early universe, because light that is coming to us from that surface, if you will, and it takes a finite time for things because the speed of light, large as it is, is still finite. This light encounters the unfolding cosmic saga of all the formation of galaxies, the first stars, the first black holes, then the bigger galaxies assembling. This light encounters everything and it carries an imprint, and I think that is the triumph of this model. This model was not only able to explain what we see, but the evolution over time, and also calculated, was able to quantify what imprints, what are the imprints you would see on the cosmic microwave background. And all of those imprints have been now detected, and so that really basically was nail… Secure nails in the coffin of alternate theories.
0:11:51.7 SC: Right.
0:11:51.7 PN: And I think this theory is very secure, partly because there are many independent lines of evidence, and this is one. Then there is the properties of galaxies, how they cluster and when they form, properties of other structures in the universe, like the ones that I work on called clusters of galaxies, which are larger assemblages, aggregates of about thousands of galaxies held together, again, all by the gravity of dark matter, as it turns out. I think that this theory has really emerged as the theory. Although it is contested, and I hope we get a chance to talk later on about how we are kind of stress testing this model, because the data has gotten so much better that we can really hone in on the very detailed predictions and challenge them with observations.
0:12:50.5 SC: Yeah, and that’s going to be the exciting thing, is looking for places where the theory doesn’t quite match the data. But let’s just be super, duper clear here, because out there on the internet, you can read all sorts of things. And I think you already said this, but maybe to emphasize it, we’re not not really open for speculation about the whole big bang expanding universe model being wrong, right?
0:13:14.1 PN: Not at all, not at all, not at all.
0:13:15.7 SC: We’re looking for tiny little variations.
0:13:17.3 PN: Absolutely, and I think that you’re absolutely right that it bears repetition that this idea of a hot dense beginning, the big bang and the sort of subsequent description that we’ve been talking about, that is very, very, securely settled.
0:13:35.4 SC: Yeah.
0:13:36.3 PN: As I said, many independent lines of evidence. And in a way, one of… That’s why it’s been really hard to falsify this theory too, or to come up with an alternative. So there are… There are researchers working on an alternate theory, partly because beautiful as this theory is and concordance with data, there’s a couple of big nagging questions. First is this nature of dark matter. Several decades on with searches directly looking for the particle, looking indirectly for a signature of the particle, they’ve all come up empty. So we are yet to find this dark matter particle. And second, the universe is not just expanding, but the expansion is accelerating. And as we all know from our experience of driving around, you kind of need to hit the gas pedal if you want to start accelerating.
0:14:29.7 PN: So clearly something, we don’t know what it is, and so we’ve called that dark energy, is propelling the accelerating expansion of the universe. And so these two key elements, and in terms of the overall composition of the universe, these are the principal components of the universe in terms of the energy contribution to overall energy contribution. And so it’s embarrassing. And the stuff that we are made of that we know quite well, and we’ve charted about 4% of the total inventory. So given that we have these two big gaps, there’s obviously room, right. However wonderful this theory is, we really don’t have an understanding of what dark energy really is.
0:15:14.6 PN: And so there are… There is one sort of alternative attempt to try to explain, is there a theory in which you could just modify the nature of gravity somehow so that you can sort of do away with the idea of dark matter, and that what you really have done is altered how the force of gravity acts over large distances. I mean, I think it’s an interesting avenue worth exploring. I’m a scientist, I have to be open-minded, although, and this model, the cold dark matter model, I have to say, one of the reasons it’s very hard to falsify, in addition to the overwhelming amount of accumulated evidence, is the fact that is the history of how this theory evolved itself and how our understanding evolved. So in many ways, it was a theory that started out with degrees of freedom, and as we made observations, you hone and refine a theory, so this theory has been honed and refined for several decades.
0:16:21.9 SC: When you say this theory, tell us exactly which of the many theories that we’re talking about you mean.
0:16:27.0 PN: It’s the cold dark matter theory.
0:16:28.9 SC: The cold dark matter standard theory, okay.
0:16:30.7 PN: Yeah, standard theory.
0:16:32.5 SC: And when you say cold dark matter theory, that means not just that there is cold dark matter, but that the universe is expanding and the cold dark matter has been responsible for the growth of galaxies and structure. Like that’s the whole theory that we’re talking about.
0:16:44.4 PN: Yeah, that the whole panoply of how the universe has started and evolved to be where it is, all of those elements form, but we refer to it shorthand as cold dark matter because we see post-facto that cold dark matter is in the driving seat. That’s what has driven everything that we really see.
0:17:04.6 SC: But the people out there listening might think you’re talking about what the dark matter particle is, but this… When we use the phrase the cold dark matter model, we mean this whole picture of cosmological evolution. And so, I mean, maybe say a couple more words, because this dark matter versus some alternative like modified gravity is obviously a crucially important question, and I think, like you say, we can imagine gravity being modified, but given all the evidence for dark matter, it’s hard these days, especially with the microwave background, etcetera.
0:17:33.7 PN: Absolutely. I think… I think what has been really difficult for these alternate theories of gravity, even though they could dispel, so there’s a lot of evidence for the existence of dark matter, sort of halos of dark matter around every galaxy, pretty much, in the universe, and that gets reflected. The presence of that dark matter gets reflected in the motions of the stuff that we do see, which is stars, because the stars are now feeling the gravity of the whole galaxy, and so they’re moving around much faster than if they would be moving if only gravity was provided by the visible gas, dust, stars that we see. So sort of the original way in which Vera Rubin and her collaborators came up with the idea of dark matter in the first place.
0:18:21.8 PN: And so to explain the… So these are called the motions of stars that you map in galaxies, it’s called the rotation curve, basically, you know, how the motion of these stars… And so the evidence for cold dark matter, compelling evidence came from these rotation curves, where basically the… Instead of seeing an edge to the galaxy when light just drops off, you see galaxies are very centrally concentrated light in stars and then they become fuzzy, fuzzy, fuzzy, they appear to kind of have an edge, at least in light. But it turns out that if you actually look at the motions of stars, doesn’t look like they have an edge, looks like actually there’s something holding up galaxies at the outer parts. And so that’s reflected in this rotation curve. So these alternate theories can actually kind of dispense this idea of dark matter holding things up and they can explain that.
0:19:16.7 PN: However, the one effect that has evaded calculation for them, and there have been attempts, is the bending of light that Einstein’s theory of general relativity predicts and is observed in abundance in the universe. So just that kind of phenomenology of seeing light bending, and this is manifest, this light-bending is manifest very clearly. You see really misshapen galaxies, you see them kind of stretched out, very unusual shapes in very particular configurations, and that theory cannot quite… The calculation hasn’t been done, to be fair, but I think that’s one of the reasons, the fact that you have all this compelling light-bending that you see, which is information-rich, right? That’s one of the areas that I work in, and there is so much data that shows you that there’s rampant light-bending, that theory has not had… Well, we’ll see in a few years. I remain open-minded to see, well, this is how we have to be as scientists, as you know, Sean. We are trained to stay nimble and sort of the provisional nature of science, the inherently provisional nature of science is at any time, it’s best to date.
0:20:50.0 SC: Sure.
0:20:50.5 PN: Your understanding is best to date, and you have to remain open to revisions and refinements.
0:20:56.4 SC: Well, and also let me just add in the temperature fluctuations in the microwave background radiation as yet another piece of data that fits very comfortably with dark matter and is very hard to explain by modifying gravity.
0:21:10.3 PN: Yeah, so they have current versions of their theory that can account for the imprint in the cosmic microwave background, but yeah, it’s not easy, but as for the cold dark matter model, it’s a very nice fit, and all the details that we see… So we see this imprint in terms of wiggles in this background radiation, so this imprint, imprinted wiggles, and the wiggles are really small fluctuations in the temperature, like the sixth or seventh decimal place, right? But those details, those very fine details are just exquisitely explained by the cold dark matter theory.
0:21:55.1 SC: And just to get it on the table, my vested interest is very much in that modified gravity should be the answer, not dark matter. I would much prefer if gravity were modified, that’d be much cooler for me, but the data are not letting me believe that.
0:22:09.1 PN: Yeah, not right now, anyway. And not this version. So it may turn out that there is a version of the… I don’t know, one of the reasons I remain open-minded, although I’m kind of… Obviously, my work is very invested in the cold dark matter model, is the fact that we know that Einstein’s theory is incomplete, because we believe that all the forces do have to be unified, so this is not the last word on gravity. It’s a compelling word on gravity, so Einstein’s theory has been tested in many, many regimes, and it comes out tops. It really comes out really fine, but you never know what may be lurking there in terms of some tiny gap or mismatch.
0:23:00.9 PN: This brings us to my favorite kind of angle. My angle is that it is… For my entire career, one of the things that’s really driven me is to find these potential tiny gaps between theory and data, because I think that there is a lot of information and potential for discovery in these gaps. I mean, not all gaps, so obviously there’s sort of the standard case from the history of science where Urbain Le Verrier was very clever and he looked at the perturbations in the orbit of Uranus, and he was able to cleverly predict that it was Neptune that was causing these perturbations. So everyone was worried, right, because at the time, because Newton’s theories were laws, were sacrosanct and here’s a little deviation, but he was able to sort of explain in a way within the context so Newton’s laws remain intact.
0:24:02.8 PN: And similarly, there were anomalous precession in the orbit of Mercury, and economical as Urbain Le Verrier was, he said, “Hey, same explanation. I think there’s another planet somewhere between the Sun and Mercury,” and he called it Vulcan and people went around looking for it and so on. It’s not there, obviously, and the resolution to that required a brand new theory, so it kind of… Although Einstein didn’t set out to start explaining that anomaly, his theory naturally explains that in his complete reimagining of gravity itself, the nature of gravity. No, I think the gaps are very enticing because you never know if you’re in the Neptune phase…
0:24:43.7 SC: Sure, you never know. And one of the great things that Einstein bequeathed to us was this idea of the bending of light, of gravitational lensing, which is where… You already mentioned it, but I really wanted to dig into the details about that for this discussion, because it’s what you work on and it’s kind of something that doesn’t get quite as much popular discussion as maybe it could. So I think we can all intuitively see how a big massive gravitational field could deflect light, that seems to make sense, but one of the great questions in cosmology is you see a galaxy, an image of a galaxy or the microwave background, I suppose, you see something out there in the sky, and you’re saying, well, it’s… Light has been deflected. How do you know that? Like, compared to what? How do you know what it should have been if there weren’t any gravitational lensing?
0:25:33.7 PN: Right. Thankfully for us, most of the universe, as we mentioned earlier on in our conversation, dark matter is very lightly smeared, so most of the universe is kind of empty of matter and then… But however, there are regions where it’s really concentrated. So when we look out into the night sky, the places where you would get dramatic light bending, the fraction, the patches on the night sky where you could expect extreme distortion, probably 10% max. So most of the time when you’re looking at the universe, you’re not seeing dramatic distortions, just a slight distortion, perhaps, so when you look through regions where there aren’t large clumps of matter, mostly dark matter, then you know that you are not seeing… You’re seeing the undistorted shapes of galaxies. So yeah, this is a big tricky issue because, you know, galaxies are born with a range of native shapes.
0:26:36.3 SC: Yeah.
0:26:38.3 PN: So we need to know the distribution of native shapes for galaxies or birth shapes. And then modulo that, we see something that is distorted, so it is with respect to this distribution, so it’s done statistically that you have a distribution of the undistorted birth shapes of galaxies, and then you look at a patch of the universe where you know there is something like a cluster of galaxies, ’cause you see all the galaxies that are clustered together to make that cluster, and you know that it’s a huge repository of dark matter, and then you go measure the shapes of galaxies behind this lens, so this matter acts as a lens, very much like our intuition with glass lenses and optical lenses, and so we are looking at a patch of the universe in the back and the light from there gets really highly distorted, we can then look at the distorted shapes you get from the stuff behind the lump versus the rest of the universe, and that is what gives us a calibration.
0:27:40.8 PN: We are able to then back out how much matter there should have been in the sort of entire cylinder from us from our eyes, all the way to where these galaxies, distant galaxies are. So what we are able to do is to sort of invert the distortion, like we can undo the distortions and figure out how much matter you needed to produce that, and in fact, it turns out this is one of the most compelling pieces of evidence for the existence of vast amounts of dark matter, because when you’re go and back out and you look at a region that has, for example, a cluster of galaxies, now you’re able to see all the light, so you can count up all the galaxies, the visible matter that is part of this inferred amount of mass that you need to produce the lensing, and you see it’s a deficit of about a factor of 10, you need almost 10 times more matter than you actually see in stars. And particularly, we now know that clusters of galaxies are the regions of the universe that are sort of the largest, contain the largest amounts of dark matter.
0:28:53.7 SC: Are most galaxies in clusters?
0:28:57.1 PN: No, a very small fraction of galaxies are actually bound in clusters. Most galaxies live a quiet life in the field. Clusters… I say so, because clusters are very violent and… It’s a very harsh environment, because there are things that are whizzing around, galaxies are falling in, the gravity is so strong that they’re getting ripped apart when they fall in, so it’s a very dynamic transformative environment.
0:29:27.0 SC: Is our galaxy, the Milky Way, inside a cluster?
0:29:29.8 PN: No, thankfully, I guess, for us. So it turns out the way structure is organized in the universe is kind of hierarchical by mass. So there’s small lumps, medium lumps, big lumps. So our galaxy is part of one of these sort of medium-ish lumps, which is called a group. So we are part of something that’s called the local group. And, but you know, so it’s more than… So it’s a few galaxies that are kind of hanging out together.
0:30:00.7 SC: Yeah. Okay, so even the lonely galaxies, if they’re big enough, will have like little satellite galaxies, and we have a friend in Andromeda, but otherwise it’s not a very big collection. Is that right? Okay.
0:30:09.4 PN: That’s right, that’s right. Well, they are very faint. There are lots of faint galaxies that are part of our local group.
0:30:15.8 SC: And do they also, by the way, have dark matter? I mean, isn’t that a… If it were modified gravity, then I would think that there’d be a sort of unbreakable relationship between how much visible or ordinary matter there was and in the gravitational field, whereas if it’s a combination of dark matter and visible matter, then you might have mixtures, whereas sometimes it’s all dark matter, very little visible, and sometimes it’s the other way around.
0:30:41.6 PN: Yeah, that’s a great question. So as it turns out in our universe, in this normal sequence of how galaxies form, assemble and evolve over cosmic time, you actually find that the fraction of dark matter that a galaxy should have is a huge range. There’s a lot of stochasticity, so you can have some galaxies that are quite bereft of dark matter, and then you can have some extremely dark matter-rich galaxies, and so there’s a real range that is allowed and you can imagine that… You can see probably how that happens, because there’s a lot of randomness in terms of physics and dynamics of things getting close to each other, two galaxies smashing in, having a close encounter and modifying each other along the way. So you can imagine that there’s a lot of room for variation.
0:31:33.2 SC: Good, and let’s go back to the… I want to dig in on even more to the gravitational lensing story, so if I understood correctly what you’re saying, we can sort of look at regions of the universe where we think there’s not a lot of stuff in the foreground nearby, and so we can see the galaxies far away, and of course, the specific image of those galaxies is not going to be duplicated somewhere else, but the statistical features of that image, how many galaxies, how close they are, etcetera. And so then we can go and compare where there is a cluster of galaxies, there will also be galaxies in the background, we know, ’cause we can look at their redshifts and we can compare their statistics. Is that what we do?
0:32:14.1 PN: That’s right, we compare their shapes.
0:32:15.4 SC: Their shapes?
0:32:17.7 PN: We literally just compare their shapes and look at how distorts. We plot out the distribution of shapes, and you see how wide that distribution is. But there’s something even cooler, though. So in parts of the universe like clusters, where I’m sort of sort of repeatedly saying, right, this is the joke about people like me who work on clusters, every grant proposal starts out with, “Clusters are the largest repositories of dark matter in the universe.” And anyway. So in clusters, the inner regions of clusters, there’s a huge amount of dark matter, and the light bending is proportional to how concentrated the dark matter is, how closely packed it is. So the inner parts of clusters are so closely packed that the light bending is so extreme that you can think of light… I find this analogy useful. Look, the thing with analogies is they stop working at some point, but they let you get to a point. So you can think of light as a tube. The bending is so extreme kind of close into the sort of concentrated part of a cluster, for example, or even a galaxy, so there are galaxies that have huge amounts of dark matter at the center that you can think of this tube of light cleaving into two.
0:33:38.4 PN: So basically what happens is you end up seeing multiple images of the same single object. So in reality, there’s only one object, but you see multiple copies of its image. And under very specific configurations that you can look at those configurations and you can figure out, ah, this one has split. But you can do something even better, because every galaxy in the universe has its unique fingerprint. The way we know these are multiple copies of the same object is they have the same spectra. You go and and take the spectrum, and they’re identical. So you know that this is in fact just an artifact in the sense that it’s a multiple image of an individual source. And this multiple imaging of how dramatic it can be really depends on the concentration of dark matter. So you can split… So you… So there’s a curious thing that happens. So you split, you have a distant faint galaxy. One other thing the lensing does, it kind of magnifies, because it changes the surface area. You can think of the tube starting out as a small tube, and then becoming thicker, if you will, so the surface area changes, so you actually magnify and you bring things into view. So that’s why it’s a lens. You bring things into view and you see these faint objects that you wouldn’t otherwise see.
0:35:03.4 PN: So not only do you magnify them, you can multiply image them, you split them up, but then you produce two, for example, so classic splitting is you have two very bright images and you have one de-magnified image in the center. You actually produce an odd number of images. So you have three, five, seven. We’ve actually seen all of these configurations now with the Hubble Space Telescope, and verified that they are copies of each other by taking their spectra.
0:35:37.9 SC: You mentioned the Space Telescope, which is good, because I wanted to sort of take a little side light onto the data and where we get it from. So the Space Telescope is one place, but I presume there are other places. And also, talk a little about the ecosystem of who is observing, who builds a telescope, who are the theorists who analyze the data. ‘Cause you’re not out there with your eyeball to the end of the telescope collecting data, right?
0:36:03.1 PN: Right. No, I’m not. In fact, sadly, now almost nobody is looking through the glass literally. The people, even the observers, are looking at computer screens at this point. But yeah. So I think the sort of the… The sort of ecosystem of how all of this work gets done is that you have these skilled teams of people with different kinds of technical expertise. So you have the observers, you have the instrument builders who’ve built the spectrographs and the light buckets, so ground-based telescopes, space telescopes that are actually collecting all the data. Then you have people who skillfully reduce this data, remove all the noise from the signal. And then they… There are people like me who then model, who have theoretical ideas, theoretical pictures and storylines, and we build our models in such a way, at least that’s the aspiration, we build models that can be directly confronted with the data.
0:37:06.8 PN: And then you have people who are actually doing computer simulations. So they’re basically doing computer models of the evolution of structure in this sort of cool dark matter paradigm or any sort of… They can change the initial conditions, they can play with the kind of dark matter particle. So these are just sort of our sort of cosmology where we can’t perform controlled experiments. These are our proxy experiments. So then you have the simulator. So then you have these sort of interpolating people like me. So I analyze simulations and then I build a theoretical model, I build a conceptual model that can connect the simulations, enabling us to mock and do observations in the simulation, then confront that with the data, and then infer what is going on.
0:37:57.3 SC: When you say simulations, what exactly are we simulating? A galaxy has 100 billion stars. Do we have 100 billion conceptual stars in a simulation, or do we include exploding stars and magnetic fields? What is the state of the art there?
0:38:15.9 PN: The state of the art now is quite remarkable, but we are still not at the level of… We are simulating aggregates of stars, we are not yet simulating individual stars. However, we have clever techniques to mimic the actual explosion of the star, because what really matters about a supernova exploding is the energy, what is happening to that energy, how is it getting distributed and the metals that are locked in in this old star, how they’re getting expelled. So those are processes that you can mimic without actually following the individual star and its life cycle, so we have abstracted these out.
0:38:58.2 PN: So there are simulations that are done on multiple scales. The kind that I was talking about are simulations, cosmological simulations, in which we have a grander, we want a grander view, so we want to see a piece, a small chunk of the universe with all the stars in galaxies locked in, and the dark matter, and we want to see the co-evolution, and now we can even drop in particles that would be like black holes, and we can see how the whole… Because we know, we know all the ingredients in the universe, so we can just pop them all in, abstract their behavior, depending on what we are interested in, and then we can look at the evolution, turn the clock on and we can look at the evolution of these components, the combined evolution, which is very hard to do theoretically, it’s very hard to do, ’cause it’s all the non-linearities and how things couple on small scales and so on, that you can’t capture with theoretical models, however abstract they are and so on.
0:40:00.3 PN: And then there are simulations where people zoom in. So there are these simulations called general relativistic magneto-hydrodynamic simulations, so these are simulations where you zoom right into the heart of a galaxy, you focus on the black hole, you focus on the gas flows around the black hole, and then you put in… You have magnetic fields that are threading it and so on. So we can abstract away portions of problems and focus and numerically tackle them, and I think this is one of the exciting challenges that has also kind of informed the arc of my work, which is, how do you bridge the scales, how do you take a simulation that looks at just the inner part and then a simulation that looks at a much larger scale, how do you put them together? How do you piece them together? In terms of the implications, in terms of the astrophysical phenomena, motions of particles, are there being flows kind of pushed out to large radii, how do you couple them? So yeah, simulations have been very powerful in cosmology, they’ve been transformative in cosmology.
0:41:06.9 SC: Well, and just to emphasize the extent to which we’ve made progress, back in my graduate school days that were formative for me, all the simulations had just dark matter in them, ’cause we figured, well, most of the matter’s dark and it’s easy to do, they don’t even collide. Just putting in ordinary matter where you could make stars, etcetera, that was a big step, and the state of the art has come a long way, even though we’re not done yet, we still have a long way to go and get even better.
0:41:34.2 PN: Yeah, absolutely.
0:41:37.1 SC: And you mentioned a little bit about the black holes at the centers of galaxies, and this is another place I think that maybe some comparisons are useful, because I think a lot of people think, “Well, there’s a supermassive black hole at the center of a galaxy,” which is a statement many people have heard, isn’t that where most of the gravity in that galaxy is coming from, but that’s nowhere near right.
0:41:57.7 PN: Absolutely. Even though the black hole has all these bizarre properties, it turns out in terms of the mass budget of the overall galaxy, it’s kind of teeny-weeny, it’s maybe even a millionth of the total mass including… So for example, the total mass of the Milky Way is 10 to the 12 times the mass of the Sun, and the mass of the black hole is 4 million times. Or is it 10 to the 6 orders of magnitude… There’s six orders of magnitude difference, right?
0:42:32.7 PN: So but they punch more than their weight because they have these bizarre properties and they sit in the centers of galaxies. It turns out you could have some of them wandering around too, that’s some exciting new work that we recently published that you could have a population of wandering black holes, but regardless, there’s only a very small region around the center of the galaxy where the gravitational effect of just the black hole dominates, the stars take over very quickly. So there’s a small region of influence and basically outside that, it’s gravitationally inconsequential, but we now know… So we have dismissed them, we thought, okay, they’re really tiny. [chuckle] They’re important, right around if you’re right there, of course it’s consequential, but otherwise not, but it turns out now they were believed to be so marginal, but now we’ve understood that they actually play an outsized role in modulating how stars form.
0:43:33.0 PN: Stars appear to form inside out in galaxies, by and large, so the inner regions are where bulk of the stars are forming, and so black holes are sitting there. So it turns out that black holes can heat the gas and they can… The energy that is falling into the black hole is a gas particle that is getting gobbled by the black hole, gets pulled in, gets heated up, get sped up, starts to radiate and gives out a lot of heat. That heat can prevent stars from forming, because for forming stars, you need to cool the gas, so this is actually adding energy. So we realized suddenly that black holes could be sources of energy in galaxies, and therefore this intricate feedback between how they either, do they promote or do they prevent stars from forming. This whole cycle has been much better understood recently, and now we believe that even though gravitationally they’re not as important, they punch much more than their weight in terms of determining the phenomenology of a galaxy, how it looks, because the how it looks is determined by the stars so, and how many stars form. So it modulates, we believe, the efficiency of formation of stars.
0:44:47.3 SC: One thing that is coming clear from what you’re saying is that, again, the whole scale of time and space in cosmology is so different than what we are used to in our everyday lives. It’s been 14 billion years since the big bang. We look at pictures of galaxies and they’re not moving, they’re just sitting there, right, like when we take a picture of them. But on the scale of, let’s say, 1 billion years, a lot happens in the universe, the universe is actually a really dynamic place, even on the scales of galaxies and clusters.
0:45:17.5 PN: Yeah, absolutely. It is… One sort of analogy that I can think of is that when you look at an anthill from really far away, you don’t see all the little ants kind of busily walking around. It seems to be a pretty stable structure that keeps its shape and stuff when you’re looking from… When you get close, you see, oh, wow, look at the amount of activity these ants are up to. So it’s somewhat like that, right? So it’s a question of scale and perspective. So if, as you said, a billion years is, it’s a fraction of the age of the universe and still… So for example, when two galaxies crash into each other, sort of the finale of them ending up as one sort of mixed up, messed up product of soup of stars and gas could take a billion years. That whole process.
0:46:14.6 SC: And that’s our future, right? We’re going to crash into Andromeda?
0:46:18.1 PN: Yes, yes, and we would make Milkomeda, as it’s called, right?
0:46:24.7 SC: I did not know that. [chuckle]
0:46:26.4 PN: I’m often asked about this and people appear somewhat worried, and then I have to say, “Well… “
0:46:32.5 SC: I’m worried.
0:46:32.9 PN: “We’re talking of billions of years in the future.” Also, come on, come on, folks, I think we need to pay attention here and now. We seem to be on course to destroy our planet within the next couple of hundred years, so…
0:46:46.6 SC: It’s the galaxies merging that I’m really worried about, yeah.
0:46:48.8 PN: Well, right, exactly.
0:46:50.4 SC: Let’s get back to these black holes. Does every galaxy have a massive black hole at the center?
0:46:56.2 PN: It appears so. It looks like it is such an essential and fundamental feature of how structure forms. Just as stars form in this cocoon of dark matter, it appears that the formation of a black hole is pretty fundamental. So most, if not all, galaxies appear to have a black hole in its center, and in the cases where we don’t see one, we see signatures of it having been kicked out. So it might have been there, it might have crashed into something, and then it got kicked out. So there are a few cases where we see that kind of signature.
0:47:31.2 SC: And what is the story of where these black holes came from? Were they there for a long time or did they slowly accumulate?
0:47:38.1 PN: So this is another thread of work, of my research, which is trying to understand the origin of the first black holes. So one thing we know, there’s a natural way to form black holes, which is just the end states of stars. If you are a star that is eight times or so more massive than the Sun, when you are born, inevitably, you finish your life cycle, all the fuel gets exhausted and then you leave behind a black hole. But those black holes are really, really tiny, and the black hole in the center of our galaxy is four million times the mass of the Sun. This one is perhaps a few times the mass of the Sun. So how do you grow from…
0:48:15.5 SC: Yeah.
0:48:16.3 PN: What are the intermediate… What do the teenage years look like? What does adulthood look like for little black holes? So it turns out that black holes can grow either by accretion of matter, so just sucking in matter, accretion is a fancy way of saying sucking in matter, or by crashing into each other. So when they obviously merge into each other, they shake up all of spacetime, they produce gravitational waves. We’ve seen that happening for little black holes. We’ve not seen that happening yet for supermassive black holes, but that’s coming with the LISA interferometer in space that is going to be put up within 20 years or so.
0:48:55.0 PN: So we know that black holes can grow in these two different ways, but there is an interesting conundrum that kind of motivated a lot of my work more than, oh, gosh, I hate to admit it, about 15 years ago now, because it reminds me of how old I am when I think about, “Oh, my God, that paper was 2005.” So we also know that these growing, kind of feeding black holes are quasars, and these are very bright beacons that we can see. So a quasar is basically the heart, the black hole at the heart of a galaxy that is feeding so rapidly that it outshines the entire galaxy, and we see them peppered everywhere in the… They are really peppered, they’re not all that common. They are rare objects, but the ones that are lit. So black holes themselves are actually fairly common, but not all of them are actively feeding and growing. So for example, the black hole at the center of our galaxy is actually fasting. You really won’t see it. It’s not a quasar.
0:50:00.4 SC: But did it use to be a quasar?
0:50:01.5 PN: Pardon?
0:50:02.1 SC: Did it use to be a quasar? Did it go though a phase?
0:50:03.7 PN: It probably, yeah.
0:50:04.6 SC: Probably, yeah?
0:50:04.7 PN: It probably cycled through. So when food was ubiquitous, it was a quasar, and then it depleted everything that it could eat in its vicinity, and in the absence of having new gas getting smashed into the center with the merger, it was sitting there fasting. So these feasting black holes you’ll see as quasars, and we are discovering… And from the brightness of those quasars, you can figure out how big the black hole is. So the bigger the black hole, the more rapidly, voraciously it can fuel itself and feed, and therefore the brighter it will be.
0:50:42.5 SC: Maybe we need to be a little bit more specific about how a black hole can be bright since, after all, they’re black.
0:50:48.4 PN: That’s right. So a black hole is bright because of the dying gasps of the matter that are actually getting eventually sucked into oblivion by the black hole. So en route, whatever the black hole is feeding on, be it a star, it could also… Black holes can also gobble stars, but if it’s basically matter coming in the form of gas, and as we said early on, everything is basically hydrogen. So hydrogen is glowing, and that’s what you see and that’s how you see a black hole. And these quasars are basically where you see very, very, very active feeding episodes. And we’re detecting these quasars now to very early on in the universe, when the universe was a fraction of its current age, 10% of its current age. And we are already seeing quasars. And once again, we know where the quasars are because we can measure their redshifts. So we know they’re very distant, and they already seem to be harboring a black hole that is a billion times the mass of the Sun.
0:51:55.4 SC: Well, but you left us hanging a little bit, because you said we know how to make a black hole from a single star, and it’ll be a few times the mass of the Sun, and you said that that could grow by either creating random matter in the universe or by coalescing with other black holes. So which is it? Or is it both, or how do you go from 5 to a million?
0:52:14.7 PN: It is both. It’s both. But it turns out that it is a challenge to grow them, regardless. And you can… They will grow as long as they are very gluttonous for periods of time, so that they are really super actively feeding, then you can start from something that’s a few times the mass of the Sun to the guy that’s sitting in the center of our… Actually, the center of our galaxy, a few million times, you can grow with just one or two episodes of gluttony. But if you have something more massive than that, then it starts to make them problematic, because then you have to really be over-feeding for a very… Throughout your lifetime, you have to be over-feeding in order to end up. And that’s hard, because to over-feed at that rate, you need a huge amount of gas supply. And we know that the gas supply in the universe kind of drops with time, so… Which is why there’s not much gas in the center of our galaxy. Most of the gas by now is locked up in stars. There’s not a whole lot of gas…
0:53:20.2 SC: Or it’s fallen into the black hole, right?
0:53:23.4 PN: Pardon?
0:53:23.8 SC: Or it’s fallen into the black hole.
0:53:25.0 PN: Exactly. It’s already gone into black holes or it’s captured in stars.
0:53:29.9 SC: So I would think that, you know, a galaxy is a big thing and it has some extent and there’s stars all over the place. So if those stars, the more massive of them exploding and creating black holes is the origin of the supermassive ones, I would be mystified as to why the supermassive ones are in the middle of all these galaxies. But there must be some reason why…
0:53:50.5 PN: It’s gravity again. It’s gravity in action.
0:53:52.4 SC: It’s gravity again.
0:53:54.0 PN: So it’s because it’s… It’s very massive, it starts… Compared to the stars. Compared to the mass of a star, it’s much more massive as it’s gathering mass. As it’s gathering mass, it slows down and it becomes plumper and plumper, and then it gets settled down to the center.
0:54:10.7 SC: So it really is kinda… It is this dynamical kind of thing, once again, where it’s not that the center of the galaxy is some special place where black holes form, but any black holes that are medium-sized would sort of drift down there in a natural sort of segregation.
0:54:25.7 PN: Yeah. And this segregation is quite efficient because of the sort of mass difference. Black holes are so much more massive than individual stars, so they can really grind our stars and settle in. But I was coming… I was talking about these bright quasars at high redshift, because that kind of posed a problem that many, including me, wanted to solve. It was exciting, because you see these billion solar mass black holes in place when the universe was a fraction of its age. So you could not have… There’s not enough time to grow from a little… A few times the mass of the Sun to a billion times the mass of the Sun within that short span of time that is available in the early universe. There’s this other fact about the passage and the flow of time in the universe. There’s not a whole lot of time in the universe early on. Lot of the time is actually kind of stretches towards later times in the universe.
0:55:24.1 PN: And so then it was pretty clear. The answer was obvious, but the question was, What do we do? How do we get this to happen? Obviously, what you had to do is you had to make seed black holes that were really massive from the get-go. So if you could make a seed black hole somehow from the physics that is 10,000 times the mass of the Sun, 100,000 times the mass of the… Get-go, just from the get-go, then you have it. Then it’s not a… There’s no feeding problem. Then these quasars that you see that are really bright are very rare. So it all works out really nicely, because you need that to happen for the rare cases, but enough of the rare cases, not just a freak case, but it’s a population. So we propose this idea of direct collapse black hole. So these are black holes that form from gas that collapses very rapidly. You bypass the formation of a traditional star.
0:56:18.4 SC: Okay.
0:56:20.0 PN: You don’t have to cycle the gas through a star, and then you leave a little wimpy black hole, but in fact, you put… The analogy that really works, at least worked for me when I was first doing the calculations, because it was really quite apt, is if you’re sitting in a bathtub, I love to take baths, so when you pull the plug, you see that vortex, the water rapidly fueling down, that is exactly what happens with the gas that forms these direct collapse black holes early on in the universe. So you… Literally what we had to do is to find the analog of the taking the plug out and a region in the universe where you could somehow cause motions, kind of… Could you whip up some motions very early on in the gas. And it turns out that there are places in the universe this can happen, because in the sequence of how galaxies and stars form, you first form a huge disk of gas, and then that gas cools and fragments and forms stars.
0:57:24.1 PN: So if you had this big disk of gas, and you prevented fragmentation and ended up creating a vortex that could cycle all the… And you can create a vortex because you can set up these sort of global kind of disturbances in the gas that could set up this vortex. So you could instigate this vortex with known astrophysical phenomena, nothing extraordinary. And then the only thing you have to make sure is that this gas disk does not fragment before the stuff falls in. If it fragments, it forms stars. Then you’re done. Then you can’t make this big seed. So we also know how to prevent the fragmentation of gas, how do you prevent the gas from cooling. It’s really neat. In the early universe, you just have to get rid of molecular hydrogen. So it’s a little bit… It’s detailed physics, but so basically, you can form these very massive black hole seeds, and that sort of alleviates this problem of how these very big quasars, bright quasars with supermassive black holes, or even ultramassive black holes are already in place early on in the universe.
0:58:36.3 SC: But when you say from the get-go already in place, just to be clear to our audience, you don’t mean from the big bang. The big bang was still very smooth.
0:58:45.8 PN: Oh, no. Not at all. No, no, no. So there is a huge gap, as we said. There’s not a whole lot happening in terms of the assembly of structure. And there’s a lot of portion of the age of the universe is just radiation-dominated, then you move on to sort of a matter-dominated era, but only very, very towards the very end, it’s very recently that we started assembling galaxies. So this is talking very much relative to the assembly, the epoch when the first stars formed. So it’s around about the same time. And then the question is, Could you form… The picture that I’ve given you, it sounds like you could form the black holes before you make the stars.
0:59:27.6 SC: Oh, okay. Yeah. Because there’s gas in there without stars.
0:59:28.9 PN: It turns out not quite, because to prevent the fragmentation, you kinda need some stars in place already at other places. You need a neighborhood of stars, little galaxies, and then you can have this gas disk make your vortex and give you your direct [0:59:46.1] ____.
0:59:46.1 SC: Because we talk about the dark ages in between when the microwave background came around between 400,000 years or so, and then the lighting up of the first stars, which was when? Do we know?
0:59:57.1 PN: Yeah. We don’t know quite precisely. There’s a little bit of uncertainty, because we don’t know exactly when the first star likely formed, but probably about 6-8 billion years ago.
1:00:10.2 SC: So yeah, I’m thinking from the big bang. So if it’s 14 billion year old, it’s still 6-8 billion years work either direction, ’cause it’s a 14 billion-year-old universe. Yeah. That’s very… That’s very good. Okay. Yeah. So that’s a lot of dark ages, actually. Right. And the final provocative thing that you mentioned there is that you can have not just supermassive black holes at the centers of galaxies, but ones that are just wandering freely through the universe that are not attached to galaxies.
1:00:35.2 PN: Right. And that is something that… So one of the big pieces of the puzzle, as we’re talking about this growth history of black holes, is that we are looking at the infant black holes, and then we look at the geriatric ones. And then so what’s happening in between? So there’s a phase called intermediate mass black holes, and that’s been pretty elusive. It’s almost like looking at a population of people where you just don’t see the teenagers, and you see either the infants and little kids, or you see adults and older people. But they are there, and like rebel teenagers, these intermediate mass black holes are not actually where we believe they might be lurking. So it turns out they are not in the centers. They’re not massive enough to have sunk into the centers. We’d been looking for them in fainter galaxies, because how big a black hole is is tied to the mass of the stars that formed around it, as we saw earlier, and so we would have to look in sort of wimpy galaxies to look for these sort of wimpier intermediate mass black holes, but we had very little success finding them.
1:01:39.8 PN: So then it turns out that they were… One was found, and we found it was not at the center, it was far out. So that prompted us to start thinking about… In the simulations. And so there was a breakthrough with the simulation that one of my collaborators and their team did, in which most simulations, these models, you had to pin down the black hole to the center of the galaxy, just a numerical technique, you had to do that. And in this simulation, they didn’t have to do that. They just let it move freely as it wants. It settles down in the center, because that’s sort of… Because of its gravity dominating. But it turns out that in the process of galaxies kinda merging and their black holes tumbling around, there are many that are left wandering in the outskirts of individual galaxies and in clusters. So we detected this population and then we’ve characterized. And that’s sort of the exciting new result.
1:02:32.3 PN: But the one question that you had earlier that, Sean, I want to come back to, it was a great question you asked about the sort of the standard cold dark matter model and the sort of the tensions. And I want to… I’m very excited about this, so I want to share this. So recently, you remember, sort of this looking for gaps thing, recently we found something really intriguing. And that is that the strength of lensing, so now we are back to lensing, away from black holes, still the dark universe, though, still the invisible universe, that the strength of lensing that is observed in the universe is incommensurate with the expectations of the cold dark matter model for the sort of the inner parts of galaxies.
1:03:20.4 PN: So you have a cluster, you have all these cluster galaxies that are part of it that are all held together by dark matter. So there’s like this smooth sea of dark matter holding everything together, but then each galaxy has its own little halo and content of dark matter around it. And from the inside out, so there’s dark matter everywhere in a galaxy. So we find that there are some new lensing signals, because there’s deeper data. So you could look and capture tinier lensing signals. So even smaller distortions on smaller scales the Hubble was able to detect. It was a long look. And there we found that the strength of lensing, therefore the number of lensing features, all these strong distortions doesn’t match with the predictions of cold dark matter theory. And the reason this is super intriguing, it’s a gap. So by a factor of 10. It’s huge.
1:04:13.8 SC: Sorry, a factor… A factor of 10 between what and what? What is the thing we’re measuring?
1:04:17.8 PN: The theory… The predictions of the number of these kind of little lenses, inside lenses, events that the cold dark matter theory predicts from simulations versus the real universe is providing for us. So it’s a factor of 10 off. So the real universe, the lenses within lenses, so we have this big lens and then you have these galaxies that act as little lenses, they are much more efficient by a factor of 10. The universe is a much more efficient lensing machine than our cold dark matter theory suggests. Of course, the discrepancy by itself is awesome. It’s factor of 10. You can’t kind of say, well, maybe you will have a slop of factor of 2 here, maybe you have something missing here. You can’t. Unless, of course, you have to conjure up like six different things that are all factors of 1.5-2 and kind of really knit them together to get you that.
1:05:15.0 PN: So the reason this is really intriguing is, previously… So as I said, the gap finders are around. Many of us are gap finders. Previous gaps that were found actually went in the opposite directions. So the real universe did not have as much matter piled in the centers of galaxies. It appeared that there was less matter piled in the centers of galaxies than the theory would predict. And so here we are finding something completely the opposite. It’s a different environment, and it’s something that you alluded to earlier, that this is the dense… This is the equivalent. Cluster is the equivalent of New York City, like a bustling city, tightly packed, densely packed.
1:06:00.8 PN: So the previous findings were sort of in the suburbs. So in kind of isolated galaxies or galaxies in groups and so on. So this is, in my opinion, this is a very intriguing mismatch. And this is going to be kind of hard to incorporate and argue away within the cold dark matter model. It could be that there’s something… Obviously, it could be that the simulations are after all… We are putting in what we are getting out of the simulation. So computationally, we may be missing some feature, maybe we’re not modeling something correctly. But it’s still pretty hard to see how it’s such a big factor. And if it’s such a big factor, then it’s kinda puzzling why everything else kinda fits so well. If you are so off, how does everything else fit so well?
1:06:54.3 SC: Well, to be clear about… You just made this point, but just to super clarify it, when we talk about a discrepancy between the theory and the observations, the theory doesn’t speak to us in an unmediated way. Making that prediction is a highly contentious thing. And like you say, there’s a lot of simulations involved, etcetera, etcetera. So the… Even if it’s a small likelihood, the option remains open that the theory is fine, but we haven’t correctly understood what it predicts.
1:07:27.3 PN: Yeah. Absolutely. Spot on. Because what we are inferring even from these simulations are mediated by a model. So there is a model, so there’s an intermediate stage. So it could be that there is something in our sort of conceptual model. But there’s, of course, the other exciting possibility that maybe something about one of our assumptions about the nature of the dark matter particle itself is not quite right. This is super interesting. It’s speculative, and I’m not pushing that. I’m not saying that we are calling the entire paradigm into question or any such thing. It’s just that it’s very interesting, because these kinds of stress tests are exciting, because once again, you never know if you’re in the Neptune situation or the Mercury situation, right?
1:08:19.7 SC: That’s right. But so let’s, again, put it into context. Again, you already did this, but this is such a rich story, and it goes back and forth, so I want to make sure we get it right. For a long time now, there’s been a… Or there was, it always seems to go away, but there are always these claims that cold dark matter theory predicts too much concentration on small scales, that it predicts that the centers of galaxies should have more dark matter than we observe there, and there should be more little satellite galaxies than we observe. But the issue always has been that you’re trying to compare a theory and observation with the dark matter theory in a regime where it’s not just the dark matter that matters. You’re looking at the regimes where the stars and the gas and everything are important, and those are the hardest things to model. Is that accurate so far?
1:09:12.3 PN: Absolutely, absolutely. And these discrepancies always tend show up in the parts of galaxies, in the regions of the universe where normal matter and dark matter are really rubbing up against each other, really closely packed. And we have made this assumption that this dark matter does not interact with the normal matter in any way, other than gravitationally. So there’s no charges, there’s no electrical forces, there’s no other kind of force. And I think, absolutely. So that’s really where all the tensions are emerging. And in the previous kind of crises in the inner parts of galaxies… So even though… So one thing you could do, you can imagine that if you need… You could just re-distribute. If you found a way to re-distribute matter, so you have… Suppose that in nature, you start out with the inner parts of galaxies being really dominated by dark matter, but then somehow, I don’t know, a supernova… Something, some explosion, something happened that pushes out some of the dark matter and sort of evacuates it, if you will, and reduces the matter.
1:10:27.5 PN: So that kind of redistribution could work, but once again… And that they believed was really sort of what worked to explain these things away, when the original finding was that dark matter appears to be heaped much more strongly compared to the real data. But I think the paradox is that now we are finding, with better data, what has really changed is much better data, higher resolution data. We’re actually finding that in the real universe, it’s actually much more heaped in the centers of galaxies that live in clusters.
1:11:08.0 SC: Right. Okay. So very specific environment. But again, I just want to go through the whole story very slowly, ’cause it does go back and forth. So the original problems were that galaxies were smoother in the center than the dark matter theory seemed to predict. And people will say, well, then why did you cling to your dark matter theory? Why don’t you try to modify it? And you and I know, well, of course people tried to modify it. That’s full employment. I’ve written papers about modifications of dark matter particles interacting with each other, interacting with ordinary matter, etcetera, etcetera.
1:11:39.7 SC: My impression, so I’m going to ask you as the expert, ’cause I’ve not followed this in recent decades, my impression is that largely those attempts failed. It wasn’t easy to add new physics to dark matter to explain these anomalies. And on the other hand, the more this data improved and the more the simulations improved, the anomalies did not tend to stick around. We mostly have accounted for them in terms of ordinary non-dark matter physics. Is that fair?
1:12:10.1 PN: Right. Yeah, yeah. And the resolution was found within the cold dark matter theory and our methods of simulation.
1:12:16.0 SC: Right. Okay. Good.
1:12:18.2 PN: Within that context of a theory. You’re absolutely right that the modifications, the kind of modifications that we’re suggesting, self-interacting dark matter or self-annihilating dark matter, all of these options, it turns out, were not really needed, because once the simulations improved and our understanding of this physics, the complex physics of kind of normal matter and dark matter kind of being clumped in the inner regions, and their potential impacts on each other, sort of the… There’s sort of an interplay. They don’t interact, but there’s an interplay. And once we could simulate that better, those anomalies kind of went away.
1:12:58.3 SC: Mostly went away, right? And so, good. But the new one that you’re pointing at, does it have a name? Is there a label for the new anomaly?
1:13:04.7 PN: It’s called GGS. In good astronomy speak, we have an acronym.
1:13:09.4 SC: That’s terrible.
1:13:10.3 PN: And it’s an acronym that’s not particularly imaginative. But it’s telegraphic. It contains information. It’s called GGSL, Galaxy-Galaxy Strong Lensing.
1:13:24.0 SC: Yeah, you need a better label for this, I’m sorry. Your branding team needs to get together and really work on this. Okay. Galaxy-Galaxy Strong Lensing. And it’s giving us the opposite direction of an anomaly in the sense that there’s more structure on these small scales, but as you say, in the particular environment of galaxies inside clusters, yeah?
1:13:47.2 PN: That’s right. Which is why we are still pushing really hard to find… There’s a lot of room for finding resolutions within cold dark matter, because these are complex places. There’s a lot happening.
1:14:00.9 SC: Yeah. So this does seem to be, again, the kind of environment that is the hardest for us to understand. So having an anomaly maybe shouldn’t be too surprising, but then your point is, but this is such a big anomaly, we’ve still got to pay attention to it.
1:14:13.9 PN: Right. And that it goes in the opposite direction, that it crucially… That it’s big, it’s a factor of 10, but it also goes in the opposite direction. So I think this is the problem that I’ve been really engaged with, trying to see if you could explain this away and account for this factor of 10 mismatch within the current theory. What could we be missing? What process or whatever could we be missing?
1:14:39.5 SC: Well, but also, it does bring us into interesting philosophy of science territory, because when you are faced with a discrepancy, you have lots of options. One is, well, your prediction isn’t very good, your simulations weren’t very good. Another is, your theory isn’t very good, you need to understand the ordinary matter more. Another is, you need a better theory. And one of the things we don’t often like to admit as scientists, but it’s certainly true, is that we’re more likely to pay attention to anomalies that have ready-made theoretical explanations. When we discovered the accelerating universe in 1998, part of the reason why it was so rapidly accepted was because Albert Einstein gave us a solution to this problem back in 1917, the cosmological constant.
1:15:23.4 PN: Inadvertently, right? Inadvertently.
1:15:26.2 SC: No, he didn’t want to do it, but he did it. But we had an instant theory that explained this. So for your anomaly, for GGSL, Galaxy-Galaxy Strong Lensing, is there an easy theoretical explanation that actually involves the physics of dark matter or something like that, or is it more likely to be found in gastrophysics, as we call it?
1:15:46.4 PN: I’m on the fence about this one in terms of where I think the resolution might lie. One can speculate. One potential, because it’s such a particular environment, a very specific set of circumstances, it could be that there is something that we have… There’s some either physical process that we missed that operates much more efficiently in these dense environments, and it could be that. It could be a physical process that we’re completely missing, or it could be that there’s an incorrect implementation of the physics process in the simulation, as you said.
1:16:30.6 PN: And so I have a sort of a dark history as a sort of historian/philosopher of science. I started a PhD, which I didn’t finish. But these are the kinds of questions that motivated me then, and which have now caused me to think much more deeply about what I do. So I was very interested in the role of simulations. So what do… What are simulations doing? So they are often thought of as tools, as ways of visualizing and a way of propagating forward, even from what I… The kind of description I gave you today when we were talking is that, well, it allows you to calculate the evolution of time, over time of a small chunk of the universe, all the complexity, we can redo it.
1:17:16.4 PN: The big question is in a field like ours, where you can’t perform controlled experiments, simulations actually occupy a much more important place, because there is a way, at least now we are seeing, like in this case, it could well be that there is something that we’re missing in the simulation, in which case… And we are able to narrow it down, and in which case, then the simulation would have been generative of a new discovery, and this is not a role that we expect simulations to have.
1:17:55.1 PN: If you look at the OED simulation, the definition of simulation, it’s pretty dodgy. So yeah, I think that it’s very, very exciting to kind come up against… I don’t have to… It’s a real pleasure to be able to talk to you, someone like you who’s thought about this a lot, the limits of knowledge, right? So these are things you think about a lot, right? What are the limits? Is everything knowable? What determines knowability? And I think I should come clean. I already, as you know, scientists fall in two camps. I believe there is no reason to think that everything ought to be knowable. [chuckle]
1:18:39.7 SC: Alright, you’re allowed to think that, I think that my… I’m very happy to imagine that not everything is knowable, but I’m much more impressed with how much we have figured out than we have any right to be, so I’m suspicious that… I don’t think we’re at the boundaries or the limits of any of those things any time soon, but I wanted to ask, since we’re past the hour mark of the podcast, this is where we let our hair down a little bit. What about the sociology of the community in the face of anomalies like this? I had Adam Riess on the podcast, we talked about the Hubble tension, we’ve talked in different places about different kinds of contentious scientific issues, and I think that there is a tendency among non-scientists to think of the scientific community as much more monolithic as it is.
1:19:29.1 SC: Either they’re all pushing the agenda of the establishment or whatever, or they’re all lone geniuses with their own theories, but really there’s a lot of heterogeneity, right? When you have a claim like this that there’s an anomaly we should be paying attention to, some people are going to be like, “Oh, yeah. We should pay attention to that.” Others are like, “Eh, it’ll go away.” What is your feeling about the different camps out there in the community?
1:19:54.9 PN: Yeah, I mean, it’s so acutely observed, Sean, unsurprisingly, right? I mean, the community is really quite heterogeneous in terms of their attitudes, and I think that there is a large portion of the community that is very risk-averse, and that is kind of all… This heterogeneity kind of mirrors the practice of science that has evolved. Since we were both students, the science has become much more collaborative, even astrophysics and cosmology that didn’t used to be run by collaborations, hundreds of people, large projects are now, that it’s become big science. It’s transitioned to a big science, and I think with that comes a certain kind of risk-averseness, because you’re asking for lots of funding, tax payer money to do a project, it better be really well articulated, it better… You’d better find something. It’s not going to be hit or miss, that’s too much of a risk to take.
1:20:55.4 PN: So I think there are people who have that sort of overall risk-averse mentality, and then there are people like me, that are sort of floating around, not parts, not really engaged totally with big science enterprise, but wanting to take creative risks, and I think they would think, “Okay, why not? Maybe this is worth exploring,” and so on. But I think the thing that interests me about the sociology of science is that I bet you, that regardless of how significant or insignificant this GGSL tension is, we don’t know yet, we just don’t know. We need a lot more people to work on this. I can bet you that I am unlikely to garner as much attention as Adam and company did for the Hubble tension, partly because of course, it has to do with how pivotal the Hubble constant is and how it anchors so much of this model and our view, our cosmic view is shaped by the Hubble constant and its value and so on.
1:22:01.4 PN: And there’s a long history of controversy about it, but I also think that this is something that I often think about, who and how new ideas are proposed matters. And I think that we are in a field where the playing field is not quite level yet, and I think that somebody like Adam Riess, who post-Nobel Prize found another sort of intriguing kind of anomaly, bringing the field to a crisis point, makes a big difference. The fact that he has a Nobel Prize to back him makes everyone sit up and say, “Okay, I think we should maybe look at it a little more carefully.” A lot more people will. But you know, on the other hand, he got the Nobel Prize because he was able to resolve something that was completely crazy at the time, it was believed to be. So I think this idea of how authority is conferred and intellectual authority is conferred, and then how a radical scientific idea gets to be accepted, it depends also obviously on the content of the idea, the significance of the idea, but I think who proposes it, which community it comes from.
1:23:19.7 PN: As you said, gravitational lensing should probably be more high profile as an area, but it’s not. Hopefully that’ll change in the next decade or two, so some finding from that field could carry more water than some other fields. So I think there are many, many factors. It’s a larger sort of ecosystem, and it’s been very interesting to watch, even in a detached way before GGSL, how these new ideas in our field are accepted, and you have experience of that, Sean, you’re extremely creative, original and take risks as well. So you’ve not joined a 1000-person collaboration as their pet theorist. [chuckle]
1:24:02.2 SC: I would not be good at that. But the other thing… I like everything you just said about the human-ness, for better or for worse, when we’re in this situation where we don’t yet know when there is an anomaly, a gap, whatever you want to call it, there’s a discrepancy between theory and observation, science is at its least objective in some sense in those cases, but then the good news is we’re going to get more data, right? Maybe this would be a good final word, like what will be the new data coming in that will help us resolve this kind of issue?
1:24:39.7 PN: Yeah, so I think the new data that can really be a game-changer in this is deeper, so at the moment, we have deep data in six clusters, just six, and so, enlarging that sample of clusters for which we have the depth of data from Hubble, where you could do this kind of analysis, so to actually detect this lenses within lenses, that’s what is sort of needed, but I think more than that, what you want is engagement, you want other people to get excited. I think that’s what I really want. I mean, I would like many other people to get excited to explore this further and see is there something about the setting, the cluster that matters or whatever. I think attracting other intellectual capital people, their attention, getting people to work on this, not attention just in terms of looking and following it, but actually engaging in the work, and I think that would be awesome, I’m really hoping…
1:25:46.9 SC: See, that’s where we differ because I work so slowly, I don’t want anyone else working on the same areas that I am, because otherwise, they will get every, all the good answers ahead of me, but looking forward to where your anomaly goes and what we learn more about the universe. So Priya Natarajan, thanks so much for being on the Mindscape podcast.
1:26:03.9 PN: Thank you so much, Sean. This was great fun. I love your podcast series, and it’s a real honor to be on it.
1:26:11.9 SC: We’re very happy to have you, thanks.
[music][/accordion-item][/accordion]
Excellent episode. May be a good idea to correct estimate of the date of formation of the first stars (a couple of hundred million years after Big Bang—not 6 to 8 billion years).
A previous comment by Stephen Demontmollin noted that the date of the first stars was a couple of hundred million years after the Big Bang. The European Space Agency (ESA) Plank Satellite reveled that the first stars in the universe started forming later than previous observations of the Cosmic Microwave Background (CMB) indicated. This new analysis shows that these stars were the only sources needed to account for reionising atoms in the cosmos, had completed half this process when the universe reached an age of 700 million years. In either case this is far from the comment made by Priya Natarajan in the interview:
0:59:57.1 PN: Yeah. We don’t know quite precisely. There’s a little bit of uncertainty because we don’t know exactly when the first star likely formed, but probably about 6-8 billion years ago.
Hopefully an explanation for the discrepancy will be given.
BTW great interview hope to hear more from Priys in future episodes.
Terrific episode, thank you Sean and Priya!
I’d love to hear more about how updates from observations make their way into the models.
Near the end of the interview Priya Natarajan brought up the topic of gravitational lensing: a phenomenon by which light rays are deflected as they transverse through curved space caused by the presence of massive astrophysical objects. These massive astrophysical objects, like galaxies or clusters of galaxies, located between the Earth and even more distant galaxies sometimes cause the image of these more distant galaxies to become magnified, distorted, and even to produce multiple images of the same galaxy. This provides a way to estimate the amount of so called ‘dark matter’ present in the galaxies which are functioning as the lensing mechanism. The problem is that the lensing effect of the intervening galaxies is up to 10 times more efficient than predicted by most cold dark matter models. It just emphasizes the fact the name ‘dark matter’ was justified, not only can we not see it, we are still very much in the ‘dark’ about what it actually is or how it behaves.
The 2014 Guardian article ‘Universe recreated in massive computer simulation’ containing the short video ‘A Virtual Universe’, posted below helps give a feel for how computer simulations can be used to help scientist test how well their theories of the universe work, by comparing the appearance of the virtual cosmos with observations made using telescopes and other instruments.
https://www.theguardian.com/science/2014/may/07/universe-recreated-computer-simulation-model-big-bang