Unravelling the Universe, Again

Every scientist dreams of making a discovery that fundamentally changes our understanding of the universe. My guest today, Nobel Prize-winning astrophysicist Adam Riess, did exactly that. What’s even more remarkable, he may just be in the process of completely upending cosmology for a second time.

RIESS: In the laws of physics, things really should match. It’s not okay for things to be off by five times the margin of error of your experiment. In fact, it’s not just not okay, we get very excited.

Welcome to People I (Mostly) Admire, with Steve Levitt.

*      *      *

 To understand Adam Riess’s contributions to cosmology, one needs to know what scientists believed prior to his findings. So I began our conversation today by asking Adam to describe the accepted scientific consensus on the expansion and contraction of the universe back in 1992, the year he started his Ph.D.

RIESS: Well, I would say for the better part of a hundred years — this goes back to Edwin Hubble and others of his time. Cosmologists, astronomers had measured that the universe was expanding just by measuring objects around us, the distance to galaxies, and the apparent speed at which they were moving away from us, something called the redshift. In the 1960s, cosmologists had discovered the radiation leftover from the Big Bang. So it became even more clear that the universe was expanding after what we call the hot Big Bang. Of course the big question then was, What is going to happen next? What is the fate of this expanding universe? And so it really comes down to the question of how much matter and gravity from that matter is there that can be slowing the expansion; like, if you launched a rocket off the surface of the earth, you would wonder, Is that rocket going to escape the earth’s gravitational pull? Or will it fall back? And it’s really a question at that point of velocity and mass. By the 1990s, the big question was, Is there enough matter in the universe to stop the expansion in the future? And they were finding not enough matter to stop the expansion.

LEVITT: So, essentially, the way cosmologists were thinking about it at the time — this push that was making the universe expand, that was a one-time explosion that came with the Big Bang. And then what you’re saying is that gravity is kind of tugging in the opposite direction, but gravity doesn’t tug once, it tugs over and over and over to infinity as long as there’s matter. And the longer the time passes, the more the tugging starts to outweigh the initial push. And then the question was just, Is the universe’s expansion slowing down or is it actually going to reverse and start to contract?

RIESS: Right. Is it slowing down enough that it would eventually stop the expansion altogether?

LEVITT: You said that we know that the universe was expanding and it was Edwin Hubble, who the telescope was named after, who was the first one who showed that. Could you talk a little bit about how we proved that? Because I think that’s going to be really relevant to thinking about what you were doing many years later.

RIESS: Sure, yeah, that’s certainly a key element. How do we know that the universe is expanding? And the answer is, We look with telescopes at distant objects. And if the universe was static, if it wasn’t expanding or contracting, then you would look at the objects around you and, yeah, maybe some would be moving towards you and some away from you because they’d have some motion that might be due to some nearby galaxy they’re being pulled to or away from. But, on average, as you look around those motions would be equal. And instead we see something very different. As you look further out, greater and greater redshifts — greater and greater motions away from us. And that is something we can only understand as the universe itself is expanding. Just to use an analogy, If you were at an airport, you might see people walking to and fro, some towards you, some away from you. That’s just the way it is. But if you saw a bunch of people who were all moving exactly together at the same speed, right? You might infer They’re on one of those people movers. There’s some sort of global story here that is making them operate together. And that’s actually what we see about the universe. When we look around, everything moves away from us. The further away, the faster, as though the universe itself is doubling in size every so often. And in order to make those measurements, we have to measure how far away things are, and that’s actually the most challenging aspect is just figuring out how far away things are.

LEVITT: Okay, so to figure out whether the universe is expanding, first you have to know how far things are away from us. And then you mentioned redshift, but you got to explain redshift to me.

RIESS: Sure, absolutely. Whenever we are receiving information from a distant object, say, you’re listening to an ambulance as it’s moving. It’s emitting some periodic phenomenon, sound in this case. And depending on whether that object is moving towards you or away from you, the pitch of the sound, the wavelength of the sound will change if it’s moving away from you, each crest of the wave will be emitted a little further away. It’ll elongate the wavelengths that you receive, and we call that the redshift, especially when it applies to light instead of sound. So if a distant galaxy is moving away from you, it’s emitting light, which is, a wave-like phenomenon with a certain wavelength that explains the colors that we see.

LEVITT: Can I ask you about the ambulance again? My entire life I’ve heard about the Doppler effect, but I actually never thought about it. So what’s actually happening is sound travels in waves. And so the ambulance starts to emit the sound, but by the time it’s done with emitting that wave, it’s actually moved, say, further away from me. And what makes the sound low pitched or high pitched is how long it takes the wave to oscillate. And because it’s moving away, it actually stretches it out a little, everything’s elongated by the fact that the ambulance itself was moving. I never really thought about why it was. Okay, that’s interesting.

RIESS: The ambulance is really boosting that wavelength to a longer wavelength.

LEVITT: And the longer the wavelength, the lower the sound.

RIESS: The lower the sound, right.

LEVITT: But the analogy to what you’re calling redshift, so now you’ve got a star way far away from us. 

RIESS: A star or a galaxy. It’s emitting light, maybe it’s emitting a very specific wavelength of light that corresponds to a very specific color. But, in this case, it’s not the galaxy that’s moving away from us, but rather the space between us and the galaxy that is expanding. And so, it’s as though the galaxy is rushing away from us and just like that ambulance, that motion is boosting or increasing the wavelength of light as we receive it.

LEVITT: And the longer it’s out there, the longer it has to go to get to us, the more and more it gets elongated and the more this redshift happens.

RIESS: And it’s the entire trip that the light makes on its way from me to you — that wavelength keeps getting longer and longer.

LEVITT: Okay. So now with that background, what did Edwin Hubble actually do — what calculations did he actually make that were super convincing? Like, basically nobody thought the universe was expanding, and he made these observations, and then everybody agreed it was expanding.

RIESS: If you look out at objects away from us, you might see motions, you might see little redshifts just because objects are moving around throughout space. But instead, what Hubble saw was a very specific pattern, the unique signature, if you will, of the expanding universe. And that is that not only did distant objects have redshifts as though everything is moving away from us, but the further away they were the faster or the greater the redshift was. A useful analogy might be, imagine a loaf of raisin bread rising in the oven, right? So you’re sitting on a raisin and you look at the other raisins around you, and they all appear to be rushing away from you. And the further away a raisin is, the faster it looks like it’s moving away because there’s more dough between you and a distant raisin. So if that raisin bread loaf is going to maintain its proportions, you will see this pattern further away raisins moving away even faster. And so, when Hubble grafted the distance of galaxies and this redshift or this apparent motion away, he saw a relationship between the two. And this is known as Hubble’s Law, but it’s the fact that the universe is expanding that produces this kind of signature.

LEVITT: Okay. But one thing we haven’t talked about yet is how does he know how far a galaxy is away from us? That is a hard problem, right?

RIESS: That is the central problem of the field of cosmology, is just, How far away is stuff? The reason it’s hard, maybe it’s useful to think about, How do we do this on Earth? How do you figure out how far away something is? We have many ways to do it. The best way, many of us use a tape measure, you actually go out to the object and stretch the tape measure. So if you can’t actually go to that object, you can very carefully observe that distant object, let’s say, relative to something even further away like a mountain. So maybe you’re looking at a tree compared to a mountain, and then you walk some certain distance and you take a second sighting and you see that tree move through some angle. And so if you measure the distance you walked and you measure that angle, you have two out of three parts of a triangle in space. And so you can use trigonometry to solve for the length or the distance to that object. 

LEVITT: And we do that every day with vision, right? ‘Cause our eyes aren’t right next to each other. And so we use that kind of parallax everyday. 

RIESS: It’s actually the reason that our eyes are separated is to give us that power. But, the problem becomes what we call the baseline, the distance between the two observing points is not large enough, then that angle becomes imperceptibly small. So I can’t use my two eyes to figure out how far away a star is because the parallax is imperceptible it’s so small.

LEVITT: And so are we able to use the motion of the earth around the sun? Is that a big enough swing? 

RIESS: We do. That’s about the best baseline we get is you can look at a star in our galaxy, let’s say, in January. And then you can look at it again in June when the earth has moved to the other end of its orbit. And, if that star is near enough, thousands of light years away, then you can detect that change in angle and measure the distance. But, unfortunately, we’re interested in the distance to galaxies, which are much, much further away, where the parallax angle becomes impossible to measure, it’s so small. So the next technique we use is the method of a lighthouse, right? When a ship captain looks at a lighthouse, they’re trying to gauge the distance to the lighthouse to make sure they’re staying away from the rocky shore. And they can do that based on the brightness of the light from the lighthouse, right? They understand that a lighthouse is very luminous, and so depending on how bright it appears, they have a good feel for about how far away they are. 

LEVITT: Okay, so in principle, the guy who’s in the boat knows that all lighthouses are the same brightness, but our problem is that the stars and the galaxies are everything…

RIESS: Right, now you’re seeing the problem, which is, we understand what a lighthouse is. We manufactured them. And we know that they’re all pretty uniform, but the stars have an enormous dynamic range. They come from hundreds of times fainter than our sun to hundreds of thousands of times more luminous than our sun. And not only that, when we look at distant galaxies, we can’t even pick out individual stars. All we see is the collection of light from maybe 10 or a hundred billion stars. We don’t even know how many stars. You’re seeing the problem, which is that up until a certain point in the history of this subject, we didn’t have a really good way to measure distances.

LEVITT: Okay, so how do we solve that?

RIESS: Yeah. Before Hubble made his important discovery, Henrietta Leavitt, who was an astronomer, what was called a computer at the time, at Harvard College Observatory——

LEVITT: No relation I will say. I wish I could say I was related to her, but I am not.

RIESS: She was studying photographs of stars in a certain cloud, called the small Magellanic Cloud in the Southern Hemisphere, which is actually outside our Milky Way. And she noticed that there was a class of stars that were changing their brightness every few weeks or months. And she found something really amazing, which is that the more luminous they were, the longer the period was between changing their brightness. This is what’s called the period luminosity relationship. And so, this is the real breakthrough that was needed. Because when you see one of these stars, which she recognized as very like a nearby star in our own Milky Way, Delta Cephei, which also does the same thing — what are called Cepheid variables — these give you a great rule for figuring out how luminous that star is and therefore, how far away it is. 

LEVITT: Again, this depends on knowing that they’re the same distance away. How did she actually know that all these stars she was looking at were roughly the same distance?

RIESS: She made a reasonable and correct assumption that because they were in this cloud, well outside the Milky Way galaxy, that their distance was all about the same as from us. It would be like looking at a bunch of people in a car traveling down the highway. The different people might be a little closer or further to you, as a whole, they’re essentially the same distance from you.

LEVITT: Okay, so now that key piece is we have this class of stars, the Cepheid variable stars, where we’ve unlocked their distance by putting together her insights, we now can pretty reasonably say how far those stars are away. And that’s what Hubble then did was to take that piece along with this redshift piece and to build that graph you talked about, that showed that the further away that the Cepheid variable star was that he looked at the more redshift in the light that it was sending.

RIESS: Correct. Just to be clear, these stars are then in galaxies and it’s the entire galaxy which has that red shift in that distance, but yes. 

LEVITT: This all has something to do with Albert Einstein’s cosmological constant, right? What he described as his greatest blunder, because he, for some reason, just assumed that the universe had a fixed size. And when he wrote down his theory of relativity, it didn’t work without just putting in a fudge factor. And then, when Hubble showed that the universe was expanding, Einstein was really embarrassed by that.

RIESS: Correct. So a decade before Hubble discovers that the universe is expanding, astronomers still think that the universe is just our own Milky Way galaxy and that the other galaxies that we see are somehow contained within the Milky Way. And Einstein is struggling to understand how to apply his new laws of gravity, general relativity, to the universe as a whole. And a very key condition is for him to understand whether the universe is in motion or not. So he asks astronomers of the day, Is the universe expanding or contracting? What’s it doing? By their not knowing that the galaxies were outside the Milky Way, they focused more on the stars in the Milky Way. And they said, “Yeah, it’s static. They’re moving this way and that way, but, there’s no real outward or inward motion.” So he thinks, That’s pretty confusing. Because if you took a collection of objects, like, galaxies or stars, and if they were not moving, if they were not expanding or contracting, then gravity acting on them would cause them to start contracting. So what is overcoming the tendency for a set of objects to start collapsing by gravity? And he made this amazing discovery really, that although the gravity of stuff is attractive, the gravity of empty space itself could be repulsive. Something he called the cosmological constant, today we would call dark energy. And he imagined that these two kinds of gravity, the attractive gravity of stuff and the repulsive gravity of empty space must have been set into a perfect balance. Over ten years later, Hubble shows that the universe is expanding and Einstein thinks, Ugh, this is a terrible blunder I made. That wasn’t the right condition and therefore I invoked this stuff, which could exist, but is unnecessary. And in physics, anything that’s unnecessary is frowned on. You like to shave with Occam’s Razor, use the simplest explanation.

LEVITT: So now let’s fast forward finally to you. So we go from 1929 when Hubble shows that the university is expanding to 1992 when you’re a budding graduate student. And now the question that somehow you’re already onto before you even get to your Ph.D. program is you’ve decided you want to measure whether or not the universe is merely slowing down or whether there’s enough matter to actually make it collapse on itself.

RIESS: Right. The Cepheid variables that astronomers were using, they’re great, but there’s a limit how far out you could see them, which is not as far as we need to be able to measure, because in order to measure if the universe is slowing in its expansion and how much, you have to look at how fast the universe is expanding today by looking at objects all around you, but then you have to look much further out so that you are effectively looking back in time because the further away something is, the longer it takes for light to reach us from that distant object. What seems kind of annoying, that the further out you look, the more delayed your information is, actually is the superpower that cosmologists have to be able to compare the past to the present. And so in order to make use of that, you need to be able to measure vast distances, billions of light years. And it really means you need a much more luminous object than a super giant star, like a Cepheid variable. You need something more like a supernova, an exploding star.

LEVITT: So as you said, you need to look really far away to figure this out. And Cepheid variables you can’t see the ones that are far enough away that would be helpful. So you needed something else. And so how in the world did anyone come to understand that supernovae would be the way to do this?

RIESS: Humans have been seeing supernovae for millennia. Sometimes it would just appear as like a new star in the sky because you couldn’t see the star until it blew up. A supernova is billions of times more luminous than the sun, but only for a short while, you know, weeks or months. And then over the last century, astronomers started realizing, Wow, perhaps they could serve as what we call standard candles. A kind of an object whose luminosity is so great but uniform enough that you can gauge its distance from its brightness.

LEVITT: Okay, but why would anyone expect that explosions would have the nature of being the same over and over? It seems like of all the things you would be suspicious of, (AR^Right, yes) it would be that every explosion would be the same.

RIESS: And that’s a great question. That really is the key question here stars come in a wide range of masses and luminosities to begin with, why would the explosion be the same? And it’s important to understand, but wasn’t really understood until the 1980s and ‘90s, that there’s really two classes of supernovae. There’s the kind where massive stars collapse and rebound, and indeed those come in a wide range of luminosities. But there’s a special class which came to be called type 1a, that go back to an idea the famous Indian astrophysicist, Chandrasekhar, showed in the 1930s that a certain kind of star called a white dwarf star can only exist, could only be stable below a certain critical mass, which came to be called the Chandrasekhar limit. Work for which he won the Nobel Prize in the 1980s. And that’s about 1.4 times the mass of our sun. So if a star exists in that state, the white-dwarf state, and is below the Chandrasekhar limit, it can be happy and go on shining. But if it has a friend that lives nearby, in a binary, then the friend may donate matter or the white dwarf may pull matter by gravity onto it until it slowly accretes, until it crosses the Chandrasekhar limit, and then explodes. And so that’s the kind of supernova that’s very uniform, because it’s always blowing up at or near about the same mass. 

LEVITT: Can I say, this is the thing about physics that blows my mind. That the theory——

RIESS: No pun intended, but yeah. 

LEVITT: Yeah, the theory is so amazing and precise. This guy, when’s he doing this? In the ‘30s? 

RIESS: Yes.

LEVITT: In the ‘30s, this guy is writing down with pencil and paper some model about stars and comes up with this idea, which 50-years later, it turns out to be exactly right. That never happens in economics. We never have that kind of precision and it’s awesome how the theory can sometimes unleash this insight that drives the empirical study. So, anyway, I just want to point that out because that’s really, truly an exceptional thing. And because these things are so bright, if a type 1a supernova happens, how far away from us can it be that we’ll still see it?

RIESS: By the 1990s, astronomers had access to four- and 10-meter class telescopes and also to new cameras that had lots of detectors on them that would allow you to image wide swaths of the sky. It sounds great, Oh, you have a type 1a supernova, just find some and measure how far away they are. They only occur in a galaxy like ours about once a century. And so if you just pick out your favorite galaxy and stare, you know you’re going to be waiting a very long time. And so the additional breakthrough is astronomers needed these wide-field cameras, and it was the invention of the C.C.D., the charge coupled device that’s in all of our smartphones now, where they became cheap and accessible and you could cover large areas with them. So now you can image regions of the sky where each single image contains tens of thousands, even hundreds of thousands of galaxies. And so then you take another image a month later and what seems like an incredibly unlikely event to find a supernova in a galaxy when it only goes off once a century. You now are staring at tens or hundreds of thousands of galaxies at once. It’s sort of like winning the lottery because you bought all the lottery tickets. And so in this case, we could find them regularly and we could find them at distances of a few billion light years, which means we are now looking back in time, billions of years, to how fast the universe was expanding in the past.

LEVITT: Can you give a sense of how bright they are on Earth when they’re these billions of light years away?

RIESS: Yeah, they’re about a trillion-times fainter than something you could see with your naked eye.

LEVITT: Wow, okay, so really looking for a needle in a haystack because it’s not like they’re jumping out and saying, “Hey, I’m here! I’m here!”

RIESS: But on the other hand, if one went off in our Milky Way, as for example, the Chinese saw in 1054 A.D., it was about as bright as the full moon. You could easily see it during the day.

LEVITT: Okay. So trying to put together the pieces, this all sounds really easy now you’ve got these supernovae and they’re hard to find, but once you find them, they are standard candles, so you know how far away they are. And you can measure the redshift of the life in the galaxy, and it seems like it’s just a very simple extension of the techniques that Hubble was using, just on a massively larger scale. But in fact, it’s really super complicated to sort out all of the mess in there, right?

RIESS: So, the concepts are, as you say, very simple and straightforward. But the reality, doing it in practice can be more complicated. For example, we didn’t mention that the supernovae is that standard candle when it’s at its peak. So it explodes and it’ll reach its peak a couple of weeks later. And then it’ll fall. And so you have to catch it before its peak so that you see it go over its peak, so you can say, “Oh, okay, I’ve now seen the peak brightness of that actual supernova.” So there’s an incredible sort of time race going on during this process. Then you have to pick out which ones are the type 1a from the other types. So you have to look for a certain spectral fingerprint. They’re not all exactly the same. Some are a little bit brighter than others. There’s additional clues in the rate at which their light rises and falls. This is something that I worked out in my thesis. And then, supernovae live in galaxies and galaxies have dust, and dust can block the light and fool you into thinking something is further away than it really is because it looks faint. It’s like looking at a lighthouse at night, but if it’s a foggy night, you could be fooled into thinking it’s further away. So you have to account for how much dust is in that distant galaxy because dust not only blocks the light, it shifts the colors of the light. Like when you look at a sunset and it looks very reddish because the light has been filtered through the dust in our atmosphere. 

LEVITT: So, there are two very different talents required to make this work. One is all the stuff about the telescopes and finding and identifying the supernovae. And the second is doing the algorithmic work, which takes the observation, the very noisy data that you are given from the telescope and extracts the truth from it. And what I understand is that one of your great contributions is devising these really smart algorithms that back out the signal from the incredible amount of noise that’s coming out of these initial images.

RIESS: This goes by a very prosaic name, we call this data reduction. But really, if you imagine you have billions and billions of pixels on the sky taking images, and at the end of the day, you have maybe a yes or no question to answer. And all those billions of pixels will weigh in on the answer to that question. And so, that’s to me one of the funnest parts of science.

LEVITT: It really sounds like it’s almost an art form. The possibilities of what you might do in the algorithm is really wide and it takes this sort of artistic touch to pull these pieces together into something that makes sense.

RIESS: I mean, certainly I would say the noisier and rougher your data is the more creativity is required. We are almost by definition dealing with data that’s just barely of the quality needed to answer the question, maybe even not quite because, you know, we’re usually ambitious and we’re trying to get to big questions you know, maybe before Moore’s law or the equivalent for telescopes is caught up. And you train and learn on easier to measure problems. And then, in some cases you are truly extrapolating what you know or how you know. One thing we worried about a lot at the time was, even if we have controlled and understood everything about the observations, what if the supernovae themselves were fooling us? What if, when the universe was younger, it produced supernovae that weren’t as bright? It would be like imagining the manufacturer of lighthouses has been slowly, over time, changing the specification of the lighthouse. And so those are things we had to tease out.

We’ll be right back with more of my conversation with astrophysicist Adam Riess after this short break.

*      *      *

LEVITT: Okay. So in 1998, so that’s only six years after you started graduate school, you and the other members of your team published this paper that literally rocks the world of cosmology. So what did that paper find?

RIESS: Yeah. So our team, the High-Z Supernova team, had collected our first large set of observations of very distant supernovae that could tell us how fast the universe was expanding in the past and compare that to how fast it’s expanding in the present. With the expectation being that the expansion would be slowing down, because of the attractive gravity of all the stuff. And the big question being, Is it slowing enough to stop the expansion? And so, when I made that measurement and cross-checked it with my colleagues, we found it was not slowing down a little. It was not slowing down a lot. It was not slowing down at all. It was actually speeding up. The universe is expanding faster and faster. It’s accelerating, which was not expected, and the only way we can understand this is go back almost a century to Einstein and this discovery he made that gravity can operate in the other direction as well. It can be repulsive, and in fact, the gravity of empty space could be repulsive if it has energy in it that could cause this acceleration. And so, we wrote this paper, saying that the expansion of the universe is speeding up and that the only way we could understand it was to associate it with Einstein’s idea of a cosmological constant. Or, as my colleagues quickly recognized, something even more general that we called dark energy.

LEVITT: You say this now, with the benefit of hindsight, like it makes sense. But I just want to emphasize, as far as I understand, this was completely and totally unexpected even by you, right? That nobody, when you started this project, ever would’ve thought this is what you’d find. Is that a fair statement?

RIESS: So I would say like a lot of times in science, the theorists are very creative and they’ve imagined all kinds of possibilities, including this possibility. But the observers are much more practical. And I was definitely living in that world. So it was like, Yeah, there’s gravity and we’re not even thinking about this cosmological constant. If Einstein made a giant blunder with that, who are we to even think about it anymore? So, you throw your keys up in the air, they fall back down. That’s what gravity does. And the question is just, How much gravity is there? And so, I would say to us, the observers, at the time this was a shock because this was not the direction that we were expecting.

LEVITT: So I’m just trying to imagine the feeling you had when you first saw the results. Was there a particular eureka moment where you understood the implication? Or was this something that slowly unfolded and built with greater and greater confidence?

RIESS: I would say the primary emotions were fear and anxiety. Because I was only a year or two out of graduate school and I had already had the experience that nine times out of 10, your first answer is wrong because you’ve made some mistake in the analysis because the analysis is complicated. And so I spent weeks going back over the steps that I did looking for the mistake or the bug checking my simulations, and I couldn’t find one. 

LEVITT: Who’d you tell first?

RIESS: Well, I talked to the other members of my High-Z Supernova team, particularly Brian Schmidt, who I ended up sharing the Nobel Prize with, but at the time, was the leader of our team and was responsible for finding the supernovae, which was really a hard part of the problem.

LEVITT: What did Brian say when you told him? Do you remember?

RIESS: It’s the usual thing, sort of like, Well, Brian, you know, I’ve clearly made a mistake ’cause the sign’s wrong, but I can’t find it, so let me just send you pieces at a time and you could check step A to step B. And finally, the last step of the analysis, which, Brian at that point lived in Australia and I was in California, so there was this very painful — you have to wait all night to get an answer back. And I went to bed that night and woke up the next morning with a message from Brian, “Well, hello Lambda.” Lambda being the symbol for Einstein’s cosmological constant.

LEVITT: He had enough faith in you to think that this was probably right.

RIESS: Well, I think we all had faith in empirics. This was an experiment, this was a measurement to do. This wasn’t how I felt or how he felt, or the rest of us felt about the universe. We were going to report what the data showed and, I don’t think that any one of us trusted each other as much as we would say, “I trust this process. You do the calculation, if I get the same thing, then that’s probably right. ‘Cause it’s unlikely we both made the same mistake. Let’s get a third person to check.” And, we would go through this until we reached that point. And I should also say there was another competing team, at the same time, doing much the same experiment. So only a couple of months later, both groups started talking about the results at conferences and it was like, Oh my gosh, you guys are seeing what we’re seeing. No, you’re seeing what we’re seeing. These are different people using, their own telescope time, observing different supernovae. But seeing the same thing. And then that’s where the community started saying, “Oh, whoa. This looks real.”

LEVITT: So your paper gets published. Was there an immediate reaction? I’ve got to think that most astrophysicists would be quick to dismiss the results as being wrong, just because they were so unexpected.

RIESS: Yeah, I think that is the nature of science. Science proceeds slowly, in some ways. Scientists are skeptical. And of course, lots of exciting new results you hear about are wrong. Because that’s the nature of work by humans. But on the other hand, you don’t want to become so cynical that you say, “Everything that I hear that is surprising is wrong,” otherwise you have no possibility to make a discovery. What that meant in practice was about half to two thirds of the community believed it, and about half to a third didn’t. But it was believed enough. It was the breakthrough of the year for Science Magazine in 1998. It was several years later when observations of the radiation leftover from the Big Bang were able to measure the state of the universe in a very different way and came to much the same conclusion — that about 70 percent of the universe is in the form of this dark energy, this kind of energy of empty space causing the expansion to accelerate.

LEVITT: So scientists have given dark energy a name, and by naming it dark energy, to an outsider like me, it gives, what I think is the illusion that you actually understand it, right? 

RIESS: That is a good thing about names.

LEVITT: I know what dark is. I know what energy is. But isn’t it really fair to say that more or less, this is just the mathematical fudge factor that you need to stick into the existing set of equations to salvage a model that otherwise has splendidly predictive power?

RIESS: I’m going to tell you sort of glass is half full, half empty aspect of this. We are making use of the same theory of physics Einstein gave us, general relativity. And a place in that theory that says, By the way, how much energy is there in empty space? ‘Cause that’s going to have repulsive gravity. And you could guess it’s zero. Okay? Or you could observe the universe and see it accelerating and say, Oh, I guess it isn’t zero. I guess it’s this amount. So that sounds much less made up in a way that just says, “Oh, there was an algebraic term here, and we didn’t know what its value was.” So previously people assumed it was zero, but it’s not. Now, that’s the glass is half full. The glass is half-empty part of the story is, And what is that stuff? What’s the physics of that? So we turn to our particle physics friends and say, Should there be energy and empty space? And they say, Yeah, by quantum theory, actually, we do expect that. We say, Great, how much should there be? And they get an answer that is about 120 orders of magnitude more than what we see, the energy density of what they see.

LEVITT: Has there ever been a worse prediction than that? I have never heard of a prediction that is so far off. 

RIESS: It’s often called, “the worst prediction in all of physics.” When you ask particle physicists just to do the naive, face-value calculation to add up all the possible energy states of empty space, they will get an answer that is so big that the universe wouldn’t even be here because the acceleration would be so great. It would push things apart before gravity at any chance to pull anything together so there wouldn’t be galaxies and stars and planets and everything. So that problem has been around a long time. Particle physicists would often say, Well then maybe it is zero because it isn’t that number. You got to be honest at that point and say, “Well, I guess we don’t really understand the physics of the quantum vacuum and how to relate that to gravity,” which is absolutely true. Two of our best theories in physics are quantum theory and gravity, general relativity, and they’re not compatible with each other. In some ways it’s awesome to do the experiment in nature, ’cause nature doesn’t accept an idea of two theories being incompatible. It knows how to do physics, anyway. And so this is where we do experiments and hopefully it teaches us something. And so what it taught us so far is there’s something interesting going on at that interface and we really don’t understand it.

LEVITT: Now, what’s so crazy to an outsider like me is that all the stuff that I experience in the world, all of the matter and the planets and elements, in this new conception of the model, the things that we can actually see, only add up to 4 percent of all the stuff. Which seems like a crazy model of the world. We have a model of the world we think is awesome, in which 96 percent of what’s there, we’ve never observed any of it.

RIESS: That’s right. I almost want to go back to my chemistry teacher in high school and say, “Why didn’t you give us a disclaimer or something? This whole periodic table of elements we’re going to spend so long studying and understanding. You’re going to kind of hint that this is everything and instead that’s 4 percent. and the other 96 percent would have its own story. Wow.”

LEVITT: Okay, so this serves as a great segue into the work you’ve done more recently, which has, once again, upended the world of cosmology. Could you describe what has come to be known as the Hubble tension?

RIESS: Sure. So after this work in 1998, and follow on work a few years later, observations of the cosmic microwave background, I would say a new model or paradigm was born in cosmology, that goes by the maybe not so sexy name, Lambda-CDM. It’s largely descriptive. The lambda part says, There’s dark energy. Let’s guess it’s the simplest form it could take, like what Einstein said. And C.D.M. stands for cold dark matter, which says there’s a lot of matter in the universe ’cause we could see it’s effects on galaxies, but it’s not emitting light either. So we’ll call that dark matter. 

LEVITT: When we say dark matter, we’re talking about some other thing that plays by different rules than our regular matter, right?

RIESS: Yeah, by the process of elimination and other observations, we’ve determined that it’s very likely to be a particle. It’s matter. But it doesn’t emit light and importantly, it doesn’t seem to have any other interesting interactions like chemistry interactions or decays or the kinds of things that could cause light to be emitted. It seems like in some ways the most boring particle in the world, but it seems plentiful. In fact, makes up most of the matter, 10-times more than the luminous matter. And without it, we wouldn’t understand how stars can whiz around the galaxies they’re in and not just be flung away. There’s more gravitational glue from matter in the galaxies.

LEVITT: And why can’t we find it? If it’s ever all around us, why can’t we find it?

RIESS: Yeah, because, it’s so reticent to interact with anything that makes it also very difficult to detect in a laboratory.

LEVITT: Okay.

RIESS: So by the late ‘90s and early 2000s, we had this great new model Lambda-CDM. And it is impressive in that it fits a lot of data. So it’s incredibly powerful just to have a model, just to organize your understanding of things and to use it to make predictions. And the sort of grand, what I’ve been calling, the end-to-end test of the universe, is to then look at the universe shortly after the Big Bang and the radiation left over from it. And that information makes a very specific prediction of how fast the universe should be expanding today given the accuracy of this model. And so we go out then with even greater precision now using not just the Hubble Space Telescope, the James Webb Space Telescope, a new European space telescope called Gaia. And we can make very precise measurements of how fast the universe is expanding. This is a number called the Hubble constant. And so, this has become a cornerstone of demonstrating how well this model works.

LEVITT: And so just to be clear, the Hubble constant differs from what you were doing back when you won the Nobel Prize, because this is trying to talk about what the universe is doing today. Whereas what you were focused on before was trying to use information from way back in time to see how the universe is changing. 

RIESS: Right, right. The work we did in the 1990s was really about measuring acceleration. Now we’re measuring velocity, we’re really measuring the actual speed of the expansion of the universe, but to great precision, or we’re trying to, and comparing that to what very precise data shortly after the Big Bang told us about the state of the universe, and then using this model, Lambda-CDM, to connect the two eras of the universe, the beginning and the end.

LEVITT: Okay, let me make sure I’ve got this. So there’s some group of researchers out there. They’re looking at this thing, the cosmic microwave background. And somehow through the genius of humanity, we’re able to understand a whole lot of stuff from this incredibly faint signal from 13 or 14 billion years ago. And you’re saying these physicists, they take this information from 13, 14 billion years ago, and they say, Given the model you talked about, Lambda-CDM, we have a very precise prediction of this thing you’re calling the Hubble constant. And they get a number. Okay. And that number is what?

RIESS: Sixty-seven plus or minus 0.5.

LEVITT: Okay. And then you say, “Well, what if I just go and use my techniques to just look directly at what’s happening around us today?” So none of this model intervening. You’re just doing an observation, you’re a very direct measure of what’s going on. And you do this magic stuff that you do that involves the supernovae and——

RIESS: Sure. All the way back to parallax and things like that. 

LEVITT: Okay. Okay. And what do you get?

RIESS: Yeah, so my colleagues and I have been measuring this more and more carefully with a series of space telescopes. And we’re sitting today at about 73 plus or minus one. So 73 plus or minus one is not the same as 67 plus or minus 0.5. Now you could say, “Eh, you guys are doing pretty good. You got the sign right. You’re same order of magnitude.” 

LEVITT: You’re not 120 orders of magnitude off. 

RIESS: You’re not 120 orders of magnitude off. Eh, you differ by 10 percent. Hey, we can get a pat on the head and a glass of milk and be pretty happy with that. But, we’re curious about the universe and we say, “This doesn’t match though.” And in the laws of physics, things really should match. It’s not okay for things to be off by five times the margin of error of your experiment. In fact, it’s not just not okay, we get very excited. This is an opportunity. This may be teaching us something about the universe. How come this model that’s worked so well, we go to the next level of precision and it’s not quite fitting, it’s straining. Maybe, and we don’t know, maybe there’s something not totally vanilla about this dark energy or dark matter or even the laws of gravity. This is the way science works, is you keep upping your game in terms of the precision, the fidelity of experiments. And ultimately you see something that required higher resolution to see.

LEVITT: So I think that this Lambda-CDM model, it assumes that the amount of dark energy has remained constant over the life of the universe. And it just seems strange to me, given that we hardly know anything about dark energy at all, that anybody would have any confidence that this is the kind of thing that would just stay constant for billions of years.

RIESS: Yeah, well, as I said, physicists like simplicity, they like elegance, and this idea that it would stay constant has a certain simplicity or elegance to it, but it doesn’t come from any real fundamental understanding of dark energy. And in fact, there’s another story, which is that dark energy could be just the temporary energy of a field in space. So you could think of any field, like a magnetic field or an electric field. It has energy associated with it, but that field itself is usually evolving or changing in some way. So that amount of energy or the energy density is changing as well. Which of those stories is right? We don’t know. There have been other observations even more recently, that suggest that the energy might not be constant. 

LEVITT: So is it your view looking at these data and others, I suppose too, that we’ve got the wrong model of the universe that we need something along the lines of the leap that was made when Einstein’s model replaced Isaac Newton’s model?

RIESS: I take the sort of trust but verify approach to cosmology. We do this empirically, we have a model, that’s a good thing to have, but it doesn’t fit everything. And that’s very important to understand. I think that the data has gotten very good. All the previous results I talked about were often based on data from the Hubble Space Telescope. A few years ago we launched the James Webb Space Telescope. Far more powerful, higher resolution, and yet it seems pretty much that the same discrepancy exists, this Hubble tension. And I think we have to take it very seriously as a clue. Are there other ways the universe could be, other models, other stories, other physics that would fit the data better? This is the way science proceeds is we have a paradigm, we have a model, and then cracks start to appear. And, hopefully those cracks provide clues if the model is inadequate of how to improve, change, evolve the model. Maybe it’s a small tweak to the model or maybe it’s something revolutionary. Sometimes you pull a loose thread on a sweater and you pull off that thread and other times it could unravel the sweater. So I don’t go in with a preconceived notion about it. I think, my part of this job is just to make the best measurements that I can and communicate those.

You’re listening to People I (Mostly) Admire. I’m Steve Levitt. After this short break, astrophysicist Adam Riess and I will return to talk about what a new model of the universe might look like.

*      *      *

 The first time Adam Riess challenged the standard model of the universe, adding dark energy was the fix. I’m curious to hear how Adam thinks the theory will need to be altered to encompass his latest findings. But before I get into that, I wanted to ask him a more personal question. Coverage of the Hubble tension has spilled over into the mainstream media, which has suggested that underlying the scientific debate is a highly charged personal battle. I asked Adam how he felt about the way he’s been portrayed in the media.

RIESS: I don’t read media descriptions of the work that I do because they often try to assign a kind of story to them. It’s hard to relate to dark energy and dark matter, so it’s easier to relate to this person and that person. I think that we live in a very complex, divided, polarized world. Science operates pretty well, scientists in that regard. We take data, we publish it, we discuss it at conferences. We don’t always agree. We don’t always see things the same way, but that’s part of the process as well. Eventually we move forward. The universe doesn’t care what we think about it. And we shouldn’t care what specific people think about it either we should really care, in this case, I think about the data. And I think that’s the only way forward.

LEVITT: So let’s just say the standard model turns out to be wrong, has to be rewritten. Are there any implications of that for anybody other than cosmologists?

RIESS: I think because the standard model is held up by deep laws of physics, like quantum theory and general relativity, our understanding of gravity. Then if there’s something that is wrong about the model, there’s the potential that it’s actually teaching us that there’s something wrong about the theory under the model. Even some of these physical laws, which, if that’s the case, in many instances it’s all bets are off. There’s a lot we could potentially learn. It could be anything from satisfying our curiosity to really understand the universe to practical things. You look back at when Einstein developed general relativity, new theory of gravity, and you might’ve thought at the time, Oh, well, Newton’s theory was good enough for everybody. Well, not really, because the GPS system that we all use in our phones and our cars, it’s precision. What allows you to not be off by kilometers every day is the differences between Einstein’s general relativity and Newton’s theory of gravity. It’s surprising, but at the end of the day, better understanding of physics often has even led to better practical applications.

LEVITT: Is there an empirical finding that could be produced by our current set of telescopes or other mechanisms through which we’re collecting astronomical data that would be really definitive in dealing essentially a fatal blow to the standard model?

RIESS: Almost by definition there always is. The universe and our understanding of it is always under the microscope or, under the gun. Nature has to know how to do every experiment that we look at. And every time somebody takes a new observation, there’s the potential. But, of course, the ones that are most different from anything we’ve done before have the greatest potential, or greater resolution. I’m particularly excited about the new generation of gravitational wave telescopes that could potentially show us another clue that could point the way forward to better understanding of this dark sector.

LEVITT: And those are operative now? 

RIESS: They are. So they started with the first discovery of a gravitational wave maybe about seven years ago. These have ability to measure distances in a completely different way than we do with light. And so they have, I think, really great potential. There are also new generations of telescopes going up in space or here on the ground that have the ability to scan the entire sky every couple of nights at great resolution, at great sensitivity. And these could certainly push the kinds of experiments I’ve been talking about to much greater fidelity and to give us new clues about the nature of the dark sector.

LEVITT: I know this is very different from what you do, but I assume there are people out there trying to do experiments to find dark energy and dark matter. Are they having any success? Do you feel like that’s going to happen anytime soon?

RIESS: It’s certainly possible, particularly with dark matter. People are looking for the dark matter particle in experiments. They’re doing experiments in mine shafts deep in the earth where they could filter out most of the normal particles using the earth’s crust and then try to detect them in a vat of xenon gas or things like that, that will give a unique signature of the discovery of a dark matter particle. They have not succeeded yet, but if they did that could be a really big deal. Other kinds of reactor, nuclear-type experiments, might potentially find a new particle, a neutrino of a very special flavor that could impact the kinds of studies that I’m talking about. When I say Lambda-CDM is our standard model, and I talk about the dark sector. There are other elements of it as well, the particular inventory of particles that we know about in the universe having certain properties, but of course if there’s something else, another particle, the laws of gravity operating as we understand them. But again, we thought we had the laws of gravity right after Newton. And then along came Einstein, showing that gravity could operate quite differently. And in fact, there had been clues there as well. You can almost sometimes see parallels throughout science that astronomers had seen the orbit of Mercury processing in the late 1800s, which did not match Newton’s theory of gravity. And the answer there ended up being we didn’t have the right theory of gravity.

LEVITT: That’s really interesting. So when you say processing, it doesn’t have the right shape, right?

RIESS: It means it’s orbit, which is an ellipse, that ellipse itself is rotating very slowly over time. And that doesn’t happen in Newton’s theory of gravity. Now, in the earlier part of the 1800s, astronomers had seen Uranus, the planet misbehaving, as well, in a different way, and had inferred the existence of a missing planet, which was Neptune. And even found it by looking where that planet needed to be. And so by the late 1800s when astronomers saw Mercury also misbehaving, they inferred that there was another planet between Mercury and the sun. They called it Vulcan. Of course Vulcan didn’t exist, but it was a good guess. And yet they looked for it. They didn’t find it, and eventually learned — well, Einstein figured out that we did not have the right theory of gravity. One of the first things Einstein did with his new theory of gravity was to calculate the expected motion of Mercury. And it matched his theory. He said he had heart palpitations when he saw that. This is the process    of astronomy and physics is, we look at the universe, we look at the motions of things around us and we try to make sense of them using the theories we have and also our understanding of the kind of stuff that’s out there. But it’s hard, and sometimes things don’t match, and that is sometime’s been a clue that we’re missing stuff, that’s out there in the universe. Sometimes it’s a clue that we’re missing the right theory of things. This is a great adventure that I think we all get to be a part of, ultimately, as humans on this planet. This just amazing ambition we have to look out into space and try to figure out what everything is, even if most of it’s not like us.

LEVITT: It seems like a really fun time to be an astrophysicist.

RIESS: It really is.

As I did my Ph.D. in economics, I was really struck by how much economists were trying to emulate physicists in the way they talked about and wrote up their research. So I had to chuckle to myself today, listening to Adam Riess talk about the Hubble tension, where a 10-percent difference between competing estimates has thrown the field into chaos. In economics, when there’s a 10-percent difference between two sets of estimates, we call that a successful replication and we go out for a beer to celebrate.

LEVITT: Now is the point in the show where I welcome on my producer Morgan to handle a listener question.

LEVEY: Hi, Steve. So we had an email from a listener named Mauricio, and Mauricio is in the process of buying a new house or apartment. In the country that Mauricio lives — he doesn’t tell us what country that is — he says it’s common that the final price is reached through an auction without a set end point. So bids are sent to the realtor, and then the realtor informs the interested parties of the highest bid, and then people can bid again or back out. Mauricio wants to know how he can optimize his chances of buying a place at the lowest possible price.

LEVITT: So this is an interesting problem because it’s a lot like what we do informally in the U.S., where when there are multiple bidders it kind of devolves into an auction setting, but it’s much more explicitly set up like an auction. And that really works to the benefit of the sellers ’cause if there’s one thing that we know from economic research, it’s that auctions are a really good mechanism for taking the surplus away from the buyers and giving it to the seller. So it’s a bad situation for Mauricio and I don’t really have a lot of great suggestions, but I do have three ideas that he might be able to use to do at least a little bit better.

LEVEY: Okay. What are they?

LEVITT: Okay, so the first one is obvious. It’s that auctions only work when you have two bidders. And so if Mauricio can find a house or an apartment where he’s the only bidder, then he can’t get into the situation where they extract all the surplus. You try to do the same thing in any market, but especially in this kind of market, I’d push to stay away from direct competition with other bidders.

LEVEY: So basically find a house that only Mauricio wants.

LEVITT: Yeah. That’s obvious but necessary to say off the bat. So my other two suggestions relate to corruption and dishonesty. So one really obvious form of corruption that could happen here is maybe Mauricio is the only one bidding, or maybe he’s already got the highest bid, and the real estate agent can tell from Mauricio’s actions or things he’s said before that he’s not yet hit his highest price. And then the real estate agent can offer up a fake bid in order to get Mauricio to go higher. So in this setting if you’re Mauricio, every single signal that you send to the real estate agent should be about how You’re right at your limit, and, Oh, this really hurts me, and I don’t want to go any further. And I think that could pay off if you’re dealing with a corrupt real estate agent on the other side.

LEVEY: So really Mauricio just needs to be protective of the information he’s giving to the real estate agent.

LEVITT: Exactly. And not only protective, I would say strategic in trying to convey that he’s really close to his limit even if he’s not, I think it is a good strategy. The real estate agent is not your friend. And don’t treat them like your friend.

LEVEY: Okay. What’s your third piece of advice?

LEVITT: Okay, so the third piece is really flipping this corruption argument on its head, which is the real estate agent has private incentives and — I don’t know what it is the real estate agent will want. It may be that a simple deal or a quick deal or a deal that’s all cash. There might be things that would appeal to the real estate agent. Obviously, bribes appeal to real estate agents too. I’m not telling Mauricio necessarily to bribe this real estate agent, but, look, that might be the easiest way to get around this auction. It’s just the flip side. If real estate agents are corrupt, and they’re people, and so they’re likely to be about as corrupt as other people, then there may be opportunities to undermine this situation. And, again, I’m not telling Mauricio to break the law, but if he wants a house cheap breaking the law might be a good way to do that.

LEVEY: What about instead of bribing the real estate agent, we just tell Mauricio to make his offer stand out in a nonfinancial way? 

LEVITT: One thing we’ve talked about before on this show is the idea of somehow making the offer look special by personalizing it. We know that selling one’s house is often a highly emotionally charged activity, especially, say, for people who’ve been in the house for a long time and are now, say, moving into a retirement facility. And so I’ve always been a big advocate of writing personal letters that explain to the owner how you have a love and respect for their house and how you’ll treat it. I think we learned in the past, Morgan, that maybe that’s illegal or something like that? I can’t remember.

LEVEY: Well, it’s illegal in some states in the U.S. and some, it’s frowned upon, but we don’t know what the laws are in Mauricio’s country. 

LEVITT: Yeah, our basic advice to Mauricio is to pull out all the stops of trying to use incentives, especially non-financial incentives, to get what he needs.

LEVEY: Mauricio, best of luck. If you have a question for us or a problem that could use an economic solution, send us an email. Our email is pima@freakonomics.com. That’s P-I-M-A@freakonomics.com. We read every email that’s sent and we look forward to reading yours.

Next week we’ve got an encore presentation of my conversation with legendary music producer Rick Rubin. And in two weeks we’re back with a brand new episode featuring Harvard economist Stefanie Stantcheva. She’s the most recent winner of the John Bates Clark Medal. As always, thanks for listening, and I’ll see you back soon.

*      *      *

People I (Mostly) Admire is part of the Freakonomics Radio Network, which also includes Freakonomics Radio and The Economics of Everyday Things. All our shows are produced by Stitcher and Renbud Radio. This episode was produced by Morgan Levey, and mixed by Jasmin Klinger. We had research assistance from Daniel Moritz-Rabson. Our theme music was composed by Luis Guerra. We can be reached at pima@freakonomics.com, that’s P-I-M-A@Freakonomics.com. Thanks for listening.

LEVITT: How do you say it? Supernovi? Supernovae — novae? What’s the plural?

RIESS: Oh, this is hotly debated. Either way is fine.