Tag Archives: astronomy

The Big Bang: What We Know and How We Know It

When most people think of the Big Bang, they imagine a single moment: a whole universe emerging from nothing. That’s not really how it worked, though. The Big Bang refers not to one event, but to a whole scientific theory. Using Einstein’s equations and some simplifying assumptions, we physicists can lay out a timeline for the universe’s earliest history. Different parts of this timeline have different evidence: some are meticulously tested, others we even expect to be wrong! It’s worth talking through this timeline and discussing what we know about each piece, and how we know it.

We can see surprisingly far back in time. As we look out into the universe, we see each star as it was when the light we see left it: longer ago the further the star is from us. Looking back, we see changes in the types of stars and galaxies: stars formed without the metals that later stars produced, galaxies made of those early stars. We see the universe become denser and hotter, until eventually we reach the last thing we can see: the cosmic microwave background, a faint light that fills our view in every direction. This light represents a change in the universe, the emergence of the first atoms. Before this, there were ions: free nuclei and electrons, forming a hot plasma. That plasma constantly emitted and absorbed light. As the universe cooled, the ions merged into atoms, and light was free to travel. Because of this, we cannot see back beyond this point. Our model gives detailed predictions for this curtain of light: its temperature, and even the ways it varies in intensity from place to place, which in turn let us hone our model further.

In principle, we could “see” a bit further. Light isn’t the only thing that travels freely through the universe. Neutrinos are almost massless, and pass through almost everything. Like the cosmic microwave background, the universe should have a cosmic neutrino background. This would come from much earlier, from an era when the universe was so dense that neutrinos regularly interacted with other matter. We haven’t detected this neutrino background yet, but future experiments might. Gravitational waves meanwhile, can also pass through almost any obstacle. There should be gravitational wave backgrounds as well, from a variety of eras in the early universe. Once again these haven’t been detected yet, but more powerful gravitational wave telescopes may yet see them.

We have indirect evidence a bit further back than we can see things directly. In the heat of the early universe the first protons and neutrons were merged via nuclear fusion, becoming the first atomic nuclei: isotopes of hydrogen, helium, and lithium. Our model lets us predict the proportions of these, how much helium and lithium per hydrogen atom. We can then compare this to the oldest stars we see, and see that the proportions are right. In this way, we know something about the universe from before we can “see” it.

We get surprised when we look at the universe on large scales, and compare widely separated regions. We find those regions are surprisingly similar, more than we would expect from randomness and the physics we know. Physicists have proposed different explanations for this. The most popular, cosmic inflation, suggests that the universe expanded very rapidly, accelerating so that a small region of similar matter was blown up much larger than the ordinary Big Bang model would have, projecting those similarities across the sky. While many think this proposal fits the data best, we still aren’t sure it’s the right one: there are alternate proposals, and it’s even controversial whether we should be surprised by the large-scale similarity in the first place.

We understand, in principle, how matter can come from “nothing”. This is sometimes presented as the most mysterious part of the Big Bang, the idea that matter could spontaneously emerge from an “empty” universe. But to a physicist, this isn’t very mysterious. Matter isn’t actually conserved, mass is just energy you haven’t met yet. Deep down, the universe is just a bunch of rippling quantum fields, with different ones more or less active at different times. Space-time itself is just another field, the gravitational field. When people say that in the Big Bang matter emerged from nothing, all they mean is that energy moved from the gravitational field to fields like the electron and quark, giving rise to particles. As we wind the model back, we can pretty well understand how this could happen.

If we extrapolate, winding Einstein’s equations back all the way, we reach a singularity: the whole universe, according to those equations, would have emerged from a single point, a time when everything was zero distance from everything else. This assumes, though, that Einstein’s equations keep working all the way back that far. That’s probably wrong, though. Einstein’s equations don’t include the effect of quantum mechanics, which should be much more important when the universe is at its hottest and densest. We don’t have a complete theory of quantum gravity yet (at least, not one that can model this), so we can’t be certain how to correct these equations. But in general, quantum theories tend to “fuzz out” singularities, spreading out a single point over a wider area. So it’s likely that the universe didn’t actually come from just a single point, and our various incomplete theories of quantum gravity tend to back this up.

So, starting from what we can see, we extrapolate back to what we can’t. We’re quite confident in some parts of the Big Bang theory: the emergence of the first galaxies, the first stars, the first atoms, and the first elements. Back far enough and things get more mysterious, we have proposals but no definite answers. And if you try to wind back up to the beginning, you find we still don’t have the right kind of theory to answer the question. That’s a task for the future.

Black Holes, Neutron Stars, and the Power of Love

What’s the difference between a black hole and a neutron star?

When a massive star nears the end of its life, it starts running out of nuclear fuel. Without the support of a continuous explosion, the star begins to collapse, crushed under its own weight.

What happens then depends on how much weight that is. The most massive stars collapse completely, into the densest form anything can take: a black hole. Einstein’s equations say a black hole is a single point, infinitely dense: get close enough and nothing, not even light, can escape. A quantum theory of gravity would change this, but not a lot: a quantum black hole would still be as dense as quantum matter can get, still equipped with a similar “point of no return”.

A slightly less massive star collapses, not to a black hole, but to a neutron star. Matter in a neutron star doesn’t collapse to a single point, but it does change dramatically. Each electron in the old star is crushed together with a proton until it becomes a neutron, a forced reversal of the more familiar process of Beta decay. Instead of a ball of hydrogen and helium, the star then ends up like a single atomic nucleus, one roughly the size of a city.

Not kidding about the “city” thing…and remember, this is more massive than the Sun

Now, let me ask a slightly different question: how do you tell the difference between a black hole and a neutron star?

Sometimes, you can tell this through ordinary astronomy. Neutron stars do emit light, unlike black holes, though for most neutron stars this is hard to detect. In the past, astronomers would use other objects instead, looking at light from matter falling in, orbiting, or passing by a black hole or neutron star to estimate its mass and size.

Now they have another tool: gravitational wave telescopes. Maybe you’ve heard of LIGO, or its European cousin Virgo: massive machines that do astronomy not with light but by detecting ripples in space and time. In the future, these will be joined by an even bigger setup in space, called LISA. When two black holes or neutron stars collide they “ring” the fabric of space and time like a bell, sending out waves in every direction. By analyzing the frequency of these waves, scientists can learn something about what made them: in particular, whether the waves were made by black holes or neutron stars.

One big difference between black holes and neutron stars lies in something called their “Love numbers“. From far enough away, you can pretend both black holes and neutron stars are single points, like fundamental particles. Try to get more precise, and this picture starts to fail, but if you’re smart you can include small corrections and keep things working. Some of those corrections, called Love numbers, measure how much one object gets squeezed and stretched by the other’s gravitational field. They’re called Love numbers not because they measure how hug-able a neutron star is, but after the mathematician who first proposed them, A. E. H. Love.

What can we learn from Love numbers? Quite a lot. More impressively, there are several different types of questions Love numbers can answer. There are questions about our theories, questions about the natural world, and questions about fundamental physics.

You might have heard that black holes “have no hair”. A black hole in space can be described by just two numbers: its mass, and how much it spins. A star is much more complicated, with sunspots and solar flares and layers of different gases in different amounts. For a black hole, all of that is compressed down to nothing, reduced to just those two numbers and nothing else.

With that in mind, you might think a black hole should have zero Love numbers: it should be impossible to squeeze it or stretch it. This is fundamentally a question about a theory, Einstein’s theory of relativity. If we took that theory for granted, and didn’t add anything to it, what would the consequences be? Would black holes have zero Love number, or not?

It turns out black holes do have zero Love number, if they aren’t spinning. If they are, things are more complicated: a few calculations made it look like spinning black holes also had zero Love number, but just last year a more detailed proof showed that this doesn’t hold. Somehow, despite having “no hair”, you can actually “squeeze” a spinning black hole.

(EDIT: Folks on twitter pointed out a wrinkle here: more recent papers are arguing that spinning black holes actually do have zero Love number as well, and that the earlier papers confused Love numbers with a different effect. All that is to say this is still very much an active area of research!)

The physics behind neutron stars is in principle known, but in practice hard to understand. When they are formed, almost every type of physics gets involved: gas and dust, neutrino blasts, nuclear physics, and general relativity holding it all together.

Because of all this complexity, the structure of neutron stars can’t be calculated from “first principles” alone. Finding it out isn’t a question about our theories, but a question about the natural world. We need to go out and measure how neutron stars actually behave.

Love numbers are a promising way to do that. Love numbers tell you how an object gets squeezed and stretched in a gravitational field. Learning the Love numbers of neutron stars will tell us something about their structure: namely, how squeezable and stretchable they are. Already, LIGO and Virgo have given us some information about this, and ruled out a few possibilities. In future, the LISA telescope will show much more.

Returning to black holes, you might wonder what happens if we don’t stick to Einstein’s theory of relativity. Physicists expect that relativity has to be modified to account for quantum effects, to make a true theory of quantum gravity. We don’t quite know how to do that yet, but there are a few proposals on the table.

Asking for the true theory of quantum gravity isn’t just a question about some specific part of the natural world, it’s a question about the fundamental laws of physics. Can Love numbers help us answer it?

Maybe. Some theorists think that quantum gravity will change the Love numbers of black holes. Fewer, but still some, think they will change enough to be detectable, with future gravitational wave telescopes like LISA. I get the impression this is controversial, both because of the different proposals involved and the approximations used to understand them. Still, it’s fun that Love numbers can answer so many different types of questions, and teach us so many different things about physics.

Unrelated: For those curious about what I look/sound like, I recently gave a talk of outreach advice for the Max Planck Institute for Physics, and they posted it online here.

What Tells Your Story

I watched Hamilton on Disney+ recently. With GIFs and songs from the show all over social media for the last few years, there weren’t many surprises. One thing that nonetheless struck me was the focus on historical evidence. The musical Hamilton is based on Ron Chernow’s biography of Alexander Hamilton, and it preserves a surprising amount of the historian’s care for how we know what we know, hidden within the show’s other themes. From the refrain of “who tells your story”, to the importance of Eliza burning her letters with Hamilton (not just the emotional gesture but the “gap in the narrative” it created for historians), to the song “The Room Where It Happens” (which looked from GIFsets like it was about Burr’s desire for power, but is mostly about how much of history is hidden in conversations we can only partly reconstruct), the show keeps the puzzle of reasoning from incomplete evidence front-and-center.

Any time we try to reason about the past, we are faced with these kinds of questions. They don’t just apply to history, but to the so-called historical sciences as well, sciences that study the past. Instead of asking “who” told the story, such scientists must keep in mind “what” is telling the story. For example, paleontologists reason from fossils, and thus are limited by what does and doesn’t get preserved. As a result after a century of studying dinosaurs, only in the last twenty years did it become clear they had feathers.

Astronomy, too, is a historical science. Whenever astronomers look out at distant stars, they are looking at the past. And just like historians and paleontologists, they are limited by what evidence happened to be preserved, and what part of that evidence they can access.

These limitations lead to mysteries, and often controversies. Before LIGO, astronomers had an idea of what the typical mass of a black hole was. After LIGO, a new slate of black holes has been observed, with much higher mass. It’s still unclear why.

Try to reason about the whole universe, and you end up asking similar questions. When we see the movement of “standard candle” stars, is that because the universe’s expansion is accelerating, or are the stars moving as a group?

Push far enough back and the evidence doesn’t just lead to controversy, but to hard limits on what we can know. No matter how good our telescopes are, we won’t see light older than the cosmic microwave background: before that background was emitted the universe was filled with plasma, which would have absorbed any earlier light, erasing anything we could learn from it. Gravitational waves may one day let us probe earlier, and make discoveries as surprising as feathered dinosaurs. But there is yet a stronger limit to how far back we can go, beyond which any evidence has been so diluted that it is indistinguishable from random noise. We can never quite see into “the room where it happened”.

It’s gratifying to see questions of historical evidence in a Broadway musical, in the same way it was gratifying to hear fractals mentioned in a Disney movie. It’s important to think about who, and what, is telling the stories we learn. Spreading that lesson helps all of us reason better.

QCD Meets Gravity 2020

I’m at another Zoom conference this week, QCD Meets Gravity. This year it’s hosted by Northwestern.

The view of the campus from wonder.me

QCD Meets Gravity is a conference series focused on the often-surprising links between quantum chromodynamics on the one hand and gravity on the other. By thinking of gravity as the “square” of forces like the strong nuclear force, researchers have unlocked new calculation techniques and deep insights.

Last year’s conference was very focused on one particular topic, trying to predict the gravitational waves observed by LIGO and VIRGO. That’s still a core topic of the conference, but it feels like there is a bit more diversity in topics this year. We’ve seen a variety of talks on different “squares”: new theories that square to other theories, and new calculations that benefit from “squaring” (even surprising applications to the Navier-Stokes equation!) There are talks on subjects from String Theory to Effective Field Theory, and even a talk on a very different way that “QCD meets gravity”, in collisions of neutron stars.

With still a few more talks to go, expect me to say a bit more next week, probably discussing a few in more detail. (Several people presented exciting work in progress!) Until then, I should get back to watching!

Truth Doesn’t Have to Break the (Word) Budget

Imagine you saw this headline:

Scientists Say They’ve Found the Missing 40 Percent of the Universe’s Matter

It probably sounds like they’re talking about dark matter, right? And if scientists found dark matter, that could be a huge discovery: figuring out what dark matter is made of is one of the biggest outstanding mysteries in physics. Still, maybe that 40% number makes you a bit suspicious…

Now, read this headline instead:

Astronomers Have Finally Found Most of The Universe’s Missing Visible Matter

Visible matter! Ah, what a difference a single word makes!

These are two articles, the first from this year and the second from 2017, talking about the same thing. Leave out dark matter and dark energy, and the rest of the universe is made of ordinary protons, neutrons, and electrons. We sometimes call that “visible matter”, but that doesn’t mean it’s easy to spot. Much of it lingers in threads of gas and dust between galaxies, making it difficult to detect. These two articles are about astronomers who managed to detect this matter in different ways. But while the articles cover the same sort of matter, one headline is a lot more misleading.

Now, I know science writing is hard work. You can’t avoid misleading your readers, if only a little, because you can never include every detail. Introduce too many new words and you’ll use up your “vocabulary budget” and lose your audience. I also know that headlines get tweaked by editors at the last minute to maximize “clicks”, and that news that doesn’t get enough “clicks” dies out, replaced by news that does.

But that second headline? It’s shorter than the first. They were able to fit that crucial word “visible” in, without breaking the budget. And while I don’t have the data, I doubt the first headline was that much more viral. They could have afforded to get this right, if they wanted to.

Read each article further, and you see the same pattern. The 2020 article does mention visible matter in the first sentence at least, so they don’t screw that one up completely. But another important detail never gets mentioned.

See, you might be wondering, if one of these articles is from 2017 and the other is from 2020, how are they talking about the same thing? If astronomers found this matter already in 2017, how did they find it again in 2020?

There’s a key detail that the 2017 article mentions and the 2020 article leaves out. Here’s a quote from the 2017 article, emphasis mine:

We now have our first solid piece of evidence that this matter has been hiding in the delicate threads of cosmic webbing bridging neighbouring galaxies, right where the models predicted.

This “missing” matter was expected to exist, was predicted by models to exist. It just hadn’t been observed yet. In 2017, astronomers detected some of this matter indirectly, through its effect on the Cosmic Microwave Background. In 2020, they found it more directly, through X-rays shot out from the gases themselves.

Once again, the difference is just a short phrase. By saying “right where the models predicted”, the 2017 article clears up an important point, that this matter wasn’t a surprise. And all it took was five words.

These little words and phrases make a big difference. If you’re writing about science, you will always face misunderstandings. But if you’re careful and clever, you can clear up the most obvious ones. With just a few well-chosen words, you can have a much better piece.

Congratulations to Roger Penrose, Reinhard Genzel, and Andrea Ghez!

The 2020 Physics Nobel Prize was announced last week, awarded to Roger Penrose for his theorems about black holes and Reinhard Genzel and Andrea Ghez for discovering the black hole at the center of our galaxy.

Of the three, I’m most familiar with Penrose’s work. People had studied black holes before Penrose, but only the simplest of situations, like an imaginary perfectly spherical star. Some wondered whether black holes in nature were limited in this way, if they could only exist under perfectly balanced conditions. Penrose showed that wasn’t true: he proved mathematically that black holes not only can form, they must form, in very general situations. He’s also worked on a wide variety of other things. He came up with “twistor space”, an idea intended for a new theory of quantum gravity that ended up as a useful tool for “amplitudeologists” like me to study particle physics. He discovered a set of four types of tiles such that if you tiled a floor with them the pattern would never repeat. And he has some controversial hypotheses about quantum gravity and consciousness.

I’m less familiar with Genzel and Ghez, but by now everyone should be familiar with what they found. Genzel and Ghez led two teams that peered into the center of our galaxy. By carefully measuring the way stars moved deep in the core, they figured out something we now teach children: that our beloved Milky Way has a dark and chewy center, an enormous black hole around which everything else revolves. These appear to be a common feature of galaxies, and many others have been shown to orbit black holes as well.

Like last year, I find it a bit odd that the Nobel committee decided to lump these two prizes together. Both discoveries concern black holes, so they’re more related than last year’s laureates, but the contexts are quite different: it’s not as if Penrose predicted the black hole in the center of our galaxy. Usually the Nobel committee avoids mathematical work like Penrose’s, except when it’s tied to a particular experimental discovery. It doesn’t look like anyone has gotten a Nobel prize for discovering that black holes exist, so maybe that’s the intent of this one…but Genzel and Ghez were not the first people to find evidence of a black hole. So overall I’m confused. I’d say that Penrose deserved a Nobel Prize, and that Genzel and Ghez did as well, but I’m not sure why they needed to split one with each other.

4gravitons, Spinning Up

I had a new paper out last week, with Michèle Levi and Andrew McLeod. But to explain it, I’ll need to clarify something about our last paper.

Two weeks ago, I told you that Andrew and Michèle and I had written a paper, predicting what gravitational wave telescopes like LIGO see when black holes collide. You may remember that LIGO doesn’t just see colliding black holes: it sees colliding neutron stars too. So why didn’t we predict what happens when neutron stars collide?

Actually, we did. Our calculation doesn’t just apply to black holes. It applies to neutron stars too. And not just neutron stars: it applies to anything of roughly the right size and shape. Black holes, neutron stars, very large grapefruits…

LIGO’s next big discovery

That’s the magic of Effective Field Theory, the “zoom lens” of particle physics. Zoom out far enough, and any big, round object starts looking like a particle. Black holes, neutron stars, grapefruits, we can describe them all using the same math.

Ok, so we can describe both black holes and neutron stars. Can we tell the difference between them?

In our last calculation, no. In this one, yes!

Effective Field Theory isn’t just a zoom lens, it’s a controlled approximation. That means that when we “zoom out” we don’t just throw out anything “too small to see”. Instead, we approximate it, estimating how big of an effect it can have. Depending on how precise we want to be, we can include more and more of these approximated effects. If our estimates are good, we’ll include everything that matters, and get a good approximation for what we’re trying to observe.

At the precision of our last calculation, a black hole and a neutron star still look exactly the same. Our new calculation aims for a bit higher precision though. (For the experts: we’re at a higher order in spin.) The higher precision means that we can actually see the difference: our result changes for two colliding black holes versus two colliding grapefruits.

So does that mean I can tell you what happens when two neutron stars collide, according to our calculation? Actually, no. That’s not because we screwed up the calculation: it’s because some of the properties of neutron stars are unknown.

The Effective Field Theory of neutron stars has what we call “free parameters”, unknown variables. People have tried to estimate some of these (called “Love numbers” after the mathematician A. E. H. Love), but they depend on the details of how neutron stars work: what stuff they contain, how that stuff is shaped, and how it can move. To find them out, we probably can’t just calculate: we’ll have to measure, observe an actual neutron star collision and see what the numbers actually are.

That’s one of the purposes of gravitational wave telescopes. It’s not (as far as I know) something LIGO can measure. But future telescopes, with more precision, should be able to. By watching two colliding neutron stars and comparing to a high-precision calculation, physicists will better understand what those neutron stars are made of. In order to do that, they will need someone to do that high-precision calculation. And that’s why people like me are involved.

QCD Meets Gravity 2019

I’m at UCLA this week for QCD Meets Gravity, a conference about the surprising ways that gravity is “QCD squared”.

When I attended this conference two years ago, the community was branching out into a new direction: using tools from particle physics to understand the gravitational waves observed at LIGO.

At this year’s conference, gravitational waves have grown from a promising new direction to a large fraction of the talks. While there were still the usual talks about quantum field theory and string theory (everything from bootstrap methods to a surprising application of double field theory), gravitational waves have clearly become a major focus of this community.

This was highlighted before the first talk, when Zvi Bern brought up a recent paper by Thibault Damour. Bern and collaborators had recently used particle physics methods to push beyond the state of the art in gravitational wave calculations. Damour, an expert in the older methods, claims that Bern et al’s result is wrong, and in doing so also questions an earlier result by Amati, Ciafaloni, and Veneziano. More than that, Damour argued that the whole approach of using these kinds of particle physics tools for gravitational waves is misguided.

There was a lot of good-natured ribbing of Damour in the rest of the conference, as well as some serious attempts to confront his points. Damour’s argument so far is somewhat indirect, so there is hope that a more direct calculation (which Damour is currently pursuing) will resolve the matter. In the meantime, Julio Parra-Martinez described a reproduction of the older Amati/Ciafaloni/Veneziano result with more Damour-approved techniques, as well as additional indirect arguments that Bern et al got things right.

Before the QCD Meets Gravity community worked on gravitational waves, other groups had already built a strong track record in the area. One encouraging thing about this conference was how much the two communities are talking to each other. Several speakers came from the older community, and there were a lot of references in both groups’ talks to the other group’s work. This, more than even the content of the talks, felt like the strongest sign that something productive is happening here.

Many talks began by trying to motivate these gravitational calculations, usually to address the mysteries of astrophysics. Two talks were more direct, with Ramy Brustein and Pierre Vanhove speculating about new fundamental physics that could be uncovered by these calculations. I’m not the kind of physicist who does this kind of speculation, and I confess both talks struck me as rather strange. Vanhove in particular explicitly rejects the popular criterion of “naturalness”, making me wonder if his work is the kind of thing critics of naturalness have in mind.

Guest Post: On the Real Inhomogeneous Universe and the Weirdness of ‘Dark Energy’

A few weeks ago, I mentioned a paper by a colleague of mine, Mohamed Rameez, that generated some discussion. Since I wasn’t up for commenting on the paper’s scientific content, I thought it would be good to give Rameez a chance to explain it in his own words, in a guest post. Here’s what he has to say:


In an earlier post, 4gravitons had contemplated the question of ‘when to trust the contrarians’, in the context of our about-to-be-published paper in which we argue that accounting for the effects of the bulk flow in the local Universe, there is no evidence for any isotropic cosmic acceleration, which would be required to claim some sort of ‘dark energy’.

In the following I would like to emphasize that this is a reasonable view, and not a contrarian one. To do so I will examine the bulk flow of the local Universe and the historical evolution of what appears to be somewhat dodgy supernova data. I will present a trivial solution (from data) to the claimed ‘Hubble tension’.  I will then discuss inhomogeneous cosmology, and the 2011 Nobel prize in Physics. I will proceed to make predictions that can be falsified with future data. I will conclude with some questions that should be frequently asked.

Disclaimer: The views expressed here are not necessarily shared by my collaborators. 

The bulk flow of the local Universe:

The largest anisotropy in the Cosmic Microwave Background is the dipole, believed to be caused by our motion with respect to the ‘rest frame’ of the CMB with a velocity of ~369 km s^-1. Under this view, all matter in the local Universe appear to be flowing. At least out to ~300 Mpc, this flow continues to be directionally coherent, to within ~40 degrees of the CMB dipole, and the scale at which the average relative motion between matter and radiation converges to zero has so far not been found.

This is one of the most widely accepted results in modern cosmology, to the extent that SN1a data come pre ‘corrected’ for it.

Such a flow has covariant consequences under general relativity and this is what we set out to test.

Supernova data, directions in the sky and dodgyness:

Both Riess et al 1998 and Perlmutter et al 1999 used samples of supernovae down to redshifts of 0.01, in which almost all SNe at redshifts below 0.1 were in the direction of the flow.

Subsequently in Astier et al 2006, Kowalsky et al 2008, Amanullah et al 2010 and Suzuki et al 2011, it is reported that a process of outlier rejection was adopted in which data points >3\sigma from the Hubble diagram were discarded. This was done using a highly questionable statistical method that involves adjusting an intrinsic dispersion term \sigma_{\textrm{int}} by hand until a \chi^2/\textrm{ndof} of 1 is obtained to the assumed \LambdaCDM model. The number of outliers rejected is however far in excess of 0.3% – which is the 3\sigma expectation. As the sky coverage became less skewed, supernovae with redshift less than ~0.023 were excluded for being outside the Hubble flow. While the Hubble diagram so far had been inferred from heliocentric redshifts and magnitudes, with the introduction of SDSS supernovae that happened to be in the direction opposite to the flow, peculiar velocity ‘corrections’ were adopted in the JLA catalogue and supernovae down to extremely low redshifts were reintroduced. While the early claims of a cosmological constant were stated as ‘high redshift supernovae were found to be dimmer (15% in flux) than the low redshift supernovae (compared to what would be expected in a \Lambda=0 universe)’, it is worth noting that the peculiar velocity corrections change the redshifts and fluxes of low redshift supernovae by up to ~20 %.

When it was observed that even with this ‘corrected’ sample of 740 SNe, any evidence for isotropic acceleration using a principled Maximum Likelihood Estimator is less than 3\sigma , it was claimed that by adding 12 additional parameters (to the 10 parameter model) to allow for redshift and sample dependence of the light curve fitting parameters, the evidence was greater than 4\sigma .

As we discuss in Colin et al. 2019, these corrections also appear to be arbitrary, and betray an ignorance of the fundamentals of both basic statistical analysis and relativity. With the Pantheon compilation, heliocentric observables were no longer public and these peculiar velocity corrections initially extended far beyond the range of any known flow model of the Local Universe. When this bug was eventually fixed, both the heliocentric redshifts and magnitudes of the SDSS SNe that filled in the ‘redshift desert’ between low and high redshift SNe were found to be alarmingly discrepant. The authors have so far not offered any clarification of these discrepancies.

Thus it seems to me that the latest generation of ‘publicly available’ supernova data are not aiding either open science or progress in cosmology.

A trivial solution to the ‘Hubble tension’?

The apparent tension between the Hubble parameter as inferred from the Cosmic Microwave Background and low redshift tracers has been widely discussed, and recent studies suggest that redshift errors as low as 0.0001 can have a significant impact. Redshift discrepancies as big as 0.1 have been reported. The shifts reported between JLA and Pantheon appear to be sufficient to lower the Hubble parameter from ~73 km s^-1 Mpc^-1 to ~68 km s^-1 Mpc^-1.

On General Relativity, cosmology, metric expansion and inhomogeneities:

In the maximally symmetric Friedmann-Lemaitre-Robertson-Walker solution to general relativity, there is only one meaningful global notion of distance and it expands at the same rate everywhere. However, the late time Universe has structure on all scales, and one may only hope for statistical (not exact) homogeneity. The Universe is expected to be lumpy. A background FLRW metric is not expected to exist and quantities analogous to the Hubble and deceleration parameters will vary across the sky.  Peculiar velocities may be more precisely thought of as variations in the expansion rate of the Universe. At what rate does a real Universe with structure expand? The problems of defining a meaningful average notion of volume, its dynamical evolution, and connecting it to observations are all conceptually open.

On the 2011 Nobel Prize in Physics:

The Fitting Problem in cosmology was written in 1987. In the context of this work and the significant theoretical difficulties involved in inferring fundamental physics from the real Universe, any claims of having measured a cosmological constant from directionally skewed, sparse samples of intrinsically scattered observations should have been taken with a grain of salt.  By honouring this claim with a Nobel Prize, the Swedish Academy may have induced runaway prestige bias in favour of some of the least principled analyses in science, strengthening the confirmation bias that seems prevalent in cosmology.

This has resulted in the generation of a large body of misleading literature, while normalizing the practice of ‘massaging’ scientific data. In her recent video about gravitational waves, Sabine Hossenfelder says “We should not hand out Nobel Prizes if we don’t know how the predictions were fitted to the data”. What about when the data was fitted (in 1998-1999) using a method that has been discredited in 1989 to a toy model that has been cautioned against in 1987, leading to a ‘discovery’ of profound significance to fundamental physics?

A prediction with future cosmological data:

With the advent of high statistics cosmological data in the future, such as from the Large Synoptic Survey Telescope, I predict that the Hubble and deceleration parameters inferred from supernovae in hemispheres towards and away from the CMB dipole will be found to be different in a statistically significant (>5\sigma ) way. Depending upon the criterion for selection and blind analyses of data that can be agreed upon, I would be willing to bet a substantial amount of money on this prediction.

Concluding : on the amusing sociology of ‘Dark Energy’ and manufactured concordance:

Of the two authors of the well-known cosmology textbook ‘The Early Universe’, Edward Kolb writes these interesting papers questioning dark energy while Michael Turner is credited with coining the term ‘Dark Energy’.  Reasonable scientific perspectives have to be presented as ‘Dark Energy without dark energy’. Papers questioning the need to invoke such a mysterious content that makes up ‘68% of the Universe’ are quickly targeted by inane articles by non-experts or perhaps well-meant but still misleading YouTube videos. Much of this is nothing more than a spectacle.

In summary, while the theoretical debate about whether what has been observed as Dark Energy is the effect of inhomogeneities is ongoing, observers appear to have been actively using the most inhomogeneous feature of the local Universe through opaque corrections to data, to continue claiming that this ‘dark energy’ exists.

It is heartening to see that recent works lean toward a breaking of this manufactured concordance and speak of a crisis for cosmology.

Questions that should be frequently asked:

Q. Is there a Hubble frame in the late time Universe?

A. The Hubble frame is a property of the FLRW exact solution, and in the late time Universe in which galaxies and clusters have peculiar motions with respect to each other, an equivalent notion does not exist. While popular inference treats the frame in which the CMB dipole vanishes as the Hubble frame, the scale at which the bulk flow of the local Universe converges to that frame has never been found. We are tilted observers.

Q. I am about to perform blinded analyses on new cosmological data. Should I correct all my redshifts towards the CMB rest frame?

A. No. Correcting all your redshifts towards a frame that has never been found is a good way to end up with ‘dark energy’. It is worth noting that while the CMB dipole has been known since 1994, supernova data have been corrected towards the CMB rest frame only after 2010, for what appear to be independent reasons.

Q. Can I combine new data with existing Supernova data?

A. No. The current generation of publicly available supernova data suffer from the natural biases that are to be expected when data are compiled incrementally through a human mediated process. It would be better to start fresh with a new sample.

Q. Is ‘dark energy’ fundamental or new physics?

A. Given that general relativity is a 100+ year old theory and significant difficulties exist in describing the late time Universe with it, it is unnecessary to invoke new fundamental physics when confronting any apparent acceleration of the real Universe. All signs suggest that what has been ascribed to dark energy are the result of a community that is hell bent on repeating what Einstein supposedly called his greatest mistake.

Digging deeper:

The inquisitive reader may explore the resources on inhomogeneous cosmology, as well as the works of George Ellis, Thomas Buchert and David Wiltshire.

When to Trust the Contrarians

One of my colleagues at the NBI had an unusual experience: one of his papers took a full year to get through peer review. This happens often in math, where reviewers will diligently check proofs for errors, but it’s quite rare in physics: usually the path from writing to publication is much shorter. Then again, the delays shouldn’t have been too surprising for him, given what he was arguing.

My colleague Mohamed Rameez, along with Jacques Colin, Roya Mohayaee, and Subir Sarkar, wants to argue against one of the most famous astronomical discoveries of the last few decades: that the expansion of our universe is accelerating, and thus that an unknown “dark energy” fills the universe. They argue that one of the key pieces of evidence used to prove acceleration is mistaken: that a large region of the universe around us is in fact “flowing” in one direction, and that tricked astronomers into thinking its expansion was accelerating. You might remember a paper making a related argument back in 2016. I didn’t like the media reaction to that paper, and my post triggered a response by the authors, one of whom (Sarkar) is on this paper as well.

I’m not an astronomer or an astrophysicist. I’m not qualified to comment on their argument, and I won’t. I’d still like to know whether they’re right, though. And that means figuring out which experts to trust.

Pick anything we know in physics, and you’ll find at least one person who disagrees. I don’t mean a crackpot, though they exist too. I mean an actual expert who is convinced the rest of the field is wrong. A contrarian, if you will.

I used to be very unsympathetic to these people. I was convinced that the big results of a field are rarely wrong, because of how much is built off of them. I thought that even if a field was using dodgy methods or sloppy reasoning, the big results are used in so many different situations that if they were wrong they would have to be noticed. I’d argue that if you want to overturn one of these big claims you have to disprove not just the result itself, but every other success the field has ever made.

I still believe that, somewhat. But there are a lot of contrarians here at the Niels Bohr Institute. And I’ve started to appreciate what drives them.

The thing is, no scientific result is ever as clean as it ought to be. Everything we do is jury-rigged. We’re almost never experts in everything we’re trying to do, so we often don’t know the best method. Instead, we approximate and guess, we find rough shortcuts and don’t check if they make sense. This can take us far sometimes, sure…but it can also backfire spectacularly.

The contrarians I’ve known got their inspiration from one of those backfires. They saw a result, a respected mainstream result, and they found a glaring screw-up. Maybe it was an approximation that didn’t make any sense, or a statistical measure that was totally inappropriate. Whatever it was, it got them to dig deeper, and suddenly they saw screw-ups all over the place. When they pointed out these problems, at best the people they accused didn’t understand. At worst they got offended. Instead of cooperation, the contrarians are told they can’t possibly know what they’re talking about, and ignored. Eventually, they conclude the entire sub-field is broken.

Are they right?

Not always. They can’t be, for every claim you can find a contrarian, believing them all would be a contradiction.

But sometimes?

Often, they’re right about the screw-ups. They’re right that there’s a cleaner, more proper way to do that calculation, a statistical measure more suited to the problem. And often, doing things right raises subtleties, means that the big important result everyone believed looks a bit less impressive.

Still, that’s not the same as ruling out the result entirely. And despite all the screw-ups, the main result is still often correct. Often, it’s justified not by the original, screwed-up argument, but by newer evidence from a different direction. Often, the sub-field has grown to a point that the original screwed-up argument doesn’t really matter anymore.

Often, but again, not always.

I still don’t know whether to trust the contrarians. I still lean towards expecting fields to sort themselves out, to thinking that error alone can’t sustain long-term research. But I’m keeping a more open mind now. I’m waiting to see how far the contrarians go.