20. Inflation - Sean Carroll - Dark Matter, Dark Energy: The Dark Side of the Universe

We have discussed our models of a smooth and persistent Dark Energy; a strictly positive vacuum energy without changes, a dynamic but slowly changing quintessence, and a modified concept of gravity in general relativity itself. All three fit our data, but none are compelling. So we need to look back to something more fundamental in order to gain insights on the nature of dark energy. Some models might then look to be more natural than others.

Alan Guth was working on his fourth post-doc in his ninth year. Only a very smart scientist could get all of these positions, but such an extended experience with no professorship could not continue. He was working on magnetic monopoles which were predicted from the Grand Unified Theory where the strong, weak, and electromagnetic forces are one. Guth wanted to get rid of monopoles to explain the fact that we don't detect any. Guth's notebook is now on display in the Adler planetarium in Chicago where he first wrote down the concept of inflation in 1980. This solved his monopole problem and included the bonus prediction of a flat universe, among others.

Initially this seemed to look true because of the trend at the time of finding more matter in the universe to reach the critical density. But the 1980s and 1990s made us realize we could only find 30% of that needed to reach the critical density. That was not enough for a flat universe, and believers in inflation were worried. There were also contradictions with the physical mechanisms for inflation.

But the supernovae observations of an accelerating universe changed everything in 1998. The dark energy provided the missing 70% to make the universe flat and the physical mechanism could be modeled. The believers of inflation were happy. An accelerating universe didn't seem like such a crazy idea than if it had happened just once, way back then or now. Later confirmations from Boomerang, COBE, and WMAP data seemed to imply inflation was really on the right track. It predicts the flatness, the smoothness, the perturbations, the physical expansion, and a possible multiverse which are all further explained by Sean.

Inflation is a model of Cosmological Initial Conditions. Unlike most experiments that set up initial conditions whose results are observed to create laws, inflation is just one version of setting up an initial condition to match the already observed universe and its known laws, and problems!

One such serious issue is the Horizon Problem. Imagine that we can actually see back to the big bang, where two different points would each see their own limited and separate areas due to the finite speed of light. These points would not have had time enough to communicate things like their temperature or time and rate of expansion. Yet we observe the entire CMB to have such an identical temperature and uniform expansion rate, that it seems as if there were some actual kind of communication to explain our observations. Inflation solves this by an ultra high energy causing an incredible acceleration that allows the two points to once be in communication with each other!

The less serious problem of flatness is also solved by inflation. The Friedmann equation ρ = H² + K, has ρ and K evolve in time, with H just adjusting accordingly to keep the equation working. But K fades less than ρ, so K should be large. This is not observed since a flat universe has K = 0. But inflation has ρ not made of matter and radiation, but of a constant dark energy. Thus ρ does not fade away, and any value of K is inflated away. At the end of inflation, the leftover dark energy is turned into matter and radiation, an amount that should be much larger than K, which is the flatness we observe.

Inflation should have a field, the Inflaton Field, with a slowly decreasing energy density and a constant rate of expansion producing acceleration. At some point there was a phase transition where the energy density turned into matter and radiation producing the big bang in a process called reheating. This field sounds like Quintessence, but with vastly different energy scales.

How could one field decrease so much? Initially dominated by dark matter, ordinary matter, and radiation, the Quintessence version of dark energy suddenly turned on to accelerate the universe, while the Inflaton Field version of dark energy had been there all along. Its easier to think of them as separate fields instead of one in the same.

Inflation produced a universe smooth enough to solve the horizon problem, but not perfectly smooth enough to prevent energy fluctuations. Quantum mechanics does not allow a perfectly smooth universe on that kind of scale. These fluctuations would produce perturbations needed to produce galaxies. The amplitude should be the same at every distance scale, whether a parsec or gigaparsec.

Imagine back to the points at the big bang where a tiny patch of dark energy dominated an Inflaton Field that led to the universe. But what about other patches not dominated by inflation? They may have difference fluctuations from different inflation fields? They may have a vacuum energy with larger values, or other values for their constants? More speculations are given on this in a few lectures, so back to inflation. Is it true and how does it work?

The flatness and fluctuations we require are predicted. But these are both requirements conjectured before inflation came around in 1980. This makes the confirmation seems very plain and points out the need for something unique prediction of inflation. This fits with Karl Popper's philosophy of a bold theory with surprising and testable predictions being the best.

One field that would also have been around all along is gravity. This would also produce fluctuations that would accumulate with the size of the universe. There will also be more on this later. Like the CMB they would fill the background of the universe. There is a polarization imprint on the CMB that we can detect, but not well enough to make any confirmations.

What was the universe like before inflation? Why did inflation start? What is inflation? We have ideas for these questions, but they are obviously far past our experimental capabilities. We need to think of how to apply our ideas to perform a test on the universe.

We've been doing a great job of describing what our universe looks like today. We have a model that fits a wide variety of data, yet in the last few lectures we try to dig in a little bit to that model, to the dark energy part of it. The part that's the ρ that doesn't change from place to place, that is smooth throughout the universe, and is persistent as the universe expands, so is more or less constant as a function of time.

We found that it's possible to come up with models that explain the dark energy. It could be a strictly positive constant of vacuum energy, one that doesn't change at all. It could by something dynamical that slowly changes. It could even be a modification of General Relativity itself. In each one of these cases, it's possible to come up with a version, some sort of model that actually fits the data. Yet in none of these cases do we have something that is actually compelling. In none of them do we have a specific version or implementation of this idea, that is not only able to fit the data, but kind of makes sense to us, so it's natural in some way and fits together with other things we know.

Therefor in this lecture and the two following, we'll step back a little bit and try to think not specifically not so much about the dark matter and dark energy, but about the fundamental laws of physics and cosmology, where the universe came from and how it works at a very deep level.

Our motivation for doing this is because we want to go back to the dark energy, and to a lesser extent the dark matter, and understand why these things have the properties they do. Maybe it's not so easy as plugging in something that seems to be something that fits the data. We need to think more deeply about what we consider to be a natural explanation versus an unnatural one.

So in particular in this lecture, we're going back to the beginning of the universe to talk about inflationary cosmology, the idea that the early universe underwent a period of extremely rapid, accelerated expansion. Of course, we've been telling you for the last several lectures that the current universe is beginning to undergo a period of accelerated expansion, so the idea that the early universe had a different idea of accelerated expansion doesn't sound so crazy!

Inflation came on the scene as a physical theory by 1980-1981. Back then the idea that the universe could accelerate was much less accepted. So it in fact predates the idea of dark energy, the idea that there was a phase of dark energy like domination in the very early universe, which we call the inflationary universe.

Now why do we even think about something like this? For our present purposes, inflation is connected to the concepts of dark matter and dark energy, in some very specific ways. First, inflation makes predictions, among which is that the universe should be spatially flat. Remember we talked about the flatness problem, the idea that it would make more sense for the universe to be flat than anything else, because it's pretty close to flat already. Yet inflation provides the dynamical mechanism to make the universe flat, and it makes it very, very flat. So that's a very strong prediction.

Secondly, inflation predicts certain kind of perturbations to the universe. Not only is the universe very flat on all scales, but there are also tiny deviations in the density from place to place. These days, we use those prediction when matching what we see in the CMB, those fluctuations at early times, to the observations we have today with large-scale structure in the distribution of galaxies.

Inflation is also of course, similar to dark energy physically. We have some feel to what is making the universe accelerate, and some ρ that doesn't go away very quickly. That is what you need for dark energy, and also separately what you need for inflation. So by thinking about one, you might better understand the other.

Finally as a spinoff of inflation, we have the concept of the multiverse. We've always known that the universe we observe right now, looks more or less the same within what we observe. Outside, what we can observe, we don't know what to say. It could be more of the same forever and ever, or it could be very different, we just have no way of saying anything.

Inflation at least, lets us ask the question scientifically and proposes that maybe the universe is very different outside what we see. That possibility, of a multiverse where conditions change from place to place in a dramatic way, will turn out to bear on the question of dark energy, the question of why the vacuum energy should be as small as it is, compared to the natural value that it should have. Our expectation for what a natural value is, might be different in a multiverse, than it is in a single, lonely, universe.

OK, so let's talk about what people were thinking about back in 1980 when they were beginning to invent inflation, and think about it seriously. The role of inflation is as a model of cosmological initial conditions. The role of initial conditions in cosmology is very different from that of all other physical sciences. If you think about physics as practiced in a lab, or chemistry or something like that, what happens is that you set up an experiment, including the initial conditions. You say, "I want to have a ball rolling down an inclined plane and I will put the ball at the top."

Yet cosmology is different than that. In cosmology, you don't get to do the experiment more than once. The experiment is being done as we live right now! The experiment is the whole universe, so a theory of cosmology, unlike one of balls rolling down planes, or chemistry, contains not only dynamical laws telling you how things evolve, but also a specification of the initial conditions. Why was the early universe in such and such a configuration, which let it lead to the universe in which we see today? That is a respectable cosmological question that just doesn't arise if you're doing particle physics or chemistry.

So we have some idea of what kind of initial conditions might seem natural, robust, or sensible to us. We have other things that seem finely tuned, that seem like for some reason, some particular quantity is very small, when we might have expected it to be big. Inflation addresses these kinds of questions directly. The entire point of inflation is to make the universe which we actually see at early times, appear very natural. It's a dynamical mechanism from which you can start in various different sets of initial conditions, and what inflation does, is to get you to a point where it looks like our Big Bang universe. A universe that looks smooth on very large-scales, with tiny fluctuations in density, and very close to spatially flat.

So lets see how that works. There are basically two geometrical problems in conventional cosmology that inflation tries to solve. One is called the horizon problem, which is actually the much more serious problem from a cosmological point of view. Think about the CMB. Think about this image we can take from satellites, such as WMAP. It's a snapshot of what the universe looked like only 400,000 years after the Big Bang.

Now we know that light travels at a finite speed, one light year per year. So when we look back in time, we don't see things arbitrarily far away. We see back to the Big Bang. We cannot see any further back than that. Perhaps we can see something right close to the Big Bang. We don't really know what the Big Bang itself is, but lets be informal for the next few minutes, and talk as if we can see to the Big Bang, yet nothing beyond that.

Well the same thing would be true of people alive at the time of the CMB. There was no one alive back then, yet imagine an observer sitting at that time in the universe. They would have a past, and be able to describe certain points in the universe as ones they can see from signals coming to them near, or at the speed of light. Other points would have different sets of things they can see in the universe.

Yet the thing is, you can do the calculation in a universe full of nothing but matter and radiation, no forms of dark energy. You find that in the CMB, widely separated points, in fact any two points we observe that are more than one or two degrees apart, share absolutely no points in their past. In other words, you take one of those points, and you extend it to the past, all the way to the Big Bang, then take another point on the CMB and extend it to the past all the way, and you get two points of the universe at early times that don't overlap.

In other words, there is nothing in the very early universe near the Big Bang, that has the ability to communicate with widely separated points on the CMB, since they would have had to travel faster than the speed of light. Nevertheless, despite the fact that, as we say, these points were never in causal contact, there's nothing that can get from one part to the other, slower than the speed of light. Yet these points on the CMB are very close to the same temperature.

That means that these different points in space began to expand at the same time. Nevertheless, they were never in contact with each other. They have their horizons which they can see in the past that don't overlap. So the question is, how do these different points know how to be at the same temperature? How did these regions of space know to start expanding at the same time? They were never in communication with each other anyway. That is known as the horizon problem.

The flatness problem, we've already talked about. If you look at the Friedmann equation of cosmology, it has three terms. It has the energy density of the universe (ρ), the Hubble parameter or expansion rate (H), and spatial curvature (K).

ρ = H² + K

Basically if you know what the universe is made of, if you know the stuff inside, let's say it's just matter and radiation (ρ), and again imagine there's no dark energy, so then you know how ρ evolves as the universe grows (H). You also know how the curvature (K) evolves as the universe grows. That's just a geometric fact. Therefor in the Friedmann equation, the Hubble parameter term (H) evolves to compensate. Basically the energy density (ρ) and curvature (K), do what they do as the universe expands, and then H just adjusts to solve this equation.

The flatness problem is the fact that the curvature term (K) goes away more slowly than the energy density term (ρ), assuming ρ is made of matter and radiation. So if both ρ and K are non-zero in the early universe, then in the late universe, the curvature should be much bigger, yet it's not. If K were exactly zero, it would stay zero and that would make sense. Yet why is it exactly zero? Why isn't it some small number or big number at early times, and therefor a very big number at late times. That's the flatness problem.

So these problems were known in the 1970s, especially to Alan Guth who invented inflation, which turns out to solve both of these problems simultaneously. Inflation says that you start in the early universe with a tiny little patch of space, dominated by some ultra high energy form of dark energy. Because it's ultra-high energy, this dark energy accelerates that little patch of universe, at a tremendous rate. It's not matter or radiation, and remains approximately constant density. This leads to a tremendously fast expansion rate in this little patch of space.

So that means two things. One is that this little patch of space which might have had a K to begin with, has it inflated away by this incredibly fast H. Remember that K goes away, and ρ goes away even faster if the ρ is matter and radiation. Yet if the ρ is dark energy, it doesn't go away as fast. So during inflation, the inflationary ρ doesn't go away, but K does. At the end of inflation, that ρ from the dark energy, turns into ordinary matter and radiation, which is not much larger than K. That's why the K in our current universe is so close to zero, it was all inflated away at early times.

The horizon problem is solved because you can imagine an incredibly tiny patch of space, one that was in causal contact, one that did share points in the past and had time to communicate, and inflation takes that patch and expands it to a tremendous size. So in other words, inflation changes the past history of the universe, in such a way that different points we observe on the CMB, did used to communicate with each other. They were very close to each other right at the Big Bang, and did know what each other were doing. There is no horizon problem in inflation.

So what you need to make that work of course, is a temporary form of dark energy at very high-energy which accelerates the universe at early times, and then goes away. In 1980 this was dramatic, since we didn't know about our current form of dark energy, since this was all brand new stuff. Yet these days, we think, "OK, that's something we can make sense of."

It was Alan Guth who was a postdoc in 1980 when he invented the theory of inflation. A postdoc is something in between a graduate student and a professor. You get a series of jobs in which you're supposed to do nothing but write papers and do research, so universities can decide whether or not they'd ever want to hire you to be a professor. These days you might do one or two postdocs before you realize either you have a professor job now, or that you should find other work.

In those days of the 1970s, you would even do less than today. So maybe one postdoc only, and even two was unusual. Yet Guth was on his fourth postdoc, in his ninth year of doing so! Everyone thought he was really smart, so they kept giving him jobs, yet he didn't write that many papers, so they didn't give him a professorship position. Finally he hit the jackpot by inventing inflation.

He was actually not trying to solve the horizon and flatness problems when he was working on inflation. There was another problem called the magnetic monopole problem, involving a set of theories called GUT (Grand Unified Theories) that tried to outdo the standard model of particle physics. It tries to take the strong force, the weak force, and the electromagnetic force, trying to unify them in a single description.

This is a very compelling idea, and still might be right, yet it made a prediction at the time that seemed incredibly incorrect. That prediction was that there were particles called magnetic monopoles, individual magnetic charges that we don't see in nature. Yet we do see individual electric charges. According to GUT the universe should be full of magnetic monopoles, yet we don't see any. How to get rid of them, that was the question Guth was trying to answer.

So not only does inflation solve the horizon and flatness problems, it also solves the monopole problem. You have a high-density of monopoles in the early universe, and all you do is inflate them away. Since then, inflation has become a great cure-all for anything that the early universe creates that you don't see in the later universe. As long as those things that were created, were done so before inflation, then inflation can dilute everything away by a tremendous amount, before it's dark energy turns into matter and radiation.

So Guth realized that he had a solution to the monopole problem, and already had in the back of his mind, knowledge about the horizon and flatness problem, so realized at once that his idea of inflation solved them all. He literally was working late at night, and in his notebook wrote, "Spectacular Realization," and put a box around it. He realized his idea could solve not only the monopole problems, but also the horizon and flatness problems.

That notebook he was writing in, is now on display in the Adler planetarium in Chicago. It was a moment in the history of cosmology when he realized that this one idea could solve a whole bunch of problems all at once. So people realized this and caught on very quickly to the idea that inflation was a great help for the various cosmological conundrums we had. Of course, among other things, Alan Guth got a faculty job very quickly, and is now a full-professor at MIT.

One of the nice things about inflation was that it provided predictions. It was a scientific theory that made scientific predictions that could come true, or be false. It's strongest prediction was that the universe should be spatially flat, since the total ρ of the universe, should be the critical density. This is an interesting prediction since it was made in 1980, and throughout the 80s and most of the 90s, it didn't look like it was true. People thought there was enough uncertainty that maybe the universe did have the critical density (Ω), yet as they measured more about the density of matter, they found out it wasn't enough. They honed in on the ρ of matter being about 30% of Ω.

So there are two really important things that this dark energy does to help the idea of inflation, to boost our confidence that something like inflation is right. First, most obviously, the dark energy provides the extra 70% of the density of the universe that we need to make it spatially flat. In other words, by 1997 if you believed in inflation, you were worried. There were some people who were actually backsliding and trying to create models of inflation which had universes without Ω, that were negatively curved spaces instead of flat spaces. Guth himself never actually went that far.

You can invent such models, but they're incredibly ugly. The true prediction of inflation is that the universe should be spatially flat. So in 1998, when the supernovae evidence came in, that there was such a thing as dark energy and you could make a spatially flat universe without only relying on matter, both ordinary and dark, it made the case for inflation much stronger. That was a prediction of inflation that came right. Then by 2000, when Boomerang and other CMB experiments said, "Yes indeed, the universe is spatially flat," inflation was of course, right on.

The other idea that was helpful to inflation from dark energy, is the very demonstration that the universe is allowed to accelerate. Remember that we have something called the cosmological constant problem. Why isn't the energy density in a vacuum, much bigger than we apparently observe it to be? We don't know the answer to that problem, yet before 1998 it was always possible that the answer was that the vacuum energy or other forms of dark energy, did not gravitate. That there was something in the laws of physics which says that the expansion of the universe just doesn't respond to things with negative pressure.

No one had a good model along those lines, but it was an allowed way to think. If that had been true, it would be difficult to understand how the universe could possibly accelerate. So the fact that we are now observing the universe to be accelerating right now, means that it is allowed to accelerate and therefor it could have been accelerating at earlier times when inflation was necessary. In other words, inflation is on much better physical grounds now than it was before.

So how do you make it work? You invent a model of inflation. Well for dynamical dark energy, we invented a field called quintessence, that slowly changes its ρ as the universe expands. Exactly the same thing is true for inflation. You invent a new field and call it the inflaton, though you have no idea what it is. It's the field that makes inflation happen. It has a huge ρ at very early times, and becomes the dominant form of energy in some patch of space, which then accelerates or inflates at a tremendous rate.

This happens when the ρ in that inflaton field gradually diminishes, very gradually so the expansion rate is continually accelerating. Then at some point, there's a phase transition where ρ in the dark energy transforms into ordinary matter and radiation. We call this reheating. In other words it's a nearly constant ρ for awhile, then it snaps and turns into matter and radiation that we know and love, which we see as the Big Bang.

That's the basic idea of inflation. So because the physics behind inflation sounds to our ears very similar to the physics underlying quintessence or dynamical dark energy, some people have asked the question if it is in fact exactly the same thing? In other words, is there one field that was the inflaton at very early times, providing the dark energy back then, and also is now the quintessence field, providing the dark energy right now? Papers are written with titles like "Quintessential inflation." The opportunity for a pun in this field is never passed by!

Well the answer is that it could happen. It could be that the same field is responsible for inflation and for the dark energy today, but probably not. For one thing, the energy scales are tremendously different form each other. The ρ of the universe near the Big Bang, when inflation was going on, was many orders of magnitude higher than it is today. It is possible that ρ was dominated by the same field then and now, yet what you have to do is make that disappear in between, or at least be dramatically sub-dominant.

At least from the time of Big Bang nucleosynthesis to the time of today, just before today, the universe was certainly dominated by ordinary matter and radiation. By ordinary, we mean matter-like particles, including dark matter. We know from the data, from Big Bang nucleosynthesis and from the CMB, that the universe wasn't dominated by dark energy all along, but that the dark energy has kicked in recently. Inflation says the dark energy was dominating way back then.

So it's actually easier to make those two periods of domination be due to two completely different fields, than to the same field that was important back then, disappears, and then comes back. It's hard to make one field be so different that it dominates at very high densities and very low densities. Yet it's still the kind of idea that people are working on, and might end up being right. We'll have to go see.

The other thing that inflation gives us is a bonus. Not only does it explain away the horizon problem, the flatness problem, and the monopole problem, but it also gives us a dynamical origin for the density fluctuations we observe in the universe. If you think about it, the universe on very large scales is in a very strange state. It's very smooth on very large scales, yet not perfectly smooth. The deviations from smoothness that we can observe, 1 part in 100,000, are certainly observable. They're not absolutely absent.

So if you think about it, why would it be that the early universe would undergo some process that made things very smooth, yet not perfectly smooth? Why isn't it either very lumpy, or even smoother? The answer in the context of inflation, comes down to quantum mechanics. Inflation tries to expand the universe, and smooth it out. As our universe expands, as it's accelerated by some form of dark energy, that acceleration smooths out bumps and ripples. Yet quantum mechanics and the Uncertainty Principle, say that you can't smooth out everything perfectly.

You're trying your best to make the universe smooth, yet the field that is doing it, the inflaton field that is driving ρ, has quantum mechanical fluctuations, a little bit of jitteriness that you can never get rid of. It's those quantum mechanical fluctuations that turn into perturbations in the ρ of matter, radiation, and dark matter, which show up in the CMB as temperature fluctuations that grow into the galaxies we have today.

In fact, there is a prediction on top of that. Namely that the amplitude of the fluctuations should be more or less the same at every distance scale. That's because inflation as it's happening, is happening at more or less the same rate at every distance scale. It is imprinting fluctuations as it goes along. This is of course exactly what we do observe. When we look at the CMB and large-scale structure in the universe, we see perturbations that seem to be about the same primordial amplitude, whether they are one parsec across, or one gigaparsec across. So inflation is at least coming close to being correct with that prediction.

The other bonus we get is a little bit less tangible. The tangible bonus we get from inflation is the fact that there are density perturbations that are predicted, and in fact have become a cottage industry amongst cosmologists, to think about how inflation leads to those perturbations, where they can come from, and how you can test them.

Yet there's a slightly more speculative outcome of inflation, which will turn out to be useful two lectures from now, when we discuss the multiverse. We alluded to this at the very beginning, where we have this picture of inflation as a tiny patch of space at early times was dominated by the dark energy in the inflaton field, and accelerated at a tremendous rate that grew up to be a universe sized thing, what we live in today.

Yet back then, what about the other patches of space? What about the other parts of the universe which were not in the little patch which initially inflated to become our universe? Well it's easy to imagine that there were all sorts of fluctuations, that the universe was very different from place to place, way back then before inflation ever happened. So inflation grabs this little piece of universe, expanding into what we see today.

Yet other regions could easily be grabbed by different kinds of inflation, different inflaton fields, or just the same inflaton field but evolving in a different way. The act of inflation may very well be to take different parts of the universe and blow them up int universe-sized pieces, yet end up in very different conditions. So we need a theory of how the universe could be in different conditions, yet inflation allows us to talk about, in a scientific way, the possibility that outside our observable universe, conditions are very different. That will change how we think about what constitutes a natural value for things like the vacuum energy, other constants of nature. So we'll have a whole multiverse, and its very possible existence will affect how we think about problems involving both dark matter and dark energy.

So the important thing then, once we're stuck with the idea of inflation, or better put, once we're granted the idea of inflation, which is a good idea, how do we know, first, whether it's true? Second, how do we make it work? These are exactly what cosmologists today are thinking about in very serious ways.

First we want to test the idea of inflation. We said that inflation makes a prediction. It takes a little patch of the universe, expands it up enough to be spatially flat, so that indeed, we'll expect a spatially flat universe. It also makes specific kinds of predictions about density fluctuations. In fact, it predicts that the amplitude of density fluctuations, should be the same at every distance scale. Both of those seem to be true enough in the universe we observe.

However, both of those were conjectured to be true, even before inflation was invented. The problem with these predictions, with the universe that should be spatially flat, and with perturbations with the same amplitude on different scales, is that they're very vanilla and not very flavorful. You could imagine that these are the simplest possibilities for what is true, even if you don't have a mechanism for inflation that makes them true.

So people imagined that the universe was spatially flat, that there were different density fluctuations with the same size, on different physical length scales, before anyone ever invented inflation. So even though they are predictions, they are not unique predictions. One could certainly imagine that they are true, without inflation necessarily being true.

What we want to test inflation, is something more unique, something that inflation says is true that other theories don't necessarily say is true. There is one known example of such a thing. Inflation, when it's expanding the universe, has the inflaton field itself, and its quantum fluctuations. The latter gets imprinted into density fluctuations,which grow under the force of gravity into galaxies and large-scale structure.

Yet there is another field lying around at that time when inflation is going on, namely the gravitational field. It will also undergo small quantum fluctuations during inflation, which will get expanded to be the size of the universe. So what is the observable form of a small fluctuation in the gravitational field, the answer is a gravitational wave, gravitational radiations, or individual gravitons. This gravitational radiation is something we're looking for in lab experiments here on earth, and haven't yet found.

Yet what inflation says, is that there should be a background of gravitational waves, filling the universe. It's possible or at least conceivable, that these gravitational waves could be detected by certain kinds of observations of the CMB. Remember that when we look at the CMB, we're measuring its temperature as it changes from place to place. Yet since we're measuring photons, we can also measure the polarization of the CMB.

If the inflationary prediction of gravitational waves is true, there will be a very specific kind of imprint on the polarization of the CMB. We have so far detected there is polarization of the CMB, yet our current measurements aren't good enough to test the predictions of inflation. This is one of the things we're shooting for in future generations of experiments to improve exactly those measurements, so we might be able to verify that either inflation or something very much like it was true on the early universe.

That will still leave us with questions. Even if we know that inflation happened, then we're still left with the question of how it happened. What was the universe really like before inflation began? How did inflation start? Why did that little patch become dominated by dark energy? For that matter, what is the inflaton? What is the field that was responsible for the ρ that led to the acceleration we witness in inflation.

All of these are good, open questions. The good news is that we have ideas for them, the bad news is that they're a little bit past our reach in terms of experimental capabilities. So we need to think more cleverly about what could have been going on at the time of inflation, and in particular, think about how to apply those ideas to things we can observe and test in the universe, so that we can turn inflation from a promising speculation into an established part of our understanding of the early universe.