You're in the market for a house. You tell the real estate agent your budget and start the shows. Naturally, the agent has one in mind that's just a titch above your budget…but I know you'll love it. As you pull up, you get your first good look at the house. It doesn't have that much curb appeal, but it has a certain charm. The entryway is neat and tidy and organized, and the tour begins in earnest.

The agent takes you through the house, and you begin to get suspicious. A gigantic kitchen. Formal and informal rooms. A mudroom. What's a mudroom, anyway? More bathrooms than you can count on one hand. Guest rooms galore. A movie theater. A parlor. A reading room. A breakfast nook. Rooms that only serve as waiting areas for other, larger rooms.

This place is huge. Way bigger than you would have guessed from the street. The agent hasn't told you how much the seller is asking, but you're realizing that this is way outside your budget. There's not a single chance you'd be able to afford even the down payment, let along the mortgage.

You're not even listening to the agent blather on about the granite this and upholstered that. A single, solitary thought occupies your entire consciousness, beating your mind like a drum: Just how big is this place, anyway?

That, my friends, was the collective state of mind by the close of the nineteenth century and the opening decades of the twentieth. There was still heaps of angst over the nature of stars and nebulae and the causes of light spectra, and we'll get to that resolution later, but the central vexing astronomical question in the pre–Great War world was the size of our home.

Astronomers had begrudgingly come to accept the fact that our universe is uncomfortably large—parallax measurements repeated on multiple stars had hammered home the vicious point that our solar system, including the mysterious, distant homes of the icy comets, was just a small, isolated island within the grand ocean of our galaxy. What's more, that galaxy—the massive collection of stars that we can and can't see with the naked eye—filled up the volume of said universe.

Or did it?

It's the natural assumption. Everywhere we look, we see stars. Yes, there might be vast gulfs of night separating the warming fires, but our universe is flooded with those fireballs. Interspersed among them are the nebulae, the vast clouds of dust and gas that serve as nesting homes or violent ends for those ferocious points of light.

The challenge with astronomy a hundred-odd years ago was that parallax is only so good. It took centuries of laborious mental exertion for the instruments and techniques to become sophisticated enough to pin down the first extrasolar distance, and once the community cracked that method, thousands upon thousands more distances were measured, verified, and cataloged. The scale of the universe was just beginning to open up before us.

But there's a limit to what we can measure with photographic plates and polished mirrors, sitting on the surface of the Earth buried under sixty miles of turbulent atmosphere. At a certain point, around a few tens of thousands of light-years, we can't get a reliable parallax measurement.

A light-year, by the way, is the distance light travels in a single (Earth) year. It's a handy unit, credited to Bessel himself for popularizing,1 because it gets severely exhausting describing even the nearest stellar distances in “hundreds of thousands of Earth-sun distances.” Instead, a typical star is, say, forty light-years away. Forty. That's much more nimble to express, so we'll keep it.

So what is a hapless astronomer to do when even the light-years stretch too far and reliable distance measurements are just a fond memory? Well, that old Newtonian idea of using the brightness of a star never quite died away, but with the stunning variety of stars on display in the sky, it always seemed like a pipe dream.

Astronomy was in need of a pioneer, and it found one in Henrietta Swan Leavitt. She had a job as a computer (back when “computer” meant a person crunching numbers by hand and not a machine that does it for you) at the Harvard College Observatory and was tasked with and/or particularly interested in the Magellanic Clouds, two cloudlike (hence the name) nebulae in the Southern Hemisphere. The clouds have been known to various cultures for millennia, but European astronomers only became aware of them in the 1400s, in part thanks to records kept during the world-girdling voyage of Magellan. Somehow his name stuck.

The Clouds contain billions of individual stars, dense stellar clumps, and smaller knots of clouds. Leavitt was particularly focusing on some specific stars within the Clouds that varied in brightness, called, well, variable stars, and here's the game she played.

If you look at a collection of stars, with some looking brighter and others dimmer, it's impossible to tell if those differences are due to variations in their true brightness (i.e., if you could examine each star close up, you would find some are blazing hot and others are dim and quiet) or whether those stars are just at different distances.

It's a mix of both in most situations—except maybe the Magellanic Clouds. They're too far away for parallax measurements, but hey, it looks like all that stuff is all clumped together in the same general vicinity, so maybe it's safe to assume that all the stars we'll find there ought to have about the same distance. (Very roughly, but hey, let's take things one step at a time and see how far we get.)

Of all the different kinds of variable stars (and yes, there are multiple kinds of variable stars), Leavitt was measuring the light output from a class known as Cepheids, named after the prototypical example located in the constellation Cepheus. These are giant stars with brightnesses at least a few thousand times that of the sun, but that brightness is, as you might have guessed, variable, dimming and brightening considerably over the course of a few days or weeks.

Painstakingly, Leavitt analyzed the photographic plates returned from a survey of the Clouds, comparing the same field at different nights, searching for any differences in the intensity of brightness from any of the pinpricks of light. It was mentally exhausting, repetitive work, but Leavitt excelled in dedication, and she identified nearly two thousand of these Cepheid creatures.

It's usually the case in astronomy that the application of large data sets reveals hidden patterns and deeper workings. When you only have a single special case or a handful of examples, it's hard to make sense of the universe. When you have a couple thousand to toss around, it's still hard, but at least you have the serum of statistics to compel nature to reveal the truth.

What Leavitt found by 1912—whether she was looking for it or not—was a remarkable pattern. If you assume that these particular Cepheids are roughly the same distance from us, then if they look brighter, they are brighter. And she must have stared in amazement at her charts and graphs when a simple relationship revealed itself to her: the brighter the Cepheid star, the longer its period between cycles of intensity.2

I know, that doesn't sound that amazing, but it is. This one simple relationship is about to hurtle astronomers—and us—to unimaginable scales.

Here's the deal with the deal. Measuring the parallax of these stars is essentially impossible. Measuring the brightness of these stars as they vary over time is merely insanely difficult—not impossible. And if you can connect the time variation to the true brightness (the astronomers among us switch to a slightly more technical term, the luminosity, to describe the brightness of a star if it can be measured from a fixed distance), then you can compare the true brightness to the brightness as the star appears in the sky.

And following the recommendation of our ancestral astronomers, we can do a little bit of geometry to compute a distance: there is only one number that connects true to perceived brightness. This allows us to go way, way further than crude parallax measures can take us.

Just how far? Well, that was the subject of an intense debate that I'm about to talk about. But all this history, as fun as it is to relate, is getting me hungry for some astrophysics. What is a Cepheid star, and how does it work?

Here's the basic picture, as far as we understand it. Take a giant star and wrap it in a layer of helium. Not hard, since helium is the second most abundant element in the universe. The star will heat up the helium and ionize it, ripping its electrons off. This stripped helium is a little bit opaque—light has a hard time making it through.

And so from our perspective the star looks a little dimmer, as the light from the surface is blocked by the enveloping helium. But that intense heat inflates the layer of helium, which causes it to expand, and in the inevitable process of the physics, it separates from the star and cools off.

Now more cool and collected, the prodigal electrons return to their homes, turning the helium neutral and the gas transparent. From here on Earth, we get to see the full blast of the star's intensity.

But cooling gases contract, right? So over time the helium collects near the surface, where it heats up, ionizes, and turns opaque, and—wax on, wax off—the cycle repeats again.

We're pretty sure. On the plausibility scale this gets a pretty high score, but of course it doesn't quite explain all the observational data.

Here's the hilarious part: it doesn't matter. We could be totally 100 percent clueless about how Cepheids work, with no understanding at all behind the cycles and variations. What matters, when it comes to cosmology, is that the relation between true brightness and period length holds fast. As long as the data support that relationship, we can use it to measure reliable distances. You don't have to know how your microwave oven works to heat up your ramen noodles, after all.

And now back to your regularly scheduled exploration of the universe.

Of course, we still didn't know the distance to the Magellanic Clouds. Leavitt only discovered a relative measure—the longer the period of variation in a Cepheid, the intrinsically brighter it is. And any place in the universe you can measure this variation, you can calculate a distance. But you need an anchor—a first step, using another method, to pin down the nearest Cepheid. From there you can step your way as far as your reliable brightness measures can take you.

Thankfully, a few years after Leavitt's work, a nearby Cepheid was found nestled in the Milky Way, with a confirmed distance using other methods, and the first rung of a distance ladder to the universe could be built.

But in the 1910s, not everybody was liking the taste of the Cepheid special sauce. With the benefit of hindsight and multiple decades of advancement, we now know that this brightness-period relation was on pretty solid ground, but of course at the time there were curmudgeonly skeptics—they thought there was too much fuzziness and uncertainty in the relations to use them reliably to gauge the spectacular distances under consideration. And good that those skeptics raised their contentious objections; they keep us all on our toes.

And so the debate raged on: just how big is this place, and where do we sit? These discussions came to a heated conclusion in a set of lectures on “The Scale of the Universe” held at the Smithsonian Museum of Natural History in 1920. Apparently one of the organizers had suggested an alternative topic, Einstein's theories of relativity, but this was quickly discouraged because not enough people understood it to even make for a decent debate.3

The two debaters, Harlow Shapley and Heber Curtis, represented the two main camps that astronomers had settled into over the past decade. It's important to relate this debate because (a) it provides a case study to set up how astronomers approach controversial issues, which will prove useful when I talk about present-day cosmological controversies, and (b) it's a fun story because Shapley and Curtis were both wrong.

In one corner, we had the Shapley camp, who thought the Cepheids were pretty spiffy and argued that our galaxy is a few hundred thousand light-years across. Plus, by noting that globular clusters (deserted clumps of old stars, or old clumps of deserted stars, take your pick) tended to be found on one side of the sky, we could imply via geometry that the sun is not at the center of the galaxy. But the galaxy, due to its great extent, filled up the entire universe, including all the mysterious spiral nebulae that dotted the evening skies.

And in the opposite corner, there was the Curtis camp, who looked askance at this Cepheid tomfoolery and insisted that our galaxy is small. The best we could do with parallax and other methods was a rough estimate of thirty thousand light-years for the diameter of the Milky Way, and no matter the direction we look, we see the same kinds of stars, implying that we're roughly in the center. But those spiral nebulae surely reside outside our own galaxy. If we assume they're the same size as the Milky Way, then they can only be extragalactic to explain their extent on the sky. And sometimes we see stars flare up—a nova—inside these nebulae. But these novae are far dimmer than those outside of nebulae. Coincidence? I think not. The spiral nebulae are “island universes,” isolated homes to populations of stars far removed from us.

Two arguments, both supported by solid lines of evidence, sound reasoning, and good old-fashioned hard thinking. We can't blame the Shapleyites or Curtisans for taking the positions they did. Both sides had weaknesses in their arguments, for sure, which their opponents exploited with relish. In a way, this single lecture event encapsulated the growing frustration with the cosmos. So much was known, but it fell short of being understood. We were squeezing useful information from our telescopes and photographic plates, but no consistent story was forthcoming.

The universe was sending us mixed signals. Who could possibly sort it all out?

I won't keep you in suspense longer than I have to: it was Edwin Hubble, in the Mount Wilson Observatory, with the one-hundred-inch Hooker telescope.

Hubble found Cepheids, forty of the variable little suckers, embedded within the Andromeda Nebula, the largest (and hence thought to be closest) of the mysterious spiral nebulae. Nobody else had seen them because nobody else had a hundred-inch-wide telescope. But with Hubble at the controls of the biggest, baddest telescope ever made, sitting outside the not-yet-insanely-bright city of Los Angeles, Hubble could resolve features never before seen to humans.

Hubble published his data and necessary analysis (remember, kids, it's important to show your work) in a very readable short paper in 1925, laying out his newfound vision for the cosmos: the Milky Way is indeed large, but it is very far away from Andromeda.4

Hubble estimated the distance between our galaxies to be about nine hundred thousand light-years.

We now know it to be three times greater.

In a single well-written, well-argued, well-researched paper, Edwin Hubble completely repainted the portrait of our universe. And to do so, he needed a much bigger canvas.

Shapley was right in his debate a decade earlier: the sun is not at the center of the galaxy. But he was also wrong: the Milky Way is much smaller than he calculated.

Curtis was also right and also wrong: he got roughly the right size for our home galaxy but missed our location within it.

It took more data and a new round of better instruments to finally sort through the confusion, and once again the implacable data showed that the universe obeys a sort of vicious and amped-up version of Copernicus's original thoughts: We are not special. We are not at the center. And the cosmos is far, far larger than we can be reasonably comfortable with.

The Milky Way galaxy is but an island of stars—hundreds of billions of stars, but an island nonetheless—separated from our nearest neighbor by vast gulfs of almost absolutely nothing.

Just a few centuries earlier, Brahe thought it absurd to place the stellar sphere seven hundred times farther than the orbit of Saturn. And here was Hubble, sitting underneath his gigantic telescope, recording the variations in brightness from a point of light sixteen trillion Saturn-orbits away from us.

The Andromeda Nebula was no banal collection of gas and dust with a dash of errant stars loosely tossed within it. It was a galaxy in its own right, one of an uncountably large multitude, spread throughout the achingly large cosmos the way a child might leave toys scattered in a room.

Nobody was really thrilled at the result. Hubble demonstrated that even the “large” universe as advocated by Shapley was too small. Let that sink in: the Cepheids used in Edwin's analysis were farther away than even the farthest possible thing a reasonable person could argue existed.

And Andromeda wasn't alone. Now that Cepheids could be reliably used to measure extragalactic distances (and “extragalactic” was now a thing), many other distances were pegged to other nebulae-cum-galaxies. Andromeda has pride of place of being the first, but it was far from the last.

But the fun didn't end there. Oh, no, child.

Not satisfied with simply rescaling our universal yardsticks and taking the first crack at a truly cosmological birds-eye view of our home, Hubble took it one step further and completely revolutionized our understanding of the dynamics of the universe writ large.

Remember how much intellectual inertia had to be overcome to finally conclude that the heavens were as violent, chaotic, and messy as the physics here on dear old Earth? Centuries, that's how long. In those intervening generations, scientists played a sort of slippery rhetorical bait and switch. Sure, stars can blow up, change their brightness, and even scoot around. Whatever. But the universe is eternal and never changing.

Individuals may come and go, but life goes on, forever into the past and forever into the future. It's the way it has been, the way it is, and the way it always will be.

Academics replaced the fixity of the firmament for a constancy of the cosmos. It's the same thinking, just with a few zeros tacked onto the end of all the numbers.

And yes, Hubble had to go and pop that bubble too.

In 1929, four years after his landmark presentation on the true distances available in our universe, he published another, highly readable paper in which he reported an interesting relationship between the distance to a galaxy and, of all things, its speed.5

Remember the spectral Doppler technique used to measure the motions of stars? (If you don't, you weren't really paying attention to chapter 3, were you?) The beauty of that method is its brutal universality. Find a star, measure its speed toward or away from us. Boom, done. Move onto the next.

Find a nebula, measure the light. Recognize any emission or absorption lines from your favorite element? Is the fingerprint correct but shifted left or right from its Earthbound counterpart? Congratulations, you've measured the speed of that nebula—even different parts of it!

Is that “nebula” really a gigantic galaxy, home to hundreds of billions of stars, as big as or bigger than the Milky Way, located hundreds of thousands of light-years away? Who cares? It's emitting light, so we can take a spectrum, either from individual stars, if we're lucky enough to resolve them, or from the generic glow of the galaxy itself.

Recognize the elements? Fingerprint shifted? You've measured the velocity.

If it's shifted toward the blue, it's coming toward us. If it's shifted toward the red, it's moving away.

Of course, just like with stars, this technique only returns the radial speed, the speed along our line of sight to the object. In other words, the speed along the in-out direction. Who knows what the up-down or left-right speed is. But hey, it's something, and we'll take it.

Specifically, Hubble took it, twenty-four times. It was the best he could do through painstaking (Have I used that word enough to describe the procedure of astronomical observation? No, I haven't.) work collecting, measuring, recollecting, and remeasuring the tediously extracted distance measurements from the Cepheids, then matching those distances to velocity measurements taken by himself and previous astronomers.

What he found was simple, surprising, and essentially inarguable: galaxies, on average, are receding away from us. And the farther away a galaxy sits from us, the greater its redshift, implying the faster it's receding from us. And this relationship is linear: double the distance, double the speed. Quadruple the distance, quadruple the speed.

Hubble could even provide a number for the recession rate: five hundred kilometers per second per megaparsec.

As we'll soon see, the numbers and counting systems we're accustomed to quickly become too cumbersome to have any utility, a fact that astronomers quickly realized probably just after the phrase astronomically large came into circulation. The standard go-to is the light-year, which as you recall is the distance that light can travel in a year—5,878,499,810,000 miles, for the curious.

But astronomers typically use a different measure: the parsec. Why? I honestly don't know—maybe because it sounds cooler, and maybe because the term light-year was initially used more for the benefit of the popular imagination and not for use by Serious Astronomers. But anyway, parsec is short for parallel arc second and comes in handy when measuring astronomically large distances. If you hold a pencil in front of your face while closing each eye individually, the pencil will appear to wiggle back and forth relative to the distant background. If you hold the pencil farther away, it will wiggle less.

If you know how far apart your eyes are, and you measure the amount of wiggle, and you know about trigonometry, you can calculate a distance. This is the parallax method.

That isn't so useful for interstellar measurements, so instead of alternating eyes, astronomers alternate seasons, repeating measurements when the Earth is at opposite sides of its orbit around the sun. This is the precise technique that Brahe used to cast doubt on the sun-centered model of the universe and Bessel used to give him a run for his money. And here comes the definition: a parsec is the distance an object has to be at in order to wiggle by one arc second (1/360th of a degree, for those not nautically inclined) when observed six months apart.

One parsec is roughly three and a quarter light-years, and Proxima Centauri, our nearest neighboring star, happens to be around one parsec away. Hmmm, I wonder why that definition was chosen?

And just as your computer can have megabytes and gigabytes, representing a million and a billion bytes respectively, distances can have megaparsecs and gigaparsecs. Start tossing those kinds of words around, and people won't even know you're not a real astronomer.

As Hubble's result implies, we're going to use those: the Andromeda Galaxy, née Nebula, our nearest major neighbor, is about three-fourths of a megaparsec away from us.

So here's what Hubble's calculation reveals: for every megaparsec you get away from the Milky Way, objects at that distance are receding from us an additional five hundred kilometers per second.

At some level, it's not surprising to know that galaxies zip and zoom around. People do it, planets do it, stars do it. Why not the largest collection of all? But something fishy was going on with Hubble's measurements. In the peculiar jargon of astronomy, the individual velocity of any particular galaxy is called its peculiar velocity. So fine, the peculiar velocities of galaxies aren't zero. I suppose it's a little bit of a big pill to swallow to contemplate the motion of these humongous cosmic assemblages, but it's one we can take.

It's the average motion that's troubling. There appeared to Hubble to be a separate, common (dare I say universal?) motion to the galaxies around us. And specifically, away from us, at the very precise rate of five hundred kilometers per second for every megaparsec in distance. This was certainly something new and potentially troubling. Given the raw data, how are we to possibly interpret this result?

The revelation of two Hubbles. Left, a modern view of the Andromeda “Nebula,” which Edwin Hubble conclusively demonstrated was really, really far away. (Image courtesy of NASA / JPL-Caltech.) Right, a long stare with the Hubble Space Telescope at a small patch of sky—equivalent to the small square next to the Moon—reveals a universe infested with these beasts. (Images courtesy of NASA / ESA.)

Option 1: A conspiracy. Galaxies move around randomly, and they just so happen to have the right velocities so that a galaxy, say, twice as far from the Earth is moving twice as fast as its nearer cousin. And they're all moving away from us. Maybe we're the center of the universe, and we're somehow repulsive?

Option 2: An illusion. Astrophysicist Fritz Zwicky, who really knew how to rock a bolo tie and whom we will meet again later, tossed this idea into the discussion. Maybe light just gets tired, like an out-of-shape guy trying to run a marathon. Who knows what the mechanism is, or what the physical implications might be? But in this case perhaps the general redshifting isn't an indication of motion but a loss of energy. Redder light is less energetic than bluer light, so this hypothesis is still quite capable of fitting all of Hubble's data. The recession is a fake; it really isn't motion but an artifact of our poor understanding of physics.6

Option 3: We live in an expanding universe.

You and I know that the answer is door number 3, but it's not easy to flesh out that deceptively simple but radical statement in a single paragraph, so it gets its own section.

One of the most amazing aspects of this saga of the 1920s is how quickly the resolution came. When Copernicus and Kepler first proposed a sun-centered universe, it took another couple of generations before Newton could offer a unifying theoretical theme, a coordinating force (“universal gravity”) that could sufficiently explain the motions and models offered by earlier thinkers.

But in this case, the theoretical basis for Hubble's observations was happening simultaneously—and had even been anticipated before he got the result! That theoretical basis is general relativity, and if you've been wondering when dear old Albert would get fully introduced into the story, well, here he is.

Einstein had long been attracted to the problem of gravity (get it?), and especially to Newton's troubled comments that he didn't fully understand why his relationship holds between massive bodies in the solar system, but that it nonetheless works, so it must have some utility.

I'll have to save a full and proper treatment of Einstein's work and legacy to another book—after all, he did basically found a few distinct branches of modern physics—so here's the short and sweet version (as short and sweet as I can make it7).

General relativity is Einstein's magnum opus, a completely radical take on the underpinnings of gravity, explaining it not as a force per se but as an effect. Instead of gravity being an instantaneous, invisible communication between massive objects, the effects of gravity in Einstein's picture are actually the consequence of a relationship between mass and space-time itself.

Right, space-time. Not space and time. Not space or time. Space-time. A single, unified thing. We don't live in a three-dimensional world; we act our little plays in a four-dimensional stage, with our three spatial dimensions (up-down, left-right, back-forth) and a fourth dimension of time (past-future). This was the revelation of special relativity, one of Einstein's first forays into reshaping our worldview in 1905, but it was another scientist (and Einstein's former teacher), Hermann Minkowski, who later made the leap into unifying the dimensions.⁸

I mention this because you're going to be seeing the word space-time a lot, and I need to explain what it really is: It's a ruler. It's a way of measuring distances (which doesn't sound like a big deal) in both space and time (which does sound like a big deal). If we agree to meet for coffee at four in the afternoon, the total description of the location (“coffee shop at 4:00 p.m.”) is a precise event, and the construction of space-time allows me to compute how far I have to travel to get to that event: I have to move, say, 3.5 miles west, and I have to wait until it's two hours into the future before the chitchat over lattes can occur.

You can imagine space-time as a grid (in four dimensions, so good luck), marking out locations throughout the entire universe, where “entire” also means the distant past and future.

It's a dance floor. The particles, forces, fields, and energies that populate our universe are the dancers, twisting and twirling away in complicated rhythms and beats. But the space-time floor stays fixed.

Or at least it did in the decade between 1905 and 1915, when Einstein brain-birthed general relativity. The special version recognized certain rules that applied throughout the universe: the speed of light is the fastest anything can travel; moving clocks run slow; different observers will disagree about lengths and intervals; mass and energy are two sides of the same coin; and so on. To make special relativity happen, Einstein had to chuck out the Newtonian framework of gravity; it simply wasn't compatible with the new set of rules.

For example, what if the sun vanished? Light takes eight minutes to make the leap across the vacuum from our star to our squinting eyes, so we wouldn't know the sun had disappeared until eight minutes after the event occurred. But would the orbit of the Earth change in those eight minutes? Would the inertia of the Earth “know” instantaneously, while light lagged behind?

Einstein didn't think so (and to be fair, had Newton been aware of the issue, he probably wouldn't think so either). Rather, something had to carry inertia, to take it from place to place in the universe—to connect motions across the vastness between us.

So what infrastructure exists that permeates the cosmos, allowing all the particles, forces, fields, and energies to interact with each other?

Bingo: space-time.

It turns out the dance floor isn't as solid as we thought it was. It doesn't just stay there, a rigid platform for the drama of the universe to play out on. Using a mathematical tool kit developed in the nineteenth century by Bernhard Riemann, Einstein was able to formulate a view of the universe where the floor—space-time itself—bends, warps, flexes, and curves.

Imagine yourself gliding down the floor in a smooth-as-butter waltz (or a sassy hip-wiggling salsa, if that's more your flavor), but the floor is a trampoline. Your very presence bends the floor underneath you, as it does to all the other dancers. Negotiating the limited space in the crowded room is a tricky thing. If you even get near the other dancers, their dance-floor depressions alter your course, veering you away from your intended movement.

And that, my friends, is the Einsteinian picture of gravity. Except in four dimensions. Sorry, folks, but analogies can only take us so far in a space-time world. The presence of matter (and energy!) distorts space-time around the object, bending it. Any other matter (and energy!) encountering that object will have its motion disturbed by that deformation in space-time. That is our gravitational experience.

This picture answers the vanishing-sun riddle: if the sun were to poof out of existence, it would take a while for space-time to “relax” back to its flat state and for the Earth to be released from the gravitational grip of the sun, so our planet would be flung out like a spinning rock cut from its string at the same time that the familiar light in the sky would wink out of existence.

So special relativity gave the world a language of space-time, and general relativity taught us that space-time itself is a dynamic, living, breathing, physical object. It's still a ruler—it's still very good at measuring intervals between events—but that ruler can stretch, flex, and bend.

The game of general relativity is then pretty straightforward. In words, that is; in the mathematics, it's insanely complicated. Gravitational interactions are formulated as a set of ten interconnected nonlinear equations, with one side of the problem describing all the possible ways that space-time can bend, flex, and twist and the other side of the problem describing all the ways the matter and energy can bunch together flow, and twist.

So in most cases you take a given physical system—say, the solar system. You count up all the matter and energy sources you care about (e.g., the sun and planets), you turn the GR crank, and out pops a configuration for space-time. Then you apply what's called an equation of motion to, well, figure out how the objects ought to move. And boom, you've got some dynamics.

What does this game have to do with the universe? Well, the universe can be considered as a single, physical system. It contains a certain arrangement of particles, forces, fields, and energies. All that stuff, when studied from the ultimate long view, smoothed out across the entire universe, will bend the cosmos at those very largest scales. In other words, the contents of our universe will bend creation itself, and that bending will influence motion at the largest scales—say, by sending galaxies flying away from each other.

Einstein was the first to apply his tools of general relativity to questions of cosmology in 1917, just a couple of years after formulating the methods in the first place.9 To be fair, he was only one of only a handful of people who actually understood how to do it, so it might as well have been him.

But 1917 was twelve years before Hubble's result, and Einstein assumed, just like everybody assumed, that we live in a static, eternal universe. The firmament was fixed, just like the ancients thought, but the word “firmament” had taken on a much larger definition.

What's interesting is that general relativity didn't automatically predict a static universe—left to their own devices, the equations naturally suggest a dynamic cosmos, one that's inclined to expand or contract but not stay still. Well, that wasn't going to work. No way was the whole entire universe moving around. So to develop a halfway decent model of the static universe as Einstein and everybody else knew it, he had to plug a somewhat awkward “bonus term” (not his words) into the equations.¹⁰

It was a perfectly reasonable decision to make at the time, and nobody gave him any gruff for it. But imagine for a moment a world where Einstein didn't feel compelled to fit the known data—where he let the simplest possible expression of general relativity predict what the universe ought to be like. He would have come out with the dynamic universe a full decade before observations would have backed him up on his bold claim.

Man, he could have been famous.

Thankfully, Einstein wasn't the only one thinking of cosmological problems and using the general relativity tool kit to think those thoughts. Many other theorists toyed and tinkered with Einstein's equations, including Willem de Sitter (a Dutchman with a pointy beard), Alexander Friedmann (a Russian with a caterpillar mustache), George Lemaître (a clean-shaven Belgian Catholic priest), Bob Robertson (an American with the tiniest mustache you've ever seen), and Arthur Walker (an Englishman with no beard).

Their story of attempting to use general relativity to describe the whole universe is long, intricate, and intertwining. The most important point is that it demonstrates that no matter your choice of facial hair arrangement, you too can be a theoretical physicist.

It also provided the mathematical framework for modern cosmology. In an expanding universe, the galaxies only appear to be physically rocketing away from each other. In fact, the fabric of space-time itself stretches like pizza dough, causing every galaxy to separate from every other galaxy (ahem, on average, and that caveat will have some interesting consequences for later chapters). The redshift that Hubble noted thus is due to not the motion of galaxies but the stretching of space-time. Make no mistake; the galaxies really are getting farther away from us, but not on their own agenda—the increasing gulfs between us are like the tectonic spreading of oceans between continents.

As light travels from a distant galaxy to, say, the aperture of a hundred-inch telescope, the expansion of the universe stretches out the light. Farther galaxy = more intervening universe = greater stretching = more redshift. Perfectly matching Hubble's results. No intergalactic conspiracy needed.

The gist is that by the time Hubble announced his findings to the world in 1929, he had done his homework and knew enough to name-drop Willem de Sitter's work in his paper as a possible explanation for the results—that we live in an expanding universe.

Modern cosmology can thus trace its lineage to two founding fathers. One was Einstein himself, the theorist's theorist, who was profoundly unconcerned with the experimental results of tests of his theories—what else could they possibly find but that he was correct? A solid mathematical argument could sway his thinking, but data would only serve to validate his reasoning. Indeed, the one time he toed the observational line by assuming a static universe, he modified his equations in a move he would later call his “greatest blunder.”

The other was the pipe-smoking, eagle-eyed observer of the cosmos, Edwin Hubble. The champion of solid analysis, good statistics, and simple but powerful writing, Hubble was extremely cautious about offering theoretical explanations for the amazing results he achieved—it was only in the closing paragraph of his landmark distance-velocity paper that he remarked that the results might be explained by an expanding universe.

In other times, both past and present, observers and theorists are often at odds, either leapfrogging each other, sneering at the opposite camp in the rearview mirror, or straight-out devolving into fistfights.

But something magical happened in the decades surrounding the world wars—for a short period of time (cosmologically speaking, and also in the sense of the timescales of human achievement and growth of understanding), those who collect data and those who try to explain data were in almost perfect lockstep, sometimes even publishing papers side by side in the same journal issue.

The universe was starting to come into focus. For a brief moment, everything felt good.