24

OUR FUTURE IN THE
UNIVERSE

J. RICHARD GOTT

This chapter is about the future of the universe. I am going to put salient events in the universe’s history, both past and future, on a timeline. This will involve some enormous times in the far future, and also some very short times in the early universe. What is the earliest time we can speak of in the early universe?

To answer that, we need to answer two related questions: What is the shortest time we can measure? What is the fastest possible clock I can imagine? Every clock, even a quartz watch, has to have something moving back and forth, like the pendulum in a grandfather clock. If I want the fastest possible clock, I just need the fastest thing to move back and forth. What should I use? Light! It’s the fastest thing I can send back and forth. In fact, all I need is that light clock in figure 17.1 with two mirrors and a light beam bouncing back and forth between them. If I want to have it tick faster, what do I do? Bring the two mirrors closer together. The closer together the two mirrors are, the faster the clock will tick. I am going to have one photon in my clock bouncing up and down.

What happens when I make my clock very small? I have a problem. At least one wavelength λ of my photon must fit inside my clock. If the distance between the mirrors in my clock is L, the smallest my clock can be is L = λ. The wavelength and the frequency of the photon are related by λ = c/ν. The smaller the wavelength is, the higher the frequency will be. As I decrease the size of my clock L, I must reduce the wavelength of the photon so it can fit inside, and so I must increase my photon’s frequency. Increasing its frequency means increasing its energy, because the photon has an energy E = hν. And we must not forget Einstein’s equation E = mc2. The energy of the photon corresponds to a certain amount of mass. So, as I make my clock smaller, the energy of the photon goes up, and the mass of the clock increases. Eventually the mass of the clock becomes so large and is compressed into such a small size L, that it falls within its own Schwarzschild radius and forms a black hole! If I try to make a clock that ticks too fast, it will collapse in this way and form a black hole when the length of the clock is about L = 1.6 × 10–33 centimeters and when it ticks once every 5.4 × 10–44 seconds. This time is called the Planck time. It is the shortest time one can measure. The length L = 1.6 × 10–33 cm is one you have heard of before. I have said that the size of the singularity at the center of the Schwarzschild black hole is not exactly zero—it is actually about 1.6 × 10–33 centimeters across, being blurred by quantum effects. This length is called the Planck length, and it is the shortest length one can measure. When I explained that the circumferences of those extra spatial dimensions predicted by string theory might be of order 10–33 cm, this is also the Planck length.

You can’t measure a time shorter than the Planck time. The length of that little time loop that Li-Xin Li and I were talking about in the beginning of the universe might be as short as this (see chapter 23). In fact, if you look at ordinary spacetime on scales of 1.6 × 10–33 cm and times of order 5 × 10–44 seconds, the geometry of spacetime should become uncertain, according to the uncertainty principle. Spacetime should become spongelike and multiply connected when seen at this scale. We can calculate the value of the Planck length, LPlanck = (Gh/2πc3)1/2 = 1.6 × 10–33 centimeters, using the fundamental constants. Here we see all our old friends: Newton’s gravitation constant G, used to calculate the Schwarzschild radius of a black hole; Planck’s constant h, used in calculating the energy of a photon E = hν; and c, the speed of light, used to calculate the mass equivalent of the energy of the photon (E = mc2). The Planck time TPlanck = LPlanck/c is equal to the amount of time it takes a light beam to cross the Planck length. Ignoring factors of order 2 and π, this is the minimum size of the fastest clock before it collapses into a black hole. The mass of this little, fastest clock is the Planck mass, or 2.2 × 10–5 grams, and the density of this little clock is the Planck density, or 5 × 1093 grams per cubic centimeter. This is the sort of density you might get up to in the singularity in a black hole, before quantum mechanics begins to smear things out. The Planck scales are where quantum mechanics comes into play in general relativity, and as mentioned before, we do not yet have a unified model for quantum gravity. Thus the Planck scale (in length or time) represents a limit beyond which we cannot go with our current understanding.

The Planck time, 5 × 10–44 seconds, is the shortest time one can measure, and it is the earliest time we can speak of in the universe. As I have discussed, our universe may be just one bubble (or patch) in one inflating funnel, making up one branch in an infinite fractal tree of universes, making up a multiverse, which may be arbitrarily old. But I am counting time after our little bubble universe has formed. Table 24.1 shows what is happening at each epoch.

As inflation ends, at about 10–35 seconds, the vacuum state that filled the early universe with high-density dark energy decays into thermal radiation. This thermal radiation is very hot and includes not only photons (carriers of the electromagnetic force), but quarks, antiquarks, electrons, positrons, muons, antimuons, taus (a heavier counterpart to the muon), antitaus, neutrinos, antineutrinos, gluons (carriers of the strong nuclear force), X-bosons (hypothetical particles, predicted in some theories, whose asymmetric decays produce the excess of matter over antimatter in the universe seen today), W and Z particles (carriers of the weak force), Higgs bosons (the particle associated with the Higgs field that gives particles their mass), and gravitons (carrier of the gravitational field, just as the photon is the carrier of the electromagnetic field). And if the theory of supersymmetry is correct, there would be supersymmetric partners for each of these particles listed above.

A word about gravitons: Einstein found that gravitational waves, ripples in the geometry of spacetime that travel through empty space at the speed of light, were a solution of his field equations of general relativity. In a similar fashion, Maxwell had previously found that electromagnetic waves traveling at the speed of light through empty space were a solution of his field equations of electromagnetism. We have indirect evidence for gravitational waves (which would consist of gravitons) from Taylor and Hulse’s binary pulsar, which is inspiraling toward a tighter and tighter orbit exactly as Einstein predicted would occur from the emission of gravitational waves as the neutron stars orbited. On September 14, 2015, the LIGO experiment made the first direct detection of gravitational waves. A laser interferometer measured with extreme accuracy (1/1,000 of the diameter of a proton) the distance between pairs of mirrors and noted the tiny oscillations in the distances between the mirrors as the gravitational wave rippled by. How fitting that gravitational waves, predicted by Einstein, would eventually be detected using lasers, since Einstein discovered the principle of the laser as well. The source of these gravitational waves was a 29-solar-mass black hole and a 36-solar-mass black hole in a tight binary orbit that inspiraled toward each other and merged to form a single black hole of 62 solar masses. So, gravitational waves exist, and the results are consistent with gravitons traveling at the speed of light. Because gravity is such a weak force, we have not detected any individual gravitons yet, but we expect they must exist because we have detected gravitational waves, and we expect a wave-particle duality for these, just as we do for electromagnetic waves and photons.

We call this epoch, when all these elementary particles are buzzing around, quark soup. Quarks are traveling freely and are not confined in tight triples. Because of the uncertainty principle, in some regions the quantum vacuum state decays slightly later and in other regions it decays slightly earlier, causing random density fluctuations in the thermal radiation that is created when the quantum vacuum state decays.

These density fluctuations are present at 10–35 seconds, when inflation ends. They form the seeds, which will, by the action of gravity over the course of 13.8 billion years, ultimately lead to the formation of the galaxies and great clusters of galaxies we see today. The spongelike pattern of galaxies we see (figure 15.4), in which great clusters of galaxies are connected by filaments (or chains) of galaxies, is called the cosmic web, and represents the (greatly expanded) fossilized remnants of these early quantum fluctuations made when the universe was just 10–35 seconds old.1

As the universe expands, this hot soup cools and massive particles decay into lighter ones. Initially the universe contains equal amounts of matter and antimatter, but it is thought that asymmetric decays of heavy X-bosons, favoring matter over antimatter, will lead to slightly more matter than antimatter in the decay products. As matter and antimatter particles collide and annihilate in equal numbers to create photons, the remainder becomes dominated by matter. The galaxies we see today are made out of matter. Antimatter particles in the universe today are rare, and always in danger of meeting up with one of the many matter particles and annihilating. Antimatter particles are greatly outnumbered by matter particles today.

At 10–6 seconds, the radiation has become so cool that the quarks bind with other quarks to form protons and neutrons. Quarks come in six different flavors: up, down, strange, charm, top, and bottom. The lightest quarks are the up and down quarks. A proton is formed by two up quarks and one down quark; they are held together by interchanging three gluons between them. A neutron is formed by two down quarks and one up quark, also held together by three gluons. (A way to remember this is that the proton has more up quarks, and up has a “p” in it for proton, while the neutron has more down quarks, and down has an “n” in it for neutron.) The up quark has an electric charge of +2/3, while the down quark has an electric charge of –1/3. Thus the proton ends up with an electric charge of +1 while the neutron is neutral with a charge of 0.

At 3 minutes, helium synthesis occurs, as discussed in chapter 15. The universe has cooled to the point where protons and neutrons can fuse to make light elements. The most common element is hydrogen (a proton), but in addition, an appreciable amount of helium is made, as well as small amounts of deuterium, and lithium. This is the epoch Gamow and his students were using to predict the existence of the CMB.

At 380,000 years, the universe has cooled to about 3,000 K. At this point, electrons can bind to protons to produce hydrogen atoms. This process, as we have discussed, is called recombination. The universe goes from being an electrically charged plasma of mostly electrically charged protons (+) and electrons (–) to being an electrically neutral gas of mostly hydrogen, where each proton has captured an electron to produce an electrically neutral hydrogen atom. Before this epoch, photons were constantly being deflected by either an electrically charged proton or electron—making them execute a random or “drunken” walk. Photons didn’t get very far, being deflected all the time. After the epoch of recombination, photons can travel unimpeded in straight lines over long distances. Because of this change to freely moving photons, we can see directly back to this epoch, when we observe the CMB radiation.

At 1 billion years, galaxies typically begin to form. The high-redshift quasars discussed in chapter 16 come from early-forming galaxies that are seen at a time a little before this.

The universe today is 13.8 billion years old.

By 22 billion years, the Sun will have finished its main-sequence lifetime and will have become a white dwarf. The Andromeda galaxy will have crashed into the Milky Way.

At about 850 billion years, the universe will cool to a constant temperature, due to a process described by Gibbons and Hawking. As discussed in chapter 23, observations indicate that the universe is filled with dark energy characterized by a pressure equal in magnitude to its energy density but negative (dynamically equivalent to Einstein’s cosmological constant). As the matter of the universe thins out due to the expansion, while the dark energy remains at the same density, the universe becomes ever more dominated by dark energy in the far future. Thus the geometry of the universe in the future should resemble that of de Sitter space, a spacetime funnel. It should be ever expanding. Two galaxies that can communicate today will flee from each other faster and faster. Eventually the space between the two galaxies will expand so fast that light cannot cross the ever-increasing distance between them. Event horizons form. A distant galaxy will look to us just like it is falling into a black hole. It will get redder and redder. If extraterrestrials in the distant galaxy were sending a signal saying “THINGS ARE GOING OK,” it would seem to us that they are saying “THINGS A. . . .  . . .  . . . R . . .  . . .  . . .  . . .  . . . E.” We would never receive the end of the signal, “GOING OK.” Events occurring at late times in the distant galaxy would be beyond our event horizon, and we would never see them (recall figure 23.2).

Hawking showed that event horizons create Hawking radiation. Gibbons and Hawking calculated that in de Sitter space at late times, any observers present would see the resulting thermal radiation, appropriately termed Gibbons and Hawking radiation. This thermal radiation, seen in the future of our universe, would have a characteristic wavelength (λmax) of about 22 billion light-years. The CMB radiation continues to increase in wavelength as the universe expands exponentially, doubling its size every 12.2 billion years. After 850 billion years, the CMB thermal radiation will have a characteristic wavelength longer than 22 billion light-years and will become unimportant compared with the Gibbons and Hawking radiation produced by the event horizons. At that point we should see the temperature of the universe stop falling and become constant at a Gibbons and Hawking temperature of about 7 × 10–31 K. That’s very cold, but still above absolute 0 K.

These ideas are actually testable. Gibbons and Hawking radiation is also produced in the early inflationary phase of the universe. This includes both electromagnetic radiation and gravitational radiation. If such gravitational radiation from the early universe is eventually detected via the imprint left on the polarization of the CMB, as discussed in chapter 23, to my mind it would constitute an important experimental verification of the Hawking radiation mechanism. These gravitational waves are not made by moving bodies, as were the gravitational waves detected by LIGO, these gravitational waves would be produced by something different, the Hawking mechanism—a quantum process. So this would be something new and exciting.

The Gibbons and Hawking radiation we expect to see in the far future is ultimately bad for intelligent life. Freeman Dyson once showed that intelligent life could last indefinitely on a finite reserve of energy if it could dump its waste heat in an ever-colder temperature bath. If I showed a movie in a theater at a temperature of 300 K, using visible light photons, it would take a certain amount of energy to show the movie. But suppose we slow everything down in the theater. Suppose the movie was shown using infrared photons having twice the wavelength of visible photons; I could show the same movie using half the energy (each photon would take half the energy) but the movie would last twice as long (because the photons have twice the wavelength). The wavelengths of the photons in the thermal radiation in the theater would also be twice as long, so the temperature in the theater would be 150 K instead of the usual 300 K. Intelligent life could conserve energy by thinking and communicating ever more S . . . L . . . O . . . W . . . L . . . Y. One could even have an infinite number of thoughts using a finite amount of energy by continuing to slow down one’s thinking. This is allowed if one can dump one’s waste heat (which all biological processes, including thought processes, generate) into the ever-cooling microwave background, hibernating from time to time and operating at ever lower temperatures as time goes on. As long as the CMB continues to cool off toward absolute 0 K, that works. But at 850 billion years, the universe will reach an equilibrium temperature equal to the Gibbons and Hawking temperature, and its temperature will remain constant after that. Then one cannot operate at temperatures lower than that to save energy. One would need refrigeration, which uses up the remaining energy very quickly. Furthermore, the other galaxies will have fled beyond the event horizon, leaving only a finite energy reserve at one’s disposal; intelligent life begins to be in energy trouble and will ultimately die out.

Here’s another trouble. At 1014 years, the stars fade as the last low-mass stars run out of hydrogen fuel and die. The universe becomes dark. Only stellar remnants are left—white dwarfs, neutron stars, and black holes. Some planets may still circle them. But by 1017 years, enough close encounters of stars will have occurred to rip the planets from their orbits and fling them into interstellar space.

At 1021 years, galactic-mass black holes form. Two-body gravitational interactions slingshot some stars out of galaxies, while the rest fall into the central black hole. Gravitational radiation causes stars close to the black hole to spiral in.

By 1064 years (if it hasn’t happened already), according to Hawking, protons should decay through a rare process of temporarily falling inside a Planck-sized black hole (via the uncertainty principle), and then having the black hole decay quickly by Hawking radiation. The black hole does not conserve baryons (protons or neutrons)—it does not remember whether it was made of a proton or a positron—but it does remember its electric charge. For that reason, a positron (which is lighter than the proton) can be emitted as one of the decay products of the black hole into which the proton has disappeared. As protons decay, we are left with electrons and positrons as the most massive particles. Protons may decay even earlier than this; perhaps on a timescale of 1034 years, but it’s likely they would have decayed in any case by 1064 years.

At 10100 years, galactic-mass black holes evaporate via Hawking radiation.

What happens after that? The standard picture physicists have is that dark energy, which today is causing an exponential expansion of the universe, represents a vacuum state with a constant positive energy density (and a negative pressure). Steven Weinberg likens our current situation to living in a valley slightly above sea level—our altitude indicating the amount of dark energy present in the vacuum. We have rolled to the bottom of this valley and are just sitting there. The amount of energy in the vacuum—the dark energy—is not changing with time. This will keep the universe doubling in size every 12.2 billion years for a very long time.

Given enough time, it would be likely that our vacuum state, which causes dark energy, could quantum tunnel (through the valley walls) into a lower energy state (of lower-altitude terrain beyond our valley). This would cause a bubble of lower-density vacuum state to form somewhere in our visible universe. The negative pressure outside the bubble would be more negative than the pressure inside, which would pull the bubble wall outward. After a short time, the bubble wall would be traveling outward at nearly the speed of light. It would expand forever. The laws of physics would be different inside the bubble, and you would be killed when the bubble wall hit you.

One can calculate the probability per unit time to quantum tunnel out of the valley to a lower-altitude region outside. We could possibly see bubbles of lower-density vacuum forming in as “little” as 10138 years due to a known instability in the Higgs vacuum. But many physicists think the Higgs vacuum will be stabilized by higher-energy effects. In that case, according to speculative calculations by Andrei Linde, bubbles of lower density vacuum should start forming only after 10^(10^34) years! These bubbles would form, and just like the bubble universes in figure 23.3, they would never percolate to fill the entire space. The ever-expanding vacuum state would continue to double in size every 12.2 billion years and have a volume that would increase endlessly—an ever-inflating sea punctuated by forming bubbles. At late times, our universe would be like eternally fizzing champagne.

Even more rarely, as Linde and Vilenkin have suggested, a quantum fluctuation could cause the entire visible universe to jump up to a high-vacuum energy density and create a new, rapidly inflating high-density inflationary universe. This would be like the high-energy inflation we saw at the beginning of our universe and would initiate a new multiverse. It might be 10^(10^120) years before this happens!

Alternatively, we don’t live in a valley at all, but on a slope, and we will slowly roll down to sea level. This is called slow-roll dark energy. As Bharat Ratra, Jim Peebles, Zack Slepian, and I, and many others, have explored, this would cause the amount of dark energy to slowly dissipate over billions of years, rolling down ultimately to a vacuum state of zero energy density. Just such a rolling down occurred once before with inflation, where a very high-density dark energy state rolled down to the low-energy vacuum that we see today. That could occur again, allowing us to ultimately roll down to sea level—a vacuum energy of zero. These scenarios can be investigated by measuring the expansion history of the universe up to now in detail. This allows us, using Einstein’s equations, to measure the ratio of the pressure to energy in the dark energy, a ratio we call w. If w turns out to be exactly –1, dynamically equivalent to Einstein’s cosmological constant, that favors the “trapped in a valley” scenario, and the dark energy will remain at its present value, and the universe will keep doubling in size every 12.2 billion years forever. If w is less negative than –1, however, we will roll slowly down to sea level, and the accelerated expansion should eventually give way to an approximately linear expansion rate. The universe will continue to expand forever but at a linear rate. In this case the universe’s expansion goes like 1, 2, 3, 4, 5, 6, . . . with time.

A radical proposal, by Robert Caldwell, Mark Kamionkowski, and Nevin Weinberg, is that w could be more negative than –1. This is called phantom energy. It would produce a vacuum energy that increased with time as the universe expanded, leading to a runaway expansion and creating a singularity in the future (a Big Rip) that would tear galaxies, stars, and planets apart in perhaps as little as a trillion years. This “phantom” energy would require a negative kinetic energy in the rolling motion of the field that controls dark energy, which seems to me to be unlikely on physical grounds. That scenario would make the dark energy we see today nothing like the dark energy that was present earlier in inflation. So, although that remains a possibility, it seems less likely to me than the other two scenarios. But many physicists take “phantom energy” quite seriously.2

As discussed in chapter 23, the best estimate of the current value of w (from the Planck satellite team using all available data, including that from the Sloan Digital Sky Survey) is w0 = –1.008 ± 0.068. Remarkably, within the errors, this is consistent with the simple value of –1 (approximating Einstein’s cosmological constant), which corresponds to the model where we are sitting at the bottom of a valley. This result strongly supports the general idea that dark energy represents a vacuum state with positive energy and negative pressure, but these observations are not yet able to truly distinguish between models in which we are sitting still at the bottom of a valley from those in which we are slowly rolling down (or up) a hill. In the latter cases, w0 would be close to, but not exactly equal to –1, slightly above or below it. If future precision measurements of w show that it is unambiguously different from –1, we could learn whether the slow-roll dark energy or phantom energy picture was favored. But, if, as measurements continue to improve and the errors continue to go down, we continue to be consistent with w0 = –1 within the errors, we may well pronounce the “sitting at the bottom of the valley” model triumphant. There are a number of experimental programs either underway or proposed for the future that can potentially lower the errors in w0 by more than an order of magnitude; it is hoped these programs can illuminate the ultimate fate of the universe.

Now you have our best predictions of what the universe is likely to be doing in the future. But what about our future in the universe? What’s likely to happen to us? How is our species Homo sapiens likely to fare in the far future? This is a question we would very much like to answer.

First, I would point out that we are living in a very habitable epoch. The universe has cooled off enough to be habitable, carbon and other elements essential to life have had enough time to form, and the stars are shining nicely today, providing warmth and energy. This is an epoch at which we might expect to find intelligent observers. After the stars have faded, it will be much more difficult for intelligent life. If we look at table 24.1, we find ourselves in a habitable epoch. The Weak Anthropic Principle, an idea proposed by Robert Dicke and later given its name and precise formulation by Brandon Carter, says that intelligent observers should, of course, expect to find themselves at habitable locations—in a habitable epoch in the universe. (Logically, they couldn’t be alive to be asking the question in an uninhabitable epoch!) In fact, we do find ourselves in the middle of what looks like the most habitable epoch in the universe.

But as the only intelligent observers we have encountered in the universe so far, we would like to know how long our future longevity as a species is likely to be. How might we think about this question?

In 1969, I visited the Berlin Wall, which separated the two sectors of the city belonging to East and West Germany. People at that time wondered how long the Berlin Wall would last. Some people thought it was a temporary aberration and would be gone quickly. But others thought the wall would remain a permanent feature of modern Europe.

Figure 24.1 shows a picture of me at the Wall in 1969. To estimate the Wall’s future longevity, I decided to apply the Copernican Principle. I remember thinking: I’m not special. My visit is not special. I am just coming to Europe after college—it was “Europe on 5 dollars a day” back then. I’m coming by to see the Berlin Wall just because I’m in Berlin and the Wall happens to be there. I could have seen it at any point in its history. But if my visit is not special, my visit should be located at some random point between the Wall’s beginning and its end. (The end comes either when the Wall ends or when there is no one left alive to see it, whichever comes first.) There should be a 50% chance, then, that I am located somewhere in the middle half of its existence—in the middle two quarters. If I were visiting at the beginning of that middle 50%, then I would have been 1/4 of the way through its history, with 3/4 still in the future. In this case, the Wall’s future longevity would be 3 times its past longevity. In contrast, if I were at the end of that middle 50%, 3/4 of its history would be past and 1/4 would remain in the future, making its future 1/3 as long as its past.

Rich Gott at the Berlin Wall in 1969. My right foot is in East Berlin, my left foot is in West Berlin, and the Berlin Wall is vertical behind me. Collection of J. Richard Gott

FIGURE 24.1. Rich Gott at the Berlin Wall in 1969. My right foot is in East Berlin, my left foot is in West Berlin, and the Berlin Wall is vertical behind me. Photo credit: Collection of J. Richard Gott

I therefore reasoned that there was a 50% chance that I would be between these two limits and that the future longevity of the Wall would be between 1/3 and 3 times as long as its past (figure 24.2). At the time of my visit, the Wall was 8 years old. While standing at the Wall, I predicted to a friend, Chuck Allen, that the future longevity of the Wall would be between 2.66 years and 24 years.

The Copernican formula (50% confidence level). J. Richard Gott

FIGURE 24.2. The Copernican formula (50% confidence level).

Credit: J. Richard Gott

Twenty years later I’m watching television and I call up my friend and say: “Chuck, you remember that prediction I made about the Berlin Wall? Well turn on your television, because NBC news anchor Tom Brokaw is at the Wall now and it’s coming down today!” Chuck did remember the prediction. The Berlin Wall had come down 20 years later, inside the range of 2.66 years to 24 years that I had predicted. My visit was in the middle of the Cold War, so an atomic bomb could have taken it (and me) out in the next millisecond. In contrast, some famous walls, such as the Great Wall of China, have lasted for thousands of years. My predicted range was rather narrow but it still gave me the right answer.

Scientists generally prefer to make predictions that are more than 50% likely to be right. They like to make predictions that have a 95% chance of being correct. That is the usual 95% confidence level used in scientific papers. How does this change the argument? When applying the Copernican Principle, keep in mind that if your location in time is not special, there is a 95% chance that you are somewhere in the middle 95% of the period of observability of whatever you are observing—that is, neither in the first 2.5% nor in the last 2.5% (figure 24.3).

Expressed as a fraction, 2.5% is 1/40. If your observation falls at the beginning of that middle 95%—only 2.5% from the start—then 1/40 of the history of what you are observing is past, and 39/40 of it remains in the future. In this case, the future is 39 times as long as the past. If you are only 2.5% from the end, then 39/40 of it is past and 1/40 remains. The future is 1/39 as long as the past. If you are in the middle 95%, between these two extremes (there is a 95% chance of that), that makes its future between 1/39 and 39 times as long as its past. Thus:

the future longevity of whatever you are observing will be between 1/39 of its past longevity and 39 times its past longevity (with 95% certainty).

I decided I’d like to apply this to something important, to the future of the human species, Homo sapiens. Our species is about 200,000 years old. That goes back to Mitochondrial Eve, in Africa, from whom we are all descended. The formula would say with 95% confidence that, if our location in the timeline of the history of our species is not special, the future longevity of our species Homo sapiens should be at least 5,100 more years (that’s 200,000/39) but less than 7.8 million more years (that’s 200,000 × 39).3 We do not have actuarial data on other intelligent species (those able to ask such questions), so arguably this is the best we can do. The range of predicted future longevity is as large as this, because one wants to be 95% certain of being correct. Yet many experts who offer their own estimates make predictions outside this range. Some apocalyptic predictions say we are likely to be extinct in less than 100 years. But if that were true, we would be very unlucky to be located at the very end of human history. Some optimists think we will go on to colonize the galaxy and last for trillions of years. But if that were true, we would be very lucky to be located at the very beginning of human history. Thus, even with its broad range, the Copernican-based formula is still highly informative, limiting the possibilities to a tighter range than those considered by many others.

The Copernican formula (95% confidence level). J. Richard Gott

FIGURE 24.3. The Copernican formula (95% confidence level).

Credit: J. Richard Gott

Certainly everything we have learned in astronomy tells us that we should take the Copernican Principle (that your location is not likely to be special) seriously. We started out thinking that we occupied a special place at the center of the universe. But then we realized that ours was just one of a number of planets orbiting the Sun. Then we found out that the Sun was just an ordinary star, not at the center of our galaxy but orbiting at a random location about halfway out. We learned that our galaxy was in an ordinary group of galaxies in a run-of-the-mill supercluster of galaxies. The more we have discovered, the less special our location has turned out to be.

The Copernican Principle is one of the most successful scientific hypotheses of all time, proving itself over and over in a variety of contexts. Christiaan Huygens used it to predict the distances to the stars. He asked why the Sun should be the brightest light in the universe. The stars, he reasoned, were just other suns like our Sun. If the other stars were as intrinsically bright as the Sun (assuming the Sun was not special), then the fact that the stars appear much dimmer in the sky than the Sun must mean they are far away. He figured that the brightest star in the sky, Sirius, was the closest. From his estimated brightness of Sirius relative to the Sun, he calculated it must be 27,664 times as far away as the Sun. He actually got the distance right to within a factor of 20, a remarkable accomplishment, given the large uncertainties involved. Huygens correctly found that the distances between the stars were vast compared to the size of our solar system.

When Hubble saw other galaxies receding from us equally in all directions, he could have concluded that we were at a special place at the center of a massive explosion. But after Copernicus, we were not going to fall for that notion. With so many galaxies, we could not be so lucky as to live in the one in the center. If it looked that way to us, it must look the same way to observers in all galaxies—otherwise we would be special. This led to the homogeneous, isotropic, Big Bang models of general relativity. Gamow, Herman, and Alpher used these to predict the existence of the CMB radiation 17 years before its discovery by Penzias and Wilson. It was one of the greatest predictions to be verified in the history of science. This success was achieved in large measure by taking the Copernican Principle seriously and then following wherever it would lead.

Interestingly, the total longevity of our species predicted by the Copernican formula agrees remarkably well with the actual longevities of other species on Earth. My 95% confidence prediction for the total longevity of Homo sapiens was between 205,100 years and 8 million years (that’s just the 200,000 years we have already had plus the additional 5,100 to 7.8 million years we are likely to have in the future). Homo erectus, our parent species, lasted 1.6 million years, and the Neanderthals lasted only about 300,000 years. Mammal species have an average longevity of 2 million years, and other groups of species on Earth have average longevities of between 1 and 10 million years. Even the fearsome Tyrannosaurus rex went extinct after existing for only 2.5 million years. It was knocked out by an asteroid strike about 65 million years ago.

Keep in mind that my Copernican prediction is only based on our past longevity as an intelligent species—that is, one that is self-conscious and able to ask questions like this—a species able to do algebra, as Neil would say. If we were really going to last another trillion years, we would be very lucky to find ourselves at such an early epoch in the history of our species, a mere 200,000 years from the beginning and, furthermore, at just such an epoch that our past longevity would predict a total longevity in line with other species. If we were observing at some random location in that trillion year history, say, 400 billion years from now, we would already know that our species had lasted far longer than other species and would rightly project a long future longevity for ourselves as well. I would be far more optimistic about our future if the human race were already 400 million years old rather than 200,000 years old, our actual observed longevity so far.

Homo sapiens could, in principle, have a far greater longevity than other species simply because we are an intelligent species. But we are still mammals, and we have a Copernican-predicted longevity quite in line with the longevities of other mammals. Even though mammals are much smarter than the average species, their longevity is not appreciably longer, and hominids (like Homo erectus and the Neanderthals) lasted no longer than typical mammal species. Intelligence and longevity do not seem to be correlated. This should give us pause.

Indeed, if we simply used actuarial data on other mammal species to forecast our future longevity, we would find a future longevity of between 50,600 years and 7.4 million years (with 95% confidence). These limits are within the limits implied by the Copernican Principle based only on our past longevity as an intelligent species. As long as we stay on Earth, we are subject to the same dangers that have caused other species to go extinct, and the fact that we have been around only 200,000 years should make us worry that our intelligence will not necessarily improve our fate relative to other species. Einstein was very smart, but he did not last longer than the rest of us. Intelligence may not be all that helpful for species longevity.

Now you might think that’s fine. Sure, Homo sapiens, will go extinct, but that’s okay, because we will give birth to an even more intelligent species in the future to replace us. Darwin noted, however, that most species leave no descendant species at all. A few species leave many descendant species; they propagate well. But most species die off without progeny. In this regard, note that all other species in our hominid family (including the Neanderthals, Homo heidelbergensis, Homo erectus, Homo habilis, and Australopithecus) have gone extinct. We are the only hominid species left. The rodent family, by comparison, has 1,600 species alive today. They are doing well and have many chances for survival. In a wonderful book called After Man, Dougal Dixon speculated on what might happen after another 50 million years of evolution, and it was not to our liking. We humans were gone in a million years. Fifty million years from now rabbits were still prevalent but they had grown up to be as big as deer and were hunted by packs of ratlike creatures, descended from present-day rodents. The scary thing about this book and the future world it imagines is that it seems so reasonable, and yet it’s clearly not what we would like to hear. None of the animals left on Earth were intelligent observers, able to ask questions like, “How long will my species last?” Of course the chances are low that Dougal Dixon’s specific animal predictions will come true, because there are many ways for evolution to proceed, but the book points out that quite plausibly, most of these ways do not include intelligent observers in the far future. Stephen Jay Gould made a similar point, calling us “just one bauble” on the Christmas tree of evolution.

The same Copernican argument applies to our entire intelligent lineage: our species plus any intelligent species which might descend from us in the future. We are the first intelligent species in our lineage (able to ask questions like this), so the current age of our entire intelligent lineage is only 200,000 years (only 1/65,000 of the age of the universe) and, therefore, our entire intelligent lineage is not likely to go on forever,4 and its future lifetime should have the same limits as those we found for our species. We might well be the only intelligent species in our lineage—given that we observe that we are the first. This accords with Darwin’s observation that most species leave no descendants when they go extinct.

There are some times you shouldn’t use the formula. Don’t use it at a wedding one minute after the vows have been said to forecast that the marriage has only 39 minutes left! You have been invited to the wedding at a special time to witness its beginning. But most of the time you can use the Copernican formula. Since I introduced it, the formula has been tested many times and successfully predicted future longevities of everything from Broadway plays and musicals, to governments, to reigns of world leaders.5 Another exception: do not use it to estimate the future longevity of the universe. You may live in a special (habitable) location, because you are an intelligent observer. (Intelligent observers were not present in the hot early universe and may die out when main sequence stars burn out.) But among intelligent observers your location in spacetime should not be special. In general, the Copernican formula works because out of all the places for intelligent observers to be, there are by definition only a few special places, and many nonspecial places. You are simply likely to live in one of those many nonspecial places. Also, your current observation is not likely to be in a special location relative to the full array of observations made by intelligent observers.

We do not have actuarial data on the longevities of intelligent species in the universe. We have no data on how long they may last. But we do know our own past longevity as an intelligent species, and we should not ignore that salient piece of data. The Copernican formula tells us how to use that information to make a rough estimate of our future longevity with 95% confidence.

If you are not special, you should expect yourself to be born somewhere randomly on the chronological list of human beings. Approximately 70 billion people have been born in the past 200,000 years. The Copernican formula gives a 95% chance that the number born in the future should be somewhere between 1.8 billion and 2.7 trillion. I found out from the referee of my 1993 Nature paper, Brandon Carter of Anthropic Principle fame, that he and John Leslie and Holgar Nielson had also pointed out that it was unlikely that you were in the first tiny fraction of all human beings ever to be born. Carter (and later Leslie elaborating on Carter’s work) used Bayesian statistics to come to this conclusion, whereas Nielson had independently arrived at the same conclusion using the idea that you should occupy a random position on the chronological list of human beings—like my own line of reasoning. I had found like-minded colleagues.

You are likely to come from a country with a population higher than the median. Half of the 190 countries in the world have populations of less than 7 million. But since more people live in the more populous countries, about 97% of all the people in the world live in countries having populations above the median. Were you born in a country with a population of more than 7 million? Just as you are likely to live in a country with a population above the median, you are likely to live in a high-population century. Indeed, you live in a century whose population is the highest it has ever been. You expect to live after some event (like the discovery of agriculture) that caused the population to soar, but before some event that causes the population to fall. You expect to live in a population spike, where the population is larger than that in a median century. This spike could occur at any random point in human history. If you want to know how many people will live after you, ask how many have lived before. If you want to know how long humanity will live in the future, ask how long it has lived in the past.

You are likely to live in an intelligent civilization that is above the median population for intelligent species in the universe—for the same reason that you are likely to live in a high-population country. Most intelligent observers live in civilizations that are above the median, and you are likely to be one of those many observers, rather than among the few who come from civilizations that are below the median in population. That means that our current population on Earth is likely to be above the median population for intelligent species in the universe. That’s not the usual situation in science fiction stories, where a big galactic civilization of extraterrestrials arrives to attack our puny Earth. Although this makes good drama—we are David facing an extraterrestrial Goliath—it does not fit the probabilities. We ourselves are likely to be one of the more successful civilizations in terms of population! High-technology civilizations are likely to have large populations, so we may expect to be one of those.

In 2015, Fergus Simpson of the University of Barcelona pointed out an interesting corollary: since we are likely to come from a planet with a larger-than-median population, most planets inhabited by intelligent observers are likely to be smaller than Earth. Thus, searches for intelligent life, or any kind of life, should focus more attention on planets smaller than Earth—where most of the examples may lie.

We can also make a Copernican, 95% confidence level upper limit for the mean longevity of radio-transmitting civilizations in the galaxy, a value to plug into the Drake equation discussed in chapter 10. This is based on the idea that you are not likely to be special among the intelligent observers living in radio-transmitting civilizations. You are likely to live in one of the longer-lived radio-transmitting civilizations, because they contain more intelligent observers over time. Also, you are unlikely to live at the start of our radio-transmitting epoch. Still, a few radio-transmitting civilizations can always be longer lived than we are, and they contribute to the average. Imagine adding the civilizations to make one long timeline whose length is equal to the total longevities of all civilizations added together. Rank the radio-transmitting civilizations in order of longevity, with the longest-lived radio-transmitting civilization at the end of the long timeline. If you are not special, you should be located randomly in that long timeline, and randomly within the Homo sapiens time segment (giving the total longevity of our radio-transmitting civilization). I was able to use this idea, and some fancy algebra, to set a 95%-confidence upper limit on the mean longevity of radio-transmitting civilizations: 12,000 years. If the mean longevity were longer than that, my 1993 paper would appear either unusually early in our radio-transmitting civilization, or unusually early in the timeline of all radio-transmitting civilizations. This yields a Copernican estimate you can plug into the Drake equation: LC < 12,000 years (with 95% confidence). Neil has used this estimate in chapter 10.

If you think intelligent species typically evolve into intelligent machine species or genetically engineered species, then you must ask yourself—why am I not an intelligent machine? Why am I not genetically engineered?

If you think intelligent species typically colonize their galaxy, then ask yourself—why am I not a space colonist? In 1950, Enrico Fermi asked a famous question about extraterrestrials: Where are they? Why haven’t they colonized Earth already, long ago? The Copernican Principle offers an answer to Fermi’s question: A significant fraction of all intelligent observers must still be sitting on their home planets (otherwise you would be special). Colonization must not occur that often. Importantly, this means that we are allowed to apply the Drake equation in the first place: it estimates the number of intelligent civilizations that arise independently on their home planets. If colonization is not common, that is approximately equal to the total number of extraterrestrial civilizations we will find.

Suppose you thought a priori that each of the following two hypotheses was equally probable:

H1. Humans would stay on Earth until they went extinct.

H2. Humans would colonize 1.8 billion habitable planets in the galaxy in the future.

Bayesian statistics says that you must multiply your prior probabilities for hypotheses H1 and H2 by the likelihood of observing what you are observing, given either H1 or H2. Under the H1 hypothesis that we stay on Earth, then there is a 100% likelihood that you as a human being would observe that you were on Earth. But if humans colonize 1.8 billion planets (i.e., if H2 is true), then as a human being, there is only 1 chance in 1.8 billion that you would find yourself on the first planet out of 1.8 billion that humans lived on. Therefore, even if you viewed the odds as 1:1 initially that we would colonize the galaxy rather than stay on Earth, after considering that you are living on Earth, Bayesian statistics would require you to reevaluate the odds as 1.8 billion:1 against colonization of the galaxy. My Copernican argument would simply say that if you are not special, there is only one chance in 1.8 billion that you would find yourself in the first 1/(1.8 billion) of all planets inhabited by humans and therefore only one chance in 1.8 billion that we would go on to colonize 1.8 billion planets, given that you are on the first one. Nevertheless, our colonizing a few more planets in the future, starting with Mars, would not be that improbable and could give us more chances to survive. We should be doing that quickly, while we still have a space program.

The goal of the human spaceflight program should be to improve our survival prospects by colonizing space. This could be achieved at reasonable cost. For example, you could begin by sending eight astronauts, both men and women, to Mars, and they could multiply there using indigenous materials. You just have to find a handful of astronauts who would be willing to take a one-way trip to Mars and stay there to have their children and grandchildren—people who would rather be founders of a Martian civilization than return to be celebrities back on Earth. It is easy to find such daring people. Story Musgrave, the astronaut I know best, once told me that he would happily volunteer for a one-way trip to Mars. The Mars One group has found a hundred serious candidates who would like to be Mars colonists. Frozen egg and sperm cells could be taken along for genetic diversity. (In this way, even though only a handful of astronauts would actually be sent, many people born on Earth could ultimately have descendants on Mars.) Mars has reasonable gravity (1/3 that of Earth), an atmosphere, water, and all the chemicals necessary for life—unlike the Moon. The atmosphere is CO2, from which oxygen for breathing can be obtained, and water is plentiful in permafrost and in Mar’s polar caps. Radiation levels could be tolerable if the colony were placed 10 meters below ground and the colonists made only brief forays on the surface. Our ancestors lived in caves—so could our Martian colonists. Our orbiters have even found some nice cave mouths on Mars worth checking out.

I have shown that placing such a colony on Mars would require us to put into orbit only as much mass in the future as we have already done in the past, not too much to ask for. According to Robert Zubrin, taking eight astronauts to Mars and supplying them with emergency return vehicles (hopefully not to be used) would require launching 500 tons into low-Earth orbit. From there they would be launched on a trajectory to Mars and aerobrake into its atmosphere prior to landing. According to Gerard O’Neill, a space colony requires as little as 50 tons per person to create a “life in a closed system” biosphere. Delivering these 400 tons to the Martian surface requires launching about 2,000 tons into low-Earth orbit. Thus a self-supporting Mars colony of eight colonists would require launching 2,500 tons into low-Earth orbit. By comparison, the Saturn V rockets of the Apollo program and the U.S. Space Shuttles have launched over 10,000 tons into low-Earth orbit, with the Russian and Chinese human spaceflight programs adding even more. NASA is currently considering building a heavy-lift vehicle capable of putting a payload of 130 tons into low-Earth orbit (a Saturn V-class vehicle). Twenty launches would be sufficient to build the colony (compared with the 18 Saturn V rockets that were built for the Apollo program). Four of these rockets could be constructed at a time in the vertical assembly building at the Kennedy Space Flight Center. If it takes a decade to develop such a rocket, and four are launched in every 26-month launch cycle, the Mars colony could be completed in an additional 9 years. Starting now, the colony would take only 19 years to finish. The human spaceflight program is 55 years old as I write this; the Copernican Principle suggests that funding for the human spaceflight program has at a 50% chance of lasting for at least another 55 years—long enough to establish a Mars colony. Asking for such a Mars colony is not unreasonable. Elon Musk, head of Space X, is interested in privately funded efforts to colonize Mars. I once shared a podium with him at a Mars conference organized by Robert Zubrin. I told of my reasons the human race should want to colonize Mars in the near future, and Elon told how he would go about doing it! Neil has made the case for going to Mars in his book Space Chronicles. Planting a colony on Mars would change the course of world history, in fact you couldn’t even call it “world” history anymore! Stephen Hawking has recently added his voice, saying, in an interview with bigthink.com: “I believe that the long-term future of the human race must be in space. It will be difficult enough to avoid disaster on planet Earth in the next hundred years, let alone the next thousand, or million. The human race shouldn’t have all its eggs in one basket, or on one planet. Let’s hope we can avoid dropping the basket until we have spread the load.”

If Martian couples had four children on average, the population could double every 30 years and reach 8 million after 600 years. (Small populations can grow—the entire original aboriginal population of Australia is thought to have descended from as few as 30 individuals who landed there by raft from Indonesia 50,000 years ago. This population had grown to between 300,000 and 1 million by the time of European settlement.) If you are worried about the funds for the space program being canceled, planting a self-supporting colony is just what you want. Don’t send astronauts to Mars and then bring them all back to Earth. Instead, leave them there, where they can help our survival prospects. Colonizing Mars would give our species two chances instead of one, and might as much as double our long-term survival prospects. It would be a life insurance policy against any catastrophe that might overcome us on Earth, from climatological disasters, to asteroid strikes, to surprise epidemics. It might also double our chances of ever getting to Alpha Centauri. Colonies can found other colonies. The first words spoken on the Moon were in English not because England sent astronauts to the Moon but because they founded a colony in North America that did.

If we look around, we can see the universe showing us what we should be doing. We live on a tiny speck in a vast universe. The universe tells us: spread out and increase your habitat to improve your survival prospects. We live on a planet littered with the bones of extinct species, and the age of our species is tiny relative to that of the universe as a whole. We should spread out before we die out. We have a space program only about half a century old that is capable of sending us to other planets. We should make the wisest possible use of it before it is gone. Will we venture out, or turn our backs on the universe? The fact that we are having this conversation on Earth is a warning that there is a significant chance that we will end up trapped on Earth.

The liftoff. J. Richard Gott

FIGURE 24.4. The Apollo 11 liftoff.

Photo credit: J. Richard Gott

In the summer of 1969, I did more than just visit the Berlin Wall. I visited Stonehenge. At that time, Stonehenge was about 3,870 years old. It’s still there! I also went to Florida specifically to see the Saturn V rocket take off, sending Neil Armstrong, Buzz Aldrin, and Michael Collins to the Moon on Apollo 11. At that time, Saturn V rockets had been taking off for the Moon for 7 months. In another 3.5 years, such Saturn V launches to the Moon would be over. The sight of the Saturn V rocket launching was spectacular (see my photo of it in figure 24.4). As it rose higher and higher, it looked like some magic sword, trailing a plume of fire much longer than itself. I had never seen anything like it. A crowd of about a million people had come to see it. They watched the launch in perfect silence, but after the rocket disappeared into a high layer of cirrus clouds, the crowd let loose with tremendous cheering. Colonizing space is what we should be doing.

Our intelligence gives us great potential, the potential to colonize the galaxy and become a supercivilization, but most intelligent species must not achieve this—or you would be special to find yourself still a member of a one-planet species. We control energy sources that are far less powerful than that of our own Sun. We are not very powerful, and we have not been around for very long. But we are intelligent creatures and we have learned a lot about the universe and the laws that govern it—how long ago it started, how its galaxies and stars and planets formed. It is a stunning accomplishment whose story we have told here.