part0003

It sounds a long and complicated story, and in a sense it is. But by the standards of stellar lifetimes, the later stages are all over in the blink of an eye. Even one of these massive stars may spend tens of millions of years burning hydrogen quietly, and shining like our Sun, then millions of years more as a helium-burning red giant. But in a star with, for example, twenty-five times the mass of our Sun, carbon burning takes just six hundred years, neon burning lasts for about a year, and oxygen burning is over in six months.* The final stage, silicon burning, is over in less than a day.

There are three reasons for this frantic speeding up of stellar processes. First, at each stage the star collapses inward a little more; its core gets hotter, and the nuclear reactions become more violent. Second, the later nuclear fuels are less efficient. A single nucleus of silicon 28, for example, weighs almost as much as 28 of its hydrogen nuclei (protons) the star began with, but can release far less energy. Third, such high temperatures are reached that neutrinos can be created; these escape freely, removing energy far faster than the stellar surface radiates.

Energy is the key to what happens next. All the energy the star has derived from nuclear reactions comes from packing protons and neutrons more tightly to-

*A11 the figures in this section are based on computer models, which calculate how stars evolve subject to the known laws of physics. They are calibrated by comparing the computer "predictions" with the appearances of different kinds of real stars, with different ages and masses. The figures are consequently quite reliable.

gether in atomic nuclei. In iron-56, they are packed as tightly together as possible, and no more energy can be provided by fusion. More massive nuclei, such as gold, lead, silver, and uranium, are less tightly packed than iron-56. To make them out of iron, more energy has to be put into the system. This, among other things, is what happens in a supernova.

The Supernova Connection

For the sake of argument, we will carry on describing what happens to a star with a mass of about twenty-five suns after silicon burning is complete and it is left with a ball of iron, about as massive as our Sun, in its centre. Only the details are different for stars with different masses. The star now has no effective means of support, since there is no more energy being produced by nuclear burning in its core. The result is dramatic. The inner regions of the star are squeezed inwards, and the pressure on the iron nuclei in the core becomes so great that electrons and protons are forced to merge into one another, forming neutrons. A ball of neutrons can indeed pack matter together more compactly than a ball of iron nuclei, and the centre of the star begins to convert into a neutron star, still with as much mass as our Sun but now occupying only as much space as Mount Everest. It becomes, in effect, a single "atomic" nucleus. Material from the inner part of the surrounding star has the floor pulled from underneath it, and plummets down onto the newly forming neutron star, reaching speeds as great as 15 percent of the speed of light. When this fast-moving material hits the neutron star, from all sides at once, the shock actually squeezes the ball of neutron material, like a golf ball being squeezed in an iron grip. But neutron stuff is very difficult to compress—it is like trying to squeeze the

nucleus of an atom—and quickly bounces back from this compression. Enormous pressures and temperatures are created in the bounce, which turns the shock wave inside out and sends it speeding back out through the giant star.

Everything has happened in less than half a second. As the shock begins to move outwards through the star, which may have a diameter of 700 million kilometres, as big as the orbit of Jupiter, it encounters resistance and begins to slow down. It is, after all, trying to move bodily about twenty-four solar masses of material. Without help, it would fizzle out. But it is followed by a flood of neutrinos produced in the neutron core of the star when it was squeezed by the infalling matter. The matter in the slowing shock wave is so dense that it actually absorbs a significant number of neutrinos. The energy from the neutrinos gives the shock wave the boost it needs to finish the job of blowing apart the outer layers of the star.

In all this energetic activity, quantities of elements heavier than iron have been formed, and many complex nuclear interactions have produced a variety of other elements from the basic products of nuclear burning. A supernova shines, for a few weeks, as brightly as a whole galaxy of normal stars put together, and the energy that makes it so bright comes from radioactivity, from unstable elements heavier than iron that were put together by the shock wave and are now breaking apart, releasing energy and forming more stable, tightly packed nuclei. From the site of this intergalactic beacon, much more than twenty times the mass of our Sun, in our specific example, is expelled completely into space, and carries with it this heavy-element legacy from the dying star. The core, at last free from the burdensome pressure of the rest of the star, settles down as a spinning neutron star, perhaps to be detected by some civilisation as a pulsar. The organic

beings that study such pulsars, and the steel girders of which their radio telescopes are manufactured (not to mention the silicon in the chips of their computers), are equally the products of supernova explosions in aeons gone by.

The story is fascinating in its own right. But where is the anthropic coincidence? It lies in that burst of neutrinos, the crucial step in helping the shock wave to blow the star apart. Computer calculations in the 1980s had shown that the shock wave alone simply could not do the job, and that neutrinos must be involved. But some researchers were sceptical, because the properties of neutrinos had to be precisely "fine-tuned" to do the job. It all hinges upon the strength of the weak interaction, one of the four fundamental forces of nature. This is the force that determines how strongly neutrinos interact with baryons. If the weak interaction were a little too weak, then even the dense shock wave would be transparent to neutrinos, and they would flood out through the star without getting involved in pushing apart the outer layers of the star. If, on the other hand, the weak interaction were a little too strong, then the neutrinos would get involved in reactions in the core itself, and would never get out to the region where the shock wave was slowing down and giving up the ghost. The weak interaction has to be just right to allow enough neutrinos both to escape from the core and to interact with the shock wave.

Some doubts about this scenario for the explosion mechanism were allayed by studies of the burst of neutrinos from supernova 1987A. The energy of these neutrinos, and the implication, from the arrival of a handful in our detectors, of how many escaped from the core of the supernova, match the requirements of the models. Studies of the supernova match very well with the computer calculations, supporting the view that neutrinos are indeed the driving force in expelling

large quantities of gas enriched with heavy elements into space—a phenomenon without which no planets like Earth or creatures like us would exist.

A Cosmic Connection

The same coincidence crops up earlier in the life of the Universe. It is the strength of the weak force that decides how much hydrogen is processed into helium in the Big Bang. It requires a rather precise fine-tuning to avoid a runaway in one direction or the other—make the force slightly stronger and no helium would have been produced; make it slightly weaker and nearly all the baryons would have been converted into helium in the Big Bang. A universe in which stars were initially made only of hydrogen might not be so very different from our own; but if all the stars were originally composed of helium, they would have burnt out more quickly, perhaps not giving life time to evolve on any planets that formed (if life can indeed develop without hydrogen present to form water). The condition that some stars go through a supernova phase (triggered by a neutrino-boosted shock) is essentially the same as the condition that there be an interesting amount of cosmological helium production. The weak force seems to be just about as weak as it can be in order to avoid all the original hydrogen being converted into helium. Supernovae might still work (exploding by a different mechanism) if the force were a little stronger, but if the force were weaker the neutrinos could not drive any kind of explosion; the Universe would be even more comfortably dominated (baryoni-cally speaking) by hydrogen if the force were a little stronger. But the window of opportunity for a universe in which there is some helium, and exploding supernovae, is very narrow.

These examples are enough to demonstrate the power of the coincidences at work in our Universe. But there is another level at which we can contemplate the puzzle of our existence. So far, we have taken the framework of our Universe for granted. We have talked of tinkering with the weak force, or the constant of gravity, and we have happily discussed bending and stretching spacetime. But is the fabric of the Universe itself something unique and special? Can we read any significance into the fact that we live in a world built up from three dimensions of space and one of time?

Space, Time, and the Universe

Take time first. At the Planck time, we have to jettison the whole idea of an arrow of time, and even that there are three dimensions of space and one of time. We can ask why there was a Big Bang—is there something uniquely self-consistent about the way it happened, or did it have coincidental features without which we would not be here—but we cannot ask what happened "before" the Bang. There's a sense in which time itself begins with the Big Bang. Similarly, time will end if there is a big crunch; and even if the Universe expands forever, time would end for any observer who is swallowed by a black hole and pulled into the central singularity.

All this assumes that time is measured by the ticking of standardised clocks. But we face a conundrum if we try to imagine how we could measure the time that elapsed before there was a big crunch. Any conceivable clock would be destroyed at sufficiently high density. We might imagine starting off measuring time in years, by orbits of planets around stars. When the density gets high and solar systems are destroyed, we could use an

atomic clock. But eventually atoms themselves would be destroyed. In approaching the singularity we would need to rely on an infinite sequence of successively smaller and sturdier clocks.* This line of thinking, however, is based on an infinite regress that is almost certainly unwarranted—time is not infinitely divisible.

Just as there may be a limit, measured in tens of billions of years, to the longest timespan that is meaningful (the time from Big Bang to big crunch), so there may be a smallest natural unit of time. Conventional physics does set such a limit. The Heisenberg uncertainty principle tells us that if we want to measure a short time interval with increasing precision, we need to use quanta of radiation with more energy, and shorter wavelengths. Because the light quanta move at finite speed, this increasing amount of energy must be fo-cussed into a smaller and smaller space. A limit arises when the concentration of energy is so great that the quantum collapses into a black hole. This occurs at about 10" 43 seconds, the Planck unit of time. It is not possible, according to quantum theory, to place events in chronological order with any greater precision than this; some physicists suggest that there may actually be a fundamental limit bigger than the Planck time, although experiments tell us that the graininess of time is certainly not on scales larger than 10" 26 seconds.

Another fundamental mystery is "time's arrow/' the irreversibility of the flow of time, from the Big Bang into the future. The laws of microphysics are reversible in time, and if we took a movie of microphysical interactions and ran it backwards we would not, in general, be able to tell that time had been "reversed." The

*This reminds us of Zeno's paradox, "proving" that motion is impossible: before an arrow can reach its target, it must first get halfway there; before that, it must get a quarter of the way; before that... and so on, in an infinite number of steps that must be completed before the arrow can wing its way.

macroscopic world is by no means time-reversible in this sense. Things age and wear out—and this ageing and wearing out of things is described by physical laws of thermodynamics. We have memories of the past only, not the future, and in general it is easier to be wise after the event than to predict the future.

The same arrow of time also seems to apply to the expanding Universe, with the Big Bang in the past. "Later" times are times when clusters of galaxies have moved farther apart, and this is a universal arrow of time that could be deduced by intelligent observers anywhere in the Universe. What would happen if the Universe were to halt its expansion and begin to recollapse? The big crunch, the counterpart to the Big Bang, would then lie in the future, not the past. "Later" times would be times when clusters of galaxies were closer together—or would they be? Thomas Gold, in the 1960s, was one of several theorists who speculated that if this happened time would run backwards. Does that mean we would "remember" the future? Would intelligent observers in a collapsing universe still think that "later" times were when clusters were farther apart?

This conjecture seems highly implausible. Nothing special happens locally, anywhere in the Universe, at the epoch of maximum expansion. Nevertheless, the idea has been taken seriously, and has been revived in the 1980s by Stephen Hawking and other researchers trying to develop a quantum-mechanical description of the Universe. Paul Davies, of the University of Newcastle upon Tyne, has recently suggested that this casts doubt on one of the most fundamental laws of physics.

This is the second law of thermodynamics, which states that disorder (entropy) is always increasing. Hawking showed that a black hole possesses entropy, and that the area of the surface of the hole is a measure of its entropy. If entropy can only increase, that means that a black hole can only get bigger (except when it is

evaporating, a quantum process that itself increases entropy). If the Universe is closed, to form the three-dimensional equivalent of the surface of a sphere, then in many ways it resembles a black hole viewed from the inside. But if the Universe is destined one day to have its expansion stop and recollapse, that implies that the "area" of the "black hole" will one day shrink— that entropy decreases and the second law of thermodynamics does indeed break down, in regions of space and time that are as normal as the Universe today. But there may be a way out of this dilemma. Roger Penrose, of the University of Oxford, has developed the idea that time's arrow is related to the difference between the dynamics of the Universe in the Big Bang and the big crunch. For reasons that are not yet understood, the Universe emerged from the Big Bang in an amazingly smooth state. Penrose argues that this makes the initial singularity a special and unusual one. The big crunch, on the other hand (if it happens), will be much more messy, chaotic, and unsynchronised. Crumpled regions of spacetime that have already formed black holes in the Universe will clump together, crumpling space still more as the Universe collapses. Penrose believes that there may be a law of physics—which nobody has yet formulated—according to which past singularities are always simpler in structure than future singularities, and that this is why we perceive time as flowing from the Big Bang into the future of the expanding Universe. Another enigma concerns closed loops in time. If it were possible to traverse such a loop and come back in your own past, there would be immediate and obvious scope for contradictions. Strangling your own grandmother in her cradle poses problems not merely of ethics but of causality—where did the strangler come from, if granny never grew up? Astoundingly, there are some solutions to Einstein's equations of general relativity that appear to represent cosmological models

that permit loops in time. These possible universes (explored mathematically by Kurt Godel in the 1940s) differ from the structure of our own Universe, but they do not seem physically impossible.

One can take either of two attitudes to these closed loops. On one hand, the fact that Einstein's equations permit them may be telling us that general relativity is incomplete, and an extra law of nature is needed to rule out such absurdities. On the other hand, one could take the view that although there are clear paradoxes for conscious observers who travel around such loops in time, they entail no obvious absurdity in a universe where there is no memory. A Godel universe might, by definition, be devoid of intelligent life—a kind of antianthropic requirement—although, in any case, since it would take almost infinite energy and a time nearly as long as the age of that universe to travel around one of these loops, they could do no practical harm. Nevertheless, physicists react against any model that permits causality to be violated. Most physicists feel equally strongly about theories that envisage the arrow of time going into reverse—but, again, this might not entail any logical absurdity, if it happened at a time when all stars and matter had decayed, black holes had evaporated, and the Universe contained nothing except pure radiation, with no conscious observers around to notice what was going on.

There is plenty of scope, it seems, for speculation about the nature of time. But there is far less scope to speculate on the possibility of universes constructed out of a different number of dimensions from our own. The fact that our Universe is composed of (three plus one) dimensions is actually such a profound observation that it caused at least a few people to begin thinking along anthropic lines long before the idea of the anthropic cosmological principle itself was articulated. Although he did not express it quite in those terms, one

of the first people to follow through the implications of the three-dimensionality of space was William Paley, an eighteenth-century philosopher and churchman.

Paley is best remembered today for his forceful expression of the argument that living things are far too complicated to have arisen by chance, and that the existence of creatures as beautifully fitted for their way of life as ourselves (or a fly, or a primrose) reveals the presence of a designer, the hand of God at work. This argument is expressed dramatically in terms of a man who finds a watch lying upon the ground, and who knows nothing about watches but perceives from an inspection of the object that it has been designed and put together for a purpose. A "blind watchmaker" who sat before a heap of watch components and put them together at random could never, the argument runs, assemble a working watch.* The argument may be correct, but it is inappropriate, since evolution by natural selection does not proceed by sticking together all the components of a living creature at random, but by building step by step on previous successes. That debate is outside the scope of our present book, but Richard Dawkins has laid the myth to rest in his superb The Blind Watchmaker, which we recommend to anyone still seduced (or confused) by the "argument from design." What is relevant to our main theme is that Paley was also intrigued by the inverse square law of gravity, described by Newton in the 1680s.

Paley realised that the inverse square law, in which the force between two objects is proportional to 1 over the square of the distance between them, is unique

*Paley's arguments came mainly from biology, but as astronomers we are amused to note the rather limited relevance he assigned to our subject: "My opinion of astronomy has always been that it is not the best medium through which to prove the agency of an intelligent creator, but that this being proved, it shows beyond all other sciences the magnificence of his operations."

in giving rise to stable orbits. If the law of gravity had, for example, been an inverse cube, then planetary orbits would be unstable, and a planet that moved a little closer to the Sun would immediately begin to fall inwards permanently, while one that moved slightly outwards in its orbit would continue receding forevermore. Tiny changes, such as those caused by the impact of a meteorite, would be disastrous. In our Universe, if the Earth's orbit, say, shifts slightly inwards or outwards because it is hit by a piece of rock from space, the natural tendency is for the planet to return close to its old, regular path. Paley saw this "choice" of the inverse square law of gravity as another example of God's work in designing a Universe suitable for human life. He did not elaborate, however, on the fact that the inverse square law is a byproduct of the fact that the Universe has three spatial dimensions—although this had been noticed by Immanuel Kant earlier in the eighteenth century.

The importance of the dimensionality of space began to interest scientists in the twentieth century, following Einstein's work that gave space (or space time) a dynamic role in physics. This showed that the dimensionality of the law of gravity is always one less than the dimensionality of space—inverse square in a space of three dimensions, inverse cube if space has four dimensions, and so on. Planetary orbits are stable only in a space with three dimensions, because an inverse square law of gravity is the natural law only in a space with three dimensions. About the same time, researchers realised that the equations of electromagnetism, discovered by the Scotsman James Clerk Maxwell in the nineteenth century, have workable solutions only if they are applied in a spacetime that has (three plus one) dimensions.

The insight was developed by G. J. Whitrow in 1955. He argued that the reason we observe the Universe to

possess three dimensions is that observers can exist only in universes that have three dimensions of space (and one of time). Life can exist only in three-dimensional space; we live, and therefore it is no surprise that we find ourselves in three-dimensional space. But in the 1950s, the way people thought about the implications of this insight was beginning to be quite different from the way Kant or Paley thought about the significance of the inverse square law. Instead of regarding our Universe as unique, and seeing its suitability for life as the result of design, some researchers began to consider the possibility that there might be many universes, in an array (an "ensemble") that included all possibilities of, for example, dimensionality. All possible worlds exist, in this picture, but life exists only in the subset of worlds suitable for life, and there is no need to invoke a designer at all. Hardly surprisingly, some of the first discussions of this idea came from researchers in the Soviet Union, where the hand of God was not regarded as a reasonable explanation for the cosmic coincidences; Fred Hoyle also considered the possibility that the energy levels in carbon, for example, might follow different laws in different parts of the Universe, or the superuniverse, so that life like us exists only in lesser regions where the coincidences work out just right (but, unlike the Soviet researchers, Hoyle has developed his own ideas about the hand of a designer at work in setting those coincidences).

Some people now take this very seriously. John Wheeler, for example, imagines an ensemble of universes with different physical laws and different values of the fundamental constants, all "laid down" at the initial singularity, the moment of creation. A kind of evolution by natural selection operates on these universes. Most of them are "stillborn," in the sense that the prevailing laws do not permit anything interesting to happen in them. But maybe in some of them com-

plex structures can evolve. It would be a major achievement if someone could show that any universe in which interesting things can happen must end up looking something like our own Universe. At present, though, the best we can do is to show how even a small change in just one of the critical numbers can make a universe that would be almost unrecognisable and probably uninhabitable. There are hundreds of ways in which we could tinker with the laws of nature; but since we have been discussing gravity, let's consider, as our single case study, an adjustment in the strength of that force.

An Alternative Universe

There is nothing truly fundamental about the units by which we choose to measure things here on Earth— pounds or kilograms, yards or kilometres, all are equally arbitrary. So when the more philosophically inclined physicists speculate about the nature of the Universe, they sometimes like to use what they call "natural" units," which are defined in terms of truly fundamental constants of nature, such as the speed of light and the quantum-mechanical constant named after Max Planck (accepting, for the moment, that these constants really are constant). Using these fundamental constants, the strengths of the various interactions can be described in terms of pure numbers, the size of which indicates their importance on a chosen scale. Usually, physicists choose to define these numbers in terms of the mass and electrical charge on a proton, one of the biggest fundamental particles. On that scale, gravity is relatively insignificant, and its strength is described by a number known as the gravitational fine structure constant. This is about 10^°—the electric fine structure constant, in the same system of units, is 1/137, a little less than 10" 2 and about 10" 38 times stronger than gravity.

The strength of gravity is tiny compared with the strength of electrical forces, but as we have seen before, it is dominant in the Universe at large, because all the electrical forces cancel out, while all the gravitational forces, from every proton and every other subatomic particle, add together. What would happen if gravity were a bit more dominant?

Consider a universe in which the gravitational fine structure constant is 10~ 30 rather than 10" 40 , but everything else is as usual. Galaxies, stars, planets, mountains, and microorganisms can all still exist, but they will be very different from their counterparts in our Universe.

The mass of a star depends on the inverse (3/2) power of the gravitational constant (that is, 1 over the square root of the cube of the constant). Lifetimes of stars, on the other hand, depend on the inverse power (1 over) the constant. In our Universe, the Sun is a typical star. It has one solar mass of matter (of course) and a lifetime of about 10 10 years. So in our alternative universe stars will typically have masses of around 10" 15 solar masses (about 10 12 tonnes, roughly the mass of an asteroid in our Universe). The lifetime of a star depends on how long it takes a photon to diffuse (or "random walk") from the centre to the surface. This time depends on the square of the distance to be travelled in a straight line. So these ministars, with roughly the same density as stars in our Universe but sizes a hundred thousand (10 5 ) times smaller, have lifetimes shorter by a factor of 10 billion (10 10 )—about one of our years. From the point of view of the origin of life, involving heavy elements produced in the first generation of stars, the universe will be at an interesting stage when it is about one year old and its Hubble radius is about one light-year; the critical density to make this compact universe flat will be 10 10 times the density of our Universe—still considerably thinner than thin air, but

more substantial than the gas between the stars of the Milky Way.

In the alternative universe, there will be as many galaxies as we see in the observable Universe (about 10 10 , 10 billion), but each one will be smaller in diameter by a factor of 10 10 than our Milky Way, and denser by the same factor. Each galaxy will contain about a hundred thousand stars—but the stars will be very different from the ones we know.

With the mass of each star, and hence its energy supply, reduced to 10~ 15 that of the Sun, and a lifetime of one year instead of 10 10 years, each star will be about 10~ 5 times the brightness of the Sun. These fast-burning stars will be a little hotter than typical stars in our actual Universe, with central temperatures of about 50 million K, compared with the 15 million K at the heart of the Sun. The whole star is about two kilometres across, a tiny nuclear furnace burning hotly with a more blue light, richer in ultraviolet, than the Sun. The energy is available for life to evolve, if these stars have any planets the right distance away.

The right distance is 10~ 2,5 times the distance the Earth is from the Sun, because the brightness of the star is down by 10" 5 , and the intensity of light at any distance from the star depends on the square of the distance, making up the extra factor of 2 in the power of 10. A planet at this distance from its parent star, about 500,000 kilometres (less than twice the distance from the Earth to the Moon) would have a comfortably Earth-like mean temperature of close to 25 degrees C. Just as the parent star has about 10~ 5 times the mass of our Sun, so this "other earth" would have a mass of about 10" 5 times (one hundred-thousandth of) the mass of our planet. This planet—about the size of a small moon in our Universe—will orbit around its parent star about once every twenty days, by our calendars; this has the curious consequence that in terms of its own

"year," that planet "lives" in a universe just eighteen "years" old.*

But a variation on the anthropic arguments that we used in chapter 1 to explain the sizes of planets in our Solar System suggests that there would be a rather large number of "days" in each of those years. If the planet rotates nearly as fast as it can without breaking apart, the number of "days" in each of its relatively long "years" will be something like 2 million. Each "day" on our miniworld is just under one of our seconds long.

Science fiction writers, no doubt, could have a field day describing the culture of a civilisation that inhabited such a planet, and evolved on a timescale set by the length of the day, while the planet orbited the star on a timescale 2 million times longer. Unfortunately, their speculations might have to be described more as fantasy than as genuine scientific fiction, for it is hard to see how life and civilisation could evolve on such planets.

The first problem is that in our hypothetical, compact, speeded-up universe stars will be packed much more closely together, even in comparison with their own size, than stars in our Milky Way. In that universe, the separation between stars is only about 10" 2 times the distance from the Earth to the Sun, smack in the zone of comfortably warm planetary orbits. Planets in those orbits will be tugged free from their parent star by the gravity of passing stars, and only planets in very close, uncomfortably hot orbits will stay tied to their own star. For those who like such speculation, however, we can envisage habitable planets as tramps be-

*The change in the strength of gravity ought, at first sight, to tie the planet more tightly to its star, and make it orbit faster; but the smaller mass of the star more than compensates for this, leaving a factor of a hundred thousand (KT 5 ) the other way.

tween the stars, wandering here and there but always about the right distance from one star or another to allow the conditions required for complex chemistry to be maintained on their surfaces. What would those surfaces, and any life forms that inhabited them, be like?

The biggest possible mountains would be those just less than the mass required to melt the material at their bases by pressure. They would be just thirty centimetres high. But the mass of a living creature that could just survive falling over without breaking (roughly equivalent to our own mass in the context of the Earth's gravity) would be so small that each living organism could contain no more than about 10 20 atoms of heavy elements such as carbon, nitrogen, and oxygen. Our own bodies contain about 10 28 such atoms, ten thousand million times more; the mass of a prospective mountain climber in our alternative universe would be roughly a thousandth of a gram, not a hundred kilograms, and it is very hard to see how such a tiny organism could contain the complexity of chemical compounds that seem to be a prerequisite for intelligent life.*

This is especially unfortunate since if the microorganism could take an intelligent interest in its surroundings it would be able to do astronomy and cosmology in "real time," watching stars and the universe evolve, and living and working for an appreciable fraction of the life of the speeded-up universe.

The relationship between gravity and life revolves

*Very hard, but perhaps not quite impossible. A bacterium here on Earth, with a diameter of about a hundred millionths of a metre (one hundred microns), weighs in at roughly half a millionth of a gram. Our hypothetical inhabitants of the miniuniverse are at least a thousand times bigger than that. Science fiction writer Greg Bear explored the possibility of intelligence on the scale of cells like bacteria in his intriguing book Blood Music.

around two features of this peculiar force that holds together individual stars and entire galaxies. These features are quite crucial for cosmogonic processes. The first point is that gravity drives things farther from equilibrium , not towards equilibrium. When gravitating systems lose energy they get hotter; for example, an artificial satellite speeds up as it spirals downwards due to atmospheric drag. Another example is the Sun. If the heat it loses were not balanced by the release of energy by nuclear fusion in its interior, the Sun would contract and deflate—but it would thereby end up hotter inside than before. It needs more pressure inside it to balance the stronger tug of gravity when it is more compressed. This runs counter to the general rule of thermodynamics, that hot objects left to their own devices (like a glowing lump of hot steel) radiate heat and get cooler. From the initial Big Bang to our present Solar System, this antithermodynamic behaviour of gravity has been amplifying density contrast and creating temperature gradients—prerequisites for the emergence of any complexity in the Universe.

The second key feature of gravity, in our Universe, is its weakness. Our Universe is large and diffuse and evolves slowly because gravity is so weak. The extravagant scale of the Universe, billions of light-years, is necessary to provide enough time for the cooking of elements inside stars and for interesting complexity to evolve around even just one star in just one galaxy. There would be less time and less scope for such evolution in the small-scale speeded-up universe discussed above, where gravity is stronger than in ours. A force like gravity is essential if structures are to emerge from amorphous starting points; but, paradoxically, the weaker that force is, the greater and more complex are its consequences.

Although gravity does play this unique role, the exact values of the strengths of the other fundamental forces

seem to be just as important for life. The example we have elaborated in detail is typical of such exercises. If we modify the value of one of the fundamental constants, something invariably goes wrong, leading to a universe that is inhospitable to life as we know it. When we adjust a second constant in an attempt to fix the problem(s), the result, generally, is to create three new problems for every one that we "solve." The conditions in our Universe really do seem to be uniquely suitable for life forms like ourselves, and perhaps even for any form of organic complexity. But the question remains— is the Universe tailor-made for man? Or is it, to extend that analogy, more a case that there is a whole variety of universes to "choose" from, and that by our existence we have selected, off the peg as it were, the one that happens to fit? If so, what are the other universes, and where are they hiding?

CHAPTER ELEVEN

Or Off the Peg?

FINE-TUNING ARGUMENTS have a long history. Lawrence Henderson, a Harvard professor, wrote an important book in this vein, titled The Fitness of the Environment, early in the twentieth century. He noted such things as the anomalous property of water, that it does not reach maximum density at the freezing point, which has played a key role in the evolution of life on Earth; he also pointed out special features of other molecules, including carbon dioxide. He argued that "we are obliged to regard this collection of properties as in some intelligible sense a preparation for the process of planetary evolution. Therefore the properties of the elements must for the present be regarded as possessing a teleological character." His contemporary Homer Smith, on the other hand, was unimpressed, saying that "the fitness of the living organism to its environment or vice versa is as the fit between a die and its mould, between the whirlpool and the riverbed."

What, though, can we make of the coincidences in the physical constants involved in nucleosynthesis? They cannot be dismissed as readily as other arguments. A complicated biological organism must indeed evolve in tune with its environment; but the basic physical laws are "given," and nothing can react back to modify

them. It does seem worthy of note that these laws permit something interesting to have happened in the Universe, where there could so easily have been a "stillborn" universe in which no complexity could evolve. The Canadian philosopher John Leslie has offered a neat analogy. Suppose you are facing execution by a fifty-man firing squad. The bullets are fired, and you find that all have missed their target. Had they not done so, you would not survive to ponder the matter. But, realising you are alive, you would legitimately be perplexed and wonder why. In its mildest form, an-thropic reasoning is simply a proper allowance for observational selection. Given the brute fact that we are a carbon-based form of life slowly evolved around a G-type star, there are some features of the Universe, some constraints on physical constants, which can be inferred quite straightforwardly.

Can we, though, go beyond a subjective expression of surprise that delicate balance seems to prevail? Some distinguished scientists have given their reactions in print, in popular articles if not in technical papers. Freeman Dyson says that in some sense "the universe knew we were coming." And, to quote from Sir Fred Hoyle's Galaxies, Nuclei, and Quasars, "the laws of physics have been deliberately designed with regard to the consequences they produce inside stars. We exist only in portions of the universe where the energy levels in carbon and oxygen nuclei happen to be correctly placed."

In response to Hoyle's comment, we might ask if the microphysical constants could be different in different parts of the Universe, or at different times. Paul Dirac suggested, half a century ago, that the gravitational constant might change as the Universe aged; this is now ruled out by observations, and there is no evidence that any other microphysical constants have varied— strict constraints are imposed by observations of the spectra of distant objects, from radioactive decay over

the geological past, and other studies; moreover, there are conceptual problems in defining what we would mean by, say, a variation in Planck's constant. The atoms and nuclei whose properties are studied in the laboratory certainly seem to behave identically to those in the most remote quasar or in the first few minutes of the Big Bang. Were this not so—were there no firm link between the cosmos at large and local physics—scientific cosmology could have made little progress (and this book would never have been written).

But even if the constants are fixed throughout the Universe we can observe, could there in some sense be other universes where they are different? This idea was outlined by a biologist, C. F. A. Pantin. He said "the properties of the material universe are uniquely suitable for the evolution of living creatures. If we could know that our universe was only one of an indefinite number with varying properties we could perhaps invoke a solution analogous to the principle of natural selection, that only in certain universes, which happen to include ours, are the conditions suitable for the existence of life, and unless that condition is fulfilled there will be no observers to note the fact."

If someone walks into a clothing shop and buys a suit that is a perfect fit to his or her body, there are two possibilities. Either the tailors who work in that shop have carefully measured that person's body and made a suit to fit it—bespoke tailoring—or the shop has such a large range of clothing available, in all shapes and sizes, that the person in question has been fitted out from stock, off the peg. The idea that the Universe is in some way constructed for our benefit, or at least designed as a fit home for intelligence, corresponds to the first possibility. In many ways, the second alternative is more attractive; but it requires the existence of a vast array of alternative universes from which we have "chosen" by the fact of our existence. In this picture,

there are myriads of other worlds in which the laws of physics and the constants of nature do differ, a little or a lot, from those we know. In most of the universes, life—certainly intelligent life—does not exist. Any universe in which our kind of intelligent life can arise must look rather like our Universe, since without the familiar coincidences and constants that life would not be there. We believe our Universe to be special because we inhabit it. But that does not mean that it is special in any deeper sense of the word.

A useful analogy is with a lottery. Suppose a million lottery tickets are sold, and then one number out of that million is selected. The holder of that number wins the prize, so that number seems special. But in a deeper sense it is no more special than any of the other numbers in the lottery. By the nature of the lottery, somebody must win, and each of the numbers has an equal chance of winning. It is only after the event that one number gains a special status. The holder may feel lucky as a result; but somebody had to get lucky!

Maybe the world is like that. There may be a multitude of universes that all start out sterile. Intelligence appears in some (or perhaps only one) of those universes as a result of the accumulation of random coincidences ("luck"). But there is no meaning to the coincidences, and that universe stands out from the rest as special only with hindsight, once intelligence has appeared to wonder over its own origins.

The Quantum Realities

The key to this off-the-peg approach to the anthropic coincidences is that there must indeed be a variety of universes, an ensemble, to choose from. Science fiction

readers will already be familiar with the concept*; it happens that it also has a perfectly respectable scientific basis, in quantum theory.

Quantum physics is all about probabilities. In an honest lottery, every ticket has an equal chance of winning; but the quantum world is not like that. The position an electron will be found in when we make a measurement on an atom, for example, depends on quantum probability. There is a very high probability that the electron will be found in one of the energy states corresponding to the "steps" on that particular atom's energy ladder, and a very small probability that it will be found somewhere else entirely, not even connected to the atom where we expect to find it. Rather as if the lottery is rigged, so that any number ending in 9 has a good chance of winning, while any other number has only a low probability of coming out of the hat. When we make a measurement, we may say that we have located the position of the electron, or at least observed which energy step it is sitting on. But that does not mean that another measurement will give us the same result. In the quantum world, as soon as we stop looking at an electron, it dissolves into a mist of probabilities, called a superposition of quantum states. It is like the lottery before the winning number is drawn. Making the same observation again will give a new answer, as if we put the winning number back into the hat and made a second draw—probably getting another number ending in 9, perhaps a different number altogether, possibly (but not certainly) the same number as before. The act of measurement forces the electron to choose among the possible states and to take a solid identity.

*01af Stapledon's classic Star Maker, for example, although first published back in 1937, contains some surprisingly modern-sounding descriptions of universes with different physical laws, and even different numbers of dimensions of both space and time.

But each identification is the result of a separate, independent coalescence, what is known as a "collapse of the wave function."

This is the basis of the standard interpretation of quantum mechanics, called the Copenhagen interpretation. It works very well if you want to apply the quantum rules to designing a laser, say, or to calculating how atoms join together to make molecules. But it falls down completely if you try to imagine a way to describe the entire Universe in accordance with quantum theory. Any quantum system, according to the Copenhagen interpretation, exists as a superposition of states, a nebulous array of probabilities, unless and until it is observed from outside. But what is there "outside" to observe the Universe and to collapse its wave function? This puzzle has led some cosmologists to embrace a different interpretation of quantum mechanics, vvhich is called the Many Worlds interpretation.

Many Worlds quantum mechanics has long been regarded with suspicion by physicists. Where Copenhagen theory says th at none of the quantum options has ^ V any re ality unless it is o bserved, Many Worlds theory says that all quantum possibilities really do exist, each /{ in its own space and time, and that each measurement we make simply identifies which branch of the multi-universe we are in. There is a separate universe, in this picture, for each possible energy level that our electron might inhabit. When we measure the atom, we find the electron in the one energy state that corresponds to the universe we live in; but our doppelgangers in the universe next door may simultaneously be making the same measurement and getting a different answer.*

*"Next door" and "simultaneously" are rather difficult concepts in this connection; it's better to think of all the different worlds being at right angles to each other, perpendicular rather than parallel. This is discussed by John Gribbin in In Search ofSchrodinger's Cat.

Both the Copenhagen and the Many Worlds interpretations give exactly the same ''answers" when applied to practical problems like the design of a laser. It is simply a matter of philosophical preference which one you choose to work with, at that level. Most physicists and engineers don't care about the philosophy and use a set of rules derived from the Copenhagen interpretation, which came along first (as John Polkinghorne puts it, "the average quantum mechanic is no more philosophical than the average motor mechanic"). But there is one problem that the Many Worlds theory can deal with but the Copenhagen interpretation cannot. This happens to be important enough to give the Many Worlds theory the edge—it is the problem of providing a quantum mechanical description of the entire Universe.

In the Copenhagen theory, the Universe cannot exist, in the everyday sense of the word, unless something outside the Universe measures it and collapses the wave function. But in the Many Worlds theory all possible universes exist. The differences between some universes are trivial—here an electron is on the first step of an energy ladder, there it is on the second step. The differences between others are more extreme—a change by a factor of 10 10 , perhaps, in the strength of the gravitational force. But each universe is "real" in the everyday sense of the word, and it is no longer a surprise to find that the Universe we inhabit fits us exactly, since we have simply selected it, off the peg, by our existence. Among the many worlds, we wear the one that fits.

But this is not the only way in which to imagine an ensemble of universes from which, by our existence, we "choose" to live in one suitable for life. If the Universe is infinite, then anything that can possibly happen may happen, or may have happened, or may be happening somewhere in that infinity. There could be a world, somewhere in the infinite Universe, where you wrote

this book, and we are the readers; a world where Virginia is still a colony of England (and one where England is a colony of Virginia); and so on. These worlds will be separated from us, in an infinite Universe, only by distance, not by any mind-numbing extra dimensions of space and time at right angles to each other. Alas, though, we cannot view them, because their light has not yet had time to reach us.

It seems, at first sight, even more like an outrageous offshoot of science fiction than the Many Worlds theory. But it is now being taken very seriously, as an offshoot not of SF but of the inflationary theory of cosmology.

Inflation in a Nutshell

"Inflation" is a generic name given to a set of theories that attempts to explain how the entire presently observable Universe could have homogenised itself while expanding from an initial superdense state at a time corresponding to 10" 43 seconds (the Planck time) after the moment of creation to the state, less than a millisecond later, when its density was roughly that of an atomic nucleus. From then on, everything can be described in terms of the laws of physics we know from our experiments and observations on Earth, within the framework of the standard Big Bang theory. But what happened before that time?

Inflation describes these events in terms of the way the original symmetries between the forces of nature broke when the Universe was young. Using the best theories we have, the grand unified theories, it is possible to calculate the energy at which this should have happened, and this corresponds to a time when the age of the Universe was about 10" 35 seconds. The simplest

thing that could have happened at that time was that the forces would quietly have gone their separate ways. But that would not have provided the expansionary boost needed to smooth out any wrinkles in the fabric of spacetime and make the Universe flat. This is where inflation comes in.

The way the Universe was cooling and changing from one state to another at that time can be compared with the way water vapour cools and condenses into liquid water. This is a change of state, from vapour to liquid, in which energy is released. The comparable change in the early Universe when symmetry was broken is known as a phase transition, and also releases energy. In 1980, Alan Guth, of MIT, suggested an extension of this analogy. Sometimes, water vapour can be cooled below 100 degrees C, the temperature at which it "ought" to become liquid. The energy that ought to be released at the transition is locked up, and the supercooled system becomes increasingly unstable as it cools further, until suddenly all the vapour changes into liquid and a burst of heat is given off. Guth suggested that the equivalent process could happen in the early Universe, with the phase transition being delayed while the whole Universe supercooled, and then flashing over into the broken-symmetry state, releasing all the locked-up energy of the phase transition in the process.

During the supercooling phase, the calculations show, the Universe will be driven into a wild, exponential expansion—inflation. It will double in size roughly once every 10~ 34 seconds, which doesn't sound too impressive until you realise that this means a hundred doublings in the space of 10~ 32 seconds, enough to expand a tennis ball to the size of the entire observable Universe. At the end of inflation, the Universe was extremely flat, with all the wrinkles ironed out; and the burst of energy from the delayed phase transition then heated it up, establishing the familiar conditions of the

standard Big Bang as the runaway expansion at last began to slow.

There are hopes—not yet fully realised—that these theories can account for the fluctuations that eventually gave rise to galaxies. The amplitude of these fluctuations (the size of the ripples perturbing the flat background Universe) is an important cosmic number that still lacks a proper explanation.

Since Guth proposed the idea of cosmic inflation, it has gone through many changes and now exists in many different forms proposed by different theorists. The basic principle is very attractive and solves many puzzles of how the Universe got to be in a Big Bang state. But there is still the puzzle of how the seed that inflated to become our Universe—a hot, dense concentration of spacetime smaller than a proton but containing all the mass-energy that became the observed Universe—came into existence at 10" 43 seconds. One possibility is that the basic state of the universe at large, the meta-universe, is one of chaos, with some regions expanding, some contracting, some hot, and some cold. In some parts of this infinite meta-universe conditions just happen to be right for inflation to begin, and so it does. A universe is born. But the most dramatic possibility is that the infinite meta-universe is, was, and always will be in a state of inflationary expansion itself, with a temperature of about 10 31 K and a density of about 10 93 grams per cubic centimetre. In its 1980s incarnation, the idea stems from the work of Richard Gott at Princeton University and Andrei Linde at the P. N. Lebedev Institute in Moscow—although it has curious echoes of the steady state theory, developed by Fred Hoyle and Jayant Narlikar, which fell from favour in the 1960s. And it all derives from one of the earliest solutions discovered, in 1917, for Einstein's equations of general relativity.

Bubbles on the River of Time

''Many and strange are the universes that drift like bubbles in the foam upon the River of Time." When Arthur C. Clarke wrote those words, almost forty years ago, as the opening to a science fiction story called "The Wall of Darkness," he can have had no idea that in the late 1980s they would stand as an accurate description of modern cosmological thought. Theorists are now being led to consider the possibility that our Universe is indeed just one bubble among many in some greater meta-universe.

One way of imagining the seed of a universe like our own coming into existence as a tiny concentration of mass-energy with an "age" of 10" 43 seconds is simply as a quantum fluctuation of the vacuum, one of those things permitted by uncertainty, like the appearance of a virtual pair of particles out of nothing at all. The idea surfaced in the early 1970s, but in its original form the trick didn't work. Such an enormously massive fluctuation could indeed occur in principle, but it would occupy a tiny volume of space, smaller than a proton, and it would, by definition, have an enormous gravitational field. The result would be an extremely rapid collapse, snuffing the embryonic universe out of existence as quickly as any pair of virtual particles produced in the vacuum today. Inflation, however, provides a way out of this dilemma. In the tiny fraction of a second it exists, inflation can set to work and blow the seed up into a full-size universe. "Full size," in this case, means something about as big as a basketball, containing as much mass-energy as the entire visible Universe today, and experiencing a big bang. All the rest follows naturally from the known laws of physics.

But why stop at one universe? If bubbles of mass-energy can appear out of nothing at all, and explode

exponentially into life as fully fledged universes, shouldn't the same processes—vacuum fluctuations—be going on in the space between the stars today? And, if so, might it not be rather uncomfortable for us if a new universe popped into existence nearby? Yes, and no. Other universes might indeed be being born all the time in the vacuum of space; but if they were, there is no way that we would know about it.

The possibilities have been investigated by several researchers, among them Edward Fahri and Alan Guth. They envisage artificially creating other universes; their scientific paper on the topic is entitled "An Obstacle to Creating a Universe in the Laboratory." You don't need a lot of mass to start with; the quantum effects will provide the mass-energy of a universe for you, once the process is kicked into starting. But you do need to create conditions of very high density and a temperature equivalent to about 10 24 K, in order for inflationary processes to do their thing. We already have the energy available to do half the trick, in the form of hydrogen bombs; the other half of the trick is to confine that energy within a very small volume (the size of an atom) which is why nobody is manufacturing universes in their basement just yet. If you could confine the energy, though, what you ought to get would be, according to the equations of general relativity, a black hole. Interesting in its own right, but not a new universe. Or is it?

Guth and others have shown that what happens inside the confined region depends on exactly how the pressure is applied. In many cases, the compressed region does turn out to be "only" a black hole. But there are solutions to the equations that, given the right initial conditions, do allow for the prospect of inflation. The confined region does not, however, expand back out into the Universe at large. Instead, it expands in a direction at right angles to our familiar

dimensions of space and time, off into a universe of its own. Exactly the same thing will happen to any inflationary seeds created by quantum fluctuations in the vacuum of our Universe.

An analogy that is often made is to describe our expanding Universe as the skin of a balloon that is steadily increasing in size (cosmologists used to say "an inflating balloon/' but we are now talking about the present-day expansion of the Universe, not its inflationary past, so we have to be careful with our choice of words). The two-dimensional skin of the balloon represents all of our familiar dimensions. As the balloon expands, the Universe gets bigger. Any "new" universes created within our Universe, either naturally or by someone with an H-bomb in his basement, are like little bubbles in the skin of the balloon. They pinch off from our spacetime (the skin of the balloon) and expand outwards in their own right, in their own space and their own time.

From our perspective, nothing seems to have happened. Perhaps a black hole has appeared, perhaps not. From the perspective of any observer able to withstand the extreme conditions inside the superdense region, however, things would be very different. The region would inflate exponentially, then go over into a big bang and expand more sedately. Stars, galaxies, and intelligent creatures could evolve, study their surroundings, and begin to wonder about the possibility of creating new universes in the basement (provided the laws of physics in the new universe permitted intelligence to evolve; this is by no means certain, since each universe may have its own set of laws and constants). Quantum cosmology allows the possibility of creating not just one universe but an infinite number of universes out of nothing at all. The universes may be interconnected, in some complex way, as new universes are born within, but then pinch off from, the vacuum of old universes,

producing a complex, multidimensional foam. Our Universe may simply be a region of spacetime that has pinched off from another bubble. But the bubbles can never communicate with one another and might have very different properties from one another. The end of inflation is linked with the breaking of symmetry between the four forces of nature; there is nothing to say, however, that the symmetry will break in the same way in every bubble. In some bubbles, the forces will have different strengths from those in our Universe; indeed, there may be three, or five, fundamental forces, or some other number, instead of the four we know.

We are back in the worlds of Arthur Clarke. If an infinite variety of universes exists, then all things are possible. There must be infinitely many universes where gravity is too weak for life to emerge, infinitely many where gravity is too strong, and infinitely many more where something else goes wrong. But it is no puzzle that we exist, since there must also be an infinite number of bubbles in which conditions closely resemble those we see in the Universe around us, and the world is, like baby bear's porridge, "just right."

Cosmic Dragons

The anthropic principle cannot claim to be a scientific explanation in the proper sense. At best it can offer a stop-gap satisfaction of our curiosity regarding phenomena for which we cannot yet obtain a genuine physical explanation. The most powerful of these insights may be the hint that our Universe is not unique, but just one among an ensemble of universes, whatever form that ensemble might take. Andrei Linde envisages an infinite universe divided into domains in each of which the physics would be different. Most of this meta-universe would be a lifeless desert; complex evolution

would occur only in "oases" where the constants—the numbers of dimensions and so on—had propitious values. Our oasis must then be at least 10 billion light-years across, because the laws of physics seem to be the same everywhere we look. But the desert regions beyond may in principle be observable in the remote future when, perhaps a thousand billion years or more from now, light from the edges of our domain has had time to reach us. This is too remote an eventuality to provide a practical way of testing the possibility—but the conceptual status of the idea is no different from the conjectures of early cosmographers about continents beyond the horizons of the then-known world. We prefer to attempt at least to sketch the outlines of the continents that might lie beyond our present horizons, rather than simply fill in the edges of the cosmological map with the legend "here be dragons." Anthropic reasoning suggests that those other worlds do indeed exist, even if we can never have direct knowledge of them.

"The most incomprehensible thing about the Universe is that it is comprehensible" is one of the best known of Einstein's sayings—it has become a cliche. He meant by it that the basic physical laws, which our brains are attuned to understand, have such broad scope that they offer a framework for interpreting not just the everyday world but even the behaviour of the remote cosmos. The physicist Eugene Wigner described this as the unreasonable effectiveness of mathematics in the physical sciences. Cosmologists start by using the physics that is validated locally, and apply this, with simplifying assumptions, to probe the workings of the Universe at large. These simple rules seem to work as a description of the Universe. There seems no reason why the Universe should be so structured that this approach permits any real progress—why the physics we study in the laboratory on Earth applies also in quasars bil-

lions of light-years away and in the early stages of the Big Bang. Unless, perhaps, there is some link between the simplicity of the Universe and its suitability as a home for intelligent life. To say that we would not be here if things were otherwise, however, need not quench our curiosity and surprise at finding that the world is as it is.

The Philosophy of Cosmology

What, then, is the physical status of anthropic reasoning— anthropic cosmology—today? Some people adopt the dismissive attitude that anthropic reasoning cannot offer scientific explanations in the proper sense. At best, they say, it can give a stop-gap satisfaction to our curiosity regarding phenomena for which we have as yet no genuine physical explanation. The world would indeed be very different if the relative strength of the nuclear and electromagnetic interactions were somewhat altered, but one still hopes for a unified physical theory that predicts the actual constants or relates them to one another. A little over a century ago, theorists might have imagined varying the electrical and magnetic forces and the speed of light—before the work of James Clerk Maxwell showed how these were interconnected. By extension, a more comprehensive theory may eventually relate all the fundamental forces. Most theorists indeed hope that the constants of nature will not forever have to be treated as numbers derived from experiments, but will be related by a unified theory. They will then be mathematically calculable, just as a circle's circumference can be calculated (rather more easily!) from its diameter.

A hostile view of the anthropic principle comes from Heinz Pagels's book, Perfect Symmetry: written in 1985:

Physicists and cosmologists who appeal to anthropic reasoning seem to me to be gratuitously abandoning the successful program of conventional physical science of understanding the quantitative properties of our universe on the basis of universal physical laws. Perhaps their exasperation and frustration ... has gotten the better of them. . . . The influence of the anthropic principle on the development of contemporary cos-mological models has been sterile. It has explained nothing, and it has even had a negative influence, as evidenced by the fact that the values of certain constants, such as the ratio of photons to nuclear particles, for which anthropic reasoning was once invoked as an explanation can now be explained by new physical laws I would opt for rejecting the anthropic principle as needless clutter in the conceptual repertoire of science.

We think this is going too far in disparaging anthropic reasoning. After all, in its weak form, little more is involved than the routine attitude of an experimenter who takes account of the limitations of laboratory techniques and equipment.

The case for the defence, however, is presented with baroque elaboration by John Barrow and Frank Tipler in their massive book The Anthropic Cosmological Principle. Without going all the way with them, we agree that the principle does deserve serious attention. Its eventual status will depend on what the laws of nature are really like. If some final unified theory yields unique numbers for all the constants, then it may be inconceivable to envisage a different kind of universe. But if the basic laws turn out to involve some random or statistical element, then the idea of an ensemble of universes, outlined in this chapter, could be put on a serious footing. It could then indeed be natural selection, not mere accident, that our Universe (that is, the part of spacetime we can observe) has the particular values of the physical constants that we measure.

The "weak" anthropic principle—the realisation that the existence of observers such as ourselves imposes some selection effects on what we see around us—is almost banal. Any more pretentious role for anthropic reasoning is controversial, and depends on the true nature of the laws of physics. To quote Steven Weinberg, from a 1984 BBC broadcast, "I certainly wouldn't give up attempts to make the anthropic principle unnecessary by finding a theoretical basis for the values of all the constants. It's worth trying, and we have to assume that we shall succeed, otherwise we surely shall fail."

So perhaps it is best, if they are to retain their scientific motivation, that theoretical physicists should not take the strong anthropic principle, the idea that the Universe is tailor-made for man, too seriously. If there is a unique "theory of everything," then there is certainly a sense in which the laws of physics could not have been otherwise. We would then have to accept it as genuinely coincidental, or even providential, that the constants determined by high-energy physics happen to lie in the narrowly restricted range that allows complexity and consciousness to evolve in the low-energy world we inhabit. The intricacy implicit in these unique laws may astonish us, but our reaction would be no less subjective than a mathematician's surprise at the rich intellectual structures that can stem from simple axioms. Everything would be a consequence of unique laws. But that need not spell the end of scientific investigation of our surroundings.

The End of Physics?

Physicists sometimes talk of a theory of everything, or TOE, as providing "the end of physics," in the form of a single package of equations to describe the Universe and all it contains. But that would not really be the end

of science, or even put all physicists out of work overnight. No set of equations explains why there is a universe. To quote Stephen Hawking, "What is it that breathes fire into the equations? Why does the Universe go to all the bother of existing?" In any case, most of the challenging questions we ask about the natural world, on the astronomical as well as the terrestrial scale (other than those involving initial conditions or "origins") involve "old-fashioned" atomic and nuclear physics. The subnuclear world and the uncertainties of high-energy physics are generally irrelevant to larger-scale phenomena, just as the atomic structure of liquids provides no practical clues to the still unexplained complexities of turbulent flows in the air and the oceans. We can in principle write down the equations (essentially the ones derived by Erwin Schrodinger in the 1920s) governing the system—but we cannot solve the equations even for a typical single molecule, let alone for any larger system. Nor, even if we could solve the equations, would we have enough accurate information about the starting conditions (the position and velocity of every molecule) to permit accurate predictions. Even if we are "reductionists," believing that all phenomena can be reduced to physical fundamentals, this does not permit us to be "constructionists," in the sense of deriving a full understanding of complex systems from their atomic constituents. Sciences will always be in a hierarchy, where each level of structure entails new irreducible concepts.

Suppose you were unfamiliar with the game of chess. Just by watching a game being played, you could infer what the rules were. The physicist, likewise, finds patterns in the natural world, and learns what dynamics and transformations govern its basic elements. But in chess, learning how the pieces move is just a trivial preliminary to the absorbing progression from novice to grand master. The whole point and interest of the

game lies in exploring the complexity implicit in a few deceptively simple rules. Uncovering a TOE would do no more (and probably much less) than the equivalent of putting us in the status of a novice chess player who has just opened the book of rules—and knowing all the rules of chess does not permit even an expert to predict the outcome of a match between two grand masters, let alone every move played.

Biologists aim to delineate 3.5 billion years of the evolutionary history of life on Earth—to learn how, in Darwin's words, "whilst this planet has gone cycling on according to the fixed law of gravity, from so simple a beginning endless forms most beautiful and most wonderful have been, and are being evolved." Astrophysicists and cosmologists aim to set the Earth, and the entire Solar System, in a broader context of cosmic evolution. Progress is meagre, but what is astonishing is that there is any progress at all. We have 10 billion years, or more, of the past history of the universe to ponder, and far more than 10 billion years of Universal future to contemplate. The task of the evolutionary biologist pales by comparison—yet nobody would seriously claim that we understand every facet of life on Earth. Breaking the problem down into smaller pieces, we can contemplate the possibility of finding a "solution" to the puzzle of galaxy formation, for example, but we have no real idea where to start in finding out how life began. We cannot say whether life is common or rare in the Universe, nor whether it is confined to a single planet. There is ample work yet for physicists, and others!

Astronomers would confidently argue that planetary systems, potential abodes for life, are widespread in our Galaxy (and presumably in every other). But, even given the right environment, what are the odds on life getting established and evolving to an "interesting" stage? We claim no expertise on this subject, but have

the impression that there is no consensus even among the experts. The chance could be high [and a "SETI" search is certainly, given the colossal pay-off if it succeeds, a worthwhile gamble]; but it could on the other hand, be so low that life is very rare. If the odds were, say, 1 in 10 20 or less, there might be no other life within the experts. The chance could be high [and a "SETI"

Brandon Carter has advanced an anthropic argument suggesting that life may indeed be rare. He notes an interesting biological coincidence, that the Sun is half way through its life, and it has taken that long for intelligence to evolve on Earth. Stellar lifetimes are a straightforward consequence of the physical laws and constants; biological evolution, on the other hand, is an immensely complex multistage process. There seems no conceivable reason why these times should be closely comparable. Carter conjectures, therefore, that the typical time taken for biological evolution is much greater than stellar ages. Evolution, he claims, must have been especially rapid on Earth.* On this hypothesis, one can make the definite prediction that intelligent life would be rare in the Universe. In typical cases, even if biological evolution got started, it would not get very far before the host star died.

Even if life is rare, however, and even if we live in just one bubble universe among many in the foam on the river of time, there may be infinite time ahead of us, and infinite space in which to develop. In this cosmic perspective, we may still be near the "simple beginning" of the evolutionary process—certainly not its culmination. So living things may eventually become an important part of the Universe, modifying the astro-

*There are other possible reasons why our planet should be unusual in this regard. To mention just one, the Earth has an unusually large Moon, making the Earth-Moon system almost a double planet. Could the unusually large tides raised on Earth as a result have influenced the evolution of life?

nomical environment in the way that mankind is already modifying the terrestrial environment— although we might hope that our descendants would take a little more care with their modifications to stars and galaxies than we have taken with our modifications to the Earth. It is sometimes argued that if life is a rare accident, it would be an irrelevant fluke in a mindless and hostile cosmos. But we take the opposite viewpoint: if there is no life elsewhere, the Earth acquires universal significance as the spark with the unique potential to spread life and consciousness through the cosmos. We should regard present life on Earth as the beginning of a process with billions of years, and perhaps a literally infinite timespan, still to run—the greening of our Galaxy and beyond by forms of life and intelligence (not necessarily all organic) seeded from Earth. To snuff out our biosphere would quench the evolutionary process when it had barely begun to realise its limitless potential. We are back in the realms of Olaf Stapledon—an entirely appropriate note on which to end our discussion of the interrelationship between humankind and the cosmos, since there is one key ingredient in science, which is highlighted most effectively by the development of anthropic cosmology—a sense of wonder.