Insight Editions · No. 229
Until the End of Time
Mind, Matter, and Our Search for Meaning in an Evolving Universe
Brian Greene
Brian Greene follows the universe from the big bang to its far future, tracing how entropy and evolution give rise to life, mind, religion, art and science, and what meaning can be in a transient cosmos.
A study guide to Until the End of Time by Brian Greene. Shared free for personal reading; not for sale and not endorsed by the authors or publishers. Please read and buy the original.
The Book on One Page
Brian Greene walks the whole timeline of the universe, from the big bang to the far future, and places every human enterprise on it: life, mind, language, religion, art and science. Two forces run through it all, entropy and evolution, and one conclusion closes it: in a lawful cosmos that will dissolve all structure, including thought itself, meaning is not found but made, by minds that know they are temporary.
Order and life arise because of the second law, not despite it; everything, life and mind included, is transient; value and meaning are human creations.
13.8 billion years back to the big bang and 10102+ years forward; physics, biology, neuroscience, anthropology, religion, art, philosophy.
1029 ways to split 100 coins evenly; ~1020 ATP made per second in you; last thought by ~1050 years; Boltzmann brains at 101068.
Readers who want the physics of time and entropy joined to the big human questions; anyone building a worldview without supernatural props.
The argument in eight moves
- Entropy is counting. Entropy measures how many microscopic arrangements look the same macroscopically. The second law says entropy almost always rises because high-entropy arrangements vastly outnumber low-entropy ones. The laws of physics do not prefer a direction of time, so the arrow of time comes from a special, low-entropy big bang.
- Order is built by the entropic two-step. A system can lower its own entropy by exporting more to its surroundings, usually as waste heat. Steam engines do it, gravitating gas clouds do it spontaneously (losing heat makes their cores hotter), stars run on it via fusion, and life is one more dissipative structure that helps the universe release the entropy locked in matter.
- Life is "physics orchestrated." All life shares one genetic code and one energy system (redox cascades, proton batteries, ATP), which shows a common ancestor about four billion years ago. Before biology there was molecular Darwinism, and thermodynamics (dissipative adaptation) may make lifelike chemistry expected wherever energy flows.
- Mind is matter, and freedom is range. Greene bets on physicalism: consciousness will be explained by ordinary particles and laws, perhaps along Graziano's attention-schema lines. Free will in the traditional sense is an illusion, but human freedom is real as an enormous repertoire of responses, unlike a rock's.
- Language and story are flight simulators. Language made imagination shareable; stories rehearse trouble and let us inhabit other minds; myths push story to the edges of death and creation, shaped by minimally counterintuitive ideas and hyperactive agency detection.
- Religion and art answer finitude. Their evolutionary origins are contested (by-product, social glue, reputation guard, death-anxiety buffer; cheesecake, courtship, bonding, cognitive training). Their common thread is awareness of death and the longing for permanence. Literal religious claims that contradict physics are false; religion as value and heritage stands; art offers a distinct truth that makes us "feel a thought."
- Everything ends. The sun dies, galaxies drift beyond the horizon, star formation stops by 1014 years, galaxies evaporate, protons may decay by 1038, and if dark energy stays constant, thought itself becomes impossible by about 1050 years because space stops accepting waste heat. Black holes evaporate by 10102; quantum oddities (Higgs decay, Boltzmann brains) haunt eternity.
- Meaning is made inward. Immortality would drain choices of weight; humanity's extinction would drain projects of point; distance in time changes neither. With no grand design, the noble direction is inward: to create purpose, beauty and connection during our brief moment.
Chapter map
| Ch | Title | One-line payload |
|---|---|---|
| 1 | The Lure of Eternity | Knowledge of death drives culture; science now shows even life itself is transient |
| 2 | The Language of Time | Entropy counts look-alikes; the arrow of time needs a low-entropy start; the entropic two-step |
| 3 | Origins and Entropy | Inflation starts the bang; gravity and fusion build stars by obeying the second law |
| 4 | Information and Vitality | One code, one power plant: life is physics orchestrated, born of molecular Darwinism |
| 5 | Particles and Consciousness | The hard problem, three theories, no quantum magic; freedom as behavioural range |
| 6 | Language and Story | Language is open-ended; story is a flight simulator; myth runs on death and agency |
| 7 | Brains and Belief | Four evolutionary theories of religion; belief as confidence versus belief in value |
| 8 | Instinct and Creativity | Art as display, cheesecake, glue or training ground; artistic truth; symbolic immortality |
| 9 | Duration and Impermanence | The Empire State Building of time; stars, galaxies and matter end; the last thought |
| 10 | The Twilight of Time | Black holes evaporate; the Higgs may decay; Boltzmann brains and how to escape them |
| 11 | The Nobility of Being | Values are human; descendants anchor meaning; the noble direction is inward |
The Lure of Eternity
Greene's opening move is to treat knowledge of our own death as a force that shapes culture, and then to set it against what physics now says about the future: nothing in the universe is permanent, probably including life and mind themselves. The book will walk the cosmic timeline from the big bang to something close to the end. Its two lead characters are entropy and evolution, and it ends with a claim about meaning: in a universe that will forget everything, the only meaning available is the meaning we make.
The theorem that stands forever
The preface opens with a remark that changed how Greene saw his own vocation. As a college sophomore he mentioned to an older friend, the man who had taught him mathematics for years, that he was writing a psychology paper on human motivation. The friend replied: "I do mathematics because once you prove a theorem, it stands. Forever." Greene had loved mathematics as a game of abstract precision. Now he saw its romance: creativity constrained by logic and axioms yields truths that cannot be overturned. Every right triangle drawn before Pythagoras, and every one drawn until the end of time, satisfies his theorem. Change the axioms (draw on a curved surface like a basketball) and the result changes, but fix the assumptions and the result is ready to be chiseled in stone. Greene realised that what drew him to physics was that it hovered above the impermanence of the everyday.
The psychology paper itself kept dissolving into vagueness; dress plausible ideas in the right jargon and, it seemed, you could make up a theory of motivation as you went. A resident adviser suggested Oswald Spengler's The Decline of the West. Greene notes the book's fame and scorn for its predictions of political collapse and its veiled sympathy for fascism, later used to support insidious ideologies, but at the time he was struck by something else. Spengler claimed that man is the only being that knows death; other animals grow old with a consciousness confined to a moment that must seem eternal to them. From that knowledge comes an "essentially human fear," and Spengler concluded that every religion, every scientific investigation, every philosophy proceeds from it.
That gave Greene a theory of motivation that did not feel improvised. Perhaps the appeal of timeless laws and permanent proofs comes from our awareness that we are anything but timeless. Others had pushed the idea further:
- Otto Rank, an early disciple of Freud, saw the artist's creative impulse as an attempt to turn ephemeral life into personal immortality.
- Jean-Paul Sartre wrote that life is drained of meaning once you lose the illusion of being eternal.
- The thread running through these thinkers is that much of human culture, from art to science, is life reflecting on the finite nature of life.
Greene then reins in his younger self. By training he distrusts one-size-fits-all explanations (physics is littered with failed unified theories, and human behaviour is harder still). He now sees awareness of death as a considerable influence, not a blanket explanation. But in one domain mortality's reach is unmistakable: the value we place on permanence, whether through absolute truth, enduring legacies, monuments, immutable laws or one or another vision of the everlasting.
"Man is the only being that knows death." Oswald Spengler
The tension the book explores
The preface sets out the three tensions that drive everything that follows. These are the book's organising contradictions:
| What the laws are | What we are |
|---|---|
| They are, as far as we can tell, timeless. | We exist for the briefest moment of time. |
| They operate without concern for destination. | We constantly ask where we are headed. |
| They seem not to require an underlying rationale. | We persistently seek meaning and purpose. |
Greene's plan is to walk the timeline of the universe: the physical principles that produce stars, galaxies, life and consciousness inside a cosmos headed for decay; the arguments that life and mind, like individual lives, have limited spans; and the ways self-reflective beings have handled that knowledge. He promises nontechnical explanation throughout, with mathematical detail pushed to the endnotes.
The worm at the core
The chapter proper begins with a plain fact. For more than three billion years death has shadowed the flowering of life. Wait long enough and the ledger of births and deaths balances exactly. The course of any given life cannot be predicted; its end is a foregone conclusion. What is unusual is that only humans seem to notice. Other animals flee storms and earthquakes, but that is instinctive fear of a present danger. Most life lives in the moment. Only we can reflect on the distant past, imagine the future and grasp the darkness ahead.
Greene calls this terrifying, but quietly so: a foreboding that we learn to tamp down, accept and joke about. William James called it the "worm at the core of all our usual springs of delight." Greene borrows a line from comedian Steven Wright: it is enough to scare you half to death, twice.
The most influential framing comes from cultural anthropologist Ernest Becker. We live under constant existential tension: a consciousness that can soar to Shakespeare, Beethoven and Einstein is tethered to a body that will decay to dust. Becker's man is "literally split in two," aware of his towering uniqueness and yet destined to go back into the ground "blindly and dumbly to rot and disappear forever." According to Becker, this awareness drives us to deny death the power to erase us, and he identifies the main strategies:
Becker's routes to symbolic immortality
- Belonging to something that outlasts you. Family, team, movement, religion or nation: constructs that survive the individual's allotted time.
- Creative legacy. Works of art and invention that extend one's presence symbolically. Emerson: "We fly to Beauty as an asylum from the terrors of finite nature."
- Winning and conquering. Pursuing stature, power and wealth as if they conferred an immunity unavailable to the common mortal.
- Gestures toward eternity. Prophecies of an afterlife, teachings of reincarnation, the deliberately impermanent sand mandala.
What is new in our era is that science can tell a lucid story of the future as well as the past. Eternity may lie beyond our equations, but our analyses already show that the universe we know is transitory: planets, stars, galaxies, black holes and nebulae all end. Not only is each life finite; so too, apparently, is life itself. Carl Sagan's "mote of dust suspended on a sunbeam" is, Greene adds, an evanescent bloom in a cosmos that will ultimately be barren.
Stories of nearly everything
Humans are a species that delights in story. We spot patterns and join them into narratives, and the plural matters: there is no single volume of ultimate understanding, only many nested stories, each parsing reality with its own grammar and vocabulary.
Greene's illustration is Cervantes. Don Quixote speaks to our yearning for the heroic through the fragile Alonso Quijano, a character imagined by a man who was, physically, a collection of bone, tissue and cells, running on atomic and molecular processes refined by billions of years of evolution, on a planet made from supernova debris, in a space that emerged from the big bang. Every level is real and connected. Yet reading the novel gives an understanding of human nature that would stay opaque if written as the motion of the knight's atoms or as the firing of Cervantes's neurons. Connected stories told at different levels give genuinely different insights; the reductionist story is complete in its domain without being the most useful one everywhere.
Perhaps one day we will move smoothly between levels, explaining Rodin's vision in The Burghers of Calais, or how the wobble of a spinning dinner plate in a Cornell cafeteria churned through Richard Feynman's mind and helped him rewrite physics. Even without that capacity, Greene argues, much is gained by tracing when each story emerged on the cosmic timeline and how it rose to explanatory prominence.
The method of the whole book: every story (physical, biological, mental, cultural) is placed on the single cosmic timeline, so that its emergence and its eventual end can be seen together. The levels are not rivals; they are chapters of one chronology told in different languages.
Two lead characters: entropy and evolution
Across all the stories two forces share the lead.
- Entropy (chapters 2 and 3). Popularly tied to disorder and the claim that disorder always rises, entropy has subtleties that let systems appear to swim against the entropic stream. After the big bang, particles seemingly flout the drive to disorder as they assemble into stars, galaxies and planets, and eventually into matter that "surges with the current of life."
- Evolution (chapter 4). Natural selection starts before life exists. In molecular Darwinism, molecules compete with molecules in an arena of inanimate matter, and round after round of chemical selection likely produced ever more robust configurations, ending in collections we would call living. The details are at the research frontier, but two decades of progress suggest the field is on the right track.
- The partnership. Entropy and evolution sound like opposites, yet recent mathematical analyses suggest lifelike behaviour may be the expected product of a long-lived energy source (like the sun) raining heat and light on molecules competing for limited resources. Greene returns to this, the dissipative-adaptation idea, in chapter 4.
A billion or so years after earth formed it was teeming with life, after which the story is standard Darwinian fare: random mutations from cosmic rays or replication errors, most neutral, some harmful, some helpful; fitness-enhancing traits spread because "more fit" means more likely to survive and reproduce. Billions of years later a suite of mutations gave some organisms enhanced cognition, and some life became aware of being aware.
From mind to culture: two guiding lights
Consciousness (chapter 5) divides researchers. One camp expects a mechanistic explanation once we know the brain in far finer detail; the other expects the hardest problem we have ever faced, requiring new ideas about mind and perhaps reality itself.
On behaviour, opinion converges. Across tens of thousands of generations in the Pleistocene, our ancestors lived in hunter-gatherer groups; growing mental dexterity gave individuals refined capacities to plan, organise, communicate, teach, judge and solve problems, and groups used those abilities. That sets up three chapters: language and story (6), religion (7) and creative expression (8).
Greene names two complementary lenses for this cultural material:
Two guiding lights for explaining human behaviour
- Darwinian evolution applied to behaviour. The brain is a biological structure shaped by selection, so ask whether a behaviour enhanced survival and reproduction among those who practised it. Caveat: unlike the opposable thumb or upright gait, many inherited brain features mould predilections rather than definite actions.
- The inner life. Human action emerges from inherited tendencies mingled with complex, deliberative, self-reflective minds. The same faculties that let us innovate also let us hover above the timeline, see past and future, and so see our own end.
His compression of the second point is the chapter's sharpest line: to achieve "I think, therefore I am" is to run headlong into "I am, therefore I will die."
How do we live with this? One line of thought holds that we build defences out of creativity. We tell stories that move us to the centre of a vast universe and make our permanent erasure unthinkable. We make painting, sculpture, dance and music in which we wrest control of creation. We invent heroes, from Hercules to Sir Gawain to Hermione, who stare death down. We develop science, turning insight into powers earlier generations reserved for gods. We get to have our cognitive cake (the nimble thought that reveals our predicament) and eat it too.
Motives don't fossilize. Greene is careful here, and the caution should be kept in view through chapters 6 to 8. The terror-management reading of culture competes with two others he names: creativity as a by-product of an over-active pattern-detecting brain, and creativity as an adaptively superfluous luxury of a large brain freed from full-time foraging. Theories abound and unassailable conclusions are elusive. What is certain is only that our works, from the Pyramids to the Ninth Symphony to quantum mechanics, point toward permanence in their durability, if not their content.
Can thought last forever?
Greene distinguishes two kinds of prediction. Futurists extrapolate short runs of exponential technology into detailed visions; over decades that is unreliable, and over cosmic time it is "a fool's errand." His aim is the opposite: broad strokes painted with confidence, using physics we have measured and calculated.
The question that rides through the book is precise. Not whether an individual mind can be preserved (cryonics, uploading and the like are irrelevant), but whether the phenomenon of thought, supported by any physical substrate at all (brains, computers, entangled particles adrift in the void) can persist arbitrarily far into the future. Like the Tasmanian tiger or the ivory-billed woodpecker, thought might rise, flourish and go extinct.
The reasoning runs as follows. Human thought arose under a fortunate set of conditions, which is why we think on earth and not on Mercury or Halley's comet, and why damage to earth's climate matters. If thought is a physical process (an assumption chapter 5 examines), then it can only happen when stringent environmental conditions hold. So the cosmic question becomes: will the evolving universe go on providing those conditions?
Key figures and scales in Chapter 1
- 3+ billion years: the span over which death has shadowed life on earth.
- ~1 billion years: roughly how long after earth formed it was teeming with life.
- Tens of thousands of generations: the Pleistocene hunter-gatherer period that shaped human cognition.
- A dozen decimal places: the agreement between calculation and experiment that defines Greene's standard of evidence.
The answer from chapters 9 and 10 is not heartening. Planets, stars, solar systems, galaxies and even black holes are transitory, each ended by its own mix of quantum mechanics and general relativity, leaving a mist of particles drifting through a cold, quiet cosmos. Entropy supplies the language for the verdict, and its most startling version is this: in the far future, anything that thinks may burn up in the heat generated by its own thoughts; thought itself may become physically impossible. Greene will also examine more hopeful futures, but the straightforward reading is that the window for self-reflective beings may be extremely narrow. Nabokov called a life "a brief crack of light between two eternities of darkness"; the description may apply to life as a whole.
That undermines the most common consolation. We console ourselves with symbolic transcendence: others will come after us, and what we create will shape how they live. In a universe ultimately empty of minds, even a whisper meant for distant descendants disappears into the void.
From dread to gratitude: the Starbucks moment
Science usually lands cognitively: a fact about black holes tickles the mind and, if impressive enough, sticks. Rarely does it land viscerally. When Greene first studied predictions of the far future, the experience was cerebral. But when he made himself really imagine all life, thought, struggle and achievement as a fleeting aberration on a lifeless timeline, the result was dark. In decades of research he had felt elation and wonder; never before had mathematics overwhelmed him with "a hollow dread."
Over time the feeling changed. More often now, contemplating the far future brings calm and connection, his own identity subsumed by gratitude for the gift of experience. The first time it happened, in a Starbucks in New York City, the rigour-trained physicist distrusted it (perhaps the soy milk in his Earl Grey was off).
Why the shift made sense
- The premise. We descend from a lineage that soothed existential discomfort by leaving a mark; the more lasting the mark, the more a life seems to matter. Robert Nozick put it sharply: to be wiped out completely, traces and all, goes a long way toward destroying a life's meaning. (Greene notes the line could have come from George Bailey in It's a Wonderful Life.)
- Greene's own case. Without a traditional religious orientation, he says, a relentless focus on endurance has pervaded everything, which is why he chose a field of space, time and laws that transcend the moment.
- The reversal. Science itself shows that life and thought occupy a minute oasis on the cosmic timeline. Taken in fully, that makes our era worthy of something like reverence.
- Outcome. The calm marked a shift from grasping at a receding future to inhabiting a breathtaking, if transient, present: a cosmological counterpart to what poets and mindfulness teachers have always said. Emily Dickinson: Forever is composed of Nows. Thoreau: "eternity in each moment."
Greene states the book's purpose as providing that clarity: a journey from our best account of the beginning to the closest science can get to the end, examining religion, creative expression, science, the quest for truth and the longing for the timeless, driven by what Franz Kafka called our need for "something indestructible."
The conclusion stated up front
Greene ends the chapter by stating the book's conclusion in advance. Reality is governed by mathematical laws indifferent to conduct, beauty, companionship, understanding and purpose. Through language, story, art, myth, religion and science, we have harnessed our small part of a mechanical cosmos to voice our need for coherence, value and meaning. That contribution is exquisite and temporary. Nothing is permanent. Nothing is absolute. Therefore the only answers of significance in the search for purpose are those of our own making: during our brief moment in the sun, we have "the noble charge of finding our own meaning."
| Thinker | Claim about mortality | How Greene uses it |
|---|---|---|
| Spengler | Only humans know death; all religion, science, philosophy proceed from it | Seed of the book's motivational theory |
| Otto Rank | The artist turns ephemeral life into personal immortality | Creativity as death-response (Ch 8) |
| Sartre | Life loses meaning when the illusion of eternity is lost | The stakes of a finite cosmos |
| William James | Death is the "worm at the core" of delight | The quiet, background terror |
| Ernest Becker | We are split between soaring mind and rotting body | Terror management, symbolic immortality |
| Emerson | We fly to Beauty as asylum from finitude | Art as refuge |
| Nozick | Total erasure destroys much of a life's meaning | Why the cosmic future stings |
| Nabokov | Life is a brief crack of light between two darknesses | Scaled up to life itself |
| Dickinson, Thoreau | Forever is made of nows; eternity in each moment | The book's resolution |
| Kafka | We need "something indestructible" | The pull of the timeless |
Broad strokes beat false precision. Greene's forecasting stance (confident on the coarse shape of the future, agnostic on the details) is the right posture for any long-horizon roadmap. In enterprise architecture and AI strategy, commit to the constraints you can actually calculate (cost curves, regulatory direction, data gravity) and refuse detailed extrapolations of short exponential runs, which is exactly the error Greene attributes to breathless futurists.
The Language of Time
This chapter builds the toolkit the whole book runs on. Entropy counts how many microscopic arrangements look the same at the macroscopic level; the second law says entropy overwhelmingly tends to rise because high-entropy arrangements vastly outnumber low-entropy ones. The laws of physics do not distinguish past from future, so the arrow of time comes from a special low-entropy starting point, and order can only be built or kept by the "entropic two-step": lowering entropy here by dumping more of it, as waste heat, there. You are a steam engine in exactly this sense, which is why the universe's ability to absorb waste will decide the fate of life and mind.
Russell's "universal death"
On 28 January 1948, between a Schubert quartet and English folk songs, BBC Radio (the Third Programme) broadcast a debate on the existence of God between Bertrand Russell and the Jesuit Frederick Copleston. Russell, who would win the 1950 Nobel Prize in Literature and who had been dismissed by both Cambridge and the City College of New York for his views, argued that the universe "has crawled by slow stages to a somewhat pitiful result on this earth" and will crawl "by still more pitiful stages to a condition of universal death." If that counted as evidence of purpose, it was a purpose that did not appeal to him, so he saw no reason to believe in any sort of God.
Greene parks the theology for later chapters and focuses on the physics behind "universal death." Its roots are humble: the nineteenth-century steam engine. Even after the leap from manual to mechanical labour, engines were wasteful: roughly 95 percent of the heat from burning wood or coal was lost. Decades of work on burning less and getting more led to the second law of thermodynamics, which in very loose terms says waste is unavoidable. Because it applies to all matter and energy, animate or inanimate, it reads like a sentence of slow cosmic decline. Greene's claim is that a precise understanding reveals something richer: a progression that explains how beauty and order arise against a backdrop of decay, and even hints at exotic ways around Russell's end.
Steam engines and Carnot
The engine's principle is simple. Steam expands when heated and pushes a piston; the thrust turns a wheel, grinds a mill or drives a loom; the steam then cools and the piston slides back, ready for the next cycle as long as fuel burns. The scientific questions were deeper than the engineering: can the engine be described with mathematical precision, is there a limit to its efficiency, and are some of its features independent of design and materials, and so universal?
The French military engineer Sadi Carnot launched thermodynamics (the science of heat, energy and work) with his 1824 Reflections on the Motive Power of Fire, which sold poorly but inspired a century of physics.
The idealised Carnot cycle (Greene's endnote)
- Hot isothermal expansion. Steam absorbs heat from a hot reservoir and pushes the piston at constant temperature.
- Adiabatic expansion. Disconnected from the source, the steam keeps pushing; temperature falls, entropy stays constant (no heat flows).
- Cold isothermal compression. Connected to a cooler reservoir, the piston is pushed back and waste heat is expelled.
- Adiabatic compression. Disconnected again, the piston completes its return and the steam is brought back to its starting temperature. The cycle repeats.
A statistical perspective
Newton's laws give ironclad predictions: give the position, velocity and forces, and the trajectory follows. But we quietly simplify, treating the moon or a baseball as a single particle. (Greene's note: this is exact for the centre-of-mass motion, because Newton's third law cancels internal forces.) Steam cannot be treated that way: perhaps a trillion trillion molecules slam into the piston, and their motion is the engine's mechanism. No one can track them all.
The rescue was a change of perspective: large collections have their own regularities. You cannot predict your next sneeze, but you can predict roughly 80,000 sneezes per second worldwide (Greene's arithmetic: about one sneeze per person per day, about 7 billion people, about 86,000 seconds in a day). James Clerk Maxwell, Rudolf Clausius and Ludwig Boltzmann applied the same idea to molecules: drop individual trajectories, compute averages over vast numbers. The bonus was that the averaged properties (pressure, density, temperature) are exactly the ones that matter. Pressure on a piston depends on the average pummelling by trillions of molecules, not on any one path.
This sounds unremarkable in an age of polls and big data, but it was a radical break from the rigid precision of physics, and some respected scientists still doubted atoms existed into the twentieth century. In 1905 Einstein explained Brownian motion, the jitter of pollen grains in water, as bombardment by water molecules. After that, "you had to be one heck of a contrarian" to doubt molecules. Statistical mechanics went on to explain earth's atmosphere, the solar corona and the interior of neutron stars.
The puzzle: laws that ignore the direction of time
Experience is lopsided. Van Gogh could not lift the paint off Starry Night with reverse brushstrokes; the Titanic could not reverse engines and unrip its hull; we cannot reclaim our youth. You would expect the laws of physics to contain some term that forbids "that to this." They do not. Through Newton (mechanics), Maxwell (electromagnetism), Einstein (relativity) and quantum physics, one feature has held steady: the equations treat evolution toward the future and toward the past identically. If the laws allow a sequence of events, they allow its reverse. Divers popping feet-first out of pools and shards reassembling into a Tiffany lamp are fully lawful. Greene found this "just shy of ludicrous" as a student.
Time-reversal is slightly subtler than stated. In a note Greene concedes that for some exotic processes (weak-force decays), reversing time alone is not a symmetry. What the known laws must respect is the combined CPT symmetry: reverse time (T), swap particles for antiparticles (C) and mirror left and right (P). The arrow-of-time argument is unaffected, because CPT still guarantees a lawful reversed history for every allowed one.
Entropy: counting look-alikes
Greene's teaching device is 100 pennies shaken in a bag and dumped on a table. All heads would astonish you. Why? There is exactly one way to get all heads, 100 ways to get one tail, and the counts explode from there.
Ways to arrange 100 pennies (the binomial count)
- 0 tails: 1 way.
- 1 tail: 100 ways.
- 2 tails: 4,950 ways = (100 × 99)/2.
- 3 tails: 161,700 ways.
- 4 tails: 3,921,225 ways (about 4 million).
- 5 tails: 75,287,520 ways (about 75 million).
- 50 tails: 100,891,344,545,564,193,334,812,497,256 ways, about 1029. A 50/50 split is about a hundred billion billion billion times more likely than all heads.
The key move is that we do not care which particular coin is heads. We sort outcomes into groups that "pretty much look the same" (same ratio of heads to tails), and those groups have wildly unequal memberships. Group size matters, because a random shake almost certainly lands in a big group.
Entropy is the size of the group: the number of microscopic configurations that look, macroscopically, like the given one. Many look-alikes means high entropy; few means low entropy. Precisely, entropy is the logarithm of that number, Boltzmann's S = k log W, so that entropies of combined systems add.
This connects to the colloquial meaning. A messy desk has high entropy because most rearrangements of its papers and pens still look messy; a pristine desk has low entropy because almost any rearrangement spoils it. Messy arrangements vastly outnumber orderly ones.
Entropy for gases: the steamy bathroom
For a gas, "looks the same" means same volume, same temperature, same pressure. All the molecular configurations sharing those values form one macrostate, and the entropy is the number of them. This is enormously powerful: by understanding the typical features of the most common configurations, we predict the macroscopic properties we care about, with statistical but fantastically reliable accuracy, without tracking a single trajectory.
Greene's case study is the steam in your bathroom after a long hot shower:
| Factor | Lower entropy | Higher entropy | Why |
|---|---|---|---|
| Volume | Steam knotted in one corner | Steam spread through a large room | More places to put molecules means more rearrangements that look alike |
| Temperature | Cold (low average molecular speed) | Hot (high average speed) | Low speeds are squeezed against the floor of zero; high averages allow a wider spread of speeds |
| Pressure / number | Short shower, fewer molecules | Long shower, more molecules | More molecules give more arrangements |
Summary rule: fewer molecules, lower temperature or smaller volume means lower entropy; more molecules, higher temperature or larger volume means higher entropy. (Greene notes that temperature is strictly proportional to average kinetic energy, the average of velocity squared; "average speed" is the intuitive stand-in.)
The practical rule of thumb follows. High-entropy states are typical, "a dime a dozen." Low-entropy states should command attention because they are rare. Uniform steam after a shower is unsurprising; steam gathered into a perfect cube hovering in front of the mirror is so unlikely that Greene would "bet my life" chance is not the explanation. An egg, an anthill or a mug is similarly a low-entropy configuration that calls out for an explanation, and the explanation is life arranging matter. Low entropy is a diagnostic: a clue that a powerful organising influence is at work.
The two laws of thermodynamics
| First law | Second law |
|---|---|
| A law of conservation: energy at the end of a process equals energy at the start. | A law of growth: entropy has an overwhelming tendency to increase. |
| Must count every form: kinetic, potential, radiation, heat. | Special configurations become ordinary ones: the pressed shirt wrinkles, the garage turns into a mess. |
| Precisely: change in internal energy = heat absorbed minus work done. | Precisely: a system not at maximum entropy will almost certainly evolve toward it. |
Why the second law is true is "a numbers game." Start the pennies at all heads and jostle them. To return to all heads the jostling would have to flip back exactly the few coins that became tails; it is overwhelmingly more likely to flip a random selection, producing more tails. The march continues until the pennies reach roughly 50/50, after which they wander among the members of the largest groups. Bread baking in the oven works the same way: there are far more ways for aroma molecules to spread than to cluster, so the smell fills the house. That is the second law, and that is why it holds.
Energy quality and the arrow of time
The two laws are linked by a principle Greene will return to repeatedly: all energy is not created equal. A stick of dynamite stores its energy in a compact, orderly chemical package that is easy to use: put it where you want and light the fuse. After the explosion the same energy exists (first law), but it is now the chaotic motion of dispersed particles, almost impossible to harness.
Low entropy = high-quality energy (concentrated, usable). High entropy = low-quality energy (dispersed, hard to use). The first law says the quantity of energy is conserved; the second says its quality deteriorates. So the future differs from the past because the energy powering the future is of lower quality: the future has higher entropy.
Boltzmann's catch and the big bang
There is a subtlety that even Boltzmann took time to absorb: the second law is statistical, not absolute. It does not forbid entropy from falling; it makes falling absurdly unlikely. For bread, the molecules could stream back into the loaf and uncook it into cold raw dough, but the odds are "closer to zero than the likelihood of splattering paint on a canvas and replicating the Mona Lisa." This had to be so: time-symmetric laws cannot forbid entropy decrease outright, since every entropy-raising history has a lawful reverse. Entropy can decrease. It's just ridiculously unlikely.
That changes what needs explaining. Given a world below maximum entropy today, the second law predicts entropy will rise tomorrow. Some stop there. Greene says the job is half done: why is today's world in such a special, low-entropy state at all, full of planets, peacocks and people? A maximum-entropy universe would just meander among its typical states (a cosmic version of the uniformly steamy bathroom) and its future would look like its past. Following the second law backward, today's entropy came from yesterday's lower entropy, and so on back to the big bang, which must have been highly ordered. The low-entropy beginning is why the universe has an eventful future at all. Why the beginning was so ordered is chapter 3's question.
A hidden assumption about "rare." In a technical note Greene flags something most popular treatments omit. Calling entropy-decreasing microstates "rare" requires choosing a measure on the space of configurations; the uniform measure works because experiment confirms it for everyday systems. For the early universe we have no comparable data, so intuitions about "special" or "generic" initial conditions lack the same empirical footing. He also notes that gravity changes what maximum entropy looks like (in a finite region it is a black hole), which chapter 3 develops.
Heat, entropy and why engines must dump waste
The link between Boltzmann's counting and the old "waste heat" version of the law runs through heat flow. When you grab a hot pan handle, nothing material flows: molecular agitation passes from the faster molecules of the handle to the slower ones in your hand, as information passes in a game of telephone. The old idea of a fluid called "caloric" was wrong.
Entropy behaves the same way. Your hand warms, the range of molecular speeds widens and its entropy rises; the handle cools and its entropy falls. That is not a statistical fluke, it happens every time you grab the pan, because the second law applies to the total system (handle, hand, pan, stove, air). Careful accounting shows the hand's gain outstrips the handle's loss. Hence the slogan: entropy rides the wave of flowing heat. To move entropy from here to there, heat must flow from here to there.
Now the steam engine. Its cycle requires that its macrostate, and so its entropy, be reset each time. All the entropy acquired in a cycle (from the fuel, from friction) must be expelled by the cycle's end, and the only way to move entropy is to move heat. So the engine must release waste heat: the old second law, derived from the new one.
Why is anything left over for useful work? Because a cooler environment turns a given amount of heat into a larger entropy increase. Greene's analogy: turn up the beat at a wild New Year's party and the dancers barely change; turn it up at the Thiksay Monastery and meditating monks start krumping. The engine takes in heat and entropy from hot fuel and can purge all that entropy by expelling only part of the heat into the cool surroundings. The remainder drives the piston.
The formula behind the space-heater analogy (Clausius)
- Entropy delivered by heat: ΔS = Q / T. The same heat Q produces a bigger entropy change at a lower temperature T.
- Engine bookkeeping: take in Qhot at Thot, dump Qcold at Tcold, with Qcold/Tcold ≥ Qhot/Thot.
- Work available: W = Qhot − Qcold, so the best possible efficiency is 1 − Tcold/Thot (the Carnot limit). The bigger the temperature gap, the more work per unit of fuel.
The entropic two-step
The steam engine is just the first example of a process that drives the entire book:
The entropic two-step: any process in which a system's entropy decreases because it exports a more-than-compensating increase to its environment. Entropy falls here, rises more there, and the total still goes up. When energy flows through a system, it carries entropy away and can sustain or even create order in its wake. This is how a universe heading toward disorder builds stars, planets and people.
You are a steam engine
What if the entropy reset failed? The engine would retain waste heat, grow hotter each cycle and break down. Inconvenient for a locomotive; existential for us, because your life is cyclical in the same way. You burn food and air to power your body and your thinking (thought is molecular motion in the brain), and you must purge waste heat to reset your entropy. You do it constantly; that is why military infrared goggles can spot people at night.
This is what was behind Russell's mood. Life wages a relentless battle against accumulating waste, and survival requires an environment that will absorb the entropy we generate. That poses the questions for the rest of the book: is the observable universe a bottomless pit for waste? Can life dance the two-step forever, or will the cosmos become "stuffed" and unable to absorb the heat of the very activities that define us? Russell's own phrasing, from his essay on a free man's worship, was that the "whole temple of Man's achievement must inevitably be buried beneath the debris of a universe in ruins." First, though, Greene must explain how entropy and the second law produced environments where life could begin, which means going back to the big bang.
Updates since 2020. Greene's sneeze estimate rests on 7 billion people; with the UN putting world population at about 8.2 billion in 2025, the figure is now nearer 95,000 sneezes per second. The physics is untouched. On thermodynamics itself, the field Greene leans on in chapter 4 (stochastic thermodynamics, which tracks entropy in single molecules and small systems) has since become a mainstream experimental subject, and work on thermodynamic uncertainty relations has tightened the link between precision, speed and dissipation in molecular machines. That strengthens Greene's point that every act of order-keeping has an entropy price.
Every system that keeps its shape pays an entropy tax. For architects, the two-step is a precise metaphor for technical debt and operational toil: a platform stays orderly only by continuously exporting disorder (patching, migration, data cleansing) to somewhere, at a cost. In AI operations the analogy becomes literal: every joule a GPU spends computing leaves as waste heat, so a data centre's ability to reject heat, not its chip count, increasingly caps the compute a site can sustain.
Origins and Entropy
If the second law drives the universe toward disorder, where do galaxies, stars, planets and brains come from? Greene's answer removes the need for a designer: ordered pockets form because the universe is obeying the second law, not defying it. Inflation (repulsive gravity powered by a uniform "cosmic fuel") starts the expansion and seeds tiny quantum variations; gravity amplifies them into clumps; and in a gravitating cloud, losing heat makes the core hotter, a self-sustaining entropic two-step that builds stars, which nuclear fusion then runs for billions of years as engines that push cosmic entropy higher.
The question and the designer reflex
Because physics can now peer back to within a fraction of a second of what may have been the beginning, people often assume science and religion are about to collide or merge there. Greene says he is asked about a creator nearly as often as about physics. This chapter takes one point of contact: how can a law of rising disorder coexist with the Milky Way's arms, earth's landscapes, the folds of the human brain and the art those brains make? The old answer was that order is hewn from chaos by a supreme intelligence, an inference that fits daily life because so much order around us is human handiwork. Greene's thesis is that a proper reading of the second law makes an intelligent designer unnecessary: concentrated energy and order, with stars as the archetype, are a natural consequence of the universe becoming more disordered, and indeed act as catalysts that help the universe reach its entropic potential.
Lemaître, Einstein and the expanding universe
1915: Einstein's general relativity, the new theory of gravity.
Mid-1920s: Georges Lemaître, a Jesuit priest with a doctorate from MIT, applies Einstein's equations to the whole cosmos and finds an expanding universe with a beginning (in parallel with Alexander Friedmann in Russia).
1927: Einstein, committed to an eternal, static cosmos, tells him: "Your calculations are correct but your physics is abominable."
Late 1920s: Edwin Hubble at Mount Wilson Observatory shows distant galaxies are rushing away, faster the farther they are, matching the equations.
Early 1930s: Einstein performs one of science's famous turnabouts and accepts a universe with a beginning.
Greene's endnote adds that Einstein's conversion had two causes: Hubble's data and Arthur Eddington's proof that Einstein's static universe was unstable (nudge it and it expands or contracts without stopping); some historians think the theory weighed more than the data. The modern summary: about fourteen billion years ago, the whole observable universe was a stupendously hot, dense nugget that rapidly expanded and cooled; particles slowed, clumped, and formed stars, planets, debris and us. The rest of the chapter refines that story to show how order arises without intent or design.
Repulsive gravity and inflation
Physics is full of opposites (matter and antimatter, positive and negative charge), yet since Newton gravity seemed purely attractive. Einstein's general relativity changed that: gravity can be repulsive. Clumpy things like stars and planets pull; but exotic, smoothly spread energy can push.
Its decisive use came more than half a century later. As a young postdoc, Alan Guth asked what force set the expansion going, since Einstein's equations described expansion but not its cause. After a late-night calculating frenzy in December 1979 he found the answer:
Inflation in four steps
- The fuel. A tiny region, perhaps 10−27 metres across ("a billionth of a billionth of a billionth of a metre"), filled with a scalar field Guth called the inflaton (the missing "i" is deliberate). A scalar field assigns a single number to each point, unlike vector fields such as electric fields.
- The condition. The field's energy is uniform, like steam of equal density throughout a sauna, not clumped.
- The push. Uniform energy of this kind produces repulsive gravity strong enough to blow the speck up, almost instantly, to the size of the observable universe or far larger.
- The bang. Repulsive gravity is the outward "kick" of the big bang.
In the early 1980s Andrei Linde, and independently Paul Steinhardt and Andreas Albrecht, built the first fully viable versions. Inflation also solved standard-cosmology puzzles (the horizon, flatness and monopole problems), but Greene, like Guth, motivates it by the more intuitive question of what drove the expansion.
The afterglow: quantum jitters written across the sky
As space stretched, the primordial heat thinned into a faint glow. Physicists predicted it in the 1940s; Arno Penzias and Robert Wilson of Bell Labs stumbled on it in the 1960s: the cosmic microwave background (CMB), just 2.7 degrees above absolute zero. Part of the static on an old television after the evening's broadcasts ended was this relic of the big bang.
Inflation sharpens the prediction by bringing in quantum mechanics. Normally quantum effects are confined to the microworld, but stretching space reveals quantum features the way stretching spandex reveals its stitches. The key ingredient is Heisenberg's 1927 uncertainty principle: properties like position and speed carry an irreducible quantum fuzziness. In Greene's image, classical physics draws a trajectory with a sharp quill; a quantum physicist runs her finger along the wet ink. He stresses that the "fog" metaphor flatters our classical bias: it is the classical picture that is the approximation. Nobody knows why reality is quantum; we accept it because a century of experiments confirms it.
Applied to inflation, uncertainty means the inflaton cannot be perfectly uniform; its energy is a tiny bit higher here and lower there. Inflation stretches these jitters across space, leaving hot and cold spots differing by about one part in 100,000 in a specific statistical pattern. Since the early 1990s, satellites above the atmosphere (COBE, WMAP, Planck) have confirmed that pattern with ever greater precision.
The chain is staggering: Einstein's equations + Guth's hypothetical field + Heisenberg's uncertainty predict a specific fossil pattern of temperature variations in the sky; thermometers built nearly fourteen billion years later by a young species in the Milky Way find exactly that pattern.
Greene's own caveat on inflation. Success is not proof. The energies involved are perhaps millions of billions of times beyond laboratory reach, and alternatives exist (including the bouncing cosmology of Steinhardt and Turok, chapter 10). His notes go further: inflation is a framework, not one theory; the simplest versions are in tension with data; critics say the framework is too flexible to be falsified; supporters call that normal theory refinement. What cosmologists broadly agree on is a phase in which the comoving horizon shrank; whether inflation caused it is less certain. On dark matter he notes that repeated failure to detect it directly is lending modified-gravity alternatives more attention.
Since 2020. The tension Greene mentions has sharpened. BICEP/Keck (2021) pushed the tensor-to-scalar ratio below r ≈ 0.036, ruling out the simplest "chaotic" inflation potentials. In 2025 the Atacama Cosmology Telescope's DR6 analysis, combined with Planck and DESI, nudged the spectral index up to about ns ≈ 0.974, putting pressure even on the long-favoured Starobinsky model and prompting a wave of "beyond Starobinsky" papers. Inflation as a framework survives; its simplest textbook versions are fading. On dark matter, the LZ experiment (2024–25, 4.2 tonne-years) set world-leading limits and still saw no WIMP signal. And JWST has found bright galaxies earlier than many models expected, which tests (so far without breaking) the gravitational snowball timeline described below.
Why was the beginning so ordered?
Leibniz's question, "Why is there something rather than nothing?", is no closer to an answer. A true answer would need no antecedent and would be immune to "why these laws and not others?" Inflation does not come close: it presupposes space, time, the inflaton, quantum mechanics, general relativity and the mathematics from calculus to differential geometry they rest on. The method is pragmatic: choose laws guided by experiment and mathematical intuition, find conditions that would trigger expansion, postulate that those conditions held, and compute forward. "It's nothing to sneeze at."
The entropy puzzle follows directly. If entropy has risen since the bang, the bang had very low entropy, and low entropy calls out for explanation. All-heads pennies have one (someone arranged them); who or what arranged the early universe? Greene admits the question has kept him up "for many a night (literally)," yet science cannot even tell whether it deserves angst: without knowing how the initial conditions were set, we cannot judge whether they were exotic or ordinary.
The boiling-vat scenario for starting inflation
- Setup. Picture the pre-inflationary universe as chaotic, the inflaton fluctuating wildly like the surface of boiling water.
- Requirement. Inflation needs a small patch where the field is uniform: like a patch of the bubbling surface suddenly going flat. That is a low-entropy fluke.
- Resolution. Wait long enough and unlikely things happen. Shake pennies enough times and they will land all heads. Before the bang there was no one "arms crossed and shoe tapping"; when the fluke finally happens, repulsive gravity takes over and the cosmos begins.
- Status. One proposal among many. Linde (personal communication, 2019) quips that for every three researchers there are at least nine opinions; his own favours quantum tunnelling from a space of all possible geometries, where time and temperature may not yet mean anything.
The principle of indifference, and the "past hypothesis." Greene's notes are frank about two assumptions. First, calling a uniform inflaton patch "unlikely" assumes every microstate is equally likely (a "flat measure"). That is empirically tested for pennies and gases, but for the universe we have one run of the experiment, so a hard-nosed empiricist could call whatever conditions produced our universe "definite," not "special." Physicists find such a tailored measure "unnatural" and want it derived from first principles, the same instinct behind thirty years of work on fine-tuning (the Higgs, the horizon and flatness problems). Greene hopes no feature must simply be accepted, but concedes there is no guarantee. Second, because the laws are time-symmetric, the second law alone predicts higher entropy in the past as well as the future, which contradicts memory and records and would undermine the very experiments supporting the laws. Physicists therefore add the past hypothesis (philosopher David Albert's term): entropy was anchored low near the big bang.
From inflaton to matter and stars
Within about 10−27 seconds (a billionth of a billionth of a billionth of a second), repulsive gravity had stretched a tiny region perhaps far beyond our cosmic horizon (the distance beyond which light has not had time to reach us). The inflaton's energy, like a stretched soap bubble, was unstable; it "popped," converting its energy into a mist of particles. Within minutes a cascade of reactions (heavy particles decaying, others binding) produced protons, neutrons and electrons, plus likely dark matter. The universe was a hot, nearly uniform gas.
"Nearly" is everything. The same quantum jitters that mark the CMB left slightly denser regions, and gravity amplified them:
A slightly denser region pulls slightly harder, gathers more matter, grows denser, pulls harder still: a gravitational snowball. After hundreds of millions of years, clumps become massive, compressed and hot enough to ignite nuclear fusion. Quantum uncertainty, magnified by inflation and concentrated by gravity, became the stars in the night sky.
Hurdles on the road to disorder: the oven explosion
Systems rarely jump straight to maximum entropy. Bread's aroma takes time to reach a far bedroom, and a closed kitchen door slows it further. Some hurdles are built into the laws themselves, and Greene tells the episode that taught him this physically.
Fourth grade, a cold pizza and an unlit oven
- What happened. Greene set the oven to 400 degrees, waited ten minutes, found the pizza still cold, realised he had turned on the gas without lighting it (no built-in pilot light), leaned in and struck a match.
- Outcome. The accumulated gas exploded; a wall of flame singed off his eyebrows and eyelashes and left second- and third-degree burns on his face and ears, followed by months of painful healing. (He now does most of the family cooking, with a twinge when his children light the oven.)
- The physics. Methane and oxygen mix peacefully; a far higher-entropy state is available (CH4 + 2O2 → CO2 + 2H2O + energy) but cannot be reached by mixing alone. It needs a catalyst. The match's energy broke some bonds; recombination released more energy, breaking more bonds: a cascade.
- Lesson. There are speed bumps on the entropic road, and catalysts let systems clear them. Chemical bonds run on the electromagnetic force, so most everyday entropy leaps are electromagnetic. In the cosmos, gravity and the nuclear forces (strong: binds nuclei; weak: drives radioactive decay) do the work, and stars, planets and life are among the transient structures formed along that path.
Gravity's reversal of intuition
Gravity is by far the weakest force. Lift a coin and your arm beats the whole earth. Between two electrons, gravity is about 1042 times weaker ("a million billion billion billion billion") than their electric repulsion. That is why chapter 2 could ignore it. But gravity is cumulative: every bit of a big body pulls on every bit of another, so on astronomical scales it dominates the entropy story.
Greene's thought experiment: a box larger than the sun containing a ball of hydrogen massive enough for its own gravity to matter. Without gravity the gas would spread evenly and reach uniform temperature. With gravity:
- Molecules leading the outward migration are tugged back, slow down, and the outer shell cools as it expands.
- Molecules nearer the centre are pulled inward, speed up as they fall, and the core heats as it shrinks.
- Heat flows from the hot core to the cool shell, as usual. But now the effect reverses: as the core loses heat it contracts and gets hotter; as the shell gains heat it expands and gets cooler.
"If your coffee behaved this way, you'd be well advised to drink it quickly." It is a self-amplifying loop, like credit-card debt: the bigger the temperature gap, the more heat flows, which widens the gap. For debt you intervene with a payment or bankruptcy; for the core nature intervenes with nuclear fusion. When nuclei (not just electrons, as in chemical burning) slam together hard enough, they fuse and release huge energy whose thermal pressure balances gravity. Contraction halts. A stable, long-lived source of heat and light results. A star is born.
| Region | Volume | Temperature | Net entropy | Why |
|---|---|---|---|---|
| Core (becomes star) | Shrinks (entropy down) | Rises (entropy up) | Decreases | Volume loss wins; a star is a pocket of order |
| Shell (surroundings) | Grows (entropy up) | Falls (entropy down) | Increases | Volume gain wins |
| Total | Increases | Shell's gain exceeds core's loss: the second law holds |
The virial back-of-envelope (Greene's note)
- Virial theorem: average kinetic energy K = −U/2, with gravitational potential U ∝ −1/R, so K ∝ 1/R and typical speed v ∝ 1/√R.
- Phase-space volume ∝ (spatial volume) × (momentum volume) ∝ R3 × (1/√R)3 = R3/2: shrinking R lowers entropy even as temperature rises.
- Because of the factor 2, potential energy falls faster than kinetic energy rises, so the contracting core loses energy too; that energy radiates to the shell, raising its entropy. A gravitating system has negative heat capacity.
| Steam engine two-step | Gravitational two-step |
|---|---|
| Emits heat and cools; peters out without fresh fuel | Emits heat and heats up; the loop sustains itself |
| Needs an intelligence to design, build and fuel it | Sculpted and powered by mindless gravity |
| Order is engineered | Order is spontaneous |
Fusion: gravity's equal partner
Without significant gravity, the second law drives things toward homogeneity. With enough gravity it makes a U-turn: matter clumps here and spreads there, energy concentrates here and disperses there, entropy falls here and rises there. Scientists often say gravity is the ultimate source of all cosmic structure. Greene argues for a more generous picture: an equal partnership with the nuclear force.
When the sun fuses hydrogen into helium (billions upon billions of times per second), the product nucleus is more complex and lower in entropy. A little mass becomes energy (E = mc²), mostly photons that heat the interior and stream from the surface. That starlight carries enormous entropy into space, more than compensating. The match that lit Greene's oven has a stellar analogue: gravity is the catalyst that crushes the core to fusion temperatures; once lit, fusion runs for billions of years on the front line of the entropic charge, making complex nuclei and a steady bath of light, both essential for life. Without gravity, uniformly spread particles are already at maximum entropy; with gravity and fusion, squeezing matter into burning balls unlocks a hidden cache of entropy and drives the total higher still.
Chapter 3 by the numbers
- ~14 billion years: age of the universe (13.8 billion in current measurements).
- 10−27 m: rough size of the pre-inflationary patch.
- 2.7 K: temperature of the cosmic microwave background.
- 1 part in 100,000: size of CMB hot/cold variations.
- Hundreds of millions of years: time for gravity to snowball the first stars.
- ~1042: how much weaker gravity is than electromagnetism between two electrons.
Structure need not be designed to be real. Greene's gravitational two-step is a clean model of emergent order under a positive feedback loop with a stabiliser (contraction amplified until fusion pushes back). The same shape appears in platform economics and data networks: a small density advantage compounds into a dominant cluster until a counter-force (regulation, congestion, cost) halts it. Architects modelling where concentration will appear should look for the amplifying loop and the eventual brake, not for an intending hand.
Information and Vitality
What separates a rabbit from a rock, when both are made of the same particles obeying the same laws? Greene follows the ingredients of life from the big bang and stars to a water-covered earth, then shows that all life runs on one information system (the DNA code) and one energy system (redox cascades charging proton batteries that build ATP), which is strong evidence for a single common ancestor. Before biology there was molecular Darwinism, and thermodynamics may have pushed matter toward it. Life breaks no physical law: it is "physics orchestrated," a dissipative structure that, like a star or a steam engine, helps the universe release the entropy locked in matter.
Schrödinger's question: rabbits and rocks
In 1953 Francis Crick wrote to Erwin Schrödinger (1933 Nobel laureate in physics): he and James Watson had both been inspired by Schrödinger's "little book" What Is Life?, and his term "aperiodic crystal" was looking very apt. Enclosed were reprints of the one-page double-helix paper that, with Maurice Wilkins of King's College (who said the book "set me in motion"), won the 1962 Nobel Prize.
The missing name. Greene's note credits Rosalind Franklin, the chemist and crystallographer whose "Photograph 51" was shown to Watson and Crick without her knowledge by Wilkins and was instrumental to the model. She died in 1958, four years before the prize, which cannot be awarded posthumously. (Archival work published in 2023 argues she was an equal contributor rather than a victim of simple theft; either way, the main text's account of three men inspired by a book omits the scientist whose data made the structure visible.)
Schrödinger gave the lectures behind the book on three Fridays in February 1943 at the Dublin Institute for Advanced Studies, in the Fitzgerald Building at Trinity College. He warned they "could not be termed popular," yet more than four hundred people, including the Irish prime minister, came while war raged on the continent. His question: how can events inside a living organism be accounted for by physics and chemistry? Rocks and rabbits are both vast collections of protons, neutrons and electrons governed by the same laws; what makes the rabbit's particles different?
Biologists often define life by what it does (takes in raw materials, powers self-sustaining functions, expels waste, reproduces). Schrödinger, a reductionist physicist, wanted the physical underpinnings. The difficulty is that life is invisible at small scales: no electron or water molecule carries a mark saying "alive." Life shows up in collective, coordinated behaviour; a single cell has over a trillion atoms. Studying life particle by particle is like hearing a Beethoven symphony one note at a time. Schrödinger made the point himself: if a stray atom could wreck a body or brain, survival would be hopeless, so living systems must be large enough to function robustly despite atomic jitter. (His epilogue invoked the Upanishads and an "omnipresent, all-comprehending eternal self," which cost him his first publisher; Greene does not share his view of free will.)
Nested stories and Ken Wilson
Physics organises itself into stories indexed by scale. To see how a baseball flattened by Mike Trout's bat springs back, study its molecules; to track its flight over the fence, ignore them. The question determines the useful story. This works because the universe is coherent at each scale: Newton knew nothing of quarks yet could compute the ball's path. If baseball trajectories required quantum mechanics, physics would never have got started.
The complementary task is synthesis. Ken Wilson (Nobel 1982) created the renormalization group, a mathematical procedure that analyses physics across scales, from below LHC distances to atomic sizes, and shows how each level's language and equations hand off to the next. It sits at the core of modern physics.
Greene's "nested naturalism" (from his notes)
- Reductionism is universal. Everything that happens admits a description in terms of fundamental constituents obeying fundamental laws. Life cannot contravene physics; nothing can.
- Reductionism is explanatorily limited. Higher-level stories wrap around it like the outer layers of a nest, and are often far more insightful.
- All levels must be mutually consistent.
- New concepts emerge at higher levels with no lower-level correlate: a water wave makes sense for many molecules, none for one.
- Neighbours on the map: Barry Stroud's expansive naturalism, John Dupré's pluralistic naturalism, Hawking and Mlodinow's model-dependent realism, David Marr's computational / algorithmic / physical levels. Greene positions himself against scientism (only science yields knowledge) and against humanists who scoff at reductionism.
The plan of the chapter follows from this: take a reductionist stance, trace life's atomic ingredients, the environment of sun and earth, and the deep unity of all life, which sharply defines what any origin-of-life theory must explain. The payoff is thermodynamic: living things share a kinship not only with one another but with stars and steam engines; life is one more way the universe releases the entropy potential locked in matter.
Where the atoms came from
Grind up anything that lived and you find six elements: carbon, hydrogen, oxygen, nitrogen, phosphorus, sulfur (SPONCH). Building atoms is simple in principle (protons and neutrons in a nucleus, electrons arranged by quantum rules) and hard in practice, because protons repel each other and only extreme temperature and pressure push them close enough for the strong force to grab hold.
Big bang nucleosynthesis. In the late 1940s George Gamow (who once tried to defect from the USSR by kayaking across the Black Sea with coffee and chocolate) and his student Ralph Alpher proposed that the early fireball built the elements. They were partly right. For the first ~90 seconds photons were too energetic and smashed any nascent nuclei. After that, fusion could proceed, but the recipe is order-dependent and no stable nucleus has five or eight nucleons, which bottlenecks the chain. By ~10 minutes the universe was too cool and thin. Result: ~75% hydrogen, ~25% helium by mass (about 92% / 8% by number of atoms), with traces of deuterium, helium-3 and lithium, exactly matching observations.
Stellar nucleosynthesis. In the 1920s Arthur Eddington (who, asked how it felt to be one of three people who understood relativity, replied "I'm trying to think who the third person is") suggested stars as cosmic Crock-Pots. Hans Bethe (Greene's office neighbour, whose 4 p.m. Sneeze was reliable enough to set a watch by) and Fred Hoyle made it a predictive theory. Hoyle also coined "big bang" on BBC radio in 1949; historian Helge Kragh suggests it was not meant as a sneer.
Life of a 20-solar-mass star
First 8 million years: hydrogen fuses to helium.
Next 1 million years: helium to carbon and oxygen.
~1,000 years: carbon to sodium and neon.
~6 months: to magnesium.
~1 month: to sulfur and silicon.
~10 days: the rest burns to iron, and fusion stops.
Iron is the ash in the fireplace. Its 26 protons and 30 neutrons are the most tightly bound of any nucleus; adding more nucleons costs energy instead of releasing it. So heavier elements (copper, nickel, silver, gold, platinum, uranium, plutonium) need two violent sources:
| Source | Mechanism | Products |
|---|---|---|
| Big bang (first ~10 min) | Fusion in the cooling fireball | H, He, traces of D, He-3, Li |
| Stellar cores | Gravity-catalysed fusion over millions of years | C, O, Ne, Mg, Si, S up to Fe |
| Core-collapse supernova | Iron core collapses, rebounds; shock wave crushes nuclei on its way out | Heavier elements, blasted into space |
| Neutron-star collision | Ejected neutron plumes clump (no charge, no repulsion), then neutrons turn into protons | Heavy elements, possibly most of the universe's supply |
Neutron stars are remnants of stars of roughly 10 to 30 solar masses. In 2017 gravitational waves from a neutron-star collision were detected, turning theory into observation (following the first black-hole merger detection in 2015).
The solar system and young earth
The sun, at just over 4.5 billion years, is a newcomer and probably third generation. The first stars ignited about 100 million years after the bang, were hundreds to thousands of solar masses, burned out fast, and either collapsed into black holes or exploded, seeding space with heavy elements and triggering new star formation with their shock waves. The sun's birthplace is uncertain; the star cluster Messier 67 (~3,000 light-years away) has matching chemistry, but trajectory studies split on whether the sun could have escaped it with its disk intact (one 2012 study says "The Sun was not born in M67"; a 2019 study says it could have been).
More confident: about 4.7 billion years ago a supernova shock compressed a cloud; it contracted over a few hundred thousand years, spun up like a skater pulling in her arms, and flattened into a disk around a core. Over 50 to 100 million years the gravitational two-step heated the core to fusion. The sun was born. The leftover few tenths of a percent of the disk snowballed into planets: volatile hydrogen, helium, methane, ammonia and water collected in the cool outer zone (Jupiter, Saturn, Uranus, Neptune); iron, nickel and aluminium formed the rocky inner planets (Mercury, Venus, Earth, Mars). Planets support themselves by atomic resistance to compression, so their cores never reach fusion: hence their temperate surfaces.
Hadean or Poseidean? Reading the zircons
- Evidence. No rocks survive from earth's first half billion years, but zircon crystals do: near-indestructible specks that trap samples of their environment and can be radioactively dated.
- Find. Zircons from Jack Hills, Western Australia, date to 4.4 billion years ago, only ~200 million years after earth formed, and suggest a "cool early earth": a calm water world with small landmasses, not a hellscape.
- Drama anyway. At 50–100 million years old, earth was probably hit by the Mars-sized Theia, vaporising the crust and forming the moon; the collision likely explains earth's tilt, and hence the seasons. Heavy meteor bombardment may have boiled off the oceans more than once.
- Why it matters. How quickly water stabilised constrains when life could start. Not every place with water has life, but no known life exists without it.
Quantum chemistry and water
By the mid-1920s Newtonian physics was failing for electrons. After a gruelling night with Niels Bohr, Heisenberg wondered in a Copenhagen park whether nature could "possibly be as absurd" as the experiments suggested. In 1926 Max Born answered yes: particles are described by probabilities, computed (as it turned out) from Schrödinger's equation, which was news even to Schrödinger. Precisely, the probability is the squared magnitude of the wave function, a correction Born added in a footnote. The solar-system atom gave way to electron probability clouds.
Greene's "quantum theater": the nucleus is the stage, electrons the audience in tiered seating. Higher tiers cost more energy; in the calmest state electrons fill lower tiers first. The "fire code" is 2 seats on tier one, 8 on tier two, 18 on tier three. (Behind it is Pauli's exclusion principle: no two electrons in the same quantum state; sub-tiers complicate the simple ordering.) And atoms have a kind of "atomic OCD": they hate partially filled tiers. Helium (2) and neon (2+8) are content; everything else barters:
- Donating and receiving. One atom gives electrons, the other takes them; opposite charges then attract (ionic bonding).
- Sharing. Each puts electrons into a communal pool (covalent bonding).
- That is chemistry. These deals are what chemical reactions are, everywhere in the cosmos.
Water is the key case. Oxygen has 2 + 6 electrons and wants two more; each hydrogen shares a pair with it. H2O forms a wide V, and because oxygen hoards the shared electrons, the vertex is negative and the hydrogen tips positive. That lopsided charge makes water a near-universal solvent: its "charged claws" pull apart table salt (oxygen grabs sodium, hydrogens grab chlorine) and most other substances. Water is about 70% of a cell's mass, life's ferrying fluid. Albert Szent-Györgyi: water is "life's matter and matrix, mother and medium"; life left the ocean only when it grew a skin to carry the water inside.
The unity of life: one code
Species estimates range from millions to trillions, yet a microscope can barely tell mouse cells from mastiff cells, or housefly from human. The reason, established in recent decades, is descent from a common single-celled ancestor. Two universal systems make the case.
Information. Rocks move when pushed; rabbits move because of internal information processing: biological software driving biological hardware. Inside cells, proteins do most of the work, built from 20 amino acids like words from letters. Random assembly is hopeless: a modest 150-amino-acid protein has about 20150 ≈ 10195 possible sequences, vastly more than the particles in the observable universe. So proteins need instructions, and those are in DNA.
The universal genetic code
- Structure. A twisted ladder whose rungs pair A–T and G–C, so each rail determines the other.
- Human genome. About 3 billion letters; you differ from Einstein, Curie or Shakespeare by less than a quarter of a percent (about one letter in 500), and from a chimpanzee by about 1% in shared genes (more if whole genomes are compared).
- Codons. Three letters = one amino acid: CTA = leucine, GCT = alanine, GTT = valine. So CTAGCTGTT means leucine–alanine–valine.
- Redundancy. 4 letters in triplets give 64 codons for 20 amino acids, plus start and stop signals ("a capital letter and a period").
- Genes. A 1,000-amino-acid protein needs a 3,000-letter gene, plus regulatory sequences.
- Universality. "Whether from seaweed or Sophocles," every DNA molecule encodes proteins the same way (strictly, "nearly" universal, with minor variants). Cracked in the 1960s by Nirenberg, Holley and Khorana (Nobel 1968), following Crick and Brenner's 1961 work.
The unity of life: one power plant
The second universal system is less famous and, to a physicist, more astonishing. Energy comes in many currencies (nuclear, electromagnetic, gravitational), yet all life on earth uses exactly one mechanism: electrons cascading down a staircase of redox reactions, starting with food (animals) or water split by sunlight on chlorophyll (plants), and ending in the arms of oxygen.
- Redox. A burning log: carbon and hydrogen lose electrons (oxidised) to oxygen, which gains them (reduced). Cells do the same but in small steps, avoiding "a cellular fire." Like a ball bouncing down stadium stairs, each step releases a bit of energy. Szent-Györgyi: "Life is nothing but an electron looking for a place to rest."
- The proton battery. Each electron hop makes the receiving molecule twitch like a ratcheting gear; the twitches pump protons across a membrane a few dozen atoms thick. The field is tens of millions of volts per metre.
- The turbines. Protons flowing back spin nanoscale turbines (ATP synthase) that jam ADP + phosphate into ATP, a molecular compressed spring.
- The delivery trucks. ATP travels anywhere in the cell and releases energy by snapping a bond, doing for power what UPS does for packages; the fragments are recycled.
The cell's power budget
- ~10 million ATP per cell per second.
- Tens of trillions of cells in a body, so ~1020 ATP per second, every second.
- ~500 million trillion ATP made while you read one sentence.
- Insight credit: Peter Mitchell, who proposed proton-driven ATP synthesis (chemiosmosis) in 1961 and won the 1978 Nobel, did it at Glynn Research, his own independent lab, having had enough of academia. (Fermentation, which skips oxygen, is the one partial exception to the universal scheme.)
Both unities point to a common ancestor living around four billion years ago. Greene's notes refine the terms: plants and animals converge on LECA (last eukaryotic common ancestor), and with bacteria and archaea, further back, on LUCA (last universal common ancestor). Like a physicist finding one force behind many, he finds it spectacular that his two dogs, the insects at his lamp, the frogs in the pond and the coyotes in the distance run on identical molecular machinery.
Since 2020. A 2024 phylogenetic study (Moody et al., Nature Ecology & Evolution) dated LUCA to about 4.2 billion years ago and reconstructed it as already sophisticated (around 2,600 proteins, an anaerobic metabolism, even a rudimentary anti-viral immune system), which squeezes the window for life's origin and supports Greene's "cool early earth" reading. In 2025, samples returned from asteroid Bennu by OSIRIS-REx contained all five nucleobases of DNA and RNA and 14 of the 20 protein amino acids, plus evaporated brines, strengthening the case that life's raw ingredients were delivered widely.
Evolution before evolution
The origin of life is a three-part question: how did genetics (storing and copying information), metabolism (extracting and using chemical energy) and compartments (cells) arise? Whatever the answer, Darwinian evolution will be part of it.
Darwin's logic is a brain-teaser solution: descent with modification plus competition for finite resources. Its acceptance rests not on logic but on overwhelming data. Darwin had no mechanism for heredity; 1953 supplied it, and Watson and Crick's famous understatement ("It has not escaped our notice...") pointed to it. The DNA ladder unzips, each rail templates a copy, and daughter cells each receive one.
Variation comes from copying errors, by chance or from UV and X-rays. Most are harmless typos ("page 413 of War and Peace"); a few help or harm. Stability is essential: the raw error rate is about 1 in 100 million base pairs (a scribe miscopying one letter per 30 Bibles), and proofreading fixes 99% of those, leaving about 1 in 10 billion.
Deep time dissolves the design intuition. If each year were a sheet of printer paper, a billion years would be a stack nearly 100 km high, a flip-book more than ten times the height of Everest. Small changes per page add up to the difference between an amoeba and a chimp. And nature ran countless flip-books in parallel, many interacting.
Evolution is innovation by trial and error, a strategy that "would bankrupt most businesses" but that nature can afford, because it has unlimited time and no bottom line. (Greene's note: computer scientists do use selection-style search when iteration is cheap or massively parallel.) Debates over tempo (gradualism vs. Punctuated equilibrium), sexual selection and gaps in the record are real but none casts doubt on evolution itself.
Molecular Darwinism. Replication, variation and competition are not unique to life. Imagine a molecule that templates copies of itself in a chemical stew, imperfectly. Limited raw materials mean the fastest, most accurate replicators win, and mutations keep offering improvements. Purely physical jostling thus becomes ever better at reproduction.
The RNA World and other origin scenarios
In the 1960s Francis Crick, Leslie Orgel and Carl Woese pointed to RNA: a single-railed cousin of DNA that carries genetic messages and also does chemistry (the ribosome, life's protein factory, is built around RNA). RNA is both software and hardware, and some RNA can catalyse its own copying, which dissolves the chicken-and-egg problem (you need instructions to build machinery and machinery to read instructions).
Supporting chemistry: in the 1950s Stanley Miller and Harold Urey sparked a mix of hydrogen, ammonia, methane and water vapour and got amino acids in a brown sludge. Their gas mix was later judged wrong for early earth, but better mixes work too, including a volcanic-gas mixture Miller made that sat unanalysed for more than half a century. Amino acids also turn up in interstellar clouds, comets and meteorites. Fatty acids spontaneously form double-layered membranes, giving natural cell walls.
| Proposal | Champion(s) | Core idea |
|---|---|---|
| RNA World | Crick, Orgel, Woese | Self-copying, catalytic RNA evolves first; proteins and DNA later |
| Metabolism first | various | Self-replicating catalytic proteins or reaction networks come before genes |
| Dual origin | various | Replicators and catalysts arise separately, then merge into cells |
| Wet–dry cycling | David Deamer | Pond edges cycling wet and dry drive lipids into membranes and link snippets into RNA-like chains |
| Clay crystals | Graham Cairns-Smith | Growing mineral lattices are the first replicators, a scaffold for organics |
| Alkaline hydrothermal vents | Mike Russell, Bill Martin | Porous limestone chimneys (some over 50 m, taller than the Statue of Liberty) channel energy-rich fluids through natural reaction chambers |
| Panspermia | Arrhenius, Hoyle, Wickramasinghe, Davies | Seeds of life arrive on rocks; but this only moves the question |
Greene's verdict: lab attempts are intriguing but inconclusive; we have not made life from scratch, though he expects we will. The narrative is clear even if details are not: once molecules replicate, molecular Darwinism takes over and, over hundreds of millions of years, can build life's chemical architecture.
RNA World progress. In 2024 Gerald Joyce's lab at the Salk Institute (PNAS) evolved an RNA polymerase ribozyme that copied a functional RNA (a hammerhead ribozyme) accurately enough for Darwinian evolution to occur in the test tube, with new variants emerging and competing. The ribozyme cannot yet copy itself in full, but this is the closest experimental approach yet to Greene's molecular Darwinism.
Information is physical
How does DNA "know" to unzip, or RNA "know" to copy a gene? It doesn't. Molecules obey blind laws. The difference from a rock is that the rabbit's particles are guided by information embodied in molecular shapes. A water slide does not override gravity; its shape guides riders along a path they would otherwise not take. Molecular shapes that attract this amino acid and repel that one, or snap together like matched Lego, do the same through the electromagnetic force.
Information in a cell is not abstract instructions that molecules read; it is encoded in arrangements that make other molecules move in particular ways. For a rock, the physics story is the whole story. For a rabbit, an information-based story rides on top, organising coordinated motion across trillions of molecules, as when you reach for a coffee mug. Life does not and cannot contravene physical law. Life is physics orchestrated.
Life and the second law
Schrödinger said organisms survive by "feeding upon negative entropy," a phrase that has drawn "persnickety criticism," but the idea is the entropic two-step. Greene, for more than half a century, has kept his entropy in check by taking in ordered structures (mostly vegetables, nuts and grains), burning them by redox, and exporting entropy as heat and waste.
Follow the low-entropy supply down the food chain and you reach sunlight:
Earth's entropy budget (Greene's note)
- In: visible photons from a ~6,000 K solar surface: few, concentrated, low entropy.
- Out: infrared photons from a ~285 K earth: many, dilute, high entropy.
- Ratio: energy per photon scales with temperature, so 6000/285 ≈ 21: earth emits roughly two dozen photons for every one it receives.
- Lesson: solar power is valuable not for its quantity (earth radiates away as much energy as it gets) but for its quality.
Trace the sun back and you return to chapter 3: gravity makes the star; fusion keeps the low-entropy photons coming for billions of years. The nuclear force, in tandem with gravity, is a fount of life-giving low-entropy fuel.
A general theory of life?
Schrödinger urged scientists to roam outside their specialities. Researchers in statistical mechanics, information theory, quantum chemistry and astrobiology have responded, aiming at a deep question: is life a fluke that happened once among hundreds of billions of galaxies, or the expected outcome of common conditions? The best tool may be thermodynamics, which Einstein said "will never be overthrown."
Bénard cells: order out of chaos
- Setup. Heat a shallow dish of viscous oil from below.
- Outcome. As the energy flow increases, random motion gives way to hexagonal convection cells: liquid rises in each cell's centre, crosses the top and sinks at the edges.
- Theory. Ilya Prigogine (Nobel 1977) showed that with a steady energy supply, certain fluctuating patterns are especially good at absorbing energy, so they persist and grow, dumping degraded energy outside. He called them dissipative structures: "order out of chaos."
- Link. Bénard cells aren't alive, but living things are dissipative structures.
Jeremy England, building on Christopher Jarzynski and Gavin Crooks's fluctuation theorems, applied the second law to driven systems. Greene's image is a playground swing that adjusts its own length until its rhythm matches your pumping: it gets "in the groove," absorbs your energy efficiently, holds steady and dissipates waste. (His daughter prefers to leap from the high point and dissipate energy by tumbling.) England's analysis says driven molecular collections can likewise rearrange into configurations that absorb and dissipate energy more efficiently: dissipative adaptation.
From dissipative adaptation to life (England's chain)
- Drive: a steady external energy source (sunlight, vent chemistry).
- Adaptation: particles drift into arrangements that resonate with the drive, absorb energy and dissipate it.
- Replication as amplification: if one such collection dissipates well, two, four, eight are better, so replicators may be an expected output.
- Molecular Darwinism: once replicators exist, selection for fitness begins.
- Information: a heritable "instruction manual" preserving battle-tested strategies gives a decisive edge.
- Life: over hundreds of millions of years, cells emerge. - Caveats (notes): much of life's waste may come from building durable structure, not just maintaining it; and energy must arrive gently (a wineglass driven too hard shatters), so living systems balance resonance and protection.
How strong is the thermodynamic case? Greene calls these ideas "in their early stages," and that remains true. Dissipative adaptation has been shown in simulations and simple physical systems, but no experiment has yet shown it producing self-replicating chemistry, and critics note that "efficient dissipation" and "fitness" can come apart. It is a promising lens, not an established mechanism.
If the physics proves general, life may be common. But life is one thing and intelligent life another: microbes on Mars or Europa would be monumental, yet as thinking, conversing beings we would still be alone. That sets the next question: how do we get from life to consciousness?
The search since 2020. NASA's Perseverance rover found the "leopard-spot" textures in the Cheyava Falls rock ("Sapphire Canyon" sample, published 2025) that NASA called the clearest potential biosignature yet found on Mars, pending return of the sample to earth. Claims of dimethyl sulfide on exoplanet K2-18b (2025) remain disputed. None of this yet answers whether life is common, but the question has moved from speculation toward measurement.
Universality is evidence of a single origin, and of lock-in. A universal code and a universal power system persisted for four billion years because every later innovation had to stay compatible with them. Enterprise architects see the same pattern: an early interface standard (a core banking ledger, a message format) becomes the "genetic code" of an estate, and its universality signals both a common origin and near-impossible replacement. Design the first shared protocols as if they will last forever, because they may.
Particles and Consciousness
Greene takes on the hard problem: how do mindless particles give rise to an inner world of colour, sound and feeling? He sets out the leading positions (Chalmers's panpsychism, Tononi's integrated information, Graziano's attention schema), rejects any special role for consciousness in quantum mechanics, and bets on physicalism: consciousness will be "demystified without being diminished." He then applies the same physics to free will. Deterministic or probabilistic, the laws leave no room for a mind that overrides them, but they leave plenty of room for a different, real freedom: a particle arrangement so richly organised that its range of responses is vast, unlike a rock's.
The third frontier
Somewhere between the first prokaryotes four billion years ago and a brain of about 90 billion neurons and 100 trillion synapses, matter learned to think and feel, love and hate, imagine and create. Camus: "Everything begins with consciousness and nothing is worth anything except through it." Yet for decades consciousness was a fringe topic in the hard sciences, fit only for "doddering researchers in the twilight of their careers," because science studies objective reality and the inner voice is audible only inside your head.
Greene calls that stance ironic. Descartes's cogito is our one certain contact with reality (Ambrose Bierce's parody, "I think that I think, therefore I think that I am," notwithstanding). And consciousness has long been humanity's main hope of outlasting death: atman, anima, immortal soul, all names for something that might transcend physical form. That is the deeper reason science kept its distance: talk of domains beyond physical law gets "an exasperated grimace" and a swift return to the lab. But the dismissal also marks a real gap: science has no robust account of subjective experience.
Three trails beyond Newton
- Smaller: quantum physics, explaining particles and life's biochemistry.
- Larger: general relativity, explaining gravity, stars and planets.
- More complex: the frontier still open, explaining how vast particle collections yield life and mind. Newton's genius was judicious simplification (planets as solid balls); "model your head as a solid ball" and you learn little about minds.
What has changed is data. fMRI tracks blood flow supporting neural activity, deep-brain probes record single neurons, EEG reads electrical waves across the brain. When such measurements mirror behaviour and reports of inner experience, the case for treating consciousness as a physical phenomenon strengthens substantially.
"Nothing but a bag of particles"
Greene once told a late-night TV host, in a heated discussion about mathematics and the universe, that he was "nothing but a bag of particles governed by the laws of physics." ("Hey, that's a great pickup line.") It was a statement of reductionist commitment, not a jibe, and it applied equally to Greene. He echoes Democritus: sweet, bitter, hot, cold and colour are conventions; "in truth there are only atoms and the void."
But consciousness is a different kind of story. Entropy, evolution and life can be told in the third person; consciousness demands a first-person account of sensations, plus a palpable sense that I control my actions: "Laws of physics be damned; I think, therefore I control." That creates two questions: can matter alone produce sensations, and can our sense of autonomy be just physics playing out? Descartes said no to both: there is physical stuff and mind stuff, and they interact but differ. Greene expects instead that consciousness will be explained by ordinary particles and ordinary laws, a revolution of its own kind: physical law's hegemony extended into the inner world.
The mind in the shadows
Most brain activity never reaches awareness. Watching a sunset, your brain processes trillions of photons per second, fills in each eye's blind spot, compensates for eye and head movement, flips the image upright and fuses the two eyes' views, all invisibly. Thinkers from the Vedas (3,000 years ago) through Augustine, Aquinas, Shakespeare and Leibniz ("music is the hidden arithmetical exercise of a mind unconscious that it is calculating") sensed an unconscious; Freud popularised it.
Evidence for unconscious processing
- Otto Loewi's dream (Easter 1921). He scribbled a dreamt experiment at night, could not read it in the morning, dreamt it again, went straight to the lab, and showed nerve signalling is chemical: the work earned a Nobel Prize.
- Patient P.S. (Halligan and Marshall, late 1980s). After right-brain damage she could not perceive the left side of images and said two drawings of a house were identical, though the left side of one was on fire. Asked which she would prefer to live in, she consistently chose the house not burning.
- Subliminal priming (Naccache and Dehaene). Images flashed for under ~40 ms between masks never reach awareness, yet they speed or slow judgements: deciding whether 4 is above or below 5 is faster after a hidden 3, slower after a hidden 7.
- Myth to discard. The 1950s "Drink Coke" cinema-sales story was invented by a struggling market researcher.
These are regulatory, functional and data-mining marvels, but not conceptual mysteries: mapping them is "merely" an enormous empirical task. The mystery is elsewhere.
The hard problem
Newton already saw it: explaining how light produces "the Phantasms of Colours" in our minds "is not so easie." Your phone can report that a banana is yellow; as far as we know, it has no inner feeling of yellow. You do. Greene, typing with popcorn and Pentatonix in the background, feels pressure on his fingertips, salt and music. Peer into his brain with superhuman vision and you would see about 1027 particles ("a thousand trillion trillion"), each obeying the same equations as in a kitchen or the corona of the North Star. Particles have mass, charge, nuclear charges and spin, and nothing in that list hints at experience. Typhoons and volcanoes are hard to compute but not mysterious, because there is nothing it is like to be a typhoon.
Thomas Nagel's question, "What is it like to be a bat?", makes the gap vivid: echolocating through the dark, locating a mosquito by its echo. A complete physical account of the bat seems unable to deliver its first-person experience. In 1994, at the Tucson consciousness conference, a long-haired young Australian philosopher, David Chalmers, named it the hard problem. The "easy" problems (memory, stimulus response, behaviour) are not easy, but we can picture what a solution looks like. For experience we cannot, and Chalmers argued a reductionist bridge will fail.
Mary's room
Greene admits his old answer ("consciousness is what certain information processing feels like") dismissed the problem too quickly. His deepest doubt came from Frank Jackson's knowledge argument (1982):
Mary the colour scientist
- Setup. Mary, colour-blind from birth, becomes the greatest neuroscientist ever and learns every physical fact about the brain: every firing and particle cascade involved in seeing a blue sky, tasting a plum or hearing Brahms's Third Symphony.
- Turn. She discovers a cure, has surgery and, bandages removed, sees red roses for the first time.
- Question. Does she learn anything new? Intuition says obviously yes: surprise, thrill, a new understanding of perception.
- Implication. If so, complete physical knowledge leaves something out. Greene felt "a sudden kinship with Mary"; his offhand physicalism was shaken.
| Mary learns something new | Mary learns nothing new |
|---|---|
| Early Jackson: physical facts are incomplete; experience is extra. | Daniel Dennett: we underestimate what complete physical knowledge means; she would predict red's look in advance. |
| Supports anti-physicalist views like Chalmers's. | David Lewis, Laurence Nemirow: she gains an ability (to recognise, recall, imagine red), not a new fact. |
| The "wow" signals a new truth. | Later Jackson: he changed his mind; deduction could get her there, just unfamiliarly. |
The stakes are high (physicalism or a new science), and no experiment or calculation has been proposed that settles it. Positions are judged by plausibility and intuition, which is why the field spans eliminativism (consciousness is an illusion) to idealism (only consciousness is real).
Two historical tales
Greene's framing device is two nineteenth-century analogies:
| Vitalism: a "hard" problem that turned easy | Electromagnetism: an "easy" problem that turned hard |
|---|---|
| Life seemed to need a nonphysical spark to animate inanimate matter. | Faraday's electric and magnetic phenomena seemed explainable with Newtonian fluids, cogs and wheels. |
| Outcome: no spark needed; atoms, molecules, cells and tissues suffice. | Outcome: new physical reality was needed: electric charge, space-filling fields, Maxwell's equations (twenty in the original form). |
| Physicalists expect consciousness to follow this path. | Chalmers expects consciousness to follow this path. |
Chalmers: consciousness on the ground floor
Chalmers takes the electromagnetic tale to heart. Perhaps particles carry an irreducible proto-consciousness alongside mass, charge and spin, and mindful particles pool their qualities into minds: a revival of panpsychism, with roots in ancient Greece, applying to "a bat's brain or a baseball bat." What is proto-consciousness? No one knows, but Greene notes he cannot say what mass or charge are either, only what they do; perhaps a theory of what proto-consciousness does will one day predict as accurately as our theories of gravity and electromagnetism. For now, it doesn't.
In information language (after John Wheeler, who popularised "black hole" and saw information as physics' basic currency), Chalmers proposes that information has two sides: the objective, third-person side physics studies, and a subjective, first-person side physics ignores. A complete theory would need laws for both. That would be a genuine unified theory, though Greene still resists "theory of everything" ("scientists would still have a hard time predicting what I'm going to have for breakfast tomorrow"). His verdict, with Sagan's rule that extraordinary claims need extraordinary evidence: our inner experience is extraordinary, but the evidence that it lies beyond conventional science is weak.
Tononi: consciousness as integrated information
Giulio Tononi, joined by Christof Koch, asks what distinguishes conscious information processing, since abacuses and thermostats process information too. The answer is integrated information theory (IIT):
IIT's two hallmarks of consciousness
- Integration. Seeing a red Ferrari, you do not see a colourless car then paint it red; shape, colour, smell and associations of power and wealth arrive as one experience, though handled by different brain areas.
- Differentiation. Each experience is one out of a vast repertoire of possible experiences.
- The camera contrast. A digital photo's pixels are independent (store each in its own file and nothing is lost): zero integration, and a fixed format: low differentiation. Hence no experience.
- The measure. Φ (phi) quantifies integration and differentiation; higher Φ, more consciousness, on a continuum that could exceed ours.
IIT is substrate-independent, so it too leans panpsychist: high Φ in neurons or neutron stars would be conscious. Scott Aaronson showed that simple logic-gate networks can be wired to have arbitrarily large Φ, so IIT must call them conscious; most find that absurd. Tononi bites the bullet: the network is conscious. Greene notes that our own consciousness, a three-pound wet grey knot, is already the "nearly ludicrous" claim; familiarity skews our sense of absurdity. His complaint is different: IIT states conditions for consciousness, not why they feel like anything, and Tononi suggests that is the wrong question. Greene's reductionist intuition "will remain unsatisfied."
IIT on trial since 2020. The Cogitate adversarial collaboration pitted IIT against global neuronal workspace theory with pre-registered predictions; results (announced 2023, published in Nature, April 2025) challenged key predictions of both: IIT's expected sustained posterior synchrony was not found, nor was the workspace theory's predicted "ignition" at stimulus offset. In June 2023 Koch paid off a 25-year bet to Chalmers that the neural correlates of consciousness would be found by then. In September 2023 an open letter from 124 researchers called IIT "pseudoscience," prompting a fierce backlash. The field is livelier and more empirical than in 2020, and no closer to consensus.
Graziano: the mind models the mind
Greene's favourite is Michael Graziano's physicalist theory (the attention schema theory). It starts from two uncontroversial facts:
- Perception is construction. "Red" is not in the light: physically, red light is an electric field oscillating ~400 trillion times per second at right angles to a magnetic field, travelling at 300 million metres per second. Red happens inside your head; so does "new car smell."
- Representations are schematic. You do not perceive 435,172,874,363,122 cycles per second at one spot on the door and 447,892,629,261,106 at the hood; you perceive "red." Brains lost in detail "would have been swiftly eaten."
We also build schematic models of other minds (theory of mind, the intentional stance): friend or foe, threat or opportunity. Graziano's key move is that we apply the same capacity to ourselves, building a simplified model of our own attention, which we call awareness.
Because the self-model omits the neural machinery that produces attention, awareness feels ethereal, unmoored, coming from nowhere. If your body schema omitted your arms, your moving hands would seem ghostly too. The hard problem seems hard only because our schematic self-models hide the brain mechanics that link thoughts and sensations to their physical basis.
The appeal is that nothing exotic is needed: no new properties of matter, just the detailed mapping of our own familiar terrain. Other physicalist models in this family include Dennett's multiple drafts and Baars's global workspace; Greene's notes also list Penrose and Hameroff's Orch OR, a quantum alternative. To those who think physicalism belittles the mind, Greene replies that a mind doing all it does with the same ingredients and forces that hold a coffee cup together is more extraordinary, not less.
Greene's notes add a personal datum. After surgery to remove a malignant tumour, his wife, Tracy, temporarily lost the ability to name common nouns; she could picture red shoes but not name them, as if one data bank had been cut out. Clinical cases like this tie mental functions tightly to physical tissue.
Consciousness and quantum mechanics
Quantum mechanics is the most accurate theory ever: never contradicted by replicable experiment, with quantum electrodynamics matching data to better than one part in a billion. Yet it has a core puzzle. Its predictions are probabilities (20% here, 35% there, 45% way over there), and unlike a coin that is already heads or tails under your hand, an unobserved electron is in a fuzzy mixture of all its possibilities. Nothing in the equations explains how one definite outcome appears; applied consistently to apparatus and brain, the fuzziness should spread to your thoughts, yet you always see one result. That is the measurement problem.
Fritz London and Edmond Bauer (1930s) and later Eugene Wigner proposed that consciousness causes collapse. Greene's assessment after decades in the field: "extraordinarily unlikely." Experiments show that interactions, not minds, produce definite outcomes; a mindless probe does the job. There is no airtight disproof, since every check requires a conscious observer, but proposed links are tenuous. Barring a stunning surprise, no textbook will carry special rules for when a mind is present.
Approaches to the measurement problem (Greene's notes)
- Copenhagen-style collapse: measurement makes the wave "snap" to one outcome; the mechanism is unexplained.
- Bohmian (de Broglie–Bohm) mechanics: particles always have definite trajectories guided by the wave; fuzziness is ignorance of initial conditions. A "dark horse" Greene won't write off.
- Many Worlds: every outcome happens, in branching worlds.
- GRW collapse: a new physical process randomly collapses each particle's wave very rarely; in large objects, collapse is almost instant.
- Consciousness-causes-collapse: London, Bauer, Wigner. Greene: extraordinarily unlikely.
Free will
We take no pride in the pancreas making chymotrypsin; what we identify with is our thoughts, feelings and choices, and with them the belief that we are the ultimate source of our actions.
~400 BCE Democritus: atoms and the void replace divine whim with immutable law.
~300 BCE Epicurus: law, "immune to all flattery," smothers freedom, so he posits the atomic swerve, random deviations from determinism.
Later centuries Augustine, Aquinas, Hobbes, Leibniz, Hume, Kant, Locke and many modern philosophers wrestle with the same knot.
The modern argument is short. You are particles; particle motion is fully governed by mathematical law; the chain of causes runs back to the big bang. "We are no more than playthings knocked to and fro by the dispassionate rules of the cosmos." Greene examines three escape routes and rejects each:
| Escape route | The idea | Greene's rebuttal |
|---|---|---|
| Anti-reductionism | Large assemblies (~1029 particles in a body) escape micro-laws | No calculation, experiment or observation has ever hinted that the laws lose control; emergence (typhoons to tigers) is always lawful |
| Quantum indeterminism | Probabilistic laws are a modern Epicurean swerve | Schrödinger's equation is as deterministic as Newton's; it fixes a unique table of probabilities |
| Measurement-problem gap | Free will hides in how one outcome is chosen | Outcomes are random and obey statistics; neither randomness nor statistical regularity is a willed choice |
Deterministic or probabilistic, the laws keep churning without pausing for human input, before and after humans existed. "We can no more intercede in this lawful progression of particles than we can change the value of pi." Our choices feel free because our senses never show us the laws operating at the particle level.
(Greene's note distinguishes his view from Aristotle's notion of the "voluntary," an act originating in the agent's deliberation: Aristotle never counted physical law as an external force that could make an act involuntary.)
Rocks, humans and freedom
You and a rock sit on a park bench. A heavy branch snaps above Greene as he walks by; you leap and tackle him to safety; the rock sits there. Neither has free will in the traditional sense, so what is the difference? Your particles are so exquisitely organised that they can respond with waves, greetings, jumping jacks, string-theory solutions or rescues; incoming photons, sound and wind trigger particle cascades that become sensations, thoughts and actions. The rock's particles merely jitter and warm slightly.
Human freedom is not freedom from physical law but freedom to exhibit an enormous range of behaviours (leaping, thinking, imagining, deliberating, explaining) unavailable to most particle collections. We are free not because we can supersede physical law, but because our prodigious internal organisation has emancipated our responses.
This is a compatibilist position, closest to Dennett's (Elbow Room, Freedom Evolves). Greene admits it can sound like bait and switch ("you are no different from a rock... But cheer up!"), yet he finds real comfort in it: "my particles just composed this very sentence and I'm pleased they did." He credits his thinking to Luise Vosgerchian, a Harvard music professor who asked him decades ago to write about freedom and creativity from the standpoint of physics.
Learning, relevance, responsibility
- Learning and creativity don't need free will. A Roomba updates its internal map as it bumps into furniture and invents routes around table legs: rudimentary learning and creativity. Greene's note adds AlphaZero (Silver et al., 2018), which went from beginner to beating world champions at chess, shogi and Go in hours of self-play, all particles obeying law. Your "software" is the same in kind, richer in degree.
- Nested stories again. The particle story of a Roomba and the "it sensed the stairs, remembered them, turned away" story are fully compatible; the higher-level one needs concepts like learning. At dinner with his wife, Greene wants her ideas and plans, not the motion of her 1029 particles.
- Actions matter. The fourth-grade match caused the oven explosion. Physics augments the human story; it does not drain it of relevance. Thoughts and actions are real links in the causal chain; what is unexpected is that they too emerge from prior causes through physical law.
- Responsibility is real. "I" is shorthand for my particle configuration, so my particles' behaviour is my behaviour, uniquely shaped by my genes, proteins, neurons and synapses. Personal identity over time is subtler (Greene favours Nozick's "closest continuer"; a duplicating transporter breaks the intuition).
- Punishment (notes). Without free will, punishment should be justified by protecting society and deterring future harm, and because people learn, consequences shape future behaviour.
Amsterdam
Intellectually convinced, Greene still feels that he freely controls his mind, but he notes how fragile that feeling is. Decades ago in Amsterdam, a mind-altering substance gave him "one of the most terrifying nights of my life": endless copies of himself, each exposing the previous one's "true" reality as fake, wiping out everything that self cared about, over and over. Physically, a few foreign particles had entered his brain; the human-level template of a stable self with free will collapsed while the particle-level template held. The experience aligned his rational grasp of the physics with his intuition.
The resolution: keep the everyday language (we do make choices; decisions do have consequences) but drop the idea that choices originate beyond physical law. The sensation of free will is real; the capacity to transcend the laws is not. Somewhere around 100,000 years ago or earlier, the archaeological record suggests our ancestors had these experiences. "Early humans had long since stood up. Now we could look around and wonder."
What the chapter leaves out. Greene's free-will verdict is philosophically mainstream among physicists but contested among philosophers: libertarians (in the philosophical sense) argue that agent causation or emergent causal powers are coherent, and some compatibilists think Greene's "freedom as behavioural range" concedes too much (a thermostat has some range). He also skips the experimental literature: the Libet readiness-potential experiments, often cited against free will, have been substantially reinterpreted since 2012 (Schurger's work suggests the potential reflects noise accumulation, not a hidden decision). And the chapter's strongest move, Graziano's theory, explains why we believe we have inner experience; critics argue that leaves the experience itself untouched, the hard problem relocated rather than dissolved.
AI consciousness is now an operational question. Greene's chapter predates large language models, but its framework applies directly. A 2023 report by Butlin, Long and colleagues derived "indicator properties" from theories like the global workspace and attention schema and concluded no current AI system is a strong candidate, though nothing in principle rules it out. For banks deploying copilots like relationship-manager assistants, Greene's two lessons are practical: fluent self-report is not evidence of experience (Graziano: a system can model itself as "aware" without being so), and learning and creativity do not require will, so accountability for an AI's outputs sits with the configuration's designers and operators, exactly where Greene places responsibility for human action.
Language and Story
Language turned private minds into a shared, open-ended system: finite sounds recombined into unlimited meanings, including things absent, past, future and imaginary. Greene surveys the competing accounts of when and why we began to speak, then asks why a species fighting for survival spends so much time on stories that never happened. His answer: story is the mind's flight simulator, a safe rehearsal for social and physical trouble and our most powerful means of inhabiting other minds. Myths push story to its limit: they cluster around death and destruction, star minimally counterintuitive beings, and exploit our tendency to see agents everywhere, which sets up religion in chapter 7.
Pattern, mathematics and words
We survive because patterns persist: the sun rises, rocks fall, water flows. Mathematics is the articulation of pattern. Galileo said the book of nature is written in mathematics. Is mathematics the source of reality or a language we invented by overindulging our taste for pattern? Greene's romantic side favours the first; his sober side allows the second, noting that few ancestors won a meal or a mate by contemplating primes or squaring the circle.
Einstein said he "very rarely" thought in words, more often "in music." Greene is the opposite: he works by fiddling with equations and writing conclusions in ordinary sentences in notebooks filling shelf after shelf, often talking to himself. He thinks Wittgenstein's "the limits of my language mean the limits of my world" overreaches, since some experience lies outside language, but words don't just express reasoning, they vitalise it. Toni Morrison: we die, but "we do language," and that may be the measure of our lives.
What language makes possible (Greene's list)
- Imagination shared: conjure images, real or fanciful, in other minds near and far.
- Instruction over discovery: pass on hard-won knowledge cheaply.
- Coordination: share plans, align intentions, act together.
- Collective creativity: pool individual inventiveness into a communal force.
- Self-transcendence: see that though shaped by evolution we can soar beyond survival needs.
- Story: a collective overlay that makes sense of experience, as in the closing lines of E. B. White's Charlotte's Web.
The fossil problem
No one knows when or why we began to speak. Darwin guessed language grew from song, with "Elvis-like" crooners winning more mates. Alfred Russel Wallace, co-discoverer of natural selection, thought selection could not explain music, art or language and invoked "a Higher Intelligence" (Quarterly Review). Darwin scrawled an emphatic "no" in the margin and wrote: "I hope you have not murdered too completely your own & my child."
We know more about the first minutes of the universe than the first words of humans, because the universe left fossils (CMB, element abundances, receding galaxies) and sound waves vanish in a moment. Absent relics, theories proliferate like "tag-team wrestling," each challenger meeting a fresh opponent.
What everyone agrees on is that human language is unique. A vervet monkey has distinct alarm calls for leopard (short high whine), eagle (low snorts) and python ("chutters"), but cannot discuss yesterday's python or tomorrow's raid on a nest. Animal communication is closed, fixed-meaning and here-and-now. Bertrand Russell: a dog "cannot tell you that his parents were honest but poor." Human language is open: a finite set of phonemes recombined into hierarchical, virtually unlimited sequences about anything, real or imagined.
Innate or learned? Sudden or gradual?
| Position | Proponents | Claim | Key point or problem |
|---|---|---|---|
| Universal grammar | Noam Chomsky (lineage back to Roger Bacon, 13th c.) | An innate, species-wide language primer; children could not otherwise extract precise grammar from fragmentary input | Any child learns any language, so the primer must target a universal core |
| Sudden origin | Chomsky | A "slight rewiring of the brain" ~80,000 years ago sparked a cognitive big bang | One mutation explaining so much is hard to test |
| Gradual adaptation | Steven Pinker, Paul Bloom | Incremental Darwinian buildup, each step useful: "wild boar at eleven o'clock," "watch out for Barney, he's got his eye on Wilma" | The standard selection story |
| By-product suite | various | Breath control, memory, symbolic thought, theory of mind and group living combined to yield language, though selected for other reasons | Language as a lucky composite |
| Thought first | Chomsky | Language began as an inner tool for thinking; speech came later, like adding speakers to early PCs | Critics: phonology and grammar look built for external communication |
Dating is equally uncertain. Hafted tools, cave art, geometric engravings and beads show planning and symbolic thought by at least 100,000 years ago. Skull and throat anatomy suggest the physical capacity to speak over a million years ago. Genetics adds FOXP2: in 2001 a single-letter change in this chromosome-7 gene was found in a British family with a three-generation disorder of grammar and articulation. Press dubbed it "the grammar gene," to experts' irritation, but it is one essential component. Chimp FOXP2 protein differs from ours by 2 amino acids out of 700+; Neanderthal FOXP2 is identical. That puts a genetic basis for speech between the chimp split (a few million years ago) and the Neanderthal split (~600,000 years ago). Overall estimates for the first words range from tens of thousands to a few million years, and capacity is not the same as use.
FOXP2 and the machine test. A 2018 reanalysis (Atkinson et al., Cell) found no evidence of a recent selective sweep on FOXP2 in humans, weakening the once-popular story that this gene's spread marked the dawn of speech. More dramatically, large language models (from 2022 on) learn fluent grammar purely from statistical exposure to text. Some linguists (e.g. Steven Piantadosi, 2023) argue this undercuts Chomsky's poverty-of-the-stimulus argument; Chomsky replied (New York Times, March 2023) that such systems are "a false promise," describing rather than explaining language. A caveat for both sides: LLMs train on trillions of words, thousands of times what a child hears, so the poverty-of-stimulus question is changed, not settled.
Why we spoke
Linguist Guy Deutscher lists candidate origins: shouts, gestures, imitation, deception, grooming, singing, dancing, even "chewing, sucking and licking," a list reflecting theorists' creativity more than history. Greene examines several:
- Motherese. Before baby slings, mothers set infants down to work with both hands; babies who babbled drew attention back, mothers cooed and hummed in reply, and infant survival selected for vocalisation.
- Gesture first. Primates gesture well; lab chimps learn hundreds of signs. As hands filled with tools and groups grew large, or the light failed, voice beat gesture. Greene, whose hands move whenever he talks, finds this especially plausible.
- Gossip as grooming (Robin Dunbar). Chimps cement alliances by grooming; as human groups grew, one-to-one grooming became too costly. Talk replaced nit-picking, letting people trade social intel in minutes. Up to 60% of conversation today is gossip.
- Guided imagination (Daniel Dor). Before language, social exchange was limited to shared experience. Language is a communally built tool for guiding one another's imaginations, letting you plant experiences in my mind I never had. As survival came to rest on cooperative hunting, fire, cooking and shared childcare, nonverbal exchange hit its limits.
Whatever the origin, once the human mind had language, its engagement with reality was poised for radical change, and the vehicle of that change was story.
The oldest story
George Smith and the flood tablet
- Who. A poor London boy who left school at fourteen to apprentice as a bank-note engraver, taught himself Assyrian cuneiform, and spent lunch breaks at the British Museum sorting clay fragments from Nineveh. The curators hired him.
- What. Reassembling a tablet, he found a flood story older than Noah's, but a crust obscured a key section and the museum's restorer, Robert Ready, was away for days.
- Greene's confession. His first telling has Smith, frantic, scrubbing the tablet himself. In fact Smith waited. Greene uses the embellishment to show what storytellers always do: rework reality for drama or joy.
- Outcome. Ready cleaned the tablet and revealed part of the Epic of Gilgamesh, composed perhaps in the third millennium BCE, likely shaped by many voices over generations. (Smith announced it in 1872.)
Five thousand years of change in food, shelter, medicine and communication, yet we recognise ourselves at once. Gilgamesh and his companion Enkidu set out on a quest that tests courage and identity ("a Neolithic Thelma and Louise"). When Enkidu dies, Gilgamesh veils his friend's face like a bride's, paces like a lioness robbed of her cubs, tears out his hair. Greene has known that place: decades ago he charged from room to room of his tiny walk-up apartment after learning his father had suddenly died. If we wrote stories 5,000 years ago, we told them far earlier. Why spend scarce time on imaginary escapades instead of more boar and roots?
Does story earn its keep?
"Because we like stories" begins an answer, as "because we like ice cream" does. The sweet tooth is the model explanation: ancestors who gorged on fruit and nuts survived lean times, so a preference for sugar and fat spread; pistachio Häagen-Dazs is a relic of calorie scavenging. Genes don't determine behaviour, but they shape predilections. "We can learn new tricks but, genetically and hence instinctually speaking, we are old dogs."
The evidence problem. Natural selection acted on ancestral behaviour, yet recorded history covers only the last quarter of 1% of the ~2 million years since the first human migrations out of Africa. Researchers rely on artefacts, extrapolation from today's hunter-gatherers, and cognitive echoes in brain architecture. This leaves room for "just so" stories. Greene's rule for a good adaptive explanation: start from a specific adaptive challenge and show the behaviour is well designed to meet it, as the sweet tooth is.
The by-product view. Stephen Jay Gould and Richard Lewontin warned that evolution sells package deals (their "spandrels"). Big social brains crave intel (who's up, who's down, who's trustworthy) and gain status by passing it on; fiction is full of such intel, so we may love it as a side effect. Greene is not persuaded: a behaviour this pervasive, central and costly would likely have been shed if adaptively useless.
The rehearsal view. Like play in many species, story is practice for the real world.
| Pinker: the chess book | Oatley: the flight simulator |
|---|---|
| "Life is like chess, and plots are like those books of famous chess games" studied for future crises | Stories are simulated worlds where we shadow characters whose experience far exceeds ours |
| We build a mental catalogue of strategies to look up | We don't look things up; intuition itself is trained, becoming sharper and more flexible |
| Critique (Jonathan Gottschall): copying fiction makes you Don Quixote or Emma Bovary | Answer: you absorb heroic passion, not windmill-tilting |
Why must stories have conflict? Because simulators are useful only when they throw trouble at you: "clear skies, textbook-perfect engines, and model passengers are not the simulations that improve pilot readiness." A Josef K. who happily confesses, or a Dorothy who hands over the ruby slippers and settles in Munchkinland, makes a short, useless story. The craft rule "no conflict, no story" and the Darwinian rule "no difficulty, no adaptive value" are the same rule.
Dreams as night-shift simulation
- Greene's high-school "Hygiene" class, taught by gym coaches, required presentations; he chose sleep and read Freud's Interpretation of Dreams.
- Michel Jouvet (late 1950s) lesioned the locus coeruleus region in cats, removing the block on dream-driven movement: sleeping cats crouched, arched, hissed and pawed at phantom prey and predators.
- Rats replay the neural patterns of a newly learned maze while dreaming, so precisely that researchers can track their progress through the dreamed maze.
- Inference (with warnings): our common ancestor with cats and rodents lived 70–80 million years ago, but dreams may be cognitive and emotional workouts. We spend about seven years of a lifetime in REM sleep consuming self-authored stories.
Story and other minds
Physics has its own stories. In November 1915, finishing general relativity, Einstein computed the anomalous precession of Mercury's orbit and felt heart palpitations: after a decade at sea in mathematics, land was sighted (Whitehead's image). Greene has had smaller versions of that thrill: a sense of deep connection with the cosmos. Mathematics confirmed by experiment lets us commune with the universe; ordinary stories let us commune with other minds. Joyce Carol Oates: reading is the one way we slip "into another's skin; another's voice; another's soul."
"With math we commune with other realities; with story we commune with other minds." We prevailed by being intensely social: not just safety in numbers but innovate, participate, delegate and collaborate in numbers. Jerome Bruner doubted collective life would be possible without narrative; Brian Boyd argues that stories make social life easier to read and richer in possibilities.
Star Trek's "Darmok" (which Greene watched with his father and later his son) imagines the Tamarians, who speak only in allusions to shared stories; Picard connects with their captain by telling the Epic of Gilgamesh. John Tooby and Leda Cosmides explain why narrative is one of our primary schemas: we evolved from organisms whose only non-innate information was personal experience, which arrives in story-shaped packets. With particle-level vision we might think in trajectories; with human senses we think in narrative.
Form is not content. Science uses narrative form but its content is judged by logic and replicable test, minimising subjectivity. That is its power and its limit: Schrödinger's equation tells us a lot about electrons and little about Schrödinger. Human stories do the opposite. Ambrose Bierce's "An Occurrence at Owl Creek Bridge", set in the instant of a hanging, as Peyton Farquhar imagines reaching his wife before the rope snaps tight, distils what Becker called the "excruciating inner yearning for life." If aliens ever visit, our physics will hold little news for them; our stories will tell them who we are.
Myth: story raised to the sacred
Some scientific findings transcend the specialist: we orbit an average star; time passes at different rates for observers in relative motion; our three-dimensional world may be a slice of a larger space. These reshape a worldview. Myths play that role for cultures: stories invoking supernatural agents to explain origins, rituals and the order of things, held sacred, defining a people.
| Theorist | Theory of myth |
|---|---|
| James Frazer | Early attempts to explain puzzling natural phenomena |
| Carl Jung | Expressions of archetypes innate to the unconscious |
| Joseph Campbell | The monomyth: call to adventure, trials and death-defying rites, return as transformed hero |
| Michael Witzel | A universal template appears only across whole traditions, running from creation to final destruction, traceable to an African proto-mythology ~100,000 years ago |
| Karen Armstrong | Myths are "nearly always rooted in the experience of death and the fear of extinction" (Greene softens to "often") |
Death is everywhere in myth. Gilgamesh crosses wilderness, scorpion monsters and the Waters of Death seeking immortality. The Hindu goddess Kali is beheaded by jealous gods. In the Kono creation myth, Sa, deity of death, decrees mortality for all after believing the god Alatangana abducted his daughter. Ma-ui of Oceania tries to win immortality by ripping out the heart of the sleeping Great Hina-of-the-Night and is shredded by her teeth. World-ending tales abound: Ragnarök, molten metal in Zoroastrian myth, Śiva's destructive dance, Maya fire and water, Atum destroying the earth, and floods worldwide.
Why? Narrative demands conflict; myth deals in the largest themes; and language-endowed minds that roam past and future cannot escape the pattern that every life ends. To reflect on origins is to raise endings.
Why giants and serpents? Three engines of myth
The cognitive recipe for myth
- Minimally counterintuitive concepts (Pascal Boyer, after Dan Sperber). Ideas are remembered and passed on best when they violate one or two deep expectations: an invisible person, yes; a river that sings calculus answers to the M*A*S*H theme, no. Mythic beings are familiar in body and personality, extraordinary in power.
- Combinatorial language. Language is a "narrative Mr. Potato Head": swap parts of ordinary descriptions (giant rocks, talking people) into talking rocks and giant people. The same power that invents myths lets minds see old problems in new ways and innovate.
- Hyperactive agency detection (Justin Barrett, Stewart Guthrie). Mistaking a moonlit shrub for a lion costs little; mistaking a leopard for wind costs your life. So we over-attribute minds, filling the world with beings that help, plot, attack and avenge.
Greene's "Wilson" On a rare camping trip, Greene was sent on a solo stint in the woods with a tarp, sleeping bag, three matches, a can, a pen and a journal. He wasted all three matches on one failed fire. At dusk, panicked, he lay staring at the tarp inches above his face while every creak became a bear or mountain lion. He drew two eyes, a splotchy nose and a slightly upturned mouth on the tarp. Still alone, he didn't feel it as fully: if every noise had a mind, so could his drawing. "I'd created my own Wilson" (the volleyball companion in Cast Away). Agency detection, on demand.
Combine death-haunted themes, minimally counterintuitive heroes, combinatorial imagination and over-detected agents, and you have the stories that guided our ancestors. The most enduring of them would seed religion.
Narrative is an interface, not decoration. Greene's flight-simulator thesis explains why scenario-based risk training, war-gaming and "pre-mortems" work better than rule lists: they train intuition through simulated conflict. In AI programmes in banking, the most effective governance artefacts are often worked failure stories (a model that drifted, a copilot that hallucinated a covenant) rather than policy text. Design them with real conflict; a scenario where nothing goes wrong teaches nothing.
Brains and Belief
Once minds could roam past and future, they buried their dead with care, painted in the deepest caves, and imagined unseen worlds. Greene weighs the main evolutionary explanations of religion: a by-product of inference systems tuned for survival (Boyer), social glue for groups larger than kin (D. S. Wilson), an invisible watcher that protects reputations (Bering) and a buffer against death anxiety (terror management). None is decisive. He then separates belief as confidence (calibrated to evidence, as in physics) from belief in value (in heritage, ritual, meaning). Literal religious claims that contradict physical law are false, but religion read for its value, not its facts, can offer coherence, connection and a lineage across the ages.
Technicians become philosophers
Any civilisation capable of building telescopes, Greene guesses, will have a history full of attempts to find meaning. "Answer enough how questions and why questions quickly follow." Our ancestors had to be technicians (stone, bronze, iron, hunting, farming), but "to survive is to kindle the search for why survival matters," so technicians became philosophers, scientists, theologians, artists and poets. The most persistent questions are existential: how did it begin, how will it end, where do we go?
Burials, dreams and caves
~100,000 years ago: Qafzeh 11, Lower Galilee (excavated from 1934). A child of four or five suffers a blow that gashes the front right of the skull, lives with brain damage to twelve or thirteen, and is buried with two deer antlers across the chest, one end resting on the palm: the excavators judged it "a deliberate, ceremonial burial." 26 others were found at the site.
40,000+ years ago: earliest dated cave paintings.
~30,000+ years ago: Sunghir, ~200 km northeast of Moscow, found in 1955 by excavator operator Alexander Nacharov of the Vladimir Ceramic Works. A boy and girl (about 10 and 12) lie head to head with arctic-fox-tooth headgear, ivory armbands, a dozen-plus ivory spears and more than 10,000 carved ivory beads, over a year of full-time work at 100 hours a week.
Such investment hints that burial was a strategy to transcend the finality of death: the body ends, but something honoured by grave goods carries on. Nineteenth-century anthropologist Edward Burnett Tylor pointed to dreams: waking from a visit with a dead friend leaves the sense that they still exist somewhere. Sumerians and Egyptians read dreams as divine directives; the Bible is full of them; Aboriginal Australians speak of the Dreamtime, an eternal world from which life comes and to which it returns; rhythmic dance and drumming produce trance journeys to other planes. Add storms, floods, earthquakes and sudden deaths, and a mind primed to see agents everywhere will populate the sky with beings wielding the powers we lack. That was "a spectacularly clever response": it turned random events into coherent stories and recast death as a passage.
Cave art (tens of thousands of images, some over 40,000 years old) shows lions, rhinos and hybrids (deer-woman, bird-man); humans appear only as sketches, and stencilled handprints abound. Greene warns against over-reading, but the effort involved argues against doodling or "art for art's sake": archaeologist David Lewis-Williams describes crawling more than a kilometre through absolutely dark passages, sliding along mud banks and wading underground rivers.
| Hunting magic (Salomon Reinach, early 1900s) | Shamanic trance (Lewis-Williams, after Mircea Eliade) |
|---|---|
| Paintings were rituals to ensure a successful hunt | Paintings record visions of shamans travelling to unseen worlds, negotiating with spirits and channelling animals |
| Problem (Greene's note, citing Jean Clottes): animals painted don't match the bones of animals eaten nearby; you'd expect to paint the bison you hunt | Fits the depth and difficulty of the sites and recurring motifs across continents |
Archaeologist Benjamin Smith: caves "were far from just 'canvases'"; they were places of ritual and communication with spirits and ancestors. The strands later called religion (ceremonial burial, art imagining other realities, myths of spirits and afterlife) were braiding together, with recognition of impermanence woven through.
Religion as a by-product: Boyer's inference systems
Pascal Boyer (Religion Explained) argues that the explanation lies in how all normal human minds work, religious or not. The brain is not a general-purpose computer awaiting any program; it is a special-purpose computer with programs shaped by natural selection, running "inference systems" for spear-throwing, courtship and alliance-building. Religious ideas hijack them.
Boyer's inference systems and what religion borrows from each
- Theory of mind + hyperactive agency detection: we readily imagine attentive minds in the sky and underground.
- Intuitive psychology and physics: give gods familiar minds and bodies, then violate expectations in one or two ways (minimally counterintuitive), making them memorable.
- Social exchange / reciprocity tracking: "I'll sacrifice, pray and do good, but come tomorrow's combat, you've got my back." This explains the transactional tone of worship, and why misfortune is read as failing divine expectations.
- Package deal: a predilection for religion need have no adaptive value of its own; it comes bundled with traits that do. As with the sweet tooth, susceptibility is not destiny.
Ghosts, gods, demons, saints and souls are "virtuoso conductors of the evolving human mind": we attend to them, act on them and spread them. Boyer doesn't deny religion explains origins and death; he denies that explains why it arose and has the features it does. "The elephant in the religious room is the human mind." Greene finds this compelling but leaves room for religion to be adaptive in its own right.
Religion as social glue
For kin, cooperation is easy to explain: Fisher, Haldane and Hamilton developed kin selection (inclusive fitness), so that saving your sister from a charging elephant helps copies of your genes. (Greene's note: inclusive fitness was challenged in 2010 by Nowak, Tarnita and E. O. Wilson, prompting a rebuttal signed by 136 researchers.) But what binds large groups of non-kin? David Sloan Wilson, following Émile Durkheim, argues that religion, meaning story plus doctrine, ritual, custom, symbol, art and moral codes, sanctified, extends the club of kinship, priming unrelated people to cooperate and defend one another.
| Within-group dynamics | Between-group dynamics |
|---|---|
| Free riders exploit cooperators, out-reproduce them and drive "trusting companions, together with their religious sensibilities," to extinction | Groups compete and fight; cohesive groups win. Darwin: a tribe of "courageous, sympathetic and faithful members" would "conquer the other" |
| Favours selfishness | Favours cooperation, all the more if fired by devotion to ancestors or watchful gods |
The religion-as-glue thesis needs between-group selection to dominate across thousands of generations, which is contested. Sceptics add game theory: in simulated tournaments of strategies, tit-for-tat ("I'll do good for you if you do good for me, and retaliate fast if you cross me") reliably beats far more selfish rivals, so cooperation can evolve without shared belief. Each side has declared victory; there is no consensus. Greene's notes add Richard Sosis's finding that the emotional and costly commitments of religious communes predict their longevity.
Religion and the individual
Jesse Bering: the invisible watcher
- Theory. After language, a single widely gossiped transgression (stealing food or partners, shirking on the hunt) could wreck a reputation and reproductive prospects. Someone who imagines a powerful witness in the wind, trees or sky transgresses less, stays out of gossip, has more children and passes on the god-fearing instinct.
- Experiment. Children left alone with a hard task often cheat; children told a friendly invisible presence is in the room cheat far less, even those who say they don't believe in it.
- Inference. Young minds, less shaped by culture, are predisposed to behave as if watched.
Terror management theory, developed by social psychologists (Sheldon Solomon, Jeff Greenberg, Tom Pyszczynski) from Becker, holds that awareness of death "would have rendered our ancestors quivering piles of biological protoplasm," and that promises of literal or symbolic immortality saved us.
The Arizona judges (Rosenblatt et al., 1989)
- Setup. Municipal judges set a bond (Greene calls it a fine) for a defendant charged with prostitution, a misdemeanour; half first answered two questions about their own death ("What emotions does thinking about your own death arouse?").
- Result. The mortality-primed judges set bonds averaging nine times those of the control group.
- Claim. If trained, dispassionate judges are swayed this much, similar stealthy effects act on us all; hundreds of later studies reported effects on voting, prejudice, creativity and religiosity. "If you scoff at such a possibility it is because culture is doing its job."
The replication problem. Greene presents terror management as well supported, but the field's centrepiece has not survived scrutiny. Many Labs 4 (preregistered, 21 labs, ~2,200 participants, published 2022) failed to replicate the basic mortality-salience effect, with or without the original authors' involvement, and later Bayesian reanalyses found evidence against it; further failed replications followed in 2023. The original judge study involved only 22 judges. Becker's philosophical picture may still have merit, but the experimental pillar Greene leans on is now weak. A related casualty: a high-profile 2019 Nature paper arguing that "moralizing gods" followed rather than enabled social complexity was retracted in 2021.
Boyer rejects the comfort thesis: religious worlds are often "every bit as terrifying," casting "a thick pall of gloom." Greene suggests a middle path: religion may have offered a softer light on death and a more enduring narrative, what William James called "an assurance of safety and a temper of peace" plus a "new zest" for life. There is no consensus on why religion arose or persists, and the evidence may be too spotty for one ever to emerge. Greene stays partial to mortality: Stephen Jay Gould said a large brain taught us "the inevitability of our personal mortality" and "all religion began with an awareness of death." Whether that awareness then became an adaptive advantage is a separate question.
| Theory | Key figure(s) | Religion is... | Main weakness |
|---|---|---|---|
| Cognitive by-product | Pascal Boyer, Justin Barrett, Scott Atran | A side effect of inference systems built for other jobs | Doesn't explain religion's costliness and persistence |
| Group cohesion | D. S. Wilson, Durkheim, Darwin | Glue that extends kinship to strangers | Needs between-group selection to dominate; tit-for-tat explains cooperation without gods |
| Reputation guard | Jesse Bering | An imagined watcher that curbs costly transgressions | Experimental base is small and child-based |
| Death-anxiety buffer | Becker; Solomon, Greenberg, Pyszczynski | Symbolic or literal immortality that keeps us functioning | Mortality-salience effects failed large replications |
| Mortality awareness | Gould, Greene | Born of the big brain's knowledge of death | Origin story, not proof of adaptive value |
The Axial Age and the great traditions
In the first millennium BCE, across India, China and Judea, thinkers re-examined old myths; Karl Jaspers called it the "beginnings of the world religions, by which human beings still live." Oral traditions were written down and sanctified, sharing the questions of this book: where did we come from and where are we going?
Greene's brother. In the late 1960s, walking through Central Park with his father and sister, the not-yet-ten-year-old Greene saw Hare Krishna devotees drumming at the Naumburg Bandshell and realised one shaven-headed drummer was his brother, supposedly away at college. Their later conversations circled the Vedas. Siddhārtha Gautama. Prince from present-day Nepal, raised on the Vedas, who gave up luxury in the mid-sixth century BCE to seek an end to suffering. The Dalai Lama. At a public forum, Greene asked whether books claiming modern physics recapitulates Eastern thought were valid. His reply: Buddhism has something important to say about consciousness, but on material reality "we need to look to you and your colleagues." William James. Science must be complemented by the inner world: "the terror and beauty of phenomena," the "sublimity" of the stars.
- Vedas and Upanishads. Sanskrit texts from as early as ~1500 BCE, with Upanishadic commentary after the eighth century BCE; the scriptures of Hinduism. The Rig Veda's creation hymn begins before existence and non-existence ("What stirred? Where?") and ends in uncertainty; the dismembered, thousand-headed Purusha forms sun, earth and moon. To Greene's brother they were a grander cosmology; to Greene, a moving poetic framework that, like physics, seeks something stable beneath shifting appearances, but with completely different standards of explanation.
- Buddhism. Perception is illusory; everything changes; there is no immutable substrate, and suffering comes from failing to recognise impermanence. The goal is release from the cycle of rebirth into a state beyond desire, suffering and self. Greene finds this even more remarkable than an afterlife: death reimagined as a step in a cycle whose very goal is permanent release from existence; impermanence becomes a sacred passage to the timeless.
- Physics and mysticism. Greene has found only "metaphorical resonance between distinct ideas vaguely construed." Mathematics is the anchor of physics; words are translations, and links built on translations rarely rise above "a poetic alliance."
- Abrahamic faiths. During the Babylonian exile, Judean leaders gathered writings and oral history into early versions of the Hebrew Bible. Genesis tells two creation stories (six days ending with man and woman; one day, with woman arriving during man's first nap) and is noncommittal about an afterlife. Christianity, about five centuries later, made eternal souls central; Islam, five centuries after that, shared its day of judgement, heaven and damnation.
Adherents (Greene's 2020 figures, with Pew's 2025 update)
- Hinduism: 1.1 billion ("one in seven"); Pew 2025 (data for 2020): ~1.2 billion.
- Buddhism: ~500 million ("one in a dozen"); Pew 2025: ~324 million, the only major group to shrink between 2010 and 2020.
- Abrahamic faiths: ~4 billion (more than one in two); Pew 2025: Christians ~2.3 billion, Muslims ~2.0 billion.
- Total in surveyed religions: more than three in four people; plus 4,000+ smaller religions.
- Unaffiliated (not in Greene's tally): ~1.9 billion, the fastest-growing category in Pew's 2025 count.
Ramtha's School of Enlightenment
Greene's keynote by accident
- How. Swamped with a project, Greene accepted a keynote in Washington State without vetting the host: Ramtha's School of Enlightenment, led by J. Z. Knight, who claims to channel Ramtha, a 35,000-year-old warrior from Lemuria (often at war with Atlantis). His young daughter, watching an old Merv Griffin Show clip, failed to suppress her giggles.
- What he saw. Hundreds of blindfolded people groping across a field for a card matching their written life's dream (one success that session); blindfolded archers; a young woman "reading" a shuffled deck ("Seven of diamonds... Six of clubs. But I was only off by one").
- What he said. We are pattern detectors because pattern recognition is how we survive, but our detectors are primed to over-fire. You'll guess a card's suit 1 time in 4 and its rank 1 in 13 by chance. "How often do you notice that a remarkable coincidence did not happen?"
- Outcome. A standing ovation, twice, which he called "appreciated but confusing." At the signing, some admitted they didn't buy much of it but came to be among others who feel "there is something else out there." Greene relates: physics repeatedly finds something strange out there, but with centuries-refined methods of mathematics and replicable experiment.
He is open to extraordinary claims given carefully designed, replicable data. Without such data, without any reason to expect it, and in flat contradiction with established physics, there is no basis for belief. That raises the bigger question about an all-powerful creator who hears prayers and dispenses rewards.
Belief as confidence
Asked about God and quantum mechanics in the same breath, Greene answers in terms of confidence: high for quantum mechanics, which predicts the electron's magnetic moment beyond the ninth decimal place; low for God, given the paucity of rigorous data.
How physicists quantify belief
- "Discovery" threshold: the chance a statistical fluke produced the signal must be below ~1 in 3.5 million (the five-sigma standard).
- Updating: new data revise confidence by formal rules (Bayesian updating), as Jack-and-Jill-seen-together raises your belief they're a couple, until you learn they're siblings.
- Caveat: even a discovery can be revised.
Evolution did not build brains to track truth. It built them to favour beliefs that produce survival-promoting behaviour: investigating every rustle for a real agent costs more than assuming one. "Verity is an important character in the drama of belief but is easily upstaged by survival and reproduction." Then add emotion: Darwin's 1872 Expression of the Emotions argued that smiling and blushing are biological universals. Fear and parental love are ancient; embarrassment, guilt and shame came later, as groups grew. And add authority: children believe parents because, as Richard Dawkins notes, that usually aids survival; adults extend trust to teachers, officials and experts because no one can re-verify millennia of knowledge.
Greene confesses a nightmare: at his PhD defence, the examiner reveals that all the experiments supporting quantum mechanics were faked, an elaborate joke. He has personally checked "a tiny fraction" of physics' key results: "You could say I've taken most results on faith." Science's protection is process: relentless testing that minimises subjectivity over time. Even so, contingency seeps in; he suspects the long dominance of the Copenhagen interpretation owed much to powerful personalities at quantum mechanics' birth.
Humans also hold incompatible beliefs comfortably. Greene knocks on wood, talks to the departed and occasionally seeks "heavenly reinforcement," and enjoys these apotropaic lapses. Boyer notes that the "judge and jury" model of belief is often wrong: religious concepts tickle inference systems below awareness rather than passing rational review. Karen Armstrong: celebrants of the Eleusinian mysteries would have been puzzled to be asked whether they "believed" Persephone descended underground, as you would if asked whether you believe in winter; myth gave a poetic schema inseparable from the reality it illuminated.
Dead metaphors and literal faith. We "sprinkle" sentences with metaphors so worn that "water evaporates, not poetry" and "horses work, not words." Perhaps religious ideas follow the same path: born as poetic ways of seeing, they lose their poetry over centuries and harden into literal claims.
Compatibility is not explanation
Greene grants that no one can rule out some god, as long as it never alters what our laws describe. But compatibility is not explanatory necessity. We use Einstein, Schrödinger, Darwin and Wallace, Watson and Crick because they give powerful, detailed, predictive explanation. On that measure religious doctrines do not register (many believers deem the measure irrelevant). A literal religious claim that contravenes established physical law is false, full stop, on a par with Ramtha.
| Belief in fact | Belief in value |
|---|---|
| Is the doctrine literally true of the world? | Does the practice enrich a life? |
| Judged by evidence and prediction | Judged by meaning, comfort, community, ritual, moral guidance, art |
| Literal claims against physics fail | Stories read poetically, symbolically or as fiction, cherry-picked and reinterpreted, stay within rational discourse |
| Science's domain: outer, objective reality | Spiritual practice's domain: the inner world, "terror and beauty," good and evil, awe, gratitude, which particles alone never show |
Religions are old, and that matters: for centuries they have held people's attention, given ritual and moral structure, inspired art, promised that death is not the end, and also intimidated, sanctioned violence and justified slavery and killing. "Some good, some bad, some utterly awful." Through it all they have provided coherence and lineage.
Greene's own religion: tefillin and Kaddish
- Childhood. Raised Jewish; in a Hebrew school that restarted the alphabet each year, he sat aside reading the Old Testament (Samuel, Absalom, Ishmael, Job) and enjoyed it despite complaining.
- Jerusalem. On a break from graduate study at Oxford, an "overzealous rabbi" tracked him down, surrounded him with Talmudic scholars "also studying the origin of the universe," and pressured him into wrapping tefillin. To the rabbi, God's will; to Greene, coercion without conviction. Unwrapping the leather bands, he "knew he was done."
- His father's death. Yet when his father, not religious himself, died, a minyan came daily to recite Kaddish in Aramaic. He wanted no translation; the words were "tribal poetry." What mattered was history and connection: "That, to me, is the grandeur of heritage. That, to me, is the majesty of religion."
Calibrated confidence is a governance tool. Greene's split between confidence (evidence-calibrated) and value (meaning-driven) maps cleanly onto how banks should talk about AI. Model risk teams need five-sigma-style thresholds and Bayesian updating for claims about accuracy and bias; change programmes need narrative, ritual and identity to make adoption stick. Confusing the two produces either hype ("we believe in AI") or cynicism ("show me the p-value for culture"). His Ramtha lesson, "how often do you notice that a remarkable coincidence did not happen?", is also the best one-line warning against cherry-picked AI success stories.
Instinct and Creativity
Art is ancient and universal, yet unlike eating or mating it does not wear its survival value on its sleeve. Greene sets out four explanations: sexual display (Darwin, Miller, Dutton), "cheesecake for the mind" (Pinker), social bonding (Dissanayake, Carroll) and training ground for flexible, innovative thought (Boyd, Greene's favourite). Whatever its origin, art delivers a distinct kind of truth: it makes us "feel a thought," reaching layers of experience that facts and equations cannot. And because we know we will die, art has always been our main road to symbolic immortality, the only immortality, Krutch says, we are likely to get.
Beethoven's Ninth, Vienna, 1824
On 7 May 1824, at Vienna's Theater am Kärntnertor, Beethoven appeared on stage for the first time in nearly twelve years to premiere his Ninth Symphony. Billed only to "assist," he could not hold back: first violinist Joseph Böhm described him throwing himself about "like a madman," crouching, stretching, flailing as if to play every instrument. Nearly deaf and tormented by tinnitus, he fell a few bars behind and was still conducting when the music ended; the contralto turned him to see the cheering, handkerchief-waving audience, and Beethoven wept. How could he have known that sounds heard only in his mind would strike a universal chord?
Myth and religion were collective attempts to make sense of the world. Individuals walked the same path, and a rare few, in every age, found new ways to articulate the transcendent: a truth "standing beyond derivation or validation."
The puzzle of art
Pattern-learning is a survival skill so basic that aliens would share it. But aliens might stare blankly at humans arranging pigment, chipping marble or vibrating air and feeling reality open up, and still more at those who spend their lives making such patterns. Nietzsche: without music, life would be a mistake. Shaw (through Ecrasia): without art, reality's crudeness would be unbearable.
Motives for creating are many: self-expression, status, and a nod to eternity (Keith Haring called art a "quest for immortality"). If art were recent or rare, like bell-bottoms and fried bananas, its history would say little about human nature. But cave paintings and grave goods date back 30,000–40,000 years, with engravings and artefacts a few hundred thousand years older still. "Because we like it" is a proximate answer, like explaining ice cream by sweetness. Can we go deeper?
Older still. Since 2020, dating has pushed figurative art further back: a cave painting of a pig and human figures in Sulawesi, Indonesia, was dated in 2024 to at least 51,200 years, the oldest known narrative art. Engraved geometric patterns on a Homo erectus shell from Trinil, Java (~500,000 years), published in 2014, support Greene's "few hundred thousand years earlier" for non-figurative marks.
Sexual selection: the peacock's tail
The flight simulator worked for story; it strains for abstract art. It is hard to see how listening to the Pleistocene's Édith Piaf or Stravinsky helped anyone survive. Darwin, troubled by the peacock's tail (beautiful, cumbersome, a liability when fleeing predators), proposed sexual selection in The Descent of Man (1871): survival is necessary but not sufficient; reproduction is what counts, so traits that win mates spread even at some cost to safety. Peahens like splendid plumes; their chicks inherit both dad's tails and mom's taste.
Sexual display across species
- Peacock: extravagant tail, a burden in flight.
- White-bearded manakin: raucous, "mosh pit" courtship dances.
- Fireflies: flashing light shows, success turning on finesse.
- Bowerbird: ornate bachelor dens of twigs, leaves, shells and candy wrappers, built solely to seduce.
- Two readings: Darwin's "taste for the beautiful" (aesthetic preference) versus fitness indicators (only a strong, healthy male can afford such a tail; beauty is a public credential). Either way, costly traits can spread.
Darwin applied this to humans: piercings, body paint and music as products of mate choice, music descending from mating calls. Geoffrey Miller and Denis Dutton developed it: art is a fitness indicator, showing a mind "firing on all cylinders" and an artist with resources to spare ("Artists of the Pleistocene, apparently, were anything but starving"). Wallace, who saw aesthetics as God's gift, found the whole idea unseemly.
Problems with art-as-courtship. Is artistic talent an honest signal of health? Might art spread simply because it is bundled with intelligence and creativity, which have obvious survival value? Why do women make art if it's a male display? And Pleistocene courtship is largely conjecture: the conquests of Lucian Freud and Mick Jagger tell us little about early hominins. Brian Boyd's verdict: "Sexual selection has been an extra gear for art, not the engine itself." (Not in Greene: ornithologist Richard Prum's The Evolution of Beauty, 2017, revived Darwin's aesthetic reading, arguing that much display is arbitrary beauty co-evolving with taste, not a fitness signal.)
Pinker's cheesecake
Steven Pinker offers the bluntest view: apart from the language arts, art is "cheesecake for the mind". Evolution made fitness-enhancing behaviours pleasurable, building feedback loops: do it, enjoy it, survive, pass it on. The arts cut the loop and stimulate pleasure centres directly, "megadoses of agreeable stimuli" with no fitness payoff. "From the standpoint of survival, the arts are junk food."
Music as "auditory cheesecake"
- Ancestral signal. Vibrating linear objects (vocal cords, a hollow-bone weapon) produce harmonically related frequencies, multiples of a common base, which betray a single identifiable source.
- Adaptive pleasure. Ancestors who enjoyed such organised sound paid closer attention, gaining awareness of predators, footfalls, thunder and cracking branches.
- The hijack. Music exploits that sensitivity for pure pleasure, as cheesecake exploits our ancient taste for calories.
- Greene's gloss. This is not a value judgement; Pinker loves the arts. Significance does not require adaptive utility. Greene recalls his perpetually dieting mother, at a restaurant, standing to salute a passing cheesecake.
Art for the group, art for the innovator
Anthropologist Ellen Dissanayake insists that in ancestral life, art (like religion) was not a weekend hobby but woven into existence. Two adaptive roles follow:
| Social bonding | Cognitive training ground |
|---|---|
| Shared rhythm, song and dance produce intense collective emotion and solidarity (Greene recommends trying prolonged group drumming) | Art is a space free of "flat-footed truth" where the mind can jump, twist and tumble through imagined novelty |
| Noël Carroll: art stirs and shapes emotions to bind people "as participants in a culture" | A mind used to crossing between real and imagined, while tracking which is which, breaks conventional thinking |
| Groups that can predict each other's emotions cooperate better | Primed for innovation: the spear, cooking, the wheel, the B Minor Mass, relativity |
Einstein and Bach: rearranging the pieces
- Einstein's move. No new data: the facts about electricity, magnetism and light were old. He inverted an assumption: instead of fixed space and time with variable light speed, fixed light speed with variable space and time.
- Bach's move. Glenn Gould praised Bach's melodic lines that, when "transposed, inverted, made retrograde, or transformed rhythmically," still yield "some entirely new but completely harmonious profile."
- The parallel. Both are genius at reconfiguring familiar building blocks. Einstein said he thought in music and images.
- Conclusion. Neither relativity nor fugues helped anyone survive directly, but both exemplify the capacities that did. Art may have been "the playground of human cognition."
The two roles work in tandem: "Innovation is the foot soldier of creativity. Group cohesion is the army of implementation." Survival needs both: ideas and the collective ability to carry them out. Art sits at their junction, joining language, story, myth and religion as the means by which minds think symbolically, reason counterfactually, imagine freely and work collaboratively. Boyd: art "fundamentally alters our relation to our world."
| Theory | Proponents | Art is... | Status |
|---|---|---|---|
| Sexual selection | Darwin, Geoffrey Miller, Denis Dutton | A costly mating display or fitness signal | "An extra gear, not the engine" (Boyd) |
| By-product / cheesecake | Steven Pinker | Pleasure-button stimulation without fitness payoff | Coherent; underrates art's ubiquity and cost |
| Social bonding | Ellen Dissanayake, Noël Carroll | Ritualised shared emotion that binds groups | Plausible; same group-selection doubts as religion |
| Cognitive training | Brian Boyd, Greene | A safe playground for imaginative, flexible thought | Greene's favoured view |
| Symbolic immortality | Otto Rank, Becker, Krutch | A way to accept death and live on in works | Fits art's obsession with death and eternity |
Music's evolution since 2020. Two lines of research bear on this debate. A 2019 Science survey (Mehr et al.) found that song forms (lullabies, dance songs, healing songs, love songs) are recognisable across cultures from acoustic features alone, supporting a biological basis. And a 2021 Behavioral and Brain Sciences target article (Savage et al.) argued that music evolved as a co-evolved system for social bonding, explicitly rebutting the cheesecake view; Pinker's camp remains unconvinced. Meanwhile, generative AI (image models from 2022, music generators from 2023–24) can now produce competent art with no mortality, mating or group to serve, a live test of which theories describe why humans make art rather than what art is.
Art and truth
About twenty years before writing, on a sunny autumn drive from New York City upstate, Greene ran over a dog that darted into the road, felt both sets of wheels thud, lifted her into the passenger seat and raced to find a vet. She sat up, pressed her head against his hand with an unblinking look (pain, terror, resignation) and died. Rationally he knows he may be reading too much into an all-too-common accident. Still, it carried "a certain kind of truth. Not a propositional truth." Something shifted in his sense of the world.
Greene's list of world-shifting moments
- Holding his first child for the first time.
- Crouching in a rock crevice outside San Francisco while a windstorm howled overhead.
- His young daughter singing solo at a school event.
- Suddenly solving an equation that had resisted months of work.
- Watching a Nepalese family cremate a relative on the banks of the Bagmati River.
- "Flailing" down a double-diamond slope in Trondheim, and surviving.
- Common thread: they lock attention and stir emotion without needing words; Greene, whose working life is entirely verbal, feels no urge to articulate them. "An examined life need not be an articulated life."
The best art produces states comparable to these encounters. Poet Jane Hirshfield: a new image that is fully right expands what is knowable. Saul Bellow, channelling Proust: "Only art penetrates" the "seeming realities of this world" to hint at "another reality, the genuine one"; without it, life shrinks to a "terminology for practical ends." Joseph Conrad: art appeals to the part of us "not dependent on wisdom," to delight, wonder, mystery, pity, beauty and pain, and to the "latent feeling of fellowship with all creation" that links the dead, the living and the unborn.
Two kinds of truth. Objective truth (facts about the world, integrated into its workings) is the raw material for survival and control, and science's domain. Artistic truth "touches a distinct layer," a higher-level story about human experience. Literature gives "heightened studies in human engagement": Odysseus (vengeance and loyalty), Lady Macbeth (ambition and guilt), Holden Caulfield (rebellion), Atticus Finch (quiet heroism), Emma Bovary (the tragedies of connection), Dorothy (self-discovery).
Wordless art works more impressionistically. For Greene: Liszt's Totentanz brings visceral foreboding; Brahms's Third, unsated longing; Bach's Chaconne, "an apotheosis of the sublime"; the "Ode to Joy" finale of Beethoven's Ninth, among the most optimistic statements the species has made; Leonard Cohen's "Hallelujah" praises the imperfect life; Judy Garland's "Over the Rainbow" captures youthful yearning; John Lennon's "Imagine" shows the power of envisioning the possible. Far from empty calories, these "megadoses" give insights hard to get otherwise.
"Words make you think a thought. Music makes you feel a feeling. But a song makes you feel a thought." Yip Harburg, lyricist of "Over the Rainbow"
To feel a thought is, for Greene, the essence of artistic truth: a kind of truth "beyond wisdom... Beyond the reach of logic... Beyond the necessity for proof." We are all bags of particles, and particle facts fully cover how we behave, but they "shed only monochrome light on the richly colored stories" of thought, perception and emotion. Proust: the only true voyage of discovery is "to possess other eyes." Physics does that with mathematics and experiment, finding the strangest lands inside familiar ones, but only by setting aside each observer's peculiarities. For the "all-too-human truths," we need art. Shaw: a mirror shows your face; works of art show your soul.
Poetic immortality
Asked for the most mind-blowing fact about the universe, Greene sometimes says relativity's malleable time, sometimes entanglement ("spooky action at a distance"), and sometimes something schoolchildren learn: we see stars as they were thousands of years ago, and some may already be dead. Light gives an illusion of presence; reflected light carries our own image across space and time, "a poetic immortality racing across the cosmos."
On earth, immortality takes symbolic form. The scale shrinks from eternity to the lifetime of civilisation, but unlike the literal kind, symbolic immortality is real. The only question is strategy: whose works will last?
~600 BCE Sappho laments age seizing her once tender body and invokes Tithonus, granted immortality without youth; a line some scholars take as the true ending suggests Eros gave her "the beauty and brightness of the sun": radiance through poetry.
~1600 Shakespeare's sonnets promise the beloved eternal life through the poet's pen, while the poet himself "must die." The joke is that Shakespeare is the one who lives on.
20th c. Otto Rank: the artist accepts mortality and transfers the urge for eternity onto creative work; making art is a path to sanity, not a symptom of torment. Joseph Wood Krutch: "Man needs eternity... But the eternity of art is, in all probability, the only sort he will ever get."
Could this drive explain why our ancestors diverted energy from survival to art, and why art persists in every culture? Greene: "Yes and yes," whether or not Rank's grand vision is exactly right. Modern literature confirms the obsession, ranging from longing to disdain:
| Writer | Stance on death and eternity |
|---|---|
| Walt Whitman | Refuses to "suspect death"; swears there is "nothing but immortality" |
| W. B. Yeats | Begs to be freed from a heart "fastened to a dying animal" and gathered "into the artifice of eternity" (Byzantium) |
| Herman Melville | "All are born with halters round their necks"; we notice only in death's sudden turn |
| Edgar Allan Poe | Premature burial: the frantic, bloody fight against death's closest embrace |
| Tennessee Williams | Big Daddy: man alone conceives of death, so the rich "buy and buy and buy" hoping one purchase will be life everlasting |
| Dostoevsky | Svidrigailov imagines eternity as a sooty country bathhouse with spiders in the corners |
| Sylvia Plath | "Eternity bores me, I never wanted it" |
| Douglas Adams | Wowbagger the Infinitely Prolonged cures immortal boredom by insulting everyone in the universe alphabetically |
"The examined life examines death." Whatever their evolutionary role, the arts are our most evocative means of expressing what matters most, and among those things are life and death, the finite and the infinite.
Music and connection
For Greene the most concentrated expression is music, which can make us feel we have stepped beyond time. Pablo Casals: music gives "wings of eternity to that which is most ephemeral." Conrad: an "invincible conviction of solidarity that knits together the loneliness of innumerable hearts."
"To keep away the loneliness"
- The interview. In the late 1960s, third-grader Greene, in Mrs. Gerber's class at P.S. 87 in Manhattan, interviewed his father for a report on an occupation. His dad, a composer and performer who joked about his "SPhD" (Seward Park High School dropout; he left in tenth grade to sing and play around the country), was asked why he chose music: "To keep away the loneliness." Then he quickly brightened his tone for a third-grade report.
- The aftermath. Few composers move the world; his father, painfully, came to accept he was not one. Hundreds of yellowing manuscript pages from the 1940s and 1950s now interest only family. Greene may be the last person who still listens, and they let him feel his father's thoughts as a young man.
Helen Keller hears the Ninth (1 February 1924)
- Event. Radio station WEAF in New York broadcast the New York Symphony Orchestra playing Beethoven's Ninth. Keller placed her hands on the diaphragm of an uncovered speaker and felt the music through its vibrations, distinguishing instruments and recognising the human voices as the chorus grew "exultant."
- Her reflection. She remembered that the composer was deaf like herself, and marvelled at the "quenchless spirit" that wrought joy out of pain: a symphony breaking "like a sea upon the silent shores of his soul and mine."
- Why it closes the chapter. It links the Vienna premiere of 1824 to a deaf-blind listener a century later: music connecting minds across disability, distance and time, the closest thing we have to transcending time.
"Feel a thought" is a design brief. Greene's distinction between objective truth and artistic truth explains why dashboards rarely change executive minds and well-told cases do. For enterprise architects selling a transformation, the numbers establish objective truth; a vivid, emotionally precise narrative is what makes leaders feel the thought and act. And the "foot soldier / army" line is a useful test for any AI innovation programme: a lab that produces ideas without a mechanism for collective implementation has built half the adaptation.
Duration and Impermanence
Greene now climbs into the future on an imaginary Empire State Building whose floor N stands for 10N years after the big bang (today: just above floor 10). The sun dies to a cold white dwarf; distant galaxies vanish over our cosmic horizon; star formation ends around floor 14; stellar near-misses fling planets and stars loose; gravitational waves drag survivors into dead suns and galactic black holes; and if protons decay, by floor 38 atoms themselves dissolve. Finally, he follows Freeman Dyson in asking whether any abstract "Thinker" can think forever. If the accelerated expansion persists, the cosmic horizon's faint, permanent warmth (about 10−30 K) eventually stops the Thinker shedding waste heat, and by roughly floor 50 the universe may have hosted its final thought.
The timeless moment
Every culture reveres something permanent: souls, sacred texts, gods, eternal laws, art, theorems. We covet permanence but never attain it. The nearest we come is a sense of time dropping away. As a teenager on a survival course in Vermont, Greene and eight others were rousted for a night hike through brush and a waist-deep mud swamp, then left in a clearing with three sleeping bags for nine people. They zipped the bags together, stripped and huddled; some cursed, some cried. Then a brilliant aurora borealis filled the sky. "Man, nature, universe." He lost track of minutes or hours: "For a brief moment, time had dissolved." Such episodes are rare and fleeting. Impermanence underlies experience, and as this chapter shows, even space, galaxies and matter lie within time's reach.
Evolution, entropy and the rules of the forecast
The tidy story that evolution builds and entropy destroys isn't quite right: through the two-step they cooperate, producing a mind that wrote the Ninth Symphony and billions who find it sublime. Will both stay in charge?
- Darwinian evolution will change form. Medicine and civilisation mean genotypes that would struggle on the savanna do fine in New York. Selection still operates (lactase persistence with dairy diets, high-altitude adaptations, mate preferences shifting average height), but the biggest force may be gene editing, adding volitional design to random variation. Greene imagines a 200-year lifespan with side effects of cyan skin, ten-foot height and a "ravenous blue-centric libido": a self-selected group of Na'vi-like humans would spread fast. Designed sentience, biological, artificial or hybrid, makes long-range forecasting of life impossible.
- Entropy will stay. The second law follows from basic probabilistic reasoning; it survived the jump from classical to quantum physics and will almost certainly survive future revisions.
Greene's forecasting assumptions
- Laws as now known, operating undirected as since the big bang.
- No changing laws or "constants", including slow drifts too small to detect today.
- No intelligence steering stars, galaxies or cosmic expansion (or making new universes). Such influences are possible but unpredictable: intelligence's vast behavioural repertoire (the freedom of chapter 5) defeats long-range prediction.
- Reading: treat what follows as what would happen absent such interventions. Greene suspects future discoveries will refine details, not rewrite the sweep.
The Empire State Building of time
Floor 1 = 10 years, floor 2 = 100 years, and so on; each floor represents ten times the one below. Climbing from floor 12 to 13 (a trillion to ten trillion years) takes nine trillion years, more than all lower floors combined. A human life is ~100 years (floor 2), a durable empire ~1,000 (floor 3). The observation deck, floor 86, is 1086 years; yet from the top landing, floor 102, that span is thinner than the paint on the last step. Today, at 13.8 billion years, we are a few steps above floor 10; everything in chapters 1–8 happened below that.
The black sun (floors 10–11)
For almost five billion years the sun has balanced gravity with the pressure from hydrogen fusion in its core, like an air pump keeping up a bounce house. In about five billion more, the core will be clogged with helium, which sinks like sand in a pond. The core is at ~15 million degrees (hydrogen fusion needs ~10 million) but helium fusion needs ~100 million, so fusion falters, gravity wins, and the sun contracts and heats. That ignites hydrogen fusion in a shell around the helium core, so fiercely that the sun swells into a red giant while shedding mass, which loosens the planets' orbits.
The race between orbits and the swelling sun
- Mercury: swallowed and vaporised.
- Venus: probably engulfed, though some simulations have the sun fall just short.
- Earth: a coin-toss; even if spared, its surface reaches thousands of degrees, oceans boil off, the atmosphere is stripped and lava floods the surface.
- Mars: safe.
- Us: almost certainly long gone, having dodged self-destruction, pandemics, asteroids "and alien invasions," and left for a better home.
About 5.5 billion years from now the core reaches helium-burning temperature, producing carbon and oxygen; after a brief eruption the sun settles. Some 100 million years later carbon and oxygen clog the core; they need 600 million degrees to fuse, which the sun can never reach. A helium shell burns; the core contracts until stopped by Pauli's exclusion principle (1925): no two identical matter particles can share a quantum state, so packed electrons "demand their personal space." (Pauli was famously caustic: "I don't mind your thinking slowly; I mind your publishing faster than you think.") The outer layers drift away and a white dwarf of carbon and oxygen remains, glowing for billions of years before cooling to a dark, frozen orb. A few steps above floor 10, the sun fades to black.
The big rip (before floor 11?)
Since Hubble, we have known space expands, and gravity should slow it, like a tossed apple. In the 1990s two teams measured the slowdown with distant supernovae and found the opposite: the expansion has been speeding up for about five billion years. The leading explanation is dark energy: a uniform energy filling space that, like inflation's field, produces repulsive gravity. It needs to be tiny: the dark energy in a cubic metre would light a 100-watt bulb for about five trillionths of a second, but there are a lot of cubic metres. The evidence is compelling but circumstantial.
Dark energy might stay constant, weaken or strengthen. If it strengthens, repulsion eventually beats every binding force:
+20 billion years: galaxy clusters torn apart (Caldwell, Kamionkowski and Weinberg's representative model).
+1 billion years later: the Milky Way's stars fly apart "like sparkles in a fireworks display."
+60 million years later: planets ripped from the sun.
A few months later: stars and planets explode as molecules are pulled apart.
30 minutes later: atoms themselves are torn apart.
Then: perhaps spacetime itself shreds. Reality, which began with a bang, ends with a rip, before floor 11.
Greene thinks this unlikely: the equations work, "barely," but don't look natural, a judgment from experience, not proof.
DESI changes the odds. In 2024 and again in March 2025 (Data Release 2), the Dark Energy Spectroscopic Instrument, combined with supernova and CMB data, found evidence at roughly 2.8 to 4.2 sigma that dark energy is weakening over time rather than holding constant. If confirmed, it cuts against the big rip and, more importantly for this chapter, it is exactly the scenario that revives Dyson's hope for endless thought (see below). The result is not yet at the five-sigma discovery threshold, and analyses are ongoing.
The cliffs of space (floors 11–12)
If dark energy stays constant, there is no rip, but in about a trillion years distant galaxies will recede faster than light. That breaks no rule: Einstein's limit applies to motion through space, and galaxies are mostly carried by swelling space, like paint specks on stretching spandex. But their light, which does travel through space, is like a kayaker paddling slower than the current: it never reaches us. Distant galaxies drop off a cliff beyond our cosmic horizon.
The Local Group (about 30 galaxies) stays bound; the Milky Way and Andromeda will likely merge into "Milkomeda" (Greene would have lobbied for "Andromilky"). But with no distant galaxies to observe, future astronomers lose the main evidence for expansion; Avi Loeb suggests tracking hypervelocity stars escaping Milkomeda, "like tossing popcorn off a raft," but concedes cosmology will suffer. By floor 12 the CMB will be redshifted beyond detection.
Knowledge can decay too. Astronomers a trillion years hence, seeing a black, static sky, might dismiss our records of an expanding universe as primitive error and conclude the cosmos is static. Accelerated expansion can carry essential information beyond reach; deep truths may beckon silently from just beyond the horizon.
Milkomeda is less certain. A 2025 study using Gaia and Hubble data (Sawala et al., Nature Astronomy) put the chance of a Milky Way–Andromeda merger within the next 10 billion years at only about 50%, with the Large Magellanic Cloud and M33 tugging the odds both ways. Greene's "likely merged" is now closer to a coin flip on that timescale, though over the far longer spans of this chapter a merger remains probable.
The twilight of stars (floors 12–15)
The first stars formed on floor 8 (~100 million years). Stars form as long as hydrogen clouds remain; accounting for gas recycled by dying stars, star formation in most galaxies ends by floor 14 (~100 trillion years). Lifetimes run inversely to mass: heavy stars burn hot and fast; the sun lasts ~10 billion years; red dwarfs (down to ~0.1 solar masses), probably the majority of stars, burn slowly and mix their hydrogen thoroughly, shining trillions of years. By floor 14 even late-born red dwarfs are "running on fumes." Galaxies become burnt-out cities of cinders, though dead stars still hold their planets by gravity.
The twilight of astronomical order (floors 15–20)
Stars are sparse: shrink the sun to a grain of sugar at the Empire State Building and Proxima Centauri sits near Greenwich, Connecticut, drifting at under a millimetre per hour, "like a game of tag played by widely dispersed slugs." Over familiar spans, collisions are rare. But by floor 15 (1015 years), close encounters are common:
- Planets are flung away. A heavy star passing close can rip earth from its orbit (an unlikely fate even before the sun dies). An ejected earth freezes; oxygen and nitrogen liquefy and rain from the sky. Yet life at deep-sea vents might carry on for billions of years, powered by radioactive decay of thorium, uranium and potassium in earth's interior: "nuclear fission in its interior" replacing fusion in the sun.
- Galaxies evaporate. In encounters, the lighter star tends to speed up (as when a ping-pong ball on a basketball rebounds high), so over time many stars are kicked out of their galaxies. Between floors 19 and 20, typical galaxies are depleted; their mostly dead stars drift through intergalactic space.
Gravitational waves and the final sweep (floors 20–30)
Why don't planets spiral in like marbles on the classic rubber-sheet picture of gravity? Marbles lose energy to friction (you can hear them roll); space has almost none. But orbiting bodies do lose energy, by shaking spacetime: gravitational waves, predicted by Einstein in 1916 and 1918, though he wavered on whether they were real or an artefact.
1960s: methods mature; gravitational waves accepted as a real prediction.
1974: Russell Hulse and Joe Taylor find the first binary neutron star; Taylor, Lee Fowler and Peter McCulloch show its orbital decay matches gravitational-wave energy loss (indirect proof).
14 September 2015: LIGO's detectors in Louisiana and Washington, upgraded and calibrated just two days earlier, twitch identically for 200 milliseconds. After months of checks (was it a prank or a hack?), the source is identified: two black holes that merged 1.3 billion years ago, releasing in a final frenzy more power than every star in the observable universe combined.
2017: Nobel Prize to Rainer (Ray) Weiss, Barry Barish and Kip Thorne.
The wave's journey, as Greene tells it
- 1.3 billion years ago: multicellular life is just emerging on earth as the black holes collide.
- ~100,000 years ago: the wave crosses the Milky Way's dark-matter halo as humans migrate out of Africa.
- ~100 years ago: passing the Hyades cluster as Einstein writes his first gravitational-wave papers.
- ~50 years ago: scientists start designing detectors.
- Two light-days out: the upgraded detector comes online. Two days later it shakes. A ripple launched before complex life met, at the instant of arrival, the instrument built by that life to catch it.
Ten years of gravitational-wave astronomy. By late 2025 the LIGO–Virgo–KAGRA network had logged around 300 candidate events. The signal GW250114 (January 2025), the clearest yet, confirmed Hawking's 1971 area theorem (merged black holes never shrink in total horizon area) and the Kerr description of spinning black holes, results that bear directly on chapter 10's black-hole thermodynamics.
The same slow drain seals the fate of surviving orbits: on floor 23, earth (if still orbiting) spirals into the dead sun; smaller planets and wider orbits take longer. Remnant stars likewise spiral into their galaxy's central black hole (millions to billions of solar masses): most are consumed by floor 24, and with the black hole's gravitational coaxing, galaxies are swept clean of stars by floor 30. Space is "dark and desolate," dotted with cold planets, dead stars and monstrous black holes.
The fate of matter: proton decay (floor 38)
Could life persist? It will need energy; exotic options exist, such as dark-matter particles annihilating into photons. But a deeper threat looms: matter itself may fall apart. Every atom has protons at its heart; if protons decay into lighter particles (positrons, photons), all complex structure dissolves.
Grand unification "slapped down by nature"
- Theory. In the 1970s Howard Georgi and Sheldon Glashow built the first grand unified theory: strong, weak and electromagnetic forces as facets of one master force at tiny distances. Linking forces links particles, allowing proton → pion + positron, with a half-life of about 1030 years.
- Test. Huge tanks of purified water (each gallon holds about 1026 protons) watched by sensitive detectors: with enough protons, even a rare decay should show up, like a lottery with enough tickets.
- Greene's view from inside. As Georgi's undergraduate student in the mid-1980s, he felt the anticipation. Year after year: no decays. The lower limit on the proton lifetime now stands at about 1034 years.
- Aftermath. Georgi called it "being slapped down by nature" and turned against unification. But nearly every successor (Kaluza-Klein, supersymmetry, supergravity, superstrings, extended GUTs) still predicts proton decay, typically at 1034 to 1037 years or longer. Avoiding decay requires contortions, so many theorists expect it.
Greene adopts a proton lifetime of ~1038 years. From floor 38, every atom of every molecule of every structure (rocks, rabbits, trees, planets, stars, you) disintegrates. What remains are electrons, positrons, neutrinos and photons, plus quiescent but hungry black holes.
Proton-decay limits since 2020. Super-Kamiokande has pushed the bound for the key channel (proton → positron + neutral pion) beyond 2.4 × 1034 years, still no decay. Its successor Hyper-Kamiokande (about eight times larger) and the US DUNE experiment are due to start taking data later this decade and will probe lifetimes approaching 1035 years, into the range many unified theories favour.
So does life hit the final wall? Perhaps, or perhaps over a billion billion billion times the present age of the universe it will have abandoned biology. The key assumption is that life and mind are processes, not substrates: "if some other arrangement of basic particles should faithfully execute the processes of life and mind, then that system will live and that system will think." Greene grants this as generously as possible, then asks: obeying only physics, can thought persist indefinitely?
The future of thought: Dyson's Thinker
Isn't this hubris, given how young the science of mind is? Greene's answer is the nested-stories point: Newton knew nothing of quantum physics yet got everyday motion right; quantum mechanics renovated rather than demolished his edifice. Today's reasoning about the far future of mind might be as wrong as Aristotle's entelechy or Empedocles's fire-in-the-eye theory of vision, or it might survive as part of a larger account.
Freeman Dyson's 1979 paper is the template. Any Thinker, of whatever construction, must, like a brain or a steam engine, expel the waste heat its thinking produces, or "burn up in its own entropic waste."
The thought–entropy–heat chain
- Entropy as information. Low entropy means few look-alike configurations, so specifying the actual one conveys little information (one soup can on a sparse shelf); high entropy, lots (one can on a vastly overstocked shelf).
- Thinking is information processing, hence entropy processing; moving entropy requires moving heat (chapter 2).
- Dyson's result: heat to expel scales with number of thoughts; the Thinker must absorb at least that much high-quality energy, which becomes ever scarcer.
- Strategy 1, slow down. Internal speed scales with temperature, so cool down and think slower. Subjectively nothing changes (all processes slow together). Efficiency rises as temperature falls, so infinitely many thoughts need only finite energy, like 1 + ½ + ¼ + ... = 2.
- Snag. A cooler body sheds heat more slowly; eventually entropy builds faster than it can be expelled.
- Strategy 2, hibernate (Dyson). Think in bursts; sleep long enough to shed all waste heat; the Thinker never notices the gaps. Thought continues forever.
Two modern complications
First, a reinterpretation. Computer science (Landauer, Bennett) shows that elementary computation can in principle be done reversibly, with no waste. But erasing memory necessarily produces heat (erasure is irreversible, like shattering glass). A finite Thinker that never erases fills its memory and is condemned to reshuffling old thoughts forever, "not a version of immortality many of us would choose." New thoughts need erasure, hence heat, hence back to hibernation.
Second, the decisive problem: accelerated expansion. A constant dark energy surrounds us with a cosmological horizon that behaves like a sphere of distant heat lamps, giving space a constant temperature of about 10−30 kelvin (explained in chapter 10; it is related to Hawking's black-hole radiation and distinct from the ever-cooling 2.7 K CMB). Tiny, but permanent. Heat flows only from hot to cold, so once the Thinker cools to 10−30 K, space stops accepting its waste. "One more thought (or, more precisely, one more erasure) and the Thinker fries."
| If dark energy stays constant | If dark energy decays |
|---|---|
| A permanent horizon temperature of ~10−30 K | No cosmological horizon; space cools without limit |
| Hibernation eventually fails; last thought around 1050 years (floor 50) | Will Kinney and Katie Freese showed Dyson's optimism returns: hibernating Thinkers think forever |
| If it strengthens: the big rip ends everything sooner | Now favoured, weakly, by DESI 2024–25 |
Even the hopeful branch rests on a "rose-colored" assumption: that free electrons, neutrinos and photons can be gathered into any thinking structure at all. "It is hard to avoid the conclusion that the far future of thought is precarious." Setting aside intelligent intervention, by floor 50 the universe may very well have hosted its final thought.
Floors of the far future (years after the big bang)
- 1010: today (13.8 billion); sun becomes a white dwarf a few steps up.
- ~1011: big-rip deadline, if dark energy strengthens.
- 1012: distant galaxies beyond the horizon; CMB undetectable.
- 1014: star formation ends.
- 1015–1020: stellar flybys strip planets; galaxies evaporate.
- 1023: earth (if still orbiting) spirals into the dead sun.
- 1024–1030: remnant stars swallowed by central black holes.
- 1034: current lower limit on proton lifetime.
- 1038: Greene's assumed proton decay; atoms dissolve.
- 10−30 K: horizon temperature that eventually halts thought.
- 1050: plausible date of the last thought.
A gleam between two nights
1050 years holds the span from the big bang to now more than a billion billion billion billion times. Yet from floor 75 it is shorter than the delay between switching on a lamp and seeing its light, and in an eternal universe any span is infinitesimal. Told from that height, the story is: a moment after the big bang, life arose, briefly contemplated its existence in an indifferent cosmos, and dissolved away. It is Pozzo's cry in Waiting for Godot: they give birth astride a grave; the light gleams an instant, then it's night once more.
Russell would call it bleak. Greene does not: the future science envisions shows how our moment of thought, our instant of light, is rare, wondrous and precious.
What rests on assumptions here. The far-future sequence depends on choices Greene states openly but a reader can lose track of: dark energy's behaviour (now genuinely in question after DESI), proton decay (unobserved; if protons are stable, matter survives far longer and the Thinker's material problems ease, though the horizon-temperature argument still applies), and the exclusion of intelligent intervention. Dyson's framework also assumes thought scales with temperature and that a Thinker can be arbitrarily large and diffuse; critics (Krauss and Starkman, 2000) argued even Dyson's original hibernation scheme fails because of the finite matter accessible within a horizon. The qualitative verdict, precarious, is robust; the specific floor numbers are illustrative.
Landauer's limit is real engineering, not cosmology. The chapter's quiet bombshell is that computing itself is free; forgetting costs energy (at minimum kT ln 2 per erased bit). Today's chips dissipate many orders of magnitude above that floor, but for AI infrastructure planners the direction of travel is clear: energy and heat rejection, not arithmetic, are the binding constraints on compute, the Thinker's problem on a data-centre scale.
The Twilight of Time
Over unlimited time, quantum mechanics guarantees that anything not forbidden eventually happens. Greene follows that principle to the end. Black holes carry entropy and glow (Bekenstein, Hawking), so even they evaporate: sun-mass ones by floor 68, the largest by floor 102. The Higgs field may tunnel to a new value, rewriting physics in a light-speed bubble. And in a forever-expanding cosmos, random fluctuations will eventually assemble Boltzmann brains, a prospect that, taken seriously, would make all our knowledge untrustworthy. Physicists escape that skeptical nightmare only if the universe changes or ends first (decaying dark energy, a Higgs transition, a crunch or a cyclic bounce); meanwhile, in an infinite multiverse, life may persist somewhere, but we cannot reach or even signal it. We can contemplate eternity; we cannot touch it.
Quantum mechanics meets eternity
Quantum mechanics is a "starry-eyed dreamer": it allows a vast range of futures and assigns each a probability. Over familiar timescales we ignore the absurdly improbable ones. Over spans that dwarf the current age of the universe, every outcome not strictly forbidden gets its moment. The chapter examines several such rare processes "biding their time."
Black holes have entropy: Wheeler, Bekenstein, Hawking
In the mid-twentieth century, nuclear and particle physics dominated (Freeman Dyson spoke of physicists' godlike power to "release this energy that fuels the stars"), and general relativity was a backwater. John Wheeler revived it, training a generation. In the early 1970s he posed a puzzle to his student Jacob Bekenstein: throw a hot cup of tea into a black hole; where does its entropy go? If it vanishes behind the horizon, the second law can be broken at will.
The black-hole entropy puzzle and its resolution
- Bekenstein's answer. The entropy isn't lost; it is transferred to the black hole, as a hot pan's entropy passes to your hand.
- The obstacle: "no hair." Werner Israel and Brandon Carter showed a black hole is fully specified by three numbers: mass, spin (angular momentum) and electric charge. Two black holes with the same three are identical, so there seem to be no look-alike configurations to count, hence zero entropy.
- Bekenstein's probe. Drop in a photon so spread out that the only information is "fell in or not": exactly one unit of entropy. Its energy raises the black hole's mass (E = mc²), enlarging the horizon by about one Planck area (~10−70 m²). Two units, two areas.
- The proposal. Black-hole entropy equals the area of its event horizon, measured in Planck units. Throwing tea in lowers entropy at your breakfast table, but the horizon grows enough to compensate: the second law survives.
- Puzzles left open. Why area, not volume (ordinary entropy lives in a volume)? And which microstates does it count? That waited until string theory in the mid-1990s.
Stephen Hawking thought the idea ludicrous: black holes, simple and mostly empty, seemed to have nothing to be disordered. Setting out to refute it with a delicate mix of general relativity and quantum mechanics, he found instead something so shocking that he took time to believe it: black holes have a temperature and they glow. They are black only if you ignore quantum physics.
How Hawking radiation works
- Quantum vacuum. Even "empty" space has energy fluctuating up and down, appearing as particle–antiparticle pairs that pop into existence and quickly annihilate (indirectly confirmed by the precision of quantum predictions).
- At the horizon. Occasionally one partner falls in and the other, with no partner to annihilate, escapes outward. All around the horizon, this looks like radiation.
- Shrinkage. The infalling partner carries negative energy, so the black hole loses mass, "eating negative calories."
- Second law. The horizon (and its entropy) shrinks, but the outgoing radiation spreads and carries more entropy away: the entropic two-step again.
- Temperature ∝ 1/mass. Big black holes are cool and calm ("Great Danes"); small ones hot and manic ("shih tzu puppies"). Precisely, T = ħc³ / (8πGMkB).
| Black hole | Mass | Hawking temperature |
|---|---|---|
| Milky Way centre (Sgr A*) | ~4 million suns | ~10−14 K |
| Stellar | 1 sun | ~10−7 K (just under a tenth of a millionth of a degree) |
| Moon-mass | ~moon | ~2.7 K, the CMB temperature (the crossover) |
| Orange-mass | ~0.2 kg | ~1024 K |
Any black hole heavier than the moon is colder than today's 2.7 K microwave background, so it absorbs more than it emits and is still growing; all black holes observed so far are far heavier, so all are "plumping up." Only when expansion cools the CMB below a black hole's temperature does it start to shrink. Don Page calculated evaporation times: sun-mass black holes vanish by floor 68 (1068 years). Their final moments (explosion? Fizzle?) remain unknown.
Area and information since 2020. In 2025 the gravitational-wave event GW250114 confirmed Hawking's area theorem: the merged black hole's horizon area exceeded the sum of the originals, as the black-hole second law demands. On the theoretical side, work from 2019–2020 (the "island" and "replica wormhole" calculations by Penington, Almheiri, Engelhardt, Marolf, Maxfield and others) showed how the entropy of Hawking radiation can follow the Page curve, rising and then falling as information escapes. That strengthens the case that black-hole evaporation preserves information, the question Greene's account of evaporation leaves implicit.
Gentle giants: the largest black holes
Galactic-centre black holes reach toward 100 billion solar masses, with horizons stretching past Neptune toward the Oort cloud (over 100 light-hours away). The recipe for a black hole is "gather mass and make it small enough," but the required density falls with mass: radius ∝ mass, volume ∝ mass³, so density ∝ 1/mass².
How small to squeeze it
- Grapefruit: to ~10−25 cm across.
- Earth: to ~2 cm.
- Sun: to ~6 km.
- 4 million suns (Milky Way centre): density about 100 times lead.
- 100 million suns: density of water.
- 4 billion suns: density of air; gather that much air and gravity alone makes a black hole, no squeezing needed.
Supermassive black holes are thus gentle giants, which is the intuition behind Hawking's result that they are coldest. With more mass and a slower glow, a 100-billion-solar-mass black hole lasts until floor 102, the top of the building, before its last burst of radiation.
An end of time
From floor 102 the universe is a thin mist of particles; now and then an electron and positron spiral together and annihilate in a pinprick of light. If dark energy drains away, particles might gather into still larger, longer-lived black holes; if it persists, accelerated expansion keeps them apart. It is "dust to dust": the early dust was dense enough for gravity to build stars; the late dust, too thin, "will be content to drift quietly through the void." Time doesn't stop, but physicists reasonably call this the end of time. Yet the improbable keeps its claim.
The disintegration of emptiness: the Higgs field
On 4 July 2012 at CERN, spokesperson Joe Incandela announced the Higgs particle. Greene watched at about 2 a.m. at the Aspen Center for Physics amid wild cheers; the camera cut to Peter Higgs wiping his eyes, nearly fifty years after his proposal.
The Higgs mechanism
- Puzzle. The equations of the strong, weak and electromagnetic forces need a symmetry to be consistent, and that symmetry demands massless particles, yet electrons and quarks resist being pushed: they have mass.
- Higgs's idea (on a long walk outside Edinburgh): particles are intrinsically massless, but space is filled with a field that drags on them, as air drags on a Wiffle ball held out of a speeding car. More drag, more mass.
- Reception. His first paper was rejected: "I was told I was talking nonsense." It gradually won over; Greene learned it in the 1980s so confidently taught he didn't realise it was untested.
- Test. Smash protons to "jiggle" the field and knock loose a droplet: a Higgs particle, a "chip off the old vacuum" (Frank Wilczek). The effort took 30+ years, 3,000+ scientists, three dozen countries and over $15 billion at the Large Hadron Collider, ending in a small bump on a smooth graph.
The future relevance: the Higgs field's value (246, in standard units) could change. Change it even slightly and particle masses change, and with them atoms, chemistry, the sun and cells: "you change more or less everything." Observations show the value has been stable for most of 13.8 billion years. But quantum tunnelling makes stability provisional.
Vacuum decay by quantum tunnelling (Sidney Coleman, mid-1970s)
- Tunnelling. A marble in a champagne flute stays put; an electron in a flute-shaped trap sometimes appears outside it. Schrödinger's equation gives a small but nonzero probability unless the walls are infinite, and experiments confirm escapes.
- Fields tunnel too. The Higgs value sits in a valley; it can tunnel through the barrier to a lower-energy value, first in one tiny random region.
- Bubble of doom. Neighbouring regions follow, domino-like; the bubble wall expands at nearly light speed. Inside: new particle masses, no familiar physics, chemistry or biology. Outside: business as usual, until it arrives. "One moment it would be life as usual. The next moment we would cease to be."
- When? Current data suggest between 10102 and 10359 years ("a range that would even challenge the reach of the Burj Khalifa"), though, being quantum, it could happen today.
- Meaning. Since even the emptiest space contains the Higgs field, this is an instability of empty space itself.
How firm is the Higgs doom? The metastability verdict hinges on the precise top-quark mass and on the assumption that no new physics enters between the energies we have probed and the Planck scale. Published lifetime estimates vary by hundreds of orders of magnitude (a widely cited 2018 calculation by Andreassen, Frost and Schwartz gave a central value near 10161 years). Greene's range is reasonable, but almost any new heavy particle could stabilise or destabilise the vacuum. Treat this as a live possibility, not a forecast.
Boltzmann brains
The second law is probabilistic: the air in your room could gather into a ball near the ceiling. Now Greene takes such entropy dips seriously.
Your sense of who you are (the mug from a New Mexico potter, your grandmother's chair, your curiosity about the universe, this book) exists only as the current arrangement of particles in your head. So if a random spray of particles in a high-entropy universe fluctuated into that exact arrangement, it would have your memories and thoughts. Such free-floating, spontaneously formed minds are Boltzmann brains. Larger fluctuations (a body, a planet, a whole universe or a new big bang) are possible but exponentially less likely; a lone brain is the cheapest fluctuation that can wonder how it came to be.
Why the far future makes Boltzmann brains likely
- Raw material. Gibbons and Hawking showed the cosmological horizon has a temperature (~10−30 K) and emits radiation, mostly photons and gravitons, forever.
- Fluctuations. Occasionally colliding particles convert into electrons, quarks, protons and neutrons; rarer still, these join into atoms; rarer still, into molecules and structures, "bobbleheads to Bentleys." Every so often, a brain. "Long extinct, thought will make a momentary comeback."
- Timescale. Roughly 101068 years: a 1 followed by 1068 zeros. Writing out 10102 takes a line and a half; writing 101068 would overflow every character of every book ever printed.
- Population. Ordinary brains can form only in a window (generously floors 10–40); Boltzmann brains form without limit. Over enough time they vastly outnumber ordinary brains, even counting only those that falsely believe they evolved.
The skeptical nightmare. By sheer numbers, the most likely explanation of your memories is that your brain just fluctuated into existence, with a fictitious past. Then no measurement, observation or calculation can be trusted, including the physics that predicted Boltzmann brains. The laws of physics, run long enough, undermine our confidence in the laws themselves. Even "I think, therefore I am" becomes "I think I thought, therefore I think I was."
How do physicists escape?
- "Look twice" (weak). A Boltzmann brain should vanish almost at once; if the world persists from glance to glance, you're probably not one. But your memory of earlier glances is exactly what a freshly fluctuated brain would also have.
- Dark energy fades. Without a permanent horizon, space cools toward zero and complex fluctuations effectively stop. Greene (2020): "no evidence for a weakening... The jury is still out."
- The Higgs tunnels first. If vacuum decay happens in 10102–10359 years, far sooner than 101068, ordinary brains dominate.
- Big crunch. If dark energy tunnels to a negative value, attractive gravity takes over and the universe collapses into a big crunch, again long before Boltzmann brains.
- Cyclic cosmology. See below.
DESI reopens the jury room. The 2024–25 DESI results (see chapter 9) are the first observational hints that dark energy is weakening, which, if confirmed, would undercut the permanent-horizon premise of the Boltzmann-brain problem. Some theorists went further: a 2025 analysis by Luu, Qiu and Henry Tye fitted the data with a dark energy that could turn negative, implying the expansion stops in about 7 billion years and a big crunch roughly 20 billion years from now. These are model-dependent and contested, but they show Greene's escape routes are no longer purely hypothetical.
The big bounce: cyclic cosmology
Paul Steinhardt, Neil Turok and Anna Ijjas turn a crunch into a bounce: regions of space expand, contract and rebound repeatedly, so the big bang is a big bounce. The idea goes back to Alexander Friedmann and Richard Tolman, who wanted to dodge the question of a beginning, but Tolman found the second law forbids it: entropy accumulating cycle after cycle means only finitely many cycles could precede ours.
| Inflation | Cyclic (bounce) cosmology |
|---|---|
| A brief burst of repulsive-gravity expansion at the start | Endless cycles; each expands far more than it contracts, diluting entropy in any finite region |
| Explains the CMB pattern | Also explains the CMB pattern |
| Predicts primordial gravitational waves possibly detectable in CMB polarisation | Predicts primordial gravitational waves too weak to see |
| Dominant: "the golden age of cosmology" | A main competitor; cycles last hundreds of billions of years, ending our climb near floors 11–12 |
A cyclic universe would replace the skyscraper with a spiral, like an endlessly soaring Guggenheim, each lap a cycle; it would also end each cycle long before Boltzmann brains form. And it would make birth, death and rebirth the cosmos's own pattern, as Hindu, Egyptian and Babylonian thinkers imagined.
The gravitational-wave test, so far. No primordial B-mode signal has been seen; BICEP/Keck (2021) limits the tensor-to-scalar ratio to r < 0.036. That squeezes simple inflation models without killing inflation, and is fully consistent with cyclic models. The Simons Observatory (operating from 2024–25) and Japan's planned LiteBIRD satellite aim to reach r ≈ 0.001–0.003, where many remaining inflation models predict a signal. A detection would strongly favour inflation; a continued null would hand the cyclic camp its best argument.
Thought and the multiverse
Is space infinite? Nobody knows, but the most-studied inflationary models say yes. We can see only ~45 billion light-years in any direction (more than 13.8 billion because space expanded while the light travelled), so an infinite space is a patchwork of independent ~90-billion-light-year regions: a multiverse of infinitely many "universes."
Garriga and Vilenkin's repetition argument
- Finite histories. Each region has finite size and energy. Nudge particles too little and quantum uncertainty makes the change meaningless; too much and they leave the region or exceed its energy. So only finitely many distinct histories ("films") are possible.
- Infinite regions. Infinitely many regions, finitely many films: every film repeats infinitely often, and random quantum jitters ensure every allowed film plays.
- Consequence. Our reality, and every variation not forbidden by law (energy and charge conservation), plays out over and over: Oswald misses, Stauffenberg succeeds, James Earl Ray doesn't. Unlike Many Worlds, these other worlds are not a matter of interpretation; if space is infinite, they're there.
- Lucky regions. For any finite duration, however long, some regions enjoy fortunate entropy dips that keep life going that long: somewhere, life and mind persist arbitrarily far into the future.
Inhabitants of such regions might conclude their physics was wrong; if Greene walked through walls repeatedly, we would revise quantum mechanics even though it technically allows it. Or they might just "go with the flow."
Could we send a message to them? Perhaps, as our end nears, we pack our knowledge and art into a capsule aimed at a luckier region, leaving "a trace on eternity." Garriga and Vilenkin, with insights from David Deutsch, say no: across infinite space and time, random fluctuations will produce far more fake capsules than we could make real ones, so any genuine record of us is lost in the quantum noise.
"Even though we can contemplate eternity, and even though we can reach for eternity, apparently we cannot touch eternity."
Where speculation takes over. This chapter moves from established physics (Hawking temperature, evaporation timescales given standard assumptions) to frontier speculation (vacuum decay rates, Boltzmann brains, infinite space, the measure problem). Many cosmologists treat the Boltzmann-brain argument as a reductio: any theory predicting that typical observers are Boltzmann brains is "cognitively unstable" and should be rejected (Sean Carroll's term), which is a methodological choice rather than a physical result. And the repetition argument depends on space being literally infinite and on how probabilities are compared across infinite collections (the unresolved measure problem), which Greene flagged back in chapter 3.
Rare events dominate over long horizons. The chapter's core lesson, that "improbable" times "enough time" equals "certain," is the foundation of operational-risk and model-risk thinking in banking. A one-in-a-million daily failure mode is a near-certainty over a decade of high-volume operations. Greene's Boltzmann-brain reductio has a practical cousin too: if a risk model, run over long horizons, implies conclusions that undermine the model's own data, the model, not reality, is what must be rejected.
The Nobility of Being
Greene closes by drawing the book's threads into a position. Nature is lawful and indifferent: particles and fields, laws and initial conditions, and no evidence for anything else. Values are human inventions, shaped by evolution and refined by thought. Imagining immortality shows that finitude is what gives choices weight; imagining humanity's extinction shows that value depends on descendants who will carry things on, and cosmic timescales do not soften that conclusion. There is no grand design and no final answer waiting in space; the only direction left to look is inward, where special collections of particles can create purpose. That is the noble direction.
The tarantula at breakfast
At Pilanesberg National Park in South Africa, a rifle-carrying guide briefed a walking-safari group: if a lion comes close, stay still; "Run from a lion? You spend the rest of your life trying to win the race." Then Greene noticed something crawling up his sleeve ("to me it was a tarantula"), flailed, sent glasses and plates flying off the breakfast table, and watched the nickel-sized creature crawl away. The guide: "The universe has spoken for our physicist friend. You travel in the Jeep." He did.
The universe hadn't spoken; it was chance. Yet because Greene was already uneasy about walking among lions, the incident felt significant, and "rational thought was a step or two from regaining full command." The episode distils a book-long theme: we detect patterns, note coincidences and assign significance, and only some of those assignments track reality; many satisfy an emotional need to impose order on chaos.
Lawful, but are the equations nature's own language?
Greene often talks as if equations are "out there," controlling everything from quarks to the cosmos. Perhaps mathematics is woven into reality. What he is surer of is that nature is lawful; our equations are our most precise statement of the laws, spectacularly accurate, but not guaranteed to be nature's intrinsic lexicon. Aliens might smile politely at our mathematics and say they started there too, before discovering "the real language of reality." Evolution gave us intuitive physics (falling rocks, snapping branches, rushing streams); curiosity and mathematics carried us far beyond anything adaptive, into domains from particles to galaxy clusters. The resulting equations may be constructs of the human mind.
Values are invented, not from whole cloth
Greene takes the same stance on right and wrong, good and evil, destiny, purpose, value and meaning: they are profoundly useful, but they do not transcend the human mind. We invent them, though not arbitrarily. Darwinian minds come predisposed: care for the young rates highly everywhere; incest is widely abhorred; fairness and loyalty to kin and compatriots are valued. As groups formed, these predispositions interacted with experience in feedback loops, yielding communal codes that themselves conferred differing survival value. Natural selection shaped our moral sense as it shaped our intuitive physics.
| Reason as engine | Reason as press secretary |
|---|---|
| Evolution gave a rudimentary moral sense; cognition lets us leap beyond it to independent attitudes | We use cognition mainly to explain moral commitments after the fact: just-so rationalisations of judgments anchored in evolution |
| Parallels how physics outgrew evolved intuition | Moral reasoning follows moral intuition |
Neither view requires free will in the traditional sense. Wishes, judgments and moral appraisals are part of the world's physical progression, not interventions in it.
The story so far, in one breath
13.8 billion years ago: a tiny ordered patch of inflaton field disintegrates, ending repulsive gravity, filling space with particles and seeding the simplest nuclei.
Next hundreds of millions of years: where quantum uncertainty left slightly higher density, gravity pulls matter into stars, planets and moons.
Billions of years: stellar fusion and collisions forge complex atoms, which rain onto at least one young planet.
~4 billion years ago: molecular Darwinism assembles self-replicating arrangements; variants that boost replication spread, including pathways for information and energy.
Since then: Darwinian evolution refines life into complex, self-directed beings; thought, self-awareness and language follow.
Particles and fields. Physical laws and initial conditions. "To the depth of reality we have so far plumbed, there is no evidence for anything else." The laws are probabilistic (quantum mechanics) but the probabilities are rigidly mathematical, and life cannot overrule them. What life can do is let particles act in concert: "marigolds grow larger and follow the sun while marbles don't." Selection favoured thought (memory, analysis, extrapolation), and thought refined over tens of thousands of generations into self-aware species with a vast repertoire of responses.
Add language, and one species sees itself as part of an unfolding from past to future. "We are no longer satisfied to merely survive. We want to know why survival is significant." Hence cosmologies, stories, imagined worlds of ancestors and gods, art, music, religion, whether adaptive (stories prepare for the unexpected; art builds imagination; music sharpens pattern sense; religion binds coalitions) or riding along with traits that were. Either way they reveal a longing to be part of something larger and lasting. "Value and meaning, decidedly absent from the bedrock of reality, become intrinsic to a restless urge that elevates us above indifferent nature."
Immortality: would we want it?
Leibniz asked why there is something rather than nothing; the personal version is that self-aware somethings dissolve into nothing. Across the timeline, stars, black holes, planets, people and atoms all fall apart. Our deaths, humanity's end and the end of life and mind are "expected, run-of-the-mill, long-term outcomes of physical law. The only novelty is that we notice."
| Source | Immortality portrayed as... |
|---|---|
| Jonathan Swift (Luggnagg's Struldbruggs, Gulliver's Travels) | Endless ageing; declared legally dead at 80, drifting into irrelevance |
| Karel Čapek (The Makropulos Affair) | Elina Makropulos, after 300+ years, lets the elixir's formula burn rather than endure more boredom |
| Jorge Luis Borges ("The Immortal") | In endless time "no one is anyone"; to be everything is to not exist |
| Bernard Williams (after Janáček's opera of Čapek) | Every desire eventually satiated; eternity becomes mind-numbing monotony |
| Aaron Smuts (after Borges) | Choices lose consequences (endless do-overs); achievement loses meaning (limited people stall, unlimited ones improve inevitably) |
| Greene | Resourceful immortals would transform their needs and invent new joys; assuming boredom is "an unduly parochial vision of the immortal mind" |
Science will lengthen lives, but immortality stays out of reach. Still, imagining it clarifies mortality: scarcity of time drives our calculus of value. "Not that we spring to our feet each morning wailing 'Carpe diem!'", but our choices of study, work, risk, partners, family and goals all reflect that opportunities are scarce because time is limited.
Descendants: the hidden anchor of value
"A year to live, or a year until earth is destroyed?"
- Setting. Years ago, Greene joined a post-show talkback at an off-Broadway play about an asteroid about to destroy earth, alongside his brother (science versus religion, the producers hoped). As his brother veered toward the ethereal, Greene got blunt: earth is a pedestrian planet around an unremarkable star; if an asteroid hits, "the universe won't so much as blink."
- Challenge. An elderly woman called this smug and asked: which news would affect you more, a year to live or earth's destruction in a year?
- Greene's later answer. His own terminal diagnosis would intensify life: focus, perspective, significance in daily moments. Humanity's end would make everything feel pointless: physics research with no one to build on it, a book with no readers, children schooled for no future.
- Insight. He had pictured the future as a land of equations and theorems. But equations are "lines and squiggles"; "their value derives from those who understand and appreciate them. Their worth derives from the minds they inhabit." Individual immortality may sap significance; the species' continuation seems necessary to secure it.
Philosopher Samuel Scheffler formalised the question: how would you feel if everyone were obliterated thirty days after your own death? (This removes one's own early death from the equation.) His conclusion: our commitments and sense of what matters presuppose that human life is "a thriving, ongoing enterprise"; we need humanity to have a future "for the very idea that things matter" to remain secure.
Samuel Scheffler: the collective afterlife; value depends on humanity's continuation. Susan Wolf: shared doom might heighten care for others, but agrees a populated future underpins the value of our projects. Harry Frankfurt: dissents; many pursuits, notably art and science, would continue for intrinsic gratification. Greene: on science, contra Frankfurt; to do research is to join "a long, rich, and ongoing human dialogue." If he were the last to speak, with no one to reflect on it, "why should I bother?"
Does distance in time change anything?
In Annie Hall, nine-year-old Alvy Singer won't do his homework because in a few billion years the expanding universe will break apart. We laugh. Scheffler shares the laugh but admits he can't justify why imminent doom should cause existential crisis and remote doom should not; he blames our inability to grasp vast timescales. Wolf agrees: if immediate extinction would drain meaning, so should distant extinction; on cosmic scales a few billion years is short.
Greene: "I agree. Forcefully so." "Long" and "short" are matters of perspective: the floor-86 span compared to floor 100 is like the blink of an eye against ten thousand centuries. The imminent-doom scenario is just a tool that uses artificial urgency to catalyse an authentic response, and the response applies to the far-future end too: viewed from a larger context, that future is a moment away.
A meditation on the timeline. Let your mind float past our epoch, past receding galaxies, stately solar systems, swirling galaxies, burnt-out stars and wandering planets, glowing and evaporating black holes, into a cold, dark, nearly empty expanse where the only evidence we existed is a particle here instead of there. The vastness doesn't soften the "shuddering yet awestruck" feeling; it intensifies the "nearly unbearable lightness of being." From the first microbes to the final thought, life and mind would occupy less than the time light takes to cross an atomic nucleus. We are ephemeral. We are evanescent.
Yet that rarity is the basis for value and a foundation for gratitude. A lasting legacy is out of reach, but the same clarity shows how wondrous it is that a few of the universe's particles can rise up, examine themselves and reality, grasp their own transience, and in a brief burst create beauty, establish connection and illuminate mystery.
Meaning
Most of us cope quietly, letting civilisation shield us from the fact that the world will hum along without us: we build community, care, laugh, grieve, love, celebrate, consecrate, regret and thrill to achievement. We habitually look outward for something to excite or soothe us. But the scientific journey shows the universe does not exist as an arena for life and mind: they are "a couple of things that happen to happen. Until they don't." Greene once hoped that answering enough how questions would reveal the whys; the more we learn, the more that stance "seems to face in the wrong direction." "Looking for the universe to hug us, its transient conscious squatters, is understandable, but that's just not what the universe does."
Why our existence is astonishing (Greene's three improbabilities)
- Cosmic. Rerun the big bang with one particle shifted or one field value tweaked and almost every alternative history lacks earth, humanity and you, though a super-intelligence viewing it coarsely, as we view pennies or air, would say the new universe "pretty much looks the same." Entropy ignores fine detail; we are made of the fine details that won.
- Genetic. Richard Dawkins's point: the nearly infinite set of possible DNA sequences means a nearly infinite crowd of potential people who will never be born. You are one of the few who were.
- Quantum. At countless junctures since the big bang, probabilistic processes could have gone otherwise, yielding a sensible universe without you. "How spectacularly unlikely. How thrillingly magnificent."
And the gift is larger: our arrangements let us step outside time, imagining past and future, exploring with reason and emotion, expressing longing and revelation in words, images and sound, and discovering the laws that govern how stars shine, how light travels, how time passes and space expands, enough to glimpse the universe's first moment and contemplate its end.
Why something rather than nothing? How did life start? How did consciousness arise? Perhaps brains built for survival can't answer; perhaps evolving intelligence will make the questions moot. Greene rejects both: the tight mathematical and logical coherence of the world as understood, and how much of it we have deciphered, suggest "we are not lacking the brainpower. We are not staring at Plato's wall."
"As we hurtle toward a cold and barren cosmos, we must accept that there is no grand design." Brian Greene
Particles are not endowed with purpose. There is no final answer hovering in the depths of space. But certain collections of particles can think, feel and reflect, and within their subjective worlds create purpose. So the only direction to look in understanding the human condition is inward: "That is the noble direction to look." It forgoes ready-made answers for the personal work of constructing meaning, which is the source of creative expression and our most resonant stories. Science is a superb tool for grasping external reality; within that understanding, everything else is "the human species contemplating itself, grasping what it needs to carry on, and telling a story that reverberates into the darkness," a story that, when it works, stirs the soul.
Critical lens on the ending. Three of Greene's closing claims deserve scrutiny. (1) "No evidence for anything else" beyond particles, fields and laws is a defensible methodological naturalism, but it settles nothing about consciousness, which chapter 5 conceded is unexplained; the confident tone outruns the evidence there. (2) "We are not lacking the brainpower" directly contests the "mysterian" position (Colin McGinn, and Chomsky's distinction between problems and mysteries); Greene offers an inference from past success, not an argument. (3) "Values are invented" is one side of a live metaethical debate; moral realists (Derek Parfit, Thomas Nagel) argue that some moral truths are discovered, not made, and Greene does not engage them.
Scheffler meets longtermism. Since 2020, the "collective afterlife" intuition Greene draws on has become the foundation of longtermism (William MacAskill's What We Owe the Future, 2022): if the value of what we do depends on humanity's continuation, then reducing existential risk (pandemics, nuclear war, misaligned AI) may be among the most important things we can do. Critics argue it licenses neglect of present suffering. Greene's chapter supplies the emotional premise of that debate, without its policy conclusions.
Meaning is an inside job, and so is purpose in organisations. Greene's last move, from looking outward for meaning to constructing it inward, has a sober corollary for leaders: no market, regulator or technology confers purpose on an institution. In AI transformation programmes in banking, purpose has to be authored (what this capability is for, whose lives it improves) and then carried by people, because, as Greene says of equations, its worth "derives from the minds" that hold it.
Master Takeaways
The ideas worth carrying out of Until the End of Time, ranked with the strongest first, followed by the book's toolkit of frameworks, a critical verdict on the whole, a who's who, and the links to AI, architecture and banking that genuinely hold.
Master takeaways
- The arrow of time is an accident of the beginning. The laws of physics run equally well forwards and backwards; the past differs from the future only because the universe started in an extraordinarily low-entropy state, and entropy has been rising ever since.
- Order is built by exporting disorder. The entropic two-step (lower entropy here, dump more there, usually as waste heat) explains steam engines, stars, cells and brains. Structure is not a defiance of the second law but one of its most efficient instruments.
- Gravity runs thermodynamics backwards. A self-gravitating cloud gets hotter as it loses heat, a self-amplifying loop that fusion eventually arrests: stars form spontaneously, with no designer needed.
- All life is one machine. A single genetic code and a single energy system (redox cascade, proton battery, ATP turbine) run every organism on earth, overwhelming evidence of one common ancestor and of life as "physics orchestrated."
- Freedom is range, not exemption. Deterministic or probabilistic, physical law leaves no room for a will that overrides it; human freedom is the vast repertoire of responses our organisation allows, and responsibility attaches to the particle configuration that is you.
- Thought itself may have an expiry date. If dark energy stays constant, the cosmic horizon's permanent ~10−30 K glow eventually stops any thinker shedding waste heat; the last thought could come by about 1050 years. Only a fading dark energy (now weakly hinted at by DESI) reopens the door.
- Low entropy is a clue. Wherever you find improbable order (an egg, an anthill, a mug), look for a powerful organising process; chance is almost never the answer.
- Energy quality, not quantity, is what matters. Earth radiates away as much energy as it receives; it lives off the difference in entropy between few hot solar photons in and ~20 cold infrared photons out.
- Evolution starts before life. Replication plus variation plus scarcity yields molecular Darwinism; driven systems may even drift toward replicators by dissipative adaptation.
- Information is physical. Molecules don't "know" anything; information is encoded in shapes that steer other molecules, as a water slide steers riders without overriding gravity.
- The hard problem may be a self-model artefact. If Graziano is right, awareness feels ethereal because the brain's model of its own attention omits the machinery; consciousness would be demystified without being diminished.
- Consciousness does not collapse wave functions. Interactions, not minds, produce definite outcomes; quantum mechanics needs no special rules for observers.
- Stories are flight simulators. Fiction rehearses trouble and lets us inhabit other minds; no conflict, no story and no difficulty, no adaptive value are the same rule.
- Myth runs on three cognitive engines: combinatorial language, minimally counterintuitive concepts, and over-active agency detection, all pointed at death and origins.
- Religion's origins remain unresolved. By-product, social glue, invisible watcher and death buffer all have support and problems (and the terror-management experiments have failed large replications).
- Separate belief in fact from belief in value. Literal claims contradicting physics are false; tradition, ritual and heritage can still carry history and connection, as Kaddish did for Greene.
- Calibrate confidence. Physicists demand five-sigma evidence and update by rule; brains evolved to favour useful beliefs over true ones, so deliberate calibration is a learned discipline.
- Art yields a different truth. It makes us "feel a thought," reaching experience that facts and equations do not; its evolutionary payoff may be flexible, innovative thinking plus group cohesion.
- Symbolic immortality is the only kind on offer, and it is bounded by civilisation's lifespan.
- The far future erases even knowledge. Accelerated expansion will push the evidence for the big bang beyond the horizon; future astronomers may wrongly conclude the universe is static.
- Forgetting, not computing, costs energy. Reversible computation can be free; erasure produces heat, which ultimately limits any thinker.
- Black holes have entropy proportional to horizon area, glow with a temperature inversely proportional to mass, and evaporate: stellar ones by 1068 years, the largest by 10102.
- Endless time breeds absurdity. Boltzmann brains would eventually outnumber evolved minds, undermining all knowledge; sane cosmology needs the universe to change or end first.
- Imminent and remote extinction should weigh the same. Our values presuppose descendants; cosmic distance does not cancel that dependence, it only hides it.
- There is no grand design, and that is the opening, not the end. Meaning is not found in the cosmos; it is constructed inward by minds that know they are brief.
Frameworks and mental models
| Name | What it says | Use it when |
|---|---|---|
| Entropy as counting (Boltzmann) | Entropy = log of the number of microstates that look the same macroscopically | Judging whether an outcome is typical or needs explaining |
| Entropic two-step | Local order is bought by exporting a larger entropy increase | Analysing any system that maintains structure: firms, platforms, organisms |
| Energy quality | First law conserves quantity; second law degrades quality | Valuing energy sources, or any resource that dissipates with use |
| Past hypothesis | Entropy was anchored low at the big bang | Explaining why records and memories point to the past |
| Gravitational negative heat capacity | Losing heat makes a self-gravitating core hotter | Spotting self-amplifying loops and the brake that stops them |
| Nested naturalism | One reductionist base; many higher-level stories, each best for its questions | Choosing the right level of explanation |
| Molecular Darwinism / dissipative adaptation | Selection and driven physics can produce replicators before life | Thinking about origins of complex systems |
| Hard problem / Mary's room | Physical facts may leave out experience | Testing claims that a system "understands" or "feels" |
| Attention schema (Graziano) | Awareness is the brain's simplified model of its own attention | Explaining why self-reports of awareness can mislead |
| Freedom as behavioural range | Freedom is the richness of possible responses, not exemption from law | Discussing agency and responsibility without free will |
| Story as flight simulator (Oatley) | Fiction trains intuition through simulated conflict | Designing training, scenarios and change narratives |
| Minimally counterintuitive concepts (Boyer) | Ideas that break one or two expectations spread best | Understanding why some ideas go viral |
| Belief as confidence vs. Value | Factual belief needs evidence; value belief needs meaning | Separating empirical claims from identity commitments |
| Empire State Building of time | Floor N = 10N years | Grasping exponential timescales |
| Dyson's Thinker | Thought requires shedding waste heat to a colder environment | Reasoning about long-run limits on computation |
| Collective afterlife (Scheffler) | Our values presuppose humanity's continuation | Thinking about legacy, risk and long-term stewardship |
Critical lens on the whole book. Greene is at his strongest where his expertise lies: chapters 2, 3, 9 and 10 give some of the clearest popular accounts of entropy, inflation and cosmic futures ever written, and he is scrupulous about flagging assumptions (the measure problem, the past hypothesis, proton decay, dark energy's behaviour). The middle chapters on language, religion and art are careful syntheses of others' work, fair-minded but secondhand, and they lean on evolutionary-psychology stories that he himself admits are hard to test; one pillar, terror management's mortality-salience effect, has since failed large replications. The book's physicalism is a bet, honestly labelled, but the closing confidence ("no evidence for anything else," "we are not lacking the brainpower") outruns what chapter 5 established. And events since 2020 have shifted several premises: DESI's hints of weakening dark energy could undo the permanent cosmic horizon on which both the "last thought" and the Boltzmann-brain arguments rest, while the simplest inflation models are now disfavoured. None of this dents the book's central achievement: placing human meaning-making squarely and humanely inside a lawful, impermanent cosmos.
Who's who
| Person | Role | Why they matter here |
|---|---|---|
| Ludwig Boltzmann | Physicist | Statistical definition of entropy; second law as probability |
| Sadi Carnot | Engineer | Founded thermodynamics with Reflections on the Motive Power of Fire (1824) |
| Bertrand Russell | Philosopher | "Universal death" as an argument against cosmic purpose |
| Georges Lemaître | Priest, physicist | First expanding-universe cosmology |
| Alan Guth, Andrei Linde | Cosmologists | Inflation: repulsive gravity drives the bang |
| Erwin Schrödinger | Physicist | What Is Life?; organisms "feed on negative entropy" |
| Watson, Crick, Wilkins, Franklin | Molecular biologists | DNA's double helix and copying mechanism |
| Peter Mitchell | Biochemist | Proton-driven ATP synthesis (chemiosmosis) |
| Ilya Prigogine, Jeremy England | Physical chemists | Dissipative structures; dissipative adaptation |
| David Chalmers | Philosopher | Named the hard problem; panpsychist proposal |
| Giulio Tononi, Christof Koch | Neuroscientists | Integrated information theory (Φ) |
| Michael Graziano | Neuroscientist | Attention schema theory, Greene's favoured view |
| Frank Jackson | Philosopher | Mary's room (and later recanted its conclusion) |
| Daniel Dennett | Philosopher | Compatibilist freedom; physicalist accounts of mind |
| Noam Chomsky, Steven Pinker | Linguist, psychologist | Innate grammar; language and art as adaptation or by-product |
| Robin Dunbar, Daniel Dor | Psychologist, linguist | Language as gossip-grooming; as guided imagination |
| Pascal Boyer | Anthropologist | Religion as by-product of inference systems; minimally counterintuitive ideas |
| David Sloan Wilson, Jesse Bering | Biologist, psychologist | Religion as group glue; the invisible watcher |
| Ernest Becker, Otto Rank | Anthropologist, psychoanalyst | Death denial; art as symbolic immortality |
| Ellen Dissanayake, Brian Boyd | Anthropologist, literary scholar | Art as bonding and cognitive training |
| Freeman Dyson | Physicist | The far future of thought; hibernating Thinkers |
| Jacob Bekenstein, Stephen Hawking | Physicists | Black-hole entropy and radiation |
| Howard Georgi, Sheldon Glashow | Physicists | Grand unification and proton decay |
| Peter Higgs, Sidney Coleman | Physicists | The Higgs field; vacuum decay by tunnelling |
| Steinhardt, Turok, Ijjas | Cosmologists | Cyclic (bouncing) universe |
| Garriga, Vilenkin | Cosmologists | Repeating histories in an infinite multiverse |
| Samuel Scheffler, Susan Wolf | Philosophers | Value depends on humanity's future |
Connect it
- Entropy budgets for AI. Greene's Thinker problem is the data-centre problem in miniature: compute is bounded by heat rejection and energy quality, and Landauer's principle sets the ultimate floor. Capacity plans that ignore cooling are planning against physics.
- Two-step architecture. Every orderly estate (clean data, current models, tidy integrations) is maintained only by continuously exporting disorder through engineering effort. Budget for the export, or the order decays.
- Calibrated confidence in model risk. Greene's split between evidence-calibrated belief and value-driven belief maps onto model validation versus adoption narrative; banks need both and must not confuse them.
- Self-report is not understanding. The attention-schema lesson for copilots and agents: a system can model itself as "aware" and say so fluently; governance should test behaviour, not testimony.
- Long horizons make rare events certain. Operational resilience in banking is Greene's quantum-eternity principle at human scale: the improbable, given enough volume and time, will happen, so design for it.
About this book
Until the End of Time: Mind, Matter, and Our Search for Meaning in an Evolving Universe by Brian Greene, 2020. This is a chapter-by-chapter study of the book, with summaries, maps, tables and diagrams; short quotations keep their page references. It is shared free for personal study, is not for sale, and is not endorsed by the author or publisher. Please read and buy the original.
Insight Editions · rajeshrmahapatra.com/insight-editions