Insight Editions · No. 226
The Elegant Universe
Superstrings, Hidden Dimensions, and the Quest for the Ultimate Theory
Brian Greene
Brian Greene explains why general relativity and quantum mechanics clash at the smallest scales and how superstring theory, with its extra dimensions, dualities and M-theory, proposes to reconcile them.
A study guide to The Elegant Universe 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's 1999 classic argues that the two pillars of modern physics, general relativity and quantum mechanics, are incompatible at the smallest scales, and that superstring theory reconciles them by replacing point particles with tiny vibrating strings. Along the way it explains why strings need extra dimensions curled into Calabi-Yau shapes, how string theory rewrites geometry (T-duality, mirror symmetry, space-tearing transitions), how five string theories merge into M-theory, and how strings account for black-hole entropy. The theory is beautiful, unifying and mathematically fertile, but as Greene concedes and twenty-seven years of data confirm, it has not yet made a prediction that experiment has tested.
Everything, matter and forces alike, may be one kind of string vibrating in different patterns in a universe of ten or eleven dimensions; this is the first framework that joins gravity and quantum mechanics without infinities.
From special relativity to M-theory: 100 years of physics, the first and second superstring revolutions, black holes and cosmology; plus an authored 2026 update chapter.
Planck length 10−33 cm; string tension ~1039 tons; 5 string theories, 1 M-theory, 11 dimensions; the LHC is still ~15 orders of magnitude short of the Planck energy.
Readers who want the logic of modern physics, not just its slogans; anyone weighing the promise and the critique of string theory.
The argument in eight moves
- Two great theories, one conflict. Special and general relativity (Ch 2–3) make space and time flexible and curved; quantum mechanics (Ch 4) makes the micro-world probabilistic and jittery. Combined, the jitters make spacetime itself froth below the Planck length, and calculations return infinity (Ch 5).
- Replace points with strings. If every particle is a tiny loop of string, each particle is a vibration pattern: mass is vibrational energy, charges are finer features, and one pattern is the graviton. Because strings have size, they smear out the foam and the infinities vanish (Ch 6).
- Supersymmetry comes with it. Only the supersymmetric string has matter and no tachyon, so string theory predicts superpartners for every particle; supersymmetry also makes the forces' strengths meet at high energy (Ch 7).
- Strings need extra dimensions. Quantum consistency requires nine space dimensions (later ten); six (or seven) are curled into tiny Calabi-Yau shapes, whose geometry fixes particle masses and charges (Ch 8). In principle the number of particle families could equal the number of holes in the shape, but no one can yet pick the shape (Ch 9).
- Geometry turns quantum. A circle of radius R and one of radius 1/R give the same physics, so space has a minimum size; differently shaped Calabi-Yaus can be mirrors with identical physics; and space can tear and repair without catastrophe (Ch 10–11).
- Five theories are one. Using dualities, Witten (1995) showed the five string theories and eleven-dimensional supergravity are limits of a single M-theory, with eleven dimensions and branes of many dimensions (Ch 12).
- Black holes are made of strings. A collapsing three-sphere turns a black hole into a massless particle; Strominger and Vafa counted the microstates of certain black holes and reproduced the Bekenstein–Hawking entropy exactly (Ch 13). String theory may also replace the big bang's infinities with a Planck-sized nugget (Ch 14).
- Open questions, honestly stated. What principle underlies the theory? Can spacetime emerge? Must quantum mechanics change? Can it be tested? Are there limits to explanation (Ch 15)? Our authored Chapter 16 reports what happened next: no superpartners, a landscape of solutions, the swampland, and the rise of holography.
Chapter map
| Ch | Title | One-line payload |
|---|---|---|
| 1 | Tied Up with String | Preface plus the plan: relativity and quantum mechanics clash; strings may unify everything |
| 2 | Space, Time, and the Eye of the Beholder | Constant light speed makes time slow and lengths shrink; E = mc² |
| 3 | Of Warps and Ripples | Gravity is curved spacetime; light bends, clocks slow, black holes, expansion |
| 4 | Microscopic Weirdness | Quanta, wave-particle duality, probability waves, uncertainty, sum over paths |
| 5 | The Need for a New Theory | The standard model; gauge symmetry; quantum foam and infinities below the Planck length |
| 6 | Nothing but Music | Strings as vibration patterns; Planck tension; strings smear out interactions |
| 7 | The "Super" in Superstrings | Spin, supersymmetry, running couplings; five superstring theories |
| 8 | More Dimensions Than Meet the Eye | Kaluza-Klein; strings need 10 dimensions; Calabi-Yau shapes fix physics |
| 9 | The Smoking Gun | Families from holes; superpartners, fractional charges, cosmic strings |
| 10 | Quantum Geometry | Winding modes, R ↔ 1/R, minimum size, mirror symmetry, "Physics Wins!" |
| 11 | Tearing the Fabric of Space | Flop transitions: space can tear; the 8.999999 story; Witten's shielding strings |
| 12 | Beyond Strings | Perturbation theory, dualities, BPS states, M-theory, 11 dimensions, branes |
| 13 | Black Holes | Conifolds, massless black holes, Hawking radiation, Strominger–Vafa entropy |
| 14 | Reflections on Cosmology | Big bang, inflation, Brandenberger–Vafa, pre-big bang, the multiverse |
| 15 | Prospects | Five questions: principle, emergent spacetime, new quantum theory, tests, limits |
| 16 | Strings Since 1999 (authored) | LHC null results, the landscape, the swampland, holography, maths, the scorecard |
How to read this edition. Greene's text dates from 1999 (with light 2003 revisions). Every chapter carries UPDATE SINCE 1999 boxes with results through October 2026, and CRITICAL LENS boxes where claims are contested, outdated or overstated; nothing has been silently corrected. Chapter 16 is not Greene's: it was written for this edition. Greene's own figures were not included in the source file, so every diagram here is newly drawn.
Tied Up with String
Modern physics stands on two pillars that cannot both be right as currently formulated: general relativity for the very large and quantum mechanics for the very small. Greene frames string theory as the resolution of the third great conflict of twentieth-century physics, after special and general relativity resolved the first two. The single idea that every particle is a tiny vibrating loop of string promises both to reconcile gravity with quantum mechanics and to explain all matter and forces as notes played by the same strings.
Einstein's unfinished dream
For the last thirty years of his life Einstein searched for a unified field theory, one framework for all of nature's forces. He was driven less by unexplained data than by the conviction that the deepest understanding would reveal the simplicity and power of the underlying principles. He failed, partly because the strong and weak forces and much of particle physics were unknown or poorly understood in his time. In the early 1940s he described himself as "a lonely old chap who is mainly known because he doesn't wear socks."
Greene's preface sets out the book's purpose and route. Superstring theory is presented as a framework that may, in principle, describe all physical phenomena. He approaches it through the evolving understanding of space and time, from Einstein and Heisenberg to the 1990s, with analogies rather than equations, flagging the harder sections of Part IV so they can be skimmed. He also warns that his first-hand accounts of discoveries he took part in may exaggerate his role in a field built by more than a thousand physicists.
The quiet crisis
For half a century physicists have been "quietly aware of a dark cloud." Each pillar has been confirmed to extraordinary accuracy, yet as formulated, general relativity and quantum mechanics are mutually incompatible. Most of the time this doesn't matter: physicists study either small, light things (atoms) or big, heavy things (stars), so they can use one theory and ignore the other. But the centre of a black hole (huge mass, tiny size) and the big bang (the whole universe in a microscopic nugget) need both, and combining the equations makes them "shake, rattle, and gush with steam like a red-lined automobile": well-posed questions get nonsensical answers.
String theory's first claim is that it resolves the conflict; its stronger claim is that in string theory, general relativity and quantum mechanics require each other for the theory to make sense. The marriage of large and small is not merely possible but inevitable.
The three conflicts
Greene's organising arc: the clash between gravity and quantum mechanics is the third in a series, and each resolution overturned our picture of space and time.
| Conflict | Clash | Resolution | What changed |
|---|---|---|---|
| 1. Late 1800s | Newton: run fast enough and you catch a light beam. Maxwell: you can't | Special relativity (1905) | Space and time depend on your state of motion (Ch 2) |
| 2. After 1905 | Special relativity forbids faster-than-light influence; Newton's gravity acts instantaneously | General relativity (1915) | Space and time warp and curve; curvature is gravity (Ch 3) |
| 3. From 1900s | Smooth curved geometry of GR versus the frantic, roiling quantum microworld | String theory (mid-1980s onward) | Extra curled-up dimensions; strings replace points (Ch 5 onward) |
"String theory, in a real sense, is the story of space and time since Einstein."
The universe at its smallest: matter
The Greeks guessed that matter is built from a few "uncuttable" ingredients, as words are built from letters. The name "atom" stuck to things that turned out cuttable: by the early 1930s J. J. Thomson, Ernest Rutherford, Niels Bohr and James Chadwick had established nucleus (protons, neutrons) plus electrons. In 1968, experiments at the Stanford Linear Accelerator Center (SLAC) showed protons and neutrons are made of quarks (named by Murray Gell-Mann from Finnegans Wake): a proton is two up quarks and a down; a neutron two downs and an up.
Early 1930s: Wolfgang Pauli predicts the neutrino.
Late 1930s: the muon, a 200-times-heavier electron, found in cosmic rays. I. I. Rabi: "Who ordered that?"
Mid-1950s: Frederick Reines and Clyde Cowan detect the neutrino; an average one passes through trillions of miles of lead unaffected, and billions from the sun pass through you every second.
1968: quarks seen inside protons at SLAC.
Later decades: charm, strange, bottom and top quarks; the tau; muon- and tau-neutrinos; every particle with an antiparticle (the positron has the electron's mass and charge +1).
Everything ever encountered, natural or made in accelerators, down to about 10−18 metres, is built from three families of matter particles:
| Family | Charged lepton | Neutrino | Up-type quark | Down-type quark |
|---|---|---|---|---|
| 1 | Electron 0.000511 GeV | Electron-neutrino (tiny) | Up ~0.002 GeV | Down ~0.005 GeV |
| 2 | Muon 0.106 GeV | Muon-neutrino (tiny) | Charm 1.27 GeV | Strange 0.093 GeV |
| 3 | Tau 1.78 GeV | Tau-neutrino (tiny) | Top ~172.6 GeV | Bottom 4.18 GeV |
(Masses in GeV; a proton is 0.938 GeV. Greene's 1999 table listed upper limits for neutrino masses; see the update box below.)
Corresponding particles across families are identical except for mass, which rises family by family. So the questions pile up: why so many particles when everyday matter needs only electrons, up and down quarks? Why three families? Why these masses: why is the tau about 3,520 times the electron, and the top about 40,200 times the up quark? Chance, divine choice, or a comprehensible explanation?
The particle table since 1999. Neutrinos have mass: oscillation experiments (Super-Kamiokande 1998, SNO 2001; Nobel Prize 2015) showed they change flavour, which requires mass; KATRIN (2025) caps the electron-neutrino mass below ~0.45 eV, about a millionth of the electron's. And the table gained its missing piece: the Higgs boson, discovered at CERN in 2012 at ~125 GeV (about 133 proton masses), the particle whose field gives the others their mass. No fourth family has appeared, and no particle outside this table has been found at the LHC.
The forces: where's the photon?
All the pushes, pulls, explosions and magnetic levitations of daily life reduce to four forces: gravity, electromagnetism, the weak force and the strong force. Mass sets how strongly a particle feels gravity; electric charge does the same for electromagnetism; particles also carry "strong charge" and "weak charge." The strong and weak forces act only over subatomic distances, which is why they were found late: the strong force glues quarks into protons and neutrons and holds nuclei together; the weak force drives radioactive decay (uranium, cobalt).
Each force comes in smallest bundles, its messenger particle:
| Force | Messenger | Mass (GeV) | Status (1999) |
|---|---|---|---|
| Electromagnetism | Photon | 0 | Established |
| Weak | W and Z bosons ("weak gauge bosons") | 80.4, 91.2 | Established by 1984 |
| Strong | Gluons (eight types) | 0 | Established |
| Gravity | Graviton | 0 (predicted) | Not detected |
The strength puzzle
- Electromagnetic repulsion vs. Gravity between two electrons: ~1042 times stronger. If your right bicep stood for gravity, your left would have to reach beyond the edge of the known universe to stand for electromagnetism.
- Strong vs. Electromagnetic: about 100 times stronger.
- Strong vs. Weak: about 100,000 times stronger.
- Why gravity still dominates big things: matter is electrically neutral, so electromagnetic forces cancel; gravity is always attractive, so it adds up.
The four forces have very different ranges and strengths, and the puzzle is not idle: change the numbers moderately and the universe changes drastically.
Fine-tuning examples Greene gives
- Strong vs. Electromagnetic balance: a small shift and most nuclei fall apart; no periodic table.
- Electron mass a few times larger: electrons and protons would combine into neutrons, eating hydrogen and blocking complex elements; stars would not form.
- Stronger gravity: stars bind tighter, burn faster, burn out too soon for life.
- Weaker gravity: matter wouldn't clump; no stars or galaxies.
- Conclusion: "the universe is the way it is because the matter and the force particles have the properties they do." Is there an explanation?
String theory: the basic idea
String theory says each "letter" of matter is not a point but a tiny one-dimensional loop, an infinitely thin, vibrating, "dancing filament" (physicists, lacking Gell-Mann's flair, call it a string). Magnify an apple: atoms, nucleus, protons and neutrons, quarks, and then, at a scale far beyond today's technology, a vibrating loop.
Chapter 6 will show how replacing points with strings removes the incompatibility between quantum mechanics and gravity: the "central Gordian knot" of theoretical physics.
A theory of everything?
String theory's unifying claim is that particle properties are vibration patterns. As violin strings have resonant frequencies heard as notes and harmonics, a string's pattern of vibration determines the mass and force charges of the particle it appears as. "The electron is a string vibrating one way, the up-quark is a string vibrating another way." Force particles, including the graviton, are vibration patterns too. So all matter and all forces are "notes" of one kind of string. For the first time, a framework could in principle explain every fundamental feature of the universe, hence the labels "theory of everything" (T.O.E.), "ultimate" or "final" theory. Many string theorists use T.O.E. In a narrower sense: a theory of the particles and their interactions.
| Strong reductionism | Its critics |
|---|---|
| Understand the ingredients and, in principle, you understand everything, "from the big bang to daydreams" | Reducing joy and sorrow to string vibrations is "fatuous and downright repugnant" |
| Steven Weinberg (Dreams of a Final Theory): the reductionist view is "chilling and impersonal" but must be accepted "because that is the way the world works" | Chaos and complexity may involve genuinely new laws at higher levels |
| Greene: complex phenomena like tornadoes reflect a "calculational impasse," not new physical laws | Some disagree |
What almost everyone agrees on: principle is not practice. A T.O.E. Would not "solve" psychology, biology, chemistry or even most of physics. "Its discovery would mark a beginning, not an end": the firmest foundation for understanding a comprehensible universe.
The state of the theory (as of 1999)
Greene is candid that this book describes a work in progress, not a theory confirmed by experiment and accepted by all.
- Achievements claimed: the harmonious union of GR and quantum mechanics; a capacity to answer primordial questions about particles and forces; and elegance, with apparently arbitrary details (number of particles, their properties) tied to the geometry of extra dimensions.
- Missing: definitive, testable predictions. Greene expected possible strong circumstantial support within about ten years (Ch 9), and cites the recent string-theory explanation of Bekenstein-Hawking black-hole entropy (Ch 13) as a major success.
- Unknown equations: no one knows the exact equations of the theory; only approximations exist, and even those are only partly solved.
- Witten's line, first said by Daniele Amati: string theory is "a part of twenty-first-century physics that fell by chance into the twentieth century", like handing a supercomputer without a manual to nineteenth-century scientists.
- Beyond strings: the late-1990s breakthroughs show that membranes and other objects also matter, and that string theory is part of a larger M-theory (Ch 12).
Reading the 1999 optimism in 2026. Greene's "within the next ten years or so" did not pan out: the LHC (running since 2009) has found no supersymmetric particles, no extra dimensions, no microscopic black holes, the circumstantial signatures he hoped for in chapter 9. The theory's mathematical and conceptual yield has been large (holography, black-hole microstates, new mathematics), but its experimental status is essentially unchanged: no confirmed prediction. Critics (Lee Smolin, The Trouble with Physics; Peter Woit, Not Even Wrong, both 2006) argue the field oversold itself; defenders reply that a theory of quantum gravity was never likely to be tested quickly. The authored final chapter of this edition reviews the whole quarter-century.
"A framework, not yet a theory" is a familiar architecture problem. String theory in 1999 resembled an enterprise target-state architecture that is elegant, unifying and internally consistent, but without a single production deployment to validate it. The lesson Greene's optimism teaches in hindsight: set falsifiable milestones early, and treat elegance as a reason to invest in a design, never as evidence that it works.
Space, Time, and the Eye of the Beholder
Two principles, the laws of physics are the same for all constant-velocity observers and light's speed is the same for everyone, force a complete rethinking of space and time. Simultaneity becomes relative, moving clocks run slow, moving objects shrink, and nothing can exceed light speed. Greene's unifying picture: everything always moves through spacetime at exactly the speed of light; motion through space is simply motion diverted from time.
The teenage paradox
In June 1905 a 26-year-old patent clerk in Bern, Albert Einstein, sent a paper to Annalen der Physik; its editor, Max Planck, saw the scientific order overturned. The puzzle had troubled Einstein since he was sixteen. In the mid-1800s James Clerk Maxwell, building on Michael Faraday's experiments, unified electricity and magnetism into the electromagnetic field (felt on a mountaintop before a thunderstorm, traced by iron filings, crackling in a wool sweater) and found that electromagnetic waves travel at one fixed speed, the speed of light. Light is such a wave, and it never stops or slows.
Consider chasing a light beam at light speed. Newton's intuition says it would appear to stand still. Maxwell's theory and all observation say there is no such thing as stationary light. Einstein, unaware of how many leading physicists were wrestling with this (often down "spurious paths"), solved it largely in private thought.
Intuition and its flaws
Special relativity is about how the world appears to observers in relative motion, and its verdict is that they disagree not just about what moves but about distances and durations themselves: identical wristwatches tick at different rates and identical tape measures disagree, not from faulty equipment but because of the nature of space and time.
Slim and Jim at the drag strip (1970, a new Trans Am)
- Time. Slim drives a one-mile strip at 120 mph. Jim, on the sideline, times 30 seconds; Slim's own stopwatch reads 29.99999999999952 seconds.
- Length. A car Slim measured at 16 feet in the showroom, Jim measures (by timing it past him) at 15.99999999999974 feet.
- At 87% of light speed (580 million mph): Jim measures the car at about 8 feet, and its run takes twice as long by his clock as by Slim's.
- Why we never notice: light travels ~186,000 miles per second (~670 million mph), circling the earth over seven times a second; everyday speeds are a negligible fraction of it.
These are time dilation and Lorentz contraction. In a world where things routinely moved near light speed, they would be as intuitive as trees appearing to move past a car window.
The two pillars
The foundations of special relativity
- Principle of relativity. Constant-velocity (force-free) motion is relative: it means something only compared with something else. George and Gracie, floating in empty space, each see themselves at rest and the other drifting past; both views are equally valid. In a smooth train with the blinds down, no experiment can reveal your motion (an idea going back to Galileo). More strongly: the laws of physics are identical for all constant-velocity observers. Acceleration, by contrast, is felt (a jet-pack firing at your back).
- Constancy of light speed. Every observer measures light at 670 million mph, whatever their motion relative to the source. Run from a thrown grenade at 12 ft/s and it approaches at 20 − 12 = 8 ft/s; flee a laser beam at 100 million mph on the Enterprise and the photons still arrive at 670 million mph, not 570 million.
- Evidence. Experiments since the 1880s; in 1913 Willem de Sitter proposed using binary stars: light from a star moving toward us and one moving away arrives at the same speed, confirmed to ever higher precision.
Physicists around 1900 tried hard to refute the constancy of light speed and couldn't. Einstein embraced it: no matter how hard you chase light, it recedes at light speed. Since speed is distance divided by duration, a fact that defies intuition about speed must defy intuition about space and time.
Time I: simultaneity is relative
The cease-fire on the train
- Setup. The UN secretary-general ends a war by placing a light bulb midway between two presidents at a long table; each signs when the light reaches them. It works.
- Second treaty, on a moving train. The president of Forwardland faces the direction of travel; Backwardland's faces backward. Same ceremony; everyone on the train sees simultaneous signatures.
- On the platform. Forwardland's president moves toward the light; Backwardland's moves away. The light travels at the same speed both ways but a shorter distance to Forwardland, so platform watchers see Forwardland sign first. Fighting breaks out over the "duping."
- Who is right? Both. Events simultaneous for one observer are not simultaneous for another in relative motion.
- Scale. A 100-ft table on a 10-mph train: about a millionth of a billionth of a second. At 600 million mph: the light takes almost 20 times longer to reach Backwardland.
Greene says that if you remember one thing from the chapter, remember this détente: it follows from the constancy of light speed with no mathematics. If light could be "kicked" faster by the train or "dragged" slower, the effects would cancel and everyone would agree; real light can't be. There is no universal cosmic clock.
Time II: the light clock
Define time pragmatically as what clocks measure, and a clock as something with perfectly regular cycles. The simplest is the light clock: a photon bouncing between two mirrors six inches apart, one round trip (a "tick") about every billionth of a second.
Watch a light clock slide past. Its photon must travel diagonally to hit the moving mirror, a longer path at the same speed, so it ticks less often. By the principle of relativity, a Rolex bolted to the light clock in a windowless train must stay in sync with it (otherwise the passengers could detect their motion), so every clock, and time itself, slows for moving objects.
The numbers behind the light clock
- Lorentz factor: γ = 1 / √(1 − v²/c²). Moving clocks tick slower by γ; moving lengths shrink by 1/γ.
- At 10 mph the sliding clock moves about 15 billionths of a foot per tick: negligible.
- At ¾ light speed: γ ≈ 1.5, so the moving clock runs at about two-thirds the rate.
- At 98% light speed: γ ≈ 5, so an object appears 80% shorter.
- At 99.5%: γ ≈ 10.
Life on the run: muons
A muon at rest decays in about two millionths of a second ("a gun to its head"). Boosted to 99.5% of light speed (667 million mph) in an accelerator, its lab-measured lifetime grows about ten-fold: its "wristwatch" runs slow. The catch: no fountain of youth. Everything the fast muon does slows equally; if a resting muon could read 100 books, so can the moving one. People zipping around as fast would live 700 years by our clocks, yet experience an ordinary lifetime.
Who is moving? The twin puzzle
Each of two passing observers says the other's clock runs slow. Contradiction? Only if they could compare clocks directly, and that requires one of them to turn around, which means accelerating, which breaks the symmetry.
George catches Gracie
- Setup. George and Gracie pass at 99.5% of light speed and both set clocks to 12:00.
- Action. After 3 years on his clock, George fires his jet-pack and chases her at the same speed, reaching her 3 years later.
- Result. George's clock shows 6 years; Gracie's shows 60 years. George, having felt the acceleration, can't claim he was at rest. He has travelled into Gracie's future.
- What about phone calls? If George and Gracie compare clocks by cell phone as they separate, the light-speed signal delay grows as they recede. Each, correcting for it, still concludes the other's clock is slow, and both are consistent, because neither signal is instantaneous.
Motion's effect on space
Jim measures Slim's speeding car by timing it past him and multiplying by its speed. From Slim's view, Jim's watch runs slow, so Jim's elapsed time is short and the car comes out shorter. Moving objects contract along their direction of motion (by 1/γ).
Motion through spacetime
Hermann Minkowski, and then Einstein, treated time as a fourth dimension: a meeting needs three spatial coordinates (9th floor, 53rd Street and 7th Avenue) and a time (3 p.m.), a point in spacetime.
The one-speed car in the desert
- Setup. A car that only goes 100 mph runs a 10-mile east–west track: 6 minutes a run. Later runs take 6.5, 7, 7.5 minutes.
- Explanation. Late-day sun glare made Slim drive at an angle: part of his fixed 100 mph went north–south, leaving less for east–west.
- The leap. The same sharing happens between space and time.
Einstein's unifying picture. Every object always travels through spacetime at exactly the speed of light. At rest, all that motion is through time (you age at the maximum rate). Move through space and some motion is diverted from time, so your clock slows. Divert all of it and you move through space at light speed with no motion through time: light doesn't age; a photon from the big bang is as old now as when it set out. That is also why nothing can go faster than light: there's no more time-motion to divert.
E = mc²
Einstein disliked the name "relativity," preferring "invariance theory" for the unchanging speed of light. His most famous equation interweaves another pair: energy and mass are convertible currencies at a fixed exchange rate, c². Because c² is huge, a little mass yields enormous energy: less than 1% of about two pounds of uranium was converted at Hiroshima; fusion might one day power the world from seawater.
It also explains the light barrier:
A muon's growing "heft" (Greene's figures)
- 99.9% of c: about 22 times heavier than at rest.
- 99.999%: about 224 times.
- 99.99999999%: over 70,000 times.
- At c: infinite, so reaching light speed would need infinite energy: nothing travels faster than light.
"Relativistic mass" is out of fashion. Greene's 1999 explanation (moving muons "weigh a lot more") uses relativistic mass, a concept most physicists now avoid. Modern usage keeps mass invariant (the rest mass in the particle tables) and says a moving particle's energy and momentum grow by the factor γ, so accelerating it further demands ever more energy. The physics is identical; the language is cleaner, and it avoids implying that a fast muon would collapse into a black hole.
Relativity in your pocket, and at the millimetre. Since 1999 special and general relativity have become everyday engineering. GPS satellites' clocks would drift about 38 microseconds a day (roughly 10 km of position error) without corrections for both motion (special relativity, −7 µs) and weaker gravity (general relativity, +45 µs). Optical atomic clocks have measured time dilation at speeds under 10 m/s (NIST, 2010) and, in 2022, gravitational time shifts across a single millimetre of height (JILA). And in the Fermilab muon g-2 experiment (final result 2025), muons circulating at γ ≈ 29.3 live about 30 times longer than at rest, exactly as Greene's muon example predicts.
The conclusion that nothing outruns light "plants the seeds" for the next conflict: Newton's gravity acts instantly.
Frames of reference are an architect's daily reality. Greene's two presidents, each right, map onto distributed systems: there is no global "now" across nodes, which is why event ordering in payments and trading systems is built on logical clocks, consensus protocols and explicit time-stamping rather than assumed simultaneity. Einstein's lesson for architects: when observers disagree about "when," don't ask who's right; design for the disagreement.
Of Warps and Ripples
Special relativity's speed limit breaks Newton's instantaneous gravity. Einstein's way out ran through two insights: gravity and acceleration are indistinguishable (the equivalence principle), and acceleration warps space and time. Put together: gravity is not a force transmitted across space; gravity is the curvature of spacetime itself, and its disturbances travel at exactly the speed of light. The theory predicted light bending, black holes and an expanding universe, and it has passed every test since, yet it collides head-on with quantum mechanics.
Newton's owner's manual
Isaac Newton (born 1642, Lincolnshire) brought mathematics to physics so forcefully that, when the mathematics he needed didn't exist, he invented it (calculus). His universal gravitation united heaven and earth: the apple and the moon obey the same law. Everything attracts everything, with a strength set by mass and distance.
Newton's law of gravity
- F = G m₁m₂ / r²: proportional to the product of the masses, inversely proportional to the square of the distance.
- Derived from a close study of Johannes Kepler's analysis of planetary orbits.
- Track record: planets, comets, the moon, eclipses, rockets to Mars, baseballs and springboard divers, all predicted with spectacular accuracy until the early twentieth century.
Newton himself admitted he had given an "owner's manual," not the inner workings. He called the idea of matter acting on distant matter through a vacuum, with nothing mediating, "so great an absurdity" that no competent thinker could accept it, and left the agent of gravity "to the consideration of my readers." Like a CD player, you can use it without knowing how it works, until it breaks.
The conflict with special relativity
Nothing, not even information or influence, travels faster than light. Sound (about 700 mph) lags behind light at a ballpark or in a thunderstorm; nothing can ever beat light. But in Newton's law the force depends only on present masses and separation, so changes are felt instantly. If the sun exploded, Newton says earth's orbit would change at once, though the light would take eight minutes to arrive across 93 million miles. That "violates this precept maximally." Einstein trusted special relativity and set out to rebuild gravity.
Einstein's happiest thought (1907)
The scale-triggered bomb (Greene's thriller, set in 2050)
- Problem. A nuclear bomb in Washington will detonate if the scale it sits on changes by more than 50%, within a week. Nowhere on earth is safe; it must be launched into space.
- Isaac's objection. As the rocket rises, gravity weakens, the bomb weighs less, the reading drops: boom.
- Albert's objection. Upward acceleration presses the bomb harder into the scale (as an accelerating car presses you into your seat), the reading rises: boom.
- Albert's insight. Adjust the acceleration moment by moment so the rise from acceleration exactly cancels the drop from weakening gravity. The reading never changes. Disaster averted.
- Lesson. Accelerated motion and gravity are indistinguishable: the equivalence principle.
In a sealed, windowless room, constant velocity is undetectable, but acceleration is felt. Einstein's realisation was that you cannot tell the feeling of acceleration from the feeling of gravity: the floor pressing on your feet is the same whether the room accelerates upward in space or sits on earth. This completes the democracy that special relativity began: every observer, accelerating or not, may claim to be at rest provided they include a suitable gravitational field. Why was Einstein so happy? Because gravity is mysterious, but acceleration is concrete; the link let him attack gravity using the physics of motion.
Acceleration warps space and time (1912)
About five years later Einstein wrote to Arnold Sommerfeld that "never in my life have I tormented myself anything like this... Compared to this problem the original relativity theory is child's play." The next step came from a spinning ride.
Slim and Jim on the Tornado
- The ride. You stand against a spinning Plexiglas wall; the spin presses you outward (in space, a spinning station mimics gravity this way).
- Space warps. We on the ground measure circumference ÷ radius = 2π ≈ 6.28. Slim measures the circumference with a ruler lying along the direction of motion, so it is Lorentz-contracted and must be laid down more times: he finds a longer circumference. Jim measures the radius with a ruler perpendicular to the motion: no contraction, same radius. Their ratio exceeds 2π.
- Euclid fails. On a flat sheet the ratio is 2π; on a sphere it's less (radii converge); on a saddle it's more (radii splay). The riders live in curved space.
- Time warps. Farther from the centre you move faster, so your clock runs slower; Jim's watch slows as he crawls out toward Slim. And the farther out, the greater the acceleration. So stronger acceleration means more warped time: time passing at different rates in different places.
Hermann Minkowski (1908): "space on its own and time on its own will decline into mere shadows, and only a kind of union between the two will preserve its independence." Chain the links together, gravity ≡ acceleration and acceleration ↔ curved spacetime, and Einstein's answer to Newton's black box follows: gravity is the warping of space and time.
How curved spacetime makes gravity
Picture (in a 2-D model, with time suppressed) space as a flat sheet. Before Einstein, space and time were an inert stage. Now mass warps the sheet, like a bowling ball on a rubber membrane, and other objects move along the warps: a ball bearing rolled past the dip curves, and at the right speed goes into orbit. Earth orbits the sun because it moves through the valley the sun makes in spacetime.
- More mass, deeper warp: stronger gravity.
- Farther away, shallower warp: weaker gravity.
- Everything warps: earth's dent holds the moon and us; a skydiver slides down earth's dent; even your body makes a tiny dimple.
John Wheeler's summary: "mass grips space by telling it how to curve, space grips mass by telling it how to move."
Three flaws in the rubber-sheet picture (Greene's own caveats). (1) The bowling ball sinks because earth's gravity pulls it down; the sun has nothing pulling it. In GR the warping is gravity, and objects follow the shortest, "easiest" paths (geodesics) through curved spacetime. (2) Real space warps in three dimensions all around the sun; it is not a solid surface, so earth doesn't "hit" it, and we do feel space: gravity is how we feel it. (3) The picture omits time, yet for slow bodies like earth around the sun, the warping of time matters far more than the warping of space in producing the orbit.
Conflict resolved: gravity at light speed
Drop a bowling ball onto the membrane and the ball bearing doesn't react instantly: ripples spread outward first. Einstein calculated that disturbances in spacetime travel at exactly the speed of light. If the sun vanished, earth would feel the change at the same moment we saw it, about eight minutes later. Gravity "keeps pace with, but does not outrun, photons."
The ripples have been heard. Gravitational waves were detected directly by LIGO in 2015 (Nobel 2017). In 2017 a neutron-star merger (GW170817) was seen in gravitational waves and, 1.7 seconds later, in gamma rays after a 130-million-year journey, pinning the speed of gravity to light speed within about one part in 1015. Greene's "keeps pace with photons" is now a measured fact.
Warped time: George on a cable
Lower George on a long cable toward the sun while Gracie stays far away: his clock runs slower the stronger the gravity. Unlike the constant-velocity case, there's no symmetry: George feels gravity tugging; both agree his clock is slow (just as his jet-pack acceleration made him age less in Ch 2).
How much gravity slows clocks
- Near the sun's surface (Gracie a billion miles away): George's clock runs at ~99.9998% of hers.
- Above a neutron star (a sun's mass at ~1015 times solar density): ~76%.
- An inch above a 1,000-solar-mass black hole's horizon: about 10,000 times slower. Hover a year, fly home, and 10,000+ years have passed on earth: a time machine into the future.
Testing general relativity
GR and Newton agree almost everywhere (as they must, given Newton's record), which makes them hard to tell apart. Einstein proposed a test: starlight grazing the sun should bend, visible only during a total eclipse.
November 1915: Einstein predicts a deflection of 1.75 arcseconds (0.00049°), the angle of a quarter seen from nearly two miles.
29 May 1919: Arthur Eddington, urged by Sir Frank Dyson (Astronomer Royal), observes the eclipse from Príncipe off West Africa; a second team (Charles Davidson, Andrew Crommelin) at Sobral, Brazil.
6 November 1919: joint meeting of the Royal Society and Royal Astronomical Society announces confirmation.
7 November 1919: The Times: "REVOLUTION IN SCIENCE... NEWTONIAN IDEAS OVERTHROWN." Einstein becomes a global celebrity.
Was Eddington right? Greene notes the 1919 measurements were hard to reproduce and drew criticism, including the charge that Eddington discarded inconvenient plates. A 2009 reanalysis (Daniel Kennefick) and a modern re-measurement of the original Sobral plates in 1979 concluded the results did genuinely favour Einstein, though with large error bars. Today radio interferometry confirms light bending to better than 0.01%.
In the forty years before 1999, precise tests (radar echo delays, planetary orbits, clock experiments) uniformly confirmed GR.
Black holes
In 1916, between artillery calculations on the Russian front, Karl Schwarzschild found the exact warping around a spherical star and sent it to Einstein, who presented it to the Prussian Academy. It revealed that if enough mass is packed small enough, nothing, not even light, can escape. First called dark or frozen stars, they were renamed black holes by John Wheeler. Schwarzschild died months later, aged 42, of a skin disease contracted at the front.
- Event horizon: the point of no return. Fall in feet-first and the pull on your feet exceeds that on your head so much that you are stretched to shreds.
- Scale: the sun becomes a black hole if squeezed from ~450,000 miles radius to just under 2 miles ("upper Manhattan"); a teaspoon of it would weigh as much as Mount Everest. Earth would have to shrink below half an inch in radius.
- Evidence (1999): infalling gas heated near light speed glows in X-rays with properties GR predicts; a black hole of about 2.5 million solar masses at the Milky Way's centre; billions of solar masses powering quasars.
Black holes since 1999. The Milky Way's central black hole, Sagittarius A*, is now pinned at ~4.3 million solar masses from decades of star orbits (Nobel Prize 2020 to Reinhard Genzel and Andrea Ghez, with Roger Penrose for the theory). The Event Horizon Telescope photographed the shadow of M87* (6.5 billion suns) in 2019 and Sgr A* in 2022. LIGO-Virgo-KAGRA now detect black-hole mergers routinely; in 2025 the clearest signal yet confirmed Hawking's area theorem. Gravity Probe B (2011) confirmed that a spinning earth drags spacetime around with it, another GR prediction.
The expanding universe and the big bang
Applied to the whole universe, Einstein's equations, built on Bernhard Riemann's nineteenth-century geometry of curved space, say space must be stretching or shrinking. Even Einstein couldn't accept that; he added a cosmological constant to hold the universe static. Alexander Friedmann (early 1920s) used the original equations to show galaxies would be carried apart by stretching space; Edwin Hubble observed exactly that (1929). Einstein dropped the constant, reportedly calling it "the biggest blunder of his life."
Run expansion backward and everything compresses: stars dissolve into hot plasma, the universe shrinks to an orange, a pea, a grain of sand, apparently a point of unimaginable density (an image Greene will challenge later). That is the big bang. But it is not an explosion into space: "there is no surrounding space." All of space was shrinking; the big bang is "the eruption of compressed space whose unfurling, like a tidal wave, carries along matter and energy even to this day."
Cosmology since the book. Greene's "about 15 billion years" is now 13.8 billion (Planck satellite). And Einstein's "blunder" came back: in 1998, two supernova teams found the expansion is accelerating (Nobel 2011), most simply explained by a cosmological constant (dark energy). Since 2024 the DESI survey has hinted that dark energy may be weakening over time, and there is a persistent "Hubble tension": local measurements give an expansion rate about 8% faster than the early-universe value. Either could signal physics beyond Einstein's equations.
Is general relativity right?
No deviation has been found. But progress doesn't come only from better measurements: Newton was dethroned not by an experiment but by a theoretical conflict with special relativity; only afterward were the tiny measurable differences sought. Internal theoretical inconsistencies can drive progress as much as data. And GR now faces such a conflict, with quantum mechanics, which blinds us to what happens at the centre of a black hole or the moment of the big bang.
Fix the model before the measurements force you. Einstein's route to GR is a lesson in architecture governance: a system that works perfectly in production (Newton's gravity) can still be structurally inconsistent with another validated component (special relativity). Mature architecture review hunts for these cross-domain contradictions (say, a real-time fraud model consuming batch-latency data) before any incident proves them. As with Mercury's orbit, the failure eventually shows up where the two theories meet.
Microscopic Weirdness
At small scales, nature is lumpy, probabilistic and uncooperative with common sense. Energy comes in quanta; light and matter are both waves and particles; a particle's wave gives only the probability of where it will be found; in Feynman's version, a particle takes every possible path at once. At the heart of it all sits Heisenberg's uncertainty principle. Nothing can be cornered: the more tightly you pin down position (or a moment in time), the more wildly momentum (or energy) fluctuates, and that frenzy, applied to space itself, is what wrecks general relativity at tiny scales.
The H-Bar
Back from their solar-system jaunt, George and Gracie stop at the H-Bar. George's cigar vanishes from his lips and turns up on the counter behind him, as if it passed through his head. The ice cubes in their glasses rattle frantically, Gracie's (in the smaller glass) even more so, and one passes straight through the glass. Leaving, they walk through a door painted on the wall, and nobody blinks. In the H-Bar, Planck's constant (ħ) is huge; in our world it is tiny, so these quantum effects stay hidden in the microworld.
Around 1900, as Conrad and Freud explored the darkness of the human heart, Max Planck shed the first light on quantum mechanics. Richard Feynman (1965): "I think I can safely say that nobody understands quantum mechanics." Relativity is strange but, once its principles are accepted, its conclusions feel inevitable. By about 1928 quantum mechanics had its rules and has since made the most precise predictions in the history of science, yet its users follow recipes "without really understanding why the procedures work." Bohr: if quantum mechanics doesn't make you dizzy, you haven't understood it. We believe it for one reason: astounding predictive accuracy, like a stranger who recounts every detail of your childhood.
The infinite oven
Heat a perfectly insulated, evacuated oven to 400°F and its walls fill it with electromagnetic radiation. Maxwell's theory says only waves with a whole number of peaks and troughs fit between the walls. Nineteenth-century thermodynamics said each allowed wave carries the same energy, set by the temperature. But there are infinitely many allowed waves (ever more peaks and troughs), so the total energy is infinite: nonsense (later dubbed the "ultraviolet catastrophe").
Wave vocabulary
- Wavelength: distance between successive peaks.
- Frequency: oscillations per second. Shorter wavelength = higher frequency (shake a rope faster to make shorter waves).
- Amplitude: height of the wave; more energy = bigger amplitude (a harder-plucked violin string sounds louder).
Planck's miserly landlord
- Setup. An infinite crowd in a cold warehouse; heat costs $1 per degree per person. Set it to 80° and the landlord expects an infinite fortune.
- Loophole. He gives no change. Tenants organise by denomination: one holds all pennies, one all nickels, and so on up to ever bigger bills.
- Payment. Pennies: 8,000; nickels 1,600; dimes 800; quarters 320; dollars 80; fives 16; tens 8; twenties 4; fifties 1. Everyone holding bigger bills can't pay without change, so pays nothing.
- Outcome. The landlord leaves with $690, not infinity.
- Physics. Planck guessed in 1900 that a wave's energy comes in lumps, multiples of a minimum that grows with frequency. High-frequency waves' minimum lump exceeds their thermal "share," so they stay dormant. Total energy becomes finite, and with one adjustable constant it matched measured oven spectra precisely. (Nobel Prize 1918.)
That constant, Planck's constant (ħ), is about a billionth of a billionth of a billionth in everyday units, so energy steps are imperceptibly small, which is why a violin's volume seems continuous. E = hν: energy per lump is proportional to frequency.
Photons and the photoelectric effect
Planck had no reason for his lumps beyond "it worked." George Gamow: it was as if nature allowed you a whole pint of beer or none, but nothing in between. In 1905 Einstein explained them (Nobel 1921) via the photoelectric effect, discovered by Heinrich Hertz in 1887: light shone on metals ejects electrons.
- Brighter light ejects more electrons, but not faster ones.
- Higher-frequency light (red to violet, then ultraviolet, X-rays) ejects faster electrons.
- Below a threshold frequency, no electrons escape however blinding the light.
The children in the basement
- Setup. The landlord confines all children to a sunken basement; leaving costs 85¢. Adults can only throw money down.
- Coins fail. Pennies, nickels, dimes and quarters rain down in vast quantities, but in the scramble no child collects 85¢.
- Dollars work. Each child catching a single dollar leaves at once, with 15¢ left over. Throw more dollars and more children leave, each still with exactly 15¢.
- Mapping. Photons are the bills: frequency sets each photon's energy (the denomination); intensity sets the number of photons. Light is a stream of particles, named photons by chemist Gilbert Lewis. A 100-watt bulb emits about 1020 photons per second.
Wave or particle? The double slit
Newton said light is particles; Christiaan Huygens said waves; Thomas Young's early-1800s double-slit experiment seemed to settle it for waves, and Maxwell gave the wave theory its mathematics. Feynman held that all of quantum mechanics can be gleaned from this one experiment.
- One slit open: a bright patch behind it.
- Particle expectation, both open: two patches added together.
- Wave result, both open: interference bands: peaks meeting peaks reinforce, peaks meeting troughs cancel (the principle behind noise-cancelling headphones).
- The shock. Fire photons one at a time, say one every ten seconds. Each lands as a single dot, yet the dots build up the interference pattern. Opening the second slit makes some places darker: adding photons reduces brightness there. Feynman: like separately fired machine-gun bullets cancelling each other.
The photoelectric effect shows light is particles; the double slit shows it interferes like a wave; it is both.
Matter is waves too
Before the mid-1920s, quantum theory (much of it led by Niels Bohr in Copenhagen, explaining hydrogen's light) was "a makeshift union" of old and new. In 1923 French nobleman Prince Louis de Broglie (Nobel 1929) reasoned: E = mc² links mass to energy; Planck and Einstein link energy to frequency; so matter should have waves. Einstein embraced it. In the mid-1920s Clinton Davisson and Lester Germer at Bell Labs bounced electrons off nickel crystals and saw interference, even firing electrons one by one. All matter has a wavelength: λ = h/p (Planck's constant over momentum), far too tiny to notice for coffee cups and baseballs, just as the huge speed of light hides relativity.
Bigger and bigger waves. Matter-wave interference has since been shown for C60 buckyballs (1999) and, in 2019, for molecules of about 2,000 atoms (over 25,000 atomic mass units) at the University of Vienna. No limit has been found where waviness stops; how big superpositions can get is now an active experimental question tied to quantum gravity (proposed tests of whether gravity itself can entangle masses).
Waves of what? Probability
Erwin Schrödinger first pictured "smeared-out" electrons, but no one ever finds a fraction of an electron. In 1926 Max Born proposed the interpretation still used: the wave's (squared) magnitude gives the probability of finding the electron there. If the probability is twice as high at A as at B, repeated identical experiments find the electron at A twice as often. This is fundamental chance, not a roulette wheel's ignorance of detail. Months after de Broglie, Schrödinger wrote the equation governing these wave functions. "By 1927, therefore, classical innocence had been lost": the equations evolve probabilities precisely, but don't say which future happens.
| Einstein | Experiment |
|---|---|
| "God does not play dice with the Universe": probability signals incompleteness in our theory | Experiments, many after his death, show Einstein was wrong |
| Particles really have definite properties we can't see | Hawking: "Einstein was confused, not the quantum theory" |
The interpretation debate continues: what probability waves are, how a particle "chooses" one outcome, or whether it splits into parallel universes living out every possibility. The meta-lesson of relativity and quantum mechanics alike: deep questions may demand flexibility in accepting the answers.
Feynman's sum over paths
Watch which slit each electron goes through, by bouncing photons off it, and you disturb it so much that the interference vanishes (the pattern reverts to two patches). So the claim "each electron goes through one slit" can't be verified without ruining the experiment. Feynman, after World War II, drew the radical conclusion:
Each electron takes every possible path simultaneously: through the left slit, the right slit, zigzagging, even via the Andromeda galaxy. Feynman assigned a number to each path; combining them reproduces the wave-function probabilities exactly. For big objects like baseballs, all paths but one cancel, leaving Newton's single trajectory. Feynman: quantum mechanics "describes nature as absurd from the point of view of common sense. And it fully agrees with experiment. So I hope you can accept nature as She is: absurd."
The uncertainty principle
Locating an electron with ever gentler light fails because light comes in photons: dim it below one photon and it's off. Use lower-frequency (gentler) photons instead and you lose resolution: a wave locates an object only to within one wavelength, as ocean waves locate a submerged rock only to within a wave cycle. Short wavelength: sharp position, big kick to velocity. Long wavelength: gentle kick, fuzzy position. Werner Heisenberg (1927) showed this trade-off is fundamental, independent of equipment.
Heisenberg's two trade-offs
- Position–momentum: Δx · Δp ≥ ħ/2. Pin down where, lose how fast, and vice versa.
- Energy–time: ΔE · Δt ≳ ħ/2. Energy can fluctuate wildly over short enough times.
- Quantum claustrophobia. Squeeze an electron in a shrinking box and it bounces ever more frantically; nature won't let its constituents be cornered. (The H-Bar's smaller glass, wilder ice.)
- Tunnelling. Like an airline letting you fly now and pay within 24 hours to claim an inheritance, a particle can "borrow" energy briefly to pass through a barrier it classically couldn't cross. For a whole pellet, every particle must tunnel together: possible, absurdly unlikely. Walk into a wall every second and you'd wait longer than the age of the universe for a good chance of passing through.
John Bell's theorem and Alain Aspect's experiments showed, contra Einstein, that particles cannot be thought of as having definite position and velocity simultaneously, hidden from view.
Bell tests made airtight. In 2015, three groups (Delft, Vienna, NIST) closed the main loopholes in Bell tests; the 2022 Nobel Prize went to Alain Aspect, John Clauser and Anton Zeilinger. Precisely, these experiments rule out local hidden variables; non-local hidden-variable theories such as Bohmian mechanics remain logically possible, so Greene's phrasing is slightly stronger than the experiments strictly allow. Quantum weirdness is now engineering: quantum computers (Google's "Willow" chip, 2024) exploit superposition and entanglement, and quantum key distribution is deployed in banking pilots and satellite links.
ħ versus h. Greene names Planck's constant ħ ("h-bar") throughout, which is a simplification. Planck's original constant is h ≈ 6.63 × 10−34 J·s; ħ = h/2π ≈ 1.05 × 10−34 J·s. The relations read E = hν (or ħω) and λ = h/p. The "billionth of a billionth of a billionth" figure matches ħ in old cgs units (erg-seconds). None of this changes the argument.
The uncertainty principle "captures the heart of quantum mechanics": definite positions, speeds and energies at definite moments are artefacts of ħ's smallness. Applied to the fabric of spacetime itself, it reveals "fatal imperfections in the stitches of gravity," the third conflict.
Measurement disturbs the system. The quantum lesson that observing an electron changes its path has a direct analogue in AI governance: monitoring changes behaviour. Models whose outputs feed back into the data they're evaluated on (credit decisions shaping who gets credit, fraud alerts shaping fraudster tactics) exhibit an "observer effect." Design evaluation to account for it, for example with hold-out populations and counterfactual logging, rather than assuming a passive measurement.
The Need for a New Theory
Quantum field theory tamed three of the four forces into the standard model, one of the most accurate theories ever built, and showed that every force is tied to a symmetry. Gravity should follow suit, but it doesn't. Uncertainty makes the gravitational field, and hence the geometry of space itself, fluctuate ever more violently at ever smaller scales. Below the Planck length (about 10−33 cm), smooth spacetime dissolves into "quantum foam," the calculations return infinity, and the two pillars of physics are revealed to be fundamentally incompatible.
Fire and gunpowder
Beings on "one planet orbiting a run-of-the-mill star in the far edges of a fairly ordinary galaxy" have understood physics from subatomic particles to the whole cosmos. Physicists still want the deepest level, what Stephen Hawking called knowing "the mind of God." Most problems need quantum mechanics or general relativity, not both. But at a black hole's centre or the big bang's first moment you need both, and mixing them is like fire and gunpowder: calculations predict probabilities not of 20% or 91% but of infinity, which is meaningless. Something is seriously wrong.
The heart of quantum mechanics: a frenzied vacuum
Uncertainty is not just observers disturbing what they measure; nature is intrinsically turbulent. Even empty space seethes.
The compulsive borrower
- Analogy. A man borrows from friend after friend, and the shorter the loan period, the bigger the loan. His bank balance swings wildly moment to moment, but over time he ends up no better off.
- Physics. In an empty box, energy and momentum fluctuate between extremes that grow as the box shrinks and the time window narrows.
- What gets borrowed? Everything. Via E = mc², a big enough fluctuation can conjure an electron–positron pair out of nothing, which then annihilate to repay the loan; likewise every particle and every force field.
- Feynman: "Created and annihilated, created and annihilated, what a waste of time."
- Why space looks calm. On average the loans cancel; macroscopic averaging hides microscopic frenzy. This frenzy is what defeats gravity.
Quantum field theory and the standard model
In the 1930s and 1940s, Paul Dirac, Wolfgang Pauli, Julian Schwinger, Freeman Dyson, Sin-Itiro Tomonaga, Richard Feynman and others found that Schrödinger's equation is only an approximation: it ignores special relativity. (Schrödinger had tried including relativity, got predictions for hydrogen that disagreed with experiment, and sensibly chose divide and conquer.) Relativity matters because energy constantly changes form (E = mc²), so particles must be creatable and destroyable.
What "relativistic quantum field theory" means
- Quantum: probability and uncertainty built in from the start.
- Field: quantum principles applied to force fields like Maxwell's electromagnetic field.
- Relativistic: special relativity built in from the start.
- Picture: an ocean of invisible field lines, but made of particles (photons for electromagnetism), with energy constantly shifting between fields as particles appear, move and vanish.
| Theory | Force(s) | Key figures | Status |
|---|---|---|---|
| Quantum electrodynamics (QED) | Electromagnetism | Dirac, Feynman, Schwinger, Tomonaga, Dyson | Most precise theory ever; electron properties match experiment to better than one part in a billion (Toichiro Kinoshita's Cornell calculations: thousands of pages, the world's biggest computers) |
| Quantum chromodynamics (QCD) | Strong | 1960s–70s | Name is just colourful ("quantum strong dynamics" would be more logical) |
| Electroweak theory | Weak + electromagnetic unified | Sheldon Glashow, Abdus Salam, Steven Weinberg (Nobel) | At high temperatures (just after the big bang) the two forces merge; as the universe cooled, symmetry breaking separated them |
Feed in about 19 measured parameters (masses, charges, force strengths) and the resulting standard model of three forces and three families matches every experiment, down to a billionth of a billionth of a metre.
The standard model completed, and then some. The model's last predicted piece, the Higgs boson, was found in 2012 (Nobel 2013 to Peter Higgs and François Englert). Neutrino masses (Ch 1) push the parameter count to about 26. QED's precision has only grown: the electron's magnetic moment was measured in 2023 to about 0.13 parts per trillion, matching theory. And the long-running muon g-2 anomaly, a possible crack in the model, essentially closed in 2025 when improved lattice-QCD calculations agreed with Fermilab's final measurement. The standard model has survived every test the LHC has thrown at it.
Messenger particles
Each force field is made of particles: photons (electromagnetism), gluons (strong), W and Z bosons (weak), all structureless in the standard model. Two charged particles interact by exchanging photons, roughly like ice-skaters pushed apart by hurling bowling balls at each other. But exchanges can also attract, so think of the photon as carrying a message: "move apart" for like charges, "come together" for opposites. Hence messenger particles. Gluons glue quarks into protons and neutrons; W and Z bosons mediate radioactive transmutations.
Gauge symmetry: forces from symmetry
Gravity enforces a symmetry: all observers, even accelerating ones, are on equal footing (Ch 3). The other forces turn out to enforce symmetries too, more abstract ones.
Quarks come in three "colours" (red, green, blue: just labels for strong charge). Physics is unchanged if you shift the colours (to "yellow, indigo, violet"), even differently at every place and moment. That is strong-force gauge symmetry. Work by Hermann Weyl (1920s) and Chen-Ning Yang and Robert Mills (1950s) showed that maintaining such a symmetry requires a force field that compensates for the shifts, like a perfect climate-control system. For colour, the compensating field is the strong force. Similar arguments give the weak and electromagnetic forces. All four forces exist to enforce symmetries.
That shared heritage made a quantum field theory of gravity, with the graviton as messenger, the obvious next step. Generations of brilliant physicists tried. "No one has succeeded."
Why gravity resists: quantum foam
Zoom into empty space. Classically it stays flat and placid forever. But quantum mechanics says the gravitational field fluctuates too, averaging zero but undulating, and the smaller the region, the larger the undulations ("nothing likes to be cornered"). Gravitational field is curvature, so space itself warps ever more violently at smaller scales until it becomes a frothing, twisted mess: John Wheeler's quantum foam, where left and right, up and down, even before and after lose meaning.
- The clash: the central feature of quantum mechanics (the uncertainty principle) directly contradicts the central feature of general relativity (smooth spacetime geometry).
- The symptom: combined calculations return infinity, "nature's way of telling us that we are doing something that is quite wrong," like a schoolteacher's rap on the wrist.
- Why GR works anyway: at larger scales the fluctuations average out, as a dot-matrix picture looks smooth from afar but is discrete up close.
The Planck length
- Formula: ℓP = √(ħG/c³), combining quantum mechanics (ħ), gravity (G) and relativity (c).
- Size: about 10−33 cm: "a millionth of a billionth of a billionth of a billionth of a centimeter."
- Scale: magnify an atom to the size of the observable universe and the Planck length grows only to the height of an average tree.
Does it matter?
The conflict shows up only in a rather esoteric domain, and physicists divide. Some note it and carry on using each theory where it works. Others are deeply unsettled: they hold, unprovably but deeply, that the universe at its deepest level is described by a logically sound, harmonious theory, not "a mathematically inconsistent patchwork." Attempts to modify either GR or quantum mechanics met "failure after failure," "until the discovery of superstring theory."
"The only game in town"? Greene's framing implies string theory is the sole viable candidate. Other research programmes pursue quantum gravity too: loop quantum gravity (Ashtekar, Rovelli, Smolin), which quantises spacetime geometry directly and predicts discrete area and volume; asymptotic safety (Weinberg's own idea, revived by Reuter); causal sets; and causal dynamical triangulations. None is complete, and string theory remains the most developed and the only one that also unifies the forces, but its rivals undercut the "no alternative" argument for it. Note too that the infinities of perturbative quantum gravity do not prove gravity can't be quantised; they show that GR is an effective theory valid below the Planck scale, which is how most physicists now frame it.
Has quantum foam been tested? Indirectly. If spacetime were foamy in simple ways, high-energy photons would travel at slightly different speeds from low-energy ones. Gamma-ray bursts observed by NASA's Fermi telescope (notably GRB 090510, in 2009) and later by LHAASO showed no such delay over billions of light-years, ruling out the simplest linear Planck-scale effects. More subtle foam remains untested.
Two systems that each pass every test can still be incompatible. Greene's central problem mirrors a common enterprise risk: a real-time decisioning platform and a batch ledger can each be "correct" yet produce contradictions only at their boundary (high-frequency edge cases, end-of-day cutovers). Like the Planck length, the conflict hides at a scale ordinary testing never probes. Architects should name the boundary conditions explicitly and test them on purpose.
Nothing but Music
String theory replaces point particles with tiny vibrating loops about the Planck length long. Each vibration pattern is a particle: its energy fixes the mass, its finer details the charges, and one pattern is exactly the graviton, so gravity is built in. Because a string has size, it cannot probe or be disturbed by structure below the Planck length, and its interactions are smeared over spacetime rather than crammed into a point; that smearing tames quantum foam and removes the infinities that wrecked every point-particle theory of quantum gravity.
An aeolian universe
From the Pythagorean "music of the spheres" onward, music has supplied the metaphors for cosmic order. In string theory the metaphor becomes literal: the microscopic world is filled with strings whose vibration patterns orchestrate the cosmos. The standard model treats particles as structureless points, has been verified down to 10−18 m, but excludes gravity, and point-particle attempts to include it drown in quantum foam. String theory changes the microscopic description so that general relativity becomes compatible with quantum mechanics, provides one ingredient for all matter and forces, and, as later chapters show, radically reshapes spacetime. Despite the name, these strings are not made of molecules; they lie deep inside the particles that make atoms.
A brief history
1968: Gabriele Veneziano at CERN notices that Euler's beta-function, a 200-year-old formula devised for pure mathematics, describes strongly interacting particles "in one fell swoop," but no one knows why.
1970: Yoichiro Nambu (Chicago), Holger Nielsen (Niels Bohr Institute) and Leonard Susskind (Stanford) show the formula describes tiny vibrating one-dimensional strings.
Early 1970s: high-energy experiments contradict the string model of the strong force; quantum chromodynamics wins; string theory is "relegated to the dustbin."
1974: John Schwarz and Joël Scherk (École Normale Supérieure) realise an unwanted messenger-like vibration is exactly the graviton: string theory is a quantum theory of gravity. "Our work was universally ignored."
Late 1970s–early 1980s: string theory is found to suffer subtle quantum inconsistencies (anomalies).
1984: Michael Green (Queen Mary College) and Schwarz show the anomalies cancel, and the theory can encompass all four forces and all matter. Hundreds of physicists drop their projects.
1984–1986: the first superstring revolution, over 1,000 papers; many features of the standard model emerge "naturally and simply."
Late 1980s–early 1990s: dry spell; equations known only approximately, approximations inadequate; many leave.
1995: at Strings '95 (University of Southern California), Edward Witten launches the second superstring revolution (Ch 12–13).
Schwarz kept faith because "the mathematical structure... Was so beautiful and had so many miraculous properties that it had to be pointing toward something deep." Michael Green: when you see "almost all of the major developments in physics over the last hundred years emerge... From such a simple starting point, you realize that this incredibly compelling theory is in a class of its own."
Oxford, October 1984 Greene started graduate school at Oxford in October 1984 amid "a pervasive feeling... That there was little or no future for particle physics": the standard model was nearly confirmed, and going beyond it (including gravity, explaining its 19 unexplained input numbers) seemed too daunting. Six months later the mood had flipped. News of Green and Schwarz reached even first-year students, who "worked deep into the night" mastering the physics and mathematics strings required.
What are strings made of?
Strings are about the Planck length, roughly 1020 times smaller than an atomic nucleus. Seeing directly that they aren't points would need an accelerator about a million billion times more powerful than any built.
- Answer 1: nothing. Strings are the true Greek "atoms," the last Russian doll. Asking their composition is like asking what letters are made of: in language, letters are the end of the line.
- Answer 2: maybe something. If string theory is wrong or incomplete, strings could be one more layer of the cosmic onion. Theorists have "intriguing hints" of substructure but no evidence.
- Greene adopts Answer 1 for most of the book.
Unification through vibration
The standard model, like a portfolio tracker that needs your initial investments, requires its particle list and 19 parameters as input and could have accommodated different ones; it is "too flexible" to explain them. String theory is "a unique and inflexible theoretical edifice," needing just one number to set the scale (the string tension).
The central fact. As a violin string's resonant patterns (whole numbers of evenly spaced peaks and troughs) sound as different notes, a fundamental string's resonant patterns appear as different particles, with different masses and force charges.
- Mass: more frantic vibration (bigger amplitude, shorter wavelength) means more energy, so via E = mc², more mass.
- Charges: electric, weak and strong charges are fixed by finer details of the vibration.
- Messengers: photons, W/Z bosons and gluons are vibration patterns too, and one pattern matches the graviton exactly.
- One "stuff." Before, each particle was "cut from a different fabric." Now every particle is one string and all strings are identical; they differ only in their notes. The universe is a cosmic symphony: a candidate theory of everything.
In the mid-1980s some declared the T.O.E. Found; a decade on, Greene calls that euphoria premature. We cannot yet compute the vibration spectrum precisely enough to compare with the particle tables. Under stated assumptions (Ch 9), string theory gives worlds qualitatively like ours; detailed numbers are out of reach.
Three consequences of stiff strings
Scherk and Schwarz could infer the string tension indirectly: the graviton pattern's force strength is inversely proportional to tension, and gravity is feeble, so the tension is colossal: ~1039 tons, the Planck tension.
Stiff strings: three consequences
- Tiny size. With nothing pinning their ends, enormous tension shrinks loops to the Planck length, ~10−33 cm.
- Huge energies. A stiffer string needs more energy to vibrate (a piano string vs. A violin string). Quantum mechanics makes vibrational energy come in whole-number multiples of a minimum proportional to the tension, so typical string masses are multiples of the Planck mass, ~1019 proton masses, about a grain of dust or a million bacteria.
- Infinitely many heavy particles. Every vibration pattern is a particle, so there's an endless tower; almost all are far heavier than the Planck mass, utterly beyond accelerators (Greene cites reach of about 1,000 proton masses), though some might survive as relics of the big bang (Ch 9).
Strings still make light particles like electrons, thanks to quantum mechanics again: a string can never be perfectly still, and its quantum jitter carries negative energy that cancels about one Planck energy from the lowest patterns. For the graviton the cancellation is perfect, giving a massless particle, as gravity's light-speed propagation requires.
The Price Is Right, Planck edition Even the heavy top quark (Greene's figure: ~189 proton masses) needs the Planck-scale energy cancelled to better than one part in 1017. It's as if Bob Barker handed you ten billion billion dollars and asked you to spend all but $189, exactly, without knowing the prices. Approximate calculations show such cancellations can occur; verifying them precisely is beyond current theory. The light particles we know are "the fine mist above the roaring ocean of energetic strings."
How strings tame gravity: the rough answer
We learn structure by throwing things at it: eyes catch bounced photons; accelerators smash particles. A probe can't resolve features smaller than itself.
Slim and Jim draw a peach pit blindfolded
- Contest. Each must draw a peach pit without looking, only by firing things at it and watching how they deflect.
- Round 1. Jim secretly loads Slim's shooter with marbles and his own with 5 mm pellets: Slim draws a smooth lump; Jim captures the major ridges.
- Round 2. Slim switches to 0.5 mm pellets and draws the finest corrugations: he wins.
- Moral. Useful probes can't be much larger than the features examined. In the quantum world a probe's resolution is its wavelength, which shrinks as energy rises: point particles can, in principle, probe ever smaller.
Strings break that rule. In 1988 David Gross (then Princeton) and his student Paul Mende showed that pumping energy into a string first sharpens it like a point particle, but beyond the Planck scale the extra energy makes the string grow, blurring it, potentially to macroscopic size at big-bang energies. A string has two sources of smearing: quantum jitter (which energy reduces) and its own extent (which energy increases). So nothing made of strings can probe below the Planck length.
Polished granite. Granite is grainy and bumpy at the microscopic level, yet feels perfectly smooth under your fingers, which "smear out" the bumps. Strings likewise smear out sub-Planckian quantum foam, enough to cure the incompatibility and eliminate the infinities. And unlike granite, which an electron microscope can expose, in string theory no finer probe exists. A positivist would say the violent sub-Planck fluctuations do not exist.
A sleight of hand?
Critics ask whether string theory has just dodged the problem. Greene answers no, for two reasons:
- The problem was of our own making. The violent sub-Planck fluctuations are an artefact of the point-particle assumption, which forced us to consider arbitrarily small distances. String theory says there is a limit to how finely distance applies.
- Extended objects are hard. Pauli, Heisenberg, Dirac and Feynman all considered "blobs" instead of points, but such theories kept violating conservation of probability (objects vanishing without trace) or causality (faster-than-light signals). String theory, after twenty years of scrutiny, respects every such principle and contains gravity.
The more precise answer: smeared interactions
When an electron and positron collide, they annihilate into a photon, which travels briefly and produces a new pair. For point particles, the annihilation happens at one definite place and time, which all observers agree on. For strings, the two incoming loops merge into one (the photon pattern), which later splits; filmed with an open shutter, the history traces a smooth tube-like surface, the string world-sheet.
Slice the world-sheet into "now" planes. George and Gracie, in relative motion, slice at different angles (Ch 2: simultaneity is relative), so they disagree on where and when the two strings first touch. A string interaction has no unique location; it is smeared across the world-sheet. Point particles cram the force's punch into a point; with gravitons, that gives infinite answers. Strings spread the punch out, and calculations give finite answers.
Beyond strings?
If points are an idealisation, infinitely thin strings might be one too. Thick "inner tubes" or "doughnuts" ran into the old Heisenberg–Dirac problems. But in the mid-1990s it emerged that string theory itself contains higher-dimensional objects: two-dimensional "Frisbees," three-dimensional "blobs" and more (the branes of Ch 12–13). String theory "is not a theory that contains only strings."
What "finite" and "unique" really mean. Two of the chapter's strongest claims need qualifiers. Finiteness: string perturbation theory is proven finite to two loops (D'Hoker and Phong, 2002) with strong arguments beyond that, but there is no complete proof at all orders, and the perturbative series itself does not converge. Uniqueness: "requires no input beyond a single number" is true of the equations, but those equations turned out (2000s) to have an enormous number of solutions, the landscape of perhaps 10500 or more vacua, each with different particles and constants. The rigid theory yields a flexible set of possible worlds, so in practice string theory has not explained the 19 (now ~26) parameters. Chapter 16 of this edition takes this up.
The energy gap, updated. The LHC (from 2010; 13.6 TeV collisions since 2022) reaches about 14,500 proton masses of collision energy, so Greene's "thousand proton masses" is out of date, but the gap to the Planck scale (~1019 proton masses) remains about fifteen orders of magnitude. The top quark mass is now measured at 172.6 GeV ≈ 184 proton masses.
Minimum resolution is a design principle. Strings say there is a smallest meaningful distance. Mature data architectures make the same move: they define a canonical grain (the transaction, the customer-day, the exposure) below which questions are declared meaningless rather than answered badly. Gross and Mende's lesson, that pushing harder past the natural scale makes your probe blurrier, is a fair description of over-granular analytics too.
The "Super" in Superstrings
Symmetry guides physics where experiment can't yet reach. After the familiar symmetries of space, time and motion there is exactly one more possible, supersymmetry, which pairs every particle with a superpartner whose spin differs by a half. Before strings, three arguments favoured it: aesthetics, taming the fine-tuning of quantum corrections, and making the three forces' strengths meet at one point. In string theory supersymmetry is not optional: only the supersymmetric string has matter particles and no fatal tachyon, so if string theory is right, supersymmetry is expected too. The embarrassment of 1985: there were five superstring theories.
"Sorry for the dear Lord"
When Hendrik Lorentz telegrammed Einstein that Eddington had confirmed light bending, a student asked what he'd have thought had it failed. Einstein: "Then I would have been sorry for the dear Lord, for the theory is correct." Aesthetics don't arbitrate science; experiment does. But while a theory is incomplete, physicists must still choose directions, guided by consistency, qualitative resemblance to the world, and a sense of elegance. As theories reach realms ever harder to test, that aesthetic becomes a steering tool, with no guarantee it leads to truth.
Symmetries of physical law
The symmetries of space, time and motion
- Time-translation: laws don't change from moment to moment (or only extremely slowly); otherwise "law" would mean nothing.
- Space-translation: laws are the same everywhere; a pogo-stick astronaut's lunar leaps reflect the moon's lower mass, not different laws.
- Constant-velocity motion (special relativity): all uniformly moving observers share the same laws.
- Accelerated motion (general relativity / equivalence principle): with gravity included, all observers are on equal footing.
- Rotation: laws don't depend on the orientation of your equipment.
- The limit: in 1967 Sidney Coleman and Jeffrey Mandula proved no other space-time-motion symmetries can be combined with these in a realistic theory, except through a loophole involving spin.
"The elegance of rich, complex, and diverse phenomena emerging from a simple set of universal laws is at least part of what physicists mean when they invoke the term 'beautiful.'"
Spin
A point has no off-axis points, so classically it can't spin. But in 1925 George Uhlenbeck and Samuel Goudsmit explained atomic spectra by giving electrons magnetism from rotation (following André-Marie Ampère's century-old link between moving charge and magnetism). Quantum spin is "like rotational motion, with a quantum-mechanical twist": it never speeds up, slows or stops. It is intrinsic, like mass or charge: "If an electron were not spinning, it would not be an electron."
| Particle type | Spin | Class |
|---|---|---|
| All matter particles (quarks, leptons) and antiparticles | ½ | Fermions (after Enrico Fermi) |
| Photon, W, Z, gluons | 1 | Bosons (after Satyendra Bose) |
| Graviton (predicted) | 2 | Boson |
| Higgs boson (found 2012) | 0 | Boson |
In string theory, spin, like mass and charge, is set by the vibration pattern. This is why Scherk and Schwarz were sure in 1974: strings necessarily contain a massless spin-2 pattern, the graviton's hallmark. "Where there is a graviton there is also gravity."
Supersymmetry and superpartners
By about 1971 physicists found that spin permits exactly one more symmetry: supersymmetry, loosely a shift of viewpoint in a "quantum-mechanical extension of space and time." Its key consequence: particles come in pairs whose spins differ by ½. It looks like it pairs matter with forces, but in the mid-1970s none of the known particles fit as each other's partners. Instead every known particle needs a new superpartner:
| Known particle | Spin | Superpartner | Spin |
|---|---|---|---|
| Electron | ½ | Selectron | 0 |
| Quarks | ½ | Squarks | 0 |
| Neutrinos | ½ | Sneutrinos | 0 |
| Photon | 1 | Photino | ½ |
| Gluons | 1 | Gluinos | ½ |
| W, Z | 1 | Winos, zinos | ½ |
| Graviton | 2 | Gravitino | 3/2 |
| Higgs | 0 | Higgsinos | ½ |
Doubling the particle list with none observed invites a variant of Rabi's quip: "nobody ordered supersymmetry."
The case for supersymmetry before strings
Three arguments
- Aesthetic completeness. It would be a shame if nature used almost every possible symmetry but not the last, like Bach leaving out the final, resolving measure.
- Taming quantum corrections (the hierarchy problem). In the standard model, quantum frenzy forces some parameters to be fine-tuned to better than one part in 1015: like aiming a rifle so the bullet hits a target on the moon within an amoeba's thickness. Bosons and fermions contribute opposite-signed quantum jitters, like two ends of a seesaw; pair them up and the pernicious effects cancel automatically.
- Grand unification. Supersymmetry makes the three non-gravitational force strengths meet exactly at one point (below).
Grand unification and running couplings
Force strengths differ hugely: electromagnetism under 1% of the strong force, the weak force about 1,000 times feebler still, gravity about 10−35 times weaker again. In 1974 Howard Georgi and Sheldon Glashow proposed a grand unified theory merging strong, weak and electromagnetic forces, visible only at about 1028 K (versus 1015 K for electroweak unification), energies about 1015 proton masses, roughly four orders below the Planck mass.
Force strengths run with distance. Around an electron, a "mist" of virtual particle–antiparticle pairs screens its charge like fog around a lighthouse; get closer and electromagnetism grows stronger. In 1973 David Gross and Frank Wilczek (Princeton) and David Politzer (Harvard) found the opposite for the strong (and weak) force: the quantum cloud amplifies it, so at shorter distances it weakens (asymptotic freedom). Georgi, Helen Quinn and Steven Weinberg (1974) showed these trends drive all three strengths together at about 10−29 cm (only 10,000 times the Planck length), when the universe was about 10−39 seconds old: three forces "melting" into one, like metals in a furnace.
In 1991 Ugo Amaldi (CERN), Wim de Boer and Hermann Fürstenau (Karlsruhe) recomputed the extrapolation with improved data and found two things: the three lines almost, but not quite, meet; and with supersymmetry's extra particles in the quantum mist, they meet exactly. To many, nature choosing "almost" unification is like a jigsaw whose last piece is slightly misshapen. The same calculations suggest superpartners are heavy, perhaps 1,000 proton masses or more, which would explain their absence from 1999's accelerators.
None of the arguments is airtight, as Greene concedes. The universe needn't be maximally symmetric; nature might walk "the fine edge between self-consistency and self-destruction"; force strengths needn't match exactly; and the simplest reason superpartners haven't been seen is that they don't exist.
Supersymmetry in string theory
Late 1960s: the original bosonic string theory: only whole-number spins (no matter particles) and a tachyon, a vibration with negative mass-squared that makes the theory inconsistent.
1971: Pierre Ramond (University of Florida) adds fermionic vibrations; André Neveu and John Schwarz extend it. Bosonic and fermionic patterns turn out to come in pairs.
1973: Julian Wess and Bruno Zumino carry supersymmetry over to point-particle field theory, launching supersymmetric field theory and later the supersymmetric standard model; so the mainstream's SUSY "owes a great debt to string theory."
1977: Ferdinando Gliozzi (Turin), Joël Scherk and David Olive (Imperial College) show the pairing is supersymmetry and that the superstring has no tachyon. Superstring theory is born.
Only the supersymmetric string avoids the tachyon and contains fermions for matter, so if string theory is right, supersymmetry should be too.
A super-embarrassment of riches
Someone who solves Amelia Earhart's disappearance is credible; someone with five equally convincing solutions is not. "In the arena of fundamental explanations, more is definitely less." By 1985 supersymmetry could be built into string theory in five ways:
The five superstring theories (1985)
- Type I (open and closed strings)
- Type IIA
- Type IIB
- Heterotic O(32) ("oh-thirty-two")
- Heterotic E8 × E8 ("e-eight times e-eight")
- Witten: "If one of the five theories describes our universe, who lives in the other four worlds?" The physicist's dream is one theory, the only possible one. Chapter 12 shows the five are five descriptions of one theory (M-theory).
Supersymmetry and the LHC: the big disappointment. Greene expected superpartners near "a thousand proton masses." The LHC searched from 2010 until Run 3 ended in June 2026, at up to 13.6 TeV, and found none: gluinos and first-generation squarks are excluded up to roughly 2–2.4 TeV, top squarks to about 1.2–1.3 TeV in simple models. The 125 GeV Higgs is compatible with supersymmetry only if top squarks are fairly heavy, which re-introduces fine-tuning, the very problem SUSY was meant to cure (the "naturalness crisis"). Dark-matter detectors (LZ, XENONnT) have also found no supersymmetric WIMP. Supersymmetry may still exist at higher energies (string theory needs it at some scale), but the 1990s hope of low-energy SUSY, Greene's main "smoking gun," has largely faded. Gauge-coupling unification with SUSY still works if superpartners sit at a few TeV or somewhat above.
Elegance is a prior, not proof. The supersymmetry story is a cautionary tale for architecture decisions made on beauty: three independent "it would be a shame if not" arguments pointed the same way, a generation invested, and the data said no (so far). In technology strategy, rank options by evidence and option value, keep elegant bets cheap to abandon, and set kill criteria before the expensive experiment runs.
More Dimensions Than Meet the Eye
Einstein settled two of the century's great conflicts and rewrote space and time each time. String theory settles the third and demands a third rewrite: the number of dimensions itself. Dimensions come in two kinds, large and extended or small and curled up, and curled-up ones can hide below any instrument's reach. Theodor Kaluza showed in 1919 that one extra dimension unites gravity with electromagnetism; string theory goes further and requires extra dimensions, six (later seven) of them, to keep its probabilities sensible. Because strings are small enough to vibrate through the curled-up dimensions, the shape of those dimensions fixes the particle masses and charges we measure: geometry becomes the source of physics.
The illusion of the familiar
Experience sets the frame for interpretation: the "wild child" raised by wolves reads the world differently from you, and so, in lesser ways, do people from different cultures. The hardest beliefs to challenge are the ones everyone shares. One of them: you can move in three independent directions (left-right, back-forth, up-down), and any location needs three numbers, like a street, a cross street and a floor. Add Einstein's time and you have four dimensions. In 1919 a little-known Polish mathematician at the University of Königsberg, Theodor Kaluza, "had the temerity to challenge the obvious." "Sometimes silly-sounding suggestions are plain silly. Sometimes they rock the foundations of physics."
Two kinds of dimension: the garden hose
The garden-hose model of hidden dimensions
- From a quarter mile away, a few hundred feet of hose stretched across a canyon looks one-dimensional: an ant on it seems to need one number (distance from the end) to locate it.
- Through binoculars, the hose's girth appears: the ant can also walk the "clockwise-counterclockwise dimension," so the surface is really two-dimensional.
- The lesson: dimensions are either large, extended and manifest or small, curled up and hard to detect. A hose as thin as a hair or capillary would hide its circular dimension far better.
- Kaluza and Klein's proposal: space has three large dimensions plus one tiny circular dimension present at every point, like loops in the pile of a tightly woven carpet. Locating a small enough ant would need four spatial numbers, five with time.
- Not a bump but a direction. The circle is not a little loop sitting in ordinary space; it is a new, independent direction existing everywhere, just as up-down exists everywhere.
How small is small? The best equipment of 1999 resolved about a billionth of a billionth of a metre (10−18 m); any dimension curled up smaller is invisible. In 1926 the Swedish physicist Oskar Klein combined Kaluza's idea with the new quantum mechanics and found the circle might be as small as the Planck length, far beyond experiment. Hence Kaluza-Klein theory.
Life in Lineland
Physicists train their intuition by imagining fewer dimensions, following Edwin Abbott's 1884 satire Flatland.
Kaluza K. Line and the end of spatial enslavement
- Lineland. Live on a garden-hose universe whose circle is too small to notice and you'd think space has one dimension. Your body must have length only, with an eye at each end forever staring along the line.
- Drudgery. Your neighbour is a single featureless dot-eye; you can never pass her, so the order of Linebeings is fixed for ever.
- The prophet. A Linebeing named Kaluza K. Line proposes that Lineland secretly has a tiny circular second dimension, which might one day grow (his colleague Linestein says it's possible).
- What would change. Your eyes are your skin: surgery in Lineland goes through the eyes. If the circle grew, a neighbour could look into your body from the side, doctors could reach straight in, and Linebeings would evolve skin and breadth, becoming Flatbeings with Abbott's rich two-dimensional culture.
- The refrain: why stop there? Flatland might hide a third curled dimension; if it grew, that could be our world. Our own three dimensions might even curl back as giant circles beyond telescope range. And our three might hide a fourth, Kaluza and Klein's proposal.
Even if extra dimensions stay small for ever, Greene stresses, their existence has profound consequences.
Unification in five dimensions
Kaluza rewrote general relativity for one extra space dimension. The equations for the three ordinary dimensions came out essentially identical to Einstein's; the extra equations were exactly Maxwell's equations of electromagnetism. By adding one curled-up dimension, Kaluza had united Einstein's gravity with Maxwell's light. Gravity is a ripple in the familiar dimensions; electromagnetism is a ripple involving the curled-up one. Before him, nothing had hinted that the two forces were related.
21 April 1919: Einstein replies that unification "through a five-dimensional [four space and one time] cylinder-world" had never occurred to him: "At first glance, I like your idea enormously."
About a week later: second thoughts: "Nowhere, so far, can I see an impossibility. On the other hand, I have to admit that the arguments brought forward so far do not appear convincing enough."
14 October 1921: "I am having second thoughts about having restrained you from publishing... If you wish, I shall present your paper to the academy after all." Kaluza has the master's approval, two years late.
1926: Oskar Klein adds quantum mechanics and sizes the circle near the Planck length.
Late 1920s: detailed study shows the simplest version predicts an electron mass-to-charge relation wildly at odds with measurement. Interest fades; Einstein dabbles on and off.
Why the idea slept for fifty years. The 1920s began a bull market for microphysics: quantum mechanics, quantum field theory, the strong and weak forces, all generating testable predictions. Speculation about structure "far too small to be probed by even the most powerful of instruments" took a back seat. By the late 1970s the standard model was largely confirmed, the bull market had lost steam, and repeated failures to quantise gravity left the community open to radical ideas. Kaluza-Klein theory was resuscitated.
Modern Kaluza-Klein theory and supergravity
Physicists argued Kaluza had been too conservative: he knew only two forces, and more forces need more dimensions. By the mid-1970s research centred on theories with several curled-up dimensions in various shapes:
- A sphere: two extra dimensions curled into the surface of a ball, attached at every point of ordinary space.
- A torus: two extra dimensions shaped like a hollow doughnut.
- More: three, four, five or more extra dimensions in exotic shapes, all smaller than any probe.
The most promising combined extra dimensions with supersymmetry, hoping superpartner cancellations would calm the quantum jitters: higher-dimensional supergravity. The extra-dimensional equations looked strikingly like those of electromagnetism and the strong and weak forces. But "the old conundrums persisted": supersymmetry softened but did not cure the short-distance quantum undulations, and no single theory captured all features of forces and matter. Pieces of a unified theory were surfacing with nothing to tie them together, until string theory took centre stage in 1984.
What Greene leaves out about supergravity. The high point was eleven-dimensional supergravity (Cremmer, Julia and Scherk, 1978), the largest dimension in which supersymmetry allows a consistent theory with gravity. It had two problems besides infinities: in 1981 Witten showed that compactifying from eleven dimensions cannot produce the left-right asymmetry (chirality) of the weak force, and its four-dimensional infinities were expected to appear at higher loops. This is why eleven dimensions fell out of favour in the 1980s, and why their return in M-theory (Ch 12) was such a surprise.
Why string theory requires extra dimensions
Postulating tiny dimensions might seem as arbitrary as postulating a microscopic civilisation of tiny green people: neither can be ruled out. String theory changes that by needing them.
From negative probabilities to ten dimensions
- Probabilities must lie between 0 and 1. Infinite probabilities signalled the GR–QM clash; strings cure those. But early string calculations gave negative probabilities, also nonsense.
- Vibration directions matter. A string confined to a surface can vibrate in two independent directions; in three-dimensional space, three; in more dimensions, more.
- The magic number. The troublesome terms are highly sensitive to the number of directions: if a string can vibrate in nine independent spatial directions, all negative probabilities cancel.
- The Kaluza-Klein loophole. Strings are tiny enough to vibrate in curled-up dimensions too, so we need three large plus six curled-up space dimensions.
- Result: nine space plus one time = ten dimensions. Kaluza's 1919 proposal "finds its most convincing and powerful forum": string theory requires extra dimensions instead of merely allowing them.
Three open questions
- Why nine? No one has an intuitive, non-technical explanation; it falls out of a calculation. Ernest Rutherford held that if you can't explain a result simply, you don't fully understand it: not that it's wrong, but that its origin and meaning remain unclear. A parenthetical warning: the ten-dimension calculation is approximate. In the mid-1990s Witten, building on Michael Duff (Texas A&M) and Chris Hull and Paul Townsend (Cambridge), argued it misses one space dimension: the full theory has ten space and one time dimension, eleven in all (Ch 12).
- Why are exactly three large? No one knows. One picture: all dimensions start curled up and, in a big-bang-like event, three space and one time dimension unfurl while the rest stay small. Rough arguments exist (Ch 14) but are "only in the formative stages." Showing this emerges from the theory is "a primary goal of modern research."
- Could some extra dimensions be time? A curled-up time dimension would mean returning to a past instant after each lap, "well beyond" experience. Some theorists explore extra time dimensions; the results are inconclusive, and Greene assumes all extra dimensions are spatial.
| Extended dimensions | Curled-up dimensions |
|---|---|
| Large, manifest, three of them (plus time) | Tiny, possibly near the Planck length; six in superstring theory, seven in M-theory |
| Where we, our hands and our telescopes live | A string can vibrate through them; your hand circles them countless times in one sweep without noticing |
| Shape set by matter and energy through general relativity | Shape fixes which vibrations are possible, so it sets particle masses and charges |
Geometry fixes physics
The chapter's deepest point. A string's resonant patterns depend on its surroundings:
- Ocean waves roam freely in open water but are shaped by depth, rocks and canals in cramped channels.
- An organ pipe or a French horn sounds the notes its interior shape allows.
- Curled-up dimensions likewise twist and constrain which vibration patterns a string can sustain.
Since vibration patterns appear as masses and force charges, the size and shape of the extra dimensions determine the fundamental properties of particles in our three large dimensions. "That's one of the most far-reaching insights of string theory." Greene says it "once again, with feeling."
Calabi-Yau shapes
The extra dimensions can't be crumpled any which way: string theory's equations tightly restrict their geometry. In 1984, Philip Candelas (University of Texas at Austin), Gary Horowitz and Andrew Strominger (UC Santa Barbara) and Edward Witten showed that a particular class of six-dimensional shapes works: Calabi-Yau spaces, named for mathematicians Eugenio Calabi (University of Pennsylvania) and Shing-Tung Yau (Harvard), whose earlier, purely mathematical work on them became central.
- Drawing them distorts them. Any picture squashes six dimensions onto a page; Greene's figure gives only "the rough idea."
- Rare but numerous. Greene cites "many tens of thousands" of qualifying shapes, a big club but tiny next to the infinity of mathematically possible shapes.
- Everywhere at once. Replace each sphere attached to ordinary space with a Calabi-Yau: six tiny dimensions at every point. Sweep your hand in an arc and you circumnavigate them an enormous number of times, returning to your start again and again; it all averages out, so you feel nothing.
- The question that follows: what particle properties come from strings vibrating through these shapes, and do they match experiment? "This is string theory's $64,000 question."
Counting hidden dimensions
- Bosonic string (1970s): 26 spacetime dimensions; no fermions, a tachyon.
- Superstring (1984–94): 10 = 3 large space + 6 curled (Calabi-Yau) + 1 time.
- M-theory (1995 on): 11 = 3 large + 7 curled + 1 time.
- Kaluza-Klein (1919–26): 5 = 3 + 1 circle + 1 time.
- Calabi-Yau threefolds known today: at least 473.8 million distinct toric constructions (Kreuzer–Skarke, 2000), with the true number of shapes possibly unbounded; Greene's "tens of thousands" was the 1999 count.
Large extra dimensions: the 1998–99 twist. Shortly before and after this book appeared, two proposals showed extra dimensions might be much bigger than the Planck length if ordinary particles are stuck to a brane (Ch 12–13) and only gravity roams the extra space:
- ADD (Arkani-Hamed, Dimopoulos, Dvali, 1998): extra dimensions as large as a millimetre could explain gravity's weakness.
- Randall–Sundrum (1999): a single warped extra dimension could do the same.
Both predicted deviations from Newton's inverse-square law at short range and missing-energy events or new resonances at colliders. The results so far are negative:
- Torsion-balance experiments (University of Washington's Eöt-Wash group) confirm the inverse-square law down to about 50 micrometres (2020).
- The LHC has found no Kaluza-Klein gravitons, microscopic black holes or missing-energy excess, pushing the fundamental gravity scale in ADD models above roughly 10 TeV.
Extra dimensions, if they exist, are small or well hidden.
The Calabi-Yau count became the landscape. Greene's "$64,000 question" assumed the theory would eventually pick out one shape. Later work found the opposite:
- Choice of shape: Calabi-Yau spaces number at least in the hundreds of millions.
- Fluxes: each shape can be threaded with fluxes (generalised magnetic fields wrapping its holes), giving perhaps 10500 or more stable configurations (Bousso–Polchinski 2000; KKLT 2003).
- F-theory geometries: one 2015 estimate (Taylor–Wang) reached 10272,000.
So the theory's equations do not obviously select our world's geometry. Since 2018, machine-learning methods have been used to scan Calabi-Yau databases and compute their metrics numerically, a live research area. The question remains open; Chapter 16 returns to it.
"Requires" deserves a footnote. The critical dimension (10 or 26) follows from demanding a consistent, Lorentz-invariant quantised string in flat background spacetime. Non-critical strings in other dimensions exist mathematically (Polyakov 1981), with an extra field compensating, though they don't give realistic physics. And the claim that the six extra dimensions form a Calabi-Yau rests on assuming unbroken supersymmetry at the compactification scale, which today's data neither confirm nor require. The extra dimensions remain a theoretical deduction with no experimental sighting.
Hidden dimensions shape visible outcomes. Greene's key move, that the unseen geometry of curled-up space fixes the measurable properties of particles, has a practical twin: in an organisation, the unseen structure (incentive design, data lineage, reporting lines) determines the visible numbers (conversion rates, losses, delivery speed). Before you try to tune the outputs, map the hidden dimensions that constrain which "vibration patterns" are even possible.
The Smoking Gun
No theory describes the world until experiment says so, and a beautiful theory that fails is "no more relevant than an elaborate game of Dungeons and Dragons." String theory may be the most predictive theory ever conceived, yet in 1999 it could not compute numbers precise enough to confront data: physicists hold "the Holy Grail" but are missing pages of the instruction manual. The obstacle is the choice of Calabi-Yau shape, which no one can derive. So string theory offers frameworks rather than predictions: the number of particle families could be the number of holes in the extra dimensions, and generic signatures (superpartners, fractional charges, cosmic strings) could reveal it, but none was in hand.
Gravity as a "prediction"
Edward Witten: "String theory has the remarkable property of predicting gravity." Newton and Einstein built gravity theories because they saw gravity; a string theorist ignorant of general relativity would be led to it inexorably by the massless spin-2 graviton vibration. Witten calls this "one of the greatest theoretical insights ever," and concedes it is really a postdiction, an accident of Earth's history: elsewhere, he muses, a civilisation might find string theory first and gravity as its consequence. Most physicists want more: a genuine prediction experiment can confirm, or a postdiction of something unexplained (the electron's mass, three families).
Crossfire: the critics and the converts
| Sceptics (mid-to-late 1980s) | Enthusiasts |
|---|---|
| Sheldon Glashow and Paul Ginsparg (Harvard): superstring theorists "pursue an inner harmony, where elegance, uniqueness and beauty define truth... Do mathematics and aesthetics supplant and transcend mere experiment?" | Edward Witten: learning how strings join gravity and quantum mechanics was "the greatest intellectual thrill" of his life |
| Glashow: "so ambitious that it can only be totally right, or totally wrong," and we "won't know which for decades"; he even asked whether string theorists should "pervert impressionable students," likening the field to medieval theology | Cumrun Vafa (Harvard): "definitely revealing the deepest understanding of the universe which we have ever had" |
| Richard Feynman: "there is more than one way to skin a cat... The fact that a theory gets rid of infinities is to me not a sufficient reason to believe its uniqueness" | Murray Gell-Mann: "a fantastic thing"; expects some version to be "the theory of the whole world" |
| Howard Georgi: the "siren call" of ultimate unification beyond experiment would lose "that... Process of pruning of irrelevant ideas" | Glashow, a decade later: "the only something else I know of is string theory" |
The split is partly physics, partly philosophy of method: "traditionalists" want theory tied closely to experiment; others think we can now tackle questions beyond present technology. By the late 1990s the criticism had softened:
- Glashow: string theorists had become "more prudent with their enthusiasm," and "we non-string theorists have not made any progress whatsoever in the last decade... The only something else I know of is string theory."
- Georgi: string theory "has gotten oversold," but its ideas proved useful in his own work.
"It used to be that as we were climbing the mountain of nature the experimentalists would lead the way. We lazy theorists would lag behind. Every once in a while they would kick down an experimental stone which would bounce off our heads... But now we theorists might have to take the lead. This is a much more lonely enterprise." David Gross
Greene frames the gap as a "historical asynchrony": theoretical ropes and crampons for the summit partly exist; experimental ones don't. The goal is to kick a theoretical stone down to the experimentalists' base camp. "No stones have as yet been dislodged... But a few tantalizing and promising pebbles certainly have."
Why strings can't be seen directly
The energy wall
- Reach in 1999: about 10−18 m, with accelerators "a few miles in size."
- The gap: the Planck length is about 17 orders of magnitude smaller, so today's technology would need an accelerator the size of the galaxy; Shmuel Nussinov (Tel Aviv) argued it would need one the size of the universe.
- Energy is not the problem; focus is. The Planck energy is about a thousand kilowatt-hours, an air conditioner running ~100 hours (more precisely about 540 kWh). The impossible part is putting it on a single string.
- Politics: Congress cancelled the Superconducting Supercollider (54-mile ring) in 1993, so "don't hold your breath" for a Planck-probing machine.
- Conclusion: any test must be indirect, through consequences visible at far larger scales than strings.
Why three families? Count the holes
Candelas, Horowitz, Strominger and Witten (1985) found that each hole in a Calabi-Yau shape (like the hole in a phonograph record, a doughnut, or a multi-doughnut; in six dimensions, holes of many dimensions) carries one family of lowest-energy string vibrations. A Calabi-Yau with three holes gives three families: the family structure would reflect the topology of the extra dimensions, not chance or divine whim. "This is the kind of result that makes a physicist's heart skip a beat."
- The catch. Known Calabi-Yau shapes have 3, 4, 5, 25 holes, some as many as 480. Nothing yet selects the right one. Choose a three-holed shape and this would be an impressive postdiction; finding the selection principle remains unsolved.
- Masses too. Strominger and Witten showed particle masses in each family depend on how the boundaries of the multidimensional holes intersect and overlap. String theory frames why the electron weighs what it does, a question earlier theories are "completely silent" on.
- Forces too. One massless spin-2 pattern (the graviton) appears for every Calabi-Yau, but the number of spin-1 messengers, their strengths and gauge symmetries depend on the shape.
- Why can't we choose? Physicists can only do perturbation theory (Ch 12), in which every Calabi-Yau looks equally good. Better, non-perturbative methods might single one out (Ch 13).
Exhausting the possibilities
One tempting shortcut is to catalogue every shape and look for a match. Two obstacles stand in the way, and one result encourages:
Why the catalogue approach stalls
- Infinite deformations. A multi-handled doughnut can be smoothly deformed through infinitely many shapes without changing its holes; the "tens of thousands" of Calabi-Yau spaces already lump each deformation family together. But masses and couplings depend on those deformations, and no graduate-student army can compute an infinite list.
- Approximate equations. Even for one shape, the approximations give only ballpark answers. Precise numbers like the electron mass need delicate Planck-scale cancellations (the Price Is Right problem of Ch 6), and small errors ruin them.
- The encouraging result. Most shapes give the wrong world (wrong family counts, wrong forces), but a few give physics qualitatively close to the standard model, with gravity stitched in. Greene calls this "the best we could have hoped for": too many matches would make none compelling; none would make string theory irrelevant.
Signature 1: superpartners
Some predictions are generic: true whatever the unknown details. The chief one is supersymmetry: every known particle should have a superpartner with known charges but unknown mass. None had been seen. Hopes rested on the Large Hadron Collider, due "before 2010." John Schwarz: "Supersymmetry ought to be discovered before too long. And when that happens, it's going to be dramatic."
Greene adds two caveats:
- Finding superpartners would not prove string theory: supersymmetry also lives in point-particle theories.
- Not finding them wouldn't kill it: they might lie beyond the LHC's reach.
He still calls a discovery "strong and exciting circumstantial evidence."
Signature 2: fractional electric charges
Known particles carry charges of 0, ±1, ±⅓, ±⅔. In string theory, if the extra dimensions contain holes that a string can disentangle from only by wrapping a set number of times, particles may carry exotic charges such as 1/5, 1/11, 1/13, 1/53: the winding number becomes the denominator. Less generic than superpartners (only some Calabi-Yau shapes do it), but far more distinctive: in point-particle theory such charges fit "as naturally as the proverbial bull in a china shop." Expected masses: probably near the Planck mass, hence unseen.
Some longer shots
Six further routes to evidence
- Cosmic strings. Big-bang energies might have made a few macroscopic strings, stretched by expansion to astronomical size, perhaps leaving a mark on the cosmic microwave background. Witten: "my favorite scenario... Nothing would settle the issue quite as dramatically as seeing a string in a telescope."
- Neutrino masses. The standard model makes neutrinos massless "for no particularly deep reason"; string theory should explain tiny non-zero masses if they're found.
- Forbidden processes. Proton decay and certain quark transmutations are forbidden in the standard model but may be allowed by strings; seeing one would signal new principles.
- New long-range forces. Some Calabi-Yau shapes give tiny new long-range force fields.
- Dark matter. String vibration spectra offer several candidates.
- The cosmological constant. Quantum vacuum fluctuations predict a value ~120 orders of magnitude too large: "a wonderful challenge and opportunity." String theory "as yet" had not met it.
An appraisal
Ideas once thought untestable, atoms, Pauli's neutrino, neutron stars and black holes, all became established. String theory, hailed as the biggest development since quantum mechanics, faces the same long road but without experimental light: quantum mechanics took ~30 years to reach logical form and ~20 more to absorb special relativity, with data at every step. "It's conceivable that one or more generations of physicists will devote their lives to... String theory without getting a shred of experimental feedback." Greene judges the risk worth taking for its simplicity, its taming of gravity, its unifying power, and its potential for limitless prediction.
The scorecard in 2026. Every signature Greene listed has now been tested to some degree.
| Signature (1999) | Status in 2026 |
|---|---|
| Superpartners at the LHC | None found at 13–13.6 TeV; gluinos excluded to ~2.3 TeV (Ch 7) |
| Fractional charges | None found; millicharged-particle searches (SLAC, milliQan at the LHC) set limits, though Planck-mass particles were never in reach |
| Cosmic (super)strings | None found; CMB (Planck satellite) limits Gμ ≲ 10−7; pulsar-timing arrays (NANOGrav and others, 2023) detected a gravitational-wave background, but merging supermassive black holes explain it best; LIGO–Virgo–KAGRA limits Gμ ≲ 10−11 for some models |
| Neutrino mass | Confirmed (Super-Kamiokande 1998, SNO 2001; Nobel 2015); KATRIN (2025) caps it below 0.45 eV; no string derivation |
| Proton decay | Not seen: Super-Kamiokande sets lifetime > 2.4 × 1034 years (p → e⁺π⁰), ruling out the simplest grand unified theories; Hyper-Kamiokande (from ~2027) and DUNE will push further |
| New long-range forces | None down to ~50 μm (Eöt-Wash); MICROSCOPE (2022) tested the equivalence principle to ~10−15 |
| Dark matter candidates | Dark matter still unidentified; WIMP searches (LZ 2024–25) empty; string-motivated axions ("the axiverse," 2009) now a major search target (ADMX, HAYSTAC) |
| Cosmological constant | Measured non-zero (accelerating expansion, 1998; Nobel 2011), ~10−122 in Planck units; DESI (2024–25) hints at evolving dark energy at 2.8–4.2σ; no string derivation of the value |
The 1998 discovery is missing. When this book appeared, two teams (Perlmutter; Riess and Schmidt) had already reported that the expansion of the universe is accelerating, implying a small positive cosmological constant. Greene's text still treats "zero or quite small" as open. The finding sharpened the problem for string theory: a positive cosmological constant is hard to build in string compactifications, which drove the KKLT construction (2003), the landscape and, later, the swampland conjectures (2018) arguing that stable de Sitter space may be impossible in string theory (Ch 16).
From "a few" to "a quadrillion" standard models. The catalogue approach was later industrialised.
- 2005: Braun, He, Ovrut and Pantev built a heterotic compactification with exactly the particle content of the supersymmetric standard model.
- 2012–14: Anderson, Gray, He and Lukas found tens of thousands of heterotic standard-model-like spectra by computer scan.
- 2019: Cvetič, Halverson, Lin, Liu and Tian found about a quadrillion (1015) F-theory geometries with the exact standard-model gauge group and three families.
But reproducing the particle list is not reproducing the numbers: none of these compactifications computes the electron mass or the quark mixing angles, and their abundance deepens the selection problem rather than solving it.
Postdiction, prediction and falsifiability. The chapter's honesty is its strength: Greene concedes string theory then had no confirmed prediction, and notes neither finding nor missing superpartners would decide the issue. Critics such as Peter Woit (Not Even Wrong, 2006) and Lee Smolin (The Trouble with Physics, 2006) argued that a framework immune to both outcomes fails the test of science. Defenders reply that string theory is a framework, like quantum field theory, not a single model, and that its tests come through consistency and spin-offs (holography, black-hole entropy, Ch 13). Twenty-seven years later, no smoking gun has appeared; the "Gross mountain" image remains apt.
Frameworks need selection rules before they make predictions. String theory's problem is a familiar one in modelling: a rich model class (all Calabi-Yau shapes) can fit almost anything, so fit alone is not evidence. Analytics teams face the same trap with over-parameterised models. The remedy is the same: pre-register what the model must predict (a number, not a qualitative match), and value a framework by the sharpness of what it forbids.
Quantum Geometry
Einstein was lucky: Bernhard Riemann's 1854 geometry of curved spaces was waiting on the shelf, ready to describe gravity. String theory has no such luck. Because strings are extended, the point-based geometry of Riemann fails at the Planck scale and a new quantum geometry must be built from scratch. Two discoveries show how strange it is. T-duality: a circular dimension of radius R and one of radius 1/R (in Planck units) give identical physics, so space cannot be squeezed below the Planck length and a cosmic crunch becomes a bounce. Mirror symmetry: two Calabi-Yau shapes with different geometry can give identical physics. String theory loosens Einstein's one-to-one link between geometry and physics, and in exchange solved problems mathematicians had been stuck on for decades.
The heart of Riemannian geometry
Riemann, building on Gauss, Lobachevsky and Bolyai, showed that the distances between all points of a shape determine its curvature: the more distance relations deviate from flat ones, the more curved the shape. Paint the Mona Lisa on a trampoline and stand on it: the image distorts most under your feet, where curvature is greatest, "yielding the hint of a grimace at the corner of her customary enigmatic smile." Einstein's genius was to declare that Riemann's mathematics is the physics of gravity.
But Riemann's framework is built on points, and the match between mathematics and physics gets better as test objects get smaller. String theory stops that: nothing is smaller than a string, so Riemannian geometry must be modified at the Planck scale. For galaxies modelled as points in cosmology the error is negligible; in the ultramicroscopic world it is total.
A cosmological playground
The big crunch, and a smaller test case
- 1999 cosmology. The universe began ~15 billion years ago and is expanding. Whether it re-collapses depended (then) on the average matter density versus the critical density, about 10−29 g/cm³, roughly five hydrogen atoms per cubic metre.
- Matter clumps like money. Using Earth's density as a guide would be like using Bill Gates's net worth to gauge the average person's finances; intergalactic space drags the average down. Visible matter falls short; dark matter's amount was unknown, so the fate was open.
- The crunch scenario. If density were high enough, galaxies would rush together until the universe shrank past a star, an orange, a grain of sand, an atom, and, in general relativity, to zero size.
- The playground. To see what strings change, Greene shrinks the circle of the Garden-hose universe (Ch 8) towards Lineland: a partial, simplified crunch.
The essential new feature: winding
A point particle on the hose can move along it or around it. A string can do both while vibrating, and something more: it can wrap around the circle, like a lasso, once or many times. This is a winding mode, with no point-particle counterpart.
Two sources of string energy on a circle
- Winding energy is proportional to R. A wrapped string has a minimum length (the circumference times the winding number), hence a minimum mass, growing with the radius. (Unwrapped strings' small minimum mass is cancelled by quantum effects, as in Ch 6; wrapped ones' is not.)
- Vibration (momentum) energy is proportional to 1/R. "Uniform vibrations," the string sliding around the circle, are confined by the circle; by the uncertainty principle's "quantum claustrophobia," a smaller circle means more energy.
- Ordinary vibrations (the wiggles that set charges) don't depend on R at all.
- So: big R means heavy windings, light vibrations; small R means light windings, heavy vibrations.
Keiji Kikkawa and Masami Yamasaki recognised the consequence in 1984: as the circle shrinks below the Planck length, all physics is identical to a circle larger than the Planck length and growing. The crunch is a bounce.
The insider-trading analogy
Two stocks that always move inversely
- The tip. A fitness-machine maker and a heart-valve maker both closed at $1. A reliable (and possibly illegal) source says tomorrow's prices will be exact reciprocals: $2 and 50¢, or $10 and 10¢, but not which goes up.
- Equal holdings. Split your money equally and you cannot lose: $4 and 25¢ make $4.25 per pair versus $2 yesterday. Which stock rose doesn't matter.
- Unequal holdings. Buy 1,000 fitness and 3,000 valve shares. Fitness at $10, valves at 10¢: $10,300. The reverse: $30,100.
- The friend's mirror portfolio. She holds 3,000 fitness and 1,000 valve. Her values are the same two numbers with the scenarios swapped. Swapping which price is high is exactly compensated by swapping the share counts.
- The string version. Winding and vibration numbers are the share counts; R and 1/R are the prices. Swap R ↔ 1/R and swap winding ↔ vibration numbers, and every total energy is unchanged.
The spectrum: R = 10 versus R = 1/10
Total energy (from these two sources) = winding number × R + vibration number ÷ R, in Planck units.
| Vibration no. | Winding no. | Energy if R = 10 | Energy if R = 1/10 |
|---|---|---|---|
| 1 | 0 | 1/10 | 10 |
| 0 | 1 | 10 | 1/10 |
| 2 | 1 | 10.2 | 20.1 |
| 1 | 2 | 20.1 | 10.2 |
| 3 | 2 | 20.3 | 30.2 |
| 2 | 3 | 30.2 | 20.3 |
| 4 | 1 | 10.4 | 40.1 |
| 1 | 4 | 40.1 | 10.4 |
Read down either energy column: the same list of numbers appears, in a different order. Swapping vibration and winding numbers maps one universe's spectrum onto the other's exactly. Add ordinary vibrations (radius-independent) and charges, and masses and charges in a universe of radius R are identical to those at 1/R. No experiment can tell the two universes apart: a fat Garden-hose universe and a thin one are the same universe.
George and Gracie argue about the radius Flattened into two-dimensional physics professors in the Garden-hose universe, George measures R = 10 and Gracie R = 1/10, each confirming a full table of string energies with "the new Planck energy accelerator." Gracie spots that the tables are identical: "what you would call a winding mode I would call a vibration mode... Nature does not care about the language we use." George, who "reasons a bit more slowly," gets there: radii related by inversion are physically indistinguishable.
Three objections, and two kinds of distance
- "Just use a tape measure!"
- "Didn't strings abolish sub-Planck distances?" Yet R = 10 implies a partner of 1/10.
- "Who cares about a toy hose?" The answer to this one is quick: with six circular extra dimensions (the simplest Calabi-Yau), each circle obeys R ↔ 1/R. Even our three large dimensions, if they close up into giant circles (an astronaut heading straight out would return "like Magellan"), have R ≈ 1061 Planck lengths, physically identical to 10−61, a universe that is tiny and shrinking as ours expands. "How can a six-foot tall human being 'fit' inside such an unbelievably microscopic universe?"
Two operational definitions of distance. In 1988 Robert Brandenberger (Brown) and Cumrun Vafa (Harvard) showed that on a circle, string theory has two ways to measure distance: time how long a probe takes to cross it, using either unwound strings or wound strings.
- Unwound probes roam the full circumference (∝ R), carry energy ∝ 1/R, and measure R.
- Wound probes have energy ∝ R and, by the uncertainty principle, are sensitive to 1/R: they measure 1/R.
- Why we never noticed: when R is far from 1, one probe is almost massless (easy) and the other billions of times heavier than a proton (impossible to make). We have always used the light probe, so our intuition knows only one distance. Photons are the light modes astronomers use.
- The human-in-a-tiny-universe puzzle dissolves. Six feet and 15 billion light-years are both measured with light modes; comparing your height with the heavy-mode size of the universe is apples and oranges.
A minimum size, and a bounce
Follow a circular universe shrinking, always measuring the easy way (with the lightest probe):
- Large R: unwound strings are light and measure a big, shrinking radius.
- R = 1 (the Planck length): wound and unwound strings have equal mass; both give 1, since 1 is its own reciprocal.
- R below 1: wound strings are now the light ones, and they measure 1/R, which is larger than 1 and growing.
The radius, measured the conventional way, falls to the Planck length and then rises: the big crunch is replaced by a bounce, and sub-Planck distances are not merely unobservable but physically identical to larger ones. This is "far more than... Semantics": the physics at sub-Planck R is identical to the physics at super-Planck 1/R. Brandenberger, Vafa and others proposed rewriting cosmology so that both the big bang and any crunch involve a Planck-sized universe, not a point (Ch 14).
How general? It works wherever strings can wind. On a sphere, a string can slip off like a rubber band off a basketball. String theorists generally believe a minimum size holds whenever a full dimension shrinks, but an isolated chunk of space can behave differently (Ch 11, 13). Proving this is an open goal.
Mirror symmetry
If radius R and 1/R are equivalent, differently shaped spaces might be equivalent too. In 1988 Lance Dixon (SLAC), amplified by Wolfgang Lerche (CERN), Vafa and Nicholas Warner (MIT), conjectured on symmetry and aesthetic grounds that two different Calabi-Yau shapes could give identical physics. A hint: the number of families depends only on the total number of holes, not how many holes of each dimension; two shapes with different hole counts per dimension but the same total would at least agree on families.
The mirror-symmetry cast
- Lance Dixon, Wolfgang Lerche, Cumrun Vafa, Nicholas Warner: the 1988 conjecture.
- Brian Greene and Ronen Plesser (Harvard): constructed mirror pairs (1989–90) and showed all physics agrees; coined "mirror manifolds."
- Doron Gepner (Princeton): exactly solvable string models that made the construction possible.
- Dixon, Jeffrey Harvey, Vafa, Witten: pioneered orbifolding (mid-1980s).
- Philip Candelas, Monika Lynker, Rolf Schimmrigk (Texas): found by computer search that almost all Calabi-Yau shapes in a large sample come in mirror pairs.
- Candelas, Xenia de la Ossa, Linda Parkes, Paul Green: used mirror symmetry to count curves (1991).
- Graham Ross: Greene's Oxford thesis adviser, an early encourager.
- Shing-Tung Yau: first insisted the result must be wrong; later helped prove it.
"Are you seated?"
- Harvard, fall 1987–88. Greene, a new postdoc with an office down the hall from Vafa, was "intrigued but also sceptical" of the conjecture and "half convinced myself that it wasn't true."
- Orbifolding. With Ronen Plesser, he was manufacturing new Calabi-Yau shapes by gluing points of an old one together. Using Gepner's insights to orbifold full Calabi-Yau shapes, they found that the right gluing produced a shape whose odd-dimensional holes equalled the original's even-dimensional holes, and vice versa: same total, so same families, but quite different geometry.
- The linchpin. Months of analysis showed all physical properties agree. Yau "politely but firmly claimed that we must have made an error." Weeks of anxiety: "If it were wrong, everyone would know."
- The phone call. Days after submission, Candelas rang: "Are you seated?" His group had found by computer that Calabi-Yau shapes come in pairs with even and odd holes swapped. Greene replied that he was still seated: he and Plesser had found the same, and shown the physics was identical. The two papers together established mirror symmetry.
Mirror symmetry as a calculating machine
Because a mirror pair gives the same physics, a calculation that is impossible on one shape can be done on its mirror, where it may be easy. "There is no simple explanation for why this happens, but... It most definitely does."
Counting oranges, and "Physics Wins!"
- The analogy. Counting oranges jumbled in a bin 50 by 50 by 10 feet one by one is hopeless. A friend says they came in boxes stacked 20 × 20 × 20 = 8,000; fill one box and multiply. Reorganising the calculation makes it easy.
- The real calculation. Candelas, de la Ossa, Parkes and Green used the mirror to count rational curves ("spheres") inside the quintic Calabi-Yau, "with a few pages of algebra and a desktop computer," and answered questions no mathematician had attempted.
- Berkeley, 1991. Candelas announced 317,206,375 curves of degree three. Norwegian mathematicians Geir Ellingsrud and Stein Arild Strømme announced 2,682,549,425. Days of debate: a "litmus test" of string theory, "the next best thing" to experiment.
- A month later, an e-mail circulated titled "Physics Wins!": the mathematicians had found an error in their computer code; the corrected answer matched Candelas.
- Proof. Building on Kontsevich, Manin, Tian, Li and Givental, Yau, Bong Lian and Kefeng Liu proved the curve-counting formulas rigorously, solving problems "that have puzzled mathematicians for hundreds of years."
Counting curves on the quintic (the mirror-symmetry numbers)
- Degree 1 (lines): 2,875 (classical, 19th century).
- Degree 2 (conics): 609,250 (Sheldon Katz, 1986).
- Degree 3 (twisted cubics): 317,206,375 (Candelas et al., via mirror symmetry; confirmed by Ellingsrud–Strømme after fixing their code).
- Every degree at once: the mirror formula gives an infinite sequence; degree 4 is 242,467,530,000.
- Topology of the pair: the quintic has h1,1 = 1 and h2,1 = 101; its mirror has 101 and 1.
The lesson: physicists long mined mathematics for tools; now "physics is beginning to repay the debt." String theory may unify mathematics as well as physics.
The author is a character here, and the bounce is a conjecture. Two cautions.
- Partly a first-person history. The mirror-symmetry account is Greene's own story. It is accurate and the Greene–Plesser and Candelas–Lynker–Schimmrigk papers (1990) are rightly co-credited, but other early contributors (for example Philip Candelas, Paul Green and Tristan Hübsch on connecting Calabi-Yau spaces, and the Batyrev toric construction of 1994 that made mirror pairs systematic) appear only in passing or not at all.
- The "minimum length" argument is narrower than it sounds. T-duality is exact for circles and tori in perturbative string theory. The claim that no space can shrink below the Planck length, and that the big bang was a Planck-sized bounce, is an extrapolation: D-branes (Ch 12–13) can probe distances shorter than the string length (Douglas, Kabat, Pouliot and Shenker, 1997), and no string calculation yet describes the big bang itself.
Mirror symmetry after 1999: from trick to theorem.
- Homological mirror symmetry (Maxim Kontsevich, 1994; Fields Medal 1998) recast it as an equivalence between two different kinds of mathematics (algebraic and symplectic geometry), now a major field of pure mathematics.
- SYZ (Strominger, Yau and Zaslow, 1996) explained mirror symmetry as T-duality applied to tori inside the Calabi-Yau, directly linking the chapter's two halves; the Gross–Siebert program has since made it constructive.
- Higher genus. Curve counts were extended to curves with handles (BCOV, 1993; genus-one proof by Aleksey Zinger, 2007–09; physics computations to genus 51 by Huang, Klemm and Quackenbush, 2007).
- Physics-to-math traffic has become routine: Witten (Fields 1990), Kontsevich and Borcherds (1998), Okounkov (2006); in 2024 Kontsevich, with Katzarkov, Pantev and Yu, announced a proof that certain four-dimensional varieties are not rational using ideas descended from mirror symmetry.
T-duality and cosmology since 1999.
- String gas cosmology (Brandenberger–Vafa, 1989; revived by Brandenberger, Nayeri and Patil, 2006) predicts a slightly blue tilt in primordial gravitational waves; no primordial gravitational waves have yet been detected (BICEP/Keck: tensor-to-scalar ratio r < 0.036, 2021).
- Double field theory (Hull and Zwiebach, 2009) makes T-duality manifest by doubling the coordinates, one set for momentum and one for winding.
- The cosmological numbers in the chapter are dated. The universe's age is 13.8 billion years (Planck satellite, 2018). Its fate is now governed by dark energy (about 68% of the cosmic budget), not matter density, and space is flat to within about 0.2%. Recent DESI results hint that dark energy may weaken over time (Ch 14).
Equivalent descriptions are a calculating tool. Mirror symmetry's practical power came from noticing that two different-looking representations encode the same thing, and then doing the work in whichever one is easy. The same move pays off in data work: a problem intractable as a join may be trivial as a graph traversal; a risk aggregation hopeless in loan space may be simple in factor space. When a calculation stalls, look for its mirror.
Tearing the Fabric of Space
General relativity forbids space to tear: Riemann's geometry needs a smooth fabric, with no creases, punctures or seams, and a tear would break Einstein's equations. Physicists have long wondered whether quantum physics relaxes this, opening the door to wormholes or smoothing out black-hole singularities. String theory gave the first definite answer. In 1987 Gang Tian and Shing-Tung Yau noticed a mathematical operation, the flop, that pinches a sphere inside a Calabi-Yau shape to a point, tears it, and regrows it into a topologically different shape. In 1992–93 Aspinwall, Greene and Morrison (via mirror symmetry) and Witten (via strings that encircle the tear) showed that string theory passes through such tears without any catastrophe: the fabric of space can rip, and physics does not notice.
Can space tear?
Stretch a rubber membrane far enough and it rips. Einstein's general relativity says space cannot: it measures distance relations between nearby points, which requires a smooth substrate. A tear would make its equations break down, "signalling some or other variety of cosmic catastrophe." Quantum foam (Ch 5) led some to suspect that rips might be commonplace at tiny scales.
Two places where space is stretched to its limit
- Wormholes. A company has headquarters on the 90th floor of one World Trade Center tower (the book was written before 2001) and a growing branch on the 90th floor of the other. Rather than ride down and up 90 floors, build a bridge. A wormhole is a shortcut between distant regions, with one key difference: the sky-bridge crosses existing space, while a wormhole creates new space. On a U-shaped two-dimensional universe, two punctures that grow tentacles and merge make the tunnel. Whether wormholes exist, microscopic or Deep Space Nine-sized, depends on whether space can tear.
- Black holes. Unlike wormholes, these have strong observational evidence, so what happens at the centre is "science, not speculation." General relativity's curvature becomes infinite there, a puncture. Roger Penrose proposed cosmic censorship: such singularities occur only hidden behind event horizons. Others hoped quantum gravity would sew up the puncture.
Tian and Yau's flop
The flop transition, step by step
- Find a sphere. Inside a Calabi-Yau shape sits a two-dimensional sphere, like the surface of a beach ball (two numbers, latitude and longitude, locate a point on it).
- Shrink it until it is pinched to a single point.
- Tear the Calabi-Yau slightly at the pinch.
- Glue in a new beach-ball sphere, oriented differently ("flopped").
- Re-inflate it to a plump shape.
- The result can be topologically distinct from the start: there is no way to deform one into the other without tearing. For mathematicians, a recipe for new Calabi-Yau shapes; for physicists, a question: can nature do this?
The mirror perspective
For a couple of years Yau nudged Greene to think about flops; Greene didn't, seeing them as "merely a piece of abstract mathematics." Chapter 10's minimum radius might seem to forbid pinching, but that argument applies to whole dimensions, not to a chunk of space such as a sphere inside a Calabi-Yau.
1987: Tian and Yau describe the flop mathematically.
1991: Paul Aspinwall (Greene's Oxford classmate) and Norwegian physicist Andy Lütken ask what a flop looks like from the mirror side, where the mathematics may be simpler; they see no sign of disaster. Greene and Plesser notice that their orbifold gluing produces exactly such pinches, and that the pinch can be repaired two ways, giving either the original or the flopped shape.
Spring 1992: Plesser describes the idea to Witten at the Institute for Advanced Study. Witten stares out of his window for a minute or two, then says that if it works "it would be spectacular."
1992: Greene, David Morrison (Duke) and Sheldon Katz (Oklahoma State) outline a proof, but the computation would take more than a century on the fastest computer.
Spring–summer 1992: Victor Batyrev (Essen), building on Shi-Shyr Roan, recasts mirror construction in systematic toric mathematics, offering the needed efficiency.
Fall 1992: Greene (on leave from Cornell), Morrison and Aspinwall gather at the Institute for Advanced Study to attack the problem.
December 1992: the decisive Saturday computation.
January 1993: Aspinwall–Greene–Morrison and Witten post their papers to the arXiv simultaneously.
| Physicists | Mathematicians |
|---|---|
| Like avant-garde composers: bend traditional rules and "brush the edge of acceptability" in search of answers | Like classical composers: work in a tighter framework, reluctant to move on until each step is rigorous |
| Batyrev called Greene and Plesser's mirror construction "black magic" | Greene could "extract the gist" of Batyrev's papers but struggled with key details |
| Neither is right or wrong: "a matter of taste and training" | Their meeting point, as string theory blurs the border, is where the breakthroughs came |
The strategy
A direct attack, following the Calabi-Yau through the tear, meant forbidding equations. Instead:
Use the mirror as a detour. In the mirror description of a flop, nothing pinches or tears; the physics looks smooth. The open question was whether the mirror relation survives the tear. If the torn-and-repaired shape still has the predicted mirror, then no catastrophe on the mirror side means no catastrophe on the original side, and space can tear. That reduced to a calculation: compute a particle mass on the post-flop shape (easy), find the exact candidate mirror (hard) and compute the same mass there (very computer-intensive), then compare.
Division of labour. Aspinwall, "a crackerjack programmer," coded Candelas's method for extracting physics from the mirror. Greene and Morrison set out to identify the mirror's precise shape, which required Batyrev's mathematics. The cultural divide slowed them, so they ran a class of one: by day they pushed the calculation; each evening Greene lectured Morrison on the physics for an hour or two, then Morrison lectured Greene on the mathematics. "School would typically let out at about 11 P.M." Meanwhile Witten, after pointing out a possible weak link in Greene's earlier work with Vafa and Warner, was developing a powerful new translation between string physics and Calabi-Yau mathematics, and it was converging on flops. "If we didn't complete our calculation soon, Witten would beat us to the punch."
A six-pack and 8.999999
- The setting. The Institute for Advanced Study, founded 1930, Einstein's home from 1933 until his death: "you can't get distracted... Because, well, there aren't any distractions." Greene (vegetarian) and Aspinwall ("as devout a meat eater") argue over dinner on Nassau Street; Aspinwall plays devil's advocate, then comes round within days.
- Aspinwall's rules. "The most civilized physicist I know": he never works past 5 P.M. Or on weekends.
- Thursday–Friday, early December. Greene and Morrison finally identify the mirror shape and code it by late Friday night, but Aspinwall's program is needed and he has gone home. After "much groaning" he agrees to come in Saturday, for a six-pack of beer.
- The run. The easy side gave a mass of 3; the computer should return something like 2.999999. Morrison presses enter: 8.999999. Greene's heart sinks, but a near-integer is no accident: at the blackboard they find they had dropped a factor of 3 weeks earlier. The true answer was 9.
- The real test. Knowing the answer makes it "all too easy" to get it, so they compute a second mass: 12 by hand. The computer: 11.999999. The mirror is the mirror; space-tearing flops are part of string theory.
- Reactions. Greene runs "an unrestrained victory lap around the office"; Morrison beams; Aspinwall: "That's great, but I knew it would work. And where's my beer?"
Witten's approach: strings that shield the tear
On Monday they told Witten, who had just established the same result by a quite different route, one that explains why tears are harmless.
Point particles can only pass by a tear; strings can encircle it. A string moving forward can wrap around the tear, its world-sheet forming a protective tube whose effects precisely cancel the catastrophe the geometric degeneration would otherwise cause.
- "What if no string is nearby?" Feynman's sum over paths (Ch 4): every string effectively "sniffs out" all possible trajectories, including infinitely many that encircle the tear. Their combined contributions do the shielding whether or not a string "is there."
- "A loop can't shield anything; it's like hiding from a cluster bomb behind a hula hoop." Same answer: the shield is the quantum sum, not one classical loop.
Consequences
The family count and particle types are unchanged by a flop; only particle masses shift, smoothly, and the moment of tearing has no observable signature at all. In six dimensions the tear is "more like the handiwork of a moth on wool than that of a deep knee bend on shrunken trousers."
- Can it happen in the large dimensions? "Almost certainly, yes." Space is space; the curled/extended distinction may itself be artificial if our big dimensions close up on a cosmic scale. The proofs used special Calabi-Yau mathematics, but the lesson is general.
- Could it be happening now? Particle masses are very stable today, so any current tear must be exceedingly slow. In the early universe, when masses did change, tears could have occurred. A slow enough rupture under way right now would go unnoticed: a rare case where "the lack of a striking observable phenomenon is cause for great excitement."
- Name: physicists call these topology-changing transitions.
How stable are particle properties? (modern limits)
- Fine-structure constant: atomic-clock comparisons (2020–21) limit any present drift to roughly 10−18 per year.
- Proton-to-electron mass ratio: molecular spectra of distant quasars show it has changed by less than about one part in 10 million over 7 billion years.
- The Oklo natural reactor in Gabon (operated 2 billion years ago) constrains nuclear physics in the deep past.
- Implication: any present-day topology change in our neighbourhood would have to be extraordinarily slow, as Greene says.
What the flop result does and doesn't show. The 1993 papers prove that one specific kind of tear (a flop, in the extra dimensions, in theories with enough supersymmetry, at a mild degeneration) is physically smooth. They say nothing direct about wormholes or black-hole singularities, the two motivating examples, which involve violent, Lorentzian, non-supersymmetric geometry; Greene's "almost certainly" for the large dimensions is a hope, not a result. The severer conifold transition, where a three-dimensional sphere collapses, needed new physics (massless black holes) to make sense (Ch 13, Strominger 1995). Note also that the chapter is again a first-person account of Greene's own work, told vividly but from one side.
Topology change, wormholes and censorship after 1999.
- Witten's method became a standard tool. His 1993 "gauged linear sigma model," the "new translation" in this chapter, is now one of the main ways to study Calabi-Yau phases, mirror symmetry and string compactifications.
- Wormholes came back through quantum information. Maldacena and Susskind's ER = EPR (2013) proposes that entangled particles are connected by microscopic wormholes. Gao, Jafferis and Wall (2016) found traversable wormholes in quantum gravity, requiring negative energy from quantum effects and too slow to beat light. A 2022 Google quantum-processor "wormhole" simulation (Jafferis et al., Nature) drew wide coverage and pointed criticism that it was a small toy model, not a real wormhole.
- Cosmic censorship is in trouble in some settings. Penrose shared the 2020 Nobel Prize for showing that black holes form generically. But simulations found naked singularities forming in five-dimensional black strings (Lehner and Pretorius, 2010) and counterexamples in four-dimensional anti-de Sitter space (Horowitz, Santos and Way, 2016; Crisford and Santos, 2017). Whether censorship holds in our universe remains unproved.
- Black-hole interiors remain the frontier: string theory has not yet described what replaces the singularity of a real (non-supersymmetric) black hole, though "fuzzball" proposals (Mathur, from 2002) and holographic methods (Ch 13, 16) attack it.
Prove safety through an equivalent system. The team could not analyse the tear directly, so they proved the mirror stayed well-behaved and that the mirror relation survived the change. Migration engineers use the same logic: when you can't test a risky cutover in place, build a provably equivalent shadow system, run it through the change, and check that invariants match at known values. And note the 8.999999 moment: a near-integer wrong answer is a clue, not a failure; check your own "easy" calculation first.
Beyond Strings
Einstein wondered "whether the necessity of logical simplicity leaves any freedom at all." The ideal final theory would be inevitable: things are as they are because any variation, like the sentence "This sentence is a lie," destroys itself. By the late 1980s string theory fell short twice: there were five string theories, and each had many solutions. Both flaws, it turned out, may be artefacts of perturbation theory, the approximation behind almost everything known about strings. In March 1995 Edward Witten used dualities, different descriptions of the same physics, to show that the five string theories and eleven-dimensional supergravity are all limits of a single theory, provisionally called M-theory, which lives in eleven dimensions and contains not only strings but membranes and higher-dimensional "branes."
The summary first
Greene offers a lifeline for readers who get lost: the second superstring revolution in two pictures.
- Before 1995: five apparently separate string theories: Type I, Type IIA, Type IIB, Heterotic O(32) ("Heterotic-O") and Heterotic E8×E8 ("Heterotic-E"). All need ten dimensions and Calabi-Yau shapes; they differ in how they incorporate supersymmetry and in their vibration patterns (Type I alone has open strings with loose ends).
- After 1995: like the five arms of a starfish, all are parts of one connected framework, which also absorbs a sixth arm, eleven-dimensional supergravity. Its two known essentials: eleven dimensions (ten space, one time; the extra dimension was missed by approximate calculations), and extended objects of many dimensions: membranes, "three-branes" and more.
- Much remains mysterious, "one suggested meaning for the 'M'"; understanding M-theory "may well constitute the central problem of twenty-first-century physics."
Perturbation theory: the mechanic's estimate
Three pictures of approximation
- The car repair. The mechanic says a new engine block will run about $900. Later: $950, with a regulator. The final bill: $987.93, adding a $27 fan belt, a $10 battery cable and a 93¢ bolt. Each detail is a small perturbation. A final bill shockingly different from the estimate is a failure of perturbation theory.
- The earth's orbit. The sun's pull dominates, so a sun-only calculation gives a ballpark; the moon and planets add ever-smaller refinements. (Even three bodies cannot be solved in neat closed form.)
- A triple star. Three comparable stars have no dominant influence; dropping the third star changes everything, and the errors cascade like dominoes. "Three people dancing the hora bears little resemblance to two people dancing the tango." Only brute-force computer simulation works.
In string theory, two strings meet and merge into one, which splits again: a zero-loop process. Quantum jitters let the middle loop briefly split into a virtual string pair and rejoin, giving a one-loop process; then two loops, three, and so on, each with its own formula. The exact answer is the sum of infinitely many diagrams, each harder than the last. Physicists compute a few and hope the rest are small.
The string coupling constant decides whether the ballpark is in the ballpark. Like the strength of a rope, a number called the string coupling constant sets how likely a string is to split into a virtual pair. The dividing line is 1:
- Below 1: more loops are ever less likely, like repeated lightning strikes; a few diagrams suffice and perturbation theory works.
- 1 or above: more loops become more important; the zero-loop "estimate" is meaningless, as in the triple star.
- Its value is unknown. Each of the five theories has its own coupling constant, and it also fixes masses and charges. "One of the most important unresolved issues in string theory."
The equations we can't solve
String theory's equations are determined by how strings interact, so they too are known only approximately. Two disappointments:
- The coupling equation. In all five theories, the approximate equation says the coupling constant is a number which, multiplied by zero, gives zero: any value works, so it gives no information.
- The geometry equation. The equation for spacetime's shape is stricter but still allows any Calabi-Yau with four large dimensions, and even other splits between large and curled dimensions.
Three readings of the flexibility
- Pessimistic: even the exact equations may have a vast spectrum of solutions, weakening the theory's predictive power (Ch 15). (This is where the field has since gone; see the update below.)
- A flaw in the method: using perturbation theory to compute the coupling assumes it is below 1; the failure may mean the coupling is larger.
- Approximations, not physics: even with a small coupling, the sum of all loops may turn loose equations into restrictive ones.
- By the early 1990s options 2 and 3 convinced most string theorists that progress required a non-perturbative approach. "As of 1994, finding such a means seemed like a pipe dream."
Strings '95 and the idea of duality
String conferences tracked the field's mood: euphoria in the mid-1980s, disillusionment in the late 1980s (some proposed cancelling the annual meeting), recovery in the early 1990s. Then, in March 1995 at the University of Southern California, Witten gave the lecture that ignited the second superstring revolution, building on Duff, Hull and Townsend and on Schwarz and Ashoke Sen.
Duality: two descriptions, one physics.
- Trivial duality: general relativity in English vs. Chinese; translation gains nothing.
- Non-trivial duality: different descriptions give different, complementary insights and methods. We have met two: R ↔ 1/R (Ch 10) and mirror symmetry (Ch 10–11).
- Witten's claim: the five string theories are five windows onto one theory, with strong coupling in one equal to weak coupling in another. The idea had seemed so outlandish physicists "would rarely if ever speak" of it.
Ice-lover meets water-lover Two sheltered people, one who knows only ice, the other only liquid water, go camping in the desert, each fascinated by the other's strange substance. In the heat, the ice melts into water; in the frigid night, the water freezes into ice. They realise the substances are one thing, H2O. In string theory the coupling constant plays the role of temperature: turn it up, and one string theory transmutes into another.
The power of symmetry: BPS states
Symmetry lets you learn things indirectly: we know the laws of physics in Andromeda without going there; a police artist can sketch a whole face from a glimpse of its right side. Led by Nathan Seiberg (Institute for Advanced Study), physicists in the 1990s turned supersymmetry into a precise tool for answering hard questions without hard calculations.
The envelope puzzle: "syzygy"
- Clue 1. An envelope holds a string of letters with exactly three y's. Endless possibilities (mvcfojziyxidqfqzyycdi...).
- Clues 2 and 3. It is an English word, and the shortest one with three y's. Now there is one answer: syzygy.
- The physics version. A box contains something with three units of electric charge: three protons? Four positrons and an electron? Nine up-quarks plus photons? Add two clues: the world is supersymmetric (the "language"), and the contents have the minimum mass for that charge (the "shortest word"). Then, following E. Bogomol'nyi, Manoj Prasad and Charles Sommerfield, the contents are uniquely fixed.
- BPS states. These minimum-mass objects have properties fixed exactly, at any coupling, without perturbation theory. Greene's mnemonic: "beyond perturbative states." They are only a small part of strong-coupling physics, but "like a few choice words in a foreign tongue... They will take us quite far."
The dualities, one by one
How the five theories connect
- Type I ↔ Heterotic-O (strong–weak duality). Witten, later with Joe Polchinski (UC Santa Barbara): Type I's BPS masses and charges at strong coupling match Heterotic-O's at weak coupling, and vice versa. Not proven (one side is always incalculable), but every test has passed.
- Type IIB ↔ itself (self-duality). Conjectured by Hull and Townsend: replacing a coupling g > 1 with 1/g gives the identical theory, the coupling analogue of R ↔ 1/R.
- Type IIA → eleven dimensions. At strong coupling, Type IIA's low-energy physics is eleven-dimensional supergravity; a new dimension opens, whose size is set by the coupling.
- Heterotic-E → eleven dimensions. Petr Hořava and Witten: at strong coupling a new dimension appears, a line segment with walls at each end.
- IIA ↔ IIB and Heterotic-O ↔ Heterotic-E (T-duality). The precise version of Chapter 10's R ↔ 1/R swaps theories: IIA on a circle of radius R equals IIB on radius 1/R; likewise the two heterotic strings. These links complete the web.
| Theory | Strings | Supersymmetry | Low-energy limit | Strong-coupling description |
|---|---|---|---|---|
| Type I | Open and closed | N = 1, gauge group SO(32) | Type I supergravity | Heterotic-O, weakly coupled |
| Type IIA | Closed | N = 2, non-chiral | IIA supergravity | M-theory (11D) on a circle |
| Type IIB | Closed | N = 2, chiral | IIB supergravity | Itself (self-dual) |
| Heterotic-O | Closed | N = 1, SO(32) | Type I/heterotic supergravity | Type I, weakly coupled |
| Heterotic-E | Closed | N = 1, E8×E8 | Heterotic supergravity | M-theory on a line segment |
| 11D supergravity | None (point particles) | Maximal | Itself | Low-energy limit of M-theory |
The supergravity story and the eleventh dimension
1976: Daniel Freedman, Sergio Ferrara and Peter van Nieuwenhuizen (Stony Brook) invent supergravity, hoping boson–fermion cancellations will tame gravity's quantum frenzy.
1978: Eugène Cremmer, Bernard Julia and Joël Scherk (École Normale Supérieure): the most promising versions live in ten or eleven dimensions; eleven is the maximum possible.
1984 on: ten-dimensional supergravity is reinterpreted as the low-energy limit of string theory. There are four such theories: three for IIA, IIB and Heterotic-E; the fourth serves both Type I and Heterotic-O, an early hint of their duality. Eleven-dimensional supergravity is "left out in the cold," seen by most "as a mathematical oddity."
1995: Witten shows strongly coupled Type IIA has eleven-dimensional supergravity as its low-energy limit. The audience is stunned.
Strings are membranes in disguise. As the Heterotic-E coupling grows, a tenth spatial dimension appears and the string stretches into a ribbon and then a deformed cylinder; it was always a two-dimensional membrane whose width is set by the coupling. For Type IIA the string inflates into an "inner tube" wrapping the new circle. Small coupling hid this because perturbation theory implicitly assumed both a thin membrane and a tiny eleventh dimension. The old count of ten dimensions is unaffected: the new dimension is inside the string, not a direction it vibrates in. A ten-dimensional universe of strings is an approximation to an eleven-dimensional universe of membranes.
What does the "M" stand for? "As many things as people you poll": Mystery, Mother (of all theories), Membrane, or Matrix (after Tom Banks, Willy Fischler, Stephen Shenker and Leonard Susskind's 1996 proposal). Whatever it is, it is "the unifying pachyderm": like the three blind men who grab a tusk, a leg and a tail and think they hold three animals, string theorists had been describing one elephant.
Democracy of branes
Following the membranes from Heterotic-E and Type IIA around the web shows that all five theories contain two-dimensional membranes. Work by Horowitz and Strominger and then Polchinski revealed what BPS states look like: strings, membranes, and objects of three, four... Up to nine spatial dimensions: p-branes ("far from euphonious"). Paul Townsend declared a "democracy of branes."
- Why strings still rule our world: in each of the five weakly coupled regions, every brane except the string has mass inversely proportional to the coupling, so at weak coupling they are far heavier than the Planck mass and nearly invisible. Only strings (or membranes curled to look like strings) are light enough to make the particles we know.
- Toward the centre of the map, branes get lighter and the theory becomes a genuine democracy. "The perturbative analyses... Were not refined enough to discover even the existence" of the heavy branes; hence the "far-from-democratic name" string theory.
Does this solve the open questions?
Yes and no. Dualities reveal the theory's logical structure; strong coupling, once "uncharted territory, potentially filled with dragons and sea monsters," turns out to lead back to weak coupling in another guise. But the coupling constant's value, the three large dimensions and the choice of curled-up shape all remain beyond present non-perturbative tools.
The map is complete, even if unexplored. Greene's cartographer: once every region of the globe is charted, "the age of geographic discovery is over," though Antarctica still deserves study. Starting from any of the five theories, every reachable theory lies on one map, so "with guarded optimism" the space of consistent theories that combine relativity, quantum mechanics, gauge forces, supersymmetry and extra dimensions is fully mapped. The remaining challenge: find the exact equations that pick our point on the map. Witten: understanding M-theory "would transform our understanding of nature at least as radically as occurred in any of the major scientific upheavals of the past."
Inevitability, unachieved. The chapter's frame, a theory that "couldn't be otherwise," sits uneasily with what followed. M-theory unified the five formulations, but it did not reduce the number of solutions. Greene's own "pessimistic" option became mainstream: flux compactifications (2000–03) yield perhaps 10500 or more stable vacua, the landscape, and many physicists now invoke the anthropic principle, which Greene treats warily (Ch 14–15). The "no blank spots" claim also has fine print: it covers supersymmetric theories reachable by duality, while non-supersymmetric ten-dimensional strings (such as the tachyon-free O(16)×O(16) heterotic string, 1986) sit off the map and remain under study.
Small factual cautions. "Even three bodies... No one has been able to solve them in full" is too strong: Karl Sundman found a convergent series solution of the three-body problem in 1912, though it converges too slowly to be useful, so Greene's practical point stands. And the strong-weak dualities, though passing every check, remain unproven conjectures, as Greene himself notes.
M-theory in 2026: still a name in search of a definition.
- No fundamental formulation. Thirty years on, M-theory has no complete equations. Partial definitions exist: BFSS matrix theory (1996) describes it in certain backgrounds; the ABJM theory (Aharony, Bergman, Jafferis and Maldacena, 2008) describes stacks of M2-branes; the six-dimensional (2,0) theory of M5-branes still has no conventional Lagrangian.
- D-branes changed everything. Joseph Polchinski's 1995 identification of branes as D-branes (surfaces where open strings end) became the key tool behind black-hole entropy (Ch 13), brane-world models and holography. Polchinski died in 2018.
- Duality's biggest child: holography. Two years after Strings '95, Juan Maldacena's AdS/CFT correspondence (1997) found a duality between string theory in a curved space and an ordinary quantum field theory without gravity on its boundary. It is the most-cited paper in high-energy physics and now underpins work on black holes, quantum information and even condensed matter (Ch 13, 16).
- Perturbation theory, revisited. Superstring amplitudes are now understood with new mathematical rigour (Witten's 2012 "Notes on supermanifolds"; D'Hoker and Phong's higher-loop results), and the amplitudes programme found hidden structures (the "amplituhedron," 2013) in related gauge theories.
- Strings conferences continue annually; Strings 2025 was held in Abu Dhabi.
When five vendors claim to be the platform, look for the shared core. The five string theories were five descriptions of one thing, visible only when you pushed each past its comfort zone (strong coupling). Platform consolidation in enterprises follows the same pattern: five "different" systems often turn out to be projections of one underlying data model, and stress-testing them at their limits (peak load, edge cases) reveals the mapping. Name the invariants (BPS states) that every description must agree on, and use them to prove equivalence before you merge.
Black Holes
Black holes and elementary particles look like opposites, yet both are described by just mass, charges and spin. String theory makes the link exact. In 1995 Andrew Strominger showed that when a three-dimensional sphere inside a Calabi-Yau collapses, a three-brane wrapped around it shields the universe from catastrophe, and that wrapped brane looks to us like a black hole whose mass shrinks to zero. Greene, Morrison and Strominger then found the conifold transition: space tears more drastically, the Calabi-Yau changes topology, and the massless black hole becomes an ordinary massless particle, a phase change like ice melting. In 1996 Strominger and Vafa built black holes from branes and counted their microscopic states. The count matched the Bekenstein–Hawking entropy exactly, the first microscopic explanation of black-hole entropy and string theory's most celebrated result. The information paradox and the singularity remained open.
Black holes have no hair
Late-1960s and early-1970s work by Demetrios Christodoulou, Werner Israel, Richard Price, Brandon Carter, Roy Kerr, David Robinson, Hawking and Penrose supported John Wheeler's slogan: "black holes have no hair." Apart from mass, force charges and spin, all black holes are identical: no "fancy hairdos." Those are exactly the traits that distinguish elementary particles, prompting speculation that black holes might be giant elementary particles.
- No minimum mass. In general relativity, crush any mass small enough and it becomes a black hole. Ever-lighter black holes should look ever more like particles.
- The catch. Below about the Planck mass (huge for a particle, ten billion billion protons; tiny for a black hole, a grain of dust), quantum mechanics matters, and the old incompatibility stymied all progress. String theory provides "the first theoretically sound connection."
The three-sphere problem and Strominger's three-brane
A Calabi-Yau contains two kinds of spheres: two-dimensional ones (the flop spheres of Ch 11) and three-dimensional ones, the "surfaces" of four-dimensional beach balls. The equations suggest three-spheres can collapse, and pre-1995 approximations warned that this would unleash infinities.
Match the brane to the sphere. A string's world-sheet can lasso a two-sphere (Ch 11), but a string cannot surround a three-sphere, just as, one dimension down, a point cannot lasso a beach ball. Strominger (building on Witten and Seiberg) used the second revolution's new objects:
- A string (one-brane) can wrap a circle.
- A membrane (two-brane) can wrap a two-sphere, like plastic wrap around an orange.
- A three-brane can wrap a three-sphere. A wrapped three-brane, he showed with "a simple and standard physics calculation," provides a tailor-made shield that exactly cancels the feared catastrophe.
The second half of the story: conifold transitions
Cornell, spring 1995: one week, one paper
- The morning after. Greene read Strominger's paper "retrieved from the World Wide Web" and suspected it told only half the story: the collapse, but not a possible tear and repair.
- That afternoon. With David Morrison, visiting Cornell, he sketched the second half within hours, drawing on late-1980s mathematics of Herb Clemens (Utah), Robert Friedman (Columbia) and Miles Reid (Warwick), as applied by Candelas, Green and Tristan Hübsch.
- Next morning. Strominger e-mailed: "it should tie in somehow with your work with Aspinwall and Morrison." A flurry of e-mails followed; within a week (by the following Wednesday) a joint draft existed.
- 11:45 P.M. Greene and Morrison submitted the paper, then, driving off campus, played devil's advocate and decided the wording overclaimed in places. The arXiv allowed changes until 2 A.M. They turned the car around, toned down a few paragraphs, resubmitted within the hour, and agreed not to talk about it on the drive home.
- The response. Enthusiastic. Ronen Plesser wrote: "I wish that I had thought of that!"
Spheres of every dimension, and the doughnut-to-ball analogy
- Two-sphere: points in 3D space at a fixed distance from a centre (a beach ball's surface).
- One-sphere: points in a plane at a fixed distance: a circle.
- Zero-sphere: points on a line at a fixed distance: just two points.
- The analogy. On a doughnut, let a circle (one-sphere) pinch, tear the surface, and plug the two holes with two points (a zero-sphere). The result is a warped banana that deforms into a beach ball: the doughnut's hole is gone, topology has changed drastically.
- The real thing. A three-sphere collapses and a two-sphere grows in its place: the conifold transition, a far more drastic change than a flop. (The analogy misleads on one point: the real transition adds a massless string pattern rather than removing a hole's worth.)
Massless black holes become particles
Two facts combine.
- Horowitz and Strominger: a three-brane wrapped on a three-sphere looks, to beings in three large dimensions, like a black hole, with mass proportional to the sphere's volume. As the sphere collapses the black hole gets lighter; at the pinch it is massless.
- Conifold transitions change the holes in a Calabi-Yau, so they change the number of massless string patterns: Greene, Morrison and Strominger found it increases by exactly one. Conclusion: the black hole loses mass until it is massless and then turns into a massless particle, such as a photon, which in string theory is just a string vibrating in a particular pattern. "For the first time, string theory explicitly establishes a direct, concrete, and quantitatively unassailable connection between black holes and elementary particles."
"Melting" black holes. The process is a phase transition. As temperature decides whether H2O is ice or water, the topology of the Calabi-Yau decides whether certain configurations appear as black holes (one phase) or elementary particles (the other). The conifold tear carries you between phases.
A bigger payoff. Before this, no one knew how to deform one Calabi-Yau continuously into another, so the unified "theory map" of Chapter 12 might have fragmented once dimensions curled up. Conifold transitions appear to connect Calabi-Yau shapes to one another, so "even after curling up all extra dimensions, the unity... Firmly holds."
Black-hole entropy: the puzzle
Entropy counts the rearrangements of a system's microscopic constituents that leave its overall appearance unchanged. A tidy desk has few such rearrangements (low entropy); a messy desk has many (high entropy). Precisely, it counts quantum microstates compatible with the same energy, pressure and so on.
Bekenstein's vacuum pump (1970–72)
- The second law. Total entropy always rises. Even tidying your desk raises it overall: you burn body fat and warm the air.
- The challenge. Jacob Bekenstein, a graduate student of Wheeler's at Princeton, asked: what if a vacuum pump sucks the warmed air, the clutter and the desk itself into a nearby black hole, leaving a cold, ordered room? Entropy outside falls.
- The proposal. To save the second law, black holes must have entropy, which grows as matter falls in.
- Hawking's clue. Hawking had proved that a black hole's horizon area never decreases (in asteroid impacts, accretion, mergers). Bekenstein proposed that horizon area measures entropy.
- The objections. Black holes seem the most ordered objects imaginable (no hair, so nowhere to put disorder), and entropy is quantum while black holes were classical.
How black is black?
Hawking, with James Bardeen and Brandon Carter, had noticed the analogy but dismissed it: if area were entropy, black holes would need a temperature, so they would glow, and black holes are black. Then, in 1974, Hawking discovered that black holes are not black.
Hawking radiation. Near the horizon, the vacuum's virtual particle pairs keep erupting. The black hole's gravity can pull a pair apart so one falls in; the other, with no partner to annihilate, is boosted outward. Repeated all around the horizon, this looks from afar like a steady glow with temperature set by the gravity at the horizon, exactly as the thermodynamic analogy required. "Black holes glow." Bekenstein was right; the black-hole laws are the laws of thermodynamics.
| Black hole | Mass | Hawking temperature | Entropy (in Boltzmann units) |
|---|---|---|---|
| Stellar (Greene's example) | 3 suns | 2 × 10−8 K | ~1078 |
| Sagittarius A* (Milky Way centre) | 4.3 million suns | 1.4 × 10−14 K | ~1090 |
| M87* (first image, 2019) | 6.5 billion suns | 10−17 K | ~1096 |
| Moon-mass | 7 × 1022 kg | 1.7 K, below the 2.7 K cosmic background, so it absorbs more than it emits | ~1062 |
| Mountain-mass primordial | ~5 × 1011 kg | ~1011 K | ~1040; would be finishing evaporating today |
Greene's figures: a three-solar-mass black hole is a hundred-millionth of a degree above absolute zero, its radiation hopelessly faint; an asteroid-mass black hole would glow like a million-megaton hydrogen bomb in gamma rays. Searches found nothing, so such small black holes are rare if they exist. Hawking joked that detection would win him a Nobel Prize. Meanwhile the entropy is enormous, "a one followed by about 78 zeros" for three suns. But entropy of what? Hawking's partial merger gave the number, not the microscopic meaning. For nearly a quarter-century the question stayed open.
String theory counts the microstates
Strominger and Vafa, January 1996, building on Susskind and Sen: "Microscopic Origin of the Bekenstein-Hawking Entropy."
- Extremal black holes: charged black holes with the minimum mass allowed for their charge, close cousins of BPS states (Ch 12).
- Designer black holes: rather than collapsing a star, they assembled black holes from a precise combination of BPS branes, as one assembles an atom from quarks and electrons.
- Count, then compare. With full control of the microscopic construction they counted the rearrangements that keep mass and charges unchanged and compared the count with the horizon area.
- Perfect agreement, for this class of black holes. "A quarter-century-old puzzle had been solved."
"When string theorists talk about black holes they are almost talking about observable phenomena, and that is impressive." Sheldon Glashow, once string theory's "archrival"
The remaining mysteries
Mystery 1: does information survive?
- Laplace's demon. Pierre-Simon de Laplace: an intelligence knowing all forces and positions "would encompass in the same formula the movements of the greatest bodies of the universe and those of the lightest atoms... The future, like the past, would be open to its eyes."
- Quantum determinism. Uncertainty kills Laplace's version, but wave functions evolve deterministically (Schrödinger, Dirac, Klein-Gordon equations): likelihoods, not outcomes, are fixed.
- Hawking, 1976: black holes break even this. Matter falling in takes its wave function along. Once black holes radiate and evaporate, the hidden region re-enters the universe: does the information come back out?
- Hawking: no; black holes introduce "a new level of uncertainty into physics." "Spacetime ties itself in knots and... Information gets lost in the folds."
- The bet. Hawking and Kip Thorne (lost) versus John Preskill (returned); stake: "an encyclopedia of the winner's choice."
- String theory's hint. In Strominger–Vafa black holes, information can be stored in and recovered from the branes. Strominger: claiming victory is "premature." Vafa: "agnostic."
Mystery 2: the singularity. Since Schwarzschild (1916), general relativity has predicted infinite curvature at a black hole's centre, where time ends, or perhaps where a new universe begins (Ch 14). String theory tames some singularities (flops, conifolds), "but if you have seen one singularity you have not seen them all." The black-hole centre still requires better non-perturbative tools.
What the entropy result does and doesn't cover. Strominger and Vafa's count works for extremal, supersymmetric black holes in five dimensions (extended to four by Maldacena, Strominger and Witten, 1997), and later for near-extremal ones. Real astrophysical black holes are neutral, spinning and far from extremal, and their entropy has not been derived by counting string states. The Kerr/CFT proposal (2008) reproduces the entropy of extremal spinning black holes from symmetry, but not from an explicit microstate count. Calling the result "solved" is accurate for the class studied; extending it to Schwarzschild black holes is still open. Also note the claim that conifold transitions connect all Calabi-Yau shapes is Reid's conjecture ("Reid's fantasy"): supported for very large families of examples, not proven in general.
Black holes, seen and heard (2015–2026).
- LIGO detected gravitational waves from merging black holes in September 2015 (Nobel 2017); by 2025 the LIGO–Virgo–KAGRA network had logged about 300 merger candidates.
- Area theorem confirmed. The exceptionally loud merger GW250114 (January 2025) tested Hawking's area theorem, the clue Bekenstein used, and confirmed that the final horizon area exceeded the initial total at high confidence.
- Event Horizon Telescope images: M87* (2019) and Sagittarius A* (2022).
- Hawking radiation itself is unobserved from real black holes. Laboratory analogues in Bose–Einstein condensates (Jeff Steinhauer, 2016 and 2019) show thermal, entangled "Hawking" emission from sonic horizons. Primordial black holes have not been found; asteroid-mass ones remain a possible dark-matter candidate.
- Penrose won the 2020 Nobel Prize for black-hole formation; Hawking died in 2018 without the prize he joked about.
The information paradox, 1999–2026.
- 2004: Hawking conceded the bet to Preskill (paying with an encyclopedia of baseball); Thorne did not concede.
- AdS/CFT (1997) made information loss hard to believe: a black hole in anti-de Sitter space is dual to an ordinary quantum system, which cannot lose information.
- 2012: the firewall paradox (Almheiri, Marolf, Polchinski and Sully) argued that preserving information might require a wall of energy at the horizon, contradicting general relativity's smooth horizon.
- 2013: ER = EPR (Maldacena and Susskind) suggested that entangled radiation and black hole are linked by wormholes.
- 2019–20: Penington and Almheiri, Engelhardt, Marolf and Maxfield computed the Page curve (entropy of the radiation rising then falling to zero) using semiclassical gravity with "islands" and replica wormholes: strong evidence that information escapes. How it is encoded, and what happens in our universe (not anti-de Sitter space), remain active questions.
- The singularity remains unresolved; fuzzball proposals (Samir Mathur, from 2002) replace the interior with horizon-sized string configurations.
Count the states to understand the summary statistic. Bekenstein and Hawking had a number (entropy = area/4) with no account of what it counts; Strominger and Vafa won by building a system they fully controlled and enumerating its microstates. In analytics, a headline metric (churn, loss rate, risk score) is only understood when you can say which underlying configurations produce it. Build a controlled "designer" case where you can count, and check the metric there first.
Reflections on Cosmology
The standard big-bang model, refined by inflation, describes cosmic history from a hundredth of a second onward with astonishing success, and plausibly back to about 10−43 seconds. Before that, general relativity predicts infinite temperature and density, a sign the theory has broken down. String theory changes the story in three ways: a smallest possible size, the R ↔ 1/R duality, and extra dimensions that must also evolve. Brandenberger and Vafa pictured the universe beginning as a hot Planck-sized nugget in which wound strings keep all dimensions small until, in exactly three of them, strings annihilate and let space grow. Gasperini and Veneziano proposed a pre-big bang past. Beyond the physics lies a harder question: if our universe is one of many, with different laws in each, some features we hoped to derive may be cosmic accidents, and the "ultimate theory" may need cosmology to explain them.
The standard model of cosmology
Einstein refused to accept that his equations implied a changing universe. Alexander Friedmann did, finding the big-bang solution. Einstein published a note claiming a fatal flaw, then, about eight months later, curtly retracted it, though he still thought the result irrelevant. About five years on, Edwin Hubble's observations of a few dozen galaxies with the 100-inch Mount Wilson telescope showed the universe is expanding. Friedmann's work, systematised by Howard Robertson and Arthur Walker, remains the foundation.
10−43 s (Planck time): about 1032 K, "10 trillion trillion times hotter than the deep interior of the sun." Before this, string theory must take over.
Before ~10−35 s: above 1028 K, the strong, weak and electromagnetic forces are one grand unified force.
~10−36 to 10−34 s: inflation: expansion by a factor of at least 1030.
~10−11 s: at about 1015 K the electroweak force splits into weak and electromagnetic.
~10−5 s: at about 10 trillion K quarks clump in threes into protons and neutrons.
0.01 s to 3 minutes: primordial nucleosynthesis: hydrogen and helium nuclei, traces of deuterium and lithium, as the universe cools to a billion degrees.
A few hundred thousand years: at a few thousand degrees, electrons join nuclei into neutral atoms; the fog lifts and the universe becomes transparent.
About a billion years (Greene's figure): galaxies, stars, planets condense.
Today: "some 15 billion or so years" in the 1999 text; now measured at 13.8 billion.
The big bang happened everywhere at once: "it took place where you are now," since all now-separate places were once the same place.
Testing the big bang
Three pillars of evidence
- Expansion. Telescopes see galaxies and quasars from a few billion years after the bang; the expansion checks out "to a 'T.'"
- The cosmic microwave background. Compressed gas heats (a pumped bicycle tyre, a pressure cooker); expanding gas cools ("blow off steam"). The photon gas released at transparency has cooled with expansion. Predicted by George Gamow, Ralph Alpher and Robert Herman (1940s–50s) and by Robert Dicke and Jim Peebles (1960s); found by accident in 1965 by Arno Penzias and Robert Wilson at Bell Labs while working on a satellite-communications antenna; measured precisely by NASA's COBE (early 1990s): 2.7 K, about 400 million photons per cubic metre. Part of the "snow" on an untuned analogue television is this afterglow.
- Light elements. Nuclear physics predicts about 23% helium (by mass), matched in stars and nebulae; deuterium, which essentially nothing but the big bang can make; and lithium.
- Verdict: the theory works from a hundredth of a second to today, "impressive almost to the point of hubris." But long-lasting features were imprinted in far smaller fractions of a second.
Symmetry breaking: the universe freezes
Water shows how cooling breaks symmetry. Steam is maximally symmetric ("molecular democracy"); liquid less so, as molecules clump; ice less still, a crystal that looks different from different angles. Each phase transition abruptly lowers symmetry. The early universe passed through at least two:
- Below ~1028 K: the grand unified force splits, strong separating from electroweak.
- Below ~1015 K (about 100 million times the sun's core temperature): the electroweak force splits into weak and electromagnetic (Glashow, Salam, Weinberg).
- Lesson: the three forces look distinct today, but "are deeply related."
The horizon problem and inflation
Why is the sky the same temperature everywhere? The microwave background is uniform to about one part in 100,000 in every direction. Hot soup cools to room temperature only through prolonged contact with the air (not in a thermos). Run the cosmic film backward: two now-distant regions do get closer, but if they are 186,000 miles apart less than a second after the bang, or one foot apart less than a billionth of a second after, light has not had time to cross between them. Gravity slows expansion like a ball tossed upward, so to halve a separation you must rewind more than halfway to the start, leaving even less time for contact. In the standard model, widely separated regions could never have exchanged heat: the horizon problem.
Alan Guth (1979; now at MIT) found a solution of Einstein's equations with a brief burst of exponential expansion, which speeds up as it goes, unlike the tossed ball. Rewound, it is a rapidly decelerating contraction, so halving a separation requires rewinding much less than halfway, leaving ample time for regions to equalise. With refinements by Andrei Linde (Stanford), Paul Steinhardt and Andreas Albrecht (Pennsylvania) and others, inflationary cosmology took over: between about 10−36 and 10−34 seconds, the universe grew by at least 1030 (versus about a hundredfold in the standard model over the same interval), more growth in a flicker than in all the billions of years since. It also solves "a number of other important problems" (the flatness problem and the absence of magnetic monopoles, among others).
What string theory changes
Before the Planck time, general relativity drives size to zero and temperature to infinity: "the most extreme signal" of breakdown. String cosmology was then "at an early stage," but three modifications stood out:
Three stringy modifications to cosmology
- A smallest size. The universe cannot shrink below about the Planck length (Ch 10).
- R ↔ 1/R duality. Shrinking through the Planck length is physically the same as bouncing back.
- Extra dimensions. Cosmology must explain the evolution of all nine space dimensions, including why only three are large.
In the beginning: a Planck-sized nugget
Brandenberger and Vafa's string cosmology (late 1980s)
- A temperature maximum. Imagine all dimensions are circles. Rewinding, the universe heats up until every radius reaches the Planck length; shrinking further is the same as expanding, so the temperature peaks and falls. Detailed calculations confirm it.
- The beginning. All nine space dimensions are curled up equally in a hot, Planck-sized nugget: high but finite temperature and energy, perfectly symmetric.
- First symmetry breaking. At about the Planck time, three dimensions are singled out to expand, then hand over to inflation and the standard history.
Why three? Strings that collide. Strings wrapped around a circular dimension act like rubber bands around an inner tube, choking its expansion. A wrapped string meeting its antistring (wrapped the opposite way) annihilates, releasing the choke. How often do they meet? Two points on a line must collide unless their velocities match; on a plane, random paths usually miss. The analogue for strings: in three or fewer dimensions, wrapped strings likely collide; in four or more, they rarely do. So a thermal fluctuation lets three dimensions grow a little; their strings collide and annihilate; with less constriction they grow further, and larger dimensions are harder to wrap. The expansion feeds on itself, in exactly three dimensions.
Choosing the Calabi-Yau through history. Since conifold transitions let any Calabi-Yau evolve into another (Ch 13), the hot early universe may have run through a frenetic dance of shapes, the fabric "ripping apart and reconnecting over and over," settling as it cooled into one shape, perhaps ours. "The issue of selecting one Calabi-Yau shape from the many may in fact be reduced to a problem of cosmology."
Vafa's caveat: the work "only provides a first look into string cosmology, and is very far from the final word."
Before the beginning? The pre-big bang
Gabriele Veneziano and Maurizio Gasperini (Turin) also use the minimum length to avoid infinities, but propose a prehistory:
- Start: the universe is cold and essentially infinite, not hot and tiny.
- Instability: string equations drive every point away from every other, an inflation-like acceleration that makes space more curved and hotter.
- Result: a millimetre-sized patch comes to look like the hot, dense region emerging from Guth's inflation, then evolves by standard cosmology into our universe. The horizon problem is solved automatically. Veneziano: "String theory offers us a version of inflationary cosmology on a silver platter."
- Status (1999): "heated, yet fruitful debate."
M-theory and the merger of gravity
The three non-gravitational forces meet near one point at high energy (Ch 7). Gravity, in the simplest perturbative string models, almost but not quite joins them; careful tuning of the Calabi-Yau could fix the mismatch, but after-the-fact fine-tuning makes physicists uncomfortable. Witten showed that letting the string coupling be large (the M-theory regime, with an eleventh dimension of finite size) nudges gravity's curve to meet the others without special tuning, a hint that "cosmological unity is more easily achieved" in M-theory.
Cosmological speculation and the ultimate theory
Two limits on what a final theory can explain
- Initial conditions. Laws alone don't say where a tossed ball lands; you also need its starting velocity. Some features of our universe, perhaps even particle properties (via which Calabi-Yau the dance ended on), may be historical accidents of how things started. Hawking and James Hartle tried to bring initial conditions under physical law (the no-boundary proposal); "all such attempts remain inconclusive."
- The multiverse. Andrei Linde: inflation may happen repeatedly, budding new island universes endlessly. Physics could vary among them: slightly (electron mass off by a thousandth of a percent), strongly (up-quark ten times heavier), or wildly (different particles, zero to ten large dimensions).
| The dream of inevitability | The anthropic multiverse |
|---|---|
| Things are as they are because they could not be otherwise (Ch 12) | Things are as they are because if they weren't, we wouldn't be here to notice |
| Every constant derived from the theory | Constants vary across universes; we find ourselves in a life-friendly one, like survivors of mass Russian roulette who shouldn't be surprised to have survived |
| "Epitome of poetic grace" | "A wildly excessive collection of universes with an insatiable appetite for variety" |
Greene's stance: we should demand a consistent quantum description of all forces and matter and a cogent cosmology for our universe; if the multiverse is real ("a huge if"), it may be asking too much to derive every particle mass and force strength. But even then, a theory of the whole multiverse might explain how the parameters are distributed.
Lee Smolin's cosmological natural selection
- Idea. The big bang and a black hole's centre are both regions of crushed matter, so perhaps every black hole seeds a new universe behind its horizon.
- Mutation. Offspring universes inherit slightly altered constants.
- Selection. Universes whose constants make more black holes (via more stars) have more offspring, so after many generations the population is dominated by black-hole-optimised universes.
- Prediction. If we're typical, any change to our constants should make black holes harder to form. "There is no consensus on its validity."
The chapter ends with hope: without Planck-energy accelerators, physicists will rely on "the cosmological accelerator of the big bang" and its relics.
| Quantity | Greene's text (1999) | Measured by 2026 |
|---|---|---|
| Age of universe | ~15 billion years | 13.8 billion years (Planck 2018) |
| CMB temperature | 2.7 K | 2.7255 K; ~411 photons per cm³ (Greene's 400 million per m³ is right) |
| CMB uniformity | 1 part in 100,000 | Fluctuations mapped in detail (WMAP 2001–10, Planck 2009–13, ACT, SPT) |
| Helium (by mass) | ~23% | ~24.5% |
| Lithium-7 | "confirmed" | Predicted abundance is ~3× too high: the unresolved lithium problem |
| Composition | Dark matter unknown; fate unclear | ~5% ordinary matter, ~27% dark matter, ~68% dark energy; space flat to ~0.2% |
| First galaxies | ~1 billion years | JWST galaxies at ~300 million years (JADES-GS-z14-0, 2024; MoM-z14, 2025) |
| Hubble constant | not given | 67.4 (CMB) vs ~73 (supernovae) km/s/Mpc: the Hubble tension |
The 1998 revolution is missing, and "ultimate" theory meets the anthropic principle. Greene's chapter predates or ignores the accelerating expansion discovered in 1998 (Nobel 2011). It matters here: Steven Weinberg had argued in 1987, on anthropic grounds, that the cosmological constant should be small but not zero, roughly what was found: the most cited success of anthropic reasoning. Combined with the string landscape (10500+ vacua, 2000–03) and eternal inflation, the multiverse moved from the "outskirts of mainstream science" (Greene's phrase) to a widely debated, and widely criticised, mainstream position. Critics (Paul Steinhardt, a co-founder of inflation; George Ellis; Joe Silk; Peter Woit) argue a theory that can accommodate any outcome forfeits testability. Note also a slip: the text gives the Planck time as both 10−43 and 10−41 seconds; the correct value is about 5 × 10−44 s.
Inflation's scorecard.
- Confirmed predictions: a flat universe; nearly scale-invariant, Gaussian, adiabatic fluctuations with spectral index ns ≈ 0.965 (slightly below 1, as most models predict); correlations on scales larger than the horizon at recombination.
- Not yet seen: primordial gravitational waves (B-mode polarisation). BICEP2's 2014 claim turned out to be dust; BICEP/Keck now set r < 0.036 (2021), ruling out the simplest textbook models (m²φ², φ⁴).
- New pressure: ACT DR6 (2025), combined with other data, prefers a slightly higher ns (≈ 0.974), putting the long-favoured Starobinsky model under mild strain.
- Critics: Ijjas, Steinhardt and Loeb (2017) argued inflation is too flexible to falsify; over 30 leading cosmologists replied in defence. LiteBIRD (Japan, planned for the 2030s), the Simons Observatory (observing since 2024) and CMB-S4 successors aim to reach r ~ 0.001.
String cosmology, 1999–2026.
- Brane inflation (Dvali and Tye, 1999; KKLMMT, 2003): inflation driven by branes moving in the extra dimensions.
- Ekpyrotic/cyclic universe (Khoury, Ovrut, Steinhardt and Turok, 2001): our universe as a brane that collides with another, a stringy rival to inflation; it predicts no detectable primordial gravitational waves.
- The pre-big bang and string-gas scenarios both predict a blue-tilted gravitational-wave background, not yet seen.
- The swampland (2018 on): Obied, Ooguri, Spodyneiko and Vafa conjectured that stable de Sitter space, and hence a true cosmological constant, may be impossible in string theory, implying dark energy must change over time. DESI results (2024, 2025) hint at evolving dark energy at 2.8–4.2σ; unconfirmed, but watched closely by string theorists.
- Gravity's merger: Witten's 1996 strong-coupling result (via Hořava–Witten M-theory) lowers the string scale toward ~1016 GeV, letting gravity meet the other forces, a mainstay of heterotic M-theory phenomenology.
- Smolin's prediction was that no neutron star should exceed about 1.6 solar masses; pulsars of ~2.0 solar masses (PSR J1614−2230, 2010; J0740+6620, 2019) are widely read as evidence against his specific proposal.
- JWST's early, bright galaxies (2022 on) have stretched galaxy-formation models but not overturned the standard cosmology.
Separate the laws from the history. Greene's tossed-ball point is a discipline for any forecasting work: decide which features of your system are structural (laws that a better model should derive) and which are path-dependent (outcomes of initial conditions and history that no model will predict). Spending effort to "derive" an accident is as wasteful as accepting a structural pattern as luck. Ask, as cosmologists now must: is this number explained, selected, or sampled?
Prospects
String theory may one day be unrecognisable, a stepping stone to something grander, or it may be the end of the search. Most string theorists expected a third and fourth revolution before its full power could be judged. Greene closes with five open questions: what principle makes string theory inevitable; what space and time really are; whether quantum mechanics must be reformulated; whether the theory can be tested; and whether there are limits to explanation. The deepest hope is that spacetime is not a stage on which strings move but something strings weave: a coherent state of gravitons that emerges from a spaceless, timeless beginning.
"I feel that we are so close with string theory that, in my moments of greatest optimism, I imagine that any day, the final form of the theory might drop out of the sky and land in someone's lap. But more realistically... Well into the twenty-first century, when I am too old to have any useful thoughts on the subject, younger physicists will have to decide whether we have in fact found the final theory." Edward Witten
Question 1: what is the fundamental principle?
The past century taught that laws come from symmetry:
| Theory | Founding principle |
|---|---|
| Special relativity | All constant-velocity observers are equivalent |
| General relativity | Equivalence principle: all observers, in any motion, are equivalent; gravity exists so that this holds |
| Strong, weak, electromagnetic forces | Gauge symmetries; the forces exist to enforce them |
| String theory | Unknown: all the above symmetries, plus supersymmetry, emerge from it as consequences |
String theory goes "another notch" deeper: had it been found a century earlier, these symmetries might have been discovered by studying it. But is string theory itself the inevitable consequence of some broader principle? No one knows. String theorists are like "an Einstein bereft of the equivalence principle," without the 1907 "happy thought" in the Bern patent office. Since Veneziano's 1968 guess, the theory has been assembled "discovery by discovery, revolution by revolution." Finding its "principle of inevitability" is "of the highest priority."
Question 2: what are space and time, and can we do without them?
| Newton: absolute space and time | Leibniz (and Mach): relational space and time |
|---|---|
| "Absolute space, in its own nature, without relation to anything external, remains always similar and immovable. Absolute, true, and mathematical time... Flows equably" (Principia) | Space and time are bookkeeping devices for relations between objects and events; a location has meaning only relative to another |
| Held sway for over 200 years, backed by Newton's successful laws | Much closer to the modern view; relativity abolished absolute space and time |
Spacetime as a fabric woven from strings. A light beam is a vast number of photons; a gravitational field is a vast number of gravitons, that is, strings in the graviton vibration pattern. Since gravity is spacetime's warping, the fabric of spacetime may be a "coherent state" of an enormous number of strings vibrating in the graviton pattern: strings as the threads of spacetime. If so, is there a raw, unwoven state? Not "a jumble of strings in space," which presupposes space and time, but a state with no space, no time, not even "before", in which strings are "shards" of spacetime that coalesce into it only through "sympathetic vibrations."
- Why it matters. Current string theory presupposes a spacetime in which strings move, like judging an artist with a paint-by-numbers kit. Since string theory's triumph is gravity, and gravity is spacetime, the theory should start from a blank canvas and create its own arena.
- The image's limit. Like Steven Wright's photographer trying to get a close-up of the horizon, we hit "a clash of paradigms."
- A first glimpse. Shenker, Witten, Banks, Fischler, Susskind and others found that zero-branes, point-like at large distances but very different up close and perhaps M-theory's most basic ingredient, replace ordinary geometry below the Planck length with non-commutative geometry (much developed by Alain Connes), where "the conventional notions of space and of distance between points melt away." Above the Planck length, ordinary space re-emerges.
Question 3: must quantum mechanics be reformulated?
The traditional recipe writes a classical theory (string theory starts with equations for a classical thread that "Newton could have written down") and then quantises it. This built the standard model. But with string/M-theory "we may well have come to the end of the line" for it.
Dualities are quantum symmetries. Duality often maps a process that depends strongly on quantum mechanics in one string theory to one that depends only weakly on it in another. So the dualities only exist in the quantum theory; a classical starting point necessarily omits them. The full theory must "spring into existence as a full-fledged quantum-mechanical theory."
- Cumrun Vafa: "a reformulation of quantum mechanics which will resolve many of its puzzles is just around the corner... A new, more geometrical framework for quantum mechanics, in which space, time, and quantum properties will be inseparably joined together."
- Edward Witten: "the logical status of quantum mechanics is going to change in a manner that is similar to the way that the logical status of gravity changed when Einstein discovered the equivalence principle."
- The prize: perhaps an answer to how the universe began, why there is space and time, and even Leibniz's question: why is there something rather than nothing?
Question 4: can string theory be tested?
Greene's case that it is testable in principle
- It contains gravity and quantum mechanics, which any unified theory must.
- It naturally produces the key ideas of the past century: spin, families, messenger particles, gauge symmetry, the equivalence principle, symmetry breaking, supersymmetry.
- It has no adjustable parameters, unlike the standard model's 19, so its predictions should be "thoroughly definitive."
- In practice: the obstacles of Chapter 9 (which Calabi-Yau?) and Chapters 12–13 (no exact equations) stand in the way.
- The big hope: supersymmetry at the LHC, which would be "compelling, albeit circumstantial" evidence, and whose superpartner masses would challenge string theorists to explain them. Ideally, string theory would predict superpartner properties before the LHC switched on: "a monumental moment in the history of science."
Question 5: are there limits to explanation?
Even with exact M-theory and a new quantum mechanics, we might still fail to calculate quark masses and force strengths, not because the theory is incomplete but because there is no explanation. "One immediate answer to all these questions is yes." Einstein: "The most incomprehensible thing about the universe is that it is comprehensible." Some features may be "happenstance, accident, or divine choice." Hitting a true limit of explanation, rather than a temporary one, "would be a singular event." The multiverse seems to impose one, but speculative theories (Ch 14) might restore prediction; cosmology will likely decide.
Reaching for the stars
The measure of a theory's depth is how far it upends our worldview: quantum mechanics and relativity brought wave functions, tunnelling, vacuum fluctuations, relative simultaneity, warped spacetime, black holes, the big bang. String/M-theory adds vibrating loops and "oscillating globules," hidden dimensions that tear and repair, perhaps a timeless, spaceless domain below the Planck length, and a multiverse of "frothing bubbles." Einstein described his own search as "the years of anxious searching in the dark, with their intense longing, their alternations of confidence and exhaustion, and final emergence into the light." Each generation "stands firmly on the shoulders of the previous," and, in Jacob Bronowski's words, "in every age there is a turning point, a new way of seeing and asserting the coherence of the world."
| Greene's question (1999) | Where it stands in 2026 |
|---|---|
| 1. Fundamental principle? | Still missing. Candidates: holography (Ch 16); bootstrap results suggesting string amplitudes are uniquely fixed by consistency (e.g. Cheung and Remmen, 2023); "string universality" in high dimensions |
| 2. Emergent spacetime? | Major progress: AdS/CFT (1997) builds spacetime from a non-gravitational quantum system; Ryu–Takayanagi (2006) and Van Raamsdonk (2010) tie geometry to entanglement; spacetime as a quantum error-correcting code (2015). Only for anti-de Sitter, not our universe |
| 3. New quantum mechanics? | No reformulation. Instead, quantum mechanics became central, via quantum information, to understanding gravity; no experiment has found a deviation from quantum theory |
| 4. Testable? | No direct test. LHC found the Higgs (2012) but no superpartners; HL-LHC from ~2030; CERN's Future Circular Collider feasibility study (2025) proposes a ~91 km ring, decision expected ~2028 |
| 5. Limits to explanation? | The landscape and multiverse (2000s) made this the field's central controversy; no consensus |
"No adjustable parameters" is the claim most changed by history. The equations have no free numbers, but their solutions do: the choice of Calabi-Yau, of fluxes, of brane configurations acts as a vast set of discrete parameters, perhaps 10500 or more. The theory therefore does not, as yet, give "an unambiguous test of whether the theory is right or wrong." Greene recognised the danger (Ch 12, 14), and later wrote about it in The Fabric of the Cosmos (2004) and The Hidden Reality (2011). The "19 free parameters" of the standard model is also now about 26 with neutrino masses.
The hoped-for "third revolution." What came closest is holography:
- Maldacena's AdS/CFT (late 1997, two years before this book) became the most-cited result in the field and the main laboratory for quantum gravity.
- Black-hole information largely resolved within AdS (islands and the Page curve, 2019–20).
- Spacetime from entanglement became mainstream through the "It from Qubit" collaboration (Simons Foundation, 2015–23).
- Non-commutative geometry, Greene's glimpse of the spaceless domain, was sharpened by Seiberg and Witten (1999), who showed it arises on branes in a background field.
- The open frontier. Holography for our own expanding, de Sitter-like universe remains unsolved, and the "principle of inevitability" has not been found. Witten, now in his mid-70s, has said the theory's ultimate meaning is still unclear.
Name the open questions explicitly. Greene ends not with conclusions but with five precise questions, each with a clear test of what would count as an answer. Mature programmes (scientific or corporate) benefit from the same discipline: publish the five questions you can't yet answer, and say what evidence would resolve each. Twenty-seven years on, his list still reads as a fair scorecard for the field.
Strings Since 1999
This chapter is not by Brian Greene. It was written for this Insight Edition to bring the book up to date at the same depth as the original chapters. Greene wrote in 1999 at a moment of high hope: dualities, M-theory, black-hole entropy, and the LHC approaching. Twenty-seven years later the scorecard is mixed and surprising. Experiment delivered the Higgs boson and gravitational waves but no superpartners, no extra dimensions and no string signatures. Theory found that string theory has a vast landscape of solutions rather than one, provoking the multiverse debate, the counter-programme of the swampland, and sharp public criticism. Meanwhile holography became the most productive idea in theoretical physics, reshaping work on black holes, quantum information, nuclear matter and mathematics. String theory has not become the tested theory of everything Greene hoped for; it has become a framework whose deepest legacy so far is the discovery that spacetime and gravity can emerge from quantum systems without either.
The quarter-century at a glance
1997: Juan Maldacena proposes the AdS/CFT correspondence (holography). Within a year Gubser, Klebanov, Polyakov and Witten make it precise.
1998: Two supernova teams discover the accelerating universe (Nobel 2011). Super-Kamiokande finds neutrino mass.
1999: The Elegant Universe is published. Randall–Sundrum warped extra dimensions; Seiberg–Witten non-commutative geometry.
2000–03: Bousso–Polchinski (2000), Giddings–Kachru–Polchinski (2001) and KKLT (2003) build stable string vacua with fluxes: the landscape, perhaps 10500 vacua. Leonard Susskind popularises the "anthropic landscape" (2003).
2005: Vafa introduces the swampland. String-inspired calculation gives the quark-gluon plasma's viscosity bound (Kovtun, Son and Starinets).
2006: Lee Smolin (The Trouble with Physics) and Peter Woit (Not Even Wrong) publish book-length critiques. Ryu–Takayanagi formula links geometry to entanglement.
2008: The LHC starts (first collisions at high energy, 2010). ABJM describes M2-branes. BCJ double copy: gravity amplitudes as "gauge theory squared."
2012: The Higgs boson is discovered at 125 GeV. First Breakthrough Prizes honour Witten, Maldacena, Arkani-Hamed, Kitaev, Seiberg, Sen, Kontsevich, Linde and Guth.
2013: ER = EPR; the amplituhedron.
2015: LIGO detects gravitational waves (announced 2016); spacetime as a quantum error-correcting code.
2016: LHC Run 2 at 13 TeV: no superpartners; the 750 GeV "diphoton bump" vanishes.
2018: De Sitter swampland conjecture. Joseph Polchinski and Stephen Hawking die.
2019: Black-hole "islands" and the Page curve. First black-hole image (M87*).
2022: The "dark dimension" proposal; LHC Run 3 begins at 13.6 TeV.
2024–25: DESI hints that dark energy evolves (2.8–4.2σ); the muon g-2 anomaly dissolves; GW250114 confirms Hawking's area theorem.
June 2026: LHC Run 3 ends; Long Shutdown 3 begins ahead of the High-Luminosity LHC (target 2030).
1. The experimental verdict
Greene's chief hope (Ch 7, 9, 15) was that the LHC would find supersymmetry, ideally after string theory had predicted superpartner masses. Neither happened.
| What 1999 hoped for | What the LHC found (2010–2026) |
|---|---|
| The Higgs boson, last piece of the standard model | Found (2012) at 125 GeV, behaving as the standard model predicts so far |
| Superpartners near 1,000 proton masses | None; gluinos above ~2.3 TeV, top squarks above ~1.3 TeV in simple models |
| Large extra dimensions (ADD, Randall–Sundrum) | None; no Kaluza-Klein gravitons, no missing-energy excess |
| Microscopic black holes | None produced |
| String resonances or fractional charges | None |
| New forces (Z′ bosons) | None up to ~5 TeV in standard benchmarks |
Why the null results matter more for supersymmetry than for strings. String theory needs supersymmetry at some energy, but nothing requires it at LHC energies. What died was the naturalness argument (Ch 7): the hope that superpartners would appear just above the W and Z to cancel quantum corrections to the Higgs mass. The 125 GeV Higgs with no partners nearby means nature is fine-tuned at roughly the one-in-a-hundred to one-in-a-thousand level in the simplest models, or naturalness was the wrong guide. That crisis is shared by all approaches to physics beyond the standard model, not only strings.
What next. The High-Luminosity LHC (from about 2030) will collect roughly ten times more data. CERN's Future Circular Collider (a ~91 km ring; feasibility study 2025) would first study the Higgs precisely, then push to ~100 TeV; a decision is expected around 2028, and China has proposed a similar machine (CEPC). None comes near the Planck scale: the gap is still about fifteen orders of magnitude (Ch 9).
2. From one theory to a landscape of solutions
Greene's Chapter 12 hoped the exact equations would select one vacuum. Work in the 2000s found the opposite.
How the landscape arose
- The moduli problem. A Calabi-Yau comes with dozens or hundreds of moduli, numbers that set its sizes and shapes. Unfixed, they would show up as massless fields and new long-range forces, which experiments exclude.
- Fluxes fix them. Threading the Calabi-Yau's holes with generalised magnetic fluxes (Giddings, Kachru and Polchinski, 2001) stabilises most moduli. Each hole can carry a whole number of flux units.
- Combinatorial explosion. With hundreds of holes and many flux choices each, Bousso and Polchinski (2000) estimated a huge number of vacua with different cosmological constants, densely enough spaced that some have the tiny value we observe. Douglas and collaborators estimated 10500; Taylor and Wang (2015) found 10272,000 in one F-theory geometry.
- De Sitter vacua. KKLT (Kachru, Kallosh, Linde and Trivedi, 2003) and the Large Volume Scenario (2005) proposed ways to get a positive cosmological constant, matching the accelerating universe.
- Eternal inflation (Vilenkin, Linde) could populate them all: the multiverse of Chapter 14.
| The landscape as discovery | The landscape as failure |
|---|---|
| Susskind, Bousso, Polchinski and others: a theory with many vacua plus eternal inflation explains why constants look fine-tuned, as Weinberg's 1987 anthropic estimate of the cosmological constant was later borne out | Smolin, Woit, Steinhardt, Ellis, Silk: a framework that fits any set of constants predicts nothing; "anthropic" reasoning is giving up on explanation |
| Like planetary orbits: Kepler sought a rule for the planets' distances; they turned out to be environmental accidents | Like epicycles: endless adjustable structure dressed up as depth |
| Measures over the multiverse may still yield statistical predictions | The measure problem (how to count infinitely many universes) is unsolved |
The methodology fight. In 2014 George Ellis and Joe Silk warned in Nature ("Defend the integrity of physics") against accepting theories on non-empirical grounds. Philosopher Richard Dawid (String Theory and the Scientific Method, 2013) argued that "no alternatives," "unexpected explanatory coherence" and "meta-inductive" track records can legitimately raise confidence. A 2015 Munich workshop ("Why Trust a Theory?") put Gross, Dawid, Ellis, Silk and others in one room; it ended without agreement. Sabine Hossenfelder's Lost in Math (2018) took the argument to a wide public, targeting the reliance on beauty and naturalness that Greene's Chapter 7 celebrates.
Is the landscape even real? The constructions rely on approximations whose control is disputed. Critics within string theory (for example Danielsson and Van Riet, 2018; Sethi; Bena, Graña and collaborators) argue that the de Sitter step of KKLT may fail when all corrections are included. Defenders (Kachru, Kallosh, Linde, McAllister and others) have produced increasingly explicit constructions: computer-assisted flux vacua with exponentially small (negative) cosmological constants (Demirtas, Kim, McAllister, Moritz and Rios-Tascon, 2021), then explicit "candidate de Sitter vacua" (McAllister, Moritz, Nally and Schachner, 2024). As of 2026 there is no consensus that string theory has even one fully controlled de Sitter vacuum like ours.
3. The swampland: what string theory forbids
If string theory allows so much, the natural next step is to ask what it rules out. In 2005 Cumrun Vafa proposed the swampland: the vast set of apparently consistent low-energy theories that cannot come from quantum gravity. The goal is to turn string theory's richness into constraints.
Main swampland conjectures
- Weak Gravity Conjecture (Arkani-Hamed, Motl, Nicolis and Vafa, 2006): gravity must be the weakest force; some particle must have charge-to-mass ratio large enough that electric repulsion beats gravitational attraction. Our world complies: for the electron, by about 1021. Now the best-supported conjecture, with proofs in many settings.
- No global symmetries in quantum gravity (long suspected; proved in AdS/CFT by Harlow and Ooguri, 2018).
- Distance conjecture (Ooguri and Vafa, 2006): moving far in the space of vacua brings down an infinite tower of light states; effective theories break down.
- De Sitter conjecture (Obied, Ooguri, Spodyneiko and Vafa, 2018): stable de Sitter vacua may not exist; dark energy must be slowly changing (quintessence). Refined versions followed after criticism.
- Trans-Planckian Censorship (Bedroya and Vafa, 2019): no fluctuation shorter than the Planck length may be stretched into a classical one, which would sharply limit inflation.
- The "dark dimension" (Montero, Vafa and Valenzuela, 2022): combining these with the observed dark energy suggests one extra dimension of about a micrometre, testable by short-range gravity experiments now reaching ~50 μm.
Why DESI matters to string theory. If dark energy is a true cosmological constant, string theory needs stable de Sitter vacua, which the swampland conjectures doubt. If dark energy weakens over time, as DESI's 2024–25 results hint (2.8–4.2σ when combined with supernovae and the CMB), the swampland programme scores one of the few observational wins any quantum-gravity idea has had. The evidence is not yet decisive; DESI's full five-year data, Euclid and the Rubin Observatory will test it over the late 2020s.
4. Holography: the revolution Greene just missed
Two years before this book appeared, Juan Maldacena found that string theory in a five-dimensional anti-de Sitter space (a negatively curved space, like a can whose interior is infinitely deep) is exactly equivalent to a four-dimensional quantum field theory without gravity living on its boundary. A theory with gravity in the bulk equals a theory without gravity on the edge: a hologram.
What holography changed.
- Quantum gravity became calculable in a well-defined setting: hard gravitational questions can be posed as ordinary quantum-mechanical ones on the boundary. With over 20,000 citations, Maldacena's paper is the most cited in high-energy physics.
- Information is not lost in AdS black holes, since the boundary theory is unitary (Ch 13). Islands and replica wormholes (2019) reproduced the Page curve from the gravity side.
- Geometry from entanglement. The Ryu–Takayanagi formula (2006): the area of a minimal surface in the bulk equals the entanglement entropy of a boundary region, a generalisation of Bekenstein–Hawking. Mark Van Raamsdonk (2010): remove entanglement and space falls apart. Almheiri, Dong and Harlow (2015): bulk spacetime works like a quantum error-correcting code.
- This realises Greene's Question 2 (Ch 15): spacetime emerges from something more primitive. But only for anti-de Sitter space; our universe is closer to de Sitter, and a working holography for it remains the field's great open problem.
Holography's practical spin-offs
- Nuclear physics. Kovtun, Son and Starinets (2005) computed from black-hole horizons that the quark-gluon plasma's viscosity-to-entropy ratio should be near 1/4π; RHIC and LHC heavy-ion data found it is the most perfect fluid known, close to that bound.
- Condensed matter. "AdS/CMT" models of strange metals and superconductors; insights, though no decisive predictions.
- Quantum chaos and simulation. The SYK model (Kitaev 2015; Maldacena and Stanford 2016) is a solvable toy of holography; black holes are the fastest scramblers of information. A 2022 Google-processor "wormhole" experiment drew strong criticism as over-sold.
5. Mathematics: string theory's most unambiguous successes
Chapter 10's "Physics Wins!" story repeated many times.
| Area | String-theory input | Status |
|---|---|---|
| Mirror symmetry | Greene–Plesser, Candelas et al. (1990–91) | Homological mirror symmetry (Kontsevich, 1994) and SYZ (1996): a central field of pure maths |
| Four-manifold invariants | Seiberg–Witten equations (1994) | Revolutionised 4D topology overnight |
| Moonshine | Vertex algebras from strings | Borcherds proved monstrous moonshine (1992; Fields 1998); Mathieu moonshine found via K3 strings (Eguchi, Ooguri and Tachikawa, 2010) |
| Geometric Langlands | Kapustin and Witten (2006): S-duality of gauge theories | A full proof of the geometric Langlands conjecture (Gaitsgory, Raskin and collaborators) was announced in 2024 |
| Curve counting | Topological strings, Gopakumar–Vafa invariants | Proved and generalised; physics computations to very high genus |
| Rationality | Mirror-symmetry ideas ("Hodge atoms") | Katzarkov, Kontsevich, Pantev and Yu announced in 2024–25 that very general cubic fourfolds are not rational, a decades-old question |
6. Amplitudes: string ideas that reached the sky
The scattering amplitudes programme, partly inspired by string theory's world-sheet methods, found unexpected structure:
- BCFW recursion (2005) and the amplituhedron (Arkani-Hamed and Trnka, 2013): amplitudes as volumes of geometric objects, with spacetime locality and unitarity emerging.
- The double copy (Bern, Carrasco and Johansson, 2008): gravity = gauge theory squared, foreshadowed by string theory's KLT relations (1986), which write closed-string amplitudes as products of open-string ones.
- Gravitational waves. These methods now compute the dynamics of merging black holes to high precision (Bern and collaborators, from 2019), feeding the waveform models LIGO, Virgo and KAGRA use. String theory's mathematics now helps interpret real observations.
- String uniqueness from consistency. Bootstrap studies (2022–24) suggest that string amplitudes are close to the only consistent way to soften gravity at high energy, a modern echo of Greene's "only game in town."
7. Rivals, critics and the sociology of a field
| String theory's position in 2026 | Its alternatives |
|---|---|
| The only framework that unifies all forces and matter with gravity; richest mathematical structure; holography | Loop quantum gravity (Ashtekar, Rovelli, Smolin): discrete area and volume, bounce cosmology; no unification, difficulty recovering smooth spacetime |
| Produces finite perturbative quantum gravity; black-hole entropy for special cases | Asymptotic safety (Weinberg, Reuter): gravity may be well-behaved at high energy on its own; it predicted a Higgs mass of about 126 GeV in 2009 (Shaposhnikov and Wetterich), close to the 125 GeV found |
| No experimental confirmation; landscape undermines uniqueness | Causal sets, causal dynamical triangulations, group field theory: none has experimental support either |
- The criticism wave (2006 on). Woit and Smolin argued the field had become too dominant in hiring and too insulated from experiment; Hossenfelder later broadened the critique to beauty-driven theorising. Many string theorists accepted part of the point: the 1980s–90s promise of an imminent, unique theory of everything was oversold, as Georgi and Glashow had warned (Ch 9).
- The field's response. Much of "string theory" research now concerns quantum field theory, holography, quantum information and mathematics, less often the search for our vacuum. Strings are increasingly described as a framework, like quantum field theory, rather than a single model.
- Recognition. Michael Green and John Schwarz won a Breakthrough Prize (2014); Polchinski, Strominger and Vafa (2017). Polchinski (2018), Hawking (2018), Gell-Mann (2019), Weinberg (2021) and Peter Higgs (2024) have died. Witten turned 75 in 2026.
- Greene's own trajectory. He followed this book with The Fabric of the Cosmos (2004), The Hidden Reality (2011, on multiverses) and Until the End of Time (2020), and co-founded the World Science Festival (2008). His later writing treats the landscape and multiverse as open, serious possibilities rather than the 1999 "outskirts."
8. The scorecard
What string theory has delivered since 1999
- A working laboratory for quantum gravity (holography), resolving the information paradox in that setting.
- Microscopic black-hole entropy for supersymmetric and near-extremal cases.
- Deep mathematics: mirror symmetry, moonshine, geometric Langlands links, enumerative geometry.
- Calculational tools now used in nuclear physics and gravitational-wave astronomy.
- Constraints (the swampland), some with possible observational links (dark energy, a micrometre dimension).
What it has not delivered
- Any confirmed experimental prediction.
- A unique vacuum, or a derivation of any standard-model parameter (the electron mass, the three families' count from first principles, the cosmological constant).
- A non-perturbative definition of M-theory.
- A theory of the big bang or of a realistic black hole's singularity.
- Holography for our kind of universe.
Critical lens on this report. This chapter is an editor's synthesis, not a consensus document, and the field is contested. String theorists would stress that no rival has done better on experiment and that holography's successes are real physics; critics would stress that 27 years without a test, plus the landscape, means the original promise has failed. Both are true at once. Readers should weigh Greene's 1999 optimism against this record, and note that the most important developments (holography, the landscape, the swampland) were not foreseen in the book.
Judge a research bet by its spin-offs as well as its headline goal. String theory missed its headline target (a tested, unique theory of everything) so far, yet its tools now solve problems in mathematics, nuclear physics and gravitational-wave astronomy. For any long-horizon programme, set two scorecards: mission success (did we reach the goal?) and option value (what capabilities did the search create?). Then be honest about which one you are reporting.
Master Takeaways
The ideas worth carrying out of The Elegant Universe, strongest first, followed by the book's toolkit of frameworks, a critical verdict on the whole, a who's who, the links to technology and finance that hold up, and a glossary.
Master takeaways
- Physics' two pillars are incompatible at the Planck scale. General relativity needs smooth spacetime; quantum mechanics makes everything, including spacetime, jitter more violently at smaller scales. Below about 10−33 cm the combination produces infinities. This, not elegance alone, is why a new theory is needed.
- Strings replace points. If the basic objects are vibrating loops, every particle is a note: energy gives mass, finer features give charges, and one note is the graviton. Gravity is not added to string theory; it is required by it.
- Size cures the infinities. A string cannot probe distances shorter than itself, and its interactions are smeared across a world-sheet rather than concentrated at a point, so string calculations come out finite (proved to low orders).
- Symmetry explains forces. Every force exists to enforce a symmetry: gravity enforces the equivalence of all observers; the other three enforce gauge symmetries. String theory produces all of them, plus supersymmetry.
- Supersymmetry was the bet, and so far it has lost. String theory needs superpartners at some energy; the hope they would appear at the LHC (and solve fine-tuning) has met null results through 2026.
- Hidden dimensions are allowed and, for strings, required. Curled-up dimensions can hide below any instrument's reach (the garden hose). Superstrings need ten dimensions, M-theory eleven.
- Geometry sets physics. The shape of the curled-up dimensions (a Calabi-Yau space) fixes particle masses, charges and families; three families could mean three holes.
- The selection problem is the central obstacle. Nothing yet picks which Calabi-Yau (and, since 2000, which flux configuration) is ours; the possibilities number at least hundreds of millions of shapes and perhaps 10500 vacua.
- Distance has two meanings for strings. Wound and unwound strings measure R and 1/R; circles of radius R and 1/R are the same universe, so space has a minimum size and a crunch becomes a bounce.
- Different shapes can give identical physics. Mirror symmetry pairs geometrically distinct Calabi-Yaus; calculations impossible on one are easy on the other ("Physics Wins!").
- Space can tear without catastrophe. Flop and conifold transitions change topology; strings and wrapped branes shield the universe from harm, and a black hole can turn into a massless particle.
- Perturbation theory hid most of the theory. The approximation assumes a small coupling; when it fails, new structure appears: membranes, an eleventh dimension, branes of every size.
- Dualities are the master tool. Strong coupling in one theory is weak coupling in another; the five string theories are five windows on one M-theory. Holography (1997) is the most consequential duality of all.
- BPS states are anchors. Objects with the minimum mass for their charge are fixed by supersymmetry at any coupling, letting physicists see beyond perturbation theory.
- Black-hole entropy has a microscopic explanation for certain black holes: counting brane configurations reproduces area/4 exactly (Strominger–Vafa, 1996).
- Black holes glow and may preserve information. Hawking radiation is established theory (unobserved); work since 2019 strongly suggests information escapes.
- Inflation solved the horizon problem; string theory may replace the big bang's infinite beginning with a Planck-sized one, and might explain why three dimensions grew large.
- A final theory may not explain everything. Initial conditions and a possible multiverse could make some constants environmental accidents; the anthropic debate is now central.
- Elegance guides but does not decide. Einstein's "sorry for the dear Lord" confidence was vindicated; supersymmetry's equally confident aesthetics so far were not.
- "No adjustable parameters" became "a landscape of solutions." The equations are rigid; their solutions are not. This is the biggest change since 1999.
- Spacetime may be emergent. Greene's hope that spacetime is woven from strings found concrete form in holography, where entanglement builds geometry, though only for anti-de Sitter space so far.
- Physics now repays mathematics. Mirror symmetry, Seiberg–Witten invariants, moonshine and geometric Langlands show string theory's mathematical fertility, its least contested success.
- The energy gap is enormous. Even after the LHC, the Planck scale is ~1015 times further; any test must be indirect or cosmological.
- Healthy scepticism is part of the story. Glashow, Georgi and Feynman in the 1980s, Smolin, Woit and Hossenfelder later: the critique that string theory lacks testability has not been answered by data.
- The questions outlive the answers. Greene's five closing questions (principle, emergent spacetime, quantum reformulation, tests, limits of explanation) still define the frontier in 2026.
Frameworks and mental models
| Name | What it says | Use it when |
|---|---|---|
| Three conflicts (Ch 1) | Newton vs. Maxwell → special relativity; special relativity vs. Newtonian gravity → general relativity; general relativity vs. Quantum mechanics → strings | Reading the history of physics as conflict resolution |
| Equivalence principle | Gravity and acceleration are indistinguishable | Turning a force into geometry |
| Uncertainty as borrowing | Shorter times allow bigger energy "loans" | Understanding vacuum fluctuations and foam |
| Probe resolution (peach pit) | A probe can't resolve features smaller than itself | Judging what a measurement can and can't see |
| Vibration = particle | Mass and charge are properties of a string's resonant pattern | Thinking about unification as one ingredient, many states |
| Running couplings | Force strengths change with energy and may meet | Reasoning about grand unification |
| Kaluza-Klein | Extra curled-up dimensions can unify forces | Seeing how hidden structure shapes visible outcomes |
| Geometry → physics | Calabi-Yau shape sets masses, charges, families | Linking model structure to observable output |
| T-duality (R ↔ 1/R) | Two radii, one physics; minimum length | Spotting equivalent descriptions that differ only in labelling |
| Mirror symmetry | Different shapes, same physics; compute on the easier one | Reframing an intractable problem |
| Perturbation theory | Ballpark plus small corrections; fails when coupling ≥ 1 | Judging when an approximation can be trusted |
| Strong-weak duality | Hard regime of one theory = easy regime of another | Translating problems between representations |
| BPS anchoring | Quantities protected by symmetry survive all regimes | Finding invariants that hold under stress |
| Bekenstein–Hawking entropy | Entropy = horizon area / 4 | Connecting information and geometry |
| Horizon problem / inflation | Uniformity needs past contact; inflation provides it | Checking whether a claimed correlation had a causal path |
| Anthropic selection | We observe constants compatible with our existence | Weighing "explained" vs. "selected" vs. "sampled" |
| Landscape vs. Swampland (Ch 16) | What a framework allows vs. What it forbids | Valuing a theory by the sharpness of its exclusions |
Critical lens on the whole book. The Elegant Universe remains one of the clearest popular accounts of relativity, quantum mechanics and their conflict ever written; Chapters 2–5 have aged well, and Greene is admirably candid about what string theory had not achieved (Ch 9, 15). Three weaknesses stand out from 2026. First, the optimism of timing: the book frames supersymmetry at the LHC as a likely near-term vindication, and the subtle uniqueness of the theory as within reach; neither materialised. Second, the "only game in town" framing underplays loop quantum gravity, asymptotic safety and other programmes, and treats the absence of alternatives as evidence. Third, Greene's first-person chapters (10, 11, 13) are vivid and accurate but present his own research area as more central than many in the field would. The book also predates or ignores the 1998 accelerating universe and Maldacena's holography (1997), the two developments that most shaped what followed: one pushed string theory toward the landscape and multiverse, the other gave it its greatest successes. Read it as a superb guide to why physicists pursue string theory, and pair it with Chapter 16 for what has happened since.
Who's who
| Person | Role | Why they matter here |
|---|---|---|
| Albert Einstein | Physicist | Special and general relativity; lifelong search for unification |
| Max Planck, Niels Bohr, Werner Heisenberg, Erwin Schrödinger | Physicists | Founders of quantum mechanics |
| Richard Feynman | Physicist | Sum over paths; QED; early string sceptic |
| Theodor Kaluza, Oskar Klein | Mathematician, physicist | Extra curled-up dimensions unify gravity and electromagnetism |
| Gabriele Veneziano | Physicist | Euler beta-function amplitude (1968), the seed of string theory; pre-big bang cosmology |
| Yoichiro Nambu, Holger Nielsen, Leonard Susskind | Physicists | Showed Veneziano's formula describes strings (1970); Susskind later championed the landscape and holography |
| John Schwarz, Joël Scherk | Physicists | Strings as a quantum theory of gravity (1974) |
| Michael Green, John Schwarz | Physicists | Anomaly cancellation (1984): the first superstring revolution |
| Pierre Ramond, André Neveu, Ferdinando Gliozzi, David Olive | Physicists | Fermionic strings and the superstring |
| Julius Wess, Bruno Zumino | Physicists | Supersymmetric field theory |
| Sheldon Glashow, Abdus Salam, Steven Weinberg | Physicists | Electroweak unification; Glashow a leading string critic |
| Howard Georgi | Physicist | Grand unification; string sceptic turned partial convert |
| Eugenio Calabi, Shing-Tung Yau | Mathematicians | Calabi-Yau spaces |
| Philip Candelas, Gary Horowitz, Andrew Strominger, Edward Witten | Physicists | Calabi-Yau compactification (1985) |
| Edward Witten | Physicist | M-theory (1995), dualities, flops; the field's central figure |
| Brian Greene, Ronen Plesser | Physicists | Mirror manifolds (1990) |
| Paul Aspinwall, David Morrison | Physicist, mathematician | Flop transitions with Greene (1992–93) |
| Cumrun Vafa | Physicist | T-duality cosmology, mirror symmetry, black-hole entropy, the swampland |
| Joseph Polchinski | Physicist | D-branes (1995); flux landscape |
| Nathan Seiberg | Physicist | Supersymmetry as an exact non-perturbative tool |
| Ashoke Sen, Chris Hull, Paul Townsend, Michael Duff | Physicists | Dualities and the eleventh dimension |
| Petr Hořava | Physicist | Heterotic M-theory with Witten |
| Jacob Bekenstein, Stephen Hawking | Physicists | Black-hole entropy and radiation |
| Robert Brandenberger | Physicist | String-gas cosmology with Vafa |
| Alan Guth, Andrei Linde | Cosmologists | Inflation; eternal inflation and the multiverse |
| Lee Smolin | Physicist | Cosmological natural selection; leading critic |
| Juan Maldacena | Physicist | AdS/CFT holography (1997), see Ch 16 |
Connect it
- Equivalent representations in data engineering. Mirror symmetry and dualities are the physics version of choosing the right representation: the same risk computation can be intractable as a join and trivial as a graph walk. Build the habit of asking "what is the dual of this problem?"
- Perturbation theory and model risk. A model linearised around normal conditions (a "small coupling") can fail completely under stress. Bank model-risk teams should state the regime of validity as Greene states the coupling < 1 condition, and keep non-perturbative checks (scenario simulations, invariants) for the stressed regime.
- BPS-style invariants for migrations. When migrating a ledger or a pricing engine, identify quantities that must be preserved in every representation (balances, notional totals, cash-flow sums) and test them at the extremes, as BPS states anchored string dualities.
- Hidden dimensions in AI systems. A model's behaviour is set by structure users can't see (training data, objective design, retrieval configuration), as particle properties are set by curled-up dimensions. Governance must inspect the hidden geometry, not just the outputs.
- The landscape lesson for platforms. A framework flexible enough to fit any requirement predicts nothing about which configuration you should run. Value architectures and vendors by what they forbid (constraints, guarantees), not by how much they could accommodate.
Glossary
| Term | Meaning |
|---|---|
| AdS/CFT (holography) | Duality between gravity in an anti-de Sitter space and a quantum field theory without gravity on its boundary (Maldacena, 1997) |
| Anomaly | A quantum inconsistency that breaks a classical symmetry; their cancellation (1984) made superstrings viable |
| Anthropic principle | The constants we observe must permit observers; used to explain fine-tuning in a multiverse |
| BPS state | A minimum-mass state for its charge, whose properties supersymmetry fixes exactly at any coupling |
| Brane (p-brane, D-brane) | Extended object with p space dimensions; D-branes are where open strings end |
| Calabi-Yau space | A six-dimensional shape that can serve as string theory's curled-up dimensions while preserving supersymmetry |
| Conifold transition | A space-tearing change where a collapsing three-sphere is replaced by a two-sphere; links black holes to massless particles |
| Coupling constant | Number setting an interaction's strength; for strings, the likelihood of splitting into virtual pairs |
| Cosmological constant | Energy of empty space; observed small and positive since 1998 (dark energy) |
| Duality | Two different-looking theories that describe identical physics |
| Equivalence principle | Gravity and acceleration cannot be told apart locally |
| Event horizon | A black hole's boundary of no return |
| Fermion / boson | Half-integer-spin matter particles / integer-spin force carriers |
| Flop transition | A mild space-tearing change in which a two-sphere shrinks and re-inflates in a new orientation |
| Gauge symmetry | A symmetry under local transformations that requires a force field to maintain it |
| Graviton | The quantum of gravity: a massless, spin-2 string vibration |
| Hawking radiation | Thermal emission from black holes due to quantum effects at the horizon |
| Inflation | A brief burst of exponential expansion in the very early universe |
| Kaluza-Klein theory | Theories with extra, curled-up spatial dimensions |
| Landscape | The vast set of string vacua (perhaps 10500 or more) with different physics |
| M-theory | The eleven-dimensional theory uniting the five superstring theories and supergravity |
| Mirror symmetry | Pairs of geometrically different Calabi-Yau spaces giving identical physics |
| Moduli | Parameters fixing the sizes and shapes of extra dimensions |
| Perturbation theory | Approximation by a simple estimate plus successively smaller corrections |
| Planck length / mass / time | Natural units where gravity and quantum effects meet: ~10−33 cm, ~1019 proton masses, ~5 × 10−44 s |
| Quantum foam | Wheeler's picture of spacetime churning violently at the Planck scale |
| Standard model | Quantum field theory of the strong, weak and electromagnetic forces and known particles |
| Supersymmetry (SUSY) | Symmetry pairing each boson with a fermion whose spin differs by ½ |
| Supergravity | Supersymmetric theory including gravity; the low-energy limit of strings and M-theory |
| Swampland | Low-energy theories that look consistent but cannot arise from quantum gravity |
| T-duality | Equivalence of a circle of radius R with one of radius 1/R (in string units) |
| Tachyon | A state with negative mass-squared, signalling instability; absent in superstrings |
| Winding mode | A string wrapped around a circular dimension; has no point-particle analogue |
| World-sheet | The two-dimensional surface a string sweeps out in spacetime |
About this book
The Elegant Universe: Superstrings, Hidden Dimensions, and the Quest for the Ultimate Theory by Brian Greene, 1999. 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