Insight Editions · No. 22
Euclid's Window
The Story of Geometry from Parallel Lines to Hyperspace
Leonard Mlodinow
Leonard Mlodinow traces the development of geometry from Euclid's axioms through non-Euclidean space, relativity, and modern theories of higher dimensions.
A study guide to Euclid's Window by Leonard Mlodinow. Shared free for personal reading; not for sale and not endorsed by the authors or publishers. Please read and buy the original.
Introduction: five windows on space
Twenty-four centuries ago Aristotle watched ships leave port and noticed that they vanished hull first, then masts and sails. On a flat earth they would shrink evenly to a dot. The order of disappearance meant the earth was curved. He had looked at the planet's large-scale shape through the window of geometry, without leaving the shore.
Mlodinow's claim is that we still do this. We will never watch a ship sail over the horizon of the universe, yet by observation and logic we have learned a great deal about its shape. The book is organised around five geometric revolutions, each answering a question about space:
| Revolution | Prophet | Question | New idea |
|---|---|---|---|
| I | Euclid | What can you prove about space? | Abstraction, axioms and proof |
| II | Descartes | How do you know where you are? | Coordinates: geometry married to number |
| III | Gauss, Riemann | Can space be curved? | Non-Euclidean geometry |
| IV | Einstein | How does space relate to matter and time? | Curved space-time is gravity |
| V | Witten and the string theorists | How many dimensions are there? | Extra dimensions, unified physics |
The introduction ends with a stunt. In June 1984 a physicist announced a theory of everything and was carried off stage by men in white coats. It was staged, but John Schwarz really had spent fifteen years on string theory while colleagues treated it like a stranger begging on the street. Mlodinow wrote in 2001 that "most physicists believe string theory is correct."
The geometry of space decides what physics is possible. Each revolution changed what counts as a proof, a place, a straight line or a law of nature, and each shattered a belief that had felt like plain fact.
The book's thesis is architectural: geometry is the window through which we see space, and each revolution changes the glass, not only the view. Mlodinow writes as a Caltech physicist who learned that the "obvious" Euclidean facts of school are historically contingent: they were argued, abandoned, restored and finally bent.
Five revolutions structure the narrative. Euclid makes proof the method. Descartes makes place algebraic. Gauss and Riemann free geometry from the fifth postulate and invent intrinsic curvature. Einstein makes space-time the dynamical stage of gravity. Witten and strings ask whether space itself is an approximation born of a deeper theory. The order is chronological and causal: each window needs the previous panes to open.
Keep three reading questions open. What counts as a proof? What counts as a place? What counts as space itself when the measuring rods bend?
FIVE-WINDOWS. Euclid (proof), Descartes (place), Gauss/Riemann (curve), Einstein (space-time), Witten (strings). Recite the five and the book's spine is the shelf.
Name one everyday act that assumes Euclid (drawing a square), one that needs coordinates (GPS pin), and one that needs curvature (satellite orbit). Write one sentence each. That is the five-window tour in miniature.
Turn it around. Treating the five windows as a progress ladder (each making the last obsolete) erases what each still does: Euclid still builds bridges; Descartes still runs CAD; Riemann still underwrites GR. Later windows open; they do not smash earlier panes.
Part I. Euclid: The Story of Proof
Euclid may not have discovered a single important theorem, yet he is the most famous geometer in history, because for two thousand years his book was the window everyone looked through first. Part I tells how "earth measurement" in Egypt and Babylon became Greek geometry: the leap from describing stone and sand to the ideal point, line and plane, and from rules of thumb to the axiom, the theorem and the proof.
Part I follows the birth of proof culture: Thales and the Milesian mutiny, Pythagorean number-geometry, Euclid's axiomatic tower, Alexandria's library and Hypatia's century. The window's first pane is method. Without proof, later revolutions have nothing to revise.
The First Revolution
Geometry becomes argument. Surveyor recipes meet Greek proof. The first window opens on space as something axioms can bind.
Egyptian and Babylonian "earth measurement" becomes Greek geometry: not only survey, but the claim that nature follows rules argument can catch. The First Revolution is proof itself. Place begins as our place, Earth; abstraction turns rods and ropes into lines and circles that exist for reason. Everything later in the book (coordinates, curvature, space-time, strings) presupposes that mutiny against mere recipe.
PROOF-NOT-RECIPE. The first revolution is method: why must it be true, not only how to compute.
Take one formula you use (interest, area, hashing). Write the assumptions that make it necessary, not merely usual. That is the Greek demand in miniature.
Treating proof culture as morally superior to craft culture sneers at engineers who keep cities alive. The claim is narrower: without proof, theory plateaus.
The geometry of taxation
The roots lie in Babylon and Egypt, whose builders "achieved purpose without ever achieving understanding." Counting is old: the Ishango bone, found by Lake Edward in what is now the Democratic Republic of Congo, carries three columns of notches and is probably the earliest known numerical record. Arithmetic needed more abstraction. Anthropologists report tribes that used a different word for "two" depending on whether it counted arrows, gazelles or hernias. Seeing that these are one concept, the abstract number 2, took thousands of years.
Egypt's driver was the Nile. Each June the river flooded for four months and laid down the black mud that gave the country its Coptic name. Settled farming brought wealth, and wealth brought taxes, assessed on the height of the flood and the area of each holding. (Borrowing was simple: 100 percent interest a year.) So the Egyptians worked out areas of squares, rectangles and trapezoids, and approximated a circle by a square with sides eight-ninths of the diameter, equivalent to taking pi as 256/81, about 3.16, only 0.6 percent high.
They built well with it. The Great Pyramid (about 2580 B.C.) rose some 480 feet from blocks of over 2 tons, laid out with wood and rope; an error of a degree at the base would have produced a four-pointed spike instead of an apex. Their surveyors, the harpedonaptae or "rope stretchers", used knotted ropes to form triangles of chosen sides, and hence chosen angles: knots at 30, 40 and 50 yards give a right angle. Mlodinow points out that the rope stretchers were really laying out geodesics on the earth's surface, and that shrinking the method to infinitesimal size is exactly how differential geometry probes curved space. The test is the Pythagorean theorem: whether it holds is the test of whether space is flat.
The Babylonians, unified under Hammurabi between 2000 and 1700 B.C., were more sophisticated. Their engineers costed a canal by computing the volume of a trapezoidal trench and dividing by a man's daily digging; their lenders computed compound interest. A few hundred clay tablets from Nippur and Kish show that they wrote everything as word problems ("four is the length and five is the diagonal, what is the breadth?"). Prose cannot be manipulated like an equation; the plus sign first appears in a German manuscript of 1481.
The tablets list Pythagorean triples as large as 3456, 3367, 4825, far too many to find by trial, so the scribes knew some number theory. They computed the diagonal of the unit square as 1.4142129, accurate to three sexagesimal places. Yet neither civilisation asked why the relation holds or whether it is exact. They were like field biologists cataloguing species, not geneticists seeking mechanism. Whether the diagonal of a square has an exact value did not bother them at all. To the Greeks it would be a religious crisis.
The most famous triples tablet, Plimpton 322, was reinterpreted in 2017 by Daniel Mansfield and Norman Wildberger (Historia Mathematica) as an "exact sexagesimal" trigonometric table built from ratios, not angles, more than a thousand years before Hipparchus. Historians, notably Eleanor Robson, reject any trigonometric reading because Old Babylonian mathematics had no concept of measured angle; she reads the tablet as a teacher's table built from reciprocal pairs. Scientific American's math blog called the "most accurate trig table" publicity hype, since the tablet contains known errors. As of 2026 it remains a live dispute, and a good example of reading modern ideas into old artefacts.
After the Nile flood erased field boundaries, harpedonaptai (rope-stretchers) re-established plots for taxation. The Rhind Papyrus (copied by Ahmes around 1650 BC from an older text) gives recipes: area of a circle as (8/9 of diameter)², which implies π ≈ 3.16; slopes of pyramids; volumes of truncated pyramids. These are worked examples, not theorems. A wrong rope length meant a wrong tax bill.
Babylonian tablets (Plimpton 322 and kin) list Pythagorean triples a millennium before Pythagoras. The triples are computational, not philosophical. Mlodinow's point is not that Greece invented measuring; it is that Greece invented demanding why the measure must be true. The tax office needed accuracy; the Academy needed necessity.
ROPE-TAX-π. Egyptian surveyors, tax after Nile flood, and a π that was good enough. Geometry begins as land law, not pure form.
Pace a rectangular room and estimate area three ways: length×width, a diagonal check with Pythagoras, and a paced perimeter / 4 squared. Note the error. That is surveyor geometry under flood conditions.
Romanticising Egyptian geometry as proto-Euclid hides the leap: rules of thumb plus tables are not axiom systems. Treating every practical craft as embryonic science flattens the later Greek invention of proof.
Nile, ropes, and triples (source depth)
Floods erase boundaries; harpedonaptae (rope-stretchers) re-measure fields for tax. Knots at 3-4-5 (or 30-40-50) yards force a right angle: Pythagorean theorem as bureaucracy. Babylonian tablets (including debates around Plimpton 322) store triples long before Greek proof. Recipe precedes reason; Greece will demand why.
Any enterprise metric (latency SLO, capital ratio) begins as a rope: a practice that works.
Thales, the first scientist
About 2,500 years ago, in the Ionian port of Miletus (on a gulf now silted up, in modern Turkey), a wealthy merchant named Thales (born around 640 B.C.) retired to study and travel. His era was the world's waking moment: the Buddha in India, Laozi and Confucius in China, and the start of Greece's Golden Age. Greek culture prized talk and inquiry; at the symposium, failing at a riddle could mean dancing naked round the room.
Thales learned Babylonian astronomy and is credited with predicting the eclipse of 585 B.C., which, Herodotus says, stopped a battle. In Egypt he stunned his hosts by measuring the height of a pyramid with similar triangles, and later the distance of a ship at sea. What mattered was method. Because he extracted the abstract principle from the practical case, he could derive one technique from another and reuse solutions.
His firsts were large:
- First proofs of geometric theorems of the kind Euclid would collect.
- First system of logical rules for what follows from what.
- Congruence: two figures are equal if you can slide and rotate one onto the other. This quietly assumes homogeneity, that figures do not warp or resize as they move, which is not true in every space, including our own.
- Nature follows regular laws: thunder is not Zeus, and claims should be checked by reason, not guessed.
- All matter is one stuff, an astonishing intuition, though he guessed water. (His student Anaximander, with similar daring, proposed that humans evolved from fish.)
In old age Thales met a long-haired youth from Samos in trousers, Pythagoras, apologised for his dimmed mind, and gave him one piece of advice: go to Egypt.
Thales of Miletus (c. 624-546 BC), one of the Seven Sages, is credited by later Greeks with predicting a solar eclipse (585 BC, likely the eclipse Herodotus links to a battle pause), measuring pyramid height by shadows, and proving that a diameter subtends a right angle in a semicircle. Whether each deed is historical matters less than the method legend: explain nature without gods as causes, and prove geometric claims from more basic ones.
Miletus sat on trade routes. Geometry, astronomy and commerce mixed. Thales allegedly cornered the olive-press market after a weather forecast; the anecdote exists to say that abstract knowledge can pay. For Mlodinow, Thales opens the first window: proof and natural explanation begin as the same habit of mind.
Miletus, by the Meander's swamp, hosts a mutiny against myth in the seventh and sixth centuries BC. Thales earns the retrospective title of first scientist/mathematician: eclipse prediction stories, water as principle, similar-triangle height measures. Sources are late and biased (Aristotle, Plato); persecution and exile stories attach to the type.
The durable claim is methodological: nature has regularities that argument can track without gods as middle terms. Geometry begins as earth-measurement and becomes proof. Mlodinow needs Thales as the first pane: without the mutiny, Euclid has no tradition to codify.
SHADOW-HEIGHT. Thales measures a pyramid by its shadow when stick equals shadow. Method over myth: similar triangles as a tool.
At a time when your shadow equals your height, measure a building's shadow and claim its height. Write the similarity assumption you used. That is Thales's pyramid trick without Egypt.
Calling Thales the first scientist is a Western origin myth. Mesopotamian and Egyptian calculators already predicted eclipses and flooded fields. The useful claim is narrower: Thales is among the first named figures who publicly argued from general principles rather than priestly recipe.
Miletus mutiny (source depth)
Seventh/sixth century BC Miletus: Thales as retrospective first scientist. Eclipse stories, water-as-principle, similar-triangle height measures. Sources late (Aristotle, Plato). Method claim survives biography fog: nature has regularities argument can track without gods as middle terms.
"Were it not for number and its nature, nothing that exists would be clear to anybody." (Pythagorean disciple, later chapter's rhyme)
Pythagoras and the secret society
In Egypt Pythagoras found no poetry. A line was a rope or the edge of a field. The Greek step, which Mlodinow sees as the romance of mathematics, was that space can be an abstraction, and one abstraction can describe many situations: the same line is the edge of a pyramid, a field boundary or the path a crow flies, so knowledge of one transfers to the others.
Legend says the ring of hammers in a smithy led him to experiment with strings and find that pitch depends on string length in simple ratios (half the length, double the pitch). It is often called the first empirical discovery of a natural law stated in mathematics, a milestone the author compares to the invention of language.
The Pythagoreans laid out numbers as pebbles. Square numbers (4, 9, 16) are sums of consecutive odd numbers; triangular numbers (3, 6, 10) are sums of consecutive integers; two neighbouring triangular numbers add to a square. They coined the word mathematics (from mathema, learning) and believed "were it not for number and its nature, nothing that exists would be clear."
Mlodinow gives the classic dissection proof, naming the two legs after his sons Alexei and Nicolai: draw a square whose side is a+b, and divide it two ways. One way leaves the square on the hypotenuse plus four leftover triangles; the other leaves the squares on the two legs plus the same four triangles. "The rest is just accounting."
Pythagoras was also a mystic and self-promoter: odd numbers masculine and even feminine, 4 meaning justice (hence a "square deal"), a golden birthmark shown as a sign of divinity. After some thirteen years in Egypt, capture by the Persians and a stay in Babylon, he founded a society at Croton in southern Italy: vegetarian, communal, dressed in white linen, believing in reincarnation, and secret. The author notes parallels between Pythagorean legend and later stories about Jesus (virgin mother, walking on water, return from the dead).
The great secret was the diagonal of the unit square. The Pythagoreans proved that no whole number or fraction equals it: √2 is irrational. For a creed of "number is everything" this was a disaster. Had Pythagoras simply named the length and called it a new kind of number, the author argues, he could have advanced the real numbers and the number line by centuries and anticipated Descartes. Instead he declared such lengths alogon, "not a ratio" and also "not to be spoken." By legend, Hippasus revealed it and was killed. Resistance to strange numbers lasted: in the nineteenth century Leopold Kronecker fought Georg Cantor's work on the real numbers and helped drive him to breakdown. Pythagoras himself was caught in Croton's politics and fled an attack around 500 B.C.
When a fact breaks your system (√2 for the Pythagoreans), you can ban it, explain it away, or name it and extend the system. The third choice produced the real numbers. Most institutional failures in this book come from the first two.
The Pythagorean brotherhood at Croton treated number as the substance of the world: "all is number." Integer ratios explained harmonies (octave 2:1, fifth 3:2). Right triangles with integer sides (3-4-5 and kin) looked like cosmic confirmation. Then the diagonal of the unit square refused to be a ratio of integers. The discovery of incommensurability cracked the creed.
Whether a member was drowned for leaking √2 is folklore. What is solid is the mathematical fork: either abandon "everything is ratio," or enlarge what "number" means. Greek mathematics chose geometry over a full arithmetic of irrationals; the cost was centuries before a comfortable real-number continuum. The secret society is Mlodinow's emblem for a deeper truth: geometry can force a worldview to change.
Croton's brotherhood mixed mathematics, diet and secrecy. Integer patterns (square and triangular numbers, pebbles in geometric arrays) look like nature's password. The discovery that the unit square's diagonal is irrational (not a ratio of integers) is a philosophical wound: the diagonal exists as length but not as "number" in the Pythagorean sense.
Legend dresses the wound as scandal; historians debate details. What matters for Euclid's window: the tension between geometric existence and numerical expressibility. That tension returns when Descartes fuses them and when modern physics asks which mathematical structures nature actually uses.
TRIPLE-CULT-√2. Integer triples as sacred; irrational √2 as scandal; silence as discipline. Math as membership.
Construct √2 on paper: unit square, diagonal. Argue in three sentences why no integers m,n satisfy m² = 2n². That is the crisis the cult supposedly buried.
The romantic story of a drowned Hippasus for revealing irrationals is late and thin. Using it as fact teaches conspiracy, not history. Keep the mathematical crisis; drop the murder mystery unless a source names it.
Hammers, pebbles, Alexei and Nicolai (source depth)
Legend: hammers in a smithy lead to string experiments; pitch scales with length (half length, double pitch). Square pebble numbers equal sums of consecutive odds; triangular numbers sum consecutive integers; neighbouring triangular numbers make a square. Geometric proof of the Pythagorean theorem builds a square on side Alexei+Nicolai, connects division points two ways, discards four leftover triangles, and equates remaining areas: Hypotenuse² = Alexei² + Nicolai².
Numerology mixed with insight: odd/even as masculine/feminine; 4 as justice ("square deal"). Egypt: hieroglyphs, priestly initiation, ≥13 years; Persian capture to Babylon; return to Samos at fifty. Croton brotherhood; politics over Sybarite exiles; attack ~500 BC; later slaughter ~460 BC; Roman revival; Justinian closes academy. Incommensurable √2 wounds "all is number."
Euclid's manifesto
Around 300 B.C. in Alexandria lived a man about whom almost nothing is known, except that he ran a school, had brilliant students, and wrote the Elements: thirteen scrolls, none surviving in original form, whose first books probably drew on an earlier Elements by Hippocrates of Chios. Euclid claimed no originality. His achievement was organisation: the first complete account of two-dimensional space by pure thought. Spinoza imitated it, Lincoln studied it, Kant defended it.
- Define every term so everyone means the same thing.
- State axioms (postulates) explicitly, so no unstated assumption sneaks in.
- Derive consequences using only logic, the axioms and earlier theorems.
Why so picky? Because mathematics is a tower in which one false brick spoils everything: from a single false statement you can prove anything. Asked to prove he was the pope given 1 = 2, Bertrand Russell replied, "The pope and I are two, therefore the pope and I are one."
Intuition must be checked at the door. A rope around the equator raised one foot needs only 2π feet (about 6 feet 3 inches) more, the same for the earth as for the sun. In the Monty Hall problem, switching curtains improves your odds, which Bayes' theorem confirms. Exactness matters too: measurement could never have revealed that √2 is irrational. Paul Curry's trick cuts a 7-by-7 square into pieces that reassemble into the same square with a hole, the missing area hidden in a 2 percent overlap. In physics, Mercury's perihelion drifts 38 arc seconds a century beyond Newton's prediction (Le Verrier, 1859), a tiny error that Einstein's curved space explained in 1915, and that brought down classical physics.
23 definitions, 5 postulates, 5 common notions (things equal to the same thing are equal to each other; the whole is greater than the part), and 465 theorems, essentially all the geometry of the day.
Some definitions are weak ("a point is that which has no part"). The five postulates carry the geometry:
- A straight segment can join any two points.
- Any segment can be extended indefinitely.
- A circle can be drawn with any centre and radius (lengths do not change as you move them).
- All right angles are equal (a line looks the same along its length, a kind of straightness and homogeneity).
- The parallel postulate: if a line crossing two others makes interior angles on one side summing to less than two right angles, the two lines meet on that side.
The fifth is Euclid's own, long-winded, and he avoided using it. Its clearest equivalent form (Playfair's axiom) says: through a point not on a line there is exactly one parallel. It could fail in two ways: no parallels, or more than one. Whether it describes our world, and whether a geometry that denies it is consistent, would later overturn both our view of the universe and of mathematics. For 2,000 years, few ideas were more universally accepted.
Euclid's Elements (c. 300 BC, Alexandria) organises plane and solid geometry, proportion and number into a chain from definitions and postulates to propositions. The famous fifth postulate: if a transversal falls on two lines so that interior angles on one side sum to less than two rights, the lines meet on that side. Equivalent forms (Playfair: through a point not on a line, exactly one parallel) came later.
For two millennia the Elements was the model of certainty. Newton framed the Principia in geometric style; Lincoln kept a copy for mental training. Mlodinow's claim: the manifesto's power is not that every line is flawless, but that truth in space was made to answer to argument. The later revolutions are lawsuits against this constitution, not arson.
The Elements is a software release that froze for two millennia: definitions, common notions, five postulates, then a tower of propositions. The parallel postulate sits oddly long among short companions; that oddity becomes history's engine. Euclid's method (assume, deduce, forbid hidden appeals to the diagram) is the real manifesto.
Later centuries find cracks: unmarked assumptions about betweenness and continuity; artificial split of postulates from common notions; missing undefined terms. Hilbert and others rebuild. The manifesto still stands as the template of proof culture: make the hidden rules explicit.
23-5-POSTULATES. Definitions, five postulates (especially the parallel one), propositions that follow. The Elements as operating system.
State Euclid's fifth postulate in your own words, then invent one alternative (many parallels, or none). Sketch what a triangle's angle sum becomes. You have just opened Gauss's door early.
Worshipping the Elements as perfect hides gaps: common notions smuggle intuition; diagrams do silent work; the parallel postulate is an empirical hostage. Hilbert's 1899 axioms repair what Euclid assumed by eye.
Elements as frozen software (source depth)
13 scrolls; 23 definitions ("a point is that which has no part"); 5 postulates (fifth oddly long); 5 common notions; ~465 propositions. Parallel postulate's oddity becomes history's engine. Method: assume, deduce, forbid hidden diagram appeals. Hilbert will rebuild; the manifesto's durable demand is make the hidden rules explicit.
| Layer | Count (traditional) |
|---|---|
| Definitions | 23 |
| Postulates | 5 |
| Common notions | 5 |
| Propositions | ~465 |
A library, a beautiful woman and the end of civilisation
Philip II of Macedon conquered Greece; his son Alexander spread Greek culture to India and in 332 B.C. began Alexandria, laid out on a grid, an accidental anticipation of coordinates. Under Ptolemy I and II it gained the Museum (the first state-run research institute) and the Library, stocked by aggressive means (Ptolemy III borrowed other kingdoms' books and kept them) to an estimated 200,000 to 500,000 scrolls.
The Alexandrians measured the world:
| Scholar | Achievement |
|---|---|
| Eratosthenes | No shadow at noon at Syene on the solstice; at Alexandria the sun is 1/50 of a circle off vertical; 500 miles x 50 = 25,000 miles, within about 4 percent |
| Aristarchus | Size and distance of the moon by trigonometry |
| Archimedes | Lever, buoyancy, a near-calculus; sphere in cylinder = 2/3 of its volume, carved on his tomb; killed by a Roman soldier at 75 |
| Hipparchus, Ptolemy | Compound circular orbits predicting eclipses; the Almagest; Geographia, where mapmaking meets the fact that no flat map preserves both area and angle |
Mlodinow could only guess at Archimedes' methods. The Archimedes Palimpsest, a prayer book written over a tenth-century copy of his works, was imaged with multispectral light and X-ray fluorescence after its sale at auction in 1998. In 2001 Reviel Netz and colleagues read Method Proposition 14 anew and found Archimedes reasoning about the "multitude" of infinitely many lines and triangles as equal, an explicit use of actual infinity that pushes the idea back almost 2,000 years before Cantor.
Then decline. Rome absorbed the Greek legacy but in some 1,100 years produced no recorded theorem. Cicero boasted that Romans limited geometry to "measuring and counting." Boethius, who abridged Euclid into a theorem list without proofs ("Euclid for Dummies"), was beheaded for politics in 524. A merchant's Topographia Christiana taught a flat, rectangular earth and stayed popular until the twelfth century.
The last famous scholar of Alexandria was Hypatia, Theon's daughter, who outshone her father, wrote commentaries on Diophantus and Apollonius, and lectured on Plato and Aristotle to large crowds. Her friendship with the prefect Orestes placed her in his feud with Archbishop Cyril, who spread rumours that she was a witch. In Lent 415 a mob of monks dragged her from her chariot and murdered her. In the book's telling, her death marks the death of Greek culture and the slide into the Dark Ages, while Islamic civilisation flourished and preserved the texts.
Mlodinow's account follows the romantic tradition. Historian Edward Watts (Hypatia, 2017) argues the killing was mob violence in a political struggle between Cyril and Orestes, not a planned attack on science, and that her importance came from her civic role. She was probably around 60, not the young beauty of legend. The great Library had already declined long before her lifetime; the Serapeum, a daughter institution, was destroyed in 391. Her death was a symbol of decline, not its cause.
- Geometry began as tax accounting and became abstraction when the Greeks asked why, not just how much.
- Proof needs explicit definitions and axioms; a single hidden false assumption ruins the system.
- Intuition and approximation can mislead; small deviations (Curry's square, Mercury's orbit) can hide large truths.
- The parallel postulate was the one assumption that felt less obvious, and it became the fault line of Part III.
- Knowledge is fragile: it depends on institutions, copying and political protection.
Alexandria's Mouseion and Library gathered texts on a scale the ancient Mediterranean had not seen. Euclid taught in that ecosystem; later, Eratosthenes measured Earth's circumference; Apollonius wrote on conics; Ptolemy built the Almagest. The library's destruction was plural and gradual. The emblematic late victim is Hypatia (died 415), mathematician and Neoplatonist, murdered in a political-religious riot under Cyril of Alexandria.
Greek geometry then travelled: Byzantine copies, Syriac and Arabic translations, Latin recoveries in medieval Europe. Without that relay, Descartes and Gauss inherit almost nothing. Mlodinow closes Part I on a warning that fits every archive age: proof culture is fragile when cities burn.
Alexander founds Alexandria (332 BC) as a planned capital; he dies before it matures; the Ptolemies finish the vision. Their book-collecting methods include coercion: seventy scholars locked to produce the Greek Septuagint (Hebrew Bible); ships searched for scrolls to copy. The Mouseion and Library make the city the ancient world's graduate school. Eratosthenes, chief librarian, measures Earth's circumference from solstice shadows at Syene and Alexandria, a geometric feat without circumnavigation.
The end is plural: fires, neglect, politics. Hypatia's murder in 415 is the emblematic late atrocity, not a single off-switch. Greek mathematical culture then lives in Byzantine copies and Arabic translation until Latin Europe relearns it. The chapter's fear is archival: proof culture is only as durable as its copies.
LIBRARY-HYPATIA-DARK. Alexandria's archive, Hypatia's murder (415), knowledge loss as politics. Geometry survives in fragments and Arabic.
List three books you rely on that exist because someone copied them across a collapse. Name the fragile link (scribe, translator, monastery, hard drive). That is the Alexandria problem in your stack.
The tidy story that Christian mobs alone ended Greek science is oversimplified. Losses stacked: Caesar's fire, neglect, conquest, paper decay, selection. Hypatia's death is real and ugly; it is not the sole off-switch of antiquity.
Alexandria to Hypatia (source depth)
Alexandria's library and museum concentrate scrolls; copy or lose. Hypatia (d. 415) stands at the end of the line in Mlodinow's telling: philosopher-mathematician murdered in Christian-political violence. Boethius later beheaded (524); Euclid without full proofs circulates. The window dims until later recoveries. Civilisation-end rhetoric is narrative heat; the operational moral is institutional: no copyists, no theorems.
Hypatia: public mathematical figure; death marks Alexandrian rupture in the book's arc. Boethius: thin Latin bridge; Euclid without the full proof culture.
Part II. Descartes: The Story of Place
How do you know where you are? Naming a spot ("Kalamazoo is in square F3") is only a dictionary. The real power comes from a grammar: a way to relate locations, paths and shapes to one another and to manipulate them with equations. That grammar is the unification of geometry and algebra, and it made calculus, Newtonian physics and modern engineering possible. The revolution trailed the last great Greek geometers by a dozen centuries.
Part II turns geometry into a language of location: grids on Earth, graphs on paper, algebra on the plane. Rome's neglect, Arabic and Latin relays, Descartes the soldier and Christina's fatal invitation fill the human plot. Place becomes coordinate; science gains a universal clipboard.
The Revolution in Place
Where becomes writable. Latitude, longitude and finally algebraic coordinates turn place into grammar, not a gazetteer.
Location becomes number. Latitude, longitude, graphs and finally orthogonal axes let "where" be computed, not only named. Kepler and Galileo lack the full coordinate toolkit later students take for granted; the counterfactual is the chapter's prod. Place, once quantified, can be transformed, differentiated and curved. Descartes is the hinge; the Revolution in Place is the door he opens.
ADDRESS-PLUS-GRAMMAR. A pin on a map is a noun; coordinates are verbs for paths and shapes.
Name a place three ways: story ("by the old mill"), grid (lat/long), equation (x²+y²=R² as a city wall). Note which tasks each form unlocks.
Reducing every human place to coordinates erases meaning. The revolution adds a grammar; it does not obsolete names.
Latitude and longitude
The earliest known maps, clay tablets from about 2300 B.C., recorded property tax, the same motive that produced Egyptian geometry. Later, navigation made maps a matter of life and death. As late as 1915, when Shackleton's Endurance was crushed in Antarctic ice, the worst danger was not knowing the way back.
Two coordinates fix a place on a sphere:
- Latitude is set by the angle between the earth's axis and the line from the earth's centre to you. Aristotle first divided the globe into climate zones. In practice you measure the angle between the north horizon and Polaris, which happens to sit near the axis. It has not always: the axis precesses in a cone every 26,000 years, so the Egyptian pyramid-builders aligned passages to Alpha Draconis, and in about 10,000 years the pole star will be Vega.
- Longitude is harder, because the sky rotates 15 degrees an hour. Two star-snapshots compare usefully only if taken at the same time, so longitude needs a clock. An error of 3 seconds a day over a six-week voyage gives more than half a degree of error. Clocks that survived the sea came only in the eighteenth century, and the Greenwich prime meridian was agreed in October 1884.
The book says Newton invented the sextant in 1700. Newton did design a double-mirror reflecting quadrant in 1699, but his manuscript was published only in 1742. The working instrument came from John Hadley, who presented his reflecting octant to the Royal Society in 1731 (Thomas Godfrey built one independently in 1730); widening its arc gave the sextant. A small slip, but a reminder that a good story often collapses a long chain of inventors into one famous name.
The Greeks advanced mapmaking fast: Anaximander's world map (about 550 B.C.), relief maps on coins, the Pythagorean spherical earth, Hipparchus's evenly spaced grid, and Ptolemy's Geographia, which assigned latitude and longitude to some 8,000 places. Then, like geometry, cartography stalled under Rome.
Latitude is readable from the altitude of the pole star or noon sun. Longitude is a time difference: Earth turns 15° per hour, so a reliable clock comparing local noon to a reference meridian solves position. Spain and later Britain posted prizes. John Harrison's marine chronometers (H1-H4, mid-18th century) won in practice; lunar-distance methods competed. The 1884 International Meridian Conference fixed Greenwich as prime meridian.
Mlodinow's bridge into Descartes: naming a place on a globe is still a dictionary entry until you have a grammar relating places. Latitude and longitude are that grammar's first global dialect. Shackleton's Endurance, crushed in 1915 ice, still lived in a world where losing chronometers meant losing the map.
Naming a spot is not locating it. Latitude (from equator along a meridian) and longitude (from an agreed prime meridian) turn the globe into a coordinate grid. Ancient and early-modern astronomy fought longitude at sea; clocks and lunar distances were engineering answers to a geometric question.
The chapter's alien-emissary joke asks what a dictionary buys you without a shared frame. Place becomes numerical. That habit, more than any single theorem, prepares analytic geometry: once Earth is a grid, the plane can be a grid too.
LAT-EASY / LONG-HARD. Latitude from stars; longitude needs time. Harrison's chronometer beats lunar distance for empire.
With only a watch set to your home noon and a stick, estimate longitude at a trip destination: local noon minus home noon, times 15 degrees per hour. Note the error sources. That is the longitude problem in a parking lot.
Crediting Newton with inventing the sextant collapses Hadley (1731) and Godfrey (1730). Hero stories that fuse inventors teach the wrong lesson about instruments: empires fund precision, they do not summon it from one genius.
Naming is not locating (source depth)
Latitude from the equator, longitude from an agreed prime meridian: Earth becomes a grid. Sea longitude demanded clocks and lunar distances. Alien-emissary joke: a dictionary without a shared frame buys nothing. Once Earth is numerical, the plane can be numerical too. Coordinate habit precedes Descartes's formal marriage of algebra and geometry.
The legacy of the rotten Romans
The revival began with Charlemagne, six foot four, illiterate despite trying, and a patron of learning on Alexander's scale. His Palace School at Aachen and the cathedral and monastery schools he fostered trained a literate clergy, copied manuscripts in the new Carolingian minuscule (still the basis of our lower-case letters), and eventually grew into universities, starting with Bologna in 1088. (For his health, Charlemagne kept 300 monks and 100 clerks praying in shifts round the clock. He died anyway, in 814.)
Contact with the Islamic world, through trade and the Crusades, brought back what Europe had lost. Arab scholars had kept faithful copies of Euclid and Ptolemy, developed all six trigonometric functions and perfected the astrolabe. Adelard of Bath, disguised as a Muslim student, translated the Elements into Latin, this time with proofs; Fibonacci brought zero and Hindu-Arabic numerals from North Africa.
Medieval scholarly life was grim:
| Obstacle | Example from the book |
|---|---|
| Climate and plague | The Little Ice Age, famine, and the Black Death from 1347, killing up to half of Europe by 1351; chronicler Villani left a blank for the plague's end date and died of it in 1348 |
| Universities | No buildings; at Bologna students hired, fined and fired professors; Leipzig had to ban throwing stones at them |
| Time and measure | Hours varied with the season (38 to 82 modern minutes in London); the first clock striking equal hours came in Milan in the 1330s; ratios were not treated as numbers |
| Authority | The Bible had to be taken as literally true |
The Scholastics tried to reconcile reason and faith, and most paid for it. Abelard argued for logical debate; he was excommunicated, his books burned (and he was castrated by his lover's uncle). Aquinas started from faith and proved it, and escaped unscathed. Roger Bacon prized experiment and spent fourteen years in prison. William of Occam gave us the razor: build theories on as few ad hoc assumptions as possible. Mlodinow notes that string theory's ambition is exactly this, to derive the electron's charge, the particle list and even the number of dimensions instead of assuming them.
Rome built roads, aqueducts and law across a continent while adding little to pure geometry. Greek mathematical culture thinned in the Latin West; it persisted longer in Byzantium and flourished in the Islamic world (al-Khwārizmī, Omar Khayyam, translations of Euclid and Ptolemy). European recovery came through Cordoba, Toledo and Sicily as much as through Rome's ghosts.
Mlodinow's tone is impatient because his story needs the next leap: without a rebirth of abstract method, coordinate geometry and calculus do not arrive. The "rotten" label is rhetorical. The structural claim stands: engineering culture without proof culture plateaus.
Rome's practical genius (roads, law, concrete) did not extend to a thriving pure-geometry culture. After the western collapse, literacy's floor was held up unevenly. Charlemagne knocked down neighbours' borders and raised church schools; his revival produced little original mathematics, yet literacy never again sank to the worst pre-Carolingian depths. The real recovery rode contact: trade, travel, Crusades (violent "contact"), Byzantium and the Islamic Mediterranean.
Twelfth-century hunters such as Adelard of Bath, travelling in disguise, brought Euclid and Arabic science into Latin. Toledo and Sicily were translation hubs. An English monk could already compare Paris to ancient Athens as a mother of arts. The stage for a new analytic geometry was not Roman virtue but relayed Greek and Arabic books. Mlodinow's impatience with "rotten Romans" is a narrative prod: without that relay, Descartes has nothing to algebraise.
ROME-USES / GREECE-PROVES. Aqueducts without axioms; Greek math hibernates; medieval Europe inherits fragments.
Walk past an engineered object (bridge, pipe, road). Write whether its builders needed Euclid or only rules of thumb. Then ask what would break if a deeper theory (stress tensors) were banned. That is the Roman bargain.
Calling Romans rotten at pure math can slide into civilisational sneer. They excelled at law, logistics and concrete. The loss was specific: little taste for abstract proof culture, not general stupidity.
Copy chains and dark-age bottlenecks (source depth)
Rome valued roads and law more than Greek proof culture; manuscripts survive through Byzantine and Islamic copyists, then Latin translators. Lose the copy chain and the window closes. Boethius transmits thin Euclid; full recovery waits on Arabic-Latin relays. "Rotten" is Mlodinow's provocation about neglect of abstract geometry, not a claim that Romans built nothing.
Civilisational decline narratives oversimplify. Engineering competence and theoretical geometry can diverge for centuries without "darkness" everywhere.
The discreet charm of the graph
Today an Avon executive reads an upward line on a profit chart without thinking. A medieval merchant would have been baffled. Graphs were the next step to a theory of place, and they needed the number line, which in turn needs the irrational numbers that Pythagoras had banned. Without them the line has infinitely many holes. Mathematicians used irrationals for centuries before Cantor and Dedekind put them on firm ground in the late nineteenth century.
"Illegal" mathematics is common in physics. Dirac's delta function (zero everywhere except one point, where it is infinite, yet integrating to one) was nonsense by the rules of its day until Laurent Schwartz rewrote the rules and founded the theory of distributions. The book notes that the quantum field theories of modern physics have still not been shown to exist rigorously. Tools that work often precede the theory that justifies them.
Around 1350, Nicole Oresme, later bishop of Lisieux, invented the graph without worrying about the gaps. A graph pictures a function: how one quantity varies with another. Every map is a graph (names or elevation against location), and a weather map turns a function into a surface whose shape can be studied. That is the deep point: types of function correspond to types of shape, so geometry and number merge. Our eyes pick out lines and circles instantly, which is why graphs reveal patterns that tables hide, the theme of Edward Tufte's classic on visual display.
Mlodinow lists three noisy data sets. As numbers they look dull. Plotted, Alexei's data form a straight line (with one outlier, "a sneeze"), Nicolai's a parabola, and Mom's a quarter circle. The parabola turns out to be New York's average monthly high temperatures, which yield a "law": degrees below 85 = 2 x (months from mid-July)². For October 15 it predicts 67; the actual average is 66.
Oresme used graphs to prove the Merton rule, proposed at Merton College, Oxford, between 1325 and 1359: a body accelerating uniformly from rest covers the same distance as one moving for the same time at half the final speed.
Distance is the area under the speed curve; the triangle and the half-height rectangle have equal area. The same picture shows that distance under uniform acceleration grows with the square of time, a law usually credited to Galileo. Oresme also anticipated relativity of motion: against his teacher Buridan, who said a rotating earth would make arrows land elsewhere, he noted that a sailor sliding his hand down a mast sees vertical motion while a watcher on shore sees a diagonal, and that motion can only be judged relative to something else. He even considered a rotating earth circling the sun, then retreated, quoting Psalm 93: the world "shall not be moved." He was rewarded with a bishopric in 1377. Galileo was his intellectual heir, though there is no evidence he read him.
A Cartesian graph (named later; invented in the 1630s climate of Descartes and Fermat) lets each point on a plane wear an address (x, y). Lines, circles and conics become equations. Intersection becomes simultaneous solution. Tangents invite calculus. The charm is discreet because it looks like school graph paper; it is actually the merger that lets Newton and Maxwell write laws as equations about space.
Oresme, in the 14th century, already related magnitudes with proto-graphs. Descartes and Fermat make the method systematic and tied to algebra. Mlodinow's line: once place has coordinates, physics can speak in formulas about where.
Medieval merchants had Roman numerals and blank stares for a coloured time series. Graphs that today compare domestic and international lotion sales rest on a conceptual marriage the Greeks partly refused: numbers and geometry as one object. The number line, negative numbers and a plane of pairs (x, y) are late cultural tools.
Coordinate thinking lets functions become curves and curves become equations. Mlodinow's "discreet charm" is the quiet power of a picture that is also a calculation. Without it, Descartes has nowhere to land, and Gauss has no charts to least-square. The graph is the everyday fossil of the Revolution in Place.
GRID-MARRIES-NUMBER. x meets y; curves become equations; geometry gains algebra's verbs.
Plot y = x² for x from −3 to 3 on scrap paper. Circle the geometric facts (symmetry, vertex, steepening) that the equation makes trivial. That is Descartes's charm in five minutes.
Treating the graph as mere visualisation misses the revolution: the graph is a dictionary between shape and formula. Software charts that never become equations reverse the gift.
From blank stares to time series (source depth)
Medieval merchants with Roman numerals could not read a coloured sales chart. Number line, negatives, and (x,y) pairs are late tools. Functions become curves; curves become equations. The graph is everyday fossil of the Revolution in Place: picture and calculation in one object.
A soldier's story: Descartes
René Descartes, born in 1596 and so sickly that doctors expected him to die, was allowed at the Jesuit college of La Flèche to stay in bed until late morning, a habit he kept almost to the end. He concluded his schooling was useless or wrong, took a law degree to please his father, gambled with mathematics in Paris, and then joined, unpaid, the army of Maurice of Nassau to travel. He never saw battle, then joined the other side and did not fight for it either.
In Breda in 1618 he saw a crowd around a posted mathematical challenge, asked an older man to translate, called it easy, and solved it. The translator was Isaac Beeckman, a leading Dutch mathematician, who became his friend and "spiritual father"; to him Descartes first described his new geometry.
Descartes found Greek geometry exhausting; each proof demanded fresh ingenuity, "greatly fatiguing the imagination." His "notorious mathematical laziness" sought a system that makes proofs routine. Compare:
Euclid: "a plane figure contained by one line such that all straight lines falling upon it from one point amongst those lying within the circle are equal to one another." Descartes: all points (x, y) with x² + y² = r². The second is shorter, and, more important, it is an equation you can manipulate.
The method: draw an x-axis and a y-axis, label each point by an ordered pair (x, y), and use plus and minus signs to separate the four quadrants (like East and West 80th Street). Fermat found the same idea, but "while Descartes had the bad habit of not including citations, Fermat had the worse habit of not even publishing." Ptolemy had used coordinates; the advance was what Descartes did with them. Every line obeys ax + by + c = 0; circles, ellipses, parabolas and hyperbolas all become simple equations. Physical science, in the author's summary, is mostly this: notice regularities in data, find the functional relation, then explain it.
Applied to the Pythagorean theorem, coordinates give the distance formula: the squared distance between two points is the sum of the squares of their horizontal and vertical separations. This looks like a small corollary. It is the hinge of the whole book, because later geometers would define a space by how it measures distance between nearby points, and changing that formula gives curved space.
Descartes also gave the modern law of refraction and the first full explanation of the rainbow, and wrote that "my entire physics is nothing other than geometry." Yet he published nothing until he was forty. In 1633 the Inquisition condemned Galileo, and Descartes, though living in Protestant Holland, shelved his book. The Discourse on Method finally appeared in 1637, anonymously, with La Géométrie as an appendix. Fermat quibbled, Pascal condemned it, and theologian Voetius at Utrecht attacked his faith in reason; within a few years the geometry was taught almost everywhere anyway. His private life held one affair and a daughter, Francine, born 1635, whom he called his niece; her death from fever in 1640 devastated him.
René Descartes (1596-1650), soldier and philosopher, published Discourse on the Method with three essays including La Géométrie in 1637. He dies in Stockholm after tutoring Queen Christina in cold dawn hours ("Snow Queen" in the next chapter). His geometry shows how to solve geometric problems by algebra and how to interpret algebra as geometry.
Coordinate thinking did not appear overnight as "the Cartesian plane" textbook myth. Descartes often used a single axis and lengths; later readers standardised orthogonal x-y. The revolution remains: place became writable as number, and analysis could inherit geometry's objects.
After letters and Paris nights, Descartes joins Prince Maurice of Nassau's army unpaid for travel and adventure. In 1618 at Breda he asks a bystander to translate a posted math challenge; he solves it; the bystander is Isaac Beekman, later called the "inspiration and spiritual father" of his studies. Within months Descartes sketches to Beekman a new way of looking at geometry.
He stays harsh on Greek geometry's awkwardness: too much ingenuity wasted on figure-chasing that algebra could straighten. La Géométrie (1637) with the Discourse shows how to encode geometric problems as equations and read equations as geometry. Later pedagogy draws the orthogonal x-y plane more tidily than Descartes himself always did. The soldier's story is the hinge from compass-and-straightedge culture to coordinate analysis.
STOVE-DREAM-METHOD. A soldier in a heated room; dreams of method; Discourse and Géométrie (1637).
Write your own four-line "method" for a hard work problem (doubt, divide, order, review), then apply it for twenty minutes. Note where geometry-style clarity helped and where it lied. That is Cartesian discipline tested.
The heated-room legend invites Great Man history. Algebraic geometry also needed Fermat, Viète's notation, and a century of algebra. Descartes is a hinge, not a sole author of place.
Army, stove, and analytic geometry (source depth)
Descartes the soldier winters with a stove-heated room and a method: doubt, clear ideas, coordinates that turn curves into equations. Algebra and geometry marry; place becomes number. Without this clipboard, Gauss has no charts and Einstein has no tensor components to index.
The famous night of dreams is biography's ornament; the durable output is a coordinate method any surveyor or physicist can steal.
Iced by the Snow Queen
Queen Christina of Sweden, twenty-three, brilliant, short-haired, deep-voiced and sleeping five hours a night, collected books and their authors. Through the diplomat Pierre Chanut she lured Descartes with promises of an academy and a house in the warmest part of Sweden. He wrote his will before leaving. The winter of 1649 to 1650 was brutal, and the queen wanted her philosophy lessons at 5 a.m. "Men's thoughts freeze here in winter just like the water," he wrote. Nursing Chanut through pneumonia, he fell ill, refused a hostile doctor's bleeding, and died on February 11, 1650. The Church banned his works in 1663, which only made them more popular. His bones went back to France in 1666, minus the skull, which arrived in 1822 and sits in the Musée de l'Homme.
In 2009 German philosopher Theodor Ebert argued from the surviving medical letters (by Descartes's servant and the physician van Wullen) that the symptoms fit arsenic poisoning, and named the embassy chaplain François Viogué, who may have feared Descartes would hinder Christina's conversion to Catholicism. A 2019 paper in Early Science and Medicine restated the case. Reviewers find the poisoning reading plausible but the suspect and motive unproven; pneumonia remains the standard account.
- Naming places is a dictionary; relating them by equations is a grammar. Coordinates supplied the grammar.
- Graphs turn functions into shapes, so patterns invisible in tables become obvious.
- Useful mathematics often runs ahead of rigour (irrationals, the delta function).
- The distance formula is the seed of every later geometry, flat or curved.
- Fear of authority delayed ideas by decades: Oresme retreated, Descartes waited.
Queen Christina of Sweden invited Descartes to Stockholm. Early lessons in a freezing castle broke his lifelong habit of late rising. He died on 11 February 1650. Mlodinow uses the episode as tonal close to Part II: the man who geometrised place is defeated by a place. The intellectual legacy moves on without him: analytic geometry, then calculus, then the physics of trajectories.
Queen Christina of Sweden sleeps five hours, collects a library and summons Descartes to make a northern Athens. The winter of 1649-50 is brutal; dawn philosophy lessons kill the bed-late mathematician. Nursing Chanut through pneumonia, Descartes catches it; a hostile physician does not help. He dies 11 February 1650.
The Snow Queen episode closes Descartes's biography without closing his method. Coordinate geometry outlives the court. Mlodinow's chill is literal and symbolic: ideas migrate through fragile human hosts.
CHRISTINA-COLD-DAWN. Stockholm mornings kill the stove-dreamer; method meets monarchy's calendar.
Schedule one hard thinking block at your worst circadian hour for three days. Journal output quality. That is Christina's dawn tutorial as an experiment on whether will beats biology.
Blaming Christina alone for Descartes's death is too neat (pneumonia, February 1650, many factors). The useful image is cost: patronage can demand bodies that method cannot override.
Christina's dawn tutorials (source depth)
Queen Christina of Sweden sleeps five hours, wants a northern Athens, summons Descartes into 1649-50 winter dawn lessons. He nurses Chanut through pneumonia, catches it, dies 11 February 1650. Coordinate geometry outlives the court. Ideas migrate through fragile hosts.
1649 arrives in Stockholm 1649-50 brutal winter; early-morning philosophy 11 Feb 1650 death; method survives the monarch
Part III. Gauss and Riemann: The Story of Curved Space
Is space really the space Euclid described and Descartes quantified? For 2,000 years ("a long time to wait for version 2") mathematicians had no quarrel with the content of the parallel postulate. They only wanted it demoted from assumption to theorem. Each attempt came close to discovering strange new spaces, and each was stopped by one belief: that the postulate was a necessary truth about space. The exception was a fifteen-year-old, Carl Friedrich Gauss, who in 1792 accepted that a consistent geometry might exist without it. This was not a patch to Euclid but, in the author's words, a new operating system, and when it was installed both physics and the foundations of mathematics had to be rebuilt.
Euclid's grip was hard to overstate. The Elements was one of the first printed books and for two centuries the only Greek work printed in the original; every building, painting and equation before 1800 was Euclidean. Yet even Euclid seemed uneasy: he proved his first twenty-eight theorems without the fifth postulate.
Part III breaks the Euclidean monopoly. Surveying, hyperbolic geometry, intrinsic curvature and Riemann's manifolds replace one true space with a family. The face-lift of axioms clears the runway for physics to ask which geometry the universe runs.
The Curved Space Revolution
Euclid's fifth postulate falls. Space may be curved; geometry becomes a family of theories. Measurement, not pure reason alone, must choose.
Version 2 of Euclid waits two thousand years. Failed parallel proofs brush new spaces; Gauss at fifteen already smells optionality. Curved geometry asks a physical question Euclid's culture did not: which geometry is ours? Mathematics loses the comfort that axioms simply mirror unique space. Physics gains the stage on which Einstein will let mass speak to metric.
V-IS-OPTIONAL. Keep postulates I-IV; swap V; inherit a new universe of theorems.
Write one physical measurement that could tell Euclidean from hyperbolic space on cosmic scales. Note today's error bars in words.
Declaring "space is curved" without saying where and how much is poster science. Locally, Euclid still builds your house.
The trouble with Ptolemy: two thousand years of failed proofs
The attempts share a pattern: an innocent-looking assumption turns out to be the parallel postulate in disguise.
| Who | When | Hidden assumption |
|---|---|---|
| Ptolemy | 2nd century | Assumed an alternate form of the postulate, then derived the original (a circular argument) |
| Proclus | 5th century | Parallel lines stay a constant distance apart |
| Thabit ibn Qurrah | 9th century, Baghdad | The curve traced at a fixed distance from a line is itself a line |
| John Wallis | 1663 | Any triangle can be scaled up or down without changing its angles |
Mlodinow turns Proclus's proof into Manhattan streets. Let Fifth and Sixth Avenues be parallel, and let a building on Sixth (the book's publisher) be the external point. Any other road through the building, call it Broadway, forms a right triangle with Sixth Avenue and a cross street; walk far enough and the cross street exceeds the gap between the avenues, so Broadway must hit Fifth. It feels watertight, but it assumes the gap is a fixed number, which cannot be proved without the postulate. (It also confuses "growing" with "growing without bound": ½, ⅔, ¾ grow forever but never pass 1.)
Thabit walked along Fifth Avenue holding a one-block stick perpendicular to it and assumed the far end traced a straight line. Only in Euclidean space is the set of points equidistant from a line also a line. If it is not, a triangle distorts when you move it. Thabit took that as proof the assumption must be right, instead of following it to a new geometry.
Wallis, a Cambridge-trained cleric who broke Royalist codes for Parliament, won Oxford's Savilian chair in 1649 and gave us the ∞ symbol, proposed replacing the postulate with the existence of similar triangles. That merely swaps one axiom for another, but read backwards it is startling: if the parallel postulate fails, similar triangles do not exist. Scale models of aircraft wings, catalogue photos of clothes and architects' blueprints would all distort when enlarged. Why have we not noticed? Because non-Euclidean spaces look Euclidean in small regions, and we live in a small corner of the universe.
A deviation too small to measure in a room can be large across a planet or a galaxy. Every non-Euclidean space looks Euclidean up close. This single idea explains why Euclid lasted 2,000 years, why surveyors and astronomers were the first to find cracks, and why later physics (relativity, cosmology) is scale-dependent.
From antiquity through Islam and early modern Europe, geometers tried to deduce the fifth postulate. Proclus, Omar Khayyam, Nasir al-Din al-Tusi, Saccheri (1733), Lambert and others built carefully and always assumed an equivalent. Saccheri's Euclides Vindicatus explored the hypothesis of the obtuse and acute angle and recoiled from the latter's strange consistency.
Mlodinow's lesson: when a postulate resists proof for two thousand years, suspect it is independent, not merely unproved. That suspicion becomes Gauss, Bolyai and Lobachevsky.
Proclus and a parade of later geometers try to prove the parallel postulate from the others. Playfair's form (exactly one parallel through an external point) makes the attempts easier to state and to fail. Street-avenue analogies in the book show how easy it is to smuggle Euclidean intuitions into a "proof."
Each near-miss brushes non-Euclidean geometry and recoils. The trouble is not Ptolemy the astronomer alone; it is two thousand years of treating one axiom as a theorem waiting to happen. The face-lift later is honesty: promote the alternatives to equal axiomatic citizenship.
FIFTH-WON'T-BUDGE. Two thousand years trying to prove Euclid V from the first four; every proof smuggles an equivalent.
Take any "proof" of the parallel postulate you can find online. Underline the hidden assumption (Playfair, angle sum 180°, similarity). You have just joined Saccheri and friends.
Calling all attempts failures can sound like they wasted time. They mapped the logical neighbourhood of Euclid V and prepared non-Euclidean geometry. Failed proofs are research.
A Napoleonic hero: Gauss
Gauss died in Göttingen on February 23, 1855, his pocket watch reportedly stopping as his heart did. He left money stuffed in drawers, an unheated bedroom and a reputation, with Archimedes and Newton, as one of the three greatest mathematicians. He was also deeply unhappy.
He was born in 1777 to "half-citizens" in Brunswick: an illiterate mother who adored him and a "domineering, uncouth" father who dug canals and sent the boy to spin flax. At about three Gauss corrected his father's payroll. At school the brutal master Büttner set the class to add 1 to 100; Gauss handed in 5050 at once, presumably by pairing 1+100, 2+99 and so on. Büttner's assistant Johann Bartels, seventeen, became his study partner, and through Bartels's connections the Duke of Brunswick paid for his education. Gauss entered Göttingen in 1795.
His rebellion, in the Sturm und Drang era, was against Euclid. At twelve he criticised the Elements; by fifteen he was the first person to accept that a consistent geometry without the parallel postulate could exist. Then came his richest years, and tragedy: marriage to Johanna Osthoff in 1805 ("I had never hoped for this bliss"), the Duke's death from wounds fighting Napoleon in 1806, and Johanna's death after childbirth in 1809, followed by their baby. (Napoleon spared Göttingen because "the foremost mathematician of all time lives there.") Tragedy, he wrote, ran "through my life like a red ribbon."
Carl Friedrich Gauss (1777-1855) surveyed Hanover, proved deep theorems in number theory and magnetism, and privately developed non-Euclidean geometry. He feared "the clamour of the Boeotians" if he published. His Theorema Egregium (1827) shows Gaussian curvature is intrinsic: bend a surface without stretching and curvature stays. That result later lets relativity talk about space from inside.
Gauss dies in Göttingen, 23 February 1855. Mlodinow paints him brilliant and cold: a hero of method who would not spend social capital on the fifth postulate's fall.
Carl Friedrich Gauss, born 30 April 1777 in Braunschweig, poor "half-citizen" stock, could sum before he could politely talk. Father Gebhard's nurture was not Montessori. Stories of prodigy arithmetic sit beside a lifelong pessimism: teaching felt burdensome; without immortality the world seemed meaningless; close friends were few. He dies 23 February 1855 in Göttingen, watch reportedly stopping with his heart, money stuffed in drawers, unheated bedroom.
The mathematical haul is enormous: number theory, astronomy, magnetism, least squares, differential geometry. The Theorema Egregium (1827) proves Gaussian curvature intrinsic. Non-Euclidean geometry he kept mostly private, fearing "Boeotian" clamour. Mlodinow's Gauss is both hero and warning: the tools of curved space arrived in a man who would not spend reputation on announcing them.
PRINCE-SURVEY-CURVE. Gauss the "Prince of Mathematics"; geodesic survey; privately sure Euclid V is optional.
On a hill or stadium, measure a large triangle's angles with a phone clinometer app (roughly). Compare the sum to 180°. Write three reasons your error bars swamp curvature. That is why Gauss needed precision.
Gauss's refusal to publish non-Euclidean work looks like cowardice in modern eyes; in his context it also looks like reputation management in a culture that equated Euclid with reason. Do not flatten him into either hero or villain.
Childhood, duke, and reluctance to publish (source depth)
Gauss the prodigy (Buettner school stories), sponsored by a duke, becomes the "Napoleonic" mathematical hero of his age. He reshapes number theory, astronomy and geodesy, yet hides non-Euclidean work. Fame without full disclosure delayed the curved-space public revolution until after his death and note-mining.
The fall of the fifth postulate
At Göttingen, Abraham Kästner collected failed proofs of the postulate (his student Klügel analysed 28 of them), yet said only a madman would doubt it. Gauss kept quiet. Between 1813 and 1816 he worked out the trigonometry of what we now call hyperbolic geometry, and in 1824 wrote to the lawyer Taurinus that assuming triangle angle sums less than 180 degrees leads to a geometry "quite different from ours, which is absolutely consistent." He forbade publication.
His fear was not the Church but the philosophers. Kant had called Euclidean space "an inevitable necessity of thought," and Kant's followers dominated an intellectual culture in which table-rapping séances were respectable and doubting Euclid was not. Gauss had watched Wallis waste twenty years in a pamphlet war with Hobbes. He read the Critique of Pure Reason five times and judged its central distinction either trivial or false, but said so only to friends.
Then others published. In 1823 János Bolyai, son of Gauss's friend Wolfgang, wrote that he had "created a new, different world out of nothing." In Kazan Nikolai Lobachevsky, once tutored by Gauss's old friend Bartels, explored the same geometry. Most historians think the spirit, not the details, of Gauss's ideas travelled through those connections, and that the two men worked independently. Nobody noticed: Lobachevsky published in the obscure Kazan Messenger, Bolyai in an appendix to his father's textbook. Gauss praised Bolyai while noting he had found it all first (crushing the young man, who never published again), and quietly had Lobachevsky elected to the Göttingen Society in 1842.
The revolution came after Gauss died. His papers revealed his work and his references to the other two; in 1867 Baltzer's textbook included Bolyai and Lobachevsky. In 1868 Eugenio Beltrami proved that if Euclidean geometry is consistent, so is the new geometry, settling the 2,000-year question: the postulate cannot be proved from the others.
Nikolai Lobachevsky (Kazan) and János Bolyai (Hungarian officer, son of Farkas) publish hyperbolic geometry in the 1820s-1830s. Gauss claims prior private discovery. In hyperbolic geometry, through a point not on a line there are infinitely many parallels; triangle angle sums fall below 180°; area couples to angle defect. Euclid V is not a secret theorem; it is a choice.
The fall is conceptual: geometry splits into a family of theories. Physical space might be any of them. Measurement, not pure reason alone, must decide.
At Göttingen in 1795, Kaestner collected parallel-postulate literature; Klügel catalogued failed proofs. Between 1813 and 1816 Gauss privately develops hyperbolic trigonometry: angle sums less than 180°, a consistent geometry "quite different from ours." He writes as much to friends (Taurinus among them) yet publishes almost nothing, fearing philosophical fashion that treated Euclidean space as necessary.
Natural philosophy had not fully split from physics; intuition and faith still competed with proof. The fall is quiet: not a public manifesto but a private recognition that Euclid's fifth is optional. Lobachevsky and Bolyai later force the issue into print. The book's drama is psychological as much as logical: the hardest step was granting permission to deny the postulate.
BOLYAI-LOBACHEVSKY-GAUSS. Three discoverers; one published boldly; Euclid V is optional at last.
Write Playfair's form of the parallel postulate, then write its two negations (no parallel, many parallels). List one theorem that must change in each case. That is the fork in the road.
Priority fights (who was first) distract from the logical event: consistency of alternatives. History of grudges is not history of geometry.
Secret journals and posthumous revolution (source depth)
Gauss kept the breakthrough secret. Göttingen 1795: Kaestner collected parallel-postulate literature; student Georg Klügel catalogued twenty-eight failed proofs. Kaestner: only a crazy person would doubt the postulate. Gauss noted ideas in a journal found forty-three years after his death; later dismissed Kaestner as "leading mathematician among poets and leading poet among mathematicians."
1813-1816: hyperbolic trigonometry privately. 6 Nov 1824 letter to lawyer F. A. Taurinus: angle sum less than 180° yields a consistent geometry "quite different from ours." He would not publish, fearing "the outcry of the Boeotians." Bolyai and Lobachevsky printed; Gauss's death and note-mining spread the news. 1867 Baltzer's Elemente includes them; 1868 Beltrami proves non-Euclidean consistency if Euclid is consistent.
Mathematical correctness was not enough; philosophical fashion still treated Euclidean space as necessary. The hardest step was social permission to deny the postulate.
Lost in hyperbolic space
In hyperbolic space there are many parallels through an external point. Triangles have angle sums below 180 degrees, by an "angular defect" that grows with the triangle's size, which is why similar triangles cannot exist. The Euclidean form is "approachable but not attainable, like the speed of light, or your ideal weight."
Neither Gauss nor his rivals could picture this space. Henri Poincaré, in the 1880s, built a model. Take a disc (Mlodinow's crêpe). Points are points inside it; lines are arcs of circles that meet the rim at right angles; angles are measured between tangents; and distance is redefined so that space is compressed near the rim, making the finite disc infinite to its inhabitants. Through a point off a Poincaré-line run many Poincaré-lines that never meet it, and no rectangles exist. The punch line: all mathematical descriptions of the hyperbolic plane are isomorphic, so the model is hyperbolic space. The other violation, no parallels at all, gives elliptic space, which had been "proved" impossible within Euclid's system. That impossibility turned out to be a flaw in Euclid's system, not in elliptic space.
Poincaré could call arcs "lines" because axioms never say what a line is, only how lines relate. As long as an interpretation satisfies the axioms, every theorem holds for it. You could, in principle, call zebra stripes lines, if the axioms survived the translation.
Models (Beltrami-Klein, Poincaré disk and half-plane) embed hyperbolic geometry in Euclidean pictures so consistency tracks Euclid's (relative consistency). Distances warp toward the boundary; "straight" lines are geodesics of the model metric. For Mlodinow the emotional payload matters: once you can be lost in a rigorously defined space that violates school intuition, space is no longer a single gift from the Greeks.
Changing the parallel postulate rewrites every theorem about the shape of space. Gauss, Lobachevsky and Bolyai had the equations; visualisation lagged. Beltrami and then Poincaré supplied models: redefine "line" and "plane" as concrete curves and disks so the axioms hold by reinterpretation. Zebra-stripe jokes in the book make the point: primitives mean what the axioms force, not what English suggests.
In hyperbolic geometry, triangle angle sums fall below 180°, similarity fails in the Euclidean way and area grows with angle defect. Poincaré's disk puts infinity on a boundary circle; geodesics are arcs meeting the boundary at right angles. The chapter's payoff for later physics: curved-space intuition before Riemann's manifolds and before Einstein's gravity.
ANGLE-DEFECT-AREA. In hyperbolic space, thinner triangles make more area; parallels behave like strangers.
Play with a Poincaré disk image (or sketch one): arcs meeting the boundary at right angles as "lines." Draw a triangle and mark angle sum visually less than 180°. That is hyperbolic space you can see.
Sci-fi metaphors ("falling forever") entertain but can block the sober point: hyperbolic geometry is as axiomatically precise as Euclid; Escher prints are aids, not proofs.
Angular defect and Poincaré's disk (source depth)
Replace "exactly one parallel" with "many parallels." Consequences: triangle angle sum always under 180° by an angular defect that grows with area; similar triangles in the Euclidean sense fail. Small triangles look nearly Euclidean (like approaching c or ideal weight).
Beltrami then Henri Poincaré (cousin of president Raymond) supply models: redefine line and plane as concrete curves so axioms hold by interpretation. Poincaré disk: infinity on the boundary circle; geodesics meet the boundary at right angles. All hyperbolic-plane models are isomorphic. Later, elliptic space (no parallels; angle sums over 180°) appears; Euclid's system had seemed to forbid it, exposing cracks in the axiomatic structure itself.
Some insects called the human race
From 1816 Gauss directed a geodetic survey of Hanover. Two problems arose. Errors accumulated along chains of short sightings, so he developed the bell curve of random errors around a mean, a foundation of statistics. And he had to fit a curved earth onto flat maps, the parent's problem of wrapping a ball in flat paper. His 1827 paper on curved surfaces founded differential geometry: describe a surface with Descartes's coordinates and analyse it with calculus.
Two ideas in it changed physics:
- A surface can be treated as a space in its own right.
- Curvature can be measured from inside the surface, using distances alone, with no reference to any larger space it might curve into (the result Gauss called his "remarkable theorem").
The second matters because we cannot step outside the universe to look at it. Mlodinow imagines Alexei and Nicolai as two-dimensional beings on the earth's surface. They have no concept of "up." Climbing Everest, they would feel a mysterious force pushing them back toward the base, and triangles enclosing the mountain would contain too much area: curvature would show up as both force and distorted geometry, a preview of general relativity. Four-dimensional beings, he adds, might view us the same way, as insects.
A flatland scholar, Noneuclid, checks the Pythagorean theorem on ever larger right triangles on the earth:
| Triangle | Sum of squares of legs | Square of hypotenuse |
|---|---|---|
| New York, Newark, New Milford (about 9 x 12 miles) | 209 | 209 |
| Libreville, Kano, Kampala | 3,444,500 | 3,404,025 |
| Libreville, Cagliari, Lérida (Colombia) | 38,264,845 | 32,455,809 |
The discrepancy grows with size. Her students then find that the shortest route from New York to Madrid is not due east along the shared latitude but a great circle curving north first, the route golden plovers fly. A triangle with two meridians and the equator has an angle sum of 270 degrees.
Gauss's intrinsic view prepares Riemann: a being who never leaves a surface can still learn its curvature from distance measurements. Humans on Earth once assumed global Euclid from local experience. The insect image warns that local Euclid does not imply global Euclid. Relativity will make the warning mandatory.
Gauss the surveyor meets instrument error: short segments chain into long lines; random errors accumulate. He invents least-squares discipline and then the harder problem: a 2D map from 3D data with elevation and Earth's curve. A sphere's surface will not flatten without distortion. Differential geometry, Descartes's coordinates plus calculus, becomes the language of that patching.
The deeper breakthrough: curvature is intrinsic. A bug living on a surface can discover Gaussian curvature without climbing into a higher-dimensional room. Mlodinow's insects are us: we measure angles and distances in our space and ask whether Euclid holds. That insect's-eye method is the bridge from survey maps to Riemannian geometry and, later, to space-time curvature that needs no embedding.
ANTS-ON-AN-APPLE. Intrinsic geometry: insects measure from inside; no need for embedding space.
Draw a triangle on an orange peel, sum its angles roughly, then flatten a peel scrap and watch distances fail. Separate intrinsic facts (on the peel) from embedding facts (in the room).
The insect parable can sound anti-human. Its point is epistemic: observers limited to internal measurements can still recover curvature. It is not a claim that humans are small-minded insects.
Survey errors, bell curves, intrinsic curvature (source depth)
1816 onward: Gauss directs a German geodetic survey. Instruments have short range; chained segments accumulate random error. Instead of hair-pulling, he invents the normal distribution of errors (bell curve around the mean), a pillar of modern statistics.
Second problem: patch a 2D map from 3D data with elevation and Earth's curve. Gift-wrap a ball with flat paper: cut into patches. 1827 paper founds differential geometry: Descartes coordinates plus calculus on surfaces. Deep claim: curvature is intrinsic; a bug on the surface can detect Gaussian curvature without embedding in higher space. Humans are the insects. Sphere geometry (elliptic) was known for mapmaking yet not owned as a full sibling of Euclid until Riemann.
| Problem | Gauss move |
|---|---|
| Segment error buildup | least squares / normal error law |
| Globe ≠ plane | differential geometry / patching |
| Bug's-eye test | intrinsic curvature |
A tale of two aliens: Riemann
Bernhard Riemann, born in 1826 near Gauss's birthplace (perhaps, the author jokes, an alien colony near Hanover was seeding geniuses), was a pastor's son, pathologically shy, and brilliant: he returned Legendre's 859-page Theory of Numbers in six days and later aced an exam on it.
For his qualifying lecture Riemann offered three topics. Custom was to pick the first; the dying Gauss chose the third, "On the Hypotheses Which Lie at the Foundations of Geometry." After weeks of paralysis Riemann wrote it in seven, and delivered it on June 10, 1854. He worked locally, describing how distance is measured in infinitesimal regions, and never mentioned non-Euclidean geometry by name, but the implication was clear: the sphere is a two-dimensional elliptic space, with great circles as lines.
The model broke more than the parallel postulate. Great circles have finite length, so Riemann reinterpreted postulate 2 as requiring lines to be unbounded, not infinite. Any two lines meet twice. "Between" becomes ambiguous: is Indonesia between Africa and South America? It depends which way round you go. On a planet of 40-mile radius, an expanding hula hoop passes over the far side and you end up outside it without moving. Gauss had warned Bolyai in 1832 that words like "between" needed clear foundations. Riemann, focused on local structure, left these loose ends; the lecture was published only in 1868, after his death from tuberculosis at 39. His differential geometry became the mathematics of general relativity.
The book appeared as hyperbolic geometry was becoming central to topology through William Thurston's geometrization conjecture, which says every three-dimensional space can be cut into pieces carrying one of eight standard geometries. In 2002 and 2003 Grigori Perelman posted proofs of geometrization, and with it the century-old Poincaré conjecture, using Richard Hamilton's Ricci flow, which smooths a space's curvature like heat spreading through metal. He declined the 2006 Fields Medal and in 2010 declined the $1 million Clay Millennium Prize, saying Hamilton's contribution was no less than his. As of October 2026 it is still the only Millennium Problem the Clay Institute has officially recognised as solved; on September 11, 2026, after a claimed proof of blow-up announced by OpenAI (with priority disputed by Alpöge and Buckmaster), the Institute said the Navier-Stokes problem had "apparently been settled" but called its verification "deliberately unhurried" and has awarded nothing.
- 2022: Maryna Viazovska won the Fields Medal for proving that the E8 lattice is the densest sphere packing in 8 dimensions (and, with collaborators, the Leech lattice in 24).
- 2023: Smith, Myers, Kaplan and Goodman-Strauss found the "hat," a single tile that covers the plane only non-periodically, then the "spectre," which needs no mirror images.
- 2025: Hong Wang and Joshua Zahl posted a proof of the Kakeya conjecture in three dimensions (sets containing a unit segment in every direction must be fully three-dimensional).
Bernhard Riemann (1826-1866) habilitation lecture, 10 June 1854, "On the Hypotheses Which Lie at the Foundations of Geometry," invents the conceptual core of Riemannian geometry: n-dimensional manifolds, metric form, curvature. Gauss sat in the audience. Riemann also planted seeds for complex analysis and number theory's ζ-function. He dies young of tuberculosis.
Mlodinow's "two aliens": beings who grow up in curved space find Euclid alien. We are the other aliens when we visit their metrics. Physics after 1915 moves into their house.
METRIC-TENSOR-MANIFOLD. Riemann's 1854 lecture: geometry from a metric at each point; dimension free; curvature free.
Write the Pythagorean theorem ds² = dx² + dy². Then write a deformed version ds² = E dx² + 2F dx dy + G dy². Circle E,F,G as "the metric." That is Riemann's seed without indexes.
Calling Riemann's lecture instantly revolutionary overstates uptake. It took decades, tensor calculus and Einstein to make the alien native. Genius lectures are not automatic paradigm shifts.
Manifolds, habilitation, metric fields (source depth)
Riemann's habilitation under aging Gauss generalizes curvature to n dimensions: manifolds with metric fields varying from point to point. Two aliens measuring triangles could disagree about "flat" if they live in different geometries. Physics inherits a menu of spaces; Einstein will let stress-energy pick the metric. The curved-space revolution is now a toolkit, not a scandal.
Geometry becomes a field. Specify g locally; global shape is integration and topology, not a single Euclidean decree.
After 2,000 years, a face-lift
Riemann's work exposed three kinds of flaw in Euclid: unstated assumptions, badly formulated assumptions, and attempts to define the undefinable.
- Undefined terms are unavoidable. A dictionary defines "space" by "place" and "place" by "space"; any finite language must leave some words undefined. They gain meaning only from the axioms. As David Hilbert put it, one must be able to say "men, women and beer mugs" in place of points, lines and circles.
- Hidden steps. Euclid's very first proposition assumes two circles intersect, though nothing guarantees continuity. He assumed that of three points on a line, one lies between the others, a hidden straightness condition. Assuming just one triangle with angles summing to 180 degrees is enough to prove the parallel postulate.
The repair came quickly: Felix Klein fixed elliptic geometry in 1871; Peano axiomatised Euclidean geometry in 1894; Hilbert published his system in 1899 and refined it until 1930, with 8 undefined terms and 20 axioms in four groups (for example, "there are at least two points on each line").
Before: axioms are self-evident truths about physical space; mathematics is a kind of physics; consistency is guaranteed by reality. After (by 1900): axioms are arbitrary starting points; mathematical spaces are abstract logical structures; which geometry describes the universe is a question for physics; consistency must be proved.
Proving consistency proved hard. Hilbert turned points into pairs of numbers, so any geometric contradiction would be an arithmetic one: geometry is relatively consistent (as consistent as arithmetic). To get at arithmetic itself, mathematicians built set theory, which hit paradoxes such as Grelling and Nelson's (1908): is the word "heterological" (not describing itself) heterological? Russell and Whitehead's Principia Mathematica (1910 to 1913) tried to derive all mathematics from logic. In 1931 Kurt Gödel proved that any system rich enough for arithmetic contains true statements that cannot be proved within it. There is still no universally accepted completion of what Euclid began.
- Two millennia of "proofs" failed because each smuggled in the conclusion as an obvious-looking assumption.
- Change one axiom and you get a new, consistent world: hyperbolic (many parallels) and elliptic (none).
- Curvature can be detected from inside, by measuring distances and triangles. That is what makes cosmology possible.
- Geometry became abstract, and which geometry is real became an empirical question.
- Formal systems have limits (Gödel): completeness and proven consistency cannot both be had.
Coordinate formulas for curvature (Christoffel, Riemann tensor, Ricci) turn philosophical manifolds into calculational engines. By 1900 the toolbox exists; Einstein will hire it for gravity. The "face-lift" is Mlodinow's joke: Euclid's window gets new glass grinding equipment, not a cosmetic selfie.
Modern axiomatics (Pasch, Hilbert and others) strips Euclid's hidden appeals. Undefined terms are allowed; diagrams are illustrations, not premises; continuity and order get explicit axioms. Common notions and postulates merge into one assumption list. Thābit's "equidistant curve" mistakes show how intuition smuggles theorems.
The face-lift does not make Euclid wrong; it makes him precise. Non-Euclidean geometries become siblings, not heresies. Physics inherits the lesson when general relativity lets geometry vary with stress-energy: axioms about space are empirical hypotheses at cosmological scale.
TENSOR-FACE-LIFT. Nineteenth-century vector and tensor calculus makes Riemann computable; geometry ready for physics.
Translate one school vector identity (divergence theorem in words) into "why curvature needs bookkeeping." One paragraph. That is the face-lift's motive.
Credit-spreading matters: Grassmann, Riemann, Christoffel, Ricci, Levi-Civita share the lift. A single-genius face-lift story lies.
Hilbert, Pasch, and honest axioms (source depth)
Pasch, Hilbert and others strip hidden diagram appeals: undefined terms allowed; betweenness and continuity stated; common notions merge with postulates. Thābit's equidistant-curve mistakes show intuition smuggling theorems. Face-lift ≠ Euclid wrong; it means precise. Non-Euclidean siblings gain equal citizenship; GR later treats axioms about space as empirical at cosmic scale.
Formal libraries (Lean, Isabelle) encode large pieces of Euclid-style geometry and Hilbert-style rebuilds. The face-lift continues as machine-checked axioms, not only prose.
Part IV. Einstein: The Story of Space-Time
Gauss and Riemann showed that space could be curved. The next questions were what kind of space we live in, and what decides its shape. Riemann had guessed the answer in 1854: the ground of space's metric relations must be sought "in the binding forces which act on it." In 1870 William Kingdon Clifford, twenty-five, told the Cambridge Philosophical Society that small regions of space are like little hills, that curvature passes from place to place like a wave, and that this variation "is really what happens in that phenomenon which we call the motion of matter." He was right, and had no theory. Without the physics of special relativity and the proper role of time, Clifford worked through the nights and got nowhere; he died in 1876 at thirty-three. The mathematics had arrived before the physics, a situation Mlodinow compares to string theory today.
Part IV welds space to time. Ether hunts fail; special relativity revises simultaneity; equivalence and field equations make gravity geometry; eclipse fame and later precision tests industrialise the curved world. Euclid's plane becomes a local chart on a dynamical space-time.
Revolution at the speed of light
For two centuries Newton seemed to explain every motion. Space was absolute, a fixed frame on which to lay Descartes's coordinates, and time merely labelled positions along a path. Newton's system had two halves: kinetics (how bodies respond to forces: acceleration equals force divided by mass) and dynamics (what the forces are; he supplied only gravity). Together they predicted orbits, and they united earthly physics with astronomy.
If Newton's space and time are right, two things are obvious. There can be no cosmic speed limit: if something approached you at the limit and you spat at it, it would approach your spit faster. And light cannot have the same speed for everyone: run towards it and it should come at you faster. Both "obvious truths" are false, and that is special relativity.
For two centuries Newton's absolute space and universal time underwrote mechanics. Maxwell's electromagnetism put a speed c into the laws. Either Maxwell holds only in an ether frame, or Newton's kinematics yields. Einstein's 1905 choice: keep Maxwell's form, rebuild kinematics. Part IV opens here.
Newton's absolute space and universal time made two "obvious" truths: no cosmic speed limit, and light's speed depending on the observer's motion. Maxwell's electromagnetism puts a single speed c into the field laws. Either those laws hold only in an ether frame, or kinematics must change. Part IV's opening is that fork. Einstein's 1905 choice (detailed in the next chapters) keeps Maxwell's form and rebuilds the meaning of space and time. The revolution is not a bomb metaphor; it is a constraint: invariants over absolute frames.
C-IS-LAW. Light speed invariant; absolute space and time fail; kinetics must change before gravity does.
Write two "obvious" Newtonian truths (no speed limit; light speed depends on observer). Strike them. Replace each with the relativistic sentence. That is the revolution's unpaid debt collection.
Starting relativity from bombs or pop "everything is relative" wrecks the subject. The theory is about invariants (c, interval), not moral relativism.
Relativity's other Albert: the ether hunt
Albert Michelson was born in Prussian Poland, grew up in the silver-boom town of Virginia City, Nevada, and at sixteen, having missed a Naval Academy place, took the new transcontinental railroad to Washington and won a special appointment from President Grant. At Annapolis he was first in optics, twenty-fifth in seamanship and last in history; the superintendent told him to spend less time on science and more on gunnery.
His textbook described the ether, a "subtle, imponderable and eminently elastic fluid" filling the universe, through which light waves were thought to travel the way sound travels through air. The history of the idea:
- 1678 Huygens revives Aristotle's fifth element as the medium for light waves.
- 1801 Thomas Young's two-slit interference makes light a wave; waves seem to need a medium.
- 1821 Fresnel shows light waves are transverse, which requires a rigid medium: the ether becomes a solid filling all space that offers no resistance to planets.
- 1865 Maxwell's electromagnetic theory; in 1878 he writes in the Britannica that interstellar space is filled with "a material substance or body."
- 1881 Michelson in Potsdam: no ether wind. 1887 Michelson and Morley in Cleveland: still none.
- 1889 to 1904 FitzGerald and Lorentz propose that moving bodies contract; Lorentz invents "local time"; Poincaré doubts absolute time.
- 1905 Einstein drops the ether.
James Clerk Maxwell's equations unified electricity, magnetism and light ("God said, let there be [four equations], and there was light," ran a Pasadena T-shirt). His original 1865 version was twenty equations in twenty unknowns, poorly explained, "a kind of intellectual jungle" (Ehrenfest). Read carefully, it implies special relativity and no ether. Nobody saw that then. Maxwell did see the test: if light moves at a fixed speed through the ether and the earth moves through it, light's measured speed should change with direction and season. A letter he wrote in 1879, dying of cancer, was published in Nature, and Michelson read it.
Dad runs north with Nicolai 10 yards ahead and Alexei 10 yards to the side. Dad shouts; each boy shouts back. Through still air, the sideways round trip beats the up-and-back one: Nicolai's leg crawls at c minus their speed even if the return leg is fast ("slow and steady wins the race"). Replace sound with light, air with ether, and running with the earth's orbital speed of 18.5 miles per second. Light is about 10,000 times faster, so the expected difference is about one millionth of a percent. Michelson's interferometer splits one beam with a half-silvered mirror (Fizeau's beam splitter), sends the halves down perpendicular arms and recombines them; any time difference shows up as shifted interference fringes. Rotating the apparatus by 90 degrees swaps the roles of the arms.
The apparatus had to keep arm lengths steady to a thousandth of a millimetre and temperatures to a hundredth of a degree; it could detect a foot stamping 100 yards away. Michelson had it built on Alexander Graham Bell's credit account at Schmidt & Haensch. In Potsdam in April 1881 he found essentially no effect. He concluded not that the ether was absent but that the earth was somehow not moving through it. Lord Kelvin declared the ether "the only substance we are confident of in dynamics." Lorentz pointed out an error that halved the expected effect, so in July 1887 Michelson and Edward Morley repeated the experiment with far better precision. Again nothing, and again little interest. They did not even make their planned seasonal follow-ups.
The fuse was lit in 1889, when George FitzGerald wrote a short letter to Science suggesting that bodies change length when moving through the ether. Lorentz reached the same idea and tried to explain it with molecular forces; without a deeper reason it was an ad hoc patch, "like Ptolemy's epicycles." By 1904 Lorentz had "local time," Kaufmann had measured mass increasing with speed, and Poincaré had written that "there is no absolute time." The pieces were on the table.
Albert A. Michelson (not Einstein), navy-trained, builds precision interferometers. With Edward Morley at Cleveland in July 1887, he finds no significant fringe shift from Earth's motion through a luminiferous ether. Reception is lukewarm; Lord Kelvin still affirms the ether in 1884 lectures. FitzGerald (1889) and Lorentz propose length contraction; Lorentz adds local time. The fuse burns toward 1905.
MICHELSON-NULL. Interferometer finds no ether wind; FitzGerald-Lorentz contraction patches; Einstein discards the ether.
Sketch a Michelson interferometer (splitter, two arms, recombination). Label what null result means for "ether wind." Fifteen minutes, no equations.
Heroising a single null result forgets how long the ether lasted afterward. Scientists rationally tried drag, contraction and local time before killing the medium.
Michelson, Morley, and the null that would not die (source depth)
Albert A. Michelson (with Morley) seeks Earth's motion through ether via interference; the null result refuses a simple wind. Lorentz contractions patch the ether; Einstein discards the stage. Michelson never fully embraces relativity. The hunt's legacy is methodological: a precise null can kill a medium.
The Stuff of Space
Trying to learn what space was made of lit the fuse to relativity. The struggle was led by people who often did not know where they were going. Like space itself, the path bent.
In 1865 a short Scottish physicist published "A Dynamical Theory of the Electromagnetic Field," then in 1873 A Treatise on Electricity and Magnetism. Born James Clerk, he gained the surname Maxwell for an inheritance clause. His theory ranks with mechanics, relativity and quantum theory among physics' cornerstones. A Caltech-area T-shirt once paraphrased Genesis: God said "let there be [Maxwell's equations]," and there was light. Radio, radar, television and a quantum electrodynamics that became the template for the Standard Model all descend from this work. A careful reading already hints at special relativity and at the absence of an ether. None of that was obvious at the time.
Students meet Maxwell as four tidy vector equations. Discovery was messier: Maxwell's 1865 form used twenty differential equations in twenty unknowns. Lorentz later complained of a missing unity; Ehrenfest called the book an "intellectual jungle." Maxwell had dumped a core, not a lesson plan. Yet on the ontology of space he was clear. In the ninth edition of the Encyclopaedia Britannica (1878) he wrote that whatever the difficulties, interplanetary space is occupied by a material substance, "the largest, and probably the most uniform body of which we have knowledge." Even Maxwell could not let the ether go.
Credit where due: he did not wave hands. If light's speed is constant relative to the ether and Earth orbits, the apparent speed of starlight should vary with season. On 23 April 1864 he tried an experiment; G. G. Stokes blocked publication, wrongly doubting the approach. In 1879, dying of stomach cancer, Maxwell sent a letter that Nature published posthumously. Michelson read it and saw a design.
Mlodinow's park parable: Dad at a right angle's vertex, Nicolai north, Alexei west, all running north. Yells through still air make the round-trip times differ; "slow and steady" wins. Replace air by ether and yells by light, and Earth's ~18.5 miles per second orbital motion should play Dad's role. Light is ~10,000× faster than that orbit, so the fractional time difference is ~10⁻⁸: hopeless without interference tricks.
Fizeau had already measured c terrestrially (1849, ~5% accuracy) and built beam splitters. Fresnel's partial-drag idea complicated expectations. Michelson, still under thirty, on a U.S. Navy fellowship in Berlin and Paris, designs an interferometer sensitive to thousandths of a millimetre and hundredths of a degree. Brasswork from Schmidt & Haensch is charged to Alexander Graham Bell's account. Potsdam 1881: null result. Michelson concludes not "no ether" but "we are not moving through it" (drag?). Kelvin in 1884 still calls the ether the one substance dynamics can be sure of. Most physicists shrug.
Lorentz (1886) and Potier note a subtle path-length error that halves the expected shift; experimental error might hide a real wind. Michelson, now at Case in Cleveland, partners with Edward Morley. After a nervous breakdown and a semester away, he returns. At noon on 8, 9, 11 and 12 July 1887 they run the definitive null experiment. Planned seasonal repeats fade; interest cools.
The fuse's first smoke: an 1889 Science letter by George Francis FitzGerald suggesting lengths contract with motion through the ether by an amount of order v²/c². Lorentz reaches the same idea and tries to derive it from molecular forces in the ether. By 1904 Lorentz has local versus universal time; Kaufman sees mass depend on velocity; Poincaré doubts absolute simultaneity and toys with c as a speed limit. The dividing line between lasting space and fleeting time is cracking. Einstein will keep Maxwell's field spirit, drop the medium, and rebuild kinematics.
Classical ether is dead. Quantum field theory fills space with fields and vacuum expectation values, not a mechanical jelly. Relativity still forbids a preferred undetectable rest frame of the Michelson kind. Geometric deep learning and numerical GR treat space-time as a dynamical mesh; that is method, not a luminiferous medium.
James Clerk Maxwell (1831-1879) publishes a dynamical theory of the electromagnetic field (1865) and the Treatise (1873). His original form used twenty equations in twenty unknowns; modern vector calculus compresses them to four. A Caltech T-shirt joke puts Maxwell's equations in God's mouth at creation: let there be light.
Maxwell still affirms a material ether in the Encyclopaedia Britannica (1878). He tries and fails to measure Earth's motion through it; a dying letter (1879) reaches Nature and seeds Michelson's idea. Fizeau's beam splitter and long-path speed-of-light work supply tools. Michelson's Potsdam 1881 attempt and the 1887 Cleveland refinement light a fuse. FitzGerald and Lorentz invent contraction; Poincaré questions absolute time. Einstein will keep the field equations' spirit and drop the medium.
This chapter is the book's missing hinge between ether hunt biography and patent-clerk theory: what space is "made of" becomes the question that kills absolute space.
MAXWELL-TWENTY-THEN-FOUR. Ether as largest body; twenty equations become four; light as EM wave; ether dies hard.
Read Maxwell's Britannica claim that space is occupied by a material ether. Write a 2026 replacement sentence about quantum fields. Note what "stuff" still means.
Mocking Maxwell for believing in ether ignores that his theory forced the crisis. Wrong ontology, right equations: a common scientific pattern.
Ether as furniture, then vacancy (source depth)
Nineteenth-century physics filled space with luminiferous ether so light waves would have a medium. Michelson-Morley and later Einstein empty that furniture: fields need no mechanical jelly. "Stuff of space" becomes geometry and fields, not a gas of undetectable wind. The chapter clears the stage for relativity's axioms.
Probationary technical expert, third class
Einstein, born in Ulm in 1879 (the year Maxwell died), was no child prodigy: a late talker, a chair-thrower, bad at rote learning, though at five he wondered what invisible force moved a compass needle. Told by his Greek teacher that he would never amount to anything, he left school at fifteen with a doctor's note, idled in Italy, failed the ETH entrance exam (French, chemistry and biology dragged down his maths), spent a year at a nearby school and got in. He skipped lectures, passed on a friend's notes (Marcel Grossmann's), and alienated professor Weber, whose poor references kept him out of academia. On June 23, 1902 he became a Probationary Technical Expert, Third Class, at the Swiss Patent Office in Bern.
- Doctoral thesis on molecular dimensions, later the most-cited pre-1912 paper in 1961 to 1975, more than his relativity papers.
- Brownian motion explained by molecular bombardment, confirmed by Perrin (Nobel 1926).
- Photoelectric effect: light comes in quanta, the work cited for his 1921 Nobel Prize.
- Special relativity, in two papers (September and November), the second giving the equivalence of mass and energy.
Albert Einstein joins the Swiss Patent Office in Bern as technical expert third class. In 1905 he publishes papers on the photoelectric effect, Brownian motion, special relativity and E=mc². Mlodinow stresses the workplace: examining electromagnetic inventions keeps light and fields in view. Geometry's next window opens from a stool, not a throne.
School Albert: strong in math and Latin, weak elsewhere; a Greek teacher called him hopeless. By 1895 he already knew Michelson-Morley, Fizeau and Lorentz, still accepted an ether, and already suspected you can never catch a light wave. Family moves to Italy; he drops out into a life that suits him until father Hermann hocks a chainik until Albert returns to study electrical engineering. Entrance exams: math fine; French, chemistry and biology drag him down.
The Swiss Patent Office stool is not a romantic accident alone. Failed academic applications, the Olympia Academy reading group and daily electromagnetic gadgets form the pressure cooker of 1905. Four papers: light quanta, Brownian motion, special relativity, E=mc². Geometry's next demand (gravity) is already latent. The chapter's moral for Mlodinow: revolutions sometimes clock in as third-class examiners.
PATENT-1905. Bern office; third-class examiner; annus mirabilis; geometry about to meet kinetics.
List three constraints of a dull day job that could help deep work (fixed hours, low prestige, concrete puzzles). Circle one you can simulate this month.
The patent-clerk myth can romanticise precarity. Einstein also had classmates, the Olympia Academy, and years of failed academic applications. Constraint helped; it was not a self-help recipe.
Ulm, late speech, and the 1905 burst (source depth)
1805: Napoleon spares Göttingen partly for Gauss; Ulm's victory site later hosts Einstein's 1879 birth, the year Maxwell dies. Einstein spoke late (some say age three), quiet, withdrew; threw a chair at a tutor. Spotty elementary record; rote school punished wrong answers with knuckle raps, so he checked twice before speaking. By thirteen, advanced math with uncle and friend; Kant on time and space.
Patent office title: probationary technical expert, third class. 1905 miracles: Brownian motion; light-quantum paper applying Planck universally (Lorentz and even Planck opposed; Nobel 1921 for the photoelectric insight); two relativity papers that begin the curved-space odyssey Gauss and Riemann had made mathematically possible.
Cast. Young Einstein: slow speaker, fast checker, compass puzzler at five. Patent director (1909): laughs when Einstein resigns for a university chair, unaware what had been stuffed into the desk drawer.
A relatively Euclidean approach: special relativity
Einstein loved his teenage geometry book ("this lucidity and certainty made an indescribable impression on me"), and he built special relativity like Euclid, from axioms:
- You cannot tell, except relative to other bodies, whether you are at rest or moving uniformly. (Oresme's and Galileo's principle; when Nicolai on his toy fire truck knocks the book from Alexei's hands, each boy's account is equally valid.)
- Light's speed is independent of its source and the same for every observer.
The second axiom is the shock. To keep both, something else must go, and Einstein found it: simultaneity. Two parents on a platform photograph two boys at the ends of a passing subway car "at the same instant"; the flashes meet midway between the parents. In the moving car the meeting point is off-centre, so the boys conclude one photo was taken first. Neither view is privileged.
Once simultaneity is relative, so are length (to measure a moving object you must mark both ends at the same time) and duration. Moving objects are contracted along their motion; moving clocks run slow. The time a clock measures along its own path is its proper time.
Relativity replaces Euclid's vocabulary: points become events (a place and a time), paths become worldlines, distance becomes a space-time interval combining time and space, and straight lines become geodesics. Newton's first law becomes: a free object follows the worldline along which its own clock records the most time. Any detour means moving faster, and moving clocks run slow.
Einstein braced for attack and got silence, then a letter from Max Planck. Promoted to "technical expert, second class" on April 1, 1906, he called himself a "venerable federal ink-shitter." Planck lectured on the theory, sent Max von Laue to Bern (von Laue dropped Einstein's cheap cigar into the Aare), wrote the first paper on relativity by anyone else, and supervised the first thesis on it. Hermann Minkowski, his old teacher, recast it as geometry in 1908: "Henceforth space by itself, and time by itself, are doomed to fade away into mere shadows." Michelson, Lorentz and Poincaré never fully accepted it.
Special relativity (1905) keeps Euclidean ideas inside inertial frames while changing the assembly: points become events; distance becomes interval; straight lines become inertial worldlines. Minkowski's 1908 space-time geometricises Einstein's kinematics. Newton's cosmic speed freedom dies; simultaneity fragments across frames.
Einstein's two axioms sound mild until combined. The principle of relativity (already in Galileo and Oresme): the laws look the same in every inertial frame; you cannot tell uniform motion from rest by an internal experiment. The second: the speed of light in vacuum is the same for all inertial observers, independent of source motion. Each axiom alone bothers nobody much. Together they outlaw absolute simultaneity.
Mlodinow's kids-and-ax parable restates Galilean relativity: whether Nicolai's truck or Alexei's book is "really" moving, Newton's laws give the same collisions. Maxwell's equations already make light's speed independent of the book, like any wave in a medium. The meat is denying the medium a detectable rest frame. If you could tell whether you are moving through ether, everyone could agree on c relative to the "stationary" object. Einstein refuses that privilege.
Railroad-car thought experiments (Einstein's 1916 Relativity) show the cut. Lightning strikes both ends of a moving train; an embankment observer says the strikes are simultaneous; a mid-train rider says the forward strike happens earlier because she rides toward its light. Simultaneity of distant events is frame-dependent. From that crack follow time dilation, length contraction and the twin "paradox" (not a paradox once acceleration and path-dependent proper time are admitted).
Minkowski's 1908 geometric language makes the bookkeeping clean: the invariant interval replaces Euclidean distance. Euclid still rules the spatial slices of an inertial frame; space-time as a whole is not Euclidean. That is the sense of Mlodinow's "relatively Euclidean" title.
INTERVAL-NOT-DISTANCE. Events, worldlines, Minkowski interval; clocks and rods renegotiate; Euclid lives in space slices carefully.
Write the Minkowski interval c²dt² − dx² − dy² − dz². Circle what stays invariant. Invent one paradox (twin, ladder) and resolve it with invariant language only.
"Relatively Euclidean" can mislead: space slices are Euclidean in inertial frames, but space-time is not. Teachers who stop at time dilation leave students Euclidean in the wrong dimension.
Two axioms and Minkowski's shadow (source depth)
26 Sep 1905 Annalen: "On the Electrodynamics of Moving Bodies"; November: inertia and energy content. Special relativity rests on two space axioms: (1) uniform motion is indistinguishable without external reference (Galilean/Oresme principle); (2) light's speed is the same for all inertial observers. Euclid's method (axioms first) impressed Gymnasium Einstein even though his later gravity would abandon Euclidean comfort.
Mlodinow's fire-truck vignette: Nicolai's plastic ax meets Alexei's book; Newton's laws cannot say who "really" moved. Add light-speed invariance and simultaneity splits. Minkowski (1908): space and time fade to shadows; only a union remains. July 1907 Planck to Einstein: advocates form "a modest crowd." Michelson kept the ether; Lorentz never fully broke; Poincaré opposed until death (1912). Acceptance pattern: Planck engaged early with a small advocate crowd. Michelson kept ether loyalty after the null. Poincaré opposed until 1912. A German core accepted gradually, then faced the harder GR shock.
Einstein's apple: the equivalence principle
In November 1907, at his patent office desk, Einstein had "the happiest thought of my life": a person falling freely does not feel his own weight. Special relativity was only a kinetics; it needed a theory of gravity, and Newton's would not fit. Newton's force acts instantaneously "at a given instant," yet nothing travels faster than light and "the same instant" is now relative. And special relativity privileged uniform motion, which bothered Einstein: who says the stadium crowd is at rest and the astronaut is spinning, or vice versa?
Acceleration announces itself. Little Hans Albert on a merry-go-round lets go of a lollipop, and from his point of view it flies away for no reason. He can rescue Newton's laws only by inventing a force pushing everything outward, which Mlodinow calls "schmavity." Newton knew such fictitious forces appear in accelerating frames. Einstein's step was to say gravity is one of them:
You cannot tell, except relative to other bodies, whether you are accelerating uniformly or at rest in a uniform gravitational field. In a falling elevator your coffee floats as in deep space; in a rocket accelerating at 32 feet per second per second it splashes as on earth. Gravity can be removed, locally, by choosing a different frame.
The principle has strange consequences. Put Alexei at the top of an accelerating rocket flashing a light each second; Nicolai at the bottom, moving toward the flashes, receives them early. Attribute the effect to gravity instead and you must conclude that clocks higher in a gravitational field run faster. The effect is tiny: time on earth runs about two parts per million faster than at the sun's surface, and a station broadcasting at 1070 kHz from the top of the World Trade Center would be received on the ground at 1070.00000000003. It took Einstein five more years to see that gravity also warps space. In Prague in 1912 he reasoned that on a rotating disc, measuring rods laid along the rim contract while those along the radius do not, so circumference divided by diameter is no longer π. "Euclidean geometry must be abandoned." For what?
November 1907, Einstein's "happiest thought": a falling person feels no weight. Gravity becomes a fictitious force like the outward shove Hans Albert feels on a carousel (Mlodinow's "schmavity"). Newton's instantaneous action at a distance cannot survive a universal speed limit. The apple points to curved space-time; the math takes until 1915.
The happiest thought (November 1907): a freely falling person feels no weight. Einstein's personal life was messy (marriage, divorce, remarriage, sour remarks on marriage), yet this physical insight was serene. Special relativity had replaced Newton's kinetics; Newton's gravity, built to match the old kinetics, could not stay. Instantaneous action at a distance is guilty under a universal speed limit, and the phrase "at a given instant" no longer names a shared fact for distant masses.
Something else bothered him: special relativity privileged uniform motion. Accelerating frames announce themselves with fictitious forces. Little Hans Albert on a merry-go-round needs "schmavity" to save Newton's laws in his spinning world. Einstein's step: gravity is one of those fictitious forces, locally indistinguishable from acceleration. Elevator thought experiments make the local equivalence vivid; tidal effects (geodesic deviation) mark where the equivalence is only local.
The apple is a seed, not the orchard. Turning equivalence into field equations took years of tensor calculus with Marcel Grossmann's help, false starts and competition with Hilbert. The payload for geometry: mass-energy tells space-time how to curve; curvature tells free bodies how to move.
HAPPIEST-THOUGHT. Free fall cancels weight; gravity equals acceleration; "schmavity" on the merry-go-round.
In an elevator thought experiment, list three experiments that cannot tell gravity from acceleration. Then name one that needs tidal gravity (higher order). That is equivalence plus its limit.
Equivalence is local. Pretending it abolishes all distinction between gravity and acceleration fails at tidal effects: geodesics converge for real mass distributions.
Happiest thought and the merry-go-round (source depth)
November 1907, Bern patent office: "if a person falls freely he will not feel his own weight." Eight hours a day, six days a week rejecting perpetual-motion machines; notes stuffed in the desk when the director approached. He called it "the happiest thought of my life" despite a hard personal ledger (divorce, adoption of firstborn, schizophrenic youngest child, exile from Nazi Europe).
Equivalence: gravity vs acceleration. Merry-go-round thought with Hans Albert: Lorentz contraction of tangential sides makes a square look rectangular; circumference/diameter ≠ π. Father concludes Euclidean geometry must go; the question is, in favor of what? That question is general relativity's homework for the next eight years.
Weightlessness in free fall is local inertial motion. Gravity is not a force in the Newtonian sense; it is the geometry of who is inertial.
From inspiration to perspiration: general relativity
Einstein needed a geometry of warped space. Riemann had built it, but Einstein had never heard of him; he remembered Gauss's theory of surfaces from a student course and went to Grossmann, now a geometer: "Grossmann, you must help me or I'll go crazy." Grossmann found Riemann's differential geometry and called it "a terrible mess which physicists should not be involved with." Einstein dived in anyway ("compared with this problem, the original theory is child's play"). Planck warned him that he would fail, and that if he succeeded no one would believe him. On November 25, 1915 he presented the field equations of gravitation to the Prussian Academy.
The core idea comes from Gauss: distance between points defines the geometry, the difference between a phone book and a map. In flat space with grid coordinates, squared distance is the Pythagorean sum. In curved space it becomes "Nonpythagoras's theorem":
distance² = g₁₁ x (east-west)² + g₂₂ x (north-south)² + g₁₂ x (east-west) x (north-south)
The g-numbers form the metric, they may vary from place to place, and they completely describe the space. (For flat space, g₁₁ = g₂₂ = 1 and g₁₂ = 0.) In four-dimensional space-time the metric has ten components. Einstein's equations relate the distribution of matter and energy to the metric: mass does not exert a force; it changes the shape of space-time, and free bodies follow its geodesics.
For a sphere of uniformly spread matter, the measured radius falls short of the Euclidean value by about 2.5 x 10⁻²⁹ cm per gram of mass inside. For the earth that is about 1.5 millimetres; for the sun, about half a kilometre. Light grazing the sun should bend by about 1.75 arc seconds, twice the value a Newtonian calculation gives.
The book mentions that GPS needs general relativistic corrections. According to NIST, GPS satellite clocks would gain about 38 microseconds a day relative to the ground: roughly 45 microseconds faster because they sit higher in earth's gravity, minus about 7 microseconds from their orbital speed (special relativity). Left uncorrected, the error would build up day after day and quickly ruin positioning. The satellites' clocks are set slightly slow before launch so they match the ground in orbit.
General relativity equates geometry with gravitation: matter-energy tells space-time how to curve; curved space-time tells matter how to move. Mercury's perihelion advance (~43" per century beyond Newton) fits. Light deflection at the Sun (~1.75") is a testable claim. For Earth, radius deficit versus Euclid is millimetres; for the Sun, order of half a kilometre in the book's figures.
Equivalence is inspiration; field equations are perspiration. With Grossmann, Einstein learns Ricci and Riemann tensors; wrong versions (Entwurf) fail Mercury or covariance; November 1915 brings the Einstein equation. Hilbert races nearby. Observational early wins: Mercury's perihelion (~43" per century beyond Newton) and light deflection (~1.75" at the Sun's limb). The book's radius-deficit figures (Earth millimetres; Sun ~0.5 km) make curvature tangible. Geometry is no longer a fixed stage; it is the dynamical actor.
GR-1915. Metric as gravity; Einstein equation; perihelion of Mercury; light bending.
State in one sentence what mass-energy does to the metric, and what the metric does to motion. That two-sentence loop is GR without indexes.
Priority and help get messy (Grossmann, Hilbert). The myth of lone Einstein in a room erases the tensor tutors. Keep the perspiration visible.
Eight years from apple to field equation (source depth)
Inspiration (1907) to finished field equations (1915) was grind: tensor calculus (Grossmann), false starts, competition with Hilbert, mercury's perihelion, light bending. Geometry returns to physics' center after Newton's calculus displaced it. Special relativity becomes the easy swallow; curved spacetime is the hard one.
1907 happiest thought (equivalence) 1912-13 Grossmann collaboration; metric formalisms 1915 field equations; Mercury perihelion 1919 Eddington eclipse; public fame
Blue hair triumphs
Two British expeditions photographed the eclipse of May 29, 1919. On November 6 the result was announced in London: starlight was bent by the Einstein amount. "REVOLUTION IN SCIENCE," said The Times; "EINSTEIN THEORY TRIUMPHS," said The New York Times, which sent its golf writer and got Einstein's age wrong. Within a year there were over a hundred books on relativity, and Scientific American offered $5,000 for the best 3,000-word explanation.
The book places Eddington at Sobral, Brazil. In fact Eddington observed from Príncipe, off West Africa; Sobral was manned by Andrew Crommelin and Charles Davidson of Greenwich. In 1980 philosophers John Earman and Clark Glymour charged that the British team discarded inconvenient Sobral plates. Historian Daniel Kennefick (Physics Today, 2009) answered that the decision was made by Dyson's Greenwich team on technical grounds (a focus change in one telescope), and a 1979 Greenwich re-measurement of the plates gave about 1.90 ± 0.11 arc seconds, supporting Einstein. Later radio measurements have confirmed the prediction to high precision.
Mlodinow's analogy is his son Alexei, the first in his school with blue hair: some classmates thought it cool, most called him names. Physics resists like a fourth-grade class. Michelson accepted the eclipse data but not the theory; a Chicago astronomer called relativity "a crazy vagary." In Germany, Nobel laureates Philipp Lenard and Johannes Stark branded it "Jewish physics"; a 1931 booklet, A Hundred Authors Against Einstein, listed 120, few of them physicists. When Hitler came to power the Nazis raided Einstein's homes and put a price on his head; he settled at the Institute for Advanced Study in Princeton in October 1933.
His last decades went into a unified theory that few colleagues believed in ("what God has torn asunder, let no one put together," said Pauli). He died on April 18, 1955, refusing surgery for an aortic aneurysm. His own summary: "When a blind beetle crawls over the surface of the globe, he doesn't realize the track he has covered is curved. I was lucky enough to have spotted it."
The author says Heisenberg "rewarded Nazi kindness" by leading their bomb effort, and lists Edward Teller and Victor Weisskopf (Austrian-born, not German) among the Americans who beat him. Historians still debate Heisenberg's motives and how far the German project really got. Read this passage as colour, not history.
- Gravitational waves: on September 14, 2015, both LIGO detectors recorded ripples from two merging black holes (GW150914), announced on February 11, 2016. Weiss, Barish and Thorne shared the 2017 Nobel Prize. Clifford's 1870 guess that curvature travels "after the manner of a wave" was directly confirmed.
- Black hole images: the Event Horizon Telescope released the first image of a black hole's shadow (M87*) in April 2019, and of our galaxy's central black hole, Sagittarius A*, in May 2022.
- Is our space curved? The Planck satellite with galaxy surveys finds the curvature parameter Ω_K = 0.001 ± 0.002 (2018), consistent with a spatially flat universe on the largest scales, even though space-time curves locally around every mass.
- The ether died hard because a wave "needed" a medium; null results were read as failures, not discoveries.
- Special relativity follows from two axioms; simultaneity, length and duration become observer-dependent.
- Gravity is geometry: matter sets the metric; free bodies follow geodesics that maximise proper time.
- Small anomalies matter (Mercury's perihelion drift, a 1.75-arc-second deflection, 38 microseconds a day).
- Acceptance is social: new ideas win as the old generation leaves, and politics can poison the process.
1919-ECLIPSE. Eddington's teams; light bending headline; Einstein celebrity; geometry goes public.
Read one modern re-analysis of the 1919 plates (quality was uneven). Write whether the triumph was data, narrative, or both. That is science as public theatre.
The blue-hair celebrity story can hide that many physicists already took GR seriously before 1919, and that later tests (Pound-Rebka, Gravity Probe B, LIGO) do the heavier epistemic lifting.
Eclipse, fame, and the window's fourth pane (source depth)
1919 eclipse measurements make Einstein a global celebrity; "blue hair" mythic images follow a man whose geometry outran common sense. Light bending, gravitational redshift, and later GPS corrections turn the abstract window into engineering. Part IV closes: space-time is dynamical; Euclid's rigid stage is gone.
GWTC-4.0 parameterized tests: no evidence beyond GR across dozens of confident events; PN deviation constraints tighten by factors ~1.2-5.5. Cosmology (Hubble tension, DESI dark-energy hints) stresses the model at largest scales without replacing the local curved-spacetime framework.
Part V. Witten and the Strings: The Weird Revolution
Einstein showed that matter bends space. In the newest theories space takes ample revenge: its most basic properties, such as the number and shape of its dimensions, decide which particles and forces can exist. "Space, the container of the universe, becomes space, the arbiter of what may be." According to string theory there are extra dimensions, curled up too small to see, and their topology (plane, sphere, donut or pretzel) determines what lives in them. Twist them differently and electrons, and so people, could not exist. Its successor, M-theory, hints that space and time themselves are only approximations.
Mlodinow is candid that this revolution is unlike the others. It is driven by the mathematics of the theory, not by data, which are lacking; physicists cheer when the theory reproduces known particles, a process they call "postdiction." Its leading figure, Edward Witten, won a Fields Medal, not a Nobel Prize, and says string theory should ultimately be a new branch of geometry. And unlike the earlier revolutions, we do not know how it ends.
String theory reopens space itself: extra dimensions, tiny and topologically rich, may shape particle physics even if we cannot wiggle into them. Many physicists laughed or screamed about tax dollars; some still do. Among high-energy theorists the programme became a major research culture.
Unlike prior revolutions, this one is unfinished. No decisive experiment crowns or kills it. Research style drifts toward mathematical consistency (Hilbert's "rules") when new data are scarce. Mlodinow's "weird" is chronological honesty: we are inside the story, not after the credits.
Part V enters unfinished business. Quantum uncertainty, unification hopes, Kaluza-Klein echoes, string birth and troubles, Witten's synthesis era: geometry tries to swallow particles and forces whole. The verdict is not in; the window is still being reglazed.
The Weird Revolution
Quantum rules and relativistic geometry refuse to merge politely. Extra dimensions, strings and M-theory are one ambitious peace offer, still lightly tested.
WEIRD=UNFINISHED. The fifth window is open glass, not a finished view.
Write the one observation that would make you take strings more seriously this decade, and the one that would demote them. Keep the list honest.
Calling the revolution "weird" can excuse low standards. Weirdness is not evidence.
Space as arbiter (source depth)
Einstein showed matter warps geometry. String and M-theory invert the revenge: the number and topology of dimensions set which particles and forces can exist. Curl a tiny extra-dimensional space into a pretzel instead of a torus and electrons (hence humans) can vanish from the spectrum. The container becomes the law-giver.
M-theory, still poorly understood even relative to strings, pushes further: space and time may be approximations of something more basic. Laugh or scream about tax dollars; many physicists did both for years. Among particle theorists the programme is de rigueur even without a Nobel-style decisive experiment. Witten's Fields Medal (not a Nobel) marks the inversion: the theory itself is the experiment, and postdiction (deriving known particles from the structure) is toasted with tequila when it works.
No LHC discovery of superpartners or large extra dimensions has crowned strings. GWTC-4.0 (through O4a) finds no evidence for deviations from GR in parameterized tests; graviton-mass bound about 1.92×10⁻²³ eV/c² (90% credibility). The weird revolution remains unfinished, as Mlodinow warned.
Ten things I hate about your theory
In 1981 Feynman would tease John Schwarz in the Caltech corridors by asking how many dimensions he was living in that day. Schwarz had been a lecturer for nine years without tenure, working on a theory that had started in 26 dimensions and was down to 10, and that had at times predicted negative probabilities and faster-than-light particles with imaginary mass. Like the heroine of Ten Things I Hate About You, he listed its faults and loved it anyway, because he saw a mathematical beauty that could not be an accident, and because it addressed the problem that defeated Einstein: reconciling quantum theory with relativity.
Quantum mechanics arrived in 1925 to 1927 in two forms. Heisenberg, working closely from experimental data with Born and Jordan, built matrix mechanics (reinventing matrix multiplication without knowing it). Schrödinger, more mathematical, wrote his wave equation during a ski-resort affair with a woman whose name is lost. Dirac proved the two equivalent and extended the theory to special relativity. Einstein, a parent of both theories, saw the clash. General relativity keeps Newton's determinism; quantum mechanics abandons it. He spent his last thirty years trying to resolve it and failed.
Part V opens on tension: general relativity is classical and geometric; quantum mechanics is probabilistic and discrete in measurement. Einstein resists the latter's incompleteness claims; the community mostly moves on. The hate list (infinities, non-renormalisability of naive quantum gravity, time's role, information puzzles) sets up strings as one proposed peace treaty.
Mlodinow's title borrows a teen movie: a hate-list that is really a love poem. John Schwarz's Caltech years (lecturer, not yet tenured, twenty-six dimensions, colleagues chuckling) fit the joke. Early string theory carried "embarrassments": negative probabilities, tachyons with imaginary mass. Schwarz and a few allies kept going because the mathematics looked too inevitable to be accident.
The hate-list in the chapter is diagnostic, not dismissal. Extra dimensions without a detection plan; a string vacua of vacua; a graviton mode that is also a feature; difficulty extracting Standard Model detail; years without decisive experiment. Beauty is a clue, not a proof. The poem's turn: those same features are why the theory refuses to die. As of October 2026 the experimental silence continues; the mathematical pressure (anomaly cancellation, dualities, AdS/CFT) also continues.
GR-vs-QM-LIST. Beautiful gravity, ugly mismatch with quantum rules; the hate-list is a research programme.
Write your own five-item hate list about a theory you use at work (what it ignores, where it breaks). Circle one item that is a real research problem, not a vibe.
Hate lists without magnitudes become aesthetics. "I hate that GR ignores quantum" is idle unless you name regimes (singularities, Planck scale, black hole entropy).
Schwarz, Feynman, and the hate-poem (source depth)
1981, Caltech corridor: Feynman ribs Schwarz, "How many dimensions are you in today?" Feynman thought strings nuts; Schwarz was used to it. A graduate student told young faculty member Mlodinow that Schwarz was "a lecturer, not a real professor," nine years without tenure, working a "crazy theory in twenty-six dimensions." By then the count was already down to ten, still "a few too many."
Embarrassments along the way: negative probabilities; tachyons with imaginary mass traveling faster than light. Career cost was real. The teen movie Ten Things I Hate About You supplies the metaphor: a hate-list that is a love poem. Schwarz stuck because of mathematical beauty that could not be accident, aimed at the problem that stumped Einstein: reconcile quantum theory with gravity.
Quantum mechanics took decades after Planck (1900) to arrive as Heisenberg/Schrödinger (1925-26); Dirac folded in special relativity and shared Nobel recognition (1932-33). Einstein spent his last thirty years seeking a unified field theory that would also tame quantum indeterminacy; he failed. About thirty years after Einstein's death, Schwarz believed he had the answer.
Stick with an ugly-looking consistent structure for a decade while peers collect tenure, or chase fashion. Schwarz chose the structure. The chapter's moral is not "beauty always wins," but that some programmes only look inevitable in retrospect.
The necessary uncertainty of being
Every measurement has error, and every McDonald's quarter-pounder weighs a slightly different amount. That is not the uncertainty principle. Heisenberg's principle says certain quantities come in complementary pairs, such as position and momentum, and the more precisely one is fixed the less precisely the other exists: beyond the limit the value is undetermined, not merely unknown. The product of the two uncertainties cannot be smaller than about Planck's constant, roughly 10⁻²⁷ in everyday (erg-second) units.
- A 1-gram ping-pong ball known to be moving slower than 1 cm/s can still be located to within about 10⁻²⁷ cm. Nobody notices.
- An electron has a mass of about 10⁻²⁷ grams. The same speed precision limits its position to about 1 cm.
- Confine an electron to an atom (about 10⁻⁸ cm) and its speed becomes uncertain by about 10⁸ cm/s, comparable to the speed itself.
Were there hidden variables that fix the values behind the scenes? John Bell showed in 1964 how to test that, and an experiment in 1982 ruled the idea out: the limit belongs to nature, not to our instruments.
In 2015 several groups, including Anton Zeilinger's in Vienna with entangled photons, reported Bell tests that closed the main remaining loopholes at once (locality, detection efficiency, freedom of choice). The 2022 Nobel Prize in Physics went to Alain Aspect, John Clauser and Anton Zeilinger for these entanglement experiments, which now underpin quantum cryptography and quantum computing research.
Heisenberg, Born, Jordan (matrix mechanics) and Schrödinger (wave mechanics) build quantum theory in 1925-27; Dirac bridges to special relativity. General relativity remains outside. Einstein's "God does not play dice" feud with Bohr is emblematic. Space in QM is usually a classical stage; that stage is what quantum gravity must renovate.
Heisenberg uncertainty is not ordinary ignorance. A McDonald's quarter-pounder might skim a hundredth of an ounce; that is measurement noise. Quantum uncertainty is structural: conjugate variables (position and momentum; energy and time) cannot be sharp together, no matter how good the apparatus. Einstein's "God does not play dice" refuses that structure; Bohr accepts it as completeness, not defect.
Mlodinow threads the joke about a flawed deity through the physicist's discomfort. An imperfect Nature bothers people who want Euclid-like certainty in the world. Quantum field theory later makes particles temporary excitations; string theory tries to push the uncertainty below a length where point-particle locality fails. Geometry's stake: if space is discrete or stringy at short distances, classical manifolds are effective, not final.
ħ-LIMIT. Uncertainty principle; determinism yields; geometry of phase space grows fuzzy.
State Δx Δp ≥ ħ/2 in words for a particle in a box. What dies: exact point trajectories or predictive science? Write one sentence.
Pop-culture uncertainty ("everything is uncertain") is false. Quantum theory is sharply predictive for distributions. The necessity is about conjugate variables, not vibes.
Heisenberg is not kitchen noise (source depth)
Indeterminacy's source is the uncertainty principle: conjugate quantities cannot all be sharp. Mlodinow's golf joke (rabbi cursing, lightning that misses and hits the nun) frames a flawed deity; that imperfect Nature bothered physicists who wanted Euclid-like exactness in the world. Einstein: the theory "hardly brings us close to the secrets of the Ancient One"; "He does not play dice."
Everyday uncertainty (McDonald's quarter-pounder off by a hundredth of an ounce) is measurement noise. Heisenberg's kind is structural. A Ping-Pong ball's speed known to ±1 cm/s still leaves position nearly sharp; the same speed uncertainty for an electron forces position uncertainty of order ±1 cm. Inside an atom (~10⁻⁸ cm), velocity uncertainty becomes comparable to the velocity itself (~10⁸ cm/s). Atomic ping-pong would be unplayable.
Apply the principle to gravity and geometry itself starts to look shaky at short distances: that is the bridge to the next chapter's clash.
| Object | Speed ±1 cm/s implies position Δ roughly |
|---|---|
| Ping-Pong ball | negligible for play |
| Electron | ~1 cm |
| Electron confined to atom | Δv ~ velocity itself |
Clash of the titans
The conflict shows up only at the Planck length, about 10⁻³³ cm. (Blow the Planck length up to the size of a human egg cell, and a marble would become as big as the observable universe.) Mlodinow stages it as a dream: Nicolai as Einstein insists that tiny regions of space are nearly flat; Alexei as Heisenberg replies that the gravitational field, like every field, fluctuates, so at small enough scales curvature goes wild and virtual black holes froth in and out. "Are." "Aren't."
Which theory gives way? Quantum theory agrees with experiment to better than a part in a million; quantum electrodynamics is "the best tested theory on earth." General relativity's triumphs are all large-scale; it has never been tested at the scale of particles. So physicists question Einstein's assumption that space is smooth at the smallest scales. The proposed remedy: extra dimensions curled up at that scale, an idea Einstein himself had considered.
The Einstein-Bohr debates sharpen what quantum theory claims about reality. EPR (1935) pushes entanglement's nonlocality; Bell (1964) and experiments decide against local hidden variables of the Einstein hope. For Mlodinow the clash matters because quantum gravity inherits both titans' demands: relativistic causality and quantum completeness.
Einstein wants a complete, deterministic description; Bohr answers with complementarity and a finished quantum formalism. EPR (1935) tries to show incompleteness via entangled correlations; Bohr replies; decades later Bell inequalities and experiments rule out large classes of local hidden variables. The titan fight is not fan fiction for team Einstein or team Bohr. Quantum gravity inherits both demands: relativistic causal structure and quantum completeness. Mlodinow parks the fight here as the doorway to speculative unifications.
EINSTEIN-BOHR. Determinism versus complementarity; EPR; titans agree on rigor, not ontology.
Steelman Bohr in four sentences, then Einstein in four. Do not pick a winner. That is the clash without fan fiction.
Turning the debate into team sports ("Einstein wrong, Bohr right") skips later work: Bell, Aspect, loophole-free tests. History continued past both titans.
Planck length versus Euclid's points (source depth)
The GR/QM clash shows only in regions so small we still cannot watch them: the Planck length ~10⁻³³ cm. Expand that length to a human egg cell's diameter and a marble balloons to the size of the observable universe. Generous versus a mathematical point, tiny versus every lab.
Mlodinow's dream-skit: Nicolai-as-Einstein claims empty space is flat; Alexei-as-Heisenberg answers that the gravitational field fluctuates, so curvature fluctuates, so tiny regions cannot be flat. At Planck scale the fluctuations are "humongous." GR on particle scales is untested (masses too small); physicists prefer to revise Einstein's approximate flatness assumption rather than throw out quantum rules that work elsewhere.
The remedy Schwarz loved and Feynman mocked: extra dimensions curled so small they were hitherto undetected, an idea Einstein himself once considered and abandoned.
FLAT-UNTIL-FLUCTUATION. Classical GR assumes tiny patches are nearly Euclidean; quantum gravity asks whether that assumption survives the Planck cut.
A message in a Kaluza-Klein bottle
In 1919 Einstein received a letter from Theodor Kaluza, an impoverished Königsberg lecturer paid by a formula of 5 x students x hours marks a semester. Kaluza had rewritten general relativity in five dimensions. In four dimensions the metric has 10 components; in five there are more, and the equations for the extra components turned out to be Maxwell's equations of electromagnetism. "The formal unity of your theory is startling," Einstein replied, and helped publish it in 1921.
Why do we not see the fifth dimension? Kaluza said it is tiny and curled into a circle, so every point of space carries a minute loop, the way a line becomes a thin garden hose. In 1926 Oskar Klein combined this with quantum theory and estimated the loop's size at about 10⁻³⁰ cm. The theory then made a wrong prediction (the electron's mass-to-charge ratio), and interest faded; Einstein last worked on it in 1938. Kaluza, with Einstein's help, finally became a professor in 1929. Extra dimensions were not taken seriously again until the 1970s.
Theodor Kaluza (1921) and Oskar Klein (1926) show that a five-dimensional metric can package four-dimensional gravity plus electromagnetism if the fifth dimension is compact and small. The message in the bottle: forces might be geometry in disguise if space has hidden shape. String theory will open the bottle again with more dimensions and consistency conditions.
Einstein reads a proposal that unites electricity with gravity by one quirk: a five-dimensional cylinder world. Extend the metric; the extra g-factors reproduce Maxwell's equations. Kaluza (and later Klein) compactifies the fifth direction to a tiny circle so we do not notice it when we "jiggle." Geometry manufactures a force.
The idea sleeps, wakes in higher-dimensional theories, and becomes everyday language in string compactifications. The bottle metaphor warns that messages from extra dimensions arrive compressed. Modern constraints (no large extra dimensions at millimetre scales from torsion-balance tests; collider bounds on Kaluza-Klein resonances) squeeze the naive versions. The durable insight remains: force content can be geometry in disguise.
5TH-DIMENSION. Kaluza-Klein: gravity plus EM from higher-dimensional metric; tiny circle we do not see.
Explain to a friend how a hose looks 1D from afar and 2D up close. That is compactification as cartoon.
Kaluza-Klein is not already a Theory of Everything. Extra dimensions without stabilisation, chirality tricks and quantum consistency are unfinished business; strings reuse the idea with new constraints.
Five-dimensional cylinder world (source depth)
The day before Einstein died he asked for his latest unified-field notes. Thirty years of failure; one early spark arrived in 1919 mail from impoverished mathematician Theodor Kaluza. Einstein replied that a "five-dimensional cylinder world never dawned on me" and that he "liked your idea enormously."
In n dimensions the metric has more independent g-factors (3D flat: 6; 4D GR: 10). Extend Einstein's equations to five dimensions and the extra g-factors reproduce Maxwell's equations. Geometry manufactures electromagnetism. Klein later compactifies the fifth direction to a tiny circle so we do not notice the jiggle.
Kaluza at thirty-four had been a privatdozent in Königsberg for ten years: salary 5 × x × y gold marks per semester (x students, y hours/week). Ten students, five hours: 100 marks/year. Einstein called the conditions schwierig ("only a dog should live that way"). With Einstein's help: professorship at Kiel (1929), Göttingen (1935). Interest in new dimensions slept until the 1970s.
Theodor Kaluza: letter that united Maxwell with Einstein's metric via one extra dimension; career rescued late. Oskar Klein: compactification that hides the fifth direction.
The birth of strings
The story begins at a summer school in Erice, Sicily, in 1967, where graduate student Gabriele Veneziano heard Murray Gell-Mann describe regularities ("dualities") in collisions of hadrons, the strongly interacting particles. The fashion was the S-matrix approach (from Heisenberg, Wheeler and Geoffrey Chew): treat collisions as a black box, like studying cars by crashing them, and see how far general principles alone can constrain the output. After a year and a half Veneziano found that a single function, Euler's beta function, had all the required properties. Why should nature's scattering be so simple? In 1970 Nambu, Nielsen and Susskind opened the box: the formula describes particles as tiny vibrating strings. Mlodinow compares this to Planck's quantum: a concoction built to fit, which might be a new principle or just a trick.
Late 1960s dual resonance models (Veneziano and others) morph into string theory. Points become loops or segments; vibrational modes look like particles. Early versions require extra dimensions and include a massless spin-2 mode: a graviton invitation. Scherk, Schwarz and others reframe strings as quantum gravity plus unified forces, not merely hadron spectroscopy.
Late-1960s dual resonance models (Veneziano and others) aimed at the strong force. The amplitudes behaved as if states were vibrations of a relativistic string. Point particles became extended objects; a massless spin-2 mode looked like a graviton. When QCD won the strong-force war, strings needed a new job. Scherk, Schwarz and others reframed them as a candidate for quantum gravity plus unification. Extra dimensions and consistency conditions (anomaly cancellation) came with the territory. Birth stories that skip the hadron-era dead end invent destiny; the real story is a theory that changed careers.
DUAL-RESONANCE-STRING. Strong-force dual models become vibrating strings; reinvented as quantum gravity candidates.
List two reasons a 1D string softens point-particle infinities (extended interaction, mode spectrum). One paragraph, no Lagrangian.
Birth stories that skip dead ends (strings as failed strong-force theory) invent destiny. The strong force went to QCD; strings survived by changing jobs.
Erice, Veneziano, and the S-matrix (source depth)
Summer 1967, Erice, Sicily, 750 feet above the Mediterranean: Centro Ettore Majorana summer school. Graduate student Gabriele Veneziano (Weizmann) hears Murray Gell-Mann on S-matrix regularities. Gell-Mann would soon win the Nobel for quarks inside hadrons.
S-matrix ("scattering"): smash particles and read debris, like studying cars by crashing them, except a Chevy-Ford smash can spray Jaguar parts. Heisenberg invented the idea; John Wheeler (1937); Geoffrey Chew championed it in the 1960s. Veneziano found a beta-function formula that matched dual resonance patterns; Nambu, Nielsen, Susskind reinterpreted the amplitudes as vibrating strings.
Bosonic string theory was a concoction with unreal features, yet also a stumble onto a possible new principle, the way Planck stumbled onto quanta. From Planck to Heisenberg/Schrödinger took ~25 years; strings have long since passed that benchmark without a final verdict.
1937 Wheeler pushes S-matrix language 1956 Gell-Mann presents related ideas at Rochester 1967 Veneziano at Erice; dual amplitudes ~1970 Nambu et al.: particles as strings 1984 Green-Schwarz anomaly cancellation (first superstring revolution)
Particles, schmarticles
The Standard Model grew from the post-1932 particle explosion (positron, neutron, then muons, pions and dozens more; Fermi said that if he could remember all their names he would have been a botanist). Its quantum field theories describe forces as exchanges of messenger particles, like two basketball players passing a ball: photons for electromagnetism (QED, 1940s: Feynman, Schwinger, Tomonaga), W and Z for the weak force, gluons for the strong force. It is a triumph of prediction, and not pretty.
- Three families of four matter particles each, identical except for mass. Why three?
- Nineteen parameters put in by hand: masses, force strengths, mixing angles, such as a fine-structure constant of 1/137.035997650.
- Mlodinow's jab: "You ever hear of a guy named Ptolemy?" With enough adjustable circles, any data can be matched.
String theorists want to derive all of this. In their theory there is one kind of object, a string about 10⁻³³ cm long, made of nothing finer. Like a guitar string, it vibrates in harmonics, the very ratios the Pythagoreans found, and each vibration pattern appears to us as a different particle. More energetic vibrations look heavier. Strings can be open or closed, and their splitting and joining is what we see as particle interactions. Detecting one directly would need an accelerator somewhere between galaxy-sized and universe-sized.
How a string can vibrate depends on how many dimensions it moves in and on their topology. A string on a cylinder has winding modes a string on a plane does not; a donut with more holes gives a different spectrum. The theory requires six extra dimensions curled into Calabi-Yau shapes (identified in 1985). Each "hole" in the shape yields a family of particles, which explains why families exist at all. The catch: there are tens of thousands of Calabi-Yau shapes, most with more than three holes, and no principle yet picks the one that gives our world.
If particles are string modes, the zoo becomes a spectrum. Supersymmetry (SUSY) pairs bosons and fermions and helps consistency. Mlodinow's playful "schmarticles" jab keeps the reader from confusing nomenclature with understanding. The serious claim is structural unity under geometry-plus-vibration.
Quantum field theory lets particles appear and disappear; quantum mechanics alone was built for fixed particle number. Interactions become exchanges of messenger particles: photons for electromagnetism (the basketball passed between players), gluons for the strong force, W and Z for the weak. QED, built in the 1940s by Feynman, Schwinger and Tomonaga, became the template; the 1970s electroweak unification and QCD completed the Standard Model's force sector.
The classification works as prediction and fails as beauty. Matter particles come in families; nineteen underived parameters (masses, mixing angles, coupling strengths) must be inserted by hand. Fermi, who already hated remembering particle names, would have hated the parameter list more. String theory's sales pitch in this chapter is structural: vibrational modes of one object replace a botanical zoo. Supersymmetry pairs bosons and fermions and helps consistency; as of October 2026 it remains unseen at collider energies. Mlodinow's "schmarticles" jab warns against mistaking a renamed zoo for understanding.
MODES=PARTICLES. Different vibrations, different particles; "schmarticles" jokes at zoo mentality.
Map three musical harmonics to three metaphorical particles. Then ask what experiment would falsify your map. That is the joke and the standard.
Mode poetry without coupling constants, symmetries and spectra is branding. A theory of everything that only renames the zoo has not earned the title.
Nuclear democracy to Calabi-Yau families (source depth)
A decade before strings, Geoffrey Chew declared field theory no good: no elementary particles, only particles composed of each other ("nuclear democracy"). He hoped one consistent S-matrix would fix the universe uniquely. Witten later called S-matrix "an approach, not a theory"; Gell-Mann said the name was pompous yet "still used today for string theory."
1932 shocks: positron and neutron. Dirac first called the positron a proton-lite; the neutron was imagined as a tight proton-electron hug. The particle zoo exploded; the Standard Model later fitted masses and charges "by hand."
String vibrations on compact spaces with holes yield particle families. Good news: families are derived, not pasted. Bad news: tens of thousands of Calabi-Yau candidates; most have more than three holes while nature shows three families. No one has found the unique space that yields the Standard Model, nor a principle that picks it. Critics are quieter than when string work was "the kiss of professional death."
Vacua-count size undercuts uniqueness rhetoric. "It must be true because beautiful and unique" collides with "tens of thousands of vacua." Hold both the derivation hope and the selection problem.
The trouble with strings
The original bosonic string was consistent only in 26 dimensions (it needs 1 - (D - 2)/24 = 0). It also predicted tachyons, negative probabilities and only bosons, the force-carrying particle type, whereas matter is made of fermions, which, unlike bosons, cannot share a quantum state (that is why atoms have a periodic table).
| Year | Step | People |
|---|---|---|
| 1971 | A string theory of fermions via an early form of supersymmetry | Pierre Ramond |
| 1971 | Spinning strings: fermions and bosons, no tachyons, 10 dimensions | André Neveu and Schwarz |
| 1972 | Gell-Mann hires Schwarz at Caltech, sight unseen, after Princeton denies him tenure | Gell-Mann |
| 1974 | Shrink strings to the Planck length and an unwanted particle becomes the graviton: strings include gravity | Joel Scherk and Schwarz |
| 1976 | Superstrings | Scherk and others |
| 1984 | Anomalies cancel: the first superstring revolution | Michael Green and Schwarz |
For a decade almost no one cared. The Standard Model was winning (W and Z bosons, 1983), and the mathematics was very hard. Scherk suffered a breakdown and took his life in 1979. Schwarz, on rolling research appointments, was compared to his adviser Chew, who spent fifteen lonely years on the S-matrix before giving up. Gell-Mann kept "a nature reserve for endangered species" at Caltech. When Green and Schwarz announced the anomaly cancellation in Aspen in 1984, a skit ended with Schwarz carried off by men in white coats. This time Witten called, and within months he and others had linked the extra dimensions to Calabi-Yau spaces. Hundreds of physicists joined. Schwarz got tenure after twelve and a half years. Witten: "Without John Schwarz, string theory would very probably have become extinct."
Critiques: energy scales out of reach; many vacua ("vacuum-space"); historic ties to unobserved SUSY. Defenders: anomaly cancellation miracles; AdS/CFT applications; mathematical payoffs. Mlodinow leaves the verdict open: Oresme waiting for Descartes, or Lorentz refining an ether?
Three troubles dominate Mlodinow's ledger. Scale: characteristic energies sit near the Planck regime, far above colliders. Vacuum space: many consistent-looking backgrounds make unique low-energy prediction hard. Null near-term tests: LHC SUSY searches through Run 2 and into Run 3 (still empty of significant excess as of October 2026) remove the easiest book-era hope; short-range gravity tests still see Newton down to tens of micrometres. Defenders point to mathematical necessity claims, dualities and AdS/CFT applications. The chapter refuses a verdict: patient Oresme, or Lorentz polishing ether?
NO-TEST-YET. Extra dimensions, string vacua, Planck-scale physics; beauty without near-term teeth.
Write one near-term observation that would raise your credence in strings and one that would lower it. If both lists are empty, you are doing aesthetics.
"Not even wrong" as a slogans kills curiosity and also kills standards. Demand routes to constraint (cosmology, tabletop gravity, dualities to calculable QFTs) without requiring a cartoon accelerator.
D=26, tachyons, and tenure (source depth)
Consistency with relativity forced a factor [1−(D−2)/24] = 0, hence D = 26 spacetime dimensions. Kaluza's five suddenly looked stingy. Further faults: negative probabilities; tachyons; extra unobserved particles. Weather forecast analogy: −50% chance of rain falling up, yet the temperature matches, so you keep the model for the hadron fits.
Worse for matter: bosonic strings lacked fermions. Electrons and quarks need fermion statistics (Pauli exclusion) or chemistry collapses. Supersymmetry and superstrings brought fermions and cut dimensions toward 10. Green and Schwarz anomaly cancellation (1984) sparked the first superstring revolution; hundreds piled in. Witten and co-workers identified Calabi-Yau candidates for the curled dimensions.
Schwarz's tenure fight: twelve and a half years. An administrator: even if he invented sliced bread, people would say he did it at Caltech so no need to keep him. Gell-Mann pushed. Witten: "Without John Schwarz, string theory would very probably have become extinct."
Corridor jokes about dimension counts → Green-Schwarz paper → sudden offers from other universities → tenure at last. Social proof lagged mathematical structure by years.
The theory formerly known as strings
By the early 1990s the field had cooled; there were five different string theories, an embarrassment for something meant to be unique. Witten, who majored in history, worked on the McGovern campaign and entered Princeton physics without having taken a physics course, is often called the Einstein of his day; Mlodinow thinks him more like Gauss, reinventing geometry and reshaping mathematics.
In March 1995, at the University of Southern California, Witten argued that all five string theories are approximations of one theory, M-theory, in eleven dimensions. (The next speaker, Nathan Seiberg, said, "I should become a truck driver.") In M-theory strings are just one kind of brane: a soap film is a two-brane. Most radically, the coordinates of branes become matrices, which look like ordinary positions only when the objects are far apart. If so, space and time are not fundamental, the deepest change in the idea of space since Euclid. "M" stands for "mystery, magic or matrix," Witten said, later adding "murky."
In 1974 Stephen Hawking showed that quantum effects make black holes glow, so they have a temperature and an enormous entropy, a count of their internal states. But a black hole was supposed to have no internal structure: where are the states? In 1996 Andrew Strominger and Cumrun Vafa built certain theoretical black holes from branes, counted their states, and matched Hawking's formula exactly. It is the strongest internal evidence for the theory, though, as the author admits, a postdiction.
The book named two possible tests: supersymmetric particles at CERN's Large Hadron Collider, and deviations from Newton's inverse-square law at short distances, then tested only down to about a centimetre. Schwarz was unworried: "I believe we have found the unique mathematical structure that consistently combines quantum mechanics and general relativity," and he would not abandon it even without supersymmetry. Is he Oresme to Witten's Descartes, or are they Lorentz, building an elegant theory on a nonexistent ether? "Has not yet been revealed."
The 2001 claim that "most physicists believe string theory is correct" overstated it then and would not hold up now. Real gains came after the book: Juan Maldacena's 1997 AdS/CFT correspondence (a duality between gravity in one space and a quantum theory on its boundary, now among the most cited papers in physics) became a widely used tool, and string ideas feed mathematics and black hole physics. But a 2025 Quanta Magazine survey and the agenda of Strings 2026 describe the same open problems: no full definition of the theory, no confirmed prediction, and an enormous set of perhaps 10^500 possible vacua (each a different low-energy world) that makes uniqueness doubtful, which the swampland programme tries to narrow. Rival approaches (loop quantum gravity, asymptotic safety) are also unconfirmed. A fair 2026 summary: the leading candidate for quantum gravity, mathematically rich and empirically unproven. (Witten, the book says, was "currently at Caltech"; he was visiting, and has been Professor Emeritus at Princeton's Institute for Advanced Study since 2022.)
- Geometry now aims to explain matter: the number and shape of dimensions would fix particles and forces.
- Quantum uncertainty is real, not a gap in knowledge (Bell tests), and it collides with smooth space at the Planck length.
- Extra dimensions are an old idea (Kaluza, Klein) revived because they unify forces.
- String theory was kept alive by a few stubborn people and spread when an influential figure endorsed it.
- Mathematical beauty is evidence of a kind, but not proof; the theory's tests so far have come back empty.
M-THEORY-1995. Witten's unification of string versions; membranes; name in flux; tests still thin.
Diagram five string theories arrowing into M-theory as a hub. Label one duality on each arrow with a guess. Accuracy optional; structure not.
Renaming is not discovery. "Formerly known as strings" jokes about marketing; the physics still owes predictions or constraints.
Five theories, Witten, and M-theory (source depth)
Early 1990s cooldown: Los Angeles Times wondered whether string theorists should be paid to "pervert impressionable young students." Andrew Strominger: "There are some big problems." Embarrassment: five different fundamental string theories, not five Calabi-Yau choices. Unaesthetic. A ten-year dry spell; Schwarz had company and a new prophet.
Edward Witten: dominant figure after Gell-Mann/Feynman. Brian Greene: intellectual roots "end at Witten's feet." History major who outpaced physics majors; Mlodinow heard Brandeis professors say "You're smart, but you're no Ed Witten." Witten's dualities knitted the five theories into facets of M-theory (the "M" stays playfully undefined: mystery, mother, membrane…).
Desk-top torsion-balance tests (Stanford, Colorado) probe gravity below about 1 cm, hunting large extra dimensions. Schwarz: even without LHC supersymmetry he would not abandon the structure that uniquely combines QM and GR. Closing image: is Schwarz Oresme and Witten Descartes, or are they Lorentz polishing ether? Unrevealed. A generation already sees through his strings.
ATLAS/CMS Run-2/3 SUSY searches keep excluding light natural superpartners (gluinos/stops often pushed above ~1.5-2.5 TeV in benchmark slices) without killing the vacua programme. Short-range gravity experiments continue; no confirmed millimetre-scale deviation that would scream large extra dimensions. M-theory remains a research culture, not a textbook final theory.
Epilogue: the joy of understanding
He sees a direction in history: the parallel postulate could not be proved, so curved space was inevitable; relativity and quantum theory contradict each other, yet a third theory seems to contain both. When Hawking's calculation and Strominger and Vafa's unrelated string calculation agree, "does not this connection imply some deeper truth?" He leaves the question open and ends with thanks to Euclid, Descartes, Gauss, Einstein and, "maybe, time will tell," Witten, who had the joy of discovery and gave the rest of us the joy of understanding.
Through Euclid's window, each generation found that space is not a stage but an actor: first a subject of proof, then of number, then of curvature, then the shape of gravity, and perhaps, at last, the source of everything in it.
Mlodinow ends on affect: the pleasure of seeing why the world must be as geometry says, or why it might differ. The joy is real when it survives those duties.
We inherit Greek tools: exact reasoning and an eye for nature's aesthetic order. Through Euclid's window we reach stars and atoms he never imagined. Mlodinow ends on three nested miracles: that regularities exist; that we can discern them; that mathematics fits so well.
Is physical law the universe's own language or one descriptive system among many? The book refuses a final metaphysics. Gratitude runs Euclid → Descartes → Gauss → Einstein → (maybe) Witten and the unnamed shoulders under them. The joy of discovery for a few becomes the joy of understanding for the rest.
JOY>MASTERY. Understanding space is pleasure, not ownership; the next window may already be grinding.
Write one geometric idea in this book you can now teach in three minutes. Teach it to someone this week. That is the joy test.
Joy without criticism becomes fandom.
Synthesis
20 takeaways
- Geometry began as accounting (Nile taxes, canal volumes) and became mathematics when the Greeks asked why.
- Abstraction pays twice: one line models a field edge, a pyramid edge and a crow's path, so knowledge transfers.
- Proof means explicit definitions, explicit axioms and nothing else; one hidden false assumption ruins everything.
- Intuition and approximation mislead (the rope around the earth, Monty Hall, Curry's square).
- Anomalies can be banned, explained away or named. Naming √2 would have given the real numbers centuries early.
- Knowledge lives in institutions (the Museum, monasteries, universities) and dies with them.
- A grammar beats a dictionary: coordinates turned shapes into equations that can be manipulated.
- Graphs turn functions into shapes that the eye reads instantly.
- Useful tools often precede their justification (irrationals, Dirac's delta, quantum field theory).
- The distance formula is the seed of every geometry: change how distance is measured and you change space.
- Two thousand years of failed proofs each smuggled the parallel postulate in disguise.
- Changing one axiom yields new, consistent worlds, hyperbolic and elliptic.
- Curved spaces look flat up close, which is why the error went unnoticed so long.
- Curvature can be measured from inside, without stepping into a higher dimension.
- After 1900 axioms became choices, not truths, and which geometry is real became a question for physics.
- Formal systems have limits: Gödel showed that some truths cannot be proved within them.
- Null results are results: Michelson and Morley's "failure" killed the ether.
- Gravity is geometry: mass shapes space-time, and free bodies follow its straightest paths.
- Fear and authority delay ideas (Oresme, Descartes, Gauss) and generational turnover lets them win.
- Beauty is a guide, not a proof: string theory is elegant and, as of 2026, still untested.
Frameworks at a glance
| Framework | What it says | Where in the book |
|---|---|---|
| Axiomatic method | Define, assume explicitly, derive | Euclid, Hilbert |
| Hidden-assumption audit | Look for the "obvious" step that carries the conclusion | Ptolemy, Proclus, Thabit, Wallis |
| Name the anomaly | Extend the system instead of banning the fact | √2, irrationals, delta function |
| Dictionary vs grammar | Labels identify; relations let you compute | Descartes |
| Local vs global | What holds in a small region can fail at scale | Non-Euclidean geometry, relativity |
| Intrinsic measurement | Judge a system from inside, by its own distances | Gauss, cosmology |
| Model, not mirror | Any structure satisfying the axioms is a valid model | Poincaré disc, "men, women and beer mugs" |
| Relative consistency | A system is as sound as the system it maps onto | Beltrami, Hilbert |
| Equivalence principle | If you cannot tell two situations apart, treat them alike | Einstein |
| Postdiction risk | Fitting known facts is weaker than predicting new ones | String theory |
A closer look
- Holds up: the mathematics, the history of the parallel postulate, the story of relativity, and the central claim that each revolution redefined space. General relativity has since passed every test thrown at it, including direct detection of gravitational waves.
- Romanticised: Hypatia's death as the end of Greek learning; the Library's destruction; Eddington at Sobral. Historians tell these more carefully now.
- Small slips: Newton "inventing the sextant"; Witten "currently at Caltech"; Weisskopf listed as German.
- Overstated: "most physicists believe string theory is correct." In 2026 string theory remains a leading but unconfirmed candidate, with no supersymmetry at the LHC and no deviation in short-range gravity.
- Style: the jokes and family anecdotes make hard ideas memorable, though they sometimes flatten historical figures (the Romans, the Church) into cartoons.
Who's who
| Person | Era | Role in the story |
|---|---|---|
| Thales | c. 640 to 546 B.C. | First proofs; nature follows laws |
| Pythagoras | 6th century B.C. | Number as nature's language; √2 secret |
| Euclid | c. 300 B.C. | The Elements; axiomatic method |
| Eratosthenes | 3rd century B.C. | Measured the earth |
| Archimedes | 3rd century B.C. | Lever, buoyancy, proto-calculus |
| Hypatia | d. 415 | Last famous Alexandrian scholar |
| Oresme | 14th century | Graphs; Merton rule proof; relative motion |
| Descartes | 1596 to 1650 | Coordinates; analytic geometry |
| Wallis | 1616 to 1703 | Similar-triangle axiom; ∞ symbol |
| Gauss | 1777 to 1855 | Non-Euclidean geometry (unpublished); intrinsic curvature |
| Bolyai, Lobachevsky | 1820s to 1830s | Published hyperbolic geometry |
| Riemann | 1826 to 1866 | 1854 lecture; geometry of curved spaces |
| Beltrami, Poincaré, Klein | 1868 to 1880s | Models and consistency of non-Euclidean geometry |
| Hilbert | 1862 to 1943 | Modern axioms of geometry |
| Gödel | 1906 to 1978 | Incompleteness |
| Maxwell | 1831 to 1879 | Electromagnetism |
| Michelson, Morley | 1881, 1887 | Null ether experiment |
| Lorentz, FitzGerald | 1889 to 1904 | Contraction hypothesis |
| Einstein | 1879 to 1955 | Special and general relativity |
| Grossmann, Minkowski | 1908 to 1915 | Geometry of space-time |
| Kaluza, Klein | 1919, 1926 | Fifth dimension unifies gravity and electromagnetism |
| Veneziano, Nambu, Nielsen, Susskind | 1968 to 1970 | Birth of string theory |
| Schwarz, Scherk, Green | 1971 to 1984 | Superstrings; gravity; anomaly cancellation |
| Witten | 1951 to | M-theory; Fields Medal 1990 |
| Strominger, Vafa | 1996 | Black hole entropy from branes |
Glossary
Angular defect. In hyperbolic geometry, the amount by which a triangle's angle sum falls short of 180 degrees; larger triangles have larger defects.
Axiom (postulate). A statement accepted without proof as a starting point for a logical system.
Black hole entropy. A measure of a black hole's internal states, predicted by Hawking and reproduced by string theory for some black holes (Strominger and Vafa).
Boson / fermion. The two classes of quantum particles. Fermions (matter) cannot share a quantum state; bosons (force carriers) can.
Brane. In M-theory, an extended object of any dimension; a string is a one-brane, a membrane a two-brane.
Calabi-Yau space. A family of six-dimensional shapes in which string theory's extra dimensions may be curled up.
Coordinates (Cartesian). Numbers that locate a point relative to perpendicular axes.
Differential geometry. The study of curved spaces using coordinates and calculus, founded by Gauss and Riemann.
Elliptic geometry. Geometry with no parallel lines; triangle angle sums exceed 180 degrees (the sphere).
Equivalence principle. Uniform acceleration and a uniform gravitational field cannot be told apart locally.
Ether. The hypothetical medium for light waves, abandoned after relativity.
Euclidean geometry. Geometry satisfying Euclid's parallel postulate; triangle angle sums equal 180 degrees.
Geodesic. The straightest possible path in a space: a straight line in flat space, a great circle on a sphere, a worldline of maximum proper time in relativity.
Gödel's incompleteness theorem. Any consistent system rich enough for arithmetic contains true statements it cannot prove.
Hyperbolic geometry. Geometry with many parallels through an external point; triangle angle sums are less than 180 degrees.
Intrinsic curvature. Curvature measurable from within a space, by distances alone (Gauss).
Irrational number. A number that cannot be written as a fraction, such as √2.
Kaluza-Klein theory. A five-dimensional extension of general relativity in which electromagnetism emerges from a curled-up extra dimension.
Length contraction / time dilation. In special relativity, moving objects are shorter along their motion and moving clocks run slow, as measured by a stationary observer.
M-theory. The proposed eleven-dimensional theory of which the five string theories are approximations (Witten, 1995).
Metric. The rule (the g-factors) giving the distance between nearby points; it fully determines a space's geometry.
Parallel postulate. Euclid's fifth postulate; equivalently, through a point not on a line there is exactly one parallel (Playfair's axiom).
Planck length. About 10⁻³³ cm, the scale at which quantum effects on space-time become large.
Poincaré disc. A model of the hyperbolic plane inside a circle, with lines as arcs meeting the boundary at right angles.
Postdiction. Deriving a theory's agreement with facts already known, rather than predicting new ones.
Relative consistency. A system is consistent if another system it can be translated into is consistent.
S-matrix. A table linking incoming and outgoing particles in a collision, studied from general principles; string theory grew out of it.
Special / general relativity. Einstein's 1905 theory of space and time for uniform motion / his 1915 theory of gravity as space-time curvature.
Standard Model. The quantum field theory of the electromagnetic, weak and strong forces, with 19 adjustable parameters.
String theory. The theory that fundamental particles are vibrations of tiny strings, requiring extra dimensions.
Supersymmetry. A proposed symmetry pairing each boson with a fermion; not observed at the LHC as of 2026.
Topology. The study of properties preserved under stretching without tearing, such as the number of holes.
Uncertainty principle. The product of the uncertainties in complementary quantities, such as position and momentum, cannot be smaller than about Planck's constant.
About this book
Euclid's Window: The Story of Geometry from Parallel Lines to Hyperspace by Leonard Mlodinow, 2001. 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