The Fourth Deduction Often Turns Out to Be the First Principle

Asked to separate mathematics from physics, Cumrun Vafa did not reach for a wall. The goals differ: mathematics does not care to describe reality, and physics does. The thought processes overlap so much that drawing a line is hard. Mathematicians, he said, care about deductive reasoning. Physicists care more about how ideas support one another, and about the puzzles that appear when they do not. History has trained that preference. Start from a principle, deduce a third or fourth consequence, and later you discover that the consequence was the real principle. Starting there opens ideas the original path never reached. Revolutions in science, in his telling, often begin with that reversal.

Newton's *F = ma* is simple and elegant. Later, physicists found the abstract idea of a potential useful: take its gradient and you recover the force, even though you do not strictly need it. Euler and Lagrange then rewrote the whole of Newtonian mechanics. To send a particle from A to B, list every possible path, compute the action — the time integral of kinetic energy minus potential — and take the path that extremizes it. In the language of the day that looked like a step backward. Why tour the space of all paths to recover an equation you already had in one line? The two formulations were mathematically equivalent. Lagrange had only recast Newton.

In the twentieth century the recasting became the law. A quantum particle does not take one path. It takes all of them, each with an amplitude proportional to *e* to the *i* times the action. Sum the complex numbers, square the modulus, and you get a probability. Newton's law survives only as an average. "We should start there as the basis of the new law," Vafa said. That is why physicists refuse to freeze the order of a deduction. The refusal is not sloppiness. He pushed back on the cartoon of physicists saying random things unbacked by logic. They stay fluid about which idea is primary. They are careful about the relations among ideas, and they look for beautiful mathematics underneath a physical law.

Fridman asked whether mathematics, used as a tool for the physical world, can lead people off track. Vafa's answer was symmetric. Math can carry you away because a piece of it is too beautiful to leave unused. A long-held physical principle can do the same. The danger is getting stuck, not the origin of the idea. New mathematics, in his view, is no more hazardous than old physics. Keep an open mind about the role of math: neither antagonistic nor overwhelmed.

Beauty, though, is not optional. Vafa does not treat it as a taste that might lead you astray. It is a requirement for the principles of physics, fundamental rather than emergent. He quoted the mathematician G. H. Hardy: there is no permanent place for ugly mathematics. The same, he said, holds in physics. An ugly-looking principle is not the end of the story. Beauty is not sufficient — he will not stop at a theory because it looks pretty — but every good physical theory has to pass that gate. The standard comes partly from centuries of scientific experience and partly from a philosophical bet about what reality should look like. In principle the universe could have been ugly. Sample after sample has given theorists the confidence to keep looking for the beautiful one.

Human aesthetic sense, he added, already contains tools you need to solve hard problems. Symmetry is one of them, and so is the breaking of symmetry. You see it in painting, in faces, in biology, in chemistry, and in the laws of particles. Scientists rarely talk about that judgment among themselves, because it sounds philosophical. Vafa thinks good scientists are amateur philosophers who dislike the label. Modern science, he said, sometimes shies away from philosophy at its peril. Science owes philosophy a great deal. A working method follows from the same aesthetic: if idea A is deeply tied to idea B, and you find a cousin A-prime, you go looking for B-prime because the picture wants to be complete. That hunt has guided scientists for centuries. Over the long arc of history, he suspects, the philosophical flavor of the ideas will be remembered longer than the equations, which will be shown incomplete or wrong.

Almost everything the Greek philosophers said about nature was incorrect, Vafa said. The way they thought is still in use. They measured the Earth's radius from how a stick's shadow changed with latitude — "very brilliant," he said twice — and then asked why the Earth does not move. Looking around, nothing seemed to. Being spherical, it had to sit at the center of the universe; any direction of motion would break that spherical symmetry, so it stayed put. The science was wrong. The symmetry argument was beautiful. Aristotle pushed further. Why should the symmetric point be preferred? Put a person at the center of a circle of loaves and tell them to stay there forever. They will get hungry and walk toward one loaf, symmetry or not. "This idea of spontaneous symmetry breaking is something we just used today to describe many physical phenomena." The idea was thousands of years old. It had been applied to the wrong world.

A modern version uses a bowl. A circular bowl has its lowest point at the center; a pebble slides there because that is the lowest energy, and the symmetric point happens to be preferred. Raise a bump in the bottom and the lowest points form a circle. The pebble stops somewhere on that circle. The overall symmetry is still there; picking a point breaks it spontaneously. The same principle explains both staying at the center and leaving it. The difference is the energy landscape. The Greeks held a correct tool and used it on the wrong problem. Two thousand years later the tool became a core idea of modern physics.

Geometry is Vafa's favorite part of mathematics, as it was for the Greeks. They tried to attach Platonic solids to air and fire, a move modern physicists laugh at until they notice how close the instinct still is. In string theory the extra dimensions are tiny compact spaces with shapes. If those shapes enjoy the same discrete rotation symmetries the Greeks loved, and you quotient the space by them, forces emanate from the singular points — electric, strong, weak. Two spheres touching and shrinking to zero size give you strong forces; one sphere gives you weak forces; other arrangements unify them. You can geometrically design a space to produce a chosen particle content. The translation of physics into geometry is, he said, one of his favorite discoveries in the subject.

Maxwell collected Faraday's discoveries into equations and noticed they did not agree with each other. He could have shrugged: nature is electricity and magnetism, mathematics can wait. The inconsistency bothered him. He added one small term — today's Maxwell term — with no experimental motivation, only mathematical consistency. The equations became clean differential equations, and they implied electric and magnetic waves moving at a calculable speed. That speed was the speed of light. He was bold enough to say that light might be those fields in motion. He did not live to see the prediction confirmed. Mathematical beauty, in Vafa's telling, drove one of the important physical connections of the nineteenth century.

Einstein looked at Maxwell's equations and asked who measures that unique speed. Move, and the speed of light should change; the equations contain no such hint. Ether experiments, including Michelson and Morley, had failed to find a medium at rest. Einstein said everyone gets the same speed for light. The mathematics underneath special relativity is linear algebra you could teach in high school. The leap is not the math. The leap is swallowing a claim that sounds ridiculous on its face until you learn that the old intuition about absolute time is wrong. Time depends on who measures it; we miss that because we do not travel near light speed. Vafa, asked which of Einstein's famous papers was the biggest leap, picked special relativity. Once the speed of light is the same for everybody, length contraction and time dilation are already in the room, and general relativity is the next step rather than a separate miracle. Einstein himself named the elevator thought experiment — inertial mass and gravitational mass as the same thing — as his aha. From outside, Vafa still votes for the speed of light.

How do physicists replace a broken intuition? His favorite example is Galileo against Aristotle. Heavier objects should fall faster; that is the gut reaction. Galileo's experiment was not enough for his contemporaries, who wanted a reason. He used symmetry. Three identical bricks dropped from the same height fall together. Sliding them horizontally does not change the time. Bring two into contact: they are now twice the mass of the third, and you have already agreed they fall together. The clash of intuitions forces a replacement. Vafa does not think a physicist can live on formalism. You build new pictures the way Galileo joined the bricks, step by step, including in ten dimensions. Einstein was willing to use tensors and four-dimensional spacetime because physics pushed him there, just as Newton invented calculus because he lacked the tools. Math and physics, he said, keep reinforcing each other.

Quantum mechanics, in the path picture Vafa had already drawn, is the statement that a particle follows every trajectory with different amplitudes. Everyday scales hide the fuzziness; atomic scales make the electron a cloud around the nucleus. Fridman asked whether that is a tool or a description. Vafa thinks reality really is fuzzy at that level, and he does not think quantum mechanics is the end of the story. He is unhappy with the usual talk of measurement and collapse — who defines a measurement, where is the boundary — and he is not among the physicists who consider the problem solved. "I don't believe any of the laws of physics we know today are exactly correct." That does not bother him. The beauty of science is that it is not dogmatic, and that it encourages skepticism toward one's own work.

Dirac tried to join Schrödinger's equation to special relativity and, roughly, took a square root of Einstein's relation so that time evolution would be first-order. Square roots come with a minus sign. Pauli pointed out that the minus sign meant negative energies with no bottom. Dirac invoked Pauli's own exclusion principle: fill the negative-energy orbits so there is nowhere to fall. Pauli then asked what happens if you kick one of those occupied states upward. You get a particle with positive energy and the opposite charge of the electron. Dirac first hoped it was the proton, then had to admit the mass was wrong, and grew uneasy about his own equation. A few years later Carl Anderson found, in cosmic-ray photographs, a particle that curved the opposite way in a magnetic field with the electron's mass: the positron. Every particle, it turned out, has an antiparticle. The math had been smarter than the person using it.

Quantizing electric and magnetic fields everywhere was harder than quantizing a point particle. Feynman, Schwinger, and others built quantum electrodynamics, then the same quantum-field-theory formalism produced the other forces. Particles and forces became exchanges: the photon mediates electromagnetism, and similarly for the strong and weak forces. Bohr's particle-wave duality, Vafa said, was one of the biggest leaps quantum mechanics forced on physics. Duality would later become the name of the game in string theory. Gravity still sat outside. Quantizing Einstein's curved spacetime gives a graviton. Feynman tried scattering gravitons with the tricks that had worked for QED and got infinities he could not tame. Gravity is too weak for those quantum effects to show up in experiment even today, so the puzzle was left open. Gravity and quantum mechanics both exist in our universe. They have to fit. The particle description of quantum gravity fails. That is the door into string theory.

String theory says the fundamental objects are not points but one-dimensional loops of energy — open intervals or closed circles — with zero thickness, a tension, and no atoms inside them. Stretch them and they get heavy; let a circle shrink toward a point and they look, from far away, like the particles we already know. Each harmonic is a particle. The lowest states behave like photons and gravitons. The typical size, depending on the model, is on the order of 10 to the minus 30 centimeters, some 22 orders of magnitude below an atom. That tiny extent is what cuts off Feynman's infinities: when the diagrams reach the dangerous region, the string starts to oscillate and the answer stays finite. Vafa's visual instruction is literal. Imagine a string. Just do not give it thickness.

Nobody sat down to unify Einstein with quantum mechanics by replacing a point with a string. In the late 1960s Gabriele Veneziano was looking at hadron collisions that showed a duality the particle models of the day could not explain. He wrote a formula, built from gamma and beta functions, whose only job was to exhibit that symmetry. A few years later people noticed the intermediate channels were full of harmonics, tried scattering two strings, and recovered Veneziano's formula exactly. Gravity was still nowhere in the brief. Then anyone computing the spectrum of the string found a massless particle with the properties of the graviton. Hadronic physics had no such particle. Theorists tried to get rid of it and could not. Joël Scherk and John Schwarz changed the perspective: string theory is a theory of quantum gravity. They were no longer describing hadrons. They were describing something larger.

The mid-1970s were the golden age of quantum field theory and the Standard Model. String theory looked like a sideshow. It also demanded extra dimensions. The version with fermions — the superstring, so named because it carries supersymmetry, a pairing of bosons and fermions — wanted ten. You do not get to set *d* equal to four. The working patch was to curl six spatial dimensions into spaces too small for today's experiments to see. Conservative physicists, Vafa said, were entitled to roll their eyes: unseen dimensions, an unseen string, what next? Extra dimensions later stopped being an apology. Their geometry mirrors the particle content of the four-dimensional world; change the holes in the compact space and you change the particles. They also hid the microstates of black holes. Bekenstein had argued that a black hole must carry entropy proportional to horizon area, or the second law fails when you throw something in. Hawking fixed the coefficient at one quarter of the area in Planck units and could not say what the states were. In 1996 Vafa and Andrew Strominger counted strings wrapping the tiny extra circles and matched Hawking's number. "The extra dimensions became useful for resolving a puzzle in four dimensions."

How do you think in ten dimensions? Vafa's training method is to steal intuition from a world you can see. Two lines in a plane typically meet at a point: one plus one is two, two minus two is zero. A plane and a line in three-space meet at a point; two planes meet in a line. The pattern is: add the dimensions of the subspaces, subtract the dimension of the ambient space, and you have the dimension of the intersection. In ten dimensions a seven-plane and an eight-plane meet in a five-plane. You still draw two planes crossing. You write 7, 8, and 5 in the margins. The picture does not change. The words do. That, he said, is how you stop being scared of the number ten.

A sphere can be drawn as an interval with a circle attached at each point, shrinking to a point at the ends. From then on an interval *means* a sphere. The product of two spheres is four-dimensional: draw a square; two circles sit at each interior point; one shrinks on each edge; both shrink at the corners. Six dimensions is a corner of a room, a cube, three spheres' product — three undrawn circles, shrinking as you hit faces, edges, and the vertex. When a mathematician says they want to blow up a point, you walk to a corner of that cube and chop it off. The triangular face that appears is the blown-up point, projective two-space. "Four plus six is ten, the dimension of string theory. We can visualize that. No problem." Without that kind of intuition, he said, the research does not move. Formalism is not physics, and it is not understanding. He has met topologists who work almost entirely outside visual space. He cannot.

Fridman described a growing cynicism about string theory: forty years without a novel, lab-tested prediction. Vafa called the criticism healthy. There has been zero experimental evidence — nobody has seen that little loop of energy. If he claimed the theory could never be checked, critics would be right to say it was not science. Science needs experimental consequences. String theory has predictions. They sit at energies far above today's colliders. That is a technology lag, not a defect in the theory. Many string theorists, he said, keep working anyway, closer to the mathematician's attitude: wanting to know how nature works even if you cannot prove it this decade.

The harder reply is that there is theoretical evidence. String theory has connected pieces of physics it had no duty to connect. Particle-physics problems that do not mention gravity have been solved with gravity, through holography. The structure is richer than anyone designed. It arrived by accident, from Veneziano's formula. "Some people say it's not fair to say we have no evidence. Graviton, gravity, is that evidence. It's predicted by string theory. We didn't put it in by hand." It is a postdiction in the sense that gravity was already known, and still a prediction in the logical sense. Condensed-matter questions, particle-physics questions, and a great deal of new mathematics have come out of the same web. Beautiful mathematics is not proof of reality, he admitted, but physical principles we already trust tend to have it underneath them.

Why three plus one large dimensions? Drawing a number at random from one to infinity, you almost never land on three or four. String theory caps the count at ten or eleven, which suddenly puts three and four near the average of a compactification. Vafa called that a weak argument and still a reason to be surprised we live in so few dimensions. A good theory should explain it, not take it as input. With Robert Brandenberger he sketched a model: a tiny nine-dimensional box, strings winding its cycles, expansion stretching those windings until they become too massive to let the box grow. Unwinding requires a winding string to find its opposite. Strings are one-dimensional; one plus one is two, plus time is four. In four-dimensional spacetime they typically meet. In higher dimensions they miss. Only three spatial directions can expand; the rest stay locked small. The same arithmetic is why four-manifolds are the hardest for mathematicians to classify: surgery uses two-dimensional tubes, and two plus two is four, so they cannot pass through each other. He would not claim this proves string theory gives 3+1. He would claim the question cannot be left sitting there.

Fridman asked who the key figures are, and who would get a Nobel if string theory turned out right. Vafa preferred the first question. A prize to one person would miss the diversity of the subject. The early epoch runs from Veneziano through Yoichiro Nambu, Leonard Susskind, Michael Green, John Schwarz, Joël Scherk. In the mid-1980s Edward Witten became the major proponent, shifted what people did, and pulled the high-energy community toward unification; in the mid-1990s he was among the drivers of the duality revolution that led toward black holes, holography, and Juan Maldacena's correspondence. Vafa has had some 180 collaborators and listed them when he received the Breakthrough Prize. Science as a lone island with a medal, he said, would be much less satisfying. Humanity likes heroes. He tries to avoid the trap. If he were betting, he thought it possible that no living physicist would get the Nobel for string theory, and that somebody eventually would, once technology catches the evidence. He would rather see something like a "string theorists without borders" prize for the little calculations that never make the newspaper.

Witten was Vafa's PhD adviser at Princeton, a Fields Medalist who used physics to move modern mathematics. As a student Vafa had a double major in math and physics at MIT and was in pain over the tension: he loved mathematical rigor and disliked the missing rigor in physics, yet a theorem sitting by itself was not enough. He went to Princeton partly to rewrite quantum field theory in strict mathematical language, fell behind in classes, and did neither. Then he watched Witten work. Witten's mathematics was excellent. Physics, for him, was not a formal system. He drew Feynman diagrams and thought of the electron going over there. Asked about relativity, he thought in invariants, not in a particular frame's momentum. "Interconnections between ideas in physics replace mathematical rigor." You cannot always define a term. You can watch it lock to another piece of physics, and another, until they make an elegant story. That story's solidity, Vafa learned, can stand in for a proof. He says he may now have gone further than Witten. He uses physical intuition to define new mathematics, and he makes claims that take mathematician friends off guard.

The example is mirror symmetry. After Harvard, studying string compactifications on Calabi-Yau manifolds, he and colleagues noticed that the physics could compute only the absolute value of the Euler characteristic, not the sign. Either the two sides were indistinguishable, or every Calabi-Yau with a positive Euler number had a partner with the negative. He took the conjecture to Shing-Tung Yau, whose name sits on the manifolds. Yau said it did not look reasonable: known examples with negative Euler characteristic far outnumbered the positive ones. Vafa answered that physics could not tell them apart. A few years later mathematicians confirmed the mirror pairs. Physical intuition had gone against the mathematical evidence and won. Instances like that, he said, are another reason the community feels string theory is pointed in the right direction — theoretical evidence of a different kind from a collider.

Feynman's infinities suggested that field theory, gravity, and quantum mechanics do not go together. That cannot be the last word, because all three exist here. String theory supplied exceptions. The broader lesson, Vafa said, is that Feynman was essentially right: almost every quantum field theory collapses once gravity is turned on. Only rare cases survive, and string theories are among them. He calls the set of consistent quantum field theories a landscape in the loose sense, then reserves *landscape* for the tiny subset that can live with gravity and *swampland* for the rest. Swampland theories are not bad. They are excellent field theories that simply cannot be coupled to gravity. The ratio is absurd. A four-dimensional theory with maximal supersymmetry is fixed by a choice of gauge group; infinitely many groups, infinitely many field theories, and only finitely many compatible with gravity — those whose rank is less than 23. Rank above 23, infinitely many pretty theories, all swamp.

The consistent subset is still humongous, though he believes it is finite. That bulk is the core difficulty of the subject and the focus of outside criticism: we do not know which corner we live in, so we cannot make sharp predictions. The compensating move is to look for features every landscape point shares. One of them is the weak gravity conjecture: in any universe consistent with gravity, gravity is the weakest force. It is easy to write a field theory that violates it. Crank up the electron mass until gravitational attraction beats electromagnetic repulsion. In every landscape example Vafa knows, that never happens. Pattern-spotting is not enough. The reason comes from black holes. Throw a charged particle into a charged black hole and wait for evaporation. If the hole cannot emit particles whose mass is small compared with their charge, evaporation never finishes. Accepted black-hole evaporation therefore demands particles light enough that gravity is weaker than the other force they feel.

The same bound, read against our universe, is a wide window rather than a bullseye. In Planck units the electron mass should be less than one and, from a large-black-hole argument, greater than about 10 to the minus 30. The actual mass sits near 10 to the minus 20 — twenty orders of magnitude below the upper edge, ten above the lower. Vafa will call that a postdiction and still a consequence of principles now understood from string theory. Other links are sharpening. Neutrino masses, from swampland arguments, seem tied to dark energy, two quantities that look unrelated on a naïve view. None of this is yet a precise quantitative prediction. It is the direction he thinks the subject has to walk.

Fridman asked about loop quantum gravity, Eric Weinstein's geometric unity, and Stephen Wolfram's computational universe. Vafa's bet is on useful ideas, not on rivals standing as equals. Loop quantum gravity is not, in his view, a complete theory of gravity, but it has nuggets. He expects the true pieces to be absorbed into string theory the way eleven-dimensional supergravity was. For a while ten-dimensional strings and eleven-dimensional supergravity looked like a contradiction, and people assumed eleven had to be wrong. Thought experiments later showed that one of the string theory circles was smaller than anyone had noticed. From the eleven-dimensional angle the picture closed, and that became M-theory. String theory still has no final formulation. Good ideas from other paths can come in. He will not put them on the same level. T-duality is his exhibit: shrink a space below the Planck scale by a factor of ten and the physics is equivalent to a space ten times larger, an *L* to *1/L* symmetry. No experiment can tell the "true" size. There are two kinds of photons, one measuring *L* and one measuring *1/L*; when the space gets small you switch description. "No physicist would have been able to design a theory where making the space smaller is as if you are making it bigger." Loop quantum gravity does not have that. Some of its instincts about discrete space may still live in corners of topological string theory; Vafa has written papers in that direction.

Can physics help us understand life and consciousness, or is "theory of everything" already a boast? Vafa's definition of physics is expansionist: any form of reality counts. He does not claim string theory currently has the tools for the origin of life, and he does not draw a border beyond which the subject may not go. Chemistry, he expects, will eventually reduce to quantum mechanics; the rules chemists use work, they have not been derived, and the delay may only be complexity — perhaps waiting on quantum computers. Biology sits in the same posture relative to chemistry. A field can refuse to reduce for now and still believe it will. Consciousness, if it needs a new ingredient, would then be a physical law that has to talk to everything else. Physics, he said, has never tolerated isolated compartments. Gravity, electromagnetism, the nuclear forces all speak to one another. Something real cannot sit in a different soup.

The analogy he likes is liquid and gas. Everyday language treats water and vapor as two states with a sharp change. In physics you can vary temperature and pressure and pass from one to the other without crossing a phase transition. There is no point you can put a finger on and say: here it is liquid, one step further it is gas. They look utterly different. The boundary is missing. He thinks many things we treat as sharp — consciousness, maybe life — might be like that. A living person and a dead one are not the same. There may be no threshold where consciousness switches on. The missing pieces might not be a boundary at all, but the unnamed analogues of temperature and pressure that would describe the phenomenon. Today consciousness and free will still live in philosophy. He does not think they always will, any more than biology stayed in natural history after DNA.

Vafa has three sons — one in biophysics, one in machine learning, one in theoretical computer science — and says he has tried not to give too much advice. Influence probably leaked anyway; none of them landed far from his world. He entered MIT planning to study economics and electrical engineering, found he was more passionate about math and physics, hesitated because those did not look like jobs, and followed the interest anyway. He calls himself lucky that society will support people doing abstract work that may never be experimentally verified or turned into technology in their lifetimes. "If people follow their interest they will find a niche that they're good at." Do not be driven by what makes a career or what parents expect. You have to live with yourself. You have one life, and it is short. He is getting there, he said, so he knows. Fridman noted how frightening it is when an interest maps to no existing job. Vafa agreed it is frightening, then said that when ability and passion line up, a path tends to appear. In a society like the United States, the floor is selling funny T-shirts.

He does not think much about death. He cannot do much about it, and it does not drive his days. Nature has an approximate time-reversal symmetry: forward is dying, backward is being born. He has no memory of birth, neither pain nor pleasure. Who knows what the moment of death feels like. Then a harder claim: mortality may be a blessing in disguise. Tell someone they are immortal and life becomes boring. Part of why we enjoy it is that it ends. Infinite time removes the reason to do anything today rather than in a billion years. Scarcity makes value, in a life as in an economy, and time is the resource you cannot print. The consolation he takes from science is a modest immortality: everyone dies, everyone before us died, and hanging on to a slightly deeper piece of an incomplete, improvable truth — a little Earth, a little galaxy, a little more order than naïve sight gives — makes living this life a bit more of a pleasure. That, he said, is how he views his role as a scientist.