DEEP READ · 112
Richard Feynman, with Robert Leighton and Matthew Sands · 1963–1965
On the surface, the three red volumes are a freshman–sophomore physics course. Underneath, they are a method for understanding nature: start from one or two rock-hard facts (everything is made of atoms, energy is conserved, experiment is the only judge), derive as much as possible from as few rules as possible, learn to see every law from several angles at once, and when something doesn't make sense, say "we don't know" rather than hide behind a name. What stays with you afterwards is less a set of formulas than a way of asking questions.
In 1961–63 Caltech asked Feynman, fresh from his work on quantum electrodynamics (QED) and two years from a Nobel Prize, to rebuild its introductory physics course from scratch. His colleagues Robert Leighton and Matthew Sands turned the tapes and blackboards into three volumes: mechanics, radiation and heat; electromagnetism and matter; quantum mechanics. As a textbook it was not much of a success; as the "second physics book" that working physicists keep going back to, it has been unrivalled for sixty years.
Chapter 1 opens with a famous thought experiment. If some cataclysm destroyed all scientific knowledge and only one sentence could be passed on to the next generation of creatures, which statement would carry the most information in the fewest words? Feynman's answer is the atomic hypothesis: all things are made of atoms — little particles that move around in perpetual motion, attracting each other when they are a little distance apart, but repelling upon being squeezed into one another.
The power of that sentence is that with a little imagination and thinking, half the world falls out of it. Feynman immediately shows how. Magnify a drop of water a billion times and you see molecules jammed together, tugging and shoving, jiggling without pause. From that picture:
How it changes the way you see: you start translating large-scale "properties" into small-scale "actions." "This is hot" becomes "its atoms are jiggling violently"; "equilibrium" becomes "coming and going at equal rates," not "nothing is happening." It is a habit of reading the large through the small: ask what the little things underneath are doing, and phenomena that looked unrelated suddenly join up. Chapter 3 has a line that captures the spirit: poets say science takes away the beauty of the stars — mere globs of gas atoms. But nothing is "mere" (paraphrased).
Chapter 4 introduces conservation of energy not with a definition but with a story about a child and his blocks. Dennis the Menace has 28 blocks, and every night his mother counts them: always 28. One day there are only 27 — one is in the toy box. Another day there are 25, and he won't let her open the box, so she weighs it and works out how many are inside. Another day some have been thrown into the dirty bathwater, so she measures how much the water level has risen... Her formula grows: visible blocks + (box weight gain ÷ weight per block) + (water rise ÷ volume per block) + … = 28.
The point of the analogy lies in its final twist: with energy, there are no blocks. All we have is a set of formulas — kinetic energy, gravitational energy, heat, electrical, chemical, mass energy — each producing a number, and when you add them up the total never changes. Feynman puts it bluntly: it is important to realize that in physics today, we have no knowledge of what energy is. It is not a substance or a fluid. It is an abstract bookkeeping number, and the miracle is that the books always balance.
Better still is what he does with the books. Assume there is no perpetual motion: no machine can keep raising weights higher on its own. Now consider ideal "reversible machines" that lower one weight to raise another. If some machine could lower 3 units of weight by 1 foot and raise 1 unit by more than 3 feet, hook it up to a reversible machine running backwards and you could lift weights for free — perpetual motion. So the most any machine can do is keep "weight × height" constant. The lever, the pulley, the inclined plane — things school teaches as separate chapters to memorize — all drop out of one sentence: there's no free lunch.
How it changes the way you see: faced with a complicated system, you can ignore the details of the process and just keep the accounts. Where does a braking car's motion go? Into heat in the brake pads. How fast is a diver going when she hits the water? The height tells you, however many somersaults she did on the way down. Conservation laws are a kind of lazy wisdom: the process can be infinitely complicated; the ledger is a single line. And when the books refuse to balance, that is usually the doorway to new physics. In the 1930s energy seemed to go missing in beta decay; rather than give up conservation, Wolfgang Pauli proposed an invisible particle that carried it away. It was eventually found: the neutrino.
What does it mean for a law to be symmetrical? Feynman borrows the mathematician Hermann Weyl's definition: a thing is symmetrical if there is something you can do to it so that, after you've done it, it looks the same as before. A circle looks the same however far you rotate it, so it is symmetric under rotation. Symmetry of physical laws means the same thing: move the whole experiment to the lab next door, run it a day later, turn it to face another way, and the results don't change.
Volume I, Chapter 52, "Symmetry in Physical Laws," ties this to conservation. Behind it lies a 1918 theorem by the mathematician Emmy Noether: every continuous symmetry corresponds to a conserved quantity.
| The laws don't change when you… | Conserved quantity | In plain words |
|---|---|---|
| shift in space (do it somewhere else) | momentum | the universe has no special place |
| shift in time (do it later) | energy | the laws are the same tomorrow as today |
| rotate (face a different way) | angular momentum | the universe has no special direction |
Table: symmetries and conservation laws (Noether's theorem). The ever-balancing ledger of idea 2 is rooted in the fact that the laws don't change over time.
Not every operation is a symmetry, though, and Feynman poses a lovely puzzle. Suppose we can talk to a Martian only by radio. How do we tell him which way is "right"? "The side opposite your heart" won't do — he may not have one. All of mechanics and electromagnetism look the same in a mirror, so in principle there's no way to tell. Then in 1956 Tsung-Dao Lee and Chen-Ning Yang suggested that the weak interaction might violate parity (mirror symmetry), and Chien-Shiung Wu confirmed it with the beta decay of cobalt-60: the electrons come out preferentially in one direction. Here, at last, nature does know left from right. So we can tell the Martian: run this cobalt experiment, and the side the electrons favor is… Feynman adds a punchline: if the Martian turns out to be made of antimatter, his experiment comes out exactly reversed, so if he holds out his left hand to shake, watch out (paraphrased) — matter and antimatter annihilate on contact.
How it changes the way you see: conservation laws stop being a list of separate facts and become shadows of the ways the world stays the same under change. When modern physicists hunt for new laws, they often begin by asking what symmetries those laws ought to have — the main thread of twentieth-century particle physics. And the places where a symmetry breaks are often more informative than the symmetry itself.
School physics teaches a local view: what force acts right now determines how you accelerate in the next instant, and you march forward step by step (Newton's second law). The Lectures keep returning to another, global view. The most intuitive case is light, in Volume I, Chapter 26. Why does light bend when it passes from air into water? Feynman's answer is Fermat's principle of least time: of all possible routes, light takes the one that takes the least time.
His analogy is a rescue. You are on the beach at A; someone is drowning at B, out in the water. You run fast on sand and swim slowly in water. Heading straight for her is not fastest — you spend too long swimming. Running along the shore until you're opposite her and then plunging in isn't fastest either — too much running. The quickest route runs a bit further on sand, then angles into the water, bending at the shoreline. Light is fast in air and slow in water, and the angle at which it bends at the surface is exactly the one given by Snell's law of refraction.
Schematic: the lifeguard and the light ray follow the same time-saving route — refraction is that bend.
Volume II, Chapter 19, is a special lecture in which Feynman recalls his high-school physics teacher, Mr. Bader, calling him over after class to tell him "something interesting": the path a thrown ball actually follows is the one that makes a quantity called the action — kinetic minus potential energy, added up over every moment of the trip — take an extreme value (usually a minimum). This is the principle of least action. It is mathematically identical to Newton's laws, yet it says something very different. Newton says "obey the force at each instant"; least action says "the whole path is the best among all possible paths."
Which raises an unnerving question: how does the particle "know" which path is best? Does it try them all first? Here Feynman gives the quantum answer: in a sense, it does. Every possible path contributes a little arrow with a phase. For paths far from the best one, the arrows of neighboring paths point every which way and cancel out; only near the best path do the arrows line up and add. So on a large scale the object seems to take just that one path. The classical "best path" is the quantum "all paths cancel except this one."
How it changes the way you see: you can reason about the same thing step by step (the local, causal language) or take in the whole trajectory at a glance (the global, optimizing language). The two descriptions are equivalent, but they take you to different places — Feynman's own reformulation of quantum electrodynamics came precisely from the global "sum over all paths" view. In any field, when one formulation gets stuck, switching to an equivalent one is often the way through.
The crucial move in Volume II is to get you to treat the field as something real, not just a bookkeeping device for computing forces. The electric field is an arrow attached to every point in space, telling you how a charge placed there would be pushed; the magnetic field likewise. James Clerk Maxwell's four equations describe how those arrows change in space and time — and one consequence is that a changing electric field creates a magnetic field and a changing magnetic field creates an electric one, the two handing off to each other and traveling through space at a fixed speed that turns out to be exactly the speed of light. Light is a wave in the electromagnetic field.
Feynman's verdict is one of the most quoted passages in the set: from a long view of the history of mankind — seen from, say, ten thousand years from now — there can be little doubt that the most significant event of the nineteenth century will be judged as Maxwell's discovery of the laws of electrodynamics, and the American Civil War will pale into provincial insignificance by comparison. The phone in your hand, the power grid, radio, radar and fiber optics are all descendants of those equations.
He also hammers home a remarkably useful meta-principle: the same equations have the same solutions. Chapter 12 of Volume II shows that the equations of electrostatics have exactly the same form as those for steady heat flow, for a stretched membrane pushed out of shape, and for certain kinds of fluid flow. Solve one and you've solved them all — the way heat "flows" outward from a source looks just like the way field lines spread out from a charge.
How it changes the way you see: you learn to look past the subject matter to the structure. Physicists cross into other fields not because they know more, but because the world has surprisingly few genuinely different mathematical structures — diffusion, oscillation, waves, exponential growth turn up again and again. Faced with an unfamiliar problem, first ask: which equation I already know how to solve does this look like? It is the most portable trick in the Feynman toolkit.
The first chapter of Volume III (quantum mechanics) walks the reader to the edge of a cliff with three experiments. A wall has two slits, with a screen behind it:
Feynman says this phenomenon is impossible, absolutely impossible, to explain in any classical way, and that it has in it the heart of quantum mechanics — in reality, it contains the only mystery. He doesn't try to "explain" it. Instead he gives three rules. The probability of an event is the absolute square of a complex number called the probability amplitude. When an event can happen in several ways that are indistinguishable, you add the amplitudes first and then square (so they can cancel — interference). When the experiment can in principle tell which way it went, you square each first and then add the probabilities (no interference).
Note that "can tell" has nothing to do with whether anyone is looking. What matters is whether any record of the path is left anywhere in the world — a single photon bounced off the electron is enough. This is also where Volume III departs from most textbooks: it doesn't open with wave functions and differential equations but with amplitudes and two-state systems (such as whether the nitrogen atom in an ammonia molecule sits "up" or "down"), going straight for the bones of the theory.
How it changes the way you see: the classical "or" — slit 1 or slit 2 — breaks down. When no record is left, "two possibilities" doesn't mean "one of them is true and we just don't know which"; the two possibilities interfere with each other. Feynman's most honest stance is this: don't ask "how can it be like that?", because nobody knows; what you can do is learn the rules exactly and test them against experiment — and the rules have never yet failed.
Starting point: experiment is the only judge; we are spectators at the gods' chess game, guessing the rules from the moves.
Layer one (matter): atoms in motion, attracting at a distance and repelling up close — heat, pressure, evaporation and phase changes all follow.
Layer two (laws): conservation laws are ledgers that ignore the process, and the books balance because the laws are symmetric under shifts in space, time and direction.
Layer three (formulations): the same law can be written as forces moment by moment or as the best path overall; the same equations have the same solutions across different fields.
Layer four (fields): fields are real, and the electromagnetic field relaying itself through space is light; one set of equations unifies electricity, magnetism and optics.
End point (quantum): the deepest rule is "add the amplitudes, then square"; classical forces, trajectories and best paths are its large-scale approximations.
So what the three volumes establish is this: nature's complexity comes from a few simple rules combined over and over, and the only way to understand it is to be able to re-derive those rules from several equivalent angles while letting experiment be the judge. They make the case not by listing results but by demonstrating, again and again, the live reasoning that gets from one sentence to a whole landscape of phenomena.
Misreading 1: "It's popular science — you can read it without the math." The tone is wonderfully conversational, but this is a real course: calculus, vectors and differential equations are all there, and Volume II in particular is hard going. Feynman's books for the general public are The Character of Physical Law and QED: The Strange Theory of Light and Matter, not these.
Misreading 2: "Least action shows that nature has purposes and picks the best path." Strictly, the action is only stationary (an extremum), not always a minimum, and the deeper reason is phase cancellation at the quantum level, not teleology. Reading it as "the universe has a will" mistakes an equivalent way of writing the mathematics for metaphysics.
Misreading 3: "The observer's consciousness collapses the electron." In the Lectures, what destroys interference is any physical interaction that distinguishes the paths — one photon will do, no human required. "Consciousness creates reality" is something added later, not something in the book.
Misreading 4: "The 'Feynman Technique' comes from this book." The popular four-step method of "explain it simply to a beginner" is a later internet packaging of Feynman's style. No such procedure appears in the Lectures.
Criticism 1: as a textbook, even Feynman thought it failed. In his preface he wrote that he didn't think he had done very well by the students, and that judging by how most of them handled the exam problems, the system was a failure — though friends pointed out that a dozen or two students had, surprisingly, understood almost everything (paraphrased). By participants' accounts, many freshmen fell away as the course went on and their seats filled with graduate students and faculty. It works best for people who have already learned physics once and want to understand it again.
Criticism 2: parts of it are frozen in the early 1960s. The lectures predate the quark model (proposed in 1964) and the Standard Model of particle physics, and quantum field theory is barely developed. The many errors in the original printings were systematically corrected only around 2010, in the "New Millennium Edition" overseen by Michael Gottlieb (now free to read on Caltech's website).
Criticism 3: it sidesteps the interpretation of quantum mechanics. Feynman's stance is close to "learn to calculate and don't ask why"; elsewhere he said he thought he could safely say that nobody understands quantum mechanics (paraphrased). Admirers call this honesty. Critics say it helped a generation of physicists lose interest in the foundational question of what quantum mechanics is actually saying — until work like Bell's inequality turned interpretation back into something experiments could test.
Summing over histories: the day job the Lectures only glimpse. The "little arrow for every path" at the end of idea 4 is really Feynman's own path integral formulation from the 1940s. From it he developed Feynman diagrams: pictures of lines and vertices showing how particles emit and absorb photons, each picture corresponding to a term you can calculate. They rescued quantum electrodynamics from a swamp of infinities and are the work for which he shared the 1965 Nobel Prize with Julian Schwinger and Sin-Itiro Tomonaga. His 1985 short book QED explains the whole scheme to non-specialists with "spinning stopwatch arrows" and almost no equations. In Volume III you see the rules; for Feynman, the path integral was where those rules come from.
"You must not fool yourself": from method to character. In his 1974 Caltech commencement address he described "cargo cult science." After the Second World War, islanders in the South Pacific imitated an airfield, with runways, fires along the sides and a man in a hut wearing wooden headphones. Every form was right, but the planes never landed. He used the image to skewer research that copies the forms of science while lacking a particular kind of utter honesty: going out of your way to report everything that might prove you wrong. The key line: the first principle is that you must not fool yourself — and you are the easiest person to fool. During the 1986 inquiry into the Challenger space shuttle disaster, he dropped a piece of O-ring seal into a glass of ice water at a public hearing to show that it lost its resilience in the cold. In the appendix to the report he wrote that for a successful technology, reality must take precedence over public relations, for nature cannot be fooled.
Taken together: "experiment is the sole judge," for Feynman, was not a textbook slogan but an ethic that ran through his research and his public life. Read only the Lectures and you may take him for a genius who happened to be a great teacher. Add these two pieces and you get the whole man: someone who invented a new way of calculating, and who never stopped guarding against fooling himself.
1. The test of all knowledge is experiment; experiment is the sole judge of scientific "truth."
2. If you could keep only one sentence, keep this: all things are made of atoms, in perpetual motion, attracting at a little distance and repelling when squeezed.
3. Heat is the jiggling of atoms, and equilibrium is not stillness — it is coming and going at equal rates.
4. We have no idea what energy is — only that there's a ledger that balances however you count.
5. Rule out perpetual motion and the laws of the lever, the pulley and the inclined plane come along for free: one prohibition is worth ten formulas.
6. Conservation laws are the shadows of symmetry: laws that don't change with time conserve energy; laws that don't change with place conserve momentum.
7. You can reason step by step or see the whole trip at once: light takes the quickest path; a ball takes the path where the action is stationary.
8. The same equations have the same solutions — solve one and you've solved a family.
9. The double slit has in it the heart of quantum mechanics; in reality, it contains the only mystery.
10. Poets say the stars are mere globs of gas atoms — but nothing is "mere" (paraphrased).