What this covers

David Tong and Sean Carroll explore unresolved puzzles within quantum field theory—the framework underlying all successful fundamental physics—arguing that meaningful progress remains possible even without new experimental discoveries. Tong, a theoretical physicist at Cambridge, contends that despite QFT's empirical success, it contains deep conceptual gaps that signal something physicists have yet to understand. The conversation ranges from specific mathematical obstacles like the Nielsen-Ninomiya theorem, which shows that the weak nuclear force's chiral gauge theory cannot be discretized onto a lattice, to broader puzzles about why the laws of physics require introducing unphysical objects to function.

The discussion centers on three interconnected themes. First, gauge redundancy—the apparent mathematical over-specification built into every major physical theory—remains fundamentally unexplained, yet it elegantly reconciles seemingly incompatible requirements like locality and unitarity. Second, the standard model's structure is far more constrained than it appears: chiral gauge theory consistency conditions force much of the particle spectrum, and Tong's work with Nakhirin Lohitsiri reveals that requiring the standard model's electric charges reduces to finding integer solutions to a Fermat equation. Third, duality and soliton phenomena—where the same physics admits competing descriptions with fundamentally different entities—suggest complementary formulations of nature may yet be discovered. The conversation also covers Tong's pedagogical project of writing foundational textbooks across theoretical physics, and both speakers debate whether the apparent slowdown in breakthroughs stems from lack of experimental guidance rather than methodological stagnation.

Sharpest takeaway

Tong argues that quantum field theory, despite underpinning all our successful fundamental physics, contains deep unresolved puzzles—chiral gauge theories that resist discretization, gauge redundancy we don't understand, and dualities/solitons—so meaningful theoretical progress remains possible even without new experimental input.

  • The weak nuclear force's parity violation (chiral gauge theory) cannot be put on a lattice per the Nielsen-Ninomiya theorem, which Tong finds suspicious and worth probing.
  • Gauge symmetry is really an unexplained redundancy in our description that lets locality and unitarity coexist.
  • Quantum field theory works empirically but is not mathematically well-defined, signaling something we don't yet understand.

The argument · threads16 threads · 36 claims
0.78

Quantum field theory describes all particles as ripples in underlying fields, and even quantum gravity appears to reduce to QFT at its core.

3 pointscentrality 5/5
  • Quantum field theory holds that what we call particles are not fundamental building blocks but are lumps of energy formed when an underlying field—a fluid-like substance permeating all of space—has a ripple in it, which the laws of quantum mechanics mold into a discrete lump of energy; this applies to all particles (electrons, quarks, neutrinos, W bosons, Higgs) just as photons are ripples of the electromagnetic field.

    “what we call particles, lumps of energy that we call particles, are ripples of an underlying field, waves of some underlying field, that this field is a fluid-like substance that spread throughout all of space”

  • String theory is the one theory of quantum gravity that demonstrably yields Einstein's general relativity compatible with quantum mechanics, and every way we know to understand it, it is actually just a quantum field theory—specifically a 1+1 dimensional QFT living on the string—suggesting that even quantum gravity's microscopic degrees of freedom may be QFT, though this could merely reflect that QFT is the only tool we have.

    “every way we know to understand string theory, it's actually just a quantum field theory. It's weirdly a quantum field theory that doesn't live in our three plus one dimensional world... It's one that lives in one plus one dimension. It's the theory that lives on a string.”

  • Our universe removes the infinities of quantum field theory not via a lattice but via quantum gravity, which provides the proper ultraviolet cutoff by smearing out physics at very small distances while preserving all the symmetries; from this perspective, the first reason to understand quantum gravity is that it is the right way to think about quantum field theory.

    “How does our world remove the infinities? Well, it removes it by quantum gravity... it must be that somehow gravity is smearing out what's going on on very, very small distances... It is the right UV cutoff for quantum field theory. It preserves all the symmetries.”

0.78

Gauge redundancy reconciles seemingly incompatible requirements but remains fundamentally unexplained by physicists.

3 pointscentrality 5/5
  • Gauge redundancy creates loopholes that reconcile seemingly incompatible features of physics: in quantum field theory you want locality (an action here doesn't affect the far side of the universe) and unitarity (probabilities sum to one), and there are descriptions where you can manifestly see one but not the other—the gauge redundancy is what allows both to hold simultaneously.

    “there are these two crucial features of the laws of physics, and you can see one of them or you can see the other one, but somehow it's very hard to see both at the same time. And it's this redundancy that allows you to see both.”

  • All current laws of physics—Maxwell's electromagnetism, Einstein's general relativity, quantum mechanics—are built on objects that cannot be measured even in principle (overall voltage, the metric's coordinate values, the wave function's overall phase); this is gauge symmetry, which is better called gauge redundancy, and physicists do not understand why the laws of physics require introducing these unphysical objects.

    “every single one of them is built on objects that we could not measure even in principle. And that's the essence of what we call gauge symmetry.”

  • The anthropic need for a massless spin-1 boson is distinct from and does not answer the question of why our description of that boson requires introducing unphysical gauge redundancy; the gauge field A-mu has components (like A0, the voltage) that are not physical yet are necessary to describe electromagnetism, and that necessity remains unexplained.

    “I agree that we need a massless spin-1 boson for anthropic reasons. It's not the same as saying why we need gauge symmetry. That's a choice of the way we've come up to describe the theory. I don't understand why we need to introduce that redundancy.”

0.64

Symmetry and conservation laws are intimately linked and central to discovering the laws of physics.

3 pointscentrality 4/5
  • Emmy Noether established about 100 years ago that there is a close relationship between symmetries and conservation laws, so everything conserved in our universe—electric charge, baryon number, momentum, energy—is conserved because of an underlying symmetry; this makes symmetry both fundamental and the key tool that lets physicists find patterns and figure out the laws of physics.

    “Emmy Noether told us that there is a close relationship between symmetries and conservation laws. And so anything in our universe that is conserved, electric charge, what's called baryon number... all of these things that are conserved are conserved because there's an underlying symmetry.”

  • Chiral gauge theories (theories that look different reflected in a mirror) must obey very subtle mathematical consistency conditions or they make no sense whatsoever; these conditions force much of the standard model's structure, so the seemingly arbitrary collection of quarks, leptons, and their weird charges fits together like a perfect jigsaw that could not be any other way, and the standard model is roughly the second or third simplest such chiral gauge theory.

    “There are these very subtle mathematical consistency conditions that chiral gauge theories have to obey. And it's the reason why the standard model is a thing of beauty... it's only when you appreciate that there's a consistency condition that suddenly the whole thing just fits together like the most perfect, beautiful jigsaw, and it could not be any other way.”

  • In a paper with student Nakhirin Lohitsiri, Tong showed that requiring the integer hypercharges of one generation of standard model particles to make the theory consistent reduces to finding integers satisfying X cubed plus Y cubed equals Z cubed (Fermat's last theorem); the only solution (1³ + 0³ = 1³) plugged back through the change of variables yields exactly the electric charges of the standard model particles.

    “we showed that in order for the theory to be consistent, you had to find three integers that obey the X, Y, and Z, that obey the following equation. X cubed plus Y cubed is equal to Z cubed. Now, that's rather famous. This is Fermat's last theorem.”

0.57

Solitons and dualities show the same physical system admits multiple competing descriptions in which fundamental and emergent roles reverse.

4 pointscentrality 3/5
  • The standard model contains no solitons except in one description: the theory of pions (states of quark-antiquark pairs) has a soliton which is the proton or neutron, so the matter we are made of can be viewed either as solitons of the pion field or, more conventionally, as bound states of three quarks—illustrating that the same system often admits competing complementary descriptions.

    “the theory of pions has a soliton, which is the proton and the neutron. So you can actually think of the stuff we're made of as solitons if you view it in a particular way. You could also view it as made of three quarks.”

  • Magnetic monopoles—a north pole without a south pole, forbidden by one of Maxwell's equations—are expected to exist somewhere in the universe because they arise as solitons in grand unified theories; they appear to be excluded by gauge symmetry (which is crucial), but Dirac found a loophole in the 1930s, so monopoles sit just on the cusp of being disallowed yet sneak through as permitted by the laws of physics.

    “if you have a grand unified theory where all the forces are unified into one, these magnetic monopoles arise as solitons... they would kill gauge symmetry that we talked about earlier... But the great physicist Paul Dirac back in the 1930s figured out a loophole to that argument.”

  • Duality is the phenomenon where the same physics admits two formulations in which what is fundamental in one is emergent in the other—e.g., in superconductors you can treat Cooper-pair electrons as fundamental and vortices as derived, or invert it so vortices are fundamental and electrons emerge from them; analogously there may be a formulation of the laws of physics written in terms of magnetic monopoles where electrons emerge as solitons, and this idea was central to the second superstring revolution.

    “You think of the vortices as being the fundamental objects that are moving around... and then the vortices themselves can start swirling. And when the vortices swirl, they give back the electrons. So there is this sort of complementary view... This is what quantum field theorists call duality.”

  • Quantum field theories produce particles two ways: ordinary particles arise when quantum mechanics molds a field ripple into a lump of energy, but solitons arise without quantum mechanics at all—when the field folds or wraps over on itself into a stable, non-dispersing lump of energy; examples include smoke rings, superconductor vortices formed by Cooper pairs resisting magnetic flux, and magnetic monopoles.

    “the field can sort of fold over on itself in a particular way, wrap around itself, and give rise to a lump of energy that doesn't disperse... It sits there in a stable way, even without invoking quantum mechanics. And that goes by the name of a soliton.”

0.56

Natural theory expectations like the hierarchy problem have proven wrong, signaling the community should explore many directions rather than converge.

2 pointscentrality 4/5
  • The naturalness argument (hierarchy problem) led most of the community, including Tong, to expect the LHC would find new particles accompanying the Higgs boson as a portal to the next layer of reality, but despite the LHC working beyond expectations and finding the Higgs, those particles are not there—so nature is telling us the well-motivated naturalness argument is wrong and the hierarchy problem has no solution.

    “There was an argument called the naturalness argument or the hierarchy problem. It's an argument that I entirely subscribe to... And yet nature is telling us it's wrong. That Higgs is not natural. That the hierarchy problem does not have a solution. I'm shocked by this.”

  • The history of physics shows everyone was confused at every step and great breakthroughs often happen by accident or for the wrong motivations, so without a strong guiding principle and absent experimental clues, the healthiest strategy is for the community to take more chances and explore many crazy different directions rather than converging on one program like supersymmetry or string theory.

    “everyone was confused at every single step of the way. And often the great breakthroughs happen by accident or they happen for the wrong motivations. And we do need experiments to guide us. We're not smart enough to do it on our own.”

0.43

Recent advances in symmetry understanding have deepened knowledge of known phenomena rather than revealing genuinely new physics.

2 pointscentrality 3/5
  • In the last decade quantum field theorists have realized there is much more to symmetry than previously appreciated, and although the mathematics is lovely and it explains long-standing subtleties (like why the pion decays to two photons and at what rate), it has mostly deepened understanding of things already known in the 1950s rather than delivering a genuine smoking gun—something we could not have understood without it.

    “rather surprisingly in the last 10 years or so, quantum field theorists have realized that there's much more to symmetry than we appreciated before... It would be nice to have a real smoking gun.”

  • The quantum field theory ideas originally developed in string theory are extremely useful and are the right way to think about condensed matter systems like topological insulators, fractional quantum Hall effects, and high temperature superconductors—incredible quantum many-body entangled systems that, while not fundamental physics, are important and as beautiful as anything in physics, and offer the advantage of actual experiments.

    “the ideas of quantum field theory that I learned about originally in string theory are extremely useful. And it's the right way to think about topological insulators or fractional quantum Hall effects or... high temperature superconductors... These are incredible quantum, many body entangled systems. They're described by quantum field theory”

0.43

Mathematical theorems constrain lattice formulations of physics, suggesting spacetime may not be fundamentally discrete despite theoretical predictions.

2 pointscentrality 3/5
  • There is a mathematical theorem from the 1980s, the Nielsen-Ninomiya theorem, stating that it is not possible to write down a discrete lattice version of the weak nuclear force; the theory as currently understood can only be formulated in the continuum, meaning the laws of physics in their full form cannot be simulated on a computer.

    “there is a mathematical theorem. It's a theorem that was written down in the 1980s. It's called the Nielsen-Ninomiya theorem. And it says you simply cannot do it.”

  • It is odd that a mathematical theorem could tell us spacetime cannot be discrete, because whether spacetime is fundamentally discrete should be a Planck-scale question requiring a collider 15 orders of magnitude more powerful than the LHC—information far above our experimental pay grade—so Tong does not believe the theorem's conclusion is right and thinks there must be a way to evade it by discarding one of its reasonable-seeming assumptions.

    “it's a bit weird that there's a theorem that says space time cannot be discreet... This should be something at the plank scale. Have to build a particle collider that's 15 orders of magnitude more powerful than the LHC before we get to answer that question.”

0.36

Black hole entropy suggests Einstein's equations may be emergent rather than fundamental, unlike typical theories that announce failure by producing infinities.

3 pointscentrality 2/5
  • When a physics theory is wrong it usually announces this by giving infinity (raising its hands to say it doesn't know), but black hole entropy is different—the theory does not admit its failing and instead 'tries to lie to you,' suggesting Einstein's equations may not be fundamental but emergent like the Navier-Stokes equations are derived from underlying atoms.

    “A theory will give infinity in physics, and that's the way of the theory putting up its hands and saying, yeah, I don't know what's going on here. And occasionally there are things where the theory tries to lie to you... the entropy of a black hole is closer to that.”

  • Both the Copenhagen and many-worlds interpretations of quantum mechanics are unsatisfactory, and the right approach is to shut up and calculate, because if you pause at the starting point having deep thoughts using words like 'epistemic' and 'ontological' you risk missing the quantum vistas that await just around the corner.

    “basically settle down on, shut up and calculate... if you pause at the starting point, having deep thoughts and saying words like epistemic and ontological, that you are in danger of missing the quantum vistas that await just around the corner.”

  • Gases are well understood because when particles are far separated you can start with 10^23 particles in a box and rigorously derive the Navier-Stokes equations, but liquids are much more complicated than gases and that derivation cannot be done for them.

    “You can start with 10 to the 23 particles in a box and just derive the Navier-Stokes equations if they're far separated... But you can't do that for a liquid. Liquids are much more complicated than gases.”

0.35

QFT is not mathematically well-defined, which physicists' inability to teach it to mathematicians reveals as a failure of physicists' understanding.

1 pointcentrality 3/5
  • Mathematicians can be taught general relativity and contribute enormously to it, but you cannot teach a mathematician quantum field theory because QFT is not well defined—it uses mathematics that mathematicians have not yet invented—and Tong now believes this inability is a sign that physicists themselves are failing to understand something, not merely that mathematicians are being pedantic.

    “You can't teach a mathematician quantum field theory... Because it's not well defined. We're using mathematics that mathematicians haven't yet invented.”

0.21

Parity violation was the most important twentieth-century particle physics discovery because it revealed standard model structure.

1 pointcentrality 2/5
  • The discovery by Chien-Shiung Wu in the 1950s that nature violates parity—that fundamental interactions can look different when reflected in a mirror, so you could tell whether a movie of such an interaction was the original or a reflection—may be the single most important particle physics experiment of the 20th century, because understanding it lays out the structure of the standard model.

    “The astonishing fact that was discovered by Chen Sheng Wu in the 1950s is that the same is true on the most fundamental level... I think that this is the single most important experiment in particle physics of the 20th century.”

0.21

Quantum gravity is not mysterious; effective field theory already explains observed phenomena, with difficulties only in extreme regimes.

1 pointcentrality 2/5
  • The claim that quantum gravity is a total mystery is oversold: gravity is currently described by a quantum field theory in which space-time is the field, you can quantize the gravitational field and explain every experiment ever done, and difficulties only arise in the most extreme situations (black hole centers, the Big Bang); the effective field theory of quantum gravity already explains why apples fall and the earth orbits the sun.

    “The idea that quantum gravity is this big mystery, that we've got no idea what's going on, that's totally oversold. The truth is you can quantize the gravitational field. You can quantize it and explain any experiment we've ever done.”

0.21

Practical lattice simulations of QFT on computers succeed perfectly and are essential because pen-and-paper methods fail in strong coupling.

1 pointcentrality 2/5
  • Replacing the continuum of space with a discrete set of points lets physicists simulate quantum field theory on computers, which is essential because QFT is unbelievably hard and pen-and-paper methods fail in strong-coupling situations like quarks binding into a proton; the lattice simulations agree perfectly with experiments.

    “once you replace the continuum of space with a discrete set of points... you're in place to start to simulate quantum field theory on a computer... it tells us what happens. And it agrees perfectly with experiments.”

0.15

Maxwell equations are best taught as fundamental starting points, and Schrodinger versus qubit pedagogy depends on valuing idea connections.

2 pointscentrality 1/5
  • Electromagnetism is better taught by starting from the Maxwell equations—if you have the vector calculus, you write down the Maxwell equations and everything else is commentary—rather than the conventional approach of building up to them through electric and magnetic charges.

    “If you've got the maths, you write down the Maxwell equations, everything else is commentary. That's literally everything you need to know about everything to do with electricity and magnetism.”

  • Quantum mechanics can be taught two ways: via the Schrodinger equation, which is mathematically more complicated (with subtleties to sweep under the rug) but connects to familiar ideas of how particles move; or via qubits, which is mathematically trivial but disconnected from prior intuition—and the right pedagogical choice depends on valuing connections between physics ideas.

    “One is mathematically more complicated, the Schrodinger equation... it's much more in contact with things that you know already... Other one is mathematically trivial, but you ask, this is to do with qubits?”

0.15

Physics allows hidden gaps and argument by personality while mathematics requires rigor, and universal scaling makes fluid mechanics astonishingly simple.

2 pointscentrality 1/5
  • Teaching physics is easier than teaching mathematics because in physics you can hide gaps, argue by force of personality, and push through with bombast, whereas mathematics (e.g., differential geometry in general relativity) must be done rigorously and it is very hard to lie.

    “It's why teaching physics is much easier than teaching mathematics... It's very hard to hide when you're teaching mathematics. Very hard. Very hard to lie.”

  • If you take anything in the universe, put a bunch of it in a box and heat it up, then regardless of what you started with, it is described by the Navier-Stokes equations of fluid mechanics—an astonishing universality where all the differences between substances evaporate in the hot regime.

    “you take anything in the universe... you throw a bunch of it in a box, you heat it up, and no matter what you started with, that stuff is described by the Navier-Stokes equation of fluid mechanics. And that's an amazing statement.”

0.07

Classical equations like Navier-Stokes and Einstein's relativity contain unresolved singularity problems mathematicians have not solved.

1 pointcentrality 1/5
  • The two preeminent equations of classical physics—the Navier-Stokes equations and Einstein's equations of general relativity—both contain subtle, poorly understood singularity problems (the Navier-Stokes millennium problem and the cosmic censorship conjecture that GR singularities are shrouded inside black hole event horizons) that mathematicians have not resolved after a century.

    “the fact that the two preeminent equations of classical physics, Navier-Stokes and Einstein's equations of GR, both have very subtle things that after at least a century, mathematicians have still not got on top of strikes is curious”

0.07

Established researchers should choose problems by importance divided by competition, though this heuristic is poor advice for students.

1 pointcentrality 1/5
  • A heuristic for choosing research problems—attributed to a famous physicist's after-dinner speech—is to estimate how important a problem is and divide by the number of people already working on it; this is terrible advice for graduate students and postdocs but is exactly how established researchers in secure jobs should pick problems.

    “you pick the problem and you try to put a value on how important that problem is... Then he said, you divide by the number of people that are already working on it... once you're settled in a nice job like you and me, this is exactly the way that we should pick our problems.”