271 | Claudia de Rham on Modifying General Relativity
What this covers
Sean Carroll interviews Claudia de Rham about whether general relativity might need modification despite its century of empirical success. De Rham, a theoretical physicist at Imperial College London who has led research into alternative gravity theories, argues that the theory may fail at cosmological scales—the very longest distances where curvature is weakest—precisely because those distances remain experimentally untested. The conversation centers on massive gravity, a framework in which the graviton carries a tiny mass rather than being massless, as a concrete alternative that sidesteps decades of "no-go theorems" and might resolve the deepest puzzle in physics: the cosmological constant problem.
The exchange moves through quantum field theory from first principles. De Rham and Carroll explore why spin 2 is the unique, inescapable choice for a force-carrying particle, how this choice led Richard Feynman and Stanley Deser to prove that general relativity is the only consistent theory of a massless spin 2 particle, and why earlier no-go theorems against massive versions—once assumed airtight—actually harbored loopholes that de Rham and Gregory Gabadadze later exploited. A central technical problem emerges: giving the graviton mass breaks symmetries that prevent dangerous ghost polarizations, negative-energy modes that would destabilize reality. De Rham explains how her work navigates that minefield. The conversation then pivots to cosmology: quantum fields predict a vacuum energy 56 orders of magnitude too large. A finite-range, finite-lifetime graviton might weaken vacuum energy's long-term grip on spacetime expansion. Future low-frequency gravitational-wave detectors could distinguish this scenario from Einstein's theory through subtle frequency-dependent delays that massive gravity would introduce.
De Rham argues that gravity is best understood as a quantum force mediated by a massless spin-2 graviton, but that giving the graviton a tiny mass ('massive gravity') is a viable, theorem-evading modification of general relativity that could explain the smallness of the observed cosmological constant by letting vacuum energy's gravitational effect weaken over cosmic time.
- General relativity is the unique consistent theory of a massless spin-2 particle, so any modification must work within strict quantum field theory constraints.
- The vacuum energy of known quantum fields overpredicts the cosmological constant by at least 56 orders of magnitude, motivating modifying gravity in the infrared.
- A massive graviton would have a finite range in space and time, potentially tuning down the long-term gravitational effect of vacuum energy.
Massive gravitons with symmetry-breaking polarizations could weaken gravity at cosmic distances and scales.
- If the graviton has a mass—no matter how small—the symmetry that forbids extra polarizations in Einstein's theory is broken, opening up additional polarizations including longitudinal modes (like sound waves) that propagate along the direction of travel; conceptually it makes no difference whether the mass is infinitesimal or large, since any non-zero mass qualitatively introduces these new channels of gravitational communication.
“if you add a small mass to the graviton, you could have those additional polarizations... it doesn't matter how small the mass is. It could be extremely small... but as soon as you have the possibility of a non-zero mass, then you open up the possibility of these additional polarizations”
- If the graviton has a finite range—particularly a finite range in time—then gravity may become 'lazy' after 14 billion years and no longer fully transmit the enormous vacuum energy into spacetime curvature, weakening its effect on cosmic acceleration over billions of years and potentially resolving the cosmological constant discrepancy.
“maybe this is because we are actually just starting to probe the fact that gravity itself has a finite range, maybe in space. But what's even more relevant is that it has a finite range in time.”
- The graviton mass envisaged in massive gravity is of the order of the Hubble parameter today, roughly 10^-32 to 10^-33 electron volts (matching the size of the observable universe), while current constraints from LIGO and neutron-star mergers only require the mass to be smaller than about 10^-22 electron volts—leaving roughly a 10-order-of-magnitude margin that future low-frequency gravitational wave observations could tighten.
“It puts a constraint on the graviton mass that it has to be smaller than roughly 10 to the minus 22 electron volts or so... we want a graviton mass which is roughly of that order of magnitude... 10 to the minus 32 electron volts”
- The Higgs mechanism demonstrates that fundamental particles carrying a force can have a mass, as with the spin-1 W and Z bosons that mediate the weak force; the weakness of the weak force is precisely because it is propagated by very massive particles, providing an intuition for how a massive graviton would weaken gravity at long ranges.
“we do know actually from the Higgs mechanism, that fundamental particle that carry a mass, sorry, that carry a force, can have a mass. That is the case of the W and the Z Boson.”
Gravity is quantum at base but perturbative quantum gravity works fine at low energies and curvatures.
- Gravity, like all the fundamental forces, is fundamentally a quantum phenomenon, and physicists can describe it very well at the quantum level as a force up to a given energy scale—this is uncontroversial and routine, distinct from the unsolved problem of full quantum gravity at the Planck scale.
“I would say that gravity also is a quantum phenomenon. And we do understand very, very well how to describe this, up to some given extent, at the quantum level, to describe gravity as a force, as a quantum force, up to a given level.”
- Perturbative quantum gravity works fine in everyday regimes of low curvature and low energy, and only becomes problematic near the Planck scale—about 10^27 electron volts, some 15 orders of magnitude beyond the LHC—where colliding particles would produce black holes and require a full theory of quantum gravity to predict the outcome.
“if you think at sufficiently low curvatures, sufficiently low energy scales, then you can quantize gravity, if you want to say, in a perturbative way”
- Gravity can legitimately be thought of as a force like electromagnetism or the weak force, but what we experience as gravitational attraction is better understood as the manifestation of the curvature of spacetime, where mass affects the curvature and that curvature in turn dictates how objects move.
“what we typically experience as gravitational attraction, let's say, it's better understood as being the representation, the manifestation, I would say, of the curvature of space-time we live in”
General relativity is the unique massless spin-2 theory and vast unexplored parameter space remains for testing gravity at cosmological scales.
- Rather than special relativity being founded on the postulate that nothing travels faster than light, modern physicists increasingly think of it the other way around: the fundamental particles and symmetries come first, and the consequence that nothing can travel faster than light emerges from those underlying symmetries.
“nowadays, I would say we almost think of it the other way around. We can very much think of the fundamental particles and the fundamental symmetries, and a lot of those things come out of that, that nothing can travel faster than light, because of the symmetry that we are relying on ourselves”
- Dark energy is merely a placeholder name for whatever causes the observed accelerated expansion, describable as a fluid with positive energy density but negative pressure; it is not 'anti-gravity' or counter-gravity—it acts entirely through gravity, with its negative pressure producing the acceleration.
“it's not at all counter-gravity in any way. It really acts with gravity, and it acts with gravity in its favor. There's nothing anti about it. But it's a fluid which we can describe with positive energy density but negative pressure.”
- The electromagnetic force is mediated by the photon, a spin-1 particle behaving like a dipole, while gravity is mediated by the graviton, a spin-2 particle behaving like a quadrupole; this difference in spin under Lorentz transformations is a technicality at the descriptive level but has dramatic consequences when building a full theory.
“One is what I would call a spin 1. And then the other one is a spin 2. And it's just a technicality, at this level, related to how they behave and rotations.”
- While general relativity is well-tested at solar-system and galactic scales, we have no experiments testing gravity at the very longest distances—the scale of the observable universe and the correspondingly low curvatures—so we cannot actually confirm that gravity is described by general relativity there; it remains an unverified premise rather than an established fact.
“it's not like I can do another experiment in the lab measuring these big distances, the behavior of gravity on those big distances... We can't compare and say, "Okay, it is well-described by Einstein's theory of general relativity," because we don't know.”
- Exploring modified gravity, even if it ultimately fails, puts physicists in a better position than before: rather than having an answer (general relativity) that they were not even allowed to question, they now have an open question—which de Rham regards as the more valuable epistemic situation when facing a puzzle as large as the vacuum energy problem.
“rather than having an answer that we weren't even allowed to question before, now we have a question that I can't answer for you, but I think that's much better”
- In particle collisions, the probability of an outcome from exchanging a mediating particle grows with energy raised to twice the spin of that particle—so photon (spin 1) effects grow like energy squared while graviton (spin 2) effects grow like energy to the fourth, meaning gravity dominates and grows uncontrollably as energies approach the Planck scale, which is why spin 2 is the highest spin that can make sense for a fundamental force-carrying particle.
“the probability of the outcome will grow the energy for that process, the energy comes to the power twice the spin. So that's to the power four.”
- The quantum fields of all known particles produce vacuum fluctuations and an associated vacuum energy that is present everywhere in space and time and acts like a cosmological constant; but integrating this contribution gives a value too large by at least 56 orders of magnitude (120 if extended to the Planck scale) compared to the observed accelerated expansion, making it arguably the biggest discrepancy in the history of physics.
“the level of the contribution of vacuum energy to this cosmological constant is too big, by at least 56 orders of magnitude, if I just consider the particles that I do know for sure do exist in nature, like the Higgs, like the electron”
- Theorems established by physicists such as Feynman and Deser show that, under certain assumptions about coupling and symmetry, the only consistent theory of a massless spin-2 particle is general relativity—meaning Einstein's theory can be derived from the ground up from particle-physics principles, and is the only possibility not merely because it is beautiful but because it is the only thing that makes sense.
“there are theorems done, for instance, by Feynman, by Deza, by all sorts of amazing physicists in the past century, that show that the only consistent theory that I can have within some given assumptions about how they couple and things like that and the symmetry level, is general relativity”
Massive gravity faces severe theoretical obstacles including ghost polarizations that destroy theory viability.
- Although a back-of-the-envelope linearized calculation suggests a massive graviton could tune down the cosmological constant, making this work in a realistic cosmological setup is extremely challenging because the standard homogeneous, isotropic cosmological solutions of general relativity no longer apply in massive gravity—which de Rham argues is actually a good sign, since a theory that merely reproduced GR's simple relations would explain nothing new.
“the typical cosmological solutions that we have, the way we construct them in general relativity, that doesn't work anymore for massive gravity”
- There were at least six no-go theorems claiming a consistent nonlinear theory of massive gravity was impossible, but each no-go theorem rests on underlying assumptions—commonly that one can extrapolate from a limited region of possibilities rather than charting all cases—so they were really no-go theorems for the simplest models, leaving loopholes that allowed de Rham and Gabadadze to construct a working theory.
“all of the no-go's theorems, you have a no-go, but there's always some level of assumptions that go beyond... one of the most common assumption was that... wasn't to chart all the possibilities, because this is really very hard to do every possible case”
- Some longitudinal polarizations of a massive graviton mess with the flow of time and, via the connection between time and energy, end up with negative energy—called 'ghosts'; because they cost negative energy to produce, they will inevitably be produced and grow as large as possible, destroying the structure of reality, which makes such theories simply unviable, unlike the merely uncomfortable tachyonic instabilities physicists know how to handle.
“you end up with some polarization which have negative energy. We call them ghosts... they will get produced, whether you want it or not... dominating the whole world and destroying the whole structure of reality along with it.”
Curvature cannot exist or be felt locally; gravity requires comparing observations across different spatial points.
- Because of the equivalence principle, gravity affects every fundamental particle in a body in exactly the same way, so there is no relative stretching, pressure, or chemical reaction by which a human could sense it at a single point—gravity cannot be felt locally because curvature itself requires comparing what happens at different points.
“it affects every single cell in a body. Every single molecule, every single atom, every single fundamental particle in your body, are affected in exactly the same way through this gravitational way we experience the curvature space-time around ourselves. So there's no stretching of any cells apart”
- Curvature, and therefore gravity, cannot exist locally at a single point; it requires comparing and communicating between different points—two observers on different parts of Earth would only detect the planet's curvature by comparing their differing views of the sky—which is why the theory is named relativity.
“it affects every single cell in a body. Every single molecule, every single atom, every single fundamental particle in your body, are affected in exactly the same way through this gravitational way we experience the curvature space-time around ourselves. So there's no stretching of any cells apart”
Current and future gravitational wave observations constrain graviton mass and can distinguish massive gravity from general relativity.
- Gravitational waves have two polarizations—plus and cross, oriented 45 degrees apart—the same number as light, but this coincidence is merely an accident of living in four dimensions; in higher dimensions the number of polarizations would differ.
“gravitational waves have two polarizations. The fact that the same number as light is just an accident of four dimensions. In higher dimensions, it would be different”
- The classical gravitational waves observed by detectors are made of an enormous number of gravitons—roughly 10^40 to 10^44—so detecting a single graviton would be extremely challenging, though in principle possible with an experiment more precise than the Heisenberg uncertainty limit would naively allow.
“The gravitational waves we have observed, they have a lot, really, a lot of those gravitons, about 10 to the 40, 44 gravitons.”
- If gravitational waves are massive, low-frequency waves are more affected by the mass and propagate slightly slower than high-frequency waves that travel near the speed of light; observations of neutron-star mergers show gravitational waves travel at the speed of light to one part in 10^15, and detecting primordial gravitational waves via B-mode polarization in the CMB could reveal a low-frequency plateau distinguishing massive gravity from general relativity.
“if gravitational waves are massive, then those at low frequency will be more affected by the mass and will start propagating at a slightly lower speed than those at a higher frequency that travel close to the speed of light”
- Gravitational waves are now established as real phenomena in nature, corresponding to a distortion of space and time around a quadrupole, propagating like waves through spacetime in close analogy to electromagnetic waves of light, and detectors use two perpendicular tubes to measure the squeezing in one direction and stretching in the perpendicular direction.
“We have observed gravitational waves... they could correspond to a distortion of space and time around what we call a quadrupole”