Gravity is Not a Force: Rethinking General Relativity Without the Rubber Sheet Metaphor

The pressure you feel on the soles of your feet right now is not gravity. It is the floor. Specifically, it is electrons in the floor repelling electrons in your shoes, hard enough to push you off the path you would otherwise be following through spacetime. General relativity says that path — the one you would take if the floor vanished — is the natural, unforced one. Falling is what it feels like to have no forces on you at all. Standing still is what it feels like to be shoved.

That is the real content of the slogan “gravity is not a force”, and it is a deep idea that the usual picture actively hides. So let me start by taking the picture apart.

§The trampoline explains gravity with gravity

You know the image: a bowling ball sits on a stretched rubber sheet, the sheet dips, and a marble rolled past curves in toward the ball. Spacetime is curved by mass, the story goes, and things roll along the curve.

Ask why the marble rolls downhill. Because something is pulling it down, toward the floor of the room the trampoline is in. The demonstration works only because Earth's gravity is acting on the sheet and the marble. It explains gravity by quietly assuming gravity. That alone should disqualify it.

The second problem is worse, because it is not a flaw in the prop but in the idea it conveys. The sheet is a picture of curved space. Put a marble on it at rest, far enough from any slope, and it stays put. But release an apple at rest above the ground and it falls. Something has to get a stationary object moving, and curved space cannot do that, because a stationary object is not going anywhere through space. It is, however, going somewhere through time — one second per second, whether it likes it or not. If spacetime is curved in the time direction, then an object moving purely through time gets tipped, slightly, into moving through space. That tipping is the apple falling.

So the trampoline leaves out the time part, and for slow objects the time part is very nearly the whole thing. I will put a number on “very nearly” below. First, the principle that makes all of this possible.

§Falling is what free looks like

Picture yourself in a windowless lift whose cable has snapped. You let go of a pen and it floats beside you. Everything in the lift falls at the same rate, so nothing moves relative to anything else, and no experiment you can do in that small box tells you that you are near a planet at all. Picture the reverse: the same box at rest in deep space, pushed by a rocket at 9.81 m/s². The pen falls to the floor just as it would on Earth. Gravity and acceleration are, locally, indistinguishable.^[1]1. Markus Pössel, “The elevator, the rocket, and gravity: the equivalence principle”, Einstein Online (Max Planck Institute for Gravitational Physics). It is the clearest short treatment I know of, including the point that the principle holds exactly only in a vanishingly small region.

This only works because everything falls the same way regardless of what it is made of. That is a checkable fact, not a philosophical stance, and it has been checked brutally. The MICROSCOPE satellite dropped titanium and platinum test masses around Earth for two and a half years, and found their accelerations agreed to around one part in 10¹⁵.^[2]2. P. Touboul et al., “MICROSCOPE mission: final results of the test of the Equivalence Principle”, Physical Review Letters 129, 121102 (2022). The Eötvös parameter for the titanium–platinum pair came out at [−1.5 ± 2.3 (stat) ± 1.5 (syst)] × 10⁻¹⁵. Contrast electricity: put a proton and an electron in the same electric field and they accelerate in opposite directions, at rates differing by a factor of nearly two thousand. The electric field acts on things according to their charge. Gravity acts on everything according to nothing. Its effect depends only on where you are and how fast you are going.

When an effect depends only on where you are and how you are moving, it is a property of the place, not a push on the thing. Einstein took that seriously: if every free body follows the same path from the same starting point, then the paths are a feature of spacetime itself. Those paths are geodesics — the straightest possible lines in a curved geometry, the way great circles are the straightest lines on a globe. A body with no forces on it follows one:

Read this as “zero acceleration, measured properly”. The first term is what you would naively call acceleration; the second, built from the Christoffel symbols $\Gamma$, corrects for the fact that your coordinates are themselves curving. There is no mass on the right-hand side and no force. An accelerometer strapped to a body on a geodesic reads zero, which is precisely what an astronaut in orbit experiences. The ISS is not out of Earth's gravity; at its altitude gravity is still about nine-tenths of what it is on the ground. Its crew is weightless because nothing is stopping them from falling.

§Most of gravity is time

The geometry lives in the metric, the rule that turns coordinate differences into real intervals of time and distance. Far from any black hole, for a weak, slowly changing field like Earth's, it looks like this:

where $\Phi$ is exactly the Newtonian gravitational potential, $-GM/r$. The first term says clocks lower in the potential run slower. The second says rulers are slightly distorted too. Now feed this into the geodesic equation for an object moving slowly, with speed $v \ll c$. Almost every term drops out, because the object moves through time at essentially $c$ and through space at a crawl. What survives is the piece that comes from the time–time part of the metric:^[3]3. The weak-field limit is in any GR textbook; two free versions are Sean Carroll's Lecture Notes on General Relativity (chapter 4) and Viktor Toth's short note “The Newtonian limit in general relativity”. Sign conventions differ between them; the physics does not.

That is Newton's law of gravity, recovered with no force in sight, purely from the way the rate of time changes with height. The spatial part of the curvature does contribute, but suppressed by a factor of about $(v/c)^2$. For a ball thrown at 10 m/s, that is $(10 / 3\times10^8)^2 \approx 10^{-15}$. So when you throw a ball and watch it arc, you are watching, to fifteen decimal places, the curvature of time.

And the curvature of time is tiny. At Earth's surface,

A clock on the ground runs slow, relative to one far away, by about seven parts in ten billion. That small slope in the rate of time, spread over the height of a room, is enough to pull everything you own toward the floor at 9.8 m/s². There is a reason the effect is so powerful despite being so small: you are moving through time very, very fast. A second of time is 300,000 km of spacetime in the units that matter, and a slight tilt over that distance becomes a large sideways drift.

The slope has been measured directly. In 1959 Robert Pound and Glen Rebka sent gamma rays down a tower at Harvard about 22.5 m tall and detected the frequency shift between top and bottom.^[4]4. “The Weight of Light”, Physical Review Focus 16, 1 (2005), on Pound and Rebka's 1960 result: within 10 percent of Einstein's prediction, improved to about one percent by 1964. The predicted shift is

and they found it, first to ten percent, later to one. You also rely on the effect every time you open a map. GPS satellite clocks, higher in Earth's potential, gain about 45 microseconds a day from gravity and lose about 7 from their orbital speed, for a net gain of about 38.^[5]5. Pratap Misra, “Inside the box: GPS and relativity”, GPS World (2023). The satellite clocks are set to 10.22999999543 MHz before launch so that, seen from the ground, they tick at 10.23 MHz. Light travels about 11 km in 38 microseconds. Without correction, the system's ranging would drift by kilometres a day.

Where the trampoline's spatial curvature does matter is for things that move fast. Light, crossing space at $c$, feels the space part as strongly as the time part. Einstein's 1911 calculation of starlight bending past the Sun, done before he had the full theory, effectively included only the time part and came out at half the value general relativity gives. So the sheet is not wrong; it is the minor half of the story, presented as the whole of it, for the one case — slow things like apples, planets and you — where it matters least.

§What is left when you fall: tides

If falling feels like nothing, how can a freely falling observer ever detect gravity? Not with one particle. With two.

Go back to the snapped lift and hold two pens, one above the other. Each follows its own geodesic toward Earth's centre, and the lower one is slightly deeper in the field, so it accelerates slightly more. The pens drift apart vertically. Hold them side by side instead and their geodesics both point at the centre of the Earth, which means they converge. Stretched in one direction, squeezed in the other. That is a tide, and it is the one gravitational effect you cannot make disappear by choosing to fall. General relativity calls it geodesic deviation, and it is governed by the Riemann curvature tensor:

Here $\xi$ is the separation between neighbouring geodesics and $u$ is their shared velocity. In the Newtonian limit this collapses to something familiar, the second derivatives of the potential:

This is the right way to say what curvature is. Not a dip in a sheet, but the rate at which initially parallel free-fall paths come together or spread apart. The Christoffel symbols in the geodesic equation, the bits that produce the “force” of gravity, can be set to zero at any single point by choosing a falling frame. The Riemann tensor cannot. It is the part of gravity that is real in every frame.

It is also small enough near Earth to be almost comic. Along the vertical, the tidal stretch is $2GM/r^3$ per metre of separation:

For a body two metres tall, that is a difference in acceleration between head and feet of about 6 × 10⁻⁶ m/s², less than a millionth of the 9.8 m/s² you think of as gravity. That ratio is the most important number in this essay. Almost all of what you call “gravity” in everyday life is the floor pushing you off a geodesic; the true, frame-independent curvature is a millionth of it. Another way to put the same number: the curvature radius of spacetime at Earth's surface, roughly $(GM/R^3c^2)^{-1/2}$, comes to about 2.4 × 10¹¹ m, more than one and a half times the distance to the Sun. Locally, spacetime here is very nearly flat. We live on the gentlest of slopes and are simply prevented from sliding down it.

§Where the slogan overreaches

I like “gravity is not a force” as a provocation, but I don't think it should be repeated as a fact without some qualifications, because it can make people think physicists who say “gravitational force” are confused. They mostly are not.

First, Newton is not wrong in any way that matters for most of life. The relativistic corrections at Earth's surface are of order 10⁻⁹. Bridges, rockets and interplanetary probes are designed with Newtonian gravity as a force, and they work. If you are in a frame tied to the ground — and engineers nearly always are — treating gravity as a force $F = m\,\nabla\Phi$ is the correct physics in that frame, in the same way centrifugal force is a perfectly usable concept on a merry-go-round. “Fictitious” in physics means “appears because of your choice of frame”, not “imaginary”.

Second, the geometric trick is not unique to gravity, and saying it is makes gravity sound more mysterious than it is. Electromagnetism and the nuclear forces are also described by mathematical objects called connections, close cousins of those Christoffel symbols. What sets gravity apart is not that it is geometric but that it is universal. Because everything falls the same way — to one part in 10¹⁵ and counting — the connection can be built into the geometry of spacetime itself, shared by all matter, rather than living in a separate internal space that only charged particles feel. A more honest slogan would be: gravity is the one interaction so universal that you can hide it in the shape of spacetime.

Third, at the level of quantum field theory, physicists routinely treat weak gravity as an interaction between particles on a nearly flat background, mediated by a hypothetical graviton, and that description agrees with general relativity where both apply. Whether gravity is “really” geometry or “really” a field is not a question the current evidence answers. The two pictures are tools that are good at different jobs.

What survives all the qualifications is the thing I find genuinely beautiful. The sensation you have called gravity your whole life is the floor holding you back. Gravity itself, the frame-independent part, is the faint tendency of things falling side by side to drift together or apart. And the pull of the Earth, the thing that keeps the Moon in its orbit and your coffee in its cup, is mostly the observation that time runs less than one part in a billion slower down here than far away, and that everything, always, is moving through time.