The Equivalence Principle

In 1907, sitting at his patent-office desk, Einstein had what he later called "the happiest thought of my life": a person in free fall does not feel their own weight. Step off a diving board, and for that whole plummeting second the bathroom scale you are clutching reads zero. Gravity, the most familiar force in the universe, can be made to vanish simply by letting yourself fall. No other force behaves like this — you cannot switch off electromagnetism by moving cleverly. That one observation is the seed of general relativity, and the principle that grows from it is what this page is about.

The claim is bolder than it first sounds. It is not merely that falling feels weightless; it is that gravity and acceleration are physically indistinguishable. Sealed inside a windowless box, no experiment you can perform will tell you whether you are sitting still on a planet or being dragged through empty space by a rocket. This is the equivalence principle, and from it — with almost no extra assumptions — light must bend near a star, and clocks must run slow where gravity is strong. We build all of that from two thought experiments and a falling elevator.

The two elevators

Einstein's argument comes in two matching halves. Picture a windowless elevator car, and an experimenter inside with a bag of apparatus.

Elevator A — the accelerating rocket. Far out in deep space, away from every star, the car is bolted to a rocket that fires steadily so the floor pushes up with acceleration a = 9.8\ \text{m/s}^2. The experimenter's feet press on the floor; a dropped apple races to meet the rising floor at 9.8\ \text{m/s}^2; a scale reads their full weight. Everything feels exactly like standing on Earth — and Einstein's point is that it is not merely similar: no measurement can tell the two apart.

Elevator B — the falling car. Now cut the cable. The car and everything in it fall together. The experimenter floats; the apple, released, hangs motionless beside them; the scale reads zero. This is exactly the experience of an astronaut in orbit, who is simply falling around the Earth and never hitting it. Inside the falling car, gravity has disappeared.

Weak form (universality of free fall). Every object falls with the same acceleration in a gravitational field, regardless of its mass or composition. Gravitational "charge" (mass) equals inertial mass.
Einstein (strong) form. In a small enough freely-falling laboratory, all the laws of physics take their special-relativistic, gravity-free form. A uniform gravitational field is locally indistinguishable from an accelerating frame.

The weak form is astonishingly well tested. Galileo rolled balls down ramps; the Apollo 15 commander dropped a hammer and a feather on the airless Moon and watched them land together; modern torsion-balance and satellite experiments (MICROSCOPE, 2022) confirm that different materials fall at the same rate to better than one part in 10^{15}. That equality — inertial mass equals gravitational mass — is a coincidence in Newton's theory and a necessity in Einstein's.

Consequence 1: gravity bends light

Here is where the principle earns its keep. Shine a laser horizontally across the accelerating rocket car (Elevator A), entering one wall at head height. Light is fast but not infinitely fast: it takes a tiny time \Delta t to cross the car. During that time the car — and its far wall — has accelerated upward by \tfrac12 a\,\Delta t^2. So the beam arrives at the far wall slightly below where it entered. To the experimenter inside, the horizontal beam has curved downward, tracing a gentle parabola.

Now invoke equivalence: if the beam bends in the accelerating rocket, it must bend in exactly the same way in a real gravitational field, because the two are indistinguishable. Therefore gravity bends light. Starlight grazing the Sun should be deflected — a prediction Eddington's 1919 eclipse expedition famously confirmed, and the moment Einstein became a household name. Light has no rest mass, yet it falls, because falling is about the geometry of spacetime, not about the light.

Consequence 2: clocks run slow in gravity

The same accelerating car forces time itself to bend. Put a clock on the floor and one on the ceiling, a height h apart, and let the floor clock flash a light pulse upward each tick. While each pulse climbs, the ceiling is accelerating away from the source, so by the Doppler effect the ceiling receives the flashes stretched out — the upper clock sees the lower clock running slow. Translate through equivalence and you have a statement about real gravity: a clock deeper in a gravitational well ticks slower than one higher up. This is gravitational time dilation, and its light-wave twin is the gravitational redshift — light climbing out of a well loses energy and reddens.

For a weak, uniform field of strength g, the fractional rate difference between two clocks separated in height by h is

\frac{\Delta f}{f} = \frac{\Delta \tau_{\text{high}} - \Delta \tau_{\text{low}}}{\Delta \tau} = \frac{g\,h}{c^2} = \frac{\Delta \Phi}{c^2},

where \Delta\Phi = g h is the difference in gravitational potential. The c^2 downstairs makes this minuscule at everyday scales — but it is real. In 1959 Pound and Rebka measured it over just the 22.5\ \text{m} height of a Harvard tower, a fractional shift of about 2.5\times 10^{-15}, using the exquisite sharpness of gamma-ray resonance. Slide the surface gravity below to see how the clock-rate gradient steepens on heavier worlds.

You can only make it disappear locally — in a lab small enough and brief enough that the field looks perfectly uniform. Real gravity is not uniform: it points toward the centre of the Earth, so it is slightly stronger at your feet than your head and slightly convergent from side to side. Two apples released a metre apart in a falling car do not hang perfectly still — they drift together, because each falls toward the Earth's centre along a slightly different line. Stretch the car tall and they drift apart vertically, since the lower apple falls faster. These leftover relative accelerations are the tidal forces, and they are what genuinely distinguishes real gravity from a mere accelerating rocket. Tidal effects are the true fingerprint of curvature — and no change of frame can transform them away. The equivalence principle is a statement about a single point and its immediate neighbourhood; curvature is what happens when you compare neighbourhoods.

A common overreach. The principle does not say gravity is an illusion or that it can be abolished globally — it says a uniform field is equivalent to acceleration over a small region. Two mistakes to avoid. First, don't drop the word "local": across a large region gravity varies, and the tidal differences (above) can never be transformed away by any choice of frame — that residue is exactly the real, coordinate-independent gravity, encoded later in the curvature tensor. Second, don't confuse the two masses casually: the equivalence of inertial mass (the m in F=ma) and gravitational mass (the m in F = mg) is an experimental fact that the principle elevates to a postulate — it is not logically forced, it is a deep clue about nature that Einstein chose to take seriously. Free fall removes the feeling of weight, not the curvature that steers your trajectory.