Equivalence Principle

Why the equivalence principle matters

• Foundation of general relativity. Einstein's 1907 insight ("the happiest thought of my life") was that an observer in free fall feels no gravity, just as someone floating in deep space feels none. From this, gravity cannot be a real force in the Newtonian sense — it must be encoded in the geometry of spacetime, with free-falling worldlines as geodesics of a curved metric. The equivalence principle is the conceptual bridge from Newton's force-of-gravity to Einstein's curved-spacetime gravity.
• GPS time dilation. GPS satellites at altitude 20,200 km sit in a weaker gravitational potential than ground clocks. The equivalence principle predicts that clocks higher up tick faster by Δt/t = ΔΦ/c², where ΔΦ is the gravitational potential difference. For GPS this gives +45 µs/day. Combined with the special-relativistic slowdown from orbital velocity (−7 µs/day), the net is +38 µs/day. Without correcting for this, GPS positions would drift ~10 km/day. Every smartphone navigation feature depends on this calibration.
• Tests of modified gravity. Many proposed alternatives to general relativity — Brans-Dicke, scalar-tensor theories, MOND, certain string-theory scenarios, and tests of dark sectors — predict tiny composition-dependent forces or violations of mass equivalence. Precision tests like MICROSCOPE (2017, 10⁻¹⁵) and proposed missions (STEP, SR-POEM, targeting 10⁻¹⁸) constrain these theories. The equivalence principle is one of the sharpest tools we have to rule out modifications to general relativity.
• Atomic clock metrology. Modern optical atomic clocks reach fractional frequency precision near 10⁻¹⁸. At this level, lifting one clock by 1 cm produces a measurable redshift difference. Such clocks are used to map the Earth's gravitational potential ("relativistic geodesy") and to test the equivalence principle by comparing isotopes that have different proton-to-neutron ratios.
• Cosmological tests. Recent observations test whether the equivalence principle holds across cosmic scales. Type Ia supernovae, large-scale structure, and CMB measurements constrain whether dark matter and dark energy obey equivalence. The latest analyses confirm the principle with high precision, ruling out long-range "fifth forces" coupling differently to different forms of matter.
• Light bending and gravitational lensing. The equivalence principle predicts that light, like everything else, falls in a gravitational field. The Eddington 1919 solar eclipse expedition measured starlight bending around the Sun by 1.75 arcseconds — twice the Newtonian prediction (full GR, of which equivalence is a part) and confirmed the modern theory. Today, gravitational lensing is a workhorse for mapping dark matter and discovering exoplanets.
• Interpretation of "g." Newton wrote F = m_g g; Einstein interpreted g not as a force but as the curvature parameter of spacetime — the worldline of a free-falling particle is a geodesic, and "g = 9.81 m/s²" is the rate at which the timelike geodesic deviates from coordinate axes near Earth's surface. The numerical value is the same, but the conceptual content is completely different.
• Quantum gravity constraints. Any consistent theory of quantum gravity must reduce to general relativity in the classical limit, hence must obey the equivalence principle there. Whether subtle quantum effects can violate the principle is an open question — discrepancies (such as the proposed Bekenstein bound or holographic considerations) are actively investigated. The principle provides one of the few experimentally accessible windows into Planck-scale physics.

The three forms and how they're tested

• Weak equivalence principle (WEP). All test bodies, regardless of composition, fall with the same acceleration in a vacuum. Equivalent statement: m_inertial / m_gravitational = constant for all materials. Tests: Galileo (Pisa tower, ~1% precision); Newton's pendulum experiments (~10⁻³); Eötvös 1908 (10⁻⁹); Roll-Krotkov-Dicke 1964 (10⁻¹¹); Adelberger group U. Washington (10⁻¹³); MICROSCOPE 2017 (10⁻¹⁵, on a satellite to eliminate seismic noise).
• Einstein equivalence principle (EEP). WEP plus local Lorentz invariance plus local position invariance. Locally, in a free-falling frame, all non-gravitational physics — including electromagnetism, weak force, etc. — looks like flat-spacetime physics. Used to derive gravitational redshift, time dilation, and bending of light. Tests: Pound-Rebka redshift, Hughes-Drever isotropy tests, atomic clock comparisons.
• Strong equivalence principle (SEP). EEP extended to gravitational self-interactions. The gravitational mass and binding energy of a body affect its motion the same way external gravity does — gravity is universal, including its own contribution. Holds in pure general relativity but is violated in scalar-tensor theories. Tested by lunar laser ranging (Nordtvedt parameter η < 4.4 × 10⁻⁴) and binary pulsar timing.
• MICROSCOPE 2017. French CNES satellite carrying two concentric capsules of different composition (titanium and platinum-rhodium alloy). Compares their free-fall acceleration around Earth using sub-nm displacement readout, free of seismic noise. Result: m_g/m_i agrees between the two materials to (1.0 ± 1.4) × 10⁻¹⁵, the most precise WEP test to date.
• Pound-Rebka 1959. Mössbauer-effect γ-rays from Fe-57 emitter at the top of a 22.5 m Harvard tower were detected at the bottom. Gravitational blueshift Δf/f = gh/c² ≈ 2.46 × 10⁻¹⁵ was tuned out by Doppler-shifting the absorber. Achieved 10% agreement with Einstein's prediction; later improved to 1% in 1964. First terrestrial test of gravitational redshift.
• Lunar laser ranging. Apollo 11/14/15 astronauts placed corner reflectors on the Moon. Pulsed lasers from McDonald Observatory (Texas) measure round-trip travel time to mm precision. Tests strong equivalence: if the Earth's gravitational binding energy fell differently from ordinary matter, the lunar orbit would show a periodic anomaly with monthly period. Constrains Nordtvedt parameter to 10⁻⁴.
• Atom interferometry. Cold-atom experiments split a single atom's wavefunction into two paths at different heights, recombine, and measure the gravitational phase shift. Compares acceleration of different atomic species — rubidium-87 vs. rubidium-85, for instance. Stanford and Wuhan groups achieve ~10⁻⁹ precision; future space-based projects (STE-QUEST, MICROSCOPE-II) aim for 10⁻¹⁵ in quantum tests.
• Pulsar timing. Binary pulsars (the Hulse-Taylor PSR B1913+16 system being the prototype) provide test beds for both WEP and SEP. Period decay matches GR's predicted gravitational-wave emission to 0.2%. Also constrains universality of free fall in strong-gravity regimes — neutron stars are highly bound (binding energy ~10% of mass-energy), so they probe SEP in a way solar-system tests cannot.

From equivalence to redshift to GR

• Einstein's elevator. Imagine a sealed elevator in deep space accelerating uniformly at g. By the equivalence principle, an observer inside cannot distinguish this from the elevator at rest on Earth. So all consequences of acceleration — for light, clocks, dropped objects — also occur in gravity.
• Light bending. A light pulse traveling horizontally across the elevator: by the time it reaches the far wall, the elevator has accelerated up by ½gt². Inside, the light appears to fall — bending downward. By the equivalence principle, light must bend in a gravitational field. This already predicts (half of) the GR light-bending angle.
• Gravitational redshift. A photon emitted at the floor with frequency ω₀ traveling upward to a detector at height h: by the time it arrives, the detector has accelerated upward at velocity v ≈ gh/c, Doppler-redshifting the photon by Δω/ω = −gh/c². By equivalence, photons climbing in gravity lose frequency by the same factor. Pound-Rebka confirmed.
• Gravitational time dilation. A clock at the bottom of a gravitational well emits ticks. Each tick is a redshifted photon. The number of ticks per second received at the top is reduced by the redshift factor. The clock at the bottom appears to run slow. Equivalent statement: clocks tick faster at higher gravitational potential.
• From local to global geometry. The equivalence principle works locally: in a small region, free fall mimics inertia. But globally, free-falling worldlines from different regions don't agree — Earth's gravitational field varies in direction. The global consistency is the curvature of spacetime: free-falling geodesics deviate from each other (tidal forces) precisely because the metric has nonzero Riemann curvature.
• Mathematical statement. EEP says: at any spacetime event, you can choose coordinates such that g_μν(P) = η_μν (Minkowski) and ∂g_μν/∂x^λ |_P = 0. The first nonzero derivatives ∂²g/∂x² encode tidal forces — Riemann curvature. The principle thus says spacetime is locally flat (Minkowski) but generically curved at second order.

Common misconceptions

• "Global equivalence — gravity always vanishes in free fall." Only locally. Tidal forces — second derivatives of gravitational potential — cannot be transformed away by any frame choice. A free-falling lab spanning a few meters near Earth notices that "gravity" still pulls its top and bottom ends slightly together (vertical tidal force). Equivalence is a statement about a single point and its infinitesimal neighborhood.
• "All forms of equivalence are tested equally well." No. WEP is tested to 10⁻¹⁵ (MICROSCOPE), EEP to ~10⁻⁵ (atomic clock comparisons of redshift), SEP to ~10⁻⁴ (lunar laser ranging). The strong form is the most fragile in modified-gravity theories and the least precisely constrained — though current experiments aim to improve all three by orders of magnitude.
• "Galileo proved the equivalence principle." He demonstrated WEP at ~1% precision through dropping experiments and inclined-plane reasoning. The principle as Einstein understood it — extending to all physical phenomena, not just dropped objects — required 20th-century reformulation. Galileo's experiments were a starting point, not a definitive test.
• "Gravity is not a force." A common slogan to convey GR. More precisely: in the curved-spacetime picture, free-falling bodies follow geodesics with no force acting; the apparent force we feel is the constraint force of the ground pushing us up, preventing geodesic motion. So saying "gravity is not a force" emphasizes the geometric viewpoint, but Newton's force description still works to enormous precision in weak fields.
• "The equivalence principle violates Newton's third law." No. In the geodesic picture, the apparent acceleration is a property of the geometry, not a force. Newton's third law applies to actual forces. When you stand on the ground, the ground pushes you up (a real force) and you push the ground down (the reaction force). What's missing is the "force of gravity pulling you down" — that's encoded in the geometry, not as a direct force.
• "Inertial mass = gravitational mass is just labeling." No — it's a deep experimental fact with no a priori justification in Newtonian mechanics. In Newton's equations, m_inertial and m_gravitational play distinct roles, and their numerical equality is an unexplained coincidence. Einstein's contribution was promoting this coincidence to a principle and using it to construct GR.
• "Equivalence implies black holes have no event horizons." Confused. Equivalence is a local statement. The event horizon of a black hole is a global, gauge-invariant feature of the spacetime — the boundary of the causal past of future infinity. Free-falling observers crossing the horizon don't notice anything locally (consistent with equivalence), but the horizon's existence is unambiguous.
• "You can derive all of GR from equivalence alone." Almost — but not quite. The equivalence principle gives gravity its geometric character, but the dynamics (Einstein's field equations linking matter to curvature) require a separate input: the choice of action (Einstein-Hilbert) or a postulate that the field equations involve at most second derivatives of the metric. Equivalence is necessary but not sufficient.