0
What Is Gravity?
The weakest force, the only one that always wins
Gravity is 10³⁸ times weaker than electromagnetism, yet it alone shapes the Universe on large scales — because it is long-ranged, universally attractive, and couples to everything. Newton described it as a force; Einstein replaced the force with the geometry of spacetime itself: mass-energy tells spacetime how to curve, and curved spacetime tells matter how to move. General relativity has passed every direct test for a century, from millimeter-scale torsion balances to merging black holes.
And yet gravity is where the deepest cracks in physics show. Taken at face value with general relativity, the Universe requires ~95% invisible content — dark matter and dark energy — inferred only through gravity. Either the inventory is real, or the theory needs modification at low accelerations or large scales. Meanwhile gravity stubbornly resists quantization, and its vacuum energy prediction misses by ~120 orders of magnitude — the worst discrepancy in science. This sheet covers the standard theory, its precision tests, and the serious alternatives:
Newtonian gravity — the working theory of galactic and planetary dynamics. General relativity — the standard model of gravitation. MOND — modified dynamics at low accelerations, no dark matter. Scalar-tensor / f(R) — extra fields and curvature terms, often aimed at dark energy. Emergent gravity — gravity as thermodynamics, not a fundamental force.
As on the companion galaxies, cosmology, black-hole, relativity, and Big Bang sheets, every equation is paired with a plain-language reading of what it physically asserts, a Use in Research column with key references, and each section ends with the open unknowns. Toggle the Dark theme at top-right for a dark background.
I
Newtonian Foundations
5 equationsNewton's 1687 law remains the working theory of gravity for nearly all of astrophysics — planetary systems, stellar dynamics, galaxy simulations. Everything later in this sheet is a correction to, or a replacement of, these five lines.
| Name | Equation | Variables | Use in Research |
|---|---|---|---|
| Universal Gravitation | \[ F = \frac{G\,m_1 m_2}{r^2} \]
Every mass attracts every other, with a force falling as the square of separation. One law unified the falling apple and the orbiting Moon — the first physical law shown to hold beyond Earth. |
G = 6.674×10⁻¹¹ m³ kg⁻¹ s⁻²; r = separation |
The force law inside every N-body code and orbit integrator; exact enough to navigate spacecraft across the Solar System.
Key referencesNewton (1687, Principia); CODATA (2022) for G.
|
| Poisson Equation | \[ \nabla^2\Phi = 4\pi G\rho \]
The field-theory form of Newton's law: a mass distribution sources a gravitational potential everywhere in space. Solve it and every orbit in the system follows. |
Φ = gravitational potential; ρ = mass density |
The equation every galaxy-dynamics model and cosmological simulation actually solves (via tree codes, particle-mesh, or multipole expansions).
Key referencesPoisson (1813); Binney & Tremaine (2008).
|
| Escape Velocity | \[ v_{\rm esc} = \sqrt{\frac{2GM}{r}} \]
The speed at which kinetic energy exactly cancels gravitational binding. Below it you fall back; above it you coast away forever. Setting it equal to c anticipates the black hole. |
M = enclosed mass; r = starting radius |
Sets spacecraft launch budgets, atmospheric escape from planets, and which stars are bound to the Galaxy — the local escape speed constrains the Milky Way's halo mass.
Key referencesMichell (1784); Deason et al. (2019).
|
| Kepler's Third Law | \[ P^2 = \frac{4\pi^2 a^3}{G(M_1+M_2)} \]
Orbital period squared scales with orbit size cubed, divided by total mass. The Universe's universal scale: measure a period and a size, and you have weighed something. |
P = period; a = semi-major axis; M = masses |
The mass-measurement engine of astronomy — binary stars, exoplanets, the black hole at the Galactic center — all are weighed with this one law.
Key referencesKepler (1619); Newton (1687); GRAVITY Collaboration (2019).
|
| Tidal Force & Roche Limit | \[ F_{\rm tid} \approx \frac{2GMm\,\Delta r}{r^3};\quad d_{\rm Roche}\approx2.44\,R\left(\frac{\rho_M}{\rho_m}\right)^{1/3} \]
Gravity's gradient stretches extended bodies — near side pulled harder than far side. Inside the Roche limit the stretching beats self-gravity and the body is torn apart. |
Δr = body size; d_Roche = disruption distance; ρ_M, ρ_m = densities |
Explains ocean tides, planetary rings, tidal disruption events around black holes, and the tidal streams that map the Milky Way's accretion history.
Key referencesRoche (1849); Murray & Dermott (1999); Rees (1988, TDEs).
|
Open unknowns · Newtonian Foundations
The Value of G
Why can't laboratories agree on Newton's constant?
G is the least precisely known fundamental constant — careful experiments disagree by ~5×10⁻⁴, far beyond their quoted errors. Whether this is unmodeled systematics or something stranger is an open embarrassment of metrology.
Is G Constant in Time?
Has the strength of gravity changed over cosmic history?
Lunar laser ranging bounds \(\dot G/G\) below ~10⁻¹³ per year today, but many alternatives (Brans–Dicke, Dirac's large-number hypothesis) predict drift. Big Bang nucleosynthesis constrains G at t ≈ 3 minutes to ~10% — a 13.8-Gyr lever arm with gaps.
Gravity at Short Range
Does the inverse-square law survive below a tenth of a millimeter?
Torsion-balance tests confirm 1/r² only down to ~50 μm. Extra compact dimensions or light scalars would show up as deviations just below current reach — the best-motivated place a fifth force could still hide.
II
General Relativity
6 equationsEinstein's 1915 theory replaces force with geometry. A century of tests — from the 1919 eclipse to LIGO — has confirmed it wherever it has been directly checked. The companion relativity sheet covers the kinematics; here are the gravitational essentials.
| Name | Equation | Variables | Use in Research |
|---|---|---|---|
| Einstein Field Equations | \[ G_{\mu\nu} + \Lambda g_{\mu\nu} = \frac{8\pi G}{c^4}T_{\mu\nu} \]
The master equation: spacetime curvature (left) is sourced by mass-energy and pressure (right). Ten coupled nonlinear equations replacing Newton's one — "matter tells spacetime how to curve." |
G_μν = Einstein curvature tensor; T_μν = stress-energy; Λ = cosmological constant |
The starting point of all relativistic astrophysics: cosmological models, black-hole and neutron-star structure, gravitational waveforms — all are solutions of this equation.
Key referencesEinstein (1915); Misner, Thorne & Wheeler (1973).
|
| Geodesic Equation | \[ \frac{d^2x^\mu}{d\tau^2} + \Gamma^\mu_{\alpha\beta}\frac{dx^\alpha}{d\tau}\frac{dx^\beta}{d\tau} = 0 \]
The other half of the theory: free bodies follow the straightest possible paths through curved spacetime. There is no gravitational force — orbits are inertia in disguise. |
Γ = Christoffel symbols (the metric's gradients); τ = proper time |
What you integrate to trace photon paths (lensing, black-hole imaging) and orbits in strong fields — every ray-traced black-hole image is this equation, pixel by pixel.
Key referencesEinstein (1915); EHT Collaboration (2019, 2022).
|
| Schwarzschild Metric | \[ ds^2 = -\left(1-\tfrac{r_s}{r}\right)c^2dt^2 + \frac{dr^2}{1-r_s/r} + r^2d\Omega^2 \]
The exact spacetime outside any spherical mass, found within weeks of Einstein's paper. The radius r_s = 2GM/c² marks the event horizon — the surface of no return. |
r_s = 2GM/c² = Schwarzschild radius; dΩ = angular element |
The baseline solution for orbits, light bending, and time dilation around stars, neutron stars, and non-spinning black holes; the template the EHT and X-ray spectroscopy test against.
Key referencesSchwarzschild (1916); Chandrasekhar (1983).
|
| Gravitational Redshift | \[ \frac{\Delta\nu}{\nu} \approx \frac{\Delta\Phi}{c^2} \quad (\text{weak field}) \]
Clocks deeper in a gravitational potential tick slower, and light climbing out loses frequency. Not an optical illusion — time itself runs at different rates at different heights. |
ΔΦ = potential difference; ν = frequency |
Corrected for in GPS, measured in white-dwarf spectra and galaxy clusters, and now detected across centimeters of height by optical lattice clocks.
Key referencesPound & Rebka (1960); GRAVITY Collaboration (2018); Bothwell et al. (2022).
|
| Light Deflection | \[ \alpha = \frac{4GM}{c^2 b} \]
Light bends around mass by twice the Newtonian value — the factor of 2 is pure spacetime curvature. Its confirmation in 1919 made Einstein famous overnight. |
b = impact parameter; α = deflection angle |
The foundation of all gravitational lensing — cluster mass maps, microlensing planet searches, weak-lensing cosmology — astronomy's only direct probe of total mass, dark or luminous.
Key referencesEinstein (1916); Dyson, Eddington & Davidson (1920); Fomalont et al. (2009).
|
| Quadrupole Formula (GW) | \[ L_{\rm GW} = \frac{G}{5c^5}\left\langle \dddot{Q}_{ij}\dddot{Q}_{ij} \right\rangle \]
Accelerating masses with a changing quadrupole radiate ripples of spacetime at the speed of light. The c⁻⁵ makes the effect fantastically weak — until masses are compact and relativistic. |
Q_ij = mass quadrupole moment; L_GW = radiated power |
Predicts binary orbital decay (the first, indirect GW detection) and underlies every LIGO/Virgo waveform; merging black holes briefly outshine the luminous Universe in this channel.
Key referencesEinstein (1918); Hulse & Taylor (1975); Abbott et al. (2016).
|
Open unknowns · General Relativity
Singularities
Does nature really contain points of infinite density?
GR predicts its own breakdown inside black holes and at the Big Bang. Penrose's cosmic censorship — that singularities always hide behind horizons — remains unproven, and what actually happens there awaits quantum gravity.
Extra Polarizations
Do gravitational waves carry only the two polarizations GR allows?
GR permits exactly two tensor modes; scalar-tensor theories add up to four more. Current detector networks weakly constrain this — a clean detection of a scalar mode would falsify GR outright.
Interior of Black Holes
Is the Kerr metric the whole story at and inside the horizon?
Ringdown spectroscopy and EHT images test the exterior; exotic horizonless alternatives (gravastars, fuzzballs) predict subtle echoes and shadow deviations. No anomaly found yet — but precision is only ~10%.
Energy of the Gravitational Field
Why can't gravitational energy be localized?
GR admits no local energy density for gravity itself — only global definitions work. This century-old conceptual gap complicates everything from quantum gravity to numerical relativity bookkeeping.
III
Precision Tests of Gravity
6 equationsGravity is the most precisely tested interaction in physics — and the framework below is how alternatives are confronted with data. Every viable rival theory must thread all of these needles simultaneously.
| Name | Equation | Variables | Use in Research |
|---|---|---|---|
| PPN Expansion | \[ g_{00} = -1 + \frac{2GM}{rc^2} - 2\beta\left(\frac{GM}{rc^2}\right)^2;\;\; g_{ij} = \left(1+2\gamma\frac{GM}{rc^2}\right)\delta_{ij} \]
The parametrized post-Newtonian framework: write the metric of any conceivable metric theory with free parameters, then measure them. GR predicts γ = β = 1 exactly. |
γ = space curvature per unit mass; β = nonlinearity of superposition |
The universal language of Solar-System tests — every experiment reports a γ or β constraint that any alternative theory must satisfy.
Key referencesNordtvedt (1968); Will (2014, Living Review); Bertotti et al. (2003).
|
| Perihelion Precession | \[ \Delta\varphi = \frac{6\pi GM}{c^2 a(1-e^2)} \;\text{per orbit} \]
Orbits in GR are not closed ellipses — they slowly rotate. The famous 43″/century excess of Mercury, unexplained for 60 years, was GR's first triumph. |
a = semi-major axis; e = eccentricity |
Still a working tool: the same formula (with spin terms) tracks S-star orbits at the Galactic center and relativistic binary pulsars, where precession reaches degrees per year.
Key referencesEinstein (1915); Le Verrier (1859); GRAVITY Collaboration (2020).
|
| Shapiro Time Delay | \[ \Delta t = \frac{4GM}{c^3}\ln\!\left(\frac{4 r_1 r_2}{b^2}\right) \]
Light passing near a mass is not only bent but delayed — spacetime near mass is "deeper," and signals take longer to cross it. The fourth classical test, found in 1964. |
r_1, r_2 = emitter/receiver distances; b = impact parameter |
Measured with planetary radar and spacecraft tracking; in binary pulsars it weighs the companion star; in lensed quasars and GW170817 it tests whether photons and gravitons fall alike.
Key referencesShapiro (1964); Reasenberg et al. (1979); Kramer et al. (2021).
|
| Equivalence Principle (Eötvös) | \[ \eta = 2\,\frac{|a_1 - a_2|}{a_1 + a_2} \]
Do all bodies fall identically, regardless of composition? The foundation stone of GR — if any two materials fall differently by even one part in 10²⁰, the geometric picture of gravity collapses. |
a_1, a_2 = accelerations of test bodies; η = Eötvös ratio |
The sharpest null test in physics, and the graveyard of fifth-force proposals; any new light scalar coupling to matter generically violates it.
Key referencesEötvös (1922); Touboul et al. (2022); Adelberger et al. (2009).
|
| Binary-Pulsar Orbital Decay | \[ \frac{dP_b}{dt} = -\frac{192\pi G^{5/3}}{5c^5}\left(\frac{P_b}{2\pi}\right)^{-5/3}\!\!f(e)\,\frac{m_1 m_2}{(m_1+m_2)^{1/3}} \]
Gravitational-wave emission drains orbital energy, so compact binaries spiral together at a precisely predicted rate. Pulsar clocks make the shrinkage measurable to seconds per century. |
P_b = orbital period; f(e) = eccentricity factor; m₁, m₂ = masses |
The strongest strong-field test of GR — radiative, not just orbital — and the original proof that gravitational waves exist, 40 years before LIGO heard them.
Key referencesHulse & Taylor (1975); Taylor & Weisberg (1982); Kramer et al. (2021).
|
| Speed of Gravity | \[ \left|\frac{c_{\rm GW}}{c} - 1\right| \;{\lt}\; 5\times10^{-16} \]
GW170817's gravitational waves and gamma-rays arrived 1.7 s apart after 130 million years in flight — gravity travels at the speed of light to one part in a quadrillion. |
c_GW = gravitational-wave speed; bound from GW170817/GRB 170817A |
A single measurement that eliminated entire families of modified gravity: any theory predicting \(c_{\rm GW}\neq c\) (many Horndeski branches, TeVeS variants, covariant Galileons built to mimic dark energy) died on 2017 August 17.
Key referencesAbbott et al. (2017, GW170817); Ezquiaga & Zumalacárregui (2017); Creminelli & Vernizzi (2017).
|
Open unknowns · Precision Tests
The Untested Regimes
Where has gravity never actually been measured?
All precision tests live in high-acceleration, small-scale environments. The low-acceleration regime (below ~10⁻¹⁰ m/s², where MOND operates) and cosmological scales are probed only with dark matter and dark energy already assumed.
Quantum Equivalence
Does the equivalence principle hold for quantum superpositions and antimatter?
ALPHA-g (2023) showed antihydrogen falls down, but only to ~25% precision. Whether superposed or entangled systems fall universally — and what "free fall" even means for them — is barely tested territory.
Dark-Sector Equivalence
Does dark matter obey the equivalence principle?
All EP tests use ordinary matter. A fifth force acting only between dark particles — dark matter falling differently from baryons — is constrained only loosely by tidal streams and galaxy-pair dynamics.
Lorentz Violation in Gravity
Is gravity exactly Lorentz invariant?
Quantum-gravity scenarios (Hořava gravity, some string vacua) predict tiny preferred-frame effects. Pulsar timing and the SME framework bound dozens of violation coefficients, but a complete test is impossible in principle — the search is open-ended.
IV
MOND — Modified Newtonian Dynamics
5 equationsMilgrom's 1983 proposal: below a critical acceleration a₀ ≈ 1.2×10⁻¹⁰ m/s², dynamics departs from Newton — and dark matter in galaxies becomes unnecessary. Forty years on, MOND's galaxy-scale predictions keep succeeding while its cluster-scale and cosmological extensions keep struggling. Whatever its ultimate status, every galaxy phenomenologist must know these equations.
| Name | Equation | Variables | Use in Research |
|---|---|---|---|
| The MOND Law | \[ \mu\!\left(\frac{a}{a_0}\right)a = a_N,\quad a_0\approx1.2\times10^{-10}\,\text{m/s}^2 \]
Above a₀, gravity is Newtonian; far below it, the interpolation function μ → a/a₀ and the true acceleration exceeds the Newtonian one. One new constant of nature replaces galactic dark matter. |
a_N = Newtonian acceleration; μ(x) = interpolation function; a_0 = Milgrom's constant |
The phenomenological law you fit to rotation curves with the baryons alone — typically with one free parameter (stellar M/L) where dark-halo fits use three.
Key referencesMilgrom (1983a,b,c); Sanders & McGaugh (2002, review); Famaey & McGaugh (2012, Living Review).
|
| Deep-MOND Limit | \[ a = \sqrt{a_N\,a_0} \;\;\Rightarrow\;\; v_{\rm flat}^4 = G M_{\rm bar}\,a_0 \]
In the low-acceleration limit, the effective force falls as 1/r instead of 1/r² — so rotation curves go exactly flat forever, and the flat velocity depends only on baryonic mass. |
v_flat = asymptotic rotation speed; M_bar = total baryonic mass |
This is the baryonic Tully–Fisher relation derived, not fitted — slope 4, zero intrinsic scatter, no dependence on size or surface brightness. It was a prediction made 17 years before the data confirmed it.
Key referencesMilgrom (1983); McGaugh et al. (2000); Lelli et al. (2016).
|
| Radial Acceleration Relation | \[ g_{\rm obs} = \frac{g_{\rm bar}}{1-e^{-\sqrt{g_{\rm bar}/a_0}}} \]
Point by point inside galaxies, the observed acceleration is a fixed function of the acceleration the baryons alone produce. The dark matter — if that is what it is — tracks the baryons with startling fidelity. |
g_obs = observed acceleration; g_bar = baryonic (Newtonian) acceleration |
The cleanest empirical statement of the "MOND phenomenology": 2,700 points across 153 SPARC galaxies collapse onto one curve with ~0.1 dex scatter. Any theory — dark matter or modified gravity — must explain this regularity.
Key referencesMcGaugh, Lelli & Schombert (2016); Lelli et al. (2017); Navarro et al. (2017, ΛCDM view).
|
| External Field Effect | \[ a_{\rm ext} \;{\gt}\; a_{\rm int} \;\Rightarrow\; \text{internal dynamics re-Newtonized} \]
Because MOND depends on total acceleration, a system's internal dynamics changes when it is embedded in an external field — even a uniform one. This violates the strong equivalence principle: in MOND, where you are changes how gravity works inside you. |
a_ext = external field; a_int = internal acceleration |
MOND's most distinctive falsifiable signature — no dark-matter model mimics it. Isolated dwarfs should behave differently from identical dwarfs near a giant host.
Key referencesMilgrom (1983); Bekenstein & Milgrom (1984); Chae et al. (2020, 2021).
|
| The a₀ Coincidence | \[ a_0 \approx \frac{cH_0}{2\pi} \approx \frac{c^2\sqrt{\Lambda/3}}{2\pi} \]
Milgrom's constant — fitted purely to galaxy rotation curves — numerically matches the cosmic acceleration scale built from the Hubble constant or the cosmological constant. Coincidence, or a clue that galactic dynamics knows about cosmology? |
H_0 = Hubble constant; Λ = cosmological constant |
The numerological fact that keeps theorists returning to MOND: it suggests the low-acceleration anomaly is tied to the dark-energy scale or the de Sitter horizon, motivating emergent-gravity derivations.
Key referencesMilgrom (1983, 1999); Verlinde (2016); Smolin (2017).
|
Open unknowns · MOND
The Cluster Problem
Why does MOND still need unseen mass in galaxy clusters?
Even in MOND, clusters show a factor ~2 mass deficit, and the Bullet Cluster's lensing peak follows the collisionless component, not the gas. MOND advocates invoke sterile neutrinos or missing baryons — an uncomfortable echo of dark matter.
Fundamental or Emergent?
Is MOND a law of nature or the residue of deeper physics?
a₀'s match to cH₀ hints the phenomenology is emergent — from dark-matter–baryon coupling, superfluid dark matter, or entanglement entropy. If so, MOND is a constraint every theory must reproduce, not a theory itself.
Is the EFE Real?
Will the external field effect survive better data?
A confirmed EFE would falsify all dark-matter interpretations at a stroke; a confirmed absence would falsify MOND. Wide binaries from Gaia were proposed as the clean test — and competing analyses currently reach opposite conclusions.
Dwarf Spheroidal Stress Tests
Do the Milky Way's satellites obey MOND?
Some dwarfs (Crater II's low dispersion was a successful MOND prediction) fit beautifully; others (Dragonfly 44, some ultra-faints) strain it. Tidal effects and the EFE complicate every case — no consensus scorecard exists.
V
Relativistic Alternative Theories
6 equationsTo compete with GR, a theory must be relativistic: it must do lensing, cosmology, and gravitational waves. The standard construction adds fields (scalars, vectors) or curvature terms to the Einstein–Hilbert action — and the 2017 neutron-star merger executed a large fraction of the candidates overnight.
| Name | Equation | Variables | Use in Research |
|---|---|---|---|
| Einstein–Hilbert Action | \[ S = \frac{c^4}{16\pi G}\int R\,\sqrt{-g}\;d^4x + S_{\rm matter} \]
All of GR from one line: vary this action and the field equations follow. Its stark simplicity — just the curvature scalar R — is the template every alternative modifies, extends, or replaces. |
R = Ricci scalar; g = metric determinant |
The reference point of gravitational theory-building. Lovelock's theorem proves it is essentially unique in 4D — so any alternative must add fields, dimensions, or higher derivatives. That theorem organizes this entire section.
Key referencesHilbert (1915); Lovelock (1971); Clifton et al. (2012, review).
|
| Brans–Dicke Scalar-Tensor | \[ S = \int\!\left[\phi R - \frac{\omega}{\phi}(\partial\phi)^2\right]\!\sqrt{-g}\,d^4x;\quad G_{\rm eff}\sim\frac{1}{\phi} \]
The prototype alternative (1961): Newton's "constant" becomes a dynamical field φ, so gravity's strength varies in space and time. The parameter ω measures how stiff the field is — ω → ∞ recovers GR. |
φ = gravitational scalar; ω = Brans–Dicke coupling |
The benchmark against which all scalar-tensor gravity is measured, and the ancestor of inflation models, dilaton gravity, and Horndeski. Solar-System data have driven its parameter into a corner.
Key referencesBrans & Dicke (1961); Bertotti et al. (2003); Will (2014).
|
| f(R) Gravity | \[ S = \frac{c^4}{16\pi G}\int f(R)\,\sqrt{-g}\;d^4x \]
Replace the curvature scalar with a function of it. Terms growing at low curvature can drive cosmic acceleration with no dark energy; an R² term at high curvature gives Starobinsky inflation — still the best-fitting inflationary model. |
f(R) = function of Ricci scalar (e.g. Hu–Sawicki, Starobinsky forms) |
The workhorse modified-gravity model in cosmology surveys: mathematically equivalent to a scalar-tensor theory, it predicts scale-dependent structure growth that DESI, Euclid, and Rubin explicitly test for.
Key referencesStarobinsky (1980); Hu & Sawicki (2007); Sotiriou & Faraoni (2010, review).
|
| TeVeS & Relativistic MOND | \[ g_{\mu\nu}^{\rm phys} = e^{-2\phi}g_{\mu\nu} - 2u_\mu u_\nu\sinh(2\phi) \]
Bekenstein's tensor-vector-scalar theory: matter feels a metric disformally built from Einstein's metric, a scalar, and a vector field — engineered so weak fields reproduce MOND while light bends as if dark matter were present. |
φ = scalar; u_μ = unit timelike vector; physical vs. Einstein metric |
The proof-of-concept that MOND could be made relativistic — and the cautionary tale: original TeVeS predicts \(c_{\rm GW}\neq c\) and was effectively killed by GW170817. Its successor is the live research front.
Key referencesBekenstein (2004); Boran et al. (2018); Skordis & Złośnik (2021).
|
| Massive Graviton | \[ \Phi(r) = -\frac{GM}{r}\,e^{-r/\lambda_g},\quad \lambda_g = \frac{h}{m_g c} \]
If the graviton has mass, gravity acquires a Yukawa cutoff at the graviton's Compton wavelength — weakening at large distances, which could mimic cosmic acceleration. GR demands the graviton be exactly massless. |
m_g = graviton mass; λ_g = Compton wavelength |
Modern "dRGT" massive gravity (2010) finally evaded the classic ghost instability, reviving the field. LIGO constrains m_g through frequency-dependent dispersion of the waveform across the inspiral.
Key referencesFierz & Pauli (1939); de Rham, Gabadadze & Tolley (2010); Abbott et al. (2021, GWTC-3).
|
| Horndeski Theory | \[ \mathcal{L} = \sum_{i=2}^{5} \mathcal{L}_i[g_{\mu\nu},\phi] \;\;(\text{most general 2nd-order scalar-tensor}) \]
The most general theory of a metric plus one scalar field with second-order equations of motion (hence no Ostrogradsky ghosts) — the master framework containing Brans–Dicke, f(R), Galileons, and quintessence as special cases. |
L_2…L_5 = Lagrangian terms with free functions of φ and (∂φ)² |
The umbrella in which modern dark-energy/modified-gravity phenomenology is parametrized (α_M, α_B, α_K, α_T functions); survey constraints are typically quoted as cuts through Horndeski space.
Key referencesHorndeski (1974); Deffayet et al. (2011); Ezquiaga & Zumalacárregui (2017).
|
Open unknowns · Relativistic Alternatives
Anything Left Standing?
Which alternatives actually survive the post-GW170817 cuts?
Theories must now deliver c_GW = c, pass Solar-System screening, fit the CMB, and remain ghost-free simultaneously. The surviving space (conformally coupled scalars, RMOND, some vector-tensor models) is small and largely untested against structure growth.
The Ghost Problem
Why do almost all modifications of GR destabilize the vacuum?
Higher-derivative terms generically introduce ghost degrees of freedom with negative kinetic energy. Whether nature exploits the narrow ghost-free corridors (Horndeski, dRGT) or the instabilities signal that GR shouldn't be modified at all is unresolved.
RMOND vs. Precision Cosmology
Can a relativistic MOND survive Planck-era data end to end?
Skordis–Złośnik fits the CMB, but its dynamical fields must also reproduce lensing of the CMB, baryon acoustic oscillations, and Lyman-α clustering without re-inventing dark matter under another name. The full gauntlet has not yet been run.
Black Holes as Discriminators
Do alternative theories predict different black holes?
Many scalar-tensor theories share the Kerr metric ("no-hair") but differ in perturbations — ringdown frequencies, scalarization of neutron stars. Next-generation detectors (Einstein Telescope, LISA) will test these at the percent level; theory predictions are incomplete.
VI
Screening Mechanisms & Cosmological Tests
5 equationsAny theory that modifies gravity enough to matter cosmologically would naively wreck the Solar System. Screening mechanisms hide the modification where density is high — making the theories viable, and making the search for them a hunt in low-density environments and large-scale structure.
| Name | Equation | Variables | Use in Research |
|---|---|---|---|
| Chameleon Mechanism | \[ m_\phi^2(\rho) = V''_{\rm eff}(\phi) \;\;\uparrow\;\; \text{with}\;\rho \]
The scalar field's mass depends on ambient density: heavy (short-ranged, invisible) in the Solar System, light (long-ranged, active) in the cosmic vacuum. The fifth force literally hides in dense environments. |
m_φ = effective scalar mass; ρ = local matter density |
What makes f(R) and similar theories viable at all — and what defines the observational strategy: compare screened (dense) and unscreened (void, dwarf-galaxy) environments for differential gravity.
Key referencesKhoury & Weltman (2004); Hamilton et al. (2015); Burrage & Sakstein (2018, review).
|
| Vainshtein Radius | \[ r_V = \left(r_s\,\lambda_g^2\right)^{1/3} \]
In massive gravity and Galileon theories, the field's own nonlinear self-interactions suppress the fifth force within a huge radius of any mass. Screening by derivative self-coupling rather than by mass. |
r_s = Schwarzschild radius; λ_g = graviton Compton wavelength |
Explains why dRGT massive gravity and braneworld models pass Solar-System tests; the residual percent-level deviations just inside r_V are targets for lunar laser ranging and planetary ephemerides.
Key referencesVainshtein (1972); Babichev & Deffayet (2013, review); Dvali, Gabadadze & Porrati (2000).
|
| Growth Index | \[ f(z) \equiv \frac{d\ln\delta}{d\ln a} \approx \Omega_m(z)^{\gamma},\quad \gamma_{\rm GR}\approx0.55 \]
How fast cosmic structure grows is a clean discriminant: GR with ΛCDM predicts γ ≈ 0.55; modified gravity changes it (DGP gives 0.68). Expansion history and growth history must agree — or gravity is modified. |
δ = matter overdensity; γ = growth index; Ω_m(z) = matter fraction |
The headline modified-gravity statistic of redshift surveys, measured via redshift-space distortions as fσ₈(z). The current generation (DESI, Euclid) aims at percent precision on γ.
Key referencesPeebles (1980); Linder (2005); DESI Collaboration (2024).
|
| E_G Statistic | \[ E_G = \frac{\nabla^2(\Psi+\Phi)}{3H_0^2 a^{-1} f\,\delta} \;\;\xrightarrow{\rm GR}\;\; \frac{\Omega_{m,0}}{f(z)} \]
The ratio of lensing (which feels both metric potentials) to galaxy velocities (which feel only one). In GR the two potentials are equal and E_G is parameter-free; most alternatives split them apart. |
Ψ, Φ = time/space metric potentials; f = growth rate; δ = overdensity |
A model-independent gravity test built entirely from observables (galaxy–galaxy lensing + clustering + RSD), designed so galaxy bias cancels. The cosmological analogue of the PPN γ test.
Key referencesZhang et al. (2007); Reyes et al. (2010); Blake et al. (2020).
|
| Dark-Energy Equation of State | \[ w(a) = w_0 + w_a(1-a);\quad w_\Lambda = -1 \;\text{exactly} \]
Is cosmic acceleration a constant of nature (Λ, w = −1 forever) or a dynamical field — possibly a symptom of modified gravity? The CPL parametrization (w₀, wₐ) is where the answer will first show. |
w = pressure/density ratio; w_0, w_a = present value and evolution |
The discriminant between Λ, quintessence, and self-accelerating modified gravity. Any confirmed deviation from (−1, 0) ends the cosmological-constant era of cosmology.
Key referencesChevallier & Polarski (2001); Linder (2003); DESI Collaboration (2024, 2025).
|
Open unknowns · Screening & Cosmic Tests
Is Dark Energy Evolving?
Will the DESI hint of w ≠ −1 survive?
A confirmed evolving equation of state would be the discovery of the decade — but the signal currently leans on combining datasets with different systematics. Independent confirmation (Euclid, Rubin, Roman) is the critical next step.
The Tensions as Gravity Clues
Are the H₀ and σ₈ tensions symptoms of modified gravity?
Local and early-Universe measurements of expansion (and of structure growth) disagree at the ~5σ and ~2–3σ levels. Most modified-gravity fixes break something else, but the tensions remain the strongest empirical motivation to keep looking.
Screening Makes It Unfalsifiable?
Can a perfectly screened theory ever be ruled out?
Screening is double-edged: it saves theories from Solar-System data but can hide them from every local test. Whether there exist theories that are cosmologically relevant yet locally undetectable in principle is a live methodological worry.
Void Gravity
Does gravity behave differently in cosmic voids?
Voids are the least screened environments in the Universe — the natural place for fifth forces to show. Void lensing and void galaxy dynamics are young, systematics-limited probes with discovery potential.
VII
Emergent & Quantum Gravity
5 equationsThe deepest alternative is that gravity is not fundamental at all — that spacetime and its dynamics emerge from quantum information, the way temperature emerges from molecules. The clues all point through black-hole thermodynamics and the Planck scale.
| Name | Equation | Variables | Use in Research |
|---|---|---|---|
| Bekenstein–Hawking Entropy | \[ S_{\rm BH} = \frac{k_B c^3 A}{4G\hbar} \]
A black hole's entropy is its horizon area in Planck units, divided by 4. Entropy scaling with area, not volume, is the single deepest clue we possess about quantum gravity — information in a region lives on its boundary. |
A = horizon area; all four fundamental constants in one formula |
The anchor of the holographic principle, string-theoretic microstate counting, and every emergent-gravity proposal. Counting what these microstates are is the benchmark problem of quantum gravity.
Key referencesBekenstein (1973); Hawking (1975); Strominger & Vafa (1996).
|
| Planck Scale | \[ \ell_P = \sqrt{\frac{\hbar G}{c^3}} = 1.6\times10^{-35}\,\text{m};\quad E_P = 1.2\times10^{19}\,\text{GeV} \]
The unique length, time, and energy built from ħ, G, and c — where a particle's Compton wavelength equals its Schwarzschild radius and quantum gravity becomes unavoidable. Sixteen orders beyond any collider. |
ℓ_P = Planck length; E_P = Planck energy; t_P = 5.4×10⁻⁴⁴ s |
Sets the stage for all quantum-gravity phenomenology: the only hopes of access are cosmological relics (primordial GWs, inflationary signatures) and cumulative effects on light from cosmological distances.
Key referencesPlanck (1899); Amelino-Camelia (2013, review); Abdo et al. (2009, Fermi).
|
| Gravity as Thermodynamics | \[ \delta Q = T\,dS \;\;\Rightarrow\;\; G_{\mu\nu} = \frac{8\pi G}{c^4}T_{\mu\nu} \]
Jacobson's 1995 derivation: demand that heat, temperature, and entropy behave correctly across every local horizon, and Einstein's equations follow — as an equation of state, not a fundamental law. Gravity as the thermodynamics of spacetime. |
δQ = energy flux; T = Unruh temperature; S = horizon entropy |
The intellectual foundation of the emergent-gravity program: if GR is an equation of state, quantizing it directly is as misguided as quantizing the ideal-gas law — a reframing of the entire quantum-gravity problem.
Key referencesJacobson (1995); Unruh (1976); Padmanabhan (2010).
|
| Entropic Gravity | \[ F = T\,\frac{\Delta S}{\Delta x} \;\;\Rightarrow\;\; F = \frac{GMm}{r^2} \]
Verlinde's proposal: gravity is an entropic force — matter changes the information content of holographic screens, and the statistical tendency toward maximum entropy manifests as attraction. Newton's law derived from information theory. |
T = screen temperature; ΔS = entropy change; Δx = displacement |
The boldest emergent claim with astronomical consequences: the 2016 extension predicts apparent extra gravity in galaxies — MOND-like, with a₀ ~ cH₀ derived rather than fitted — directly testable against lensing data.
Key referencesVerlinde (2011, 2016); Brouwer et al. (2017); Lelli et al. (2017, critique).
|
| Semiclassical Gravity & Its Limit | \[ G_{\mu\nu} = \frac{8\pi G}{c^4}\langle\hat T_{\mu\nu}\rangle \]
The current working compromise: classical spacetime sourced by the quantum expectation value of matter. It underlies Hawking radiation and inflationary perturbations — and it cannot be the final word, because measurement collapses ⟨T⟩ discontinuously. |
⟨T̂_μν⟩ = expectation value of the quantum stress-energy operator |
The framework behind every result connecting quantum theory and gravity to date (black-hole evaporation, the CMB fluctuation spectrum). Where it breaks down is precisely where full quantum gravity must take over.
Key referencesMøller (1962); Hawking (1975); Mukhanov & Chibisov (1981); Planck Collaboration (2020).
|
Open unknowns · Emergent & Quantum Gravity
The Cosmological Constant Problem
Why is the vacuum's gravity wrong by ~120 orders of magnitude?
Quantum field theory predicts a vacuum energy density that should curve the Universe into a ball smaller than an atom; the observed Λ is ~10¹²⁰ times smaller. The worst prediction in the history of physics is also gravity's deepest unsolved problem.
Quantize or Emerge?
Is gravity a quantum force at all?
String theory and loop quantum gravity quantize it; Jacobson and Verlinde say that's a category error. No experiment has ever probed gravity's quantum nature — proposed tabletop tests of graviton-mediated entanglement between masses may give the first data this decade.
The Hierarchy Problem
Why is gravity 10³⁸ times weaker than the other forces?
Supersymmetry and large extra dimensions were the leading explanations; the LHC found neither. The weakness of gravity — equivalently the lightness of the Higgs against the Planck scale — currently has no accepted explanation.
Is Spacetime Fundamental?
Do space and time themselves emerge from entanglement?
Holography (AdS/CFT) suggests geometry is built from quantum entanglement ("ER = EPR"), but the construction works only in unrealistic universes so far. Extending it to our accelerating cosmos is a defining open problem of theoretical physics.
Observable Quantum Gravity
Will any quantum-gravity effect ever be measured?
Primordial B-mode polarization in the CMB would be gravitons from inflation — quantum gravity seen. Current bounds (r < 0.03) keep tightening without a detection; next-generation experiments (CMB-S4, LiteBIRD) are the field's best observational hope.
Gravity reference values: \(G = 6.674\times10^{-11}\) m³ kg⁻¹ s⁻² (least-known fundamental constant, ~2×10⁻⁵ relative uncertainty); \(r_s = 2GM/c^2\): Sun 2.95 km, Earth 8.9 mm; solar light deflection 1.75″; Mercury precession 43″/century; PPN: \(|\gamma-1| {\lt} 2.3\times10^{-5}\) (Cassini), \(|\beta-1| {\lt} 10^{-4}\); equivalence principle \(\eta {\lt} 3\times10^{-15}\) (MICROSCOPE); double-pulsar GR agreement 1.3×10⁻⁴; \(|c_{\rm GW}/c-1| {\lt} 5\times10^{-16}\) (GW170817); graviton mass \({\lt}1.27\times10^{-23}\) eV/c²; \(\dot G/G {\lt} 10^{-13}\)/yr (LLR); inverse-square law verified to ~50 μm; \(a_0 = 1.2\times10^{-10}\) m/s² ≈ cH₀/2π; growth index \(\gamma_{\rm GR}\approx0.55\); \(\ell_P = 1.616\times10^{-35}\) m, \(E_P = 1.22\times10^{19}\) GeV; vacuum-energy discrepancy ~10¹²⁰.