Gravity & the equivalence principle

gravitational bending of light

Gravity bends the path of light. A ray of starlight grazing the edge of the Sun does not travel in a perfectly straight line; it is deflected by a tiny angle, so the star appears slightly shifted from where it really sits. We do not normally notice because the bending is minute and the Sun's glare hides nearby stars — but during a total eclipse, with the Sun blacked out, those stars become visible and their displacement can be measured.

You can feel why it must happen from the equivalence principle. Picture a sealed lab accelerating upward and a laser beam crossing it horizontally. By the time the light reaches the far wall, the lab has moved up a little, so the beam strikes lower than where it entered — its path bends downward inside the lab. Since acceleration is locally equivalent to gravity, a beam crossing a real gravitational field must bend too. Light has no rest mass, but it still carries energy and momentum, and gravity acts on it; more deeply, light is simply following the straightest available path through a spacetime that mass has curved.

This prediction made Einstein world-famous overnight. In May 1919, expeditions led by Arthur Eddington photographed stars near the eclipsed Sun and found the shift matched general relativity, which predicts about 1.75 arcseconds at the Sun's edge — twice what a naive Newtonian 'light as falling particles' estimate gives. The doubling matters: it is not enough that light bends, it must bend by the full relativistic amount, and it does. Today the same effect, magnified by whole galaxies, is a workhorse tool of astronomy called gravitational lensing.

θ = 4GM/(c² b) ≈ 1.75″ at the Sun's edge (twice the Newtonian estimate)

A ray passing a mass M at distance b is deflected by 4GM/(c²b) — exactly double the naive Newtonian value.

A Newtonian 'falling corpuscle' picture predicts only half the bending. General relativity's extra factor of two comes from the curvature of space itself, not just time; measuring the full value is what made the 1919 result a genuine test.

Also called
deflection of light by gravitylight bending光线偏折