running of the electromagnetic coupling
We are told the strength of electromagnetism is fixed: the fine-structure constant is about 1/137. But that turns out to be only half the story. The number you get depends on how hard you look. Probe two charges gently, at low energy and large separation, and you measure 1/137. Smash them together at very high energy, peering at tiny distances, and the effective strength climbs — to around 1/128 at the energies of the largest colliders. This change of the coupling with the scale of probing is called the running of the coupling.
The reason is vacuum polarization. A real charge is wrapped in a screening cloud of virtual electron-positron pairs that hide part of it from distant observers. The closer you get, the more you penetrate the cloud and the larger the unscreened charge you see, so the effective coupling grows at short distances. Because in particle physics short distances are reached only with high energy, physicists usually phrase it as: the coupling runs, growing with energy. The mathematics that tracks this change is the renormalization group.
Running couplings are central to modern particle physics, not a curiosity. They mean the 'constants' in the Standard Model are really sliding scales whose values you must quote together with the energy at which they were measured. They also fuel a grand dream: as you extrapolate to enormous energies, the electromagnetic, weak, and strong couplings drift toward one another, hinting they might merge into a single force in a grand unified theory — a tantalizing if still unproven idea.
At a low-energy tabletop experiment, alpha measures about 1/137. In high-energy electron-positron collisions at energies near the Z boson's mass (as at the LEP collider), the effective alpha measures about 1/128 — the same fundamental quantity, simply seen at a finer resolution.
Alpha is not truly constant: it grows from ~1/137 to ~1/128 as the energy rises.
Running does not mean the coupling changes over time or place; it changes with the energy or distance scale at which you probe — a fixed physical fact, not random drift.