Wigner Rotation: Why Two Boosts Don't Commute

What the Wigner Rotation Is: Boosts That Refuse to Add Up

A Lorentz boost is a change of velocity between inertial frames — it mixes time and one spatial direction while leaving perpendicular directions alone. Rotations, meanwhile, mix spatial directions and leave time alone. Together, boosts and rotations generate the full Lorentz group. The crucial structural fact: rotations form a closed subgroup, but boosts do not.

Multiply two boosts in the same direction and you get a third boost (rapidities simply add). But multiply two boosts in different directions, B(u) then B(v), and the product is no longer a pure boost. It factors as a boost composed with a genuine spatial rotation:

• B(v)·B(u) = R(ε)·B(w), where w is the relativistically-combined velocity and R(ε) is the Wigner rotation by angle ε.
• Reverse the order and you get a different combined velocity direction — boosts don't commute.

The leftover rotation R(ε) is the Wigner rotation. It is the price the geometry charges for the fact that the set of pure boosts is not closed under composition — a purely kinematic consequence of the hyperbolic geometry of velocity space.

The Mechanism: Rapidity, Hyperbolic Geometry, and Curvature of Velocity Space

The cleanest way to see the rotation is through rapidity, φ, defined by tanh φ = v/c, so that φ = artanh(β). Collinear boosts add their rapidities linearly, which is why 1D velocity addition looks so tidy. But rapidity space is not flat Euclidean space — it is a hyperbolic space of constant negative curvature (the Lobachevsky/Minkowski velocity space).

On a curved surface, parallel-transporting a direction around a closed loop rotates it — this is holonomy. When you boost out along one rapidity vector and back along another, the round trip encloses a small hyperbolic triangle, and the frame comes back rotated by the triangle's angular defect. For two orthogonal boosts a compact exact result is:

• tan(ε/2) = tanh(φ₁/2)·tanh(φ₂/2), with φ₁, φ₂ the two rapidities.

The general (non-orthogonal, unequal-speed) angle is:

• cos ε = (1+γ+γ_u+γ_v)² / [(1+γ)(1+γ_u)(1+γ_v)] − 1, with the composite factor γ = γ_u γ_v (1 + u·v/c²).

Both encode the same message: no curvature (c → ∞, β → 0) means no rotation.

Characteristic Numbers and a Worked Example

Take two perpendicular boosts of equal speed β = 0.5 (v = 1.5×10⁸ m/s each). Rapidity φ = artanh(0.5) = 0.5493. Then tanh(φ/2) = tanh(0.2747) = 0.2679, and:

• tan(ε/2) = (0.2679)² = 0.07178 → ε/2 = 4.10° → ε ≈ 8.2° (0.143 rad).

The general closed-form formula gives the same ε ≈ 8.2° for this symmetric perpendicular configuration, matching the rapidity result. The key point is scale: at β = 0.5 the misalignment is several degrees, decidedly not negligible.

In the non-relativistic limit, the leading behavior is:

• ε ≈ ½ (v₁ × v₂)/c² — proportional to the cross product of the two velocities.

The famous Thomas factor ½ lives right here. For everyday speeds it is tiny — a satellite at v = 7.7 km/s changing direction by 90° accrues ε ≈ ½·(7.7×10³/3×10⁸)² ≈ 3×10⁻¹⁰ rad — but integrated over an electron's orbit it is decisive.

How It Shows Up: Thomas Precession, Fine Structure, and g−2 Rings

A particle on a curved trajectory is being boosted through an infinite sequence of infinitesimally-different-direction frames. Each step contributes a tiny Wigner rotation; summed, they make the particle's rest frame — and any spin it carries — precess. That is Thomas precession, with angular velocity:

• ω_T = (γ²/(γ+1)) · (a × v)/c², where a is proper acceleration, v the velocity, γ the Lorentz factor.

Its most celebrated appearance is atomic fine structure. The naive spin–orbit coupling energy, computed in the electron's instantaneous rest frame, comes out twice too large. Thomas's 1926 insight was that the electron's rest frame is itself Thomas-precessing, and this halves the result — the Thomas ½ — bringing predicted doublet splittings into agreement with spectra.

It is also built into muon and electron g−2 storage rings (CERN, Fermilab, BNL): the measured spin precession relative to momentum is the anomalous precession precisely because the Thomas–Wigner (T-BMT) kinematics have been subtracted out. Get the Wigner rotation wrong and the extracted anomaly is wrong.

Distinguishing It From Its Cousins

The Wigner rotation is easy to confuse with several neighbors; the distinctions matter:

• Wigner rotation vs. Thomas precession: The Wigner rotation is the discrete rotation from composing two boosts. Thomas precession is its continuous rate, the time-derivative accumulated along a worldline. Same physics, discrete vs. differential.
• vs. ordinary aberration: Aberration changes the direction of light/velocity between frames; the Wigner rotation is the extra twist of the frame's axes beyond what aberration alone accounts for.
• vs. geodetic/de Sitter precession: That is a general-relativistic effect from spacetime curvature (mass), whereas Wigner/Thomas is flat-spacetime kinematics. Gravity Probe B measured both geodetic and frame-dragging precession; neither is the special-relativistic Thomas term.
• vs. Berry phase: Both are holonomies (round-trip rotations from curved parameter spaces), and the analogy is exact — Wigner rotation is the holonomy of hyperbolic velocity space, a genuine geometric phase in the rotation group.

Significance, History, and Open Threads

Llewellyn H. Thomas derived the precession in 1926, rescuing the electron-spin model of fine structure that was otherwise off by a factor of two — a result Bohr called it hard to believe such an elementary point had been missed. Eugene Wigner, in his landmark 1939 paper on unitary representations of the Poincaré group, gave the group-theoretic home for the effect: the little group of stabilizing transformations for a massive particle is the rotation group SO(3), and the rotation induced on a particle's spin states under a boost is exactly the Wigner rotation. This is why spin transforms the way it does under Lorentz transformations.

Modern significance runs deep:

• It underlies relativistic quantum information: a boost can rotate a particle's spin, entangling spin with momentum and altering measured spin entanglement between observers in relative motion.
• It appears in accelerator spin dynamics, polarized-beam transport, and precision g−2 measurements.

Open and pedagogical threads persist: the cleanest geometric derivations (hyperbolic-triangle defect, quaternion methods), the massless-particle limit where the little group becomes ISO(2), and debates over the most physically transparent way to teach the Thomas ½ without sleaning on cancellation-heavy algebra.

Wigner (Thomas–Wigner) rotation angle for two equal-speed perpendicular boosts of speed β = v/c, compared across regimes. ε is exact from the closed-form formula; the small-angle estimate ε ≈ β²/2 (radians) is the leading term.

Boost speed β = v/c	Lorentz factor γ	Exact Wigner angle ε	Small-β estimate β²/2
0.10	1.005	0.29°	0.29°
0.30	1.048	2.7°	2.6°
0.50	1.155	8.2°	7.2° (est. drifts)
0.80	1.667	28°	18° (est. fails)
0.99	7.09	74°	28° (est. useless)
→ 1	→ ∞	→ 90°	diverges