Pitfalls in Spin, Symmetry, and Geometric Phase

A wrong rotation sign, an unstated field transformation, or a missing hypothesis can change the meaning of a symmetry calculation. The examples below show how to locate the assumption that fails.

Required background. Use notation and conventions to fix the active-rotation and angular-momentum signs.

Confusing a Transformation with a Symmetry

Mistake. Calling every transformation a symmetry of the system.

Correction. A transformation is a symmetry only when it preserves the relevant structure. For a time-independent Hamiltonian and unitary $U$, the test is

$$UHU^\dagger=H.$$

A rotation can act on a state even when it is not a symmetry of the Hamiltonian. Check the specific rotation or group being claimed; invariance under rotations about one axis is weaker than full rotational invariance. For an unbounded Hamiltonian, its domain and boundary conditions must also be preserved.

Mixing Active and Passive Transformations

Mistake. Losing a sign in rotation formulas because active rotations of states and passive rotations of coordinate axes are treated as the same operation.

Correction. This volume uses active rotations by default:

$$U(\hat{\mathbf n},\theta) = \exp\left( -\frac{i}{\hbar}\theta\,\hat{\mathbf n}\cdot\mathbf J \right).$$

Passive coordinate changes often use the inverse transformation. State which viewpoint is being used before comparing formulas.

For example, $U_y(\pi/2)|+z\rangle=|+x\rangle$ with this active convention. The inverse rotation gives $|-x\rangle$. See active and passive transformations for the wavefunction version.

Treating Spin as a Tiny Rotating Ball

Mistake. Explaining spin-$1/2$ as literal rotation of an extended classical object.

Correction. Spin is an intrinsic angular momentum degree of freedom described by representations of the rotation group or its double cover. Spinors have transformation properties that no ordinary three-dimensional vector has, including a sign change under a $2\pi$ rotation.

Confusing SO(3) and SU(2)

Mistake. Assuming that the rotation group acting on physical space and the group acting on spinors are identical.

Correction. $SO(3)$ is the group of proper rotations of ordinary three-dimensional vectors. $SU(2)$ is its double cover and naturally acts on spin-$1/2$ spinors. Two elements of $SU(2)$ correspond to the same element of $SO(3)$. See SU(2) versus SO(3) for the canonical comparison.

Forgetting the Spinor Sign Under a Full Rotation

Mistake. Assuming every $2\pi$ rotation acts as $+I$ on state vectors.

Correction. For spin-$1/2$,

$$U(2\pi)=-I.$$

The physical ray is unchanged, so this sign is not directly observable for a single isolated state vector. Relative signs in interference and multi-path settings, however, can matter.

Treating Antiunitary Symmetries Like Ordinary Unitaries

Mistake. Moving complex numbers through time reversal as if the operator were unitary.

Correction. An antiunitary operator $T$ conjugates complex scalars:

$$T(a\lvert\psi\rangle+b\lvert\phi\rangle) = a^*T\lvert\psi\rangle+b^*T\lvert\phi\rangle.$$

This is why time reversal changes $i$ to $-i$ in operator equations. In particular, if $\Theta H\Theta^{-1}=H$, then $\Theta e^{-iHt/\hbar}\Theta^{-1}=e^{+iHt/\hbar}$: it reverses evolution rather than commuting with the propagator at the same time.

Mixing Angular-Momentum Phase Conventions

Mistake. Combining Clebsch–Gordan coefficients, spherical harmonics, and ladder-operator formulas from sources with different phase conventions.

Correction. State the convention. This volume uses the standard Condon–Shortley convention unless otherwise declared. A different convention is not wrong, but mixing conventions silently can flip signs in matrix elements and selection-rule calculations.

Misusing Parity

Mistake. Assuming that parity remains a conserved energy-state label when the Hamiltonian is not invariant under inversion.

Correction. Parity is useful when the Hamiltonian satisfies

$$PHP^{-1}=H.$$

If the potential is not inversion symmetric, parity may still be a transformation, but it is not a symmetry of that Hamiltonian and need not label energy eigenstates. One can still prepare parity eigenstates or use an even/odd basis; the Hamiltonian can mix those sectors during evolution.

Treating Berry Phase as a Dynamical Phase

Mistake. Attributing every phase accumulated during evolution to the energy integral.

Correction. The dynamical phase depends on energy and elapsed time. Berry phase depends on the geometry of a path in parameter space. For a cyclic adiabatic process, the total phase contains both pieces.

Calling every geometric phase quantized

Mistake. Assuming that a Berry phase cannot change continuously because it is geometric.

Correction. For a spin-$1/2$ eigenstate followed adiabatically around a field-direction loop, the Berry phase is proportional to the enclosed solid angle and changes continuously as the loop changes. An integer Chern number instead integrates curvature over an appropriate closed two-dimensional parameter surface. Its stability requires that the isolated band or eigenbundle remain well-defined. Geometry and topology are related, but their invariants are different.

Assuming Degeneracy Always Means Symmetry

Mistake. Inferring a symmetry from any observed degeneracy.

Correction. Symmetry is a powerful explanation for robust degeneracy, but accidental degeneracies can occur. Identify the symmetry representation and ask whether symmetry-preserving perturbations may split the level. A nonabelian multiplet can enforce degeneracy; a one-dimensional parity representation does not. Every degenerate subspace admits mathematical transformations commuting with the Hamiltonian, so the physical question is which independently specified symmetry explains and protects the degeneracy.

Exercises

1. A Hamiltonian $H=p^2/(2m)+V(x)$ is transformed by parity. What condition on $V(x)$ is needed for parity to be a symmetry?

Solution

Parity maps $x$ to $-x$ and $p$ to $-p$. The kinetic term is unchanged. The potential term is unchanged only if

$$V(-x)=V(x).$$

Thus the potential must be even. The spatial domain and boundary conditions must also be inversion invariant; an even differential expression on an asymmetric interval is not enough.

2. Explain why $U(2\pi)=-I$ for spin-$1/2$ does not contradict the fact that physical states are rays.

Solution

The vectors $\lvert\psi\rangle$ and $-\lvert\psi\rangle$ represent the same ray, so a single isolated spinor is physically unchanged by the overall sign. The sign becomes meaningful only when compared with another amplitude, for example when only one arm of an interferometer undergoes the spin rotation. A $2\pi$ rotation on a fixed spin-$1/2$ tensor factor alone still multiplies the entire joint state by $-1$; making the system composite does not by itself make a global phase observable.

References

• M. V. Berry, “Quantal phase factors accompanying adiabatic changes,” Proceedings of the Royal Society A 392, 45–57, 1984.
• A. R. Edmonds, Angular Momentum in Quantum Mechanics, Princeton University Press, 1957.
• D. J. Griffiths and D. F. Schroeter, Introduction to Quantum Mechanics, 3rd ed., Cambridge University Press, 2018.
• J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, 3rd ed., Cambridge University Press, 2020.
• E. P. Wigner, Group Theory and Its Application to the Quantum Mechanics of Atomic Spectra, Academic Press, 1959.