Anyons: Beyond Bosons and Fermions

In three or more spatial dimensions, identical particles come in exactly two kinds: bosons, whose wavefunction is symmetric under exchange, and fermions, whose wavefunction is antisymmetric. In two spatial dimensions this dichotomy breaks down, and a richer continuum of possibilities — anyons — becomes available. Understanding why requires looking at exchange not as an algebraic sign but as a physical process of moving particles around one another.

Statistics from topology of paths

Consider exchanging two identical particles by physically dragging them along a path that swaps their positions. The amplitude for this process is governed by the topology of the swap. In $d \ge 3$ dimensions, any two such exchange paths can be continuously deformed into one another, and doing the exchange twice can always be unwound to the identity. Group-theoretically, the relevant object is the permutation group $S_n$ : a double exchange is the identity, so an exchange operator $R$ must satisfy $R^2 = \mathbb{1}$ , forcing eigenvalues $\pm 1$ — bosons or fermions.

In $d = 2$ the situation changes qualitatively. A worldline cannot be lifted "over" another, so a clockwise exchange is genuinely distinct from a counterclockwise one, and a double exchange (a full $2\pi$ braid of one particle around another) cannot be undone. The relevant group is no longer $S_n$ but the braid group $B_n$ , which is infinite. Because $R^2 \neq \mathbb{1}$ in general, the exchange phase is no longer restricted to $\pm 1$ .

Abelian anyons: any phase

For the simplest anyons, exchanging two identical particles multiplies the state by a phase:

\psi \;\longmapsto\; e^{i\theta}\,\psi.

The boson case is $\theta = 0$ and the fermion case is $\theta = \pi$ , but in two dimensions $\theta$ may take any value in $[0, 2\pi)$ — hence "any-on." A particle with $\theta = \pi/2$ , say, is a genuinely new kind of identical particle, intermediate between boson and fermion. Because the operations all act as commuting phases, these are called abelian anyons.

Non-abelian anyons: matrices, not phases

A still richer possibility arises when a collection of anyons at fixed positions does not have a unique quantum state but spans a degenerate multidimensional space $\mathcal{V}$ . Then exchanging two of them acts as a unitary matrix on $\mathcal{V}$ rather than a mere phase. Successive exchanges generally do not commute,

R_1 R_2 \neq R_2 R_1,

so these are non-abelian anyons. The braid group is represented by noncommuting unitaries, and the sequence of braids — not just the net winding number — determines the final state. This degenerate, braid-controlled space is exactly the resource topological quantum computation exploits.

Why two dimensions is special, restated

The deep reason is the topology of the configuration space of $n$ distinct points. In $d \ge 3$ its fundamental group is the permutation group $S_n$ (finite); in $d = 2$ it is the braid group $B_n$ (infinite). Quantum statistics is a one- (abelian) or higher- (non-abelian) dimensional unitary representation of that fundamental group. The infinite, noncommutative braid group is what opens the door to fractional and non-abelian statistics.

What to take away

Anyons are quasiparticles of two-dimensional topologically ordered media whose exchange is governed by the braid group rather than the permutation group. Abelian anyons pick up an arbitrary phase $e^{i\theta}$ under exchange; non-abelian anyons live in a degenerate space and exchange as noncommuting unitary matrices. The next lessons make both cases concrete.