Anyons and Fractional Statistics

Every student learns that identical particles come in exactly two kinds — bosons, whose wavefunction is unchanged when two are swapped, and fermions, which pick up a minus sign. That dichotomy feels like a law of nature. It is not. It is a fact about the topology of three-dimensional space. Swap two particles, and what you have really done is drag one along a path that loops halfway around the other. In three dimensions any such loop can be lifted out of the plane and unwound, so swapping twice is the same as doing nothing: the exchange operator squares to the identity, and only the phases $+1$ and $-1$ survive. Take away that third dimension and the argument collapses. In a plane, a loop that encircles another particle is genuinely stuck — you cannot contract it without passing through the particle. Clockwise and counterclockwise exchanges become physically different operations, and swapping twice is not the identity. The permutation group $S_N$ , whose only one-dimensional pictures are the two we know, gives way to the braid group $B_N$ , which admits a continuous circle of exchange phases $e^{i\theta}$ . Particles that carry such a phase — anything between boson and fermion — are anyons, and the angle $\theta$ is their statistics. This is not a mathematical curiosity. Anyons are the real, measured excitations of two-dimensional electron gases in the fractional quantum Hall regime, where quasiparticles carry a fraction $e/m$ of the electron's charge and braid with angle $\pi/m$ . Push further and the exchange stops being a mere phase and becomes a matrix acting on a protected, degenerate space — non-Abelian anyons, whose topological rigidity is the physical foundation of one entire proposal for fault-tolerant quantum computing. This article closes the identical-particles chapter by showing where the two-sector story ends and a richer one begins.

Key takeaways

Particle statistics is set by the topology of the exchange path, not by any local force — the swap is a physical loop of one particle around another.
In three or more dimensions that loop unwinds, forcing the exchange operator to square to the identity, so only bosons ( $+1$ ) and fermions ( $-1$ ) exist.
In two dimensions the loop is topologically trapped; the braid group $B_N$ replaces the permutation group $S_N$ , and any exchange phase $e^{i\theta}$ becomes legal — these are anyons.
Abelian anyons appear as fractional quantum Hall quasiparticles with charge $e/m$ and statistics $\theta=\pi/m$ (for $\nu=1/3$ : charge $e/3$ , angle $\pi/3$ ); the flux-tube model tunes $\theta$ by binding charge to magnetic flux.
Non-Abelian anyons live in a degenerate ground space where braiding is a non-commuting unitary — the basis of topological quantum computation.
Key idea: exchange is a braid, not a swap. In $\geq 3$ D every braid is its own inverse and only $\pm1$ survives; in $2$ D clockwise and counterclockwise part ways, and the full continuum $e^{i\theta}$ of anyonic statistics opens up.

The core insight

Statistics is a topological property of the exchange path: in three or more dimensions a double exchange contracts to nothing, permitting only bosons and fermions, whereas in two dimensions the braid group allows a continuous statistical angle $\theta$ and a whole family of anyons.

Where this lives today

Anyons stopped being a thought experiment around 2020, when two groups reported direct signatures of fractional braiding statistics: an "anyon collider" measuring cross-correlated current noise from colliding $\nu=1/3$ quasiparticles, and a Fabry–Pérot interferometer resolving the discrete $2\pi/3$ phase jumps as anyons are added inside the loop. Together with decades-old shot-noise measurements of the $e/3$ charge, these close the loop on Abelian anyons: the phase predicted by topology in 1977 is now on an oscilloscope. The frontier is non-Abelian. Ising anyons — Majorana zero modes at the ends of engineered superconducting wires, or in the $\nu=5/2$ quantum Hall state — would encode a qubit non-locally and process it by braiding, giving hardware-level protection against decoherence. This is the physical premise behind topological quantum computing programs. The experimental status is genuinely open: the theory is clean, but unambiguous demonstration of non-Abelian braiding remains a leading challenge in condensed-matter physics.

The physical picture

Hold two identical particles in your mind and swap them. The naive picture is instantaneous relabeling, but physically nothing is instantaneous: you have to move one particle along a continuous path to where the other was, and vice versa. The swap is a loop, and the question "what phase does the wavefunction pick up?" becomes "what loops are inequivalent, and what phases can they carry?" That is a question of topology, answered by the fundamental group of the space of distinct configurations. In three dimensions the answer is dull, and its dullness is exactly why we only ever see two kinds of particle. A path that exchanges two particles twice — sending each around the other in a full loop — can always be lifted into the third dimension and shrunk to a point, like sliding a rubber band off the end of a pencil. So a double exchange must act trivially, the exchange operator obeys $\hat P^2=\mathbb{1}$ , its eigenvalues are $\pm1$ , and we recover bosons and fermions. The Permutation Symmetry and the Exchange Operator article proves this two-sector result in full and states its own frontier explicitly: "in two dimensions the symmetric group gives way to the braid group." This article is that frontier.

Figure

Why 3D erases the braid but 2D keeps it — a double exchange loop

Where does a continuous phase come from physically? The cleanest model glues each particle to an infinitesimally thin solenoid carrying magnetic flux $\Phi$ — a "flux tube." A charge $q$ carried once around a flux $\Phi$ acquires the Aharonov–Bohm phase $q\Phi/\hbar$ , and a single exchange is exactly half of such a loop. The statistics is then literally an interference phase, $\theta = q\Phi/(2\hbar)$ , tunable to any value by dialing the flux. There is nothing mysterious about a particle "knowing" how to be an anyon; it is winding a charge around a flux.

Figure

The statistical angle interpolates: bosons, anyons, fermions on one circle

For non-Abelian anyons the ground state at fixed particle positions is not unique. There is a protected, degenerate Hilbert space whose dimension grows like $d^N$ with the number of anyons, and braiding does not merely multiply by a phase — it rotates the state within that space. Because the rotation depends only on which worldline crosses over which, not on the detailed shape of the paths, small wiggles leave it unchanged. That topological rigidity is the whole point: information stored in the pattern of braids is immune to local noise.

The core equations

The braid group is defined by generators $\sigma_i$ — a counterclockwise exchange of neighboring strands $i$ and $i+1$ — subject to two relations. The crucial one is the braid (Yang–Baxter) relation, which says a three-strand braid can be slid past itself; the second says distant exchanges commute. What is conspicuously absent is any relation like $\sigma_i^2=\mathbb{1}$ .

\sigma_i\,\sigma_{i+1}\,\sigma_i = \sigma_{i+1}\,\sigma_i\,\sigma_{i+1},\qquad \sigma_i\,\sigma_j=\sigma_j\,\sigma_i\ \ (|i-j|\ge 2) \tag{1}

where: $\sigma_i$ is the generator braiding strand $i$ over strand $i+1$ ; $|i-j|\ge 2$ marks non-adjacent pairs. Read physically: the first relation is the only constraint braids in a plane must obey, and it does not make $\sigma_i$ its own inverse — the source of everything that follows. The one-dimensional (Abelian) unitary representations of $B_N$ send every generator to the same phase, because the braid relation forces all $\sigma_i$ to have equal image. That single free phase is the statistical angle.

\sigma_i \mapsto e^{i\theta},\qquad \theta\in[0,2\pi) \tag{2}

where: $\theta$ is the statistical angle — $0$ for bosons, $\pi$ for fermions, and every other value an Abelian anyon. Because the value is continuous, the anyon "sits between" the two familiar cases rather than defining a third discrete kind. Concretely, exchanging two anyons counterclockwise multiplies the many-body wavefunction by that phase:

\psi(\dots \mathbf r_j \dots \mathbf r_i \dots) = e^{i\theta}\,\psi(\dots \mathbf r_i \dots \mathbf r_j \dots) \tag{3}

where: $\mathbf r_i,\mathbf r_j$ are the positions of the two exchanged anyons. Unlike the boson/fermion case, the sense of the exchange matters: a clockwise swap gives $e^{-i\theta}$ , the inverse, and the two are not equal unless $\theta\in\{0,\pi\}$ . The physical origin of a tunable $\theta$ is the flux tube. One full counterclockwise encirclement of one composite about another is topologically two exchanges, so it carries twice the statistical angle, and that phase is Aharonov–Bohm.

\Delta\varphi_{\text{loop}} = \frac{q\Phi}{\hbar} = 2\theta \quad\Longrightarrow\quad \theta = \frac{q\Phi}{2\hbar} \tag{4}

where: $q$ is the bound charge, $\Phi$ the attached flux, and $\Delta\varphi_{\text{loop}}$ the phase for one full loop (two exchanges). Setting $q\Phi/\hbar = 2\pi$ recovers $\theta=\pi$ , a fermion — the composite-fermion picture that underlies flux attachment in the quantum Hall problem. For the fractional quantum Hall effect at filling $\nu=1/m$ , the many-electron ground state is Laughlin's, and its charged excitations are the anyons. We quote the wavefunction and its consequences rather than re-derive the FQHE spectrum:

\psi_m = \prod_{i<j}(z_i-z_j)^{m}\,\exp\!\Big(-\tfrac{1}{4\ell_B^2}\sum_k |z_k|^2\Big),\qquad e^* = \frac{e}{m},\quad \theta = \frac{\pi}{m} \tag{5}

where: $z_k=x_k+iy_k$ is the complex planar coordinate of electron $k$ , $m$ an odd integer, and $\ell_B=\sqrt{\hbar/eB}$ the magnetic length. The Laughlin quasiholes carry a fraction $e/m$ of the electron charge and braid with angle $\pi/m$ ; at $\nu=1/3$ that is charge $e/3$ and statistics $\pi/3$ . When several anyons share a degenerate ground space of dimension $\dim\mathcal H_N \sim d_a^{\,N}$ , the generator $\sigma_i$ is a matrix $U_{\text{braid}}$ , and the quantum dimension $d_a$ measures how fast that space grows ( $d=1$ Abelian, $d=\sqrt2$ Ising/Majorana, $d=\varphi$ Fibonacci, with $\varphi$ the golden ratio).

Derivation ATwo-particle configuration space and the emergence of θ

We want to see, from topology alone, why a plane permits a continuous statistical angle while three-dimensional space forbids it. The strategy is to strip the two-particle problem down to its relative coordinate, remove the forbidden coincidence point, and ask what phases a single-valued wavefunction may carry around the resulting hole. The payoff is a clean statement: in 2D a half-rotation may multiply $\psi$ by any $e^{i\theta}$ , while in 3D the same rotation is constrained to $\pm1$ . Watch the puncture — it is the whole story.

Step 1. Separate center-of-mass and relative motion; the statistics lives entirely in the relative coordinate $\mathbf r = \mathbf r_1-\mathbf r_2$ . In the plane write it in polar form $\mathbf r=(\rho,\phi)$ , and remove the point $\rho=0$ : two identical particles cannot sit on top of each other, so the relative space is the punctured plane $\mathbb R^2\setminus\{0\}$ . Why puncture it? Because the exchange operation must be a smooth motion in the space of allowed configurations, and $\rho=0$ is not an allowed configuration.

Step 2. An exchange sends $\mathbf r\to-\mathbf r$ , i.e. $\phi\to\phi+\pi$ — a half-turn about the puncture. The punctured plane deformation-retracts onto a circle, so its fundamental group is $\mathbb Z$ : loops are classified by an integer winding number about the hole.

\pi_1\big(\mathbb R^2\setminus\{0\}\big)=\mathbb Z

Why this matters: a nontrivial fundamental group is exactly what allows a nontrivial phase. A simply connected space would force every loop, hence every double exchange, to be trivial.

Step 3. A single-valued phase representation of $\mathbb Z$ assigns to the half-turn any phase $e^{i\theta}$ , so the wavefunction transforms as

\psi(\rho,\phi+\pi)=e^{i\theta}\,\psi(\rho,\phi),\qquad \theta\in[0,2\pi).

A full $2\pi$ rotation is two exchanges and gives $e^{2i\theta}$ , which need not equal $1$ — the loop encircles the puncture and cannot be undone. Every $\theta$ is consistent.

Step 4. Now redo the count in three dimensions. The relative space is $\mathbb R^3\setminus\{0\}$ , which retracts onto the sphere $S^2$ ; that sphere is simply connected, so a double exchange is contractible and

e^{2i\theta}=1\ \Longrightarrow\ \theta\in\{0,\pi\}.

Only bosons and fermions survive. The single extra dimension is precisely what kills the continuum. Why the sphere and not a circle? In 3D you may swing the exchange path out of any plane; the set of directions from one particle to the other is a full $S^2$ , whose double cover is trivial at the level of loops, unlike the circle's $\mathbb Z$ . *Verified: the structural claim — braid relation holds with $\sigma^2\neq\mathbb 1$ in 2D but collapses to the permutation group's $\pm1$ under $\sigma^2=\mathbb 1$ — is machine-checked in the verifier (Burau rep of $B_3$ vs the standard rep of $S_3$ ), and the 3D constraint $e^{2i\theta}=1$ selects exactly $\theta\in\{0,\pi\}$ .*

Go deeper · Derivation B · 1 minFractional charge and statistics of Laughlin quasiholes

Derivation BFractional charge and statistics of Laughlin quasiholes

The goal is to show that the Laughlin quasihole is genuinely fractional — charge $e/m$ , statistics $\pi/m$ — by transporting one quasihole around a closed loop and reading off the Berry phase. The trick, due to Arovas, Schrieffer and Wilczek, is to split that phase into an Aharonov–Bohm piece (sensitive to enclosed flux) and a statistical piece (sensitive to enclosed quasiholes). We quote the FQHE ground state rather than derive it; the fractional numbers fall out of the adiabatic transport. The subtle part is bookkeeping which enclosed charge contributes what.

Step 1. Insert a quasihole at $\eta$ by multiplying the Laughlin state by $\prod_i(z_i-\eta)$ . This pushes electron density away from $\eta$ , digging a localized correlation hole.

\psi_\eta=\prod_i (z_i-\eta)\,\psi_m

Why this operator? Multiplying by $(z_i-\eta)$ forces the wavefunction to vanish as any electron approaches $\eta$ , which is exactly a depletion of charge there.

Step 2. Counting the depleted density shows the hole is missing a fraction $1/m$ of an electron, so its charge is

e^*=\frac{e}{m},\qquad\text{e.g. }e^*=\tfrac{e}{3}\ \text{at}\ \nu=\tfrac13.

Step 3. Adiabatically drag the quasihole around a loop enclosing area with $N_{\text{enc}}$ electrons. Its Berry phase is the Aharonov–Bohm phase of a charge $e^*$ in the field $B$ :

\gamma_{\text{AB}} = \frac{e^*}{\hbar}\Phi_{\text{enc}} = 2\pi\,\frac{N_{\text{enc}}}{m}.

Why $N_{\text{enc}}/m$ ? Each enclosed electron contributes one flux quantum, and a charge $e/m$ around one flux quantum accrues $2\pi/m$ .

Step 4. Now place a second quasihole inside the loop. The extra Berry phase beyond the smooth background is the full-braid (double-exchange) phase, and half of it is the exchange angle:

\Delta\gamma = \frac{2\pi}{m}=2\theta \ \Longrightarrow\ \theta=\frac{\pi}{m}.

For $m=1$ (a filled Landau level) this gives $\theta=\pi$ — ordinary fermions — a reassuring consistency check that the machinery reduces correctly. Verified: $e^*=e/m$ and $\theta=\pi/m$ are tabulated for $m=1,3,5$ in the verifier, with $m=1\to\theta=\pi$ (fermion) and $m=3\to$ charge $e/3$ , angle $\pi/3$ , full-braid $2\pi/3$ .

Limiting cases

The single angle $\theta$ organizes every case; the table reads off the whole classification.

Case	Result	Meaning
Bosons	$\theta=0,\ e^{i\theta}=+1$	Symmetric wavefunction; the zero-flux limit of the composite model. The Bosons and Fermions article owns this sector.
Fermions	$\theta=\pi,\ e^{i\theta}=-1$	Antisymmetric wavefunction; one flux quantum attached. Connects to the Pauli exclusion principle.
Abelian anyon, $\nu=1/m$	$\theta=\pi/m,\ e^*=e/m$	For $m=3$ : $\theta=\pi/3$ , charge $e/3$ — a genuinely fractional phase strictly between boson and fermion.
Non-Abelian (Ising/Majorana)	braiding $\to U$ , $d=\sqrt2$	Exchange is a matrix on a degenerate space; the $\nu=5/2$ Moore–Read candidate. Quantum dimension $\sqrt2$ .
Three or more dimensions	$\sigma_i^2=\mathbb 1\Rightarrow\theta\in\{0,\pi\}$	The braid group collapses to the permutation group; only bosons and fermions remain.

Worked examples

Example 1 — Laughlin quasiholes at $\nu=1/3$ : charge and braiding phase (enumerate → apply → number).

Step 1. Identify the filling and the odd integer: $\nu=1/3$ means $m=3$ .

Step 2. Apply the quasihole formulas of Eq. (5). The fractional charge is $e^* = e/m = e/3 \approx 0.333\,e$ .

Step 3. The exchange (statistical) angle is $\theta = \pi/m = \pi/3 \approx 1.047$ rad, i.e. $60^\circ$ .

Step 4. The measurable quantity in an interferometer is the full braid phase — one quasihole all the way around another — which is a double exchange, $\Delta\varphi = 2\theta = 2\pi/3 \approx 2.094$ rad ( $120^\circ$ ).

Answer: a $\nu=1/3$ Laughlin quasihole carries charge $\mathbf{e/3}$ and braids two-particle exchange phase $\boldsymbol{\pi/3}$ , giving a full-encirclement phase of $\mathbf{2\pi/3}$ . Both numbers have been measured — the charge by shot noise, the braiding phase by anyon interferometry and collider experiments.

Example 2 — flux attachment: what flux turns a boson into a fermion?

Step 1. Start from Eq. (4): the statistical angle from a charge $q$ bound to flux $\Phi$ is $\theta = q\Phi/(2\hbar)$ .

Step 2. A fermion needs $\theta=\pi$ . Solve: $q\Phi/(2\hbar)=\pi \Rightarrow q\Phi = 2\pi\hbar = h$ .

Step 3. Read this as flux quanta. With $q=e$ , the required flux is $\Phi = h/e = \Phi_0$ , exactly one flux quantum. Attaching one flux quantum to a boson makes it a fermion — the statistical transmutation at the heart of composite-particle theory.

Step 4. Check the half-way point: attaching half a flux quantum, $\Phi=\Phi_0/2$ , gives $\theta=\pi/2$ — a "semion," an anyon exactly midway between boson and fermion, whose double exchange is $e^{i\pi}=-1$ .

Answer: one flux quantum $\boldsymbol{\Phi_0=h/e}$ converts a boson to a fermion; half a flux quantum makes a semion with $\boldsymbol{\theta=\pi/2}$ . This is why "flux attachment" is the standard language for building anyons and composite fermions.

Origins · 1 minFrank Wilczek and how this was discovered

Frank Wilczek

b. 1951 · MIT theorist who named the anyon

Frank Wilczek introduced the word "anyon" in a 1982 paper on the quantum mechanics of fractional-spin particles, giving the flux-tube model that makes a continuous statistical angle concrete. The topological groundwork had been laid five years earlier by Jon Magne Leinaas and Jan Myrheim, who first showed that two-dimensional identical particles obey braid- rather than permutation-statistics. Wilczek's Nobel Prize (2004, shared with Gross and Politzer) was for asymptotic freedom in quantum chromodynamics — an entirely different achievement; the anyon work, though enormously influential, was not what the prize recognized.

History

1977

— Leinaas and Myrheim show 2D identical particles obey braid statistics, a continuum of exchange phases

1982

— Wilczek coins "anyon" and gives the flux-tube model of fractional statistics

1983

— Laughlin writes the ν = 1/m wavefunction; its quasiholes carry charge e/m

1984

— Halperin and, independently, Arovas–Schrieffer–Wilczek establish θ = π/m for the quasiholes

1991

— Moore and Read propose non-Abelian anyons in the ν = 5/2 Pfaffian state

2020

— anyon-collider and interferometry experiments directly measure the fractional braiding phase

Going further

References. Griffiths & Schroeter, Introduction to Quantum Mechanics, 3rd ed., identical-particle chapter · Sakurai & Napolitano, Modern Quantum Mechanics, 2nd ed., Ch. 7 · Shankar, Principles of Quantum Mechanics, 2nd ed., Ch. 10 · Wilczek, Phys. Rev. Lett. 49, 957 (1982) · Leinaas & Myrheim, Nuovo Cimento B 37, 1 (1977) · Arovas, Schrieffer & Wilczek, Phys. Rev. Lett. 53, 722 (1984) · Nayak, Simon, Stern, Freedman & Das Sarma, Rev. Mod. Phys. 80, 1083 (2008).

Provenance. The braid-vs-permutation classification and the 2D continuum of statistics follow Leinaas–Myrheim (1977) and Wilczek (1982); the verifier confirms the braid relation with $\sigma^2\neq\mathbb 1$ (non-Abelian Burau rep of $B_3$ ) and its collapse to the $\pm1$ permutation statistics under $\sigma^2=\mathbb 1$ (standard rep of $S_3$ ). The Laughlin wavefunction, fractional charge $e/m$ , and statistics $\theta=\pi/m$ are standard (Laughlin 1983; Arovas–Schrieffer–Wilczek 1984) and are recomputed for $m=1,3,5$ . Non-Abelian quantum dimensions ( $d=\sqrt2$ Ising, $d=\varphi$ Fibonacci) per Nayak et al. (2008). The AllQFT base was physically sound; no numeric or sign error was found — every displayed value was independently recomputed and the topological arguments cross-checked against the Permutation Symmetry article's stated frontier. Wilczek's 2004 Nobel is correctly attributed to asymptotic freedom (QCD), not to anyons.

Builds on: Indistinguishability of Identical Particles · Permutation Symmetry and the Exchange Operator · Bosons and Fermions · Spin-Statistics Theorem · Symmetric and Antisymmetric States Leads to: Fractional Quantum Hall Effect · Topological Quantum Computation · Pauli Exclusion Principle

Is an anyon a fundamental particle like an electron or a photon?

No. Anyons are emergent quasiparticles — collective excitations of a two-dimensional many-body system, most famously the fractional quantum Hall fluid. They are not elementary particles in vacuum. What makes them "real" is that they carry sharply defined fractional charge and a definite braiding phase that experiments measure, so within their host system they behave as genuine particles with anyonic statistics. The vacuum of our $3{+}1$ -dimensional world admits only bosons and fermions.

If exchange gives e^{iθ}, why does a clockwise swap matter?

Because in two dimensions clockwise and counterclockwise exchanges are topologically distinct operations — one is the inverse of the other, giving $e^{-i\theta}$ versus $e^{i\theta}$ . For bosons ( $\theta=0$ ) and fermions ( $\theta=\pi$ ) the two agree, since $e^{i\pi}=e^{-i\pi}=-1$ , which is exactly why we never had to specify a direction before. For a generic anyon they differ, so you must fix a convention (usually counterclockwise as $\sigma_i$ ) and track the sense of every braid.

How can a fraction of an electron's charge exist? Isn't charge quantized?

The electron charge is quantized, and no fraction of an electron ever appears. The $e/3$ is the charge of a collective excitation of the whole electron fluid — a localized rearrangement of many electrons that carries, in total, one-third of an elementary charge relative to the uniform background. It is an emergent quantum number of the many-body state, fully consistent with every individual electron carrying charge $-e$ . Shot-noise experiments measure the effective charge of the current-carrying excitations, and they read $e/3$ .

Why does the flux-tube phase involve 2θ for a full loop but θ for an exchange?

A single exchange moves one particle halfway around the other — a half-loop. A full encirclement, where one particle returns to its start after going all the way around, is two successive exchanges. So the Aharonov–Bohm phase for a complete loop, $q\Phi/\hbar$ , equals $2\theta$ , and the exchange angle is half of it, $\theta=q\Phi/(2\hbar)$ . Interferometry measures the full-loop phase (it is the topologically robust, gauge-invariant quantity), so experimental papers usually quote $2\theta$ .

What exactly is "non-Abelian" about non-Abelian anyons?

When the ground state at fixed anyon positions is degenerate, each braid generator $\sigma_i$ is represented by a matrix acting on that degenerate space, not by a number. Two such matrices generally do not commute: $\sigma_1\sigma_2 \neq \sigma_2\sigma_1$ . So the order in which you braid anyons changes the final state — "non-Abelian" just means these braiding matrices form a non-commutative group. Abelian anyons, by contrast, have a one-dimensional (non-degenerate) space, so braiding is multiplication by a phase, and order does not matter.

Do non-Abelian anyons let you do universal quantum computation by braiding alone?

It depends on the anyon type. Braiding Ising (Majorana) anyons generates only the Clifford gates — powerful and error-protected, but not universal on their own; you need one extra non-topological "magic state" gate to complete the set. Fibonacci anyons, by contrast, are braiding-universal: their braid group is dense in the relevant unitary group, so braiding alone approximates any gate to arbitrary accuracy. This is why Fibonacci anyons are the theorists' favorite even though Ising/Majorana platforms are experimentally closer.

Where does the golden ratio come from as a "quantum dimension"?

The quantum dimension $d_a$ measures how fast the protected Hilbert space grows as you add anyons: $\dim \sim d_a^N$ . For Fibonacci anyons the fusion rule $\tau\times\tau = 1+\tau$ (two Fibonacci anyons can fuse to either the vacuum or another Fibonacci anyon) makes the dimensions obey the Fibonacci recursion, whose growth rate is exactly the golden ratio $\varphi=(1+\sqrt5)/2\approx1.618$ . So $d_\tau=\varphi$ is not a coincidence of naming — it is the literal asymptotic growth of the computational space, and its irrationality is tied to braiding-universality.