Projective Braid Group Representations from a Microscopic Long-Range Kitaev Chain

#1. Introduction

Topological phases of matter support exotic quantum statistics: when quasiparticles are exchanged adiabatically, the wave function transforms by a unitary matrix, and for non-Abelian anyons these matrices form a representation of the braid group $B_{N}$. A recurring theme across exactly soluble field theories is that such representations are generically projective: the unitaries satisfy the group relations only up to phases, because the phase ambiguity is the fingerprint of a nontrivial central extension or 2-cocycle. Canonical Chern–Simons theory on a Riemann surface provides the canonical example: its wave functions carry projective representations whose cocycles depend on the (possibly rational) Chern–Simons level [1, 2], and the symplectic approach to the Wess–Zumino–Witten model yields projective representations of the loop group by the same mechanism [4].

The purpose of this paper is to derive such a projective braid representation not from a topological field theory but from a microscopic lattice model: the Kitaev chain, generalized to include long-range (power-law decaying) hopping and pairing. The motivation is threefold.

First, microscopic derivations anchor the abstract theory. Chern–Simons derivations [1, 2] are continuum and topological; a lattice derivation shows precisely which microscopic operators generate the braid action and where the projective phase comes from — it comes from the Clifford algebra {$\gamma _{i}, \gamma _{j}$} = $2\delta _{ij}$, i.e., from fermionic anticommutation, not from any gauge field.

Second, long-range hopping changes the locality structure. In the short-range Kitaev chain the MZMs are exponentially localized; with hopping t_ℓ ~ $t_{0}/\ell ^\alpha$, localization and splitting scale differently, and for $\alpha \leq 1$ the zero modes can fail to be localized at all. We quantify this and show that the algebraic structure of the braid representation is independent of $\alpha$, while its physical realizability is not.

Third, there is a live question about what braid representations do and do not determine. Recent analyses indicate that braid group representations of non-Abelian anyons need not uniquely fix the modular data (S and T matrices) of the underlying topological order, with the Ising theory as the canonical test case [12, 13]. A microscopic lattice construction of the braid representation, decoupled from any modular tensor category, is a useful control experiment for such uniqueness questions.

Our main results are: (i) an explicit verification that the Majorana braid matrices satisfy the braid relation exactly and the relation $\sigma ^{4} = 1$ only projectively, with $U^{4}$ = −𝟙 in the fixed-parity sector; (ii) explicit numerical estimates of zero-mode localization, gap, and cumulative phase drift for representative long-range parameters, showing class stability for $\alpha \gt 2$; and (iii) a roadmap for completing the derivation, including the marginal regime $\alpha$ ∈ (1, 2] where the standard short-range arguments require modification.

#2. Background and Related Work

Chern–Simons braid representations. The canonical route to anyon statistics is canonical quantization of Chern–Simons theory coupled to non-dynamical matter. In [1], the quantization is generalized to rational Chern–Simons coefficient k, arbitrary genus, and arbitrary total charge, and the question is posed of which braid group representations on a Riemann surface the wave functions carry. In [2], an explicit solution of the Schrödinger equation for Chern–Simons theory coupled to charged particles is given for rational k and zero total charge, and the wave functions are shown to carry a projective representation of the group of large gauge transformations. These works establish the template our lattice construction mirrors: a quantization ambiguity (rational k there, fermion parity here) produces a projective rather than honest representation.

Symplectic and loop-group perspectives. The symplectic approach to the Wess–Zumino–Witten model [4] obtains the quantum theory as a projective representation of the loop group, with Gauss constraints imposed at a finite set of points. This is conceptually parallel to our construction: the Majorana zero modes are precisely "Gauss constraints at points" (defects), and the projectivity arises from quantizing a symplectic space whose prequantum line bundle has nonzero curvature. The Kashaev quantization of universal Teichmüller space similarly yields projective representations of the Ptolemy–Thompson group and associated dilogarithmic central extensions [5]; the structural lesson — that quantization of a classical symmetry group naturally produces central extensions — is the same one we instantiate for $B_{N}$.

Projective statistics beyond braiding. The analysis in [3] of non-Abelian braid statistics versus projective permutation statistics clarifies that non-Abelian projective representations of the permutation group can serve as statistics valid in any dimension, but that these are strictly different structures from braid statistics. This distinction matters for us: a strictly 1D chain can at best realize the projective-permutation sector unless the chain is embedded in a network that permits genuine exchange. In a related direction, [8] analyzes the three-dimensional model of Teo and Kane, in which defect lines carry Majorana zero modes whose exchange and twist implement a "ghostly" remnant of 2D Ising-anyon braiding; the governing group $T_{2n}$ (projective ribbon permutation group) is computed there. Our work is the 1D-chain analogue: the Kitaev chain's MZMs, moved around a network, implement braid-like unitaries whose projectivity is the same remnant phenomenon.

Algebraic machinery. The Drinfeld associator machinery [6] associates to representations of a certain Hopf algebra braid group representations over a Laurent-series field and studies their dependence on the associator for GT-rigid representations; this rigidity is the algebraic shadow of our physical claim that the cohomology class of the statistics cocycle cannot change without a phase transition. The cluster-geometric construction of [7] endows moduli spaces of G-local systems on decorated surfaces with a cluster Poisson structure equivariant under the mapping class group, providing a geometric arena in which braid-type actions arise from quantization of cluster varieties. Braid group actions on representation categories also appear in [11], where an action of the braid group of a simple Lie algebra on imaginary roots of a quantum affine algebra determines cyclicity of tensor products and zeros of R-matrices; and in [9], where brane quantization connects Lagrangian branes on the $SL(2,\mathbb{C})$ character variety to representations of the spherical double affine Hecke algebra of $C^\vee C_{1}$, with an affine braid group action on the category. Both illustrate that braid actions on zero-mode/defect Hilbert spaces are a robust algebraic phenomenon.

Modular data and uniqueness. Two recent analyses [12, 13] address whether the braid group representation carried by N non-Abelian anyons uniquely determines the modular data of the underlying topological order, with the Ising/Majorana fusion rules as the central test case; the emerging picture is one of non-uniqueness. Our microscopic construction is directly relevant: since the Majorana braid representation we derive in Section 4 is fixed entirely by the Clifford algebra, any "modular data" attributed to it must come from extra structure, not from the braid representation alone. Finally, [14] studies ultrametric relaxation dynamics in topological quantum memory, which bears on the stability of any memory built from our zero modes: the hierarchical relaxation barriers are a plausible model for error processes in a Majorana-based qubit whose protection is weakened by long-range hopping.

We note that reference [10] (a QCD collider-physics report) is part of the provided bibliography but is topically unrelated to this work and is not cited substantively.

#3. Methods

#3.1 The long-range Kitaev chain

The standard Kitaev chain on L sites is $H_{0} = -\Sigma _{j}$ [ $\mu$ $a\dagger _{j}a_{j}$ + t(a†ⱼaⱼ₊₁ + h.c.) + Δ(aⱼaⱼ₊₁ + h.c.) ], with $t, \Delta$ real. We generalize to long-range hopping and pairing with algebraic decay: $H_{\alpha } = -\Sigma _{j} \Sigma _{\ell =1}^{L-j}$ [ (t/ℓ^α)(a†ⱼaⱼ₊ℓ + h.c.) + (Δ/ℓ^α)(aⱼaⱼ₊ℓ + h.c.) ] $- \mu \Sigma _{j} a\dagger _{j}a_{j}$, with power-law exponent $\alpha \gt 0$. The model is in symmetry class BDI (or D with a phase on $\Delta$); in the topological regime $|\mu | \lt 2t \Sigma _\ell \ell ^{-\alpha } = 2t \zeta (\alpha )$ (valid for $\alpha \gt 1$, where $\Sigma \ell ^{-\alpha } = \zeta (\alpha )$), the open chain hosts unpaired MZMs $\gamma _{1}, \gamma _{2}$ at its ends, with $\gamma _{1} = \Sigma _{j} v_{\alpha }(j)(a_{j} + a\dagger _{j}), \gamma _{2} = \Sigma _{j} (-1)_{j} v_{\alpha }(j)(a_{j} + a\dagger _{j})$, where $v_{\alpha }$ is a normalized Majorana wave function whose decay is controlled by $\alpha$. For $\alpha \gt 1$ the decay is exponential with a length $\xi (\alpha )$ that grows as $\alpha \to 1^{+}$; for $\alpha \leq 1$ the zero modes delocalize and topological protection is lost. We treat $\xi (\alpha )$ as a quantity to be bounded, not assumed.

#3.2 From zero modes to a braid action

Given 2n well-separated MZMs $\gamma _{1}$, …, $\gamma _{2n}$ with {$\gamma _{i}, \gamma _{j}$} = $2\delta _{ij}$, an adiabatic exchange of modes i and j implements, on the zero-mode Hilbert space, the unitary $U_{ij} = \exp((\pi /4) \gamma _{i} \gamma _{j})$, whose action on the Majorana operators we verify directly in Section 4. The generators $\sigma _{i}:= U_{i,i+1}$ are the candidate braid generators. The Hilbert space of 2n Majoranas is the Fock space of n complex fermions, with $2^{n}$ states; fixing total parity (superselection) gives a representation of dimension $2^{n-1}$.

#3.3 Exchange protocol in a chain

In a strictly 1D chain, "exchange" is not a native motion; we assume the standard network implementation: the chain is patterned into a T-junction network in which Majoranas are moved by tuning local chemical potentials, or exchanges are implemented by parity measurements. The braid unitary is the same either way; only the adiabaticity condition changes.

#3.4 Cocycle extraction and stability criterion

For a loop ℓ in the ground-state manifold representing a braid $\sigma _{i}$, the geometric phase $\varphi (\ell )$ defines a normalized 2-cocycle $c(\sigma _{i}, \sigma _{j}) = \varphi (\sigma _{i}\sigma _{j}) - \varphi (\sigma _{i}) - \varphi (\sigma _{j}) \bmod 2\pi$ on $B_{N}$ with values in U(1); its class in $H^{2}(B_{N}, U(1)) \cong \mathbb{Z}_{2}$ (for $N \ge 5$, since $H_2(B_N; \mathbb{Z}) \cong \mathbb{Z}_2$ and $H^1(B_N) = 0$, so the universal-coefficient theorem gives $H^2(B_N; U(1)) \cong \mathrm{Hom}(H_2(B_N), U(1)) \cong \mathbb{Z}_2$) is the invariant of interest. The Ising statistics corresponds to the nontrivial element of this $\mathbb{Z}_{2}$; the projective phase $e^{i\pi /8}$ is a representative-level quantity, with the class fixed only by the parity of the cocycle. The stability criterion is: the class is invariant under deformation of the Hamiltonian provided the bulk gap never closes along the path. For the long-range chain, the couplings $t/\ell ^\alpha$ are perturbatively irrelevant for the zero-mode sector when $\alpha \gt 2$ (the analogue of the familiar locality threshold $\alpha \gt d + 1$ in dimension $d = 1$), so the gap stays open and the class is protected; only the cocycle representative can drift, by an amount controlled by the zero-mode overlap.

#4. Analysis

All numbers in this section are computed explicitly from stated inputs.

#4.1 The braid unitary and its action on Majoranas

Let $A = \gamma _{i}\gamma _{j}$ with $i \neq j$. Then $A^{2} = \gamma _{i}\gamma _{j}\gamma _{i}\gamma _{j} = -\gamma _{i}^{2}\gamma _{j}^{2} = -1$, so A has eigenvalues $\pm i$, and $U_{ij} = e^{(\pi /4)A} = \cos(\pi /4)$ 𝟙 $+ \sin(\pi /4)$ A = (1/√2)(𝟙 + $\gamma _{i}\gamma _{j}$). Using the adjoint action: [A, $\gamma _{i}] = A\gamma _{i} - \gamma _{\mathrm{iA}} = -\gamma _{j} - \gamma _{j} = -2\gamma _{j}$ and [A, $\gamma _{j}] = \gamma _{i} - (-\gamma _{i}) = 2\gamma _{i}$. The series $e^{\theta \cdot ad_{A}}\gamma _{i} = \gamma _{i} \cos 2\theta - \gamma _{j} \sin 2\theta$ then gives, at $\theta = \pi /4$: $U_{ij} \gamma _{i} U_{ij}\dagger = -\gamma _{j}, U_{ij} \gamma _{j} U_{ij}\dagger = \gamma _{i}$. The exchange acts as i ↔ j up to a sign convention (either sign is a valid orientation convention for the braid generator).

#4.2 Explicit matrix computation: braid relation and projective lift

Take three Majoranas $\gamma _{1}, \gamma _{2}, \gamma _{3}$; fix the parity sector spanned by {$|0\rangle , \gamma _{3}|0\rangle$}, where $|0\rangle$ is the vacuum of the complex fermion $f = (\gamma _{1} + i\gamma _{2})/2$. In this basis, choosing phases consistently, $\gamma _{1}\gamma _{2} = i\sigma _{3}$ and $\gamma _{2}\gamma _{3} = i\sigma _{1}$. With $\theta = \pi /4$: $U_{12} = e^{(\pi/4)(i\sigma_3)} = \mathrm{diag}(e^{i\pi/4}, e^{-i\pi/4})$, $U_{23} = e^{(\pi/4)(i\sigma_1)} = \tfrac{1}{\sqrt{2}}\begin{pmatrix}1 & i \\ i & 1\end{pmatrix}$.

Fourth power (projectivity). $U_{12}^4 = e^{i\pi\sigma_3} = \mathrm{diag}(e^{i\pi}, e^{-i\pi}) = -I$. So the braid relation $\sigma ^{4} = 1$ is satisfied only up to the phase −1: the representation is projective. Equivalently, $U_{12}^{2} = i\sigma _{3} = i\gamma _{1}\gamma _{2}$: the square of an exchange is the fermion parity operator, and four exchanges differ from the identity by the fermion parity $(-1)^{F}$ — the well-known "fermion parity anomaly" of Majorana braids, here derived rather than assumed.

Braid relation, exact. Direct 2×2 multiplication, first ordering: $\mathrm{U}_{12}\mathrm{U}_{23} = (1/\sqrt{2})[[e^{i\pi /4}, i e^{i\pi /4}], [i e^{-i\pi /4}, e^{-i\pi /4}]]$; right-multiplication by the diagonal $U_{12}$ rescales the columns by $e^{i\pi /4}$ and $e^{-i\pi /4}$, giving $\mathrm{U}_{12}\mathrm{U}_{23}\mathrm{U}_{12} = (1/\sqrt{2})[[e^{i\pi /2}$, i], [$i, e^{-i\pi /2}]] = (1/\sqrt{2})[[i$, i], [i, −i]] = $(i/\sqrt{2})[[1, 1], [1, -1]] = (i/\sqrt{2})(\sigma _{1} + \sigma _{3})$. Second ordering: $\mathrm{U}_{23}\mathrm{U}_{12} = (1/\sqrt{2})[[e^{i\pi /4}, i e^{-i\pi /4}], [i e^{i\pi /4}, e^{-i\pi /4}]]$; left-multiplication by $U_{23}$ gives entrywise $(1,1) = (1/2)(e^{i\pi /4} - e^{-i\pi /4}) = i/\sqrt{2}, (1,2) = (i/2)(e^{i\pi /4} + e^{-i\pi /4}) = i/\sqrt{2}, (2,1) = i/\sqrt{2}, (2,2) = (1/2)(e^{-i\pi /4} - e^{i\pi /4}) = -i/\sqrt{2}$, so $\mathrm{U}_{23}\mathrm{U}_{12}\mathrm{U}_{23} = (i/\sqrt{2})[[1, 1], [1, -1]] = (i/\sqrt{2})(\sigma _{1} + \sigma _{3})$. Both orderings therefore agree exactly, $\mathrm{U}_{12}\mathrm{U}_{23}\mathrm{U}_{12} = \mathrm{U}_{23}\mathrm{U}_{12}\mathrm{U}_{23} = (i/\sqrt{2})[[1, 1], [1, -1]]$, using the same orientation convention ($\gamma _{1}\gamma _{2} = i\sigma _{3}, \gamma _{2}\gamma _{3} = i\sigma _{1}$, forcing $\gamma _{1}\gamma _{3} = -i\sigma _{2}$) for both. The braid relation $\sigma _{1}\sigma _{2}\sigma _{1} = \sigma _{2}\sigma _{1}\sigma _{2}$ therefore holds exactly (not merely projectively) in this representation, while $\sigma ^{4} = 1$ holds only projectively. This is precisely the Ising anyon structure: the exchange phases in the two fusion channels are $R^{\sigma \sigma }_{1} = e^{-i\pi /8}$ and $R^{\sigma \sigma }_{\psi } = e^{3i\pi /8}$ (consistent with the Ising test case of [12]), and the per-exchange projective phase relative to a bosonic baseline is $\varphi _{\mathrm{Ising}} = \pi /8 = 3.14159/8 = 0.3927$ rad.

#4.3 Input numbers

#4.4 Localization and drift bounds

Localization parameter. For a semi-infinite chain with $t(r) = t_{0}r^{-\alpha }$, the MZM decay factor $\lambda$ satisfies, at small $\mu , \lambda \approx \mu /(2t_{0}\zeta (\alpha ))$. At $\alpha = 3, \mu /t_{0} = 0.1$: $\lambda = 0.1/(2 \times 1.2020569) = 0.1/2.4041138 = 0.0416$. The localization length is $\xi$ = 1/|ln $\lambda | = 1/3.1792 = 0.3145$ sites — essentially on-site localized — so the overlap of MZMs separated by s sites is ~ $\lambda ^{s}$.

Drift projection at $\alpha = 3$ (assumption-labeled). We assume, without derivation, that the long-range correction to the exchange phase scales as $\varepsilon$ ~ $\lambda ^{s} \cdot s^{-\alpha }$. No Berry-connection calculation or perturbative derivation of the overlap-to-phase coupling is given here, and no prefactor is fixed; this is a projection, not a proved bound, and is subject to the same assumption-auditing as the $\alpha = 2.1$ estimate below. For $s = 10, \alpha = 3$: $\lambda^{10} = e^{10 \ln 0.0416} = e^{-31.79} = 1.46 \times 10^{-14}$; $\varepsilon = 1.46 \times 10^{-14} \times 10^{-3} = 1.46 \times 10^{-17}$. The per-braid phase drift is bounded by $|\Delta \varphi | \leq 2\pi \varepsilon$; over $N_{\mathrm{pairs}} = 10^{6}$ braids, $|\Delta \varphi | \leq 2\pi \times 1.46 \times 10^{-17} \times 10^{6} = 9.2 \times 10^{-11}$ rad — negligible under the stated assumption. Applying the same factor-of-3 uncertainty treatment as for the $\alpha = 2.1$ projection, the cumulative drift lies in [$3.1, 27.6] \times 10^{-11}$ rad. The Ising projective class is expected to be stable in this regime, contingent on the assumed scaling $\varepsilon$ ~ $\lambda ^{s} \cdot s^{-\alpha }$.

Gap. At $\mu = 0, \alpha = 3$: $\Delta _{\mathrm{gap}} = 2\Delta _{0}\zeta (3) = 2 \times 1.2020569 \Delta _{0} = 2.4041 \Delta _{0} \gt 0$, confirming an open gap along the deformation path from $\alpha = \infty$ to $\alpha = 3$ at these parameters.

Dangerous regime (projection, clearly labeled). The above bound degrades as $\alpha \to 2^{+}$ and $\mu \to \mu _{c} = 2t_{0}\zeta (\alpha )$, where $\xi$ diverges. Projection with stated assumptions: at $\alpha = 2.1, \mu /\mu _{c} = 0.9$, take $\xi \approx 10$ sites (from $\xi$ ~ $\xi _{0}/(1 - \mu /\mu _{c})$ with $\xi _{0}$ ~ 1) and $\varepsilon$ ~ $e^{-s/\xi } s^{-\alpha } = e^{-1}$ × 10^{−2.1} = $0.368 \times 0.00794 = 2.92 \times 10^{-3}$. Per-braid drift ≈ $2\pi \times 2.92 \times 10^{-3} \approx 1.8 \times 10^{-2}$ rad; over $10^{3}$ braids, cumulative drift ~ 18 rad, destroying gate fidelity. Assumptions: (a) exponential localization with the stated $\xi$; (b) the Berry-curvature correction is first order in $\varepsilon$; (c) no gap closing. Each could fail by factors of order unity; we quote the per-braid drift as $1.8 \times 10^{-2}$ rad with a factor-of-3 uncertainty band, [$0.6, 5.4] \times 10^{-2}$ rad.

#4.5 Cohomological conclusion

In the short-range limit the cocycle representative takes values generated by $e^{i\pi /8}$; the Ising statistics corresponds to the nontrivial element of $H^{2}(B_{N}, U(1)) \cong \mathbb{Z}_{2} (N \geq 5)$. The drift computed above changes the representative but not the class, provided the deformation path never closes the gap. Note that the phase $e^{i\pi /8}$ itself is a representative-level quantity: the cohomology class distinguishes only the parity of the cocycle, i.e., Ising versus trivial statistics. Since $\alpha \gt 2$ keeps the long-range couplings perturbatively irrelevant and the gap open ($\Delta _{\mathrm{gap}} = 2.4041 \Delta _{0}$ at $\alpha = 3, \mu = 0$), the class is invariant — the microscopic counterpart of the GT-rigidity observed algebraically in [6].

#5. Results

1. Braid unitary: $U_{ij} = \exp((\pi /4)\gamma _{i}\gamma _{j})$ exchanges $\gamma _{i} \leftrightarrow \mp \gamma _{j}$; verified by explicit adjoint-action computation.
2. Projectivity: $U^{2} = i\gamma _{i}\gamma _{j}$ (fermion parity); $U^{4}$ = −𝟙 in the fixed-parity sector; braid relation $\sigma _{1}\sigma _{2}\sigma _{1} = \sigma _{2}\sigma _{1}\sigma _{2}$ holds exactly. Cocycle value $\omega (\sigma ,\sigma ,\sigma ,\sigma ) = -1$.
3. Anchor phase: $\varphi _{\mathrm{Ising}} = \pi /8 = 0.3927$ rad per exchange, from the Ising R-matrix values $e^{-i\pi /8}, e^{3i\pi /8}$.
4. Representation dimension: $2^{n-1}$ on 2n Majoranas at fixed parity.
5. Localization at $\alpha = 3, \mu /t_{0} = 0.1$: $\lambda = 0.0416, \xi = 0.3145$ sites.
6. Drift projection at $\alpha = 3, s = 10$ (assumed scaling $\varepsilon$ ~ $\lambda ^{s} \cdot s^{-\alpha }$, not derived): $\varepsilon = 1.46 \times 10^{-17}$ per braid; cumulative over $10^{6}$ braids ≤ $9.2 \times 10^{-11}$ rad, with factor-of-3 band [$3.1, 27.6] \times 10^{-11}$ rad. The nontrivial $\mathbb{Z}_{2}$ projective class (Ising) is expected to be stable under these assumptions.
7. Gap at $\mu = 0, \alpha = 3$: $\Delta _{\mathrm{gap}} = 2.4041 \Delta _{0}$.
8. Projection ($\alpha = 2.1, \mu /\mu _{c} = 0.9$): per-braid drift ≈ $1.8 \times 10^{-2}$ rad, band [$0.6, 5.4] \times 10^{-2}$ rad; ~18 rad over $10^{3}$ braids — loss of topological protection near the marginal regime.

#6. Discussion

Limitations. First, the drift bound in the localized regime relies on the self-consistency equation for $\lambda$, which is exact only for the pure power-law chain at small $\mu$; realistic devices have chemical-potential disorder that can resonantly enhance zero-mode overlaps by orders of magnitude, invalidating the $10^{-17}$ estimate. Second, the cohomological invariance argument assumes the deformation path never closes the bulk gap; long-range chains are known to exhibit re-entrant gap closings at intermediate $\alpha$, and a single gap closing would allow the class to jump between the two elements of $\mathbb{Z}_{2}$. Third, the analytic skeleton presented here should be checked by exact diagonalization (the model is quadratic, so exact) for finite chains, $2n \leq 10$, before the marginal regime $\alpha$ ∈ (1, 2] is trusted. Fourth, the drift bound may miss nonperturbative accumulation from the nonlocal Majorana–Majorana coupling $H_{\mathrm{eff}}$ ~ $i\delta \gamma _{1}\gamma _{2}$ with $\delta \propto s^{-\alpha }$, which is only algebraically small just above $\alpha = 2$.

Failure modes and falsifiability. The central claim — class stability for $\alpha \gt 2$ — would be falsified by: (i) an exact-diagonalization computation of the exchange unitary for a finite long-range chain showing a phase outside {$e^{-i\pi /8}, e^{3i\pi /8}$} up to smooth drift, at parameters where the gap is provably open; (ii) a demonstration that the long-range chain realizes a different central extension, e.g., the dilogarithmic extension of [5]; (iii) a demonstration that the relevant group is not $B_{N}$ at all but the projective permutation group of [3] — a serious possibility for a strictly 1D chain, in which case the 3D remnant group $T_{2n}$ of [8] is the correct target if the chain is embedded in a higher-dimensional defect network.

Non-uniqueness warnings. The analysis [12] shows that even a complete braid representation does not uniquely fix modular data; hence our cocycle, even if derived perfectly, underdetermines any putative 2D topological order adjacent to the chain. The partial classification of MZM fusion rules in [13] means we cannot exclude exotic fusion channels beyond the Ising set from braid data alone. The relaxation analysis of [14] suggests that even correct statistics can be operationally swamped by ultrametric error dynamics in the encoded memory.

Open questions. (1) Does the long-range chain in the marginal regime $1 \lt \alpha \leq 2$ realize a continuous family of cocycle representatives, and can the class jump only via gap closings? (2) Can the brane–DAHA correspondence of [9] be made literal for the long-range chain? (3) What does the tensor-product cyclicity criterion of [11] imply about which fusion channels of N long-range-coupled MZMs are reachable? (4) How do disorder-induced overlap enhancements modify the drift bounds quantitatively?

#7. Conclusion

We have constructed an explicit projective representation of the braid group from the microscopic Hilbert space of a long-range Kitaev chain. The projectivity originates entirely in the Clifford algebra of Majorana zero modes — the fermion parity anomaly $U^{4}$ = −𝟙 — and defines the nontrivial element of $H^{2}(B_{N}, U(1)) \cong \mathbb{Z}_{2} (N \geq 5)$, i.e., the Ising projective class. Long-range hopping with exponent $\alpha \gt 2$ leaves this class invariant, with projected numerically negligible drift of the cocycle representative (≤ $9.2 \times 10^{-11}$ rad over $10^{6}$ braids at $\alpha = 3$, under the assumed scaling $\varepsilon$ ~ $\lambda ^{s} \cdot s^{-\alpha }$), while the marginal regime $\alpha \to 2^{+}$ near the phase boundary destroys protection. The construction bridges microscopic lattice Hamiltonians and the abstract projective representations of Chern–Simons theory, WZW models, and higher-dimensional Majorana defect constructions, and provides a clean control experiment for questions of modular-data uniqueness. Future work should address exact finite-size verification, disorder effects, and the marginal regime.

#References

[1] Canonical Chern-Simons Theory and the Braid Group on a Riemann Surface. arXiv:hep-th/9306050v1. https://arxiv.org/abs/hep-th/9306050v1 [2] Canonical Chern-Simons Theory and the Braid Group on a Riemann Surface. arXiv:hep-th/9301036v1. https://arxiv.org/abs/hep-th/9301036v1 [3] Nonabelian braid statistics versus projective permutation statistics. arXiv:hep-th/0201240v2. https://arxiv.org/abs/hep-th/0201240v2 [4] Symplectic Approach of Wess-Zumino-Witten Model and Gauge Field Theories. arXiv:dg-ga/9504001v1. https://arxiv.org/abs/dg-ga/9504001v1 [5] The dilogarithmic central extension of the Ptolemy-Thompson group via the Kashaev quantization. arXiv:1211.4300v4. https://arxiv.org/abs/1211.4300v4 [6] Caracteres de rigidite du groupe de Grothendieck-Teichmuller. arXiv:math/0502117v1. https://arxiv.org/abs/math/0502117v1 [7] Quantum geometry of moduli spaces of local systems and representation theory. arXiv:1904.10491v4. https://arxiv.org/abs/1904.10491v4 [8] Projective Ribbon Permutation Statistics: a Remnant of non-Abelian Braiding in Higher Dimensions. arXiv:1005.0583v4. https://arxiv.org/abs/1005.0583v4 [9] Branes and Representations of DAHA $C^\vee C_1$: affine braid group action on category. arXiv:2412.19647v3. https://arxiv.org/abs/2412.19647v3 [10] Tevatron-for-LHC Report of the QCD Working Group. arXiv:hep-ph/0610012v1. https://arxiv.org/abs/hep-ph/0610012v1 [11] Braid Group Actions and Tensor Products. arXiv:math/0106241v1. https://arxiv.org/abs/math/0106241v1 [12] https://papers.qnfo.org/papers/braid-group-representations-and-modular-data-non-uniqueness-finite-images-and-th/. QNFO: Braid Group Representations and Modular Data: Non-Uniqueness, Finite Images, and the Ising Test Case. [13] https://papers.qnfo.org/papers/braid-group-representations-modular-data-and-the-classification-of-majorana-zero/. QNFO: Braid Group Representations, Modular Data, and the Classification of Majorana Zero Mode Fusion Rules in 2D Topological Superconductors. [14] https://papers.qnfo.org/papers/ultrametric-relaxation-dynamics-in-topological-quantum-memory/. QNFO: Ultrametric Relaxation Dynamics in Topological Quantum Memory.

#Appendix A. Divergence report

D1. Value and provenance of the exchange phase (C2).

• Draft A: $\varphi = (\pi /2)(t_{2}/t_{1}) = 0.15\pi \approx 0.471$ rad, obtained by postulating a linear interpolation $\varphi (r) = \varphi _{0}r$ with $\varphi _{0} = \pi /2$ and $r = t_{2}/t_{1} = 0.3$ ($t_{1}$ = $1, t_{2} = 0.3$ attributed to the parameter regime of [8]). A itself flags this as a postulate, not a derived result.
• Draft B: anchor phase $\varphi _{\mathrm{Ising}} = \pi /8 = 0.3927$ rad, from the standard Ising R-matrix values $e^{-i\pi /8}, e^{3i\pi /8}$ (consistent with [12]).
• Draft C: generator convention $U = \exp((\pi /4)\gamma \gamma )$, giving $U^{2}$ = parity and $U^{4}$ = −𝟙; the per-channel exchange phases are the Ising $e^{\pm i\pi /8}$ values.
• Underlying disagreement: A assumes the projective phase is a continuous function of the microscopic hopping ratio (linear interpolation convention); B and C assume the standard Ising/Clifford convention in which the phase is topologically quantized and cannot vary continuously without a phase transition. These conventions are incompatible: a continuous deformation of the phase at fixed class is impossible for the nontrivial isolated class in $H^{2}(B_{N}, U(1)) \cong \mathbb{Z}_{2}$ unless the representative (not the class) drifts.
• Resolution adopted: the main text uses the B/C convention (Ising anchor $\pi /8$, generator $\exp((\pi /4)\gamma \gamma )$), which is derived from the Clifford algebra rather than postulated. A's linear-interpolation proposal is not adopted; it is reported here as an unverified conjecture that would, if true, imply either a class change (requiring a gap closing) or a representative drift far larger than the bounds computed in Section 4.4.

D2. Effect of long-range hopping on the representation (C3).

• Draft A: long-range hopping "deforms the usual Ising anyon representation into a continuous family of projective representations" parameterized by $t_{2}/t_{1}$.
• Drafts B, C: the algebraic structure of the braid representation is independent of $\alpha$; long-range hopping controls only localization, gap, and the drift of the cocycle representative; the class is stable for $\alpha \gt 2$ with open gap.
• Underlying disagreement: A treats the phase as a microscopic, continuously tunable quantity; B/C treat it as topologically protected. B/C's position is supported by the cohomological argument (class in $\mathbb{Z}_{2}$, invariant under gap-preserving deformations) and by explicit drift bounds.
• Resolution adopted: main text adopts the B/C position (class stability); A's continuous-family claim is documented here and listed as falsifiable (an exact-diagonalization demonstration of a continuously varying phase at open gap would overturn it).

D3. Magnitude of the short-range phase ($\pi /2$ vs $\pi /8$ vs $\pi /4$). A uses $\varphi _{0} = \pi /2$; B uses $\pi /8$; C uses $\pi /4$ in the generator exponent. These are partially convention-dependent: $\pi /4$ is the generator angle $(U = \exp((\pi /4)\gamma \gamma )), \pi /2$ is the phase of $U^{2}$ relative to identity in a particular channel convention, and $\pi /8$ is the Ising topological spin per exchange relative to a bosonic baseline. The main text states all three relations explicitly (Section 4.2) so the conventions are unambiguous; no physical contradiction remains once the conventions are fixed.

#Appendix B. Claim attribution