---
title: Higher-Form Kramers–Wannier Duality
url: https://www.emergentmind.com/topics/higher-form-kramers-wannier-duality
type: topic
---

# Higher-Form Kramers–Wannier Duality

Higher-form Kramers–Wannier duality is the extension of the ordinary Kramers–Wannier paradigm from a \(0\)-form symmetry in \(1+1\) dimensions to settings in which gauging produces dual higher-form symmetries or, at self-dual points, non-invertible topological defects. In recent lattice constructions, gauging a \(\mathbb Z_2^B\) spin-flip symmetry or a \(\mathbb Z_2^F\) fermion parity produces \(\mathbb Z_2\) gauge theories with dual \((n-1)\)-form symmetries in \(n\) spatial dimensions [2510.20893]. In continuum and lattice gauge theories, gauging a self-dual higher-form symmetry on only half of spacetime yields the higher-dimensional analogue of the Kramers–Wannier defect, with non-invertible fusion rules and order/disorder conversion for extended operators [2111.01139]. Bosonization frameworks sharpen the same structure in arbitrary dimensions by relating parity-gauged fermions, flat gauge fields, modified Gauss laws, and projections onto fixed higher-form charge sectors [2508.20167].

## 1. Duality defects beyond \(1+1\) dimensions

The defect-theoretic formulation begins from the observation that the familiar \(1+1\)-dimensional Kramers–Wannier defect line can be generalized to even spacetime dimensions by gauging only half of spacetime. For a \(q\)-form symmetry in \(d\) dimensions, the self-dual case occurs when
\[
q=\frac{d-2}{2},
\]
so that gauging returns a theory of the same type. In \(3+1\) dimensions this gives the central example \(q=1\): a \(\mathbb Z_N^{(1)}\) one-form symmetry and a codimension-one topological defect supported on a \(3\)-manifold \(M\) [2111.01139].

The construction splits spacetime into regions \(L\) and \(R\) separated by an interface \(M\), gauges the higher-form symmetry only in \(R\), and imposes a topological Dirichlet boundary condition
\[
a^{(q+1)}\big|_M=0.
\]
For a discrete \(\mathbb Z_N^{(q)}\) symmetry, the gauging sector is the topological gauge theory
\[
\frac{iN}{2\pi}\, b^{(d-q-2)}\, d a^{(q+1)}.
\]
The full interface theory is therefore
\[
\int_L \mathcal L_{\mathcal T}^{(q)} +\int_R \mathcal L_{\mathcal T}^{(q)}[a^{(q+1)}] +\frac{iN}{2\pi}\int_R b^{(d-q-2)} d a^{(q+1)},
\qquad
a^{(q+1)}|_M=0.
\]

This defect is the direct higher-form analogue of the Kramers–Wannier line. Crossing it takes charged operators to operators attached to higher-dimensional symmetry operators. In \(3+1\) dimensions, a charged line operator crossing the defect becomes a line bounded by a topological surface operator. This is the higher-form version of order/disorder conversion in the Ising model. The defect is non-invertible because fusing it with itself does not return the identity defect; instead, the fusion produces a sum over higher-form symmetry defects [2111.01139].

## 2. Gauging on lattices and the emergence of dual \((n-1)\)-form symmetries

A lattice realization of higher-form Kramers–Wannier structure arises from gauging \(\mathbb Z_2\) symmetries in a way that treats bosonic and fermionic systems in parallel. One construction gauges the \(\mathbb Z_2^B\) spin-flip symmetry of the transverse-field Ising model by inserting \(\mathbb Z_2\) gauge qubits and imposing Gauss law plus flatness. A second construction gauges \(\mathbb Z_2^F\) fermion parity by inserting Majorana fermions and using a fermionic analogue of the disentangling unitary. These procedures produce, respectively, the usual Ising-dual \(\mathbb Z_2\) gauge theories and Majorana-dual \(\mathbb Z_2\) gauge theories [2510.20893].

In two dimensions, the bosonic gauging of the transverse-field Ising model yields the standard \(\mathbb Z_2\) lattice gauge theory with dual \(1\)-form symmetry generated by loop operators
\[
V_B=\prod_{e\subset \Gamma} X_e,
\]
while Wilson loops
\[
W=\prod_{e\subset C} Z_e
\]
are the charged operators. The fermionic version places Majoranas on edges, with local parity
\[
(-1)^{F_e}=i\gamma'_e\gamma_e,
\]
and the dual \(1\)-form symmetry becomes a loop of fermion parity,
\[
V_F=\prod_{e\subset\Gamma} (-1)^{F_e}.
\]
At \(J=0\), the resulting fermionic dual Hamiltonian is described as a Majorana stabilizer code [2510.20893].

The higher-dimensional statement is explicit: gauging a \(\mathbb Z_2^B\) \(0\)-form symmetry yields a \(\mathbb Z_2\) gauge theory with a dual \((n-1)\)-form symmetry, and gauging \(\mathbb Z_2^F\) fermion parity yields a gauge theory whose dual \((n-1)\)-form symmetry is generated by loops of emergent fermions. In this sense, the original \(0\)-form symmetry and the dual \((n-1)\)-form symmetry are exchanged under gauging. The same construction extends to general polyhedral decompositions of space, provided the relevant faces are even-edged in the fermionic case. In three dimensions, the product of the three Gauss-law operators around a cube is the identity, so only two are independent; this enforces that excitations form loops [2510.20893].

A notable feature of this lattice framework is that the bosonic and fermionic gauge theories are unitarily equivalent, but not trivially so. The equivalence is implemented by a linear-depth local unitary circuit and induces a direction-dependent anyonic transmutation connecting bosonic and fermionic toric-code-like theories. In one direction a dressed flux-creation operator maps to a bare \(X\); in the other direction the map introduces a large Wilson-loop attachment. The result is a transmutation rather than a simple anyon permutation [2510.20893].

## 3. Fermionic bosonization, spin structure, and translation-induced duality

A complementary arbitrary-dimensional formulation treats higher-form Kramers–Wannier duality as the bosonized image of a minimal Majorana translation. The construction starts with a complex fermion on each top-dimensional cell and a \(\mathbb Z_2\) gauge spin on each codimension-one cell. Gauging fermion parity imposes a Gauss law such as
\[
G_c \equiv (-1)^{F_c}\prod_{f\subset c} Z_f = 1
\]
in three dimensions, or
\[
G_f \equiv i\gamma_{v_1(f)}\gamma_{v_2(f)}\prod_{e\subset f} Z_e = 1
\]
in two dimensions. A local disentangling unitary
\[
U=\stackrel{\leftarrow}{\prod}_{e}\Big(P_e^+ + P_e^- S_e\Big),
\qquad
P_e^\pm=\frac{1\pm Z_e}{2},
\]
maps the gauge constraint to pure fermion parity, \(U G_c U^{-1}=(-1)^{F_c}\), so that the fermions and gauge spins disentangle [2508.20167].

To recover the original ungauged fermion theory, the gauge field must be flat. In two dimensions this requires
\[
\prod_{e\supset v} X_e=1,
\]
and in three dimensions
\[
\prod_{f\supset e} X_f=1.
\]
After the disentangling unitary, flatness becomes a modified Gauss law in the spin system. The resulting bosonization duality is therefore
\[
\text{ungauged fermions}
\;\longleftrightarrow\;
\text{spin system with a Gauss law}.
\]
The sign structure of that Gauss law depends on a Kasteleyn orientation, equivalently on a discrete spin structure. In two dimensions, inequivalent Kasteleyn orientations are in bijection with \(H^1(M,\mathbb Z_2)\), and the obstruction is the second Stiefel–Whitney class \(w_2\) [2508.20167].

Within this framework, a minimal translation of a Majorana lattice becomes a higher-dimensional Kramers–Wannier duality after bosonization. On a hypercubic lattice in arbitrary dimension, the canonical map is
\[
(-1)^{F_c}\to W_c=\prod_{f\subset c} Z_f,
\qquad
S_{f_x^c}\to X_{f_x^c},
\]
so the translation-induced duality takes the form
\[
W_c \to X_{f_x^c}\to W_{c+a_x}.
\]
This is the direct higher-dimensional analogue of the one-dimensional exchange between a spin operator and a domain-wall operator [2508.20167].

Non-invertibility is built into the lattice operator. On periodic lattices, the bosonized theory carries higher-form symmetry operators on nontrivial cycles, and the Kramers–Wannier operator includes projection onto a fixed higher-form symmetry sector. In two dimensions, for example,
\[
D_{KW}\sim T_x^b \left(\prod_m \frac{U_m^x+1}{2}\right).
\]
Because the projector annihilates the \(-1\) sector, the duality has no inverse on the full Hilbert space. The duality is therefore an isomorphism only after restricting to a sector with fixed higher-form charge [2508.20167].

## 4. Four-dimensional gauge theories and non-invertible fusion

In four-dimensional gauge theory, higher-form Kramers–Wannier duality appears as a non-invertible topological defect tied to self-duality under gauging a discrete one-form symmetry. One route begins with a theory carrying an anomaly-free \(\mathbb Z_N^{(1)}\) one-form symmetry. Gauging that symmetry on half of spacetime yields a \(3\)-dimensional defect \(\mathcal D\), and for a connected orientable \(3\)-manifold \(M\) its fusion rule is
\[
\mathcal D\times \overline{\mathcal D}
=
\frac{1}{N}\sum_{S\in H_2(M;\mathbb Z_N)} \eta(S),
\]
where \(\eta(S)\) is the symmetry surface operator on the \(2\)-cycle \(S\). On \(S^3\),
\[
\langle \mathcal D\rangle_{S^3}=\frac{1}{\sqrt N},
\]
which is a direct indicator of non-invertibility [2111.01139].

The same paper shows that such defects constrain infrared phases. If the theory is invariant under gauging the relevant one-form symmetry, then a trivially gapped infrared phase would have to be a one-form SPT phase compatible with the self-duality. For bosonic \(\mathbb Z_N^{(1)}\) symmetry, the required arithmetic condition includes
\[
4p^2\equiv -1 \pmod N.
\]
As a result, bosonic theories invariant under gauging \(\mathbb Z_N^{(1)}\) can flow to a trivially gapped phase only if every prime factor of \(N\) is \(1 \bmod 4\); for even \(N\), bosonic theories cannot flow to a trivially gapped phase. Fermionic theories obey modified constraints in which \(N\) must be even and every prime factor of \(N/2\) must be \(1 \bmod 4\) [2111.01139].

A related \(3+1\)-dimensional construction starts from a mixed anomaly between a \(\mathbb Z_2^{(0)}\) symmetry and a \(\mathbb Z_2^{(1)}\) one-form symmetry,
\[
\pi \int_{X_5} A^{(1)} \cup \frac{(B^{(2)})}{2}.
\]
After gauging the one-form symmetry, the original codimension-one symmetry wall is no longer gauge-invariant. Gauge invariance is restored by attaching the spin TQFT \(U(1)_2\) on the defect worldvolume, producing a genuine non-invertible defect
\[
(M_3):=D(M_3,b^{(2)})\,{}^{2,1}(M_3,b^{(2)}).
\]
Its fusion is
\[
(M_3)\times (M_3)
=
\frac{1}{|H^0(M_3,\mathbb Z_2)|}
\sum_{\Sigma\in H_2(M_3,\mathbb Z_2)}
(-1)^{Q(\Sigma)}L(\Sigma),
\]
where \(L(\Sigma)=e^{i\pi\oint_\Sigma b^{(2)}}\) generates the one-form symmetry and \(Q(\Sigma)\) is the mod-\(2\) triple intersection number inside \(M_3\) [2111.01141].

This framework gives explicit self-duality defects in \(SO(3)\) Yang–Mills at \(\theta=\pi\), \(\mathcal N=1\) \(SO(3)\) super Yang–Mills, and \(\mathcal N=4\) \(SU(2)\) super Yang–Mills at \(\tau=i\). The self-duality transformation is \(TST\) when the \(\mathbb Z_2^{(0)}\) symmetry is linear and \(TKST\) when it is anti-linear. The characteristic Kramers–Wannier feature remains unchanged: the square of the duality defect is not the identity, but a sum over higher-form symmetry operators [2111.01141].

## 5. Continuum, homological, and infrared reformulations

Higher-form Kramers–Wannier duality also admits a continuum Landau-type formulation in which the fundamental variable is not a point field but a functional field \(\phi[C_p]\) defined on closed \(p\)-surfaces. In this approach, the \(p\)-dimensional Wilson-surface operator of a lattice higher gauge theory is promoted to the basic field, charged under the \(p\)-form global symmetry. The continuum kinetic term is written using an area derivative rather than an ordinary derivative, and the field dimension is
\[
[\phi[C_p]]=\frac{D-2(p+1)}{2},
\qquad
D_c=4(p+1)
\]
for the quartic coupling [2507.06555].

The classical solutions reproduce the expected order/disorder behaviour of higher gauge theory. In the strong-coupling limit, the solution takes the form
\[
f(z)\approx c\,e^{-T_p\,\mathrm{Vol}[M_{p+1}]},
\]
which is the area law and corresponds to the unbroken \(p\)-form symmetry. In the weak-coupling limit, a nonzero expectation value develops,
\[
\langle\phi[C_p]\rangle=\frac{v}{\sqrt2}\neq 0,
\]
the low-energy mode is a phase modulation
\[
\phi[C_p]=\frac{v}{\sqrt2}\exp\!\left(i\int_{C_p}A_p\right),
\]
and the classical behaviour is perimeter-like. Compact \(U(1)\) and finite \(\mathbb Z_N\) theories further admit topological defects that generalize vortices and domain walls [2507.06555].

At the duality level, the lattice higher gauge theory with gauge group \(G\) is mapped to a gauged \((D-p-2)\)-form theory with dual group \(\widehat G\cong G\). Near criticality, this induces an infrared duality between the corresponding Landau theories, exchanging order and disorder operators and relating a \(p\)-form theory to a gauged \((D-p-2)\)-form theory. The order/disorder dictionary is
\[
\phi[C_p]\ \leftrightarrow\ e^{i\int_{C_p}B_p},
\qquad
U[C_{D-p-1}]\ \leftrightarrow\ e^{i\int_{C_{D-p-1}}A_{D-p-1}}.
\]
This is the higher-form counterpart of the standard Kramers–Wannier exchange between local order and disorder variables [2507.06555].

A distinct but complementary topological reformulation uses normal factor graphs to encode chains, cochains, boundary maps, and homology. In that framework, the \(2\)-torus decomposition underlying finite-size Ising duality appears as four homology sectors, while on the \(3\)-torus one has \(\dim H_1=3\), so the duality relation contains \(2^3=8\) topological sectors. The same language explains why nontrivial homology forces sums over topological sectors in dual partition functions and why higher-dimensional dual models need not remain pairwise-interaction models [1607.02361].

## 6. Adjacent generalizations and conceptual delimitations

Several nearby constructions are often grouped with higher-form Kramers–Wannier duality but are technically distinct. In \(2+1\) dimensions with subsystem \(\mathbb Z_2\) symmetry, gauging line-like row and column symmetries produces a subsystem Kramers–Wannier transformation whose duality operator obeys
\[
(\mathcal N^{\rm sub})^\dagger\times\mathcal N^{\rm sub}
=
\mathsf{Grid}_{\{\widehat t^y_i,\widehat t^x_j\}},
\]
a sum over a grid of subsystem symmetry operators. The corresponding defects are mobile in both spatial directions by local unitaries. The same work stresses that subsystem symmetry is not simply higher-form symmetry: the distinction is not only codimension, but also mobility and topologicalness [2304.09886].

A related operator-level program constructs Kramers–Wannier duality in subsystem-symmetric lattice models as a sequential circuit followed by projection onto the symmetric subspace. In two dimensions and higher-dimensional hypercubic models, the duality squares to translation times projection,
\[
(\mathbf D^{(2)})^2\propto \mathbf T_{(1,1)}\mathbf P^{(2)},
\qquad
(\mathbf D^{(d)})^2\propto \mathbf T_{(1,\dots,1)}\mathbf P^{(d)},
\]
and the authors explicitly present this as a non-invertible symmetry operator. The same paper remarks that the philosophy should extend to higher-form symmetries, but its main body is devoted to subsystem symmetry rather than to higher-form gauge symmetry proper [2402.09520].

In \(1+1\) dimensions, gauging finite Abelian modulated symmetries yields further Kramers–Wannier-like dualities in which the dual symmetry generators are reflected versions of the original ones,
\[
U_q^\vee = M U_q M^{-1},
\]
and the non-invertible symmetry operator is a reflection-dressed sequential circuit,
\[
D := M\,\tilde D_{\mathrm{KW}}.
\]
These constructions generalize the Ising story to dipole, quadrupole, and other modulated symmetries, but they are not higher-form constructions in the strict sense [2406.12962].

A final conceptual distinction concerns fermionic higher-dimensional self-dualities obtained from minimal Majorana translations. In the Majorana-dual \(\mathbb Z_2\) gauge theories, the resulting Kramers–Wannier self-dualities are direction-dependent and inherited from the translation and folding structure of the Majorana lattice. They are not simply the same as gauging the \(0\)-form symmetry of the folded Ising chain, nor are they identical to gauging along a space-covering path [2510.20893].

Taken together, these distinctions delimit the higher-form case. The strict higher-form Kramers–Wannier framework concerns dualities generated by gauging higher-form or parity-related symmetries so that extended operators, higher-form charges, and non-invertible defect fusion become the primary structures. Subsystem, modulated, and translation-induced variants are closely related generalizations, but the cited works treat them as structurally different categories of duality.

Source: https://www.emergentmind.com/topics/higher-form-kramers-wannier-duality