Spontaneous breaking of generalized symmetries
Use charged-loop scaling and the order of limits to distinguish an exact generalized symmetry from its unbroken or broken realization.
On this page
The question for today
What you’ll understand
- Separate explicit endpoint breaking from spontaneous symmetry realization.
- Interpret area and perimeter terms with ultraviolet renormalization stated.
- State dimensional and dynamical assumptions behind generalized Goldstone reasoning.
Knowing the group of a generalized symmetry does not determine how a ground state realizes it. A one-form symmetry can be unbroken in a confining phase or broken in an appropriate deconfined phase. Extended charged operators replace local order parameters, and their size dependence requires ultraviolet subtractions before long-distance behavior is interpreted. Thermodynamic limits and allowed dynamical endpoints matter just as much as the formal line algebra. A finite-volume expectation equal to zero can coexist with symmetry breaking selected in a different limiting procedure.
A charged extended operator needs a limiting prescription
In a symmetric finite-volume state, a charged operator that transforms by a nontrivial phase can have zero expectation by a selection rule. Spontaneous breaking is instead a statement about infinite-volume states and the order in which a symmetry-selecting source is removed. For an extended operator, its support also grows, so the appropriate asymptotic scaling must be specified. A raw small-loop expectation is not an unambiguous order parameter for that limit.
First establish that the symmetry is exact in the theory under discussion. If a line can end on dynamical fundamental matter, its one-form symmetry may already be explicitly broken and its area law may eventually be spoiled by string breaking. Heavy endpoints can make an intermediate-distance regime appear confining without restoring an exact ultraviolet symmetry. A spontaneous-breaking classification only makes sense for the symmetry that actually survives the allowed endpoints and interactions.
Area and perimeter terms have different meanings
For a large rectangular Wilson loop in a conventional confining regime, a positive string tension gives a contribution proportional to enclosed area. A local self-energy associated with the contour contributes a perimeter term. Under simultaneous rescaling of both side lengths, those contributions grow quadratically and linearly respectively. Subtracting or fitting the perimeter contribution is necessary before an area coefficient is interpreted as an infrared string tension.
For a suitable exact electric one-form symmetry, an area law diagnoses the unbroken confining realization. Broken realizations can have perimeter behavior after local subtraction, with additional scale dependence such as Coulomb contributions in a gapless phase. A perimeter coefficient itself is regulator dependent, and a near-perimeter fit over a narrow range is not a proof of a phase. The shape, charge sector, string-breaking scale, and extrapolation to large loops all belong in the claim.
Continuous and discrete breaking need different infrared descriptions
A broken continuous p-form symmetry can yield a massless p-form gauge field as a generalized Goldstone mode under the usual locality, Lorentz, and infrared assumptions. In four spacetime dimensions a one-form Goldstone description is the Maxwell photon with two physical polarizations. Infrared fluctuations can obstruct long-range order in lower dimensions, so the ordinary Goldstone slogan cannot be applied without checking dimensionality and the form of the order diagnostic.
Breaking a finite higher-form symmetry need not produce a gapless Goldstone particle. In familiar gapped deconfined discrete gauge models it instead accompanies topological sectors and extended operator algebra. The precise relation to topological order requires assumptions about locality, geometry, and which symmetries are available; it is not an unrestricted equivalence for every model with a generalized symmetry. Phase diagnostics combine symmetry realization with spectrum and state information, while a topological effective action is an additional low-energy description that must be justified.
Work it through
Separate scaling contributions in a loop model
Use the illustrative ansatz with σ = 0.50 and μ = 0.10 in consistent inverse-length units. For R = 2 and T = 3, compute −ln|⟨W⟩|. Then double both lengths and separate area and perimeter contributions.
Initially the area term is , while the perimeter term is , so the total is 4.
For R = 4 and T = 6, the area term is 12 and the perimeter term is 2, giving total 14.
Area grew by four and perimeter by two. After the stated perimeter subtraction the remaining logarithms are 3 and 12, rather than 4 and 14.
The modeled logarithms are 4 and 14; their separated area pieces are 3 and 12. This arithmetic does not establish that measured data obey the asymptotic ansatz.
The living notebook
Explore the model
Interactive lab · function
Area scaling differs from perimeter scaling
Reset and set BOTH Coefficient a to 0.5 and Position x to 1; predict y = 0.5. Then Reset and set BOTH a to 0.5 and x to 2; predict y = 2. The exact quadratic illustrates a contribution growing with squared size. It contains no Wilson loop, linear perimeter term, thermodynamic limit, or confinement dynamics.
- y at chosen x
- 0.608
- Tangent slope · 2ax
- 1.35
- Signed area · ax³/3
- 0.182
This lab uses only y = ax². Playback traces the curve within the fixed integration interval from 0 to x; it does not estimate numerical error or change the chosen point. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.
Look forA scaling exponent can distinguish contributions only after the physical observable and subtractions are identified.
Pause & reason
Check your understanding
Your turn
Try it yourself
In a regime where the rectangular-loop ansatz applies at large T, a fitted static potential gives V(2) = 1.2 and V(4) = 2.2 in consistent units. Infer σ and the constant 2μ. Explain why an added dynamical endpoint may invalidate extrapolation to arbitrarily large R.
Take a moment to make a prediction or work it out on paper. Open a hint when you need one.
Give me a hint
The large-T ansatz gives a linear potential plus a constant; use the difference first.
Compare with the solution
The slope is . Then 2μ = V(2) − σ(2) = 0.20. If dynamical charges can screen the line, producing endpoints may eventually cost less than stretching the string, so the potential can saturate and the exact one-form symmetry may be absent. The fit describes the tested regime, not automatically the infinite-distance limit.
Keep the idea
Take it with you
- Exact generalized symmetry and its spontaneous realization are separate questions.
- Loop area terms and local perimeter counterterms must be disentangled before phase inference.
- Continuous Goldstone fields, finite-group topological sectors, and infrared dimension bounds require their own assumptions.
Read further
Original sources and deeper explanations to follow at your own pace.
Optional progress, saved in this browser when available. Revisit any lesson at any time.
Lesson not marked complete.