Goldstone Counting, Low-Dimensional Obstructions, and Spacetime Exceptions

The slogan “one Goldstone mode per broken generator” is a theorem only for exact continuous internal symmetries in a Lorentz-invariant broken vacuum, under the usual locality, current, and spectral assumptions. Without Lorentz invariance, two broken internal directions can become one canonically paired type-B mode. Broken spacetime generators require a separate test of whether their long-wavelength fluctuations are independent. In a local relativistic theory in $1+1$ dimensions, meanwhile, infrared fluctuations obstruct the ordinary continuously broken phase to which the relativistic counting theorem would apply.

This page separates those three issues: existence of the broken phase, counting of independent internal-symmetry modes, and redundancy among spacetime-symmetry fluctuations. It states the general internal count but leaves developed finite-density dynamics and effective theories to the Many-Body volume.

Required background. Goldstone’s Theorem: Hypotheses and Pole Argument supplies the regulated-charge criterion for breaking and proves the existence of massless spectral support without assuming that the global charge creates a normalizable state.

Helpful background. Lie Groups, Lie Algebras, and Exponential and Adjoint Maps supplies independent generators, Lie brackets, and basis changes. Vacuum Orbits and Unbroken Subgroups identifies the continuous broken directions with the tangent space of $G/H$.

Relativistic internal symmetries

Let an exact ordinary continuous internal global symmetry $G$ be spontaneously broken to a closed subgroup $H$. The number of independent broken directions is

$$n_{\mathrm{br}}=\dim(G/H).$$

This counts tangent directions, not points in the vacuum orbit. Disconnected vacua related by a finite symmetry do not contribute to $n_{\mathrm{br}}$.

Relativistic counting theorem. Suppose:

• the theory is a local, unitary, relativistic QFT in spacetime dimension $d>2$;
• the selected infinite-volume vacuum is invariant under translations and Lorentz transformations;
• $G$ is an exact ordinary zero-form internal global symmetry with local conserved currents;
• the broken charge directions act independently on the vacuum;
• the positive-energy spectral, completeness, and current hypotheses of the Goldstone pole argument hold; and
• $G/H$ has finite dimension.

Then the number of symmetry-required Nambu–Goldstone modes is

$$n_{\mathrm{NG}}=n_{\mathrm{br}}=\dim(G/H),$$

and every one of these modes is type A. This is the precise setting in which the familiar slogan is correct. A full current-algebra proof must show that the matrix of broken variations has rank $n_{\mathrm{br}}$ and that the corresponding massless residues contain that many independent directions; a single diagnostic operator proves existence, not the complete count. The theorem and its relativistic QFT realization are developed in Weinberg 1995, § 19.2, pp. 167–177 and Schwartz 2014, § 28.2, pp. 563–575; the modern type-A/type-B classification and its zero-rank relativistic specialization are summarized in Watanabe 2020, accepted manuscript, § 2.2, pp. 3–4, Open PDF.

Lorentz invariance also explains why the pairing correction introduced below vanishes here. The expectation value of an internal charge density is the time component of a Lorentz vector. A Lorentz-invariant vacuum admits no nonzero preferred vector, so the relevant charge densities vanish and the commutator-density matrix has zero rank.

The theorem counts modes required by broken symmetry. An accidental gapless branch may coexist with them, but it is not part of $n_{\mathrm{NG}}$. Conversely, the equality does not cover a finite-volume symmetric state, a finite-density or nonrelativistic phase, a gauge redundancy, a broken spacetime symmetry, or a sufficiently nonlocal or long-range system.

Charge-density pairing without Lorentz invariance

Consider now exact broken internal symmetries in a translation-invariant thermodynamic-limit phase, without assuming Lorentz invariance. Restrict indices $a,b$ to the $n_{\mathrm{br}}$ broken charges and define

$$\rho_{ab} =-i\lim_{V\to\infty} \frac{\langle\Omega|[Q_a,Q_b]|\Omega\rangle}{V},$$

where $V$ is spatial volume and the phase is selected before the limit is evaluated. The limit must exist, be finite, and have a stable rank on the broken subspace.

For Hermitian charges, $[Q_a,Q_b]$ is anti-Hermitian, so $\rho_{ab}$ is real. Interchanging $a$ and $b$ reverses its sign; hence $\rho$ is antisymmetric and has even rank. Under a nonsingular change of broken-generator basis,

$$\rho\longmapsto M\rho M^{\mathsf T},$$

so individual entries can change sign or normalization while $\operatorname{rank}\rho$ cannot.

Under the general internal-symmetry counting hypotheses, the exact count is

$$\begin{aligned} n_{\mathrm{NG}} &=n_{\mathrm{br}} -\frac12\operatorname{rank}\rho, \\ n_{\mathrm B} &=\frac12\operatorname{rank}\rho, \\ n_{\mathrm A} &=n_{\mathrm{br}} -\operatorname{rank}\rho. \end{aligned}$$

Equivalently,

$$n_{\mathrm A}+2n_{\mathrm B}=n_{\mathrm{br}}.$$

The result, including the hypotheses behind the thermodynamic commutator density, is stated in Watanabe 2020, accepted manuscript, § 2.2, pp. 3–4, Open PDF.

To see the pairing, bring the real antisymmetric matrix to canonical form. Each nonzero sector is a block

$$\begin{pmatrix} 0&\lambda_r\\ -\lambda_r&0 \end{pmatrix}, \qquad \lambda_r\neq0.$$

The two associated broken directions form a conjugate pair and produce one type-B mode rather than two independent modes. Every remaining unpaired broken direction produces a type-A mode. This symplectic pairing—not the power of momentum in the dispersion relation—defines the A/B distinction. In a generic local analytic low-energy theory, type-A modes commonly have linear dispersion and type-B modes commonly have quadratic dispersion, but tuning, extra degeneracy, or altered locality can change those powers.

If the charge algebra is

$$[Q_a,Q_b]=if_{ab}{}^{c}Q_c,$$

then

$$\rho_{ab}=f_{ab}{}^{c}q_c, \qquad q_c=\lim_{V\to\infty} \frac{\langle Q_c\rangle}{V}.$$

Thus a density of an unbroken charge can pair broken charges. For the bounded algebraic example $SU(2)\to U(1)$, take $Q_x$ and $Q_y$ broken and $Q_z$ unbroken. If

$$[Q_x,Q_y]=iQ_z, \qquad \lim_{V\to\infty}\frac{\langle Q_z\rangle}{V}=m\neq0,$$

then the broken $2\times2$ matrix has rank two. The formulas give

$$n_{\mathrm B}=1, \qquad n_{\mathrm A}=0, \qquad n_{\mathrm{NG}}=1.$$

The two broken directions therefore supply one type-B mode. The developed ferromagnetic, superfluid, and finite-density realizations belong to Finite-Density Goldstone Counting.

Spacetime generators require an independence test

There is no universal replacement of the internal rank formula for arbitrary broken spacetime symmetries. A spacetime generator acts on coordinates as well as fields, and its Noether density can be locally dependent on the densities of other broken generators. Counting every broken spacetime generator as a new field can therefore overcount physical fluctuations.

A spinless crystal gives the standard warning. It breaks continuous translations and orbital rotations, but its long-wavelength spectrum needs phonons for the broken translations, not an additional mode for every broken orbital rotation. For spinless fields—or, in relativistic notation, after using a symmetric Belinfante-improved stress tensor—the rotation current can be written in terms of the stress tensor:

$$M^{\mu ij} =x^iT^{\mu j}-x^jT^{\mu i}.$$

An infinitesimal local rotation of the ordered pattern is consequently expressible through a position-dependent translation, and the rotational fluctuation is contained in gradients of the displacement field. More generally, redundancy is established by an actual relation of the form

$$\sum_a c^a(\mathbf x)\, q_a(\mathbf x)|\Omega\rangle=0$$

among the relevant local Noether densities acting on the state. An independently ordered intrinsic-spin or orientational sector must be tested separately. Watanabe explains this local-independence criterion and the crystal example in Watanabe 2020, accepted manuscript, § 3.2, p. 10, Open PDF.

An algebraic relation such as

$$[P_\mu,X_a]\supset i c_{\mu a}{}^bX_b$$

can signal that an inverse-Higgs constraint is possible in an effective description: one Goldstone coordinate may be expressible through derivatives of another. The commutator alone does not prove that a physical mode is absent. One must demonstrate the operator or order-parameter redundancy and check that the proposed constraint is compatible with the low-energy realization. This qualification is discussed in Watanabe 2020, accepted manuscript, § 3.2, p. 10, Open PDF. Scale and conformal breaking are developed in Conformal Field Theory and Bootstrap; systematic spacetime-coset and effective-action construction belongs to Renormalization and Effective Field Theory.

The 1+1-dimensional obstruction

The low-dimensional exception occurs before mode counting. In a local or sufficiently short-range relativistic QFT in $1+1$ dimensions, with the usual positivity, clustering, and infrared assumptions, an ordinary continuous internal type-A order parameter cannot remain nonzero in the infinite-volume vacuum. Because Lorentz invariance would force $\rho=0$, every broken internal direction would have to be type A; the assumed broken phase is therefore obstructed. This scope is summarized in Watanabe 2020, accepted manuscript, § 2.6, p. 8, Open PDF.

The infrared mechanism is visible in the equal-time fluctuation of a type-A angular field. Its long-wavelength contribution behaves as

$$\begin{aligned} &\langle[\pi(x)-\pi(0)]^2\rangle \\ &\quad\sim \int_0^\Lambda \frac{\mathrm dk}{k} \bigl(1-\cos kx\bigr) \\ &\quad\sim \log(\Lambda|x|) \qquad (|x|\to\infty). \end{aligned}$$

The growing fluctuation washes out a fixed continuous phase at long distance. This is why the premise of the relativistic Goldstone theorem fails rather than its conclusion becoming a conventional massless particle.

The statement is deliberately narrow. It does not forbid massless fields, algebraically decaying correlations, Berezinskii–Kosterlitz–Thouless-type or other quasi-long-range sectors, discrete symmetry breaking, or topological gapless sectors without a local order parameter. Type-B systems lie outside the relativistic conclusion and can have different infrared power counting. Nonlocal or sufficiently long-range interactions can also invalidate the locality or dispersion assumptions, but there is no universal decay-exponent criterion within this page’s scope. The type-A obstruction and type-B qualification are summarized in Watanabe 2020, accepted manuscript, § 2.6, p. 8, Open PDF.

Complex U(1) and discrete-subgroup check

Return to the neutral complex scalar in $3+1$ dimensions after the temporary phase-selecting source has been removed following the infinite-volume limit. For exact

$$U(1)\longrightarrow\{1\},$$

there is one broken generator:

$$n_{\mathrm{br}}=1.$$

Every $1\times1$ antisymmetric matrix vanishes, so $\rho=0$ and

$$(n_{\mathrm A},n_{\mathrm B},n_{\mathrm{NG}}) =(1,0,1).$$

The single type-A mode is the angular fluctuation of the scalar. The explicit weak-coupling calculation appears in Schwartz 2014, § 28.2.1, pp. 563–568. Notice that a lone broken $U(1)$ generator cannot form a type-B pair even if Lorentz invariance is later abandoned; its dispersion would still require a separate dynamical analysis.

Now keep a permanent term

$$\begin{aligned} \Delta\mathcal L &=h\phi^N+h^*(\phi^\dagger)^N, \\ N&\in\mathbb Z_{\geq2}, \qquad h\neq0. \end{aligned}$$

The exact subgroup inherited from the original $U(1)$ is then $\mathbb Z_N$; the full theory may retain additional discrete transformations. Assuming the full potential is stable, this finite subgroup may itself be broken and produce disconnected vacua. Regardless, it contributes no continuous tangent direction:

$$n_{\mathrm{br}}=0, \qquad n_{\mathrm{NG}}=0.$$

This is not a failure of Goldstone counting. There is no exact continuous generator to count. If $h$ is small and the stable theory remains continuously connected to the $U(1)$-broken phase, the lifted angular excitation is a pseudo-Goldstone mode; its mass and Ward identity belong to Explicit Breaking and Pseudo-Goldstone Modes.

Choosing the correct count

Use the following sequence before quoting a number.

1. Identify the exact physical global symmetry. Do not count gauge redundancies or explicitly broken generators.
2. Select the thermodynamic-limit phase, determine its unbroken subgroup $H$, and count only the independent continuous tangent directions of $G/H$.
3. If all broken generators are internal, compute $\rho$ on the broken subspace. Zero rank gives one type-A mode per direction; every nonzero $2\times2$ block replaces two directions by one type-B mode.
4. If a spacetime generator is broken, test the local Noether densities and order-parameter variations for independence. Do not apply the internal formula mechanically.
5. Recheck whether the broken phase exists in the stated spatial dimension and whether locality, range, finite-volume limits, or finite density invalidate the theorem being used.
6. Keep symmetry-required Nambu–Goldstone modes distinct from accidental gapless branches.

Common pitfalls

“The number of vacua is the number of Goldstone modes.” Goldstone modes correspond to continuous tangent directions, not to disconnected points. A broken finite symmetry can produce domain walls but no continuous Goldstone coordinate.

“Type A means linear and type B means quadratic.” Those dispersions are common generic outcomes, not the definitions. The invariant distinction is whether broken directions are unpaired or paired by the nonzero-rank blocks of $\rho$.

“An inverse-Higgs commutator automatically removes a field.” The algebra can indicate a possible covariant constraint. Removing a mode requires a demonstrated local redundancy in the physical realization.

“Coleman’s theorem says that no massless excitation exists in 1+1 dimensions.” It obstructs ordinary continuous internal spontaneous breaking under its assumptions. Massless fields and quasi-long-range sectors can still exist without a nonzero local order parameter.

“The finite-volume expectation value can be inserted directly into $\rho$.” A symmetric finite-volume state may erase the order parameter and the relevant densities. Select the phase and take limits in the order established on the finite-volume page.

Check the Goldstone count

These questions are for self-study and are not graded.

1. A phase has $n_{\mathrm{br}}=7$ broken internal directions and $\operatorname{rank}\rho=4$. Find $n_{\mathrm A}$, $n_{\mathrm B}$, and $n_{\mathrm{NG}}$.
2. For $SU(2)\to U(1)$ with broken $Q_x,Q_y$, explain why a nonzero density of $Q_z$ changes the count from two unpaired directions to one paired mode.
3. Why can an exact $\mathbb Z_N$ theory have several vacua but no Goldstone mode?

Check

1. The rank-four matrix has two nonzero canonical blocks. Therefore

   $$\begin{aligned} n_{\mathrm B}&=\frac12(4)=2, \\ n_{\mathrm A}&=7-4=3, \\ n_{\mathrm{NG}}&=3+2=5. \end{aligned}$$

   The weighted identity gives the cross-check $3+2(2)=7$.

2. The algebra $[Q_x,Q_y]=iQ_z$ gives $\rho_{xy}=\langle Q_z\rangle/V$ in the thermodynamic limit. If this density is nonzero, the broken $2\times2$ block has rank two, so $n_{\mathrm B}=1$ and $n_{\mathrm A}=0$. With zero density, the block vanishes and the two directions are unpaired.

3. A finite vacuum orbit has no tangent space and a finite group has no infinitesimal generator. Its disconnected vacua can support defects, but there is no continuous flat direction whose slow variation would be a Goldstone field.

What follows

• Cosets and Nonlinear Realizations constructs fields for independent continuous broken directions.
• Explicit Breaking and Pseudo-Goldstone Modes follows approximate symmetries and lifted angular modes.
• Elitzur’s Theorem and the Gauge-Invariant Higgs Mechanism explains why gauge generators are not physical broken generators.
• Finite-Density Goldstone Counting develops nonrelativistic examples, dispersion checks, and effective dynamics.

References

• Schwartz, Matthew D. Quantum Field Theory and the Standard Model. Cambridge: Cambridge University Press, 2014. DOI.
• Watanabe, Haruki. “Counting Rules of Nambu–Goldstone Modes.” Annual Review of Condensed Matter Physics 11 (2020): 169–187. DOI. Open PDF.
• Weinberg, Steven. The Quantum Theory of Fields, Volume II: Modern Applications. Cambridge: Cambridge University Press, 1995. DOI.