12.18.02 · quantum / gauge-and-symmetry

The Goldstone theorem and effective Goldstone Lagrangians

shipped3 tiersLean: none

Anchor (Master): Weinberg, S., *The Quantum Theory of Fields, Vol. 2: Modern Applications* (Cambridge, 1996), Ch. 19 (spontaneously broken global symmetries; the Goldstone theorem; the structure of the effective Lagrangian; pions as Goldstone bosons; soft-pion theorems and the Adler zero); Coleman, S., *Aspects of Symmetry* (Cambridge, 1985), Ch. 2 (the invariance of the vacuum is the invariance of the world)

Intuition Beginner

Some collective motions cost almost no energy. Imagine a long line of compass needles, each gently coupled to its neighbours so that they all prefer to point the same way. The system has to choose a direction, and once it does, the whole line points, say, east. Now nudge one end so the needles slowly swing from east toward north along the line. If the twist is very gradual, neighbouring needles barely disagree, so the energy cost is tiny. A very long, very gentle twist costs almost nothing. That nearly-free swinging motion is a wave that ripples through the system at vanishingly small energy when its wavelength is long.

This is the heart of the Goldstone idea. A system whose underlying laws do not care which direction is chosen, but which is forced to pick one, always has these cheap "swing the choice around" waves. In the language of particles, a wave that costs no energy at long wavelength is a massless particle: the longer the wavelength, the lower the energy, all the way down to zero. Such a particle is called a Goldstone boson, after Jeffrey Goldstone, who showed in 1961 that it must appear whenever a continuous symmetry is spontaneously broken.

The needles only had one way to swing, so they give one Goldstone mode. A system with more independent directions to swing the choice around has one Goldstone boson for each one. Counting the broken directions counts the massless particles. This counting rule is exact, and it is the same whether we are talking about magnets, crystals, superfluids, or the pions of nuclear physics.

Visual Beginner

Picture the same wine-bottle-punt energy surface that describes a chosen vacuum: a central bump and a circular valley of lowest energy. The system rests at one point of the ring. Two motions live at that point. Climbing the inner wall, straight up out of the valley, costs real energy: that stiff radial motion is a massive particle. Sliding around the flat valley floor costs nothing, because every point of the ring has the same height: that flat motion is the Goldstone mode.

The flatness of the valley floor is the whole story. The symmetry of the underlying law is what makes the floor perfectly level: rotating the chosen point to any other point of the ring gives a state of exactly the same energy. A perfectly level direction means a motion with no restoring force, and no restoring force means no mass. If the floor had even a slight tilt, the would-be Goldstone particle would pick up a small mass; that gentler case describes the pions, whose tiny masses come from a small tilt added by the quark masses.

Worked example Beginner

Count the Goldstone bosons for a familiar example: a magnet that picks a direction in ordinary three-dimensional space. The underlying rule is symmetric under all rotations, and rotations in three dimensions have three independent axes. Once the magnet points along a chosen axis, how many of those rotations actually change the state?

Step 1. List the symmetry directions. Rotations in three-dimensional space come in three independent kinds: spin about the $x$ axis, about the $y$ axis, and about the $z$ axis. That is three symmetry directions to start with.

Step 2. Find the ones that leave the chosen state alone. Suppose the magnet points along the $z$ axis. Spinning the whole magnet about the $z$ axis does nothing visible: the arrow still points along $z$ . So one of the three rotations leaves the state unchanged.

Step 3. Subtract. The directions that leave the state unchanged are the unbroken ones; the rest are broken. Three total minus one unbroken leaves two broken directions.

Step 4. Read off the count. The Goldstone rule gives one massless mode per broken direction, so there are two Goldstone modes. For a magnet these are the two independent long-wavelength spin waves, the gentle twists of the arrow away from $z$ toward $x$ or toward $y$ .

What this tells us: the answer came purely from counting, not from solving any equation. Three symmetry directions, one of them unbroken, two broken, two massless spin waves. The same subtraction, "total symmetry directions minus unbroken directions," gives the Goldstone count in every example.

Check your understanding Beginner

Formal definition Intermediate+

Throughout, the metric is mostly-minus, $η_{μν} = diag (+, -, -, -)$ , matching the convention of the sibling unit 12.18.01. Let a relativistic quantum field theory have a continuous global symmetry group $G$ (a compact Lie group), realised by a set of conserved Noether currents $j_{a}^{μ}$ , $a = 1, \dots, dim G$ , satisfying $\partial_{μ} j_{a}^{μ} = 0$ with associated charges $Q_{a} = \int d^{3} x j_{a}^{0} (x)$ . These currents are the global counterpart of the Lie-algebra-valued currents of 03.07.05.

The symmetry is spontaneously broken when the vacuum $∣0 ⟩$ is not invariant under all of $G$ , in the sense of 08.02.02. Concretely, there is a local field $ϕ$ (or composite operator) whose vacuum expectation value transforms non-invariantly: for some generators,

⟨ 0∣ [Q_{a}, ϕ (0)] ∣0 ⟩ \neq = 0.

The generators $T_{a}$ for which this order-parameter condition holds are the broken generators; they span a subspace $m \subset g$ . The generators that annihilate the order parameter, $⟨ 0∣ [Q_{a}, ϕ (0)] ∣0 ⟩ = 0$ , span the Lie algebra $h$ of the unbroken subgroup $H = {U \in G : U fixes the vacuum}$ . The number of broken generators is $dim G - dim H = dim (G / H)$ , the dimension of the coset space $G / H$ .

Goldstone's theorem (relativistic form) states: for each broken generator there is a massless, spin-zero, single-particle state in the spectrum — a Goldstone boson. The number of Goldstone bosons equals $dim (G / H)$ .

The effective Goldstone Lagrangian organises the low-energy dynamics of these massless modes. Parametrise the Goldstone fields by a map $π (x) = π^{a} (x) X_{a}$ into the broken subspace $m$ , packaged as a coset representative $U (x) = exp (i π^{a} (x) X_{a} / f)$ taking values in $G / H$ , where $f$ is a constant with mass dimension one (the decay constant) and $X_{a}$ are the broken generators. The leading low-energy Lagrangian invariant under $G$ is

L_{eff} = \frac{f ^{2}}{4} ⟨ \partial_{μ} U^{†}, \partial^{μ} U ⟩ + \dots,

with the dots denoting terms of higher order in derivatives. Every term carries at least one derivative on each Goldstone field, the algebraic fingerprint of masslessness: a constant Goldstone field is a symmetry transformation of the vacuum and must drop out.

Counterexamples to common slips

A discrete broken symmetry produces no Goldstone boson. Goldstone's theorem requires a continuous symmetry: there must be a flat valley-floor direction to move along, and a discrete symmetry (such as a sign flip $ϕ \to - ϕ$ ) has isolated minima with no connecting flat direction. The kink of a broken $Z_{2}$ symmetry is a domain wall, not a massless particle.
Breaking a gauge symmetry does not give physical Goldstone bosons; that is the Higgs case of 12.18.01, where the would-be Goldstone mode becomes the longitudinal polarisation of a massive gauge boson. The theorem as stated here is for global symmetries, where the Goldstone bosons are genuine massless particles.
In $d \leq 2$ spacetime dimensions the theorem fails: infrared fluctuations of the would-be Goldstone field are so violent that no symmetry-breaking vacuum survives at nonzero temperature (Mermin-Wagner-Coleman). The relativistic theorem assumes $d \geq 3$ and a genuine isolated one-particle pole.

Key theorem with proof Intermediate+

Theorem (Goldstone, relativistic form). Let a Lorentz-invariant local field theory have a conserved current $j_{a}^{μ}$ and a local field $ϕ$ with $⟨ 0∣ [Q_{a}, ϕ (0)] ∣0 ⟩ = v_{a} \neq = 0$ for a broken generator $a$ . Then the spectrum contains a massless single-particle state $∣ π_{a} (k)⟩$ with $⟨ 0∣ j_{a}^{μ} (0) ∣ π_{a} (k)⟩ = i f k^{μ}$ for a nonzero constant $f$ , and $⟨ 0∣ ϕ (0) ∣ π_{a} (k)⟩ \neq = 0$ . There is one such massless scalar per broken generator.

Proof. Consider the Fourier transform of the current-field commutator vacuum expectation value,

F^{μ} (k) = \int d^{4} x e^{ik \cdot x} ⟨ 0∣ [j_{a}^{μ} (x), ϕ (0)] ∣0 ⟩ .

Insert a complete set of momentum eigenstates and use Lorentz invariance and translation covariance $j_{a}^{μ} (x) = e^{i P \cdot x} j_{a}^{μ} (0) e^{- i P \cdot x}$ . The spectral representation reads

F^{μ} (k) = (2 π) k^{μ} [ρ (k^{2}) θ (k^{0}) - \tilde{ρ} (k^{2}) θ (- k^{0})],

where the structure $k^{μ} ρ (k^{2})$ is forced because the only available Lorentz vector built from the intermediate-state momentum is $k^{μ}$ and the spectral density $ρ$ can depend only on the invariant $k^{2}$ . Current conservation $\partial_{μ} j_{a}^{μ} = 0$ gives $k_{μ} F^{μ} (k) = 0$ , hence $k^{2} ρ (k^{2}) = 0$ , so $ρ (k^{2}) = c δ (k^{2})$ for some constant $c$ : the spectral density is supported only at $k^{2} = 0$ .

Now evaluate the equal-time charge-field commutator. Integrating $F^{0}$ over $k$ at fixed setup recovers

⟨ 0∣ [Q_{a}, ϕ (0)] ∣0 ⟩ = v_{a} = \int \frac{d ^{3} k}{( 2 π ) ^{3}} (contribution of ρ) \propto c .

Because $v_{a} \neq = 0$ by the breaking assumption, the constant $c \neq = 0$ , so the delta function $δ (k^{2})$ at $k^{2} = 0$ carries nonzero weight. A nonzero spectral weight concentrated at $k^{2} = 0$ is precisely a massless single-particle pole: there exists a state $∣ π_{a} (k)⟩$ with $k^{2} = 0$ and $⟨ 0∣ j_{a}^{μ} (0) ∣ π_{a} (k)⟩ = i f k^{μ}$ with $f \propto c \neq = 0$ . Lorentz invariance of the matrix element $⟨ 0∣ j_{a}^{μ} ∣ π_{a} ⟩ \propto k^{μ}$ identifies the state as spin zero. Repeating the argument for each broken generator (each with its own $v_{a} \neq = 0$ ) produces one massless scalar per broken generator. $□$

Bridge. This builds toward the effective Lagrangian of the Master tier and appears again in 12.18.01, where the very same massless pole, once the symmetry is gauged, is absorbed into a massive vector. The foundational reason the Goldstone boson is massless is that the broken charge moves the vacuum along a degenerate valley floor at no energy cost, and this is exactly the flatness that forces $ρ (k^{2}) \propto δ (k^{2})$ . The central insight is that the current matrix element $⟨ 0∣ j_{a}^{μ} ∣ π_{a} ⟩ = i f k^{μ}$ identifies the broken-symmetry current with the creation of a single massless quantum: putting these together, the decay constant $f$ is at once the coefficient of the massless pole, the normalisation of the Goldstone kinetic term, and the strength with which the Goldstone couples to the broken current. The bridge is the recognition that the geometry of degenerate vacua and the analyticity of the current two-point function are two faces of one fact, and this pattern recurs whenever a continuous symmetry is spontaneously broken, generalising from the magnet's spin waves to the pions of the strong interaction.

Exercises Intermediate+

Exercise 4 (medium, short-answer).

Explain why the Goldstone theorem requires a continuous symmetry and gives nothing for a discrete one.

Hint

The massless pole came from a current. Does a discrete symmetry have a conserved Noether current generating a flat valley direction?

Answer

The proof relies on a conserved Noether current $j_{a}^{μ}$ , which exists only for a continuous symmetry (Noether's theorem applies to one-parameter families of transformations). The current Ward identity is what forces the massless pole. A discrete symmetry, such as $ϕ \to - ϕ$ , has no associated conserved current and no flat valley-floor direction connecting the degenerate minima; the minima are isolated points. Moving between them is a finite jump (a domain wall / kink), not a long-wavelength massless wave. Hence no Goldstone boson. Rubric: full credit for "no conserved current / no flat direction for a discrete symmetry."

Exercise 5 (medium, short-answer).

Goldstone bosons couple derivatively: every interaction vertex carries a factor of the Goldstone momentum. Argue from the symmetry why a constant shift $π^{a} \to π^{a} + c^{a}$ must be a symmetry of the effective Lagrangian, and why this forces derivative coupling.

Hint

A constant Goldstone field is the same as acting on the vacuum with a broken symmetry transformation. What does a symmetry do to the action?

Answer

A spatially constant Goldstone configuration $π^{a} (x) = c^{a}$ is exactly a global broken-symmetry transformation applied to the vacuum, which produces another degenerate vacuum of identical energy. Because all these vacua have the same energy, the effective Lagrangian must be invariant under the constant shift $π^{a} \to π^{a} + c^{a}$ (the shift symmetry). A Lagrangian invariant under arbitrary constant shifts of a field can depend on that field only through its derivatives $\partial_{μ} π^{a}$ , since the field itself can be shifted away pointwise but its gradient cannot. Hence every term carries at least one derivative per Goldstone field, and every vertex carries a momentum factor. Rubric: full credit for the shift symmetry plus the derivative-only consequence.

Exercise 7 (hard, short-answer).

For $G$ broken to $H$ by an order parameter, prove that the number of Goldstone bosons equals $dim G - dim H$ by characterising the broken generators.

Hint

Consider the linear map $T_{a} \mapsto ⟨ 0∣ [Q_{a}, ϕ] ∣0 ⟩$ . Its kernel is $h$ ; its image counts the broken directions.

Answer

Define the linear map $Φ : g \to V$ , $T_{a} \mapsto δ_{a} := ⟨ 0∣ [Q_{a}, ϕ (0)] ∣0 ⟩$ , where $V$ is the (real) vector space in which the order parameter lives. The kernel of $Φ$ consists of generators with $δ_{a} = 0$ , i.e. those leaving the vacuum invariant — this is precisely the Lie algebra $h$ of the stabiliser $H$ , of dimension $dim H$ . By rank-nullity, the image has dimension $dim g - dim h = dim G - dim H$ . Each independent broken direction (each basis vector of a complement $m$ to $h$ ) yields, by the Goldstone theorem, one massless scalar. Hence the number of Goldstone bosons is $dim G - dim H = dim (G / H)$ . Rubric: full credit for the kernel-equals- $h$ argument and the rank-nullity count.

Exercise 8 (hard, short-answer).

The lowest-order chiral Lagrangian is $L = \frac{f _{π}^{2}}{4} tr (\partial_{μ} U^{†} \partial^{μ} U)$ with $U = exp (2 i π^{a} T^{a} / f_{π})$ . Expand to quadratic order and verify that the $π^{a}$ are canonically normalised massless fields.

Hint

Expand $U \approx 1 + 2 i π^{a} T^{a} / f_{π} + \dots$ , compute $\partial_{μ} U$ , and use $tr (T^{a} T^{b}) = \frac{1}{2} δ^{ab}$ .

Answer

With $U = 1 + \frac{2 i}{f _{π}} π^{a} T^{a} + O (π^{2})$ , one has $\partial_{μ} U = \frac{2 i}{f _{π}} (\partial_{μ} π^{a}) T^{a} + \dots$ and $\partial_{μ} U^{†} = - \frac{2 i}{f _{π}} (\partial_{μ} π^{a}) T^{a} + \dots$ . Then $tr (\partial_{μ} U^{†} \partial^{μ} U) = \frac{4}{f _{π}^{2}} (\partial_{μ} π^{a}) (\partial^{μ} π^{b}) tr (T^{a} T^{b}) = \frac{4}{f _{π}^{2}} (\partial_{μ} π^{a}) (\partial^{μ} π^{a}) \cdot \frac{1}{2} = \frac{2}{f _{π}^{2}} (\partial π^{a})^{2}$ . Multiplying by $f_{π}^{2} /4$ gives $\frac{1}{2} (\partial_{μ} π^{a}) (\partial^{μ} π^{a})$ , the canonical kinetic term, with no mass term: the pions are massless and canonically normalised at this order. A pion mass enters only when an explicit symmetry-breaking quark-mass term tilts the potential. Rubric: full credit for the canonical $\frac{1}{2} (\partial π)^{2}$ and the absence of a mass term.

Advanced results Master

The single-pole statement of the theorem refines into a complete low-energy framework: the Goldstone fields are coordinates on the coset $G / H$ , their dynamics is fixed order-by-order in derivatives by the unbroken symmetry, and the leading interactions are universal. The following results assemble the structure and its physical incarnation in the strong interactions.

Proposition (coset structure of the Goldstone manifold). Let $G$ break to $H$ . The Goldstone bosons parametrise the homogeneous space $G / H$ : a coset representative $U (x) \in G / H$ is acted on by $g \in G$ via $U \mapsto g U h^{- 1} (g, U)$ , where $h (g, U) \in H$ is the compensating transformation, and the unbroken $H$ acts linearly while the broken generators act by nonlinear shifts. The lowest-derivative $G$ -invariant built from $U$ is the Maurer-Cartan term $\frac{f ^{2}}{4} ⟨ \partial_{μ} U^{†}, \partial^{μ} U ⟩$ .

This is the Callan-Coleman-Wess-Zumino (CCWZ) construction. The geometry of $G / H$ determines the entire low-energy Lagrangian up to a finite number of constants at each derivative order. The action of $G$ is realised nonlinearly: the broken generators shift the Goldstone fields, so $G$ is not represented by ordinary matrices on the $π^{a}$ but by field-dependent transformations. The single number $f$ at leading order, the decay constant, fixes both the normalisation of the kinetic term and the leading Goldstone scattering amplitudes; higher-derivative terms (the Gasser-Leutwyler constants $L_{i}$ at order $p^{4}$ for chiral $S U (3)$ ) are independent low-energy constants.

Proposition (Adler zero / soft-Goldstone theorem). Any scattering amplitude with an external Goldstone boson of four-momentum $k$ vanishes linearly as $k \to 0$ , provided no other particle becomes on-shell in the limit: $M \to 0$ as $O (k)$ .

The Adler zero is the amplitude-level shadow of derivative coupling: each Goldstone emission vertex carries a factor of the Goldstone momentum, so the amplitude is analytic in $k$ and vanishes at $k = 0$ . This is the soft-pion theorem, the experimental signature distinguishing genuine Goldstone bosons from ordinary light scalars. It underlies the low-energy theorems of pion physics, including the Weinberg $π π$ scattering lengths predicted purely from $f_{π}$ and chiral symmetry.

Proposition (pseudo-Goldstone mass from explicit breaking). If $G$ is broken not only spontaneously but also explicitly by a small term $L \supset - ϵ O$ , the would-be Goldstone bosons acquire a small mass-squared $m^{2} \propto ϵ$ , the Gell-Mann-Oakes-Renner relation $m_{π}^{2} f_{π}^{2} = - (m_{u} + m_{d}) ⟨ \overset{q}{ˉ} q ⟩$ in the chiral case.

The pions are not exactly massless because the quark masses $m_{u}, m_{d}$ explicitly break chiral symmetry; the valley floor is slightly tilted. The smallness of $m_{π}$ relative to other hadron masses ( $m_{π} \approx 140$ MeV versus $m_{ρ} \approx 770$ MeV) is the measure of how good the approximate chiral symmetry is. The relation $m_{π}^{2} \propto m_{q}$ , not $m_{π}^{2} \propto m_{q}^{2}$ , is the characteristic signature of a pseudo-Goldstone boson.

Proposition (Mermin-Wagner-Coleman dimensional obstruction). In $d \leq 2$ spacetime dimensions a continuous symmetry cannot be spontaneously broken (at nonzero temperature, or in the relativistic vacuum for $d = 2$ ): the would-be Goldstone field has infrared-divergent fluctuations $⟨ π^{2} ⟩ \sim \int d^{d - 1} k / k^{2}$ that destroy long-range order. The Goldstone mode that the theorem would produce is precisely what forbids the breaking in low dimensions; the massless propagator $1/ k^{2}$ integrated over too few dimensions diverges in the infrared.

Synthesis. The foundational reason a Goldstone boson exists is that a spontaneously broken continuous symmetry leaves a manifold of degenerate vacua — the coset $G / H$ — and the central insight is that the massless modes are exactly the fluctuations that move the system along this flat valley floor. Putting these together, the current-algebra pole $⟨ 0∣ j_{a}^{μ} ∣ π_{a} ⟩ = i f k^{μ}$ , the coset geometry of the CCWZ Lagrangian $\frac{f ^{2}}{4} ⟨ \partial U^{†}, \partial U ⟩$ , the Adler soft zero, and the Gell-Mann-Oakes-Renner pseudo-Goldstone mass are one phenomenon viewed from four angles: spectral, geometric, kinematic, and perturbed. This is exactly the structure that appears again in 12.18.01, where gauging the broken symmetry feeds these same Goldstone modes into the longitudinal polarisations of massive gauge bosons — the Goldstone boson is dual to the longitudinal gauge mode, and the decay constant $f$ becomes the gauge-boson mass scale. The whole framework generalises the magnet's spin waves to the pions of the strong interaction, and it builds toward the effective-field-theory paradigm in which symmetry and its breaking, not a fundamental Lagrangian, dictate the low-energy physics.

Full proof set Master

Proposition (Goldstone modes are derivatively coupled — shift symmetry), proof. Let the broken-symmetry transformation act on the Goldstone fields as $π^{a} \to π^{a} + ϵ c^{a} + O (π)$ , where the inhomogeneous constant shift $c^{a}$ comes from the nonlinear action of the broken generators. A spatially and temporally constant configuration $π^{a} (x) = c^{a}$ is the image of the vacuum under a global broken-symmetry transformation, hence is itself a vacuum of equal energy; the potential $V (π)$ is constant along the coset, $V (π) = V (0)$ , so $V$ contributes no mass term. Invariance of the action under $π^{a} \to π^{a} + c^{a}$ for arbitrary constant $c^{a}$ means $L$ depends on $π^{a}$ only through $\partial_{μ} π^{a}$ : writing $L = L (π, \partial π)$ , the shift invariance $\partial L / \partial π^{a} = 0$ at constant field forces all explicit (non-derivative) $π$ -dependence to cancel. The lowest term consistent with Lorentz invariance and the shift symmetry is therefore $\frac{1}{2} G_{ab} (π) \partial_{μ} π^{a} \partial^{μ} π^{b}$ with $G_{ab}$ the metric on $G / H$ , whose expansion about $π = 0$ begins $\frac{1}{2} δ_{ab} \partial_{μ} π^{a} \partial^{μ} π^{b}$ . No mass term is possible; the Goldstone bosons are massless and derivatively coupled. $□$

Proposition (one massless pole per broken current, spectral form), proof. From the Theorem's spectral representation $F_{a}^{μ} (k) = (2 π) k^{μ} [ρ_{a} (k^{2}) θ (k^{0}) - \tilde{ρ}_{a} (k^{2}) θ (- k^{0})]$ and the conservation constraint $k^{2} ρ_{a} (k^{2}) = 0$ , the density is $ρ_{a} (k^{2}) = c_{a} δ (k^{2})$ . The Källén-Lehmann positivity of the spectral function in the channel created by $j_{a}^{μ}$ guarantees $ρ_{a} \geq 0$ , so $c_{a} \geq 0$ . The breaking condition $v_{a} = ⟨ 0∣ [Q_{a}, ϕ] ∣0 ⟩ \neq = 0$ fixes $c_{a}$ to be strictly positive through the sum rule $v_{a} = c_{a} \cdot ⟨ 0∣ ϕ ∣ π_{a} ⟩$ (up to normalisation), forcing $c_{a} > 0$ for each broken $a$ . A strictly positive spectral weight located at $k^{2} = 0$ is, by the definition of the Källén-Lehmann decomposition, a one-particle massless state in the spectrum: the channel $j_{a}^{μ} ∣0 ⟩$ has a stable massless single-particle contribution $∣ π_{a} ⟩$ . Distinct broken generators give linearly independent order parameters $v_{a}$ (Exercise 7), hence linearly independent massless states, one per broken generator. $□$

Proposition (Adler zero from derivative coupling), proof. Consider an amplitude $M = ⟨ β π_{a} (k) ∣ α ⟩$ for emission of a single Goldstone boson of momentum $k$ . By the LSZ reduction the pion field is replaced by its interpolating current; using the broken-current matrix element $⟨ 0∣ j_{a}^{μ} ∣ π_{a} (k)⟩ = i f k^{μ}$ , the amplitude is proportional to $k_{μ} ⟨ β ∣ j_{a}^{μ} ∣ α ⟩ / f$ . Current conservation $\partial_{μ} j_{a}^{μ} = 0$ implies $k_{μ} ⟨ β ∣ j_{a}^{μ} ∣ α ⟩ = 0$ for the connected piece, save for pole contributions where an external leg goes on-shell as $k \to 0$ . Away from such poles the amplitude is therefore $O (k)$ and vanishes as $k \to 0$ : this is the Adler zero. The exceptions are exactly the pole terms in which the soft Goldstone attaches to an external line, accounted for separately in the soft-pion expansion. $□$

Proposition (pseudo-Goldstone mass, Dashen formula), proof. Add a small explicit breaking $L \to L - H^{'}$ with $H^{'}$ transforming in a known representation of $G$ . The Goldstone potential, formerly flat along $G / H$ , acquires a tilt $V_{eff} (π) = ⟨ H^{'} (π)⟩ = ⟨ Ω∣ H^{'} ∣ Ω_{π} ⟩$ , where $Ω_{π}$ is the vacuum rotated by the Goldstone field. Expanding to quadratic order, $m_{ab}^{2} = \partial^{2} V_{eff} / \partial π^{a} \partial π^{b} ∣_{0} = - ⟨ 0∣ [Q_{a}, [Q_{b}, H^{'}]] ∣0 ⟩ / f^{2}$ , the Dashen formula, manifestly linear in the breaking $H^{'}$ . For chiral $S U (2)$ with $H^{'} = (m_{u} + m_{d}) \overset{q}{ˉ} q /2$ this reduces to the Gell-Mann-Oakes-Renner relation $m_{π}^{2} f_{π}^{2} = - (m_{u} + m_{d}) ⟨ \overset{q}{ˉ} q ⟩$ , so $m_{π}^{2} \propto m_{q}$ . The mass-squared, not the mass, is linear in the symmetry-breaking parameter, the defining signature of a pseudo-Goldstone boson. $□$

Connections Master

Spontaneous symmetry breaking 08.02.02. The Goldstone theorem is the dynamical consequence of the broken-vacuum structure of 08.02.02: a symmetric Lagrangian with a degenerate manifold of vacua. Where 08.02.02 supplies the order parameter and the broken-vacuum picture, this unit extracts the spectral consequence — one massless excitation for each flat valley-floor direction. The coset $G / H$ here is the explicit geometry of the vacuum manifold that 08.02.02 introduces.
The Higgs mechanism 12.18.01. When the spontaneously broken symmetry is gauged, the Goldstone bosons of this unit do not survive as physical massless particles; each is absorbed as the longitudinal polarisation of a massive gauge boson, the content of 12.18.01. The decay constant $f$ that normalises the Goldstone kinetic term here becomes the gauge-boson mass scale $m_{A} = g f$ there. The two units are the global and gauged faces of the same broken-symmetry structure: Goldstone's theorem states the global case, the Higgs mechanism is its gauged evasion.
Yang-Mills action 03.07.05. The conserved currents $j_{a}^{μ}$ whose broken charges generate the Goldstone modes are built from the same Lie-algebra generators $T_{a}$ that structure the Yang-Mills covariant derivative of 03.07.05; the nonlinear realisation of $G$ on the coset $G / H$ uses the group geometry that 03.07.05 develops for the gauge connection. The CCWZ Maurer-Cartan term is the global-symmetry analogue of the Yang-Mills field strength's geometric origin.

Historical & philosophical context Master

Jeffrey Goldstone, drawing an explicit analogy to the theory of superconductivity, showed in 1961 that a relativistic field theory whose Lagrangian possesses a continuous symmetry broken by the vacuum must contain a massless spin-zero boson (Goldstone, Nuovo Cimento 19, 154, 1961) ^{[source pending]}. The analogy was not incidental: Yoichiro Nambu, in the same period, had recognised that the energy gap of a superconductor is the manifestation of a broken symmetry and that the would-be massless mode is the collective phase oscillation. Nambu and Jona-Lasinio (Phys. Rev. 122, 345, 1961) ^{[source pending]} built a dynamical model in which chiral symmetry is broken by a fermion condensate and the pion emerges as the near-massless Goldstone boson, predicting the Goldberger-Treiman relation. The theorem was then made rigorous by Goldstone, Salam, and Weinberg (Phys. Rev. 127, 965, 1962) ^{[source pending]}, who gave three independent proofs, the most general assuming only Lorentz invariance and a conserved current with nonvanishing vacuum-to-one-particle matrix element. Coleman's dictum, "the invariance of the vacuum is the invariance of the world," captures the philosophical core: it is the symmetry of the state, not of the Lagrangian, that governs the observed spectrum.

The effective-Lagrangian program that organises the Goldstone dynamics was completed by Coleman, Wess, and Zumino and by Callan, Coleman, Wess, and Zumino (Phys. Rev. 177, 2239 and 2247, 1969), who proved that the most general phenomenological Lagrangian is fixed by the coset geometry $G / H$ . Adler's soft-pion theorem (Phys. Rev. 137, B1022, 1965) and the partially-conserved-axial-current (PCAC) hypothesis turned the abstract theorem into quantitative predictions for pion physics. The dimensional obstruction of Mermin and Wagner (Phys. Rev. Lett. 17, 1133, 1966), with Coleman's relativistic counterpart, showed that the very infrared fluctuations the theorem produces forbid symmetry breaking in low dimensions — a striking instance of a theorem's conclusion undermining its own hypothesis. Together these results made spontaneous symmetry breaking, rather than any specific fundamental Lagrangian, the organising principle of modern low-energy field theory.

Bibliography Master

@article{Goldstone1961,
  author  = {Goldstone, Jeffrey},
  title   = {Field Theories with `Superconductor' Solutions},
  journal = {Nuovo Cimento},
  volume  = {19},
  year    = {1961},
  pages   = {154--164}
}

@article{GoldstoneSalamWeinberg1962,
  author  = {Goldstone, Jeffrey and Salam, Abdus and Weinberg, Steven},
  title   = {Broken Symmetries},
  journal = {Phys. Rev.},
  volume  = {127},
  year    = {1962},
  pages   = {965--970}
}

@article{NambuJonaLasinio1961,
  author  = {Nambu, Yoichiro and Jona-Lasinio, Giovanni},
  title   = {Dynamical Model of Elementary Particles Based on an Analogy with Superconductivity. I},
  journal = {Phys. Rev.},
  volume  = {122},
  year    = {1961},
  pages   = {345--358}
}

@article{CCWZ1969a,
  author  = {Coleman, Sidney and Wess, Julius and Zumino, Bruno},
  title   = {Structure of Phenomenological Lagrangians. I},
  journal = {Phys. Rev.},
  volume  = {177},
  year    = {1969},
  pages   = {2239--2247}
}

@article{CCWZ1969b,
  author  = {Callan, Curtis G. and Coleman, Sidney and Wess, Julius and Zumino, Bruno},
  title   = {Structure of Phenomenological Lagrangians. II},
  journal = {Phys. Rev.},
  volume  = {177},
  year    = {1969},
  pages   = {2247--2250}
}

@article{Adler1965,
  author  = {Adler, Stephen L.},
  title   = {Consistency Conditions on the Strong Interactions Implied by a Partially Conserved Axial-Vector Current},
  journal = {Phys. Rev.},
  volume  = {137},
  year    = {1965},
  pages   = {B1022--B1033}
}

@article{MerminWagner1966,
  author  = {Mermin, N. David and Wagner, Herbert},
  title   = {Absence of Ferromagnetism or Antiferromagnetism in One- or Two-Dimensional Isotropic Heisenberg Models},
  journal = {Phys. Rev. Lett.},
  volume  = {17},
  year    = {1966},
  pages   = {1133--1136}
}

@book{WeinbergQTFVol2,
  author    = {Weinberg, Steven},
  title     = {The Quantum Theory of Fields, Vol. 2: Modern Applications},
  publisher = {Cambridge University Press},
  year      = {1996}
}

@book{ColemanAspects,
  author    = {Coleman, Sidney},
  title     = {Aspects of Symmetry},
  publisher = {Cambridge University Press},
  year      = {1985}
}

@book{PeskinSchroeder,
  author    = {Peskin, Michael E. and Schroeder, Daniel V.},
  title     = {An Introduction to Quantum Field Theory},
  publisher = {Addison-Wesley},
  year      = {1995}
}

Prerequisites

08.02.02
03.07.05

Tier anchors

beginner: Lancaster, T. & Blundell, S. J., *Quantum Field Theory for the Gifted Amateur* (Oxford, 2014), Chs. 26 (Goldstone modes) and 39 (broken symmetry, spin waves, conceptual); Zee, A., *Quantum Field Theory in a Nutshell*, 2nd ed. (Princeton, 2010), §I.10 and §IV.6 (Goldstone bosons, intuitive)
intermediate: Peskin, M. E. & Schroeder, D. V., *An Introduction to Quantum Field Theory* (Addison-Wesley, 1995), §11.1 (spontaneous symmetry breaking and Goldstone's theorem); Schwartz, M. D., *Quantum Field Theory and the Standard Model* (Cambridge, 2014), Ch. 28 (the Goldstone theorem and the chiral Lagrangian); Srednicki, M., *Quantum Field Theory* (Cambridge, 2007), Ch. 83
master: Weinberg, S., *The Quantum Theory of Fields, Vol. 2: Modern Applications* (Cambridge, 1996), Ch. 19 (spontaneously broken global symmetries; the Goldstone theorem; the structure of the effective Lagrangian; pions as Goldstone bosons; soft-pion theorems and the Adler zero); Coleman, S., *Aspects of Symmetry* (Cambridge, 1985), Ch. 2 (the invariance of the vacuum is the invariance of the world)

References

Weinberg, S. — The Quantum Theory of Fields, Vol. 2: Modern Applications (Cambridge, 1996) · Ch. 19 §19.1 (the Goldstone theorem from the broken charge: $Q|0\rangle$ ill-defined, one massless scalar per broken generator, the spectral / current-Ward-identity proof); §19.2 (the general structure of the broken-symmetry effective Lagrangian; the Goldstone fields as coordinates on the coset $G/H$); §19.4–19.5 (pions as the Goldstone bosons of chiral $SU(2)_L\times SU(2)_R \to SU(2)_V$; the lowest-order chiral Lagrangian $\tfrac{f_\pi^2}{4}\mathrm{tr}(\partial_\mu U^\dagger \partial^\mu U)$; the Adler soft-pion zero and PCAC)
Goldstone, J. — Field theories with 'superconductor' solutions · Nuovo Cimento 19, 154 (1961) — the original demonstration that a relativistic field theory whose Lagrangian has a continuous symmetry broken by the vacuum contains a massless spin-zero boson; the linear-sigma-model example $V = -\mu^2|\phi|^2 + \lambda|\phi|^4$ whose angular fluctuation is massless
Goldstone, J., Salam, A. & Weinberg, S. — Broken symmetries · Phys. Rev. 127, 965 (1962) — three independent proofs of the relativistic Goldstone theorem, including the one assuming only Lorentz invariance and the existence of a conserved current with $\langle 0|j^\mu|\text{state}\rangle \neq 0$; one massless particle per broken generator, with the massless pole forced by the current Ward identity
Nambu, Y. & Jona-Lasinio, G. — Dynamical model of elementary particles based on an analogy with superconductivity · Phys. Rev. 122, 345 (1961) and 124, 246 (1961) — the dynamical (fermion-condensate) origin of chiral symmetry breaking; the pion identified as the near-massless Goldstone boson of broken chiral symmetry, with the Goldberger-Treiman relation as a consequence
Coleman, S., Wess, J. & Zumino, B. — Structure of phenomenological Lagrangians I · Phys. Rev. 177, 2239 (1969); Callan, C. G., Coleman, S., Wess, J. & Zumino, B., II, Phys. Rev. 177, 2247 (1969) — the CCWZ construction: the most general effective Lagrangian for the Goldstone fields is fixed by the geometry of the coset $G/H$, with the Goldstone bosons as coordinates on $G/H$ and the symmetry realised nonlinearly
Adler, S. L. — Consistency conditions on the strong interactions implied by a partially conserved axial-vector current · Phys. Rev. 137, B1022 (1965) — the soft-pion theorem: amplitudes for emission of a Goldstone pion vanish as the pion four-momentum goes to zero (the Adler zero), the field-theoretic signature of the derivative coupling of Goldstone bosons
Mermin, N. D. & Wagner, H. — Absence of ferromagnetism or antiferromagnetism in one- or two-dimensional isotropic Heisenberg models · Phys. Rev. Lett. 17, 1133 (1966) — the infrared divergence of Goldstone-mode fluctuations forbids spontaneous breaking of a continuous symmetry in $d \le 2$; the dimensional counterpart to the Goldstone theorem

Estimated time

beginner: 18m
intermediate: 45m
master: 85m