Skip to content

Chiral Effective Theory and Nonlinear Symmetry

Chiral effective theory is the low-energy expansion for Goldstone bosons of a spontaneously broken approximate symmetry. Its fields are coordinates on the vacuum manifold, its symmetry acts nonlinearly on those coordinates, and its small parameters are derivatives and explicit-breaking scales divided by the first omitted scale. For low-energy QCD this architecture produces the pion Lagrangian and its loop expansion; it does not by itself determine how shallow nuclear channels must be iterated.

Required background. A Map of Effective-Theory Architectures supplies the common architecture card. Cosets and Nonlinear Realizations supplies the G/HG/H construction and compensating unbroken transformation used below.

Helpful background. Explicit Breaking and Pseudo-Goldstone Modes explains why weak breaking gives the Goldstones small masses rather than removing the nonlinear realization.

Goldstone coordinates turn symmetry into an expansion

Section titled “Goldstone coordinates turn symmetry into an expansion”

Take two-flavor QCD in the limit mu=md=0m_u=m_d=0. Ignoring the anomalous singlet axial transformation, its global chiral group and spontaneous breaking pattern are

G=SU(2)L×SU(2)RH=SU(2)V.G=SU(2)_L\times SU(2)_R \longrightarrow H=SU(2)_V.

The three pion fields coordinatize G/HG/H. A convenient matrix coordinate is

U(x)=exp ⁣(iτaπa(x)F),ULUR,U(x)=\exp\!\left(\frac{i\tau^a\pi^a(x)}{F}\right), \qquad U\longrightarrow LUR^\dagger,

where τa\tau^a are Pauli matrices and FF is the pion decay constant in the chiral limit. Another regular coordinate—stereographic coordinates, a square-root field uu with u2=Uu^2=U, or a different exponential convention—describes the same on-shell physics. The pion fields themselves are chart coordinates, not invariant observables.

In the exact symmetry limit, no nonconstant potential can be invariant under independent LL and RR: UU=1U U^\dagger=1 reduces every derivative-free invariant to a constant. Goldstone interactions therefore vanish with momentum. At two-derivative order the parity-even invariant is, up to normalization,

L2,kin=F24Tr ⁣(DμUDμU).\mathcal L_{2,\mathrm{kin}} =\frac{F^2}{4}\, \operatorname{Tr}\!\left(D_\mu U D^\mu U^\dagger\right).

With no external sources, DμU=μUD_\mu U=\partial_\mu U. Background left- and right-handed sources replace the ordinary derivative by a chiral covariant derivative, allowing current correlation functions and Ward identities to be generated without choosing a pion-field coordinate as an observable. Leutwyler proves that, subject to locality, clustering, Lorentz invariance, and the stated anomaly qualification, the low-energy generating functional can be represented by such a symmetry-invariant Goldstone action to all orders in the derivative expansion in Leutwyler 1994, §§ 2–9, printed pp. 2–26, Open PDF.

This is the nonlinear symmetry manifold branch of the architecture map. The branch selects the form of the fields and invariants, but the complete card still needs counting, matching inputs, an observable, an uncertainty, and a breakdown condition.

The organizing feature of a low-energy problem selects one or more EFT architecture branches, but every branch must complete the same card before producing a controlled prediction and linking onward to detailed applications.

An EFT name is not a construction. Starting from the observable, state, scale hierarchy, and target accuracy, identify the dominant low-energy organizing structure, then declare degrees of freedom, symmetry and state, counting, matching or input, observables, uncertainty, and breakdown. Branches may be nested; the diagram is schematic and not to scale.

The light-quark mass term couples left- and right-handed quarks,

Lm=qˉLMqRqˉRMqL,M=diag(mu,md).\mathcal L_m =-\bar q_L\mathcal M q_R -\bar q_R\mathcal M^\dagger q_L, \qquad \mathcal M=\operatorname{diag}(m_u,m_d).

Temporarily assign the mass matrix the transformation

MLMR.\mathcal M\longrightarrow L\mathcal M R^\dagger.

One may then build formally invariant operators and only afterward set M\mathcal M to its physical constant value. Define

χ=2BM,χLχR,\chi=2B\mathcal M, \qquad \chi\longrightarrow L\chi R^\dagger,

where BB is a low-energy constant. The complete leading even-parity meson Lagrangian is

L2=F24Tr ⁣(DμUDμU+χU+Uχ).\mathcal L_2 =\frac{F^2}{4}\operatorname{Tr}\!\left( D_\mu U D^\mu U^\dagger + \chi U^\dagger+U\chi^\dagger \right).

The spurion is a construction device, not a claim that the physical mass matrix transforms. Once it is frozen, it records precisely how the symmetry is broken and prevents symmetry-violating operators from appearing with unrelated coefficients. Leutwyler derives this extension of the invariance argument and the assignment of explicit-breaking parameters as spurions in Leutwyler 1994, § 10, printed pp. 27–29, Open PDF.

For the isospin-symmetric choice M=mq1\mathcal M=m_q\mathbf 1, expand UU to quadratic order:

L2=12μπaμπa12M2πaπa+O ⁣(π4F2),M2=2Bmq.\mathcal L_2 =\frac12\partial_\mu\pi^a\partial^\mu\pi^a -\frac12M^2\pi^a\pi^a +O\!\left(\frac{\pi^4}{F^2}\right), \qquad M^2=2Bm_q.

Because an on-shell pseudo-Goldstone obeys p2M2p^2\sim M^2, chiral counting assigns

μ=O(p),χ=O(p2),U=O(1).\partial_\mu=O(p), \qquad \chi=O(p^2), \qquad U=O(1).

Thus two derivatives and one quark-mass insertion enter at the same chiral order. For unequal masses the same spurion operator produces isospin breaking; electromagnetic sources and other weak breakings require their own transformation rules and counting. The power assignment is justified only while the induced pseudo-Goldstone masses remain well below the first omitted resonance or other non-Goldstone scale.

Let a connected mesonic diagram contain LL loops and ViV_i vertices of chiral order did_i, where one mass-spurion insertion counts as two derivatives. Four-dimensional loop integration contributes four powers of momentum, an internal pion propagator contributes p2p^{-2}, and topology eliminates the number of internal lines. The net chiral order is

ν=2+2L+iVi(di2).\nu =2+2L+\sum_i V_i(d_i-2).

Every vertex has di2d_i\ge2. Consequently only finitely many loop and vertex combinations contribute at fixed ν\nu:

  • O(p2)O(p^2) consists of trees from L2\mathcal L_2;
  • O(p4)O(p^4) consists of one-loop graphs made only from L2\mathcal L_2 and trees with one L4\mathcal L_4 insertion; and
  • each additional loop raises the order by two powers of pp.

The one-loop divergences from L2\mathcal L_2 are local and have the form of the symmetry-allowed O(p4)O(p^4) operators. Their renormalized low-energy constants cancel the subtraction-scale dependence of the chiral logarithms. Schematically,

Ap4=p4F4{116π2[clogln ⁣p2μ2+cloop]+iir(μ)fi},\mathcal A_{p^4} =\frac{p^4}{F^4}\left\{ \frac{1}{16\pi^2} \left[c_{\log}\ln\!\frac{-p^2}{\mu^2}+c_{\mathrm{loop}}\right] +\sum_i \ell_i^r(\mu)\,f_i \right\},

where the process-dependent tensors fif_i are fixed by the O(p4)O(p^4) basis. The logarithm is not a complete prediction by itself: the scale-dependent local constants are required at the same order. Manohar derives the loop formula, its topological reduction, and the estimate Λχ4πF\Lambda_\chi\lesssim4\pi F in Manohar 1996, §§ 11–13, printed pp. 34–45, Open PDF.

A practical expansion parameter is

ϵχ=max ⁣(pΛb,MπΛb),\epsilon_\chi =\max\!\left( \frac{|p|}{\Lambda_b}, \frac{M_\pi}{\Lambda_b} \right),

with Λb\Lambda_b set by the lowest omitted resonance, inelastic threshold, or other state rather than automatically by 4πF4\pi F. A truncation claim should vary the kinematics and compare successive orders; changing μ\mu only probes part of the omitted contribution. An anomaly can also require a Wess–Zumino–Witten term whose transformation reproduces the microscopic anomaly. It is an allowed, quantized addition to the architecture, not a failure of nonlinear symmetry.

First application: the leading pion amplitude

Section titled “First application: the leading pion amplitude”

The construction can now be tested without fitting an operator list. Expanding the unique O(p2)O(p^2) Lagrangian through four pion fields gives the tree amplitude

Mab;cd(s,t,u)=δabδcdA(s,t,u)+δacδbdA(t,s,u)+δadδbcA(u,t,s),\begin{aligned} \mathcal M^{ab;cd}(s,t,u) ={}&\delta^{ab}\delta^{cd}A(s,t,u) +\delta^{ac}\delta^{bd}A(t,s,u)\\ &+\delta^{ad}\delta^{bc}A(u,t,s), \end{aligned}

with

A(s,t,u)=sM2F2+O(p4),s+t+u=4M2.A(s,t,u)=\frac{s-M^2}{F^2}+O(p^4), \qquad s+t+u=4M^2.

Here FF and M2=2BmqM^2=2Bm_q are leading-order parameters; replacing them by physical FπF_\pi and MπM_\pi consistently shifts terms at higher order. This result passes four independent checks:

  1. Crossing: the three Kronecker structures permute with ss, tt, and uu.
  2. Soft behavior: in the chiral limit the amplitude vanishes when an external momentum is taken soft, reflecting derivative Goldstone couplings.
  3. Dimensions: s/F2s/F^2 is dimensionless, as a four-dimensional scattering amplitude must be in this normalization.
  4. Coordinate independence: a regular redefinition of UU changes off-shell vertices but not this on-shell amplitude.

Gasser and Leutwyler organize the current generating functional, the leading effective Lagrangian, its one-loop completion, and pion scattering in Gasser and Leutwyler 1984, §§ 3–7 and 17–18, pp. 147–159 and 179–182. The detailed extraction of FF, BB, higher-order constants, quark-mass information, and comparison with hadron data belongs to Chiral Lagrangians and Low-Energy QCD.

The chiral architecture card and its boundary

Section titled “The chiral architecture card and its boundary”
EntryChiral-EFT choice
Degrees of freedomGoldstone fields on G/HG/H, with matter fields added only when the target process requires them.
Hierarchy and stateQ/Λχ1Q/\Lambda_\chi\ll1, including derivatives and explicit-breaking scales, about the symmetry-breaking vacuum.
Symmetry and localityNonlinear realization of GG with local invariants and spurions for controlled explicit breaking.
Power countingDerivatives, light masses, loops, and any declared matter-field expansion are assigned a homogeneous chiral order.
Matching and inputsLow-energy constants are matched to the microscopic theory, lattice calculations, or data at a stated scale and scheme.
ObservablesSoft amplitudes, current matrix elements, and low-energy hadronic observables within the declared field content.
UncertaintyFirst omitted chiral order, low-energy-constant inputs, numerical errors, and the proximity of omitted resonances or inelastic thresholds.
Detailed applicationChiral Lagrangians and Low-Energy QCD.

Nonlinear chiral symmetry continues to constrain theories containing nucleons, but it no longer supplies the entire infrared organization. A single baryon introduces a large mass and recoil expansion. Two or more nucleons can introduce anomalously large scattering lengths and shallow poles, so reducible diagrams are enhanced and selected interactions must be iterated. The counterterms required by that iteration depend on the few-body counting and regulator analysis, not on the pion derivative expansion alone.

This is why “chiral EFT” and “nuclear EFT” are not interchangeable. The former statement identifies Goldstone symmetry, spurions, and chiral order. The latter must additionally state whether pions are explicit, which channels are nonperturbative, how forces and currents are ordered, and how cutoff independence is demonstrated. Epelbaum, Hammer, and Meißner separate the natural mesonic hierarchy from the infrared enhancement of nuclear amplitudes in Epelbaum, Hammer, and Meißner 2009, §§ I.D and II.A, pp. 1778–1780 and 1784–1787. The next architecture is developed on Nuclear and Few-Body Effective Theory.

The same caution applies across subjects. HEFT also uses a nonlinear symmetry manifold, but its electroweak coset, scalar content, hierarchy, and loop counting differ from those of low-energy QCD; see SMEFT and HEFT: Linear versus Nonlinear Electroweak Realizations. Sharing a CCWZ construction does not make two EFTs the same architecture card.

Writing a pion potential in the exact symmetry limit. A nonconstant derivative-free function of UU violates the independent left–right action. A potential becomes allowed only through spurions or other declared explicit breaking.

Counting quark masses as one derivative. Pseudo-Goldstone masses obey Mπ2mqM_\pi^2\propto m_q, so mqm_q and χ\chi count as p2p^2. Counting them as pp destroys the loop hierarchy and mixes nominal orders.

Keeping only chiral logarithms at next order. One-loop logarithms and renormalized local O(p4)O(p^4) constants occur together. The logarithm alone is generally scale dependent and is not a complete amplitude.

Treating a field parametrization as physical. Exponential and other regular coordinates change off-shell vertices. Only invariant currents, on-shell amplitudes, and consistently transformed sources may be compared.

Exporting mesonic counting unchanged to nuclei. Chiral symmetry still constrains operators, but shallow poles and reducible propagation can promote interactions and require iteration. That additional infrared decision belongs to the nuclear/few-body architecture.

  • Epelbaum, Evgeny, Hans-Werner Hammer, and Ulf-G. Meißner. 2009. “Modern Theory of Nuclear Forces.” Reviews of Modern Physics 81 (4): 1773–1825. DOI. Open PDF.

  • Gasser, Jürg, and Heinrich Leutwyler. 1984. “Chiral Perturbation Theory to One Loop.” Annals of Physics 158 (1): 142–210. DOI. CERN record.

  • Leutwyler, Heinrich. 1994. “On the Foundations of Chiral Perturbation Theory.” Annals of Physics 235 (1): 165–203. DOI. Open PDF.

  • Manohar, Aneesh V. 1996. “Effective Field Theories.” Lectures at the Schladming Winter School. arXiv:hep-ph/9606222. Open PDF.