Intrinsic parity of Dirac fermions

Parity is the simplest of the discrete symmetries — reflect the world through the origin, $\mathbf{x}\to-\mathbf{x}$ , leave the clock alone — and it hides a surprise only a relativistic theory can see. In non-relativistic quantum mechanics a particle's parity is a property of its spatial wavefunction and nothing more. The Dirac equation adds what the Schrödinger equation cannot: a fixed, internal parity with nothing to do with the wavefunction's shape, an intrinsic parity stamped on the field itself. The argument is short and inescapable. Demand that the free Dirac equation keep its form when you invert space, and the four-component spinor cannot simply come along for the ride — it must be rotated by a matrix. Anticommutation of the gamma matrices forces that matrix to be $\gamma^0$ (up to a sign convention), and $\gamma^0$ is diagonal in the rest frame, $\mathrm{diag}(1,1,-1,-1)$ . So the two particle spinors are eigenstates with eigenvalue $+1$ and the two antiparticle spinors with eigenvalue $-1$ : a Dirac fermion and its antiparticle carry opposite intrinsic parity. This is not a measurable absolute — the overall sign is a convention, and particles are always made in pairs — but the relative minus sign is real, and it decides why the pion is a pseudoscalar and why para- and ortho-positronium annihilate into different numbers of photons.

Key takeaways

Parity is spatial inversion, $\mathbf{x}\to-\mathbf{x}$ , $t\to t$ , with $\hat P^2=\mathbb{1}$ so eigenvalues are $\pm1$ .
Requiring the free Dirac equation to keep its form fixes the spinor parity operator to $\hat P=\eta_P\,\gamma^0$ , conventionally $\hat P=+\gamma^0$ .
In the rest frame $\gamma^0=\mathrm{diag}(1,1,-1,-1)$ , so particle spinors have parity $+1$ , antiparticles $-1$ .
A fundamental fermion and its antifermion carry opposite intrinsic parity — a structural fact, not a measurable absolute.
A fermion–antifermion pair with orbital angular momentum $L$ has parity $P=(-1)^{L+1}$ , so $L=0$ $q\bar q$ mesons are pseudoscalars ( $J^P=0^-$ ).
Key idea: demanding the Dirac equation keep its form under spatial inversion forces the spinor parity operator to be $\gamma^0$ , whose rest-frame eigenvalues make particle and antiparticle carry opposite intrinsic parity.

The core insight

Demanding the free Dirac equation keep its form under spatial inversion forces the spinor parity operator to be $\gamma^0$ , whose rest-frame eigenvalues $\mathrm{diag}(1,1,-1,-1)$ make particle and antiparticle carry opposite intrinsic parity.

Where this lives today

Intrinsic parity is quiet infrastructure: exact in the strong and electromagnetic interactions, so every hadron in the Particle Data Group tables carries a measured or inferred $J^P$ , and the architecture of the meson and baryon spectra — pseudoscalar and vector nonets, the parity ordering of excited states — rests on the bound-state rule $P=(-1)^{L+1}$ and its baryonic cousin. Positronium spectroscopy remains a precision laboratory for the same physics, pushing the ortho three-photon rate and searching for forbidden decays as tests of QED. Intrinsic parity turns from bookkeeping into frontier science exactly where it fails: because parity is maximally violated by the weak interaction, precision parity-violating measurements have become a tool to hunt for new forces. Parity-violating electron scattering (the Qweak experiment at Jefferson Lab) measured the proton's weak charge; atomic parity violation in cesium probes the same electroweak couplings; and neutron and electron electric-dipole-moment searches look for the joint violation of parity and time reversal that any explanation of the matter–antimatter asymmetry seems to require. The opposite parity of a fermion and its antifermion, meanwhile, is not negotiable — a theorem of any Lorentz-invariant local field theory, tied by the same $\gamma^0$ to spin-statistics and CPT.

The physical picture

Parity does two things at once, and the confusion in this topic almost always comes from mixing them. It flips the spatial coordinate — so a momentum $\mathbf p$ becomes $-\mathbf p$ while an angular momentum (an axial vector) is untouched — and, because the object inverted is a spinor, it must also multiply the four components by $\gamma^0$ , diagonal in the Dirac–Pauli representation: it leaves the "upper" pair of components alone and flips the sign of the "lower" pair. Since a rest-frame particle lives entirely in the upper components and an antiparticle entirely in the lower, the two are eigenstates with opposite eigenvalues. The first figure puts the spatial flip and the matrix action side by side.

Figure

Parity inverts space through the origin, sending (t, x) to (t, −x) so that momentum p reverses while spin is unchanged and helicity flips; on a Dirac spinor it acts as the matrix γ⁰ = diag(1, 1, −1, −1), giving the rest-frame particle spinors u₁, u₂ eigenvalue +1 and the antiparticle spinors v₁, v₂ eigenvalue −1

The relative sign is invisible in isolation but decisive the moment two fermions meet. Bind a fermion to its antifermion with orbital angular momentum $L$ and the parities multiply: the intrinsic $-1$ combines with the $(-1)^L$ of the spatial wavefunction to give $P=(-1)^{L+1}$ . That formula predicts that the lightest $q\bar q$ mesons — pion, kaon, eta — are all pseudoscalars, $J^P=0^-$ , and it sorts the two ground states of positronium into partners forced by symmetry to annihilate into different numbers of photons. The second figure follows that split to the measured lifetimes.

Figure

The two ground states of positronium sorted by the relative-parity rule P = (−1)^(L+1) and charge conjugation C = (−1)^(L+S): para-positronium ¹S₀ is 0⁻⁺ and annihilates to 2 photons with mean life 0.125 ns, while ortho-positronium ³S₁ is 1⁻⁻, needs 3 photons, and lives 1136 times longer at 142 ns

The core equations

Parity begins as a statement about coordinates and is promoted to one about the field. Inverting space through the origin sends $\mathbf x\to-\mathbf x$ ; acting twice restores everything, so the operator squares to the identity and its eigenvalues can only be $\pm1$ . On a Dirac field the transformation also carries a matrix and an overall phase.

\mathsf{P}:\ (t,\mathbf x)\to(t,-\mathbf x),\qquad \psi(t,\mathbf x)\ \longrightarrow\ \psi'(t,\mathbf x)=\eta_P\,\gamma^0\,\psi(t,-\mathbf x),\qquad \hat P^2=\mathbb{1} \qquad (1)

where: $\mathsf P$ — the parity (spatial-inversion) operation; $\psi$ — the four-component Dirac spinor field; $\eta_P$ — the intrinsic-parity phase, a constant; $\gamma^0$ — the timelike Dirac matrix; $\hat P$ — the operator implementing $\mathsf P$ on states, with $\hat P^2=\mathbb{1}$ forcing eigenvalues $\pm1$ . The matrix $\gamma^0$ is fixed by covariance, as the next equation shows. The form of the operator is forced, not assumed. Requiring the Dirac equation to hold in the inverted frame demands $\hat P$ commute with $\gamma^0$ and anticommute with the three $\gamma^k$ — only $\gamma^0$ (times a phase) does both.

\hat P\,\gamma^0\hat P^{-1}=\gamma^0,\qquad \hat P\,\gamma^k\hat P^{-1}=-\gamma^k\ \ (k=1,2,3)\quad\Longrightarrow\quad \hat P=\eta_P\,\gamma^0,\ \ \eta_P=\pm1 \qquad (2)

where: $\gamma^k$ — the three spacelike Dirac matrices; commuting with $\gamma^0$ and anticommuting with $\gamma^k$ is exactly the algebra of $\gamma^0$ itself. The residual phase $\eta_P$ is pure convention; the near-universal $\eta_P=+1$ gives $\hat P=\gamma^0$ . Derivation A establishes both conditions. With $\hat P=\gamma^0$ in hand, the eigenvalues follow by inspection in the rest frame, where $\gamma^0$ is diagonal and the spinors are simple.

\gamma^0\,u^s(m,\mathbf 0)=+\,u^s(m,\mathbf 0),\qquad \gamma^0\,v^s(m,\mathbf 0)=-\,v^s(m,\mathbf 0),\qquad \gamma^0=\mathrm{diag}(1,1,-1,-1) \qquad (3)

where: $u^s(m,\mathbf 0)$ — the two ( $s=1,2$ ) rest-frame particle spinors, in the upper components; $v^s(m,\mathbf 0)$ — the two antiparticle spinors, in the lower. The particle eigenvalue is $+1$ , the antiparticle $-1$ : fermion and antifermion have opposite intrinsic parity. Only the relative sign is physical. For a particle in motion the operator reverses the momentum while leaving the spin untouched, returning the same intrinsic eigenvalue attached to the reflected momentum state.

\hat P\,u^s(E,\mathbf p)=+\,u^s(E,-\mathbf p),\qquad \hat P\,v^s(E,\mathbf p)=-\,v^s(E,-\mathbf p) \qquad (4)

where: $u^s(E,\mathbf p)$ , $v^s(E,\mathbf p)$ — the moving particle and antiparticle spinors of energy $E$ and momentum $\mathbf p$ . Because $\mathbf p\to-\mathbf p$ with spin fixed, the helicity $h=\mathbf S\cdot\hat{\mathbf p}$ flips sign — with sharp consequences for massless fermions and the weak interaction. Derivation B verifies (3)–(4). Finally, the observable content lives in composite systems. For a fermion bound to an antifermion with orbital angular momentum $L$ and spin $S$ , the intrinsic and orbital parities multiply, with charge conjugation following a companion rule.

P_{f\bar f}=\eta_f\,\eta_{\bar f}\,(-1)^L=-(-1)^L=(-1)^{L+1},\qquad C_{f\bar f}=(-1)^{L+S} \qquad (5)

where: $\eta_f\,\eta_{\bar f}=-1$ is the relative intrinsic parity of the pair; $(-1)^L$ is the parity of the spatial wavefunction $Y_L^m$ ; $C_{f\bar f}=(-1)^{L+S}$ is the charge-conjugation eigenvalue. Setting $L=0$ gives $P=-1$ : the ground-state $q\bar q$ meson is a pseudoscalar. Derivation C applies this rule.

Derivation AWhy the spinor parity operator must be gamma-zero

The goal is to show, with no appeal to convention beyond a single overall sign, that covariance of the free Dirac equation under spatial inversion forces the spinor parity operator to be $\gamma^0$ . We start from the Dirac equation, write the inverted field, demand the equation survive the inversion, and let the anticommutation of the gamma matrices dictate the answer — the one loose thread, an overall phase, being all that convention gets to choose.

Step 1. Write the free Dirac equation as $\big(i\gamma^0\partial_t+i\gamma^k\partial_k-m\big)\psi(t,\mathbf x)=0$ , and define the parity-transformed field $\psi'(t,\mathbf x)=\hat P\,\psi(t,-\mathbf x)$ , with $\hat P$ a constant $4\times4$ matrix to be found. We require $\psi'$ to satisfy the same Dirac equation. Why: "parity is a symmetry" means the transformed field obeys the identical law of motion — if it did not, an experiment could tell the mirror world from ours for free.

Step 2. Compute the derivatives of $\psi'$ . Because its argument is $-\mathbf x$ , the chain rule flips every spatial derivative: $\partial_t\psi'=\hat P\,(\partial_t\psi)(t,-\mathbf x)$ but $\partial_k\psi'=-\hat P\,(\partial_k\psi)(t,-\mathbf x)$ . Substituting into the Dirac equation for $\psi'$ and premultiplying by $\hat P^{-1}$ gives $i\,\hat P^{-1}\gamma^0\hat P\,\partial_t\psi-i\,\hat P^{-1}\gamma^k\hat P\,\partial_k\psi-m\psi=0$ . Why: the sign flip on the spatial derivative is the entire mechanism — it is what must be undone by $\hat P$ acting on the gammas.

Step 3. Compare with the original equation, $i\gamma^0\partial_t\psi+i\gamma^k\partial_k\psi-m\psi=0$ . The two agree for all fields only if $\hat P^{-1}\gamma^0\hat P=\gamma^0$ and $\hat P^{-1}\gamma^k\hat P=-\gamma^k$ : $\hat P$ must commute with $\gamma^0$ and anticommute with each $\gamma^k$ . Why: the mass term already matches, so covariance reduces to a pair of algebraic conditions on how $\hat P$ threads through the gammas.

Step 4. Solve the conditions. The Clifford algebra $\{\gamma^\mu,\gamma^\nu\}=2g^{\mu\nu}$ says $\gamma^0$ commutes with itself and anticommutes with every $\gamma^k$ — exactly the required pattern. Any matrix satisfying both equals $\gamma^0$ up to a scalar, and $\hat P^2=\mathbb{1}$ pins the scalar to a phase with $\eta_P^2=1$ , so $\hat P=\eta_P\gamma^0=\pm\gamma^0$ ; conventionally $\hat P=+\gamma^0$ . Why: the structure that gives particle and antiparticle opposite parity is not put in by hand — it is squeezed out of the demand that the mirror image of a Dirac particle is still a Dirac particle.

Go deeper · Derivation B · 1 minOpposite parity of particle and antiparticle

Derivation BOpposite parity of particle and antiparticle

The aim is to read the intrinsic parities directly off the spinors, show they are opposite for particle and antiparticle, then check that a moving particle simply reverses its momentum under parity while keeping its spin. We use the explicit Dirac–Pauli spinors, act with $\gamma^0$ , and track what happens as the momentum is switched on — the central claim of the article, at the cost of two matrix multiplications.

Step 1. In the Dirac–Pauli representation $\gamma^0$ is block-diagonal, $\mathrm{diag}(\mathbb 1,-\mathbb 1)$ , and the rest-frame spinors are $u^s(m,\mathbf 0)=\sqrt{2m}\,(\xi^s,\,0)^{\top}$ and $v^s(m,\mathbf 0)=\sqrt{2m}\,(0,\,\eta^s)^{\top}$ . Acting with $\gamma^0$ : $\gamma^0 u=+u$ and $\gamma^0 v=-v$ . Why: a particle at rest lives in the upper block that $\gamma^0$ leaves alone, an antiparticle in the lower block that $\gamma^0$ negates — the block structure is the whole story.

Step 2. Read off the eigenvalues. With $\hat P=\gamma^0$ the particle has intrinsic parity $+1$ and the antiparticle $-1$ ; their product is $-1$ . Choosing $\hat P=-\gamma^0$ swaps both signs but leaves the product unchanged. Why: the individual signs depend on the convention $\eta_P=\pm1$ and are not observable, but their ratio is fixed — the relative parity of a fermion and its antifermion is $-1$ .

Step 3. Switch on the momentum. The moving spinor is $u^s(E,\mathbf p)=\sqrt{E+m}\,\big(\xi^s,\ \tfrac{\boldsymbol\sigma\cdot\mathbf p}{E+m}\xi^s\big)^{\!\top}$ . Applying $\gamma^0$ flips the sign of the lower block, which is algebraically identical to negating $\mathbf p$ : $\gamma^0 u^s(E,\mathbf p)=\sqrt{E+m}\,\big(\xi^s,\ \tfrac{\boldsymbol\sigma\cdot(-\mathbf p)}{E+m}\xi^s\big)^{\!\top}=u^s(E,-\mathbf p)$ . Why: parity turns the $\mathbf p$ state into the $-\mathbf p$ state while never touching the spin label $\xi^s$ — momentum reverses, spin stays.

Step 4. Note the consequence for helicity. Since $\mathbf p\to-\mathbf p$ with $\mathbf S$ fixed, $h=\mathbf S\cdot\hat{\mathbf p}$ changes sign: a right-handed state maps to a left-handed one of reversed momentum. For a massive fermion both helicities exist and parity connects them; for a massless fermion of a single helicity there is no partner, so parity is not a symmetry — the seed of weak parity violation. Why: $\gamma^0\gamma^5\gamma^0=-\gamma^5$ means parity exchanges left- and right-handed Weyl components, so a theory built from only one cannot be parity-symmetric.

Go deeper · Derivation C · 1 minParity of a fermion-antifermion bound state

Derivation CParity of a fermion-antifermion bound state

The goal is to derive the working formula $P=(-1)^{L+1}$ for a fermion bound to its antifermion, then apply it to the two cases that matter most: the pseudoscalar mesons and the two states of positronium. We combine the intrinsic parities from equation (3) with the parity of the orbital wavefunction and feed in the quantum numbers — the single line that turns an abstract sign into a testable prediction about the hadron table.

Step 1. Write the two-body state as a product of intrinsic and spatial parts. Each constituent contributes its intrinsic parity, and the relative-motion wavefunction $\psi(\mathbf r)=R(r)\,Y_L^m(\hat{\mathbf r})$ contributes the parity of the spherical harmonic, $Y_L^m(-\hat{\mathbf r})=(-1)^L Y_L^m$ . Why: parity factorises — internal and orbital degrees of freedom transform independently, so their parities multiply.

Step 2. Multiply. With $\eta_f\eta_{\bar f}=(+1)(-1)=-1$ , the total is $P_{f\bar f}=\eta_f\,\eta_{\bar f}\,(-1)^L=-(-1)^L=(-1)^{L+1}$ . Why: the extra minus from the antiparticle is what makes the ground state ( $L=0$ ) parity-odd — without it every $q\bar q$ meson at $L=0$ would be a scalar, and the observed pseudoscalar octet would be impossible.

Step 3. Apply to mesons. The lightest $q\bar q$ mesons sit at $L=0$ , giving $P=-1$ ; with $S=0$ they form the $J^P=0^-$ pseudoscalar nonet (pion, kaon, eta), while $S=1$ gives the $J^P=1^-$ vector nonet ( $\rho$ , $\omega$ , $K^*$ ). Why: the formula organises the entire low-lying meson spectrum, and its first term — the pseudoscalar pion — was the historic test of the whole scheme.

Step 4. Apply to positronium, the cleanest $f\bar f$ system. Both ground states have $L=0$ , so both have $P=-1$ ; they differ in spin. The rule $C=(-1)^{L+S}$ gives $C=+1$ for the singlet $^1S_0$ (para) and $C=-1$ for the triplet $^3S_1$ (ortho). Since a photon carries $C=-1$ , an $n$ -photon state has $C=(-1)^n$ , forcing para $\to 2\gamma$ and ortho $\to 3\gamma$ . Why: parity fixes the states as $0^{-+}$ and $1^{--}$ , and charge conjugation dictates the photon count — a symmetry argument alone predicts the factor-of-a-thousand lifetime gap.

Limiting cases

Case	Result	Meaning
Rest frame	$\gamma^0=\mathrm{diag}(1,1,-1,-1)$ , eigenvalues $\pm1$	intrinsic parity read off cleanly: particle $+1$ , antiparticle $-1$
Moving particle	$\hat P\,u(E,\mathbf p)=u(E,-\mathbf p)$	momentum reverses, spin fixed, so helicity flips sign
Massless / single chirality	parity swaps $L\leftrightarrow R$ ( $\gamma^0\gamma^5\gamma^0=-\gamma^5$ )	no definite parity for one Weyl spinor — root of weak parity violation
Fermion–antifermion, orbital $L$	$P=(-1)^{L+1}$	ground state ( $L=0$ ) is parity-odd; pseudoscalar mesons, $0^{-+}$ para-positronium
Scalar (Klein–Gordon) boson	$\phi(t,\mathbf x)\to\pm\phi(t,-\mathbf x)$ , no $\gamma^0$	a spin-0 particle equals its own antiparticle-parity; no relative minus sign
Convention flip $\eta_P\to-\eta_P$	all intrinsic signs reverse together	absolute parity is unobservable; only relative parity survives

Common trapThe absolute sign of a fermion's intrinsic parity is physically meaningful

It is not. The choice $\hat P=+\gamma^0$ versus $-\gamma^0$ is pure convention, and flipping it reverses every fermion's intrinsic parity at once, so no experiment can detect it. What is real — and measurable — is the relative parity between states a conserved-parity interaction can connect. The parity of a single free fermion is not an observable; you can only measure ratios, such as the fermion–antifermion relative parity ( $-1$ ) or the parity of a meson relative to its decay products — and because particles and antiparticles are always made in pairs, the convention drops out of every amplitude automatically. You are then free to assign a reference: the proton, neutron, electron and lambda are conventionally given intrinsic parity $+1$ , and every other hadron's parity is quoted relative to them.

Worked examples

1 · The parity of the pion from π⁻ capture (fully worked)

A stopped $\pi^-$ is captured by a deuteron from an atomic $S$ -state, producing two neutrons, $\pi^-+d\to n+n$ . Given the deuteron has $J^P=1^+$ and the strong interaction conserves parity, find the intrinsic parity of the pion.

Step 1. Fix the initial angular momentum. The pion is captured from an atomic $S$ -state, so the relative orbital angular momentum is $L_i=0$ . The deuteron carries spin $1$ and the pion spin $0$ , so the total angular momentum is $J=1$ . Both parity and angular-momentum conservation apply, since the capture is a strong process.

Step 2. Write the initial parity. Parity $=P_\pi\,P_d\,(-1)^{L_i}=P_\pi\,(+1)\,(-1)^0=P_\pi$ : the deuteron and the $S$ -wave capture contribute nothing. Initial parity $=P_\pi$ .

Step 3. Constrain the final $nn$ state. Two identical neutrons require a totally antisymmetric wavefunction. Spatial exchange gives $(-1)^L$ , spin exchange $(-1)^{S+1}$ , so antisymmetry demands $L+S$ even. With $J=1$ , the only allowed term is $^3P_1$ ( $L=1$ , $S=1$ ): the candidates $^3S_1$ , $^1P_1$ ( $L+S=1$ ) and $^3D_1$ ( $L+S=3$ ) are all odd and forbidden. The Pauli principle does the selecting; only $^3P_1$ survives.

Step 4. Match parities. The final $^3P_1$ state has parity $(-1)^{L}=-1$ . Strong-interaction parity conservation sets initial $=$ final: $P_\pi=-1$ . The pion is a pseudoscalar, $J^P=0^-$ — exactly what $P=(-1)^{L+1}=-1$ predicts for an $L=0$ $q\bar q$ state, and the historic confirmation of that rule.

2 · Positronium: parity, C-parity, and the photon count (fully numerical)

Using $P=(-1)^{L+1}$ and $C=(-1)^{L+S}$ , assign $J^{PC}$ to the two ground states of positronium and predict the annihilation-photon number for each, then compare with the measured lifetimes.

Step 1. List the ground states. Both have $L=0$ ; they split by spin into the singlet $^1S_0$ ( $S=0$ , "para", $J=0$ ) and the triplet $^3S_1$ ( $S=1$ , "ortho", $J=1$ ).

Step 2. Parity. With $L=0$ , $P=(-1)^{0+1}=-1$ for both states — parity alone does not distinguish them. $P=-1$ for para and ortho alike.

Step 3. Charge conjugation and photon count. $C=(-1)^{L+S}$ gives $C=+1$ for para and $C=-1$ for ortho. A photon has $C=-1$ , so $n$ photons carry $C=(-1)^n$ ; conservation of $C$ forces para into an even photon number and ortho into an odd one. The lowest allowed counts are para $\to 2\gamma$ and ortho $\to 3\gamma$ (single-photon decay is forbidden by momentum conservation). Full assignments: $J^{PC}=0^{-+}$ (para), $1^{--}$ (ortho).

Step 4. Compare with experiment. The three-photon channel is phase-space and coupling suppressed, so ortho lives far longer: the measured mean lifetimes are $\tau_{\text{para}}=0.125\ \text{ns}$ and $\tau_{\text{ortho}}=142\ \text{ns}$ , a ratio of about $1136$ . The symmetry bookkeeping — parity fixing the states, $C$ -parity the photon number — predicts both the $2\gamma$ / $3\gamma$ split and the thousand-fold lifetime gap, with no dynamics beyond counting signs.

Physics of the portrait

Intrinsic parity is quantum bookkeeping so routine that the PDG prints a $J^P$ next to every particle, but it began as a genuinely new idea: that a symmetry of space could be a conserved quantum number in its own right. The person who put it there was an atomic spectroscopist reaching for group theory.

Origins · 1 minEugene Wigner and how this was discovered

Eugene Wigner

1902–1995 · parity as a conservation law and the symmetries of quantum mechanics

Eugene Paul Wigner — Hungarian-born, chemical-engineering trained, and one of the great mathematical physicists of the twentieth century — introduced parity into physics in 1927, before Dirac's antiparticles and a generation before anyone doubted it. Studying atomic spectra, Wigner realised that invariance under spatial reflection implied a genuinely conserved quantity, a $\pm1$ label that selection rules must respect — the first time a discrete spacetime symmetry was treated on the same footing as continuous rotations and translations. It was part of a larger programme, Wigner's systematic application of group representation theory to quantum mechanics, which in 1939 classified relativistic particles by the irreducible representations of the Poincaré group — the modern definition of "spin" and "mass" themselves. When Dirac's equation supplied the concrete parity operator $\gamma^0$ , and hence an intrinsic parity carried by the field, it slotted directly into that framework. Parity's reign as an exact symmetry ended in 1956–57, when Lee and Yang questioned it and Wu's cobalt-60 experiment found it violated by the weak force — but the intrinsic parity of the strong and electromagnetic sectors is still exact, still tabulated, and still Wigner's. He shared the 1963 Nobel Prize "for his contributions to the theory of the atomic nucleus and the elementary particles, particularly through the discovery and application of fundamental symmetry principles."

History

1927

— Wigner introduces parity as a conserved quantum number in atomic spectra

1928

— Dirac's relativistic equation supplies the spinor parity operator γ⁰, giving fermion and antifermion opposite intrinsic parity

1939

— Wigner classifies relativistic particles by representations of the Poincaré group, fixing what spin and mass mean

1951

— Panofsky's measurements of π⁻ capture point toward a pseudoscalar pion

1954

— Chinowsky and Steinberger fix the pion parity as negative via π⁻ d → n n

1956

— Lee and Yang propose that parity may not be conserved in the weak interaction

1957

— Wu's cobalt-60 experiment confirms parity violation in beta decay

Going further

Texts. Thomson, Modern Particle Physics, §4.7–4.9 (parity of Dirac spinors and antiparticles) · Griffiths, Introduction to Elementary Particles, 2nd ed., §4.4 and Ch. 4 Appendix (discrete symmetries, parity of the pion) · Halzen & Martin, Quarks and Leptons, §5.5–5.7 (bilinear covariants and parity) · Peskin & Schroeder, An Introduction to Quantum Field Theory, §3.6 (discrete symmetries of the Dirac field) · Schwartz, Quantum Field Theory and the Standard Model, §11.5 (P, C and T on fermions); Bjorken & Drell, Relativistic Quantum Mechanics, §5. The key results of this article are: the commutation/anticommutation conditions of equation (2) force $\hat P\propto\gamma^0$ ; acting on rest-frame spinors gives $\gamma^0 u^s(m,\mathbf 0)=+u^s$ and $\gamma^0 v^s(m,\mathbf 0)=-v^s$ with $\gamma^0=\mathrm{diag}(1,1,-1,-1)$ ; for a moving spinor, $\gamma^0 u^s(E,\mathbf p)=u^s(E,-\mathbf p)$ ; the bound-state rules $P=(-1)^{L+1}$ and $C=(-1)^{L+S}$ give para-positronium $0^{-+}\to2\gamma$ and ortho $1^{--}\to3\gamma$ , and the pion $J^P=0^-$ from the $\pi^-d\to nn$ selection to the $^3P_1$ state. The positronium lifetimes $\tau_{\text{para}}=0.125\ \text{ns}$ and $\tau_{\text{ortho}}=142\ \text{ns}$ are standard tabulated values.

Prerequisites. The Dirac Equation · Free-Particle Dirac Solutions · The Covariant Dirac Equation. Leads to. Antiparticles and the CPT Picture · Spin and Helicity.

Ask the tutor: Ask me to run the $\pi^-d\to nn$ selection rule step by step, or to build the $J^{PC}$ of any $q\bar q$ meson or positronium state from the parity and C-parity rules.

Why do a fermion and its antifermion have opposite intrinsic parity?

Because the parity operator on a Dirac field is $\gamma^0$ , diagonal in the rest frame with the block structure $\mathrm{diag}(1,1,-1,-1)$ . A particle at rest occupies the upper two components, which $\gamma^0$ leaves alone (eigenvalue $+1$ ); an antiparticle occupies the lower two, which $\gamma^0$ negates (eigenvalue $-1$ ). The opposite sign is not a physical input but a direct algebraic consequence of the one matrix that covariance forced upon us. It is a theorem more generally: any spin- $\tfrac12$ particle and its antiparticle must have opposite intrinsic parity, whereas a spin-0 boson and its antiparticle have the same parity — a difference knitted together with spin-statistics and CPT.

If the absolute parity is just a convention, what is actually measured?

Only relative parities. You can never measure the parity of an isolated free fermion, because shifting the convention $\eta_P\to-\eta_P$ flips every fermion's intrinsic parity simultaneously and cancels out of every amplitude. What you can measure is the parity of one state relative to others a parity-conserving interaction connects — the pion relative to the nucleon (via $\pi^-d\to nn$ ), or the relative parity of a fermion and its antifermion ( $-1$ ). By convention the proton, neutron, electron and lambda are assigned $+1$ , and all other hadron parities are quoted against that reference.

How is the pion known to be a pseudoscalar rather than a scalar?

By the classic $\pi^-d\to nn$ capture experiment (worked example 1). A stopped $\pi^-$ is caught by a deuteron from an atomic $S$ -state, so the initial parity is just $P_\pi$ times the deuteron's $+1$ . The final two neutrons are identical fermions, so the Pauli principle allows only a $^3P_1$ state for total $J=1$ , and that state has parity $(-1)^1=-1$ . Strong-interaction parity conservation then gives $P_\pi=-1$ : the pion is $J^P=0^-$ , consistent with $P=(-1)^{L+1}=-1$ for an $L=0$ quark-antiquark meson. It is one of the cleanest determinations of an intrinsic parity in all of physics.

Does parity flip a particle's helicity or its chirality?

Both, and this is where intrinsic parity meets the weak interaction. Parity reverses momentum but not spin, so the helicity $h=\mathbf S\cdot\hat{\mathbf p}$ changes sign — a right-handed electron of momentum $\mathbf p$ maps to a left-handed electron of momentum $-\mathbf p$ . At the field level, $\gamma^0\gamma^5\gamma^0=-\gamma^5$ means parity interchanges the chiral projections $\psi_{L,R}=\tfrac12(1\mp\gamma^5)\psi$ . A theory with only one chirality — like the Standard Model's left-handed neutrino — therefore cannot be parity-symmetric, since parity would demand a right-handed partner that does not exist. This is the structural reason the weak interaction violates parity maximally.

What sets the number of photons in a positronium decay?

Charge conjugation, working alongside parity. Both ground states are parity-odd ( $P=-1$ ), so parity alone does not separate them; the discriminator is $C=(-1)^{L+S}$ , which is $+1$ for para ( $^1S_0$ ) and $-1$ for ortho ( $^3S_1$ ). A photon carries $C=-1$ , so an $n$ -photon state has $C=(-1)^n$ ; conservation of $C$ then forces para into an even number of photons ( $2\gamma$ ) and ortho into an odd number ( $3\gamma$ ). Single-photon decay is separately forbidden by energy-momentum conservation. The upshot — para lives $0.125\ \text{ns}$ and ortho $142\ \text{ns}$ — is pure symmetry, and any anomalous rate would signal new physics.

Is intrinsic parity always conserved?

No — it is conserved by the strong and electromagnetic interactions but violated, maximally, by the weak interaction. That is why the $J^P$ of hadrons is a good quantum number (their spectroscopy is governed by strong dynamics) yet a weak decay such as beta decay does not respect it. Parity violation, discovered by Wu in 1957, is a statement about the dynamics of the weak force, not about whether intrinsic parity exists — the label is still defined; the weak Hamiltonian simply fails to commute with $\hat P$ . The opposite parity of a fermion and its antifermion remains a rigorous theorem regardless.