# Is Parity really violated? (Even though neutrinos are massive)

The weak force couples only to left-chiral fields, which is expressed mathematically by a chiral projection operator $P_L = \frac{1-\gamma_5}{2}$ in the corresponding coupling terms in the Lagrangian.

This curious fact of nature is commonly called parity violation and I'm wondering why? Does this name make sense from a modern point of view?

My question is based on the observation:

A Dirac spinor (in the chiral representation) of pure chirality transforms under parity transformations:

$$\Psi_L = P_L \Psi = \begin{pmatrix} \chi_L \\ 0 \end{pmatrix} \rightarrow \Psi_L^P = \begin{pmatrix} 0\\ \chi_L \end{pmatrix} \neq \Psi_R$$

Chirality is a Lorentz invariant quantity and a left-chiral particle is not transformed into a right-chiral particle by parity transformations.(The transformed object lives in a different representation of the Lorentz group, where the lower Weyl spinor denotes the left-chiral part.)

In understand where the name comes from historically (see the last paragraph) but wouldn't from a modern point of view chirality violation make much more sense?

Some background:

Fermions are described by Dirac spinors, transforming according to the $(\frac{1}{2},0) \oplus (0,\frac{1}{2})$ representation of the (double cover of the) Lorentz group. Weyl spinors $\chi_L$ transforming according to the $(\frac{1}{2},0)$ representation are called left-chiral and those transforming according to the $(0,\frac{1}{2})$ representation are called right-chiral $\xi_R$. A Dirac spinor is (in the chiral representation)

$$\Psi = \begin{pmatrix} \chi_L \\ \xi_R \end{pmatrix}$$

The effect of a parity transformation is

$$(\frac{1}{2},0) \underbrace{\leftrightarrow}_P (0,\frac{1}{2}),$$ which means the two irreps of the Lorentz group are exchanged. (This can be seen for example by acting with a parity transformation on the generators of the Lorentz group). That means a parity transformed Dirac spinor, transforms according to the $(0,\frac{1}{2}) \oplus (\frac{1}{2},0)$ representation, which means we have

$$\Psi = \begin{pmatrix} \chi_L \\ \xi_R \end{pmatrix} \rightarrow \Psi^P = \begin{pmatrix} \xi_R \\ \chi_L \end{pmatrix}$$

Now we can examine the effect of a parity transformation on a state with pure chirality:

$$\Psi_L = P_L \Psi = \begin{pmatrix} \chi_L \\ 0 \end{pmatrix} \rightarrow \Psi_L^P = \begin{pmatrix} 0\\ \chi_L \end{pmatrix}$$

This means we still have a left-chiral spinor, only written differently, after a parity transformation and not a right-chiral. Chirality is a Lorentz invariant quantity. Nevertheless, the fact that only left-chiral particles interact weakly is commonly called parity violation and I'm wondering if this is still a sensible name or only of historic significance?

Short remark on history

I know that historically neutrinos were assumed to be massless, and for massless particles helicity and chirality are the same. A parity transformation transforms a left-handed particle into a right-handed particle. In the famous Wu experiment, only left-handed neutrinos could be observed, which is were the name parity violation comes from. But does this name make sense today that we know that neutrinos have mass, and therefore chirality $\neq$ helicity.

• Particles in a world consisting of nothing but left-handed particles move differently than particles in an otherwise identical world consisting of nothing but right-handed particles. Right? That's all that's meant by partity violation. Oct 12 '15 at 3:35

Yes, parity is really violated, even if neutrinos are massive. You seem to be confusing the relationship between parity, helicity, and chirality in the modern standard model with the physical symmetry operation of spatial inversion.

Wu's experiment did not measure neutrino helicity. Wu and collaborators prepared a thin layer of a beta-emitting nucleus with rather high spin, polarized the nuclei, and observed that the beta particles were more likely to be emitted from the "south pole" of the nucleus than from the "north pole." Why is this a parity violation? The way you define the "north pole" of a rotating object is to grab it with your right hand, so that the fingers of your right hand curl around in the same sense as the rotation; the "north pole" is the one where your thumb goes, and the "south pole" is the other one. Mirror reflection (which is a special case of the parity transformation) turns your right hand into a left hand, and changes which pole you label as "north" for the same sense of rotation.

The following paper in that issue of PRL, by Garwin et al, shows that muons produced in the decay $\pi\to\mu+\nu$ are polarized. Such polarization is a violation of parity symmetry because the pion has zero spin, and a stopped pion can therefore express no preference about the spin directions of its daughters. This experiment was also the first to measure the magnetic moments for the muon and antimuon.

The statement that one can produce polarized $\mu$ from spinless $\pi$ in the decay $\pi\to\mu+\nu$ suggests that in that decay the neutrinos should also be produced polarized. However the first measurement of neutrino helicity is generally taken to be the one by Goldhaber and collaborators, later in the same year. Analysis of the Goldhaber experiment requires you to spend a little time thinking carefully about nuclear spins.

It was actually discovered in 1927 that unpolarized beta sources produce slightly left-handed polarized electrons, though the significance was not understood at the time and the paper was forgotten until unearthed by Grodzins in 1958. Allan Franklin calls it "the non-discovery of parity non-conservation."

Parity violation is real in the visceral sense that if you were to show me a (sufficiently detailed) photograph of a weak interaction experiment with polarized particles, I could in principle tell you whether the photograph had been reflected or not.

I guess it is because you first of all change sign of $\vec x$ to $- \vec x$ in physical space(this is parity transformation in a nutshell). All this peculiar algebra concerning left and right chirality fields comes from $J = 1/2$ Lorenz group representation, so transformation rules are defined as representatives of parity transformation of physical space.

• Hi @Vladimir, unfortunately, I'm not sure what you're trying to say. What does parity violation mean for you? And in what sense is it then really violated?
– jak
Dec 10 '14 at 7:55
• Hi, @JakobH, for me parity violation means that inversion of space coordinates does not keep law of physics the same. People used to expect law of physics to remain the same under this transformation, e.g. top rotating clockwise under this transformation will rotate counterclockwise; all observables will "revert". However, parity violation means that some laws shouldn't change this way. Wu experiment link itself is a good example. What I try to say that notion of flip of sign of coordinate (parity inversion) is primal for parity transformation, Dec 10 '14 at 12:30
• ..., and specific laws for neutrinos (and other fields of spin 1/2) are just representation of physical notion of inversion. Your discussion concerns formalism of this transformation (transformation matrices, definitions of left and right spinors etc). This formalism is constructed in such way that appropriate observables change sign. By the way, this transformation is applied to all fields, even those for which notion of chirality makes no sense (for example, for scalar fields wavefunction is just plain function and parity transformation is just f(x) -> f(-x)). Dec 10 '14 at 12:44
• I know about the Wu experiment, but what do you mean by "some laws shoudn't change this way"? As far as I know the result of the Wu eperiment is, that only left-chiral particles interact weakly. I'm wondering why this is called parity violation.
– jak
Dec 10 '14 at 12:58
• There is a illustration in article of Wu experiment, that demonstates violation of parity. We would expect (implying parity conservation) that experiment which in our world has preferred direction in mirror world has same preferred direction, but it is not. In our world, only one sort of spin 1/2 particles interact with W-boson. In mirror world (after parity transformation), this sort of particles do not, but other sort do interact with W-boson. So physics are different after parity transformation. Dec 10 '14 at 13:24

Weak interactions with $W$ and $Z$ gauge bosons violate parity simply because the righ-handed and the left-handed fermions coupled differently to $W$ and $Z$. For example, the $W$'s couple only to the left-handed fields. A parity inviariant dynamics would require that both left- and right- fields couple in the same way to the gauge vector since they get exchanged under parity transformation.

Slightly more technically , the gauge bosons must transform as polar vectors under parity (because they form a covariant derivative with $\partial_\mu$ which is polar) and hence in order to preserve parity they should couple to polar vector current $J^\mu$ such as $\bar{\Psi}\gamma^\mu\Psi$. Instead the $W$'s couple to $V-A$ current of the type $\bar{\Psi}\gamma^\mu(1-\gamma^5)\Psi$. While the $\gamma^\mu$ term transforms as polar vector, it's not difficult to see that the $\gamma^\mu\gamma^5$ term transform instead as an axial vector. To restore parity one should add a right-current $\bar{\Psi}\gamma^\mu(1+\gamma^5)\Psi$ that couples with the same strength to $W$ so that the $\gamma^5$ would drop.

• Hi, thanks for you answer. Do you refer in the first paragraph to helicity or chirality when speaking of left- and right-handed fermions? As far as I know, a left-handed (helicity) field is transformed into a right-handed field, but a left-chiral field, stays left-chiral. In the standard model "parity violation" is incorporated by using $1-\gamma_5$, which is the Chirality projection operator. If I understand you correctly you say that parity is violated because under parity $\bar{\Psi}\gamma^\mu(1-\gamma^5)\Psi \rightarrow \bar{\Psi}\gamma^\mu(1+\gamma^5)\Psi$ ?!
– jak
Dec 15 '14 at 8:18
• In my opinion this is not correct, because we have $\Psi \rightarrow \gamma_0 \Psi$, which cancels the minus sign we get from $\gamma_5 \rightarrow - \gamma_5$ and yields $\bar{\Psi}\gamma^\mu(1-\gamma^5)\Psi \rightarrow \bar{\Psi}\gamma^\mu(1-\gamma^5)\Psi$...
– jak
Dec 15 '14 at 8:21
• @JacobH I am sorry but in fact the V-A current gets transformed into a V+A current. It is not a question of opinin, but just of doing the transformation correctly. Think about it, a term $\bar{\Psi}\gamma^\mu\gamma^5\Psi$ isnt called an axial current for nothing. The $\gamma_5$ term flips sign because of the trasformation rule of $\Psi$, there is no extra minus to be inserted. As for your question, I am referring to the states projected with $1-\gamma_5$, but it down matter much since you can ask the same question even in the limit of massless particles when the Higgs vev is turned off. Dec 15 '14 at 18:54
• @JakobH I had a problem with my automatic corrector... 'Down matter' was meant to be 'doesn't matter'... I hope the rest is clear enough. Dec 15 '14 at 19:20

Let's check that parity is violated by the weak interaction lagrangian: $$\mathcal{L}(x) = \bar{\psi}(x) \gamma^\mu \frac{(1-\gamma^5)}{2} \psi(x) W_\mu(x)$$ Saying that parity is violated means that the transformed lagrangian $\mathcal{L}'(x)$ is not equal to the old lagrangian resulting from new coordinates $\mathcal{L}(x')$ where $x'^0 = x^0$ and $\vec{x'} = -\vec{x}$. In order to evaluate $\mathcal{L}'(x)$, one has to know the transformations of the fermion field $\psi(x)$ and the boson field $W_\mu(x)$. One can show that $\psi'(x) = \eta \gamma^0 \psi(x')$ with $\eta = \pm 1$ the intrinsic parity of the field $\psi$ and $W'_0(x) = \eta_W W_0(x')$, $W'_i(x) = -\eta_W W_i(x')$ (usual transformation of a vector field) with $\eta_W = \pm 1$ the intrinsic parity of the vector field. Hence: $$\mathcal{L}'(x) = \bar{\psi'}(x) \gamma^0 \frac{(1-\gamma^5)}{2} \psi'(x) W'_0(x)+\bar{\psi'}(x) \gamma^i \frac{(1-\gamma^5)}{2} \psi'(x) W'_i(x)$$ giving: $$\mathcal{L}'(x) = \eta^2 \eta_W \left( \bar{\psi}(x') \gamma^0 \gamma^0 \frac{(1-\gamma^5)}{2} \gamma^0 \psi(x') W_0(x') - \bar{\psi}(x') \gamma^0 \gamma^i \frac{(1-\gamma^5)}{2} \gamma^0 \psi(x') W_i(x')\right)$$ Knowing the commutation rules of the $\gamma$ matrices, $(\gamma^0)^2 = 1$ and $\eta^2 =1$, it becomes: $$\mathcal{L}'(x) = \eta_W \left( \bar{\psi}(x') \gamma^0 \frac{(1+\gamma^5)}{2}\psi(x') W_0(x') + \bar{\psi}(x') \gamma^i \frac{(1+\gamma^5)}{2}\psi(x') W_i(x')\right) = \eta_W \bar{\psi}(x') \gamma^\mu \frac{(1+\gamma^5)}{2} \psi(x') W_\mu(x')$$ As you see, whatever the value of $\eta_W = \pm 1$, the transformed lagrangian cannot be equal to $\mathcal{L}(x')$: $$\mathcal{L}(x') = \bar{\psi}(x') \gamma^\mu \frac{(1-\gamma^5)}{2} \psi(x') W_\mu(x')$$ justifying the violation of the parity by the weak interaction.

Okay, I think I have an idea why the terminology is used, but I think this argument makes little sense:

The Lagrangian term describing weak interactions is of the form

$$\bar \Psi \gamma_\mu P_L \Psi W^\mu$$

Under parity transformations $\Psi \rightarrow \gamma_0 \Psi$ and $\bar \Psi \rightarrow \bar \Psi \gamma_0$, therefore

$$\bar \Psi \gamma_\mu P_L \Psi W^\mu \rightarrow \bar \Psi \gamma_0 \gamma_\mu P_L \gamma_0 \Psi W^\mu$$

We can transform the parity transformed Lagrangian, by using the explicit form of $P_L = \frac{1-\gamma_5}{2}$ and $\{\gamma_5, \gamma_\mu \}=0$ :

$$\bar \Psi \gamma_0 \gamma_\mu P_L \gamma_0 \Psi W^\mu = \bar \Psi \gamma_0 \gamma_\mu \gamma_0 P_R \Psi W^\mu = \bar \Psi \gamma_\mu P_R \Psi W^\mu$$

which shows that the weak interaction term in the Lagrangian is not invariant under parity transformations. I think this argument is wrong!

The above discussion misses that under parity transformations $\gamma_5 \rightarrow - \gamma_5$. If we take this in account the corresponding term in the Lagrangian is invariant und weak interactions are invariant under parity transformations.

We can see this, because for an ordinary Dirac spinor we have the projection operator:

$$P_L \Psi = \begin{pmatrix}1 & 0\\ 0 &0 \end{pmatrix} \begin{pmatrix} \chi_L \\ \xi_R\end{pmatrix} = \begin{pmatrix} \chi_L \\ 0 \end{pmatrix} = \Psi_L$$

and for the parity transformed Dirac spinor $\Psi^P$ the left-chiral projection operator is

$$P_L^P \Psi^P = \begin{pmatrix}0 & 0\\ 0 &1\end{pmatrix} \begin{pmatrix} \xi_R \\ \chi_L\end{pmatrix} = \begin{pmatrix} 0\\ \chi_L \end{pmatrix} = \Psi_L^P$$

The parity transformed Dirac spinors live in a different representation and therefore we need different projection operators. The discussion above shows that under parity transformation

$$P_L \rightarrow P_L^P = \frac{1+\gamma_5}{2} = P_R$$ and therefore the parity transformed in the Lagrangian reads:

$$\bar \Psi \gamma_0 \gamma_\mu P_R \gamma_0 \Psi W^\mu = \bar \Psi \gamma_0 \gamma_\mu \gamma_0 P_L \Psi W^\mu = \bar \Psi \gamma_\mu P_L \Psi W^\mu$$

which shows the invariance.

• To anyone who downvoted, a short comment about whats wrong would help me a lot!
– jak
Dec 15 '14 at 8:22