I want to show $${\gamma^0}^\dagger=\gamma^0\\ {\gamma^i}^\dagger=-\gamma^i.$$

To do this I consider the Dirac equation $$ (i\gamma^\mu\partial_\mu-m)\psi=0$$

and I write it as

$$ i\partial_t \psi=(-i\gamma^0\gamma^i\partial_i+m\gamma^0)\psi:=H\psi$$

where I defined the Hamiltonian $H$. We require $H=H^\dagger$ i.e. $$-i\gamma^0\gamma^i\partial_i +m\gamma^0=i(\gamma^0\gamma^i)^\dagger\partial_i+m{\gamma^0}^\dagger.$$

The second term makes it clear that ${\gamma^0}^\dagger=\gamma^0$, the first term becomes

$$ -i\gamma^0\gamma^i=i{\gamma^i}^\dagger{\gamma^0}$$


$$\gamma^i\gamma^0={\gamma^i}^\dagger\gamma^0 $$

which means ${\gamma^i}^\dagger=\gamma^i$

Why doesn't this work? In this I'm ignoring the operator $\partial_i$ when taking the adjoint because the adjoint is taken in spinor space, not in $L^2$. Is this incorrect? Should I take $\partial_i^\dagger=-\partial_i$? Why?

  • $\begingroup$ duplicate. $\endgroup$ – Cosmas Zachos Jan 4 at 11:59
  • $\begingroup$ Hi, thanks for the reference. I don't see how the question you linked answers mine. The answer starts with "knowing ${\gamma^\mu}^\dagger=\gamma^0\gamma^\mu\gamma^0$", which is what I want to show. $\endgroup$ – user2723984 Jan 4 at 12:01
  • $\begingroup$ In your first step you seem to use the Dirca algebra, i.e. $\gamma^0 \cdot\gamma^0=\mathbb{1}$. Then the result is more or less obvious. The original drivation assumes some coefficients $\gamma^i$ and derives the algebra from the requirement that the $H^2$ is the Klein-Gordon operator. $\endgroup$ – Toffomat Jan 4 at 12:14
  • $\begingroup$ I derive the Dirac algebra by requiring that $(i\gamma^\mu\partial_\mu +m)(i\gamma^\mu\partial_\mu -m)=\partial^2+m^2$ but this doesn't involve the hermitian conjugate of the gamma matrices anywhere $\endgroup$ – user2723984 Jan 4 at 12:23
  • $\begingroup$ Yes, the momentum is Hermitean, so the derivative anti Hermitean, as in that proof. Follow the signs. $\endgroup$ – Cosmas Zachos Jan 4 at 12:31

Yes. You sould use $\partial_i^\dagger= -\partial_i$ because that is the correct adjoint of the derivative in $L^2[\mathbb R]$.

Recall that the adjoint $A^\dagger$ with respect to an inner product $<\phantom x,\phantom y>$ of an operator $A$ is defined so that $$ <A^\dagger \phi, \chi>= <\phi, A\chi>. $$ When $<\phi,\chi>= \int_{-\infty}^\infty \phi^*\chi\,dx$ an integration by parts gives $$ <\phi, \partial_x \chi>= <- \partial_x \phi, \chi> $$ so $\partial_x^\dagger=-\partial_x$.

  • $\begingroup$ Then I am confused about what we're talking about. The adjoint of $H$ is the operator $H^\dagger$ such that $\langle\psi, H\psi\rangle=\langle H^\dagger \psi, \psi\rangle$, but is this the inner product on $\mathbb{C}^4$, on $L^2$ or something else? $\endgroup$ – user2723984 Jan 4 at 13:38
  • $\begingroup$ @user2723984 I'm not sure what you are confused about, but I added a bit to my answer about how the adjoint is defined. Ah I see your response. The inner product is on ${\mathbb C}^4 \otimes L^2 [\mathbb R^4]$ $\endgroup$ – mike stone Jan 4 at 13:44
  • $\begingroup$ mmh I think I understand, since in this case $H$ contains matrices, aren't we actually dealing with $L^2[\mathbb{C^4}]$ and the inner product $\langle \phi, \chi \rangle = \int dx \phi^\dagger \chi$, with the adjoint in the integral wrt the standard inner product of $\mathbb{C^4}$? $\endgroup$ – user2723984 Jan 4 at 13:46
  • $\begingroup$ @user2723984 Yes. I think we are both being imprcise about the apce, but it is the space of square integrable, complex, four component vector-valued function in ${\mathbb R}^3$, with the inner product you indicate. $\endgroup$ – mike stone Jan 4 at 13:49
  • $\begingroup$ Thank you, my confusion was ultimately about which space and inner products I was dealing with $\endgroup$ – user2723984 Jan 4 at 13:52

One can choose gamma-matrices in such a way that they do not satisfy the hermiticity condition.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.