I am following Sidney Coleman's lectures of Quantum Field Theory.

At the end of ch.32, he derived the Ward identity for the 1PI generating functional $\Gamma[\psi,\bar{\psi},A_{\mu}]$ for QED:

\begin{equation} ie \bar{\psi} \frac{\delta \Gamma}{\delta \bar{\psi}(x)} - ie \frac{\delta \Gamma}{\delta \psi(x)} \psi(x)- \partial^{\mu} \frac{\delta \Gamma}{\delta A^{\mu}(x)} = \frac{-1}{\xi} (\partial_{\nu}\partial^{\nu})(\partial_{\mu} A^{\mu}). \end{equation}

The term on the RHS is the gauge fixing term in the original QED Lagrangian. I am now wondering that whether all of the fields involved in the Ward identity are $c$-number fields. As $\psi$ and $\bar{\psi}$ represent Fermi field, it seems like we should interpret their classical correspondence as Grassmann fields. However, it is clear on the RHS, we have a $c$-number function. Then this equation seems to have both $c$-number and Grassmann number involved, which I think may not make sense?

I am wondering whether we should interpret both $\psi$ and $\bar{\psi}$ also as $c$-number fields? But if that is the case, how does the Fermi minus sign issue be properly handle under the above Ward identity?


1 Answer 1


No, $\psi$ and $\bar{\psi}$ are $a$-number fields, while $A_{\mu}$ is a $c$-number field. The above Ward-Takahashi identity (WTI) is a supernumber-valued identity.

The WTI is often used by differentiating it a number of times wrt. the fields, and then set the remaining fields to zero.

See also this related Phys.SE post.


  1. Bryce DeWitt, Supermanifolds, Cambridge Univ. Press, 1992.

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.