This question does not follow from reading any text, but I was watching Shiraz Minwalla's CFT lectures on YouTube. At 51:36 into the lecture, he raises the question that, as an operator statement, we know $$\partial_\mu j^\mu = 0,$$ where $j^\mu$ is a conserved current of the theory. However, we know that the Ward Identities are

\begin{equation} \langle \partial_\mu j^\mu(x) O_1(x_1)\cdots O_n(x_n)\rangle + \sum_i \langle \cdots O_i(x_i)\delta(x-x_i)\cdots\rangle = 0, \end{equation}

where $\langle \cdots \rangle$ denote a path integral average. Now he says that this is consistent with the fact that $\partial_\mu j^\mu = 0$ as an operator statement simply because path integral calculates time ordered correlators. He continues to say that "having $\partial_\mu j^\mu$ inside a path integral really means \begin{equation} \partial_\mu\langle j^\mu O_1(x_1)\cdots O_n(x_n)\rangle \end{equation} i.e. computing the average at two infinitesimally close insertions and then taking the difference, further dividing by the infinitesimal distance between the insertions. Now what I do not understand is how having $\partial_\mu j^\mu$ 'inside' the path integral actually means the above, since if I just pull out the $\partial_\mu$ from the Ward Identities as well, I get something completely different. Is this really just some abuse of notation and I am unnecessarily confused? I hope I was able to explain my question.

  • $\begingroup$ You need to remember that the correlators calculated by the path integral correspond to vacuum expectations of time ordered products in the operator language. $\endgroup$
    – mike stone
    Commented Oct 21, 2023 at 16:53

1 Answer 1


The main point is that each correlator $\langle\cdots\rangle$ implicitly contains a covariant time-ordering $T_{\rm cov}$, i.e. time-differentiations inside its argument should be taken after/outside the usual time ordering $T$ in the correlator. This is necessary in order to correctly translate between the operator formulation and the path integral formulation, cf. e.g. this related Phys.SE post.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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