In the book Gauge/Gravity Duality: Foundations and Applications by M. Ammon and J. Erdmenger they derive the Schwinger-Dyson equation by considering the generating functional $J[Z]$ and a change of variables $\phi(x) \rightarrow \widetilde{\phi}(x)= \phi(x)+\delta \phi(x)$, with $\delta\phi(x)$ being an arbitrary infinitesimal shift. This leads to the identity $$0=\delta Z[J]=i\int D\phi e^{i(S+\int d^dJ(x)\phi(x)}\int d^dx \left(\frac{\delta S}{\delta\phi(x)}+J(x) \right)\delta \phi(x) $$ The next step is then to take functional derivatives with respect to $J(x_i)$ and setting $J$ to zero. The book notes the result in eq. 1.234 as $$0=i\left\langle \frac{\delta S}{\delta\phi(x)}\phi(x_1)...\phi(x_n)\right\rangle +\sum_{j=1}^n\left\langle \phi(x_1)...\phi(x_{j-1})\delta(x-x_j)\phi(x_{j+1})...\phi(x_n) \right\rangle$$ However when I try calculate the functional derivatives I find (and I now go to the specific case of $n=2$) $$0=\frac{\delta^2}{\delta J(x_1) \delta J(x_2)} \left[i\int D\phi e^{i(S+\int d^dJ(x)\phi(x)}\int d^dx \left(\frac{\delta S}{\delta\phi(x)}+J(x) \right)\delta \phi(x) \right]_{J=0}$$

$$=\frac{\delta}{\delta J(x_1)} \left[i\int D\phi e^{i(S+\int d^dJ(x)\phi(x)}\int d^d\delta(x-x_2)\phi(x)\int d^dx \left(\frac{\delta S}{\delta\phi(x)}+J(x) \right)\delta \phi(x)\\+i\int D\phi e^{i(S+\int d^dJ(x)\phi(x)}\int d^dx \delta^d(x-x_2)\delta \phi(x) \right]_{J=0}$$

$$=\left[i\int D\phi e^{i(S+\int d^dJ(x)\phi(x)}\left( \left(\frac{\delta S}{\delta \phi(x)}+J(x)\right)\delta\phi(x)\phi(x_1)\phi(x_2)+\delta\phi(x_1)\phi(x_2)+\phi(x_1)\delta\phi(x_2) \right)\right]_{J=0}$$

$$=i\int d^dx\left\langle \frac{\delta S}{\delta\phi(x)}\delta\phi(x)\phi(x_1)\phi(x_2)\right\rangle +\left\langle \phi(x_1)\delta\phi(x_2)\right\rangle+\left\langle \delta\phi(x_1)\phi(x_2)\right\rangle$$ Comparing this result with the book I see that there are some unwanted $\delta\phi(x_i)$'s as well as an integral that is also not in the equation in the book. I am almost certain that this result is true, but there is a missing final step that I am not realizing. I have tried taking the functional derivative $\frac{\delta}{\delta(\delta\phi(y))}$, which would at the very least give a dirac-delta function in all the terms as well as getting rid of all the shifted fields $\delta\phi(x_i)$. $$0=\frac{\delta}{\delta(\delta\phi(y))}\left[i\int d^dx\left\langle \frac{\delta S}{\delta\phi(x)}\delta\phi(x)\phi(x_1)\phi(x_2)\right\rangle +\left\langle \phi(x_1)\delta\phi(x_2)\right\rangle+\left\langle \delta\phi(x_1)\phi(x_2)\right\rangle\right]$$ $$=i\int d^dx\left\langle \frac{\delta S}{\delta\phi(x)}\delta^d(y-x)\phi(x_1)\phi(x_2)\right\rangle +\left\langle \phi(x_1)\delta(y-x_2)\right\rangle+\left\langle \delta(y-x_1)\phi(x_2)\right\rangle$$

Howvever its still doesn't completely work as I am now missing a field $\phi(x_i)$ in the second and third term. Where am I going wrong here?


The first equation in the question holds for arbitrary infinitesimal $\delta\phi(x)$, so it implies $$ 0 = \int D\phi\ e^{i\left(S+\int J\phi\right)} \left(\frac{\delta S}{\delta\phi(x)}+J(x)\right). \tag{1} $$ If you calculate functional derivatives of this with respect to $J$, then everything should work out.

By the way, equation (1) is just the fundamental theorem of calculus: $$ 0 = \int D\phi\ \frac{\delta}{\delta\phi(x)} e^{i\left(S+\int J\phi\right)}. \tag{2} $$


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.