Consider the 4-point correlation function, $G(x_1,x_2,x_3,x_4)$, in $\phi^4$ theory. Let us consider the term which is $\propto \lambda^2$, represented by the following Feynman's diagram:

enter image description here

According to symmetry rules, I should get a $1/2$ symmetry factor, due to the middle lines. Hence, I would suppose that

$$G_2(x_1,x_2,x_3,x_4) = -\frac{\lambda}{2}\int\int d^4x_5 d^4x_6\Delta_F(x_1-x_5)\Delta_F(x_2-x_5)\Delta_F(x_4-x_6)\Delta_F(x_3-x_6)\left(\Delta_F(x_5-x_6)\right)^2$$

where $x_5$ and $x_6$ denote de vertices. However, if I carry out the calculations by hand, I should have something like $\frac{1}{2\times 24}\times 4 \times 3 \times 2 \times 6 \times 2 = 6 \neq \frac{1}{2}$. What am I doing wrong here?


1 Answer 1


The factor is $\frac{1}{2!} \times \frac{1}{4!} \times \frac{1}{4!} \times 8 \times 3 \times 4 \times 3 \times 2 = \frac{1}{2}$.

The $\frac{1}{2!}$ comes from expanding $e^x$ to second order. The two factors of $\frac{1}{4!}$ come from the interaction term (which is $\frac{\lambda}{4!}$). The remaining factors come from counting contractions so lets look at that. The fields to be contracted are $$ \phi_1 \phi_2 \phi_3 \phi_4 \, \, \phi_x\phi_x\phi_x\phi_x \, \, \phi_y\phi_y\phi_y\phi_y $$ First $\phi_1$ can contract with 8 different fields. Everything is completely symmetrical, so lets assume it contracts with $\phi_x$. $\phi_2$ must then contract with $\phi_x$ also which happens in 3 different ways. $\phi_3$ contracts with $\phi_y$ in 4 different ways and $\phi_4$ contracts with another $\phi_y$ in 3 different ways. Finally, the remaining un-contracted fields are $\phi_x \phi_x \,\, \phi_y \phi_y$. The first $\phi_x$ can contract with a $\phi_y$ in 2 different ways and the final contraction is then fixed. Multiplying all the numbers in this paragraph, we find that the number of contractions is $8\times3\times4\times3\times2$.

  • $\begingroup$ Ahhhh! This makes much more sense than what I was thinking. Thank you so much. $\endgroup$
    – miniplanck
    May 2, 2021 at 21:17

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.