4
$\begingroup$

Im currently trying to understand the form of the 4 point function in CFT, i.e. how to derive equation 4.62 in Di Francesco et al. In particular, the coefficients of the $x_{ij}=|x_i-x_j|$. For four scalars $O_i(x_i)$ with dimensions $\Delta_i$, I know the general form should be:

$$<O_1(x_1)O_2(x_2)O_3(x_3)O_4(x_4)>=\frac{f(u,v)}{x_{12}^a x_{13}^b x_{14}^c x_{23}^d x_{24}^e x_{34}^f}$$

with $u$ and $v$ the crossing ratios, based on rotation and translation invariance. Scale invariance gives $$a+b+c+d+e+f=\Delta=\Sigma_{i=1}^4 \Delta_i,$$ then special conformal invariance gives me 4 more equations similar to those in 4.59 of Di Francesco et al for the 3 point function, leaving me with 5 equations on 6 unknowns that i can't solve uniquely. Am i missing something?

$\endgroup$

1 Answer 1

4
$\begingroup$

The problem is that there is too much freedom in your parametrization of the 4-point function:

If we take the cross ratios to be (different conventions give rise to similar arguments) $$ u = \frac{x_{12} x_{34}}{x_{14} x_{23}}, \qquad v = \frac{x_{13} x_{24}}{x_{14} x_{23}}, $$ then you can rewrite $$\frac{f(u,v)}{x_{12}^a x_{13}^b x_{14}^c x_{23}^d x_{24}^e x_{34}^f} = \frac{f'(u,v)}{x_{12}^{a-f} x_{13}^{b-e} x_{14}^{c+e+f} x_{23}^{d+e+f}}$$ where $$ f'(u,v) = u^{-f} v^{-e} f(u,v) $$ Now you can apply the logic that you were mentioning in the question: special conformal invariance gives you 3 equations that fix completely the coefficients $(a-f)$, $(b-e)$, $(c+e+f)$ and $(d+e+f)$, and you find that the 4-point function is fixed up to an unknown function of the two cross-ratios $u$ and $v$.

$\endgroup$
2
  • $\begingroup$ Sorry to reactivate this question 5 years later, but I have similar issues. Reparametrizing your numerator allows you to solve the system given by SCT, should we continue with all possible reparametrization for example $f'(u,v)=u^av^bf(u,v)$? I am confused because for the moment I don't have the correct form for the exponent, that should be $F(u,v) x_{ij}^{(\Delta_1+\Delta_2 + \Delta_3+\Delta_4)/3 -\Delta_i - \Delta_j}$. This is not what we obtain by solving the system. $\endgroup$ Commented Oct 15 at 9:32
  • $\begingroup$ $F(u,v) \prod_{i \neq j} x_{ij}^{(\Delta_1 + \Delta_2 + \Delta_3 + \Delta_4)/3 - \Delta_i - \Delta_j}$ is indeed one possible parametrization of the 4-point function, and it is a nice one because it is symmetric under the exchange of any two operators. But what is the question then? $\endgroup$
    – M.Jo
    Commented Oct 17 at 7:40

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.