I was going through Witten's paper on AdS and holography , and am confused in section 2.4. He starts by considering a massless scalar action in Euclidean AdS spacetime, with a boundary value $\phi_0$. He then looks for a "Green function" $K$, but says that it satisfies the following condition:

$$L K = 0,$$

where $L$ is the Laplacian for the scalar field. But in general say when we have a differential equation $$L \phi = f,$$

in order to solve it we construct a Green function $G$ for the differential operator $L$, it has to satisfy

$$L_x G_{xy} = \delta(x-y)$$

with proper boundary conditions and the solutions to $L$ are given by $\phi$ such that $\phi = \int_M G f $. The solution and not the Green function satisfies

$$L \phi =0.$$

Is there a deeper reason behind calling it a Green function? Does this have any relation to the fact that he is working in Euclidean AdS and not Lorentzian AdS?


  • $\begingroup$ Minor comment to the post (v1): In the future please link to abstract pages rather than pdf files. $\endgroup$
    – Qmechanic
    Mar 1, 2020 at 11:39

1 Answer 1


Here $x'$ denotes a point on the boundary, and $x$ denotes a point in the bulk. The problem he is looking at is given some boundary value $\phi_0(x')$ of the scalar field $\phi(x)$, how do you solve the wave equation.

$$L \phi(x) = \phi_0(x')$$

In order to do that he is solving

$$L_x K_{xx'} = \delta(x-x').$$

One way to solve this is to just consider the solutions to the equation $L K = 0$, and see which component of this equation blows up at the boundary point $x'$, which is here given by $x' = \infty$, and show that this blowing up corresponds to a delta function. Here the blowing up is given by

$$K(x_0) = C x_0^d.$$

In equation (2.18) he shows that this blowing up corresponds to a delta function, by implementing an isometry transformation

$$x^i \to \frac{x^i}{x_0^2 + \sum_{i=1}^d x_i^2}$$

which maps $x_0'=\infty$ to the origin. He then shows that the Green function

$$K(x) = C \left( \frac{x^i}{x_0^2 + \sum_{i=1}^d x_i^2} \right)^d$$

satisfies all properties of a delta function

$$L_x K(x-x') = \delta(x-x')$$ in the limit $x_0\to 0, x_i =0$. Therefore the solution of the equation of motion is given by

$$\phi(x_0,x_i) = \int_M \phi_0(x') K(x-x')$$


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.