Worldsheet CFT$_2$ and Celestial CFT$_2$ : An AdS$_3$-CFT$_2$ perspective

Celestial CFT$_d$ is the putative dual of quantum gravity in asymptotically flat $(d+2)$ dimensional space time. We argue that a class of Celestial CFT$_d$ can be engineered via AdS$_{d+1}$-CFT$_d$ correspondence. Our argument is based on the observation that if we zoom in near the boundary of (Euclidean) AdS$_{d+1}$ then the conformal isometry group of EAdS$_{d+1}$, which is SO$(d+2,1)$, contracts to the Poincare group ISO$(d+1,1)$. This suggests that the near boundary scaling limit of a theory of \textit{conformal} gravity on EAdS$_{d+1}$ should be dual to a boundary CFT$_d$ with ISO$(d+1,1)$ symmetry. This dual CFT$_d$, since the symmetries match, is an example of a Celestial CFT$_d$. Similarly, if we have a \textit{non-conformal} theory of gravity on EAdS$_{d+1}$ then the near boundary scaling limit of such a theory is dual to a (boundary) Celestial CFT$_d$ with \textit{only} (SO$(d+1,1)$) Lorentz invariance. Celestial CFTs with only Lorentz invariance have been recently studied in the literature. Now following this logic we discuss, among other things, the near boundary scaling limit of the bosonic string theory on Euclidean AdS$_3$ in the presence of the NS-NS B field. The AdS$_3$ part of the worldsheet theory is free in this limit and has been studied in the literature in different contexts. This limit describes a ``long string'' which wraps the (Euclidean) AdS$_3$ boundary and it has been argued that the space-time CFT$_2$ which describes the radial fluctuations of a long string is a Liouville CFT. According to our proposal, the dual CFT$_2$ which describes the \textit{long string sector} is an example of a \textit{Celestial} CFT$_2$ with \textit{only} (SO$(3,1)$)Lorentz invariance. We do not get a full ISO$(3,1)$ invariant Celestial CFT$_2$ in this way because the string theory does not have target space conformal invariance.