Erratum: Holographic scattering and non-minimal RT surfaces

From our paper, one might expect that the u < d condition on extremal surfaces is necessary and sufficient for holographic scattering in all pure states under consideration.In this erratum, we would like to emphasize a small caveat to this statement. 1 Consider a defect geometry with |M | > 1/4.Due to the effect discussed in figure 25, there exist configurations of boundary regions such that the u candidate does not exist.In such cases, it does not make sense to say that the u < d relation controls holographic scattering.(Indeed, one can check in examples that the absence of u is consistent with both scattering and not-scattering.)In Version 2 of our paper on the arXiv, we have added explanations of this point in footnote 3 on page 8, at the bottom of page 10, and in footnote 9 on page 21.