FALSE: The EPRL amplitude is supported on flat connections

TL;DR

The authors correct a previous claim that the EPRL amplitude is supported on flat connections, clarifying an error and withdrawing the preprint.

Contribution

They identify and correct an error in their earlier work regarding the EPRL amplitude's support on flat connections.

Findings

Original claim about EPRL amplitude supported on flat connections is false.

The preprint with the original claim will be withdrawn from arXiv.

The correction clarifies the physical implications of the EPRL model.

Abstract

There is an error in our recent preprint ``The EPRL amplitude is supported on flat connections''. The error is in Section 3. Here we leave the original text unchanged, but add a note in Section 3 pointing out exactly what the error is. We apologize for the false alarm. After one month, this preprint will be withdrawn from the arxiv. {\bf Original Abstract:} For the version of the EPRL model based on the original vertex amplitude and the face amplitude selected by its gluing properties, we prove that the EPRL amplitude of any region with the topology of a 4-ball is supported on flat connections. We state immediate consequences of this result, comment on some applications, and discuss physical implications. The results hold in general; they do not rely on a semiclassical analysis.