Bi4Ti3O12 is a well-known and studied layered perovskite; such phases are ferroelectrics, which have been used in random-acccess memories and other applications. Understanding the atomic-scale nature of the mechanism driving ferroelelctricity requires very careful and detailed crystallographic analysis, with powder neutron diffraction being the technique of choice. We undertook the first such analysis of Bi4Ti3O12 in 1999, using HRPD at ISIS. Since then, theoretical understanding in the field has moved on, and while our initial study is not 'incorrect' it is certainly 'incomplete'. We are now in a position to re-examine this material, using methods that have become available in recent years. Several possible distinct mechanisms of ferroelectricity in Bi4Ti3O12 that were not considered fully in the earlier work, will re-addressed and the full mechanism will be confirmed.