Can Poincaré gauge theory be saved?

Recent investigations on the field equations of Poincaré gauge theory seem to indicate that the only subclass of theories well founded on a reasonable translation limit are actually ruled out by the lack of a well-defined initial-value problem, and, in particular, by tachyonic torsion solutions. We show, however, that these nonphysical torsion fields may be transformed into mere Lagrange multipliers by symmetries that necessarily come along with any 3+1 decomposition of space-time. We even invert the argument by pointing out that the incriminated theories are the only ones that allow for the nonphysical fields to be trivialized in that way, and the canonical variables to be reduced to a set with a well-defined initial-value problem.