An example of Serre shows that in the strong form of his modularity conjecture, one can't also fix the nebentypus. Serre and Carayol independently explained that this obstruction is due
to nontrivial isotropy groups on certain modular orbifolds, hence only occurs for the primes 2 and 3 and certain Galois representations that we'll call badly dihedral.
Curiously, when studying the deformation theory of a mod p modular Galois representation for an odd prime p, the same badly dihedral representations for p = 3 arise: it is exactly for
these that the minimal deformation ring does not appear to be a flat local complete intersections over the ring of Witt vectors.
We explain this link via a derived version of a minimal R = T theorem. As a corollary, we can characterize when these badly dihedral representations admit minimal lifts. This is joint
work in progress with Preston Wake.
