[whatshisface]

If there's a method for selecting the right one for the situation, it's all one self-consistent theory

[simonh]

Would you recommend using Newton’s laws to model the behaviour of colliding toy trains at a few feet per second, or Relativity?

[whatshisface]

Newton's laws are a part of relativity, in the way a map of California is a part of a map of the US.

[__MatrixMan__]

I don't see why the two should be connected. Consistency and utility are somewhat orthogonal.

Here's a counterexample:

Einstein chose hyperbolic geometry for arguments about space ships traveling near light speed, but we use spherical geometry for arguments about the shortest path an airplane should take. Those theories disagree about the playfair postulate, so they're inconsistent. Yet we can pretty reliably pick the right one for the job.