Were you using the anisotropic lighting model?
Since Birefringence is often caused by a distortion in an otherwise isotropic material (an extreme bend in a fiberoptic line, for example), The anistropic effect might yield something similar; though I'm not sure that Carrara has the ability to simulate the prismatic dispersion.

Also, I have been able to get birefringent refractions by using: Operaters -> Overlay, in the refraction slot, and using two different indices of refractions. It shows up as a dual refractive image through the transparency in question, but I couldn't sat what effect that would have on the caustic emmissions, if any.

Friends don't let friends use booleans.