We provide methods for inference on a finite dimensional parameter of interest, theta in Re^{d_theta}, in a semiparametric probability model when an infinite dimensional nuisance parameter, g, is present. We depart from the semiparametric literature in that we do not require that the pair...