ON SOME DEFINITIONAL PROBLEMS WITH THE METHOD OF CERTAINTY EQUIVALENTS