I have some preliminary ideas on how to describe the cartesian closed subcategories of Set. Is there some well-known characterization I might be rediscovering? (I am interested in subcategories that are not "full," so, for example, Hom(1,a) might be smaller than the set a.) John Mitchell