?? roughly speaking, "(one...) purpose" of "restricted yoneda embedding" as ... ??"to come close to being an equivalence, in which ideal case you actually recognize the original category as being equivalent to presheaves on the restricted domain" ... ... ???which is sort of saying (when you work it out...) that the prototypical functor along which to "restrict" (...) the yoneda embedding is in fact the yoneda embedding... in that orientation that in fact tautologically works out to the identity functor ... ??? .... ???so ... ??well, maybe this puts in an interesting light the idea that "isbell conjugastion" (...??...) is essentially just the "_other_ orientation" case ... ?? ...