so what about specializing the idea of "generalized legendrian submanifold" to the case where there are no non-trivial invariant distributions (or something... ??relationship to "hermitian symmetric space" and so forth... ??something about grassmanian vs partial flag variety... ??...) ?? so maybe arbitrary submanifolds qualify as generalized legendrian in such a context?? seems like a case where "geometric quantization" develops a somewhat "classical" flavor... ??or something...