I'm afraid I did not quite catch that. If I interpret correctly, the relationship between Vegetable and Vitamin is described as a "containment" relationship, ie the Vitamins do not exist on their own but share the life cycle of their "parent" Vegetable. On the other hand, common sense guides me to the observation that in real-life Vitamins seem to exist independently from Vegetables - eg fish? or meat? So this is not necesarily this strong kind of "containment" relationship but rather a "membership" type of relationship (even though I would say "vegetables contain vitamins"! It is still not a "containment" type of relationship in the data modeling sense) : Vitamins are "members of" Vegetables, a much looser coupling than "Vitamins are contained within Vegetables". So what can be "marked as deleted" (and resurrected) is the membership of a Vitamin within a Vegetable, but not the Vitamin itself. This distinction seems crucial to the further oberservations, because if this is the case, Vitamins can be updated at any time without effecting Vegetables that refer to them, Vegetables that do not refer to them or Vegetables that once did but do no longer refer to them. But now I have the feeling that there is more behind the "resurrection pattern" and I completly missed the point.