Do you have a reference to have a wider context?

Basically, he take a subset U of A which is invariant under H (I guess, not stated) and claim it must contains a, b and a’. (I don’t understand what the adjective « vrai » adds here. And it seems that U could also contain none of the elements a,b,a’. If b does not belong to A, this must be the case)

« Renfermer » means « contains », even though we mostly use (now?) « contient » for sets. (We mostly use « renfermer » for treasure chest, or secret documents.) This is the first weird thing from my point of view.