I take it as trivially true that they can be separated shallowly. Then the question is how big the informal "leaps of faith" are. If they are big, the next thin

I take it as trivially true that they can be separated shallowly. Then the question is how big the informal "leaps of faith" are. If they are big, the next thing is to look for "strands" of reasoning (lemmas that may not be connected) inside them, as a sort of formal proof reverse-marginalia.