>Given a type T, if we can prove some property exists for all of T's minimal substructures, and all of T's immediate substructures, then this property must also hold for T itself.
I'm pretty sure people don't like this explanation because it implicitly assumes a directed acyclic graph and a reduction operation that takes a set of T and produces a single composite T.
I'm pretty sure people don't like this explanation because it implicitly assumes a directed acyclic graph and a reduction operation that takes a set of T and produces a single composite T.