Inner Approximation Schemes, A-Proper Operators, and the Galerkin Method