I don't think "certain control flow structures are inexpressible" is the same assertion as "certain programs are unexpressible"... After all, any turing comple

I don't think "certain control flow structures are inexpressible" is the same assertion as "certain programs are unexpressible"... After all, any turing complete system will have a similar set of computable programs as another, but they will not all have the same control structures. For example, you can have a turing complete language with no loops, only recursion. Both can express the same programs but not in the same way. That said, it's also a bit of a tautology so.