No, in principle is exactly when it's a lot less trustworthy than the statically-typed one. Because in principle you could execute some code (untested; data-driven; "is_testing?function(x){return true;}:1.5"; etc.) that passes a non-function instead of a function, and this code would be considered valid (until it goes pop at runtime). But a language that checked this property statically would give an error before execution even started.
> But a language that checked this property statically would give an error before execution even started.
Like C or C++ does.
But you are right - I shouldn't use 'in principle'. It was rather a decorative idiom here, because it is true only for simple functions like map or filter or whatever you take from a standard prelude.
What I mean is that we are not talking about run-time in this context. Yes, one could pass by reference (or even by value) some crap at run-time, but this is a quite different problem. Machine code has this problem too.
