[Haskell-beginners] function - argument termination problem

Hello, Is it possible to implement in haskell a function f (A) such that if A does not ever terminate then f always terminates, and if A always terminates then f does not ever terminate? I've been thinking it for a while, with no results unfortunately. the actual problem is that i can think of no way to force both ifs: a function that always terminates independently from its argument could be f (a) = 1, and a function that does not terminate even if its argument terminates could be: f (a) = f (a + 1), but i can't figure out a "hybrid" version... Is there any idea on this? Thanks in advance :-), yannis