lazyness

hello!

i try to get my head around the implications of if a language is lazy
or not.

therefore i constructed a strange "sandwich" function in haskell:

let f x y g = (x+y, g(fst (f x y g)))

f 10 20 (^2)
(30, 900)

in the definition of f x y g the same "f x y g" is "called" again and
still it is solvable for GHCi.
i think this is quite remarkable.

at first only the "fst output" of that function f is evaluated, then a
function g is run over that "first output" and after that the "second
output" of that function is evaluated.

is it true that no strict language will be able to reduce such a
function?

thanks,
sebastian