Ok, here's where the original question came from...maybe this helps.
Say I have a function F(a,b,c) = G(d,e). Assume the Implicit Function Theorem conditions are satisfied. So, I can solve for c as follows: c = H(G(d,e),a,b). Now, in this case, can I write c as c = N(d,e,M(a,b))? Why or why not?