The "cubic" part of the inverse function is defined for y less than 0 not y greater than 3 - see http://web2.0calc.com/questions/math-problem_17#r4
Look at the graph of the original function. It is defined for all x and all y, and has a one-to-one mapping x onto y and a one-to-one mapping y onto x, so the inverse function does also.
.
You have interpreted the inverse function by just interchanging x and y. I've interpreted it as as a mapping back from the original f(x) onto the original x, which means, for me, the "x" in f-1(x) is on the original "y" axis, and f-1(x) is on the original "x" axis.
.
Is this confusing or what?!!!!!
.