A function is 1-1 (or injective) if it maps different arguments to different values. For a function to have an inverse, it must be injective: if where , then is not defined because it cannot be both and . Therefore, you need to find a restriction where the function is 1-1, as skeeter said.