How sufficient is this proof for the inverse of an inverse being the identity relation?

43 Views Asked by At

I was studying mathematical analysis with the book Introduction to Mathematical Analysis (Parzynski and Zipse, 1987) and had a question regarding an exercise problem. It's actually the same one as this question but I was curious if my approach was also valid. The exercise is:

Show that if $f: X \rightarrow Y$ is one-to-one then $(f^{-1})^{-1} = f$.

The approach that I took is:

$$f: X \rightarrow Y\\f^{-1}: Y \rightarrow X \\ (f^{-1})^{-1} : X \rightarrow Y \\ \therefore (f^{-1})^{-1} = f$$

It seems a bit...insufficient and inconvincing.

The approach laid out in the question I linked above is:

Since we know that if $f$ is one-to-one, $f \circ f^{-1} = i_X$ and $f^{-1} \circ f = i_X$.

If we say that we have some function $g$ where $g \circ f^{-1} = i_X$, then $g = f$ since inverses are unique. Since the inverse of $f^{-1}$ is $g$, we can conclude that $(f^{-1})^{-1} = f$.

I was wondering if my approach also works, or if there should be a bit more rigor to it. Thanks.