This is a question about a step in the proof of the inverse function theorem.
Say we have a function $f\colon V\to W$, that is $C^1$ and where $V,W\subset R^n$. The derivative of $f$ is assumed to be invertible.
My book says the following: let $y,b\in W$, with $y=f(x)$ and $b=f(a)$. Then $$ f^{-1}(y)-f^{-1}(b)=x-a=f’(a)^{-1}(y-b)+f’(a)^{-1}o(\Vert x-a\Vert). $$ I don't understand how they got the second equality. I know that $$ f(x)-f(a)=f’(a)(x-a)+o(\Vert x-a\Vert), $$ and it seems that they kind of applied this, but it doesn’t seem exactly the same to me. First of all, why do they multiply $o(\Vert x-a\Vert)$ by $f’(a)^{-1}$? And why don't they just write something like $$ f^{-1}(y)-f^{-1}(b)=(f(b)^{-1})'(y-b)+o(\Vert y-b\Vert). $$ Maybe it's equivalent. It seems we need to show that $f'(a)^{-1}=(f(b)^{-1})'$ and $f'(a)^{-1}o(\Vert x-a\Vert)=o(\Vert y-b\Vert)$. I know the second is true. So I'm only having problems with $$ f'(a)^{-1}=(f(b)^{-1})'. $$ The inverse of the derivative is the derivative of the inverse?
EDIT
The last question can be answered by noting that $f^{-1}f(x)=x$, so $f^{-1}(f(x))'f'(x)=I$, which implies that $(f'(x))^{-1}=f^{-1}(y)'$. So that basically answers everything.