Let $X$ and $Y$ be Banach spaces and $A:X\to Y$ a linear operator. I found the following statement on Wikipedia, https://en.wikipedia.org/wiki/Unbounded_operator#Closed_linear_operators which confuses me.

If $A$ is closed (i.e its graph $\Gamma(A)=\{(x,Ax)):x\in A\}$ is closed) and injective, then its inverse $A^{-1}$ is also closed.

Why is this true? For $A$ to have an inverse we need $A$ to be surjective as well? Since $A$ is closed it is bounded by the closed graph theorem. If $A$ would be bijective, then by the inverse mapping theorem, the inverse would be bounded and linear and so closed.


Note that this is a statement about unbounded operators. Those are not necessarily defined everywhere. The statement, as a matter of fact, is completely trival, since the graph of $A$ is mapped to the graph of $A^{-1}$ under the homeomorphism between $X\oplus Y$ and $Y\oplus X$ which flips the coordinates.

Note also that if $Dom(A)\subsetneq X$, then $A$ being closed need not imply $A$ being bounded. Consider for instance $X=Y=L^2([a,b]),$ $A=\frac{d}{dx}$ and $Dom(A)=H^1([a,b])$. This operator is closed due to completeness of Sobolev Spaces. It is, of course, not very injective.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.