Adjoint of a linear isomorphic functional is an isomorphism

1.2k Views Asked by At

I have this exercise to solve:

Let $E$ and $F$ be normed spaces and let $T \in L(E,F)$, where $L(E,F)$ denotes the set of all bounded linear operators from $E$ to $F$. $T^*$ is the dual operator of $T$ with $T^*: F^* \to E^*$ and $T^*\varphi=\varphi \circ T$. Assume $T$ is an isomorphism, i.e. $T$ is bijective and $T^{-1}\in L(F,E)$. Prove that $T^*$ is also an isomorphism and $(T^*)^{-1}=(T^{-1})^*$.

My work

  • Proof of $(T^*)^{-1}=(T^{-1})^*$:

We know that $(T \circ G)^*= T^*\circ G^*$ since $((T \circ G)^*\varphi)(x)=\varphi((T\circ G)(x))=(G^*\varphi)(Tx)=((T^*\circ G^*)\varphi)(x)$

Now plug in $T^*$ and $(T^{-1})^*$ to the equation, we have:

$T^*\circ (T^{-1})^* = (T\circ T^{-1})^*= Id^* = Id$ and $(T^{-1})^* \circ T^* = (T^{-1}\circ T)^*= Id^* = Id$

Hence, $(T^*)^{-1}=(T^{-1})^*$.

  • Proof of isomorphism:

    Injectivity:

Let $\varphi$ and $\psi \in F^*$. Let $T^*\varphi=T^*\psi$ then we have $\varphi \circ T=\psi \circ T$. Here can I conclude that since $T$ bijective we have $\varphi = \psi$?

It remains to prove the surjectivity and I'm still struggling with this part.

2

There are 2 best solutions below

1
On

Hong thank you for tagging me in the other question: Let $X$ and $Y$ be Banach spaces, show that if they are isomorphic, then $X$ is reflexive iff $Y$ is reflexive. But the answer is just that $(f^{-1})^*=(f^*)^{-1}$ and $f^*$ linear.

Do you need to see either of these facts spelled out further? The point is that $f^*$ is an isomorphism if and only if $(f^*)^{-1}$ is continuous (or more generally, $g:X\to Y$ is an isomorphism if and only if $g^{-1}$ is continuous). But continuity is equivalent to being bounded. So we just have to show that $(f^{-1})^*$ is bounded. Does that clarify things?

See this for example: http://mathonline.wikidot.com/isomorphism-linear-operators-on-normed-linear-spaces

0
On

Hint for surjectivity:

Let $\psi \in E^*$ be arbitrary. You need to find a functional $\varphi \in F^*$ such that $T^* \varphi = \psi$. That is, such that $\varphi \circ T = \psi$. Can you guess how to define a functional $\varphi$ that might work? You'll have to verify that your guess works, and that it's actually a continuous functional.