Let $X$ be a Banach Lattice and denote by $\mathcal{B}(X)$ the banach space of bounded linear endomorphisms. An operator $T \in \mathcal{B}(X)$ is called positive if $Tx \geq 0$ whenever $x \geq 0.$ We say that a positive operator is a lattice homomorphism if $T(x\vee y) = Tx \vee Ty$ for every $x,y \in X$, where $x\vee y = \sup \{x,y\}.$

Now, consider $T$ a positive operator and assume $T$ is an isomorphism. I wonder if the positiveness of $T^{-1}$ is a suficient condition for $T$ to be a lattice homomorphism (and applying the same condition to $T^{-1}$, a lattice isomorphism). I think this is the most intuitive way of thinking about a lattice homomorphism, but I couldn't find such a result in any Banach Lattices book, so my intuition may be wrong.

Could anyone give me a proof or a counterexample?

Thank you very much


It is true, and quite easy to prove. Abstractly, if $T$ is a positive isomorphism with positive inverse, then $T$ preserves the order structure, so of course it preserves suprema.

To make it more explicit: Since $T$ is positive and $x\vee y\geq x,y$, we have $T(x\vee y)\geq Tx, Ty$. Thus $T(x\vee y)\geq Tx\vee Ty$. With the same reasoning for $T^{-1}$ instead of $T$ and $Tx,Ty$ instead of $x,y$ we obtain $T^{-1}(Tx\vee Ty)\geq x\vee y$. Now apply $T$ to get $Tx\vee Ty\geq T(x\vee y)$. Altogether, we have proven $T(x\vee y)=Tx\vee Ty$.


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.