If we consider the space of integrable functions $L^1([0,1];\mathbb{R})$, it can be ordered by the convex cone of positive integrable functions $L^1([0,1];\mathbb{R}_+)$. It is known that the corresponding dual cone is $L^\infty ([0,1];\mathbb{R}_+)$. I wonder if the same kind of statement holds when replacing $\mathbb{R}$ by an arbitrary Banach space $Y$?

More formally :

Let $L^1([0,1];Y)$ be the Banach space of Bochner integrable functions from $[0,1]$ to $Y$ (identifying the functions a.e. equal on $[0,1]$). Its topological dual space is known to be $L^\infty_{w^*}([0,1];Y^*)$, the space of $w^*$-measurable functions from $[0,1]$ to $Y^*$ (with also an identification there).

Suppose now that $Y$ is ordered by a closed convex cone, with non-empty interior, that we note $Y_+$. We note $Y_+^*$ the correspoding dual cone, defined by :

$$Y_+^* := \{ y^* \in Y^* \ | \ \langle y^*,y\rangle \geq 0 \ \forall y \in Y_+ \}.$$

Can we prove that the dual cone of $L^1([0,1];Y_+)$ is $L_{w^*}^\infty([0,1];Y^*_+)$ ?

By just applying the definitions, we directly see that $L_{w^*}^\infty([0,1];Y^*_+)$ is included in the dual cone of $L^1([0,1];Y_+)$. But the reverse inclusion seems tricky...

  • $\begingroup$ Why is the reverse inclusion tricky? Take a separating hyperplane, that leads to a contradiction. $\endgroup$ Commented Mar 31, 2016 at 7:47

1 Answer 1


Take some element in the dual cone $g\in L^1([0,1],Y_+)^*\subset L^\infty_{w^*}([0,1],Y^*)$. Then, by definition, for every $f\in L^1([0,1],Y_+)$

$$\int_0^1 \langle f(t),g(t)\rangle\mathbb{d}t\geq0$$

In particular, for any $y\in Y_+$ and $[a,b]\subset [0,1]$, we can take $f \equiv y\cdot \chi_{[a,b]}$ so that

$$\int_a^b \langle y,g(t)\rangle \mathbb{d}t\geq 0.$$

Hence, for every $y\in Y_+$, and every $t\in[0,1]$ we have $\langle y,g(t) \rangle\geq 0$. Thus the image of $g$ is in the positive cone which means that $$g\in L^\infty([0,1],Y_+^*).$$

  • 1
    $\begingroup$ Nice! A minor detail (which does not invalidate the proof): in the last part the inequality holds for almost every t (not every t). $\endgroup$
    – Guillaume
    Commented Nov 20, 2018 at 11:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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