Consider the Banach space $C_c(\mathbb{R})$ of continuous functions with compact support equipped with the uniform norm $||f||_\infty := \sup_{x \in \mathbb{R}} |f(x)|$. Then it is known (Riesz representation theorem) that the continuous dual of $C_c$ can be identified the vector space of all finite signed Radon measures on $\mathbb{R}$, that is set-functions $\mu : \mathscr{A} \to \mathbb{R}$ (that are regular) and $\mathscr{A}$ is the Borel $\sigma$-algebra on $\mathbb{R}$ (or a $\sigma$-algebra that contains the Borel sets).

Now, if we equip $C_c = \bigcup_{K \text{ compact}} C_c(K)$ with the inductive limit topology where $C_c(K)$ is the space of continuous functions $f \in C_c$ (i.e. $f$ has domain $\mathbb{R}$) with $\text{supp}(f) \subseteq K$ normed by $||f||_K := \sup_{x \in \mathbb{R}} |f(x)| = \sup_{x \in K} |f(x)|$, then $C_c$ is a locally convex space. The dual space is larger than in the case above and consists of all Radon "signed measures" (e.g. the Lebesgue measure $dx$ which is clearly not finite on $\mathbb{R}$ or $\sin(x) dx$ which is not a signed measure on a $\sigma$-algebra). I want to identify this dual space with an object coming from abstract measure theory. I think, the correct object is also a signed measure which is also countably additive BUT is defined on merely the $\delta$-ring generated by bounded intervals in $\mathbb{R}$ instead of a $\sigma$-algebra (the $\delta$-ring does not contain unbounded sets like the whole space).

Is this correct intuition? And if yes, does anyone know a reference for this fact?

(Of course one can generalize the ground space to be a more general topological space then $\mathbb{R}$, e.g. a locally compact second-countable Hausdorff space.)

EDIT: $C_c$ was primarily equipped with the wrong topology (the compact convergence). It is now correctly equipped with the inductive limit topology from its subspaces $C_c(K)$.


Edited following the edited question.

Yes, now the dual are all measures on the $\delta$-ring of bounded Borel sets. In your case they can be also considered as $\sigma$-finite measures on $\mathbb{R}$, since it is a countable union of compacts. In general, it's measures on the $\delta$-ring of Borel subsets of compact sets. In the topology of inductive limit, functionals are continuous when they are continuous on each $C_c(K)$, that is - measures on each $K$, from where it all follows.

  • $\begingroup$ Oh, thank you for the hint. I have equipped $C_c$ with the wrong topology ("compact convergence"). It must be the inductive limit topology from its subspaces $C_c(K)$ with compact support in $K$. I have edited the question. $\endgroup$ – yadaddy Jan 21 '15 at 9:02

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.