Canonical: https://knowlpedia.com/functional-analysis/topological-dual/

# Topological dual

The vector space of continuous linear functionals on a topological vector space.

Let \(E\) be a
[topological vector space](https://knowlpedia.com/functional-analysis/topological-vector-space/) over
\(\mathbb K=\mathbb R\) or \(\mathbb C\). Its **topological dual**,
**continuous dual**, or **continuous linear dual** is
\[
E'=\{\varphi:E\to\mathbb K:\varphi\text{ is a continuous linear map}\}.
\]
With pointwise addition and scalar multiplication, \(E'\) is a [vector space](https://knowlpedia.com/linear-algebra/vector-space/). Evaluation gives the canonical [duality pairing](https://knowlpedia.com/convex-analysis/dual-space-and-duality-pairing/)
\[
\langle \varphi,x\rangle=\varphi(x).
\]
The topological dual is generally smaller than the algebraic dual. Moreover, the notation \(E'\) specifies a set of functionals, not a topology on that set; weak, weak-star, strong, and operator-norm topologies are additional choices determined by the setting.

## Continuity criteria

For a normed space \(E\), [a linear functional is continuous](https://knowlpedia.com/functional-analysis/continuous-linear-map/) exactly when it is bounded on the unit ball, and
\[
\lVert\varphi\rVert=\sup_{\lVert x\rVert\leq1}|\varphi(x)|
\]
defines the dual norm. For a [locally convex space](https://knowlpedia.com/functional-analysis/locally-convex-space/) whose topology is generated by seminorms, continuity means that \(|\varphi(x)|\) is bounded by a constant times the maximum of finitely many generating seminorms.

## Natural dual topologies

The [weak-star topology](https://knowlpedia.com/functional-analysis/weak-star-topology/) \(\sigma(E',E)\) is [pointwise convergence](https://knowlpedia.com/real-analysis/pointwise-convergence/) on \(E\). The [strong dual topology](https://knowlpedia.com/functional-analysis/strong-dual/) is [uniform convergence](https://knowlpedia.com/real-analysis/uniform-convergence/) on bounded subsets of \(E\). When \(E\) is normed, the dual norm topology is uniform convergence on its unit ball. These topologies need not coincide, so statements about convergence or continuity on \(E'\) must name the chosen topology.

## Separation and examples

For a Hausdorff locally convex space, the Hahn–Banach theorem ensures that \(E'\) [separates points](https://knowlpedia.com/real-analysis/separates-points/): if \(x\neq0\), some \(\varphi\in E'\) satisfies \(\varphi(x)\neq0\). For \(1\leq p<\infty\), the familiar \(L^q\)-pairing describes the dual of \(L^p\) under standard measure-theoretic hypotheses. By contrast, the continuous dual can be trivial for non-locally-convex topological vector spaces.

## References

1. Walter Rudin, *Functional Analysis*, 2nd ed., McGraw–Hill, 1991. [WorldCat record](https://search.worldcat.org/title/Functional-analysis/oclc/21163277). Relevant: Chapter 3.
2. François Trèves, *Topological Vector Spaces, Distributions and Kernels*, Academic Press, 1967. [Elsevier book record](https://shop.elsevier.com/books/topological-vector-spaces-distributions-and-kernels/treves/978-1-4831-9859-0). Relevant: Chapter 13.
