Canonical: https://knowlpedia.com/functional-analysis/frechet-space/

# Fréchet space

A Hausdorff locally convex space whose topology is induced by a complete metric.

A **Fréchet space** is a Hausdorff [locally convex topological vector space](https://knowlpedia.com/functional-analysis/locally-convex-space/) whose topology is induced by a translation-invariant metric with respect to which the space is [complete](https://knowlpedia.com/convex-analysis/complete-metric-space-complete-subset/). Equivalently, it is a complete, metrizable, locally convex space. Completeness belongs to the compatible uniform structure and is independent of which compatible translation-invariant metric is chosen. A Fréchet space need not be normable: one countable family of seminorms may be required to describe its topology, and no single norm may generate the same open sets.

## Seminorm description

The topology of a Fréchet space can be generated by a countable separating family of seminorms \((p_n)_{n\geq1}\). One compatible translation-invariant metric is
\[
d(x,y)=\sum_{n=1}^{\infty}2^{-n}
\frac{p_n(x-y)}{1+p_n(x-y)}.
\]
Conversely, a Hausdorff locally convex space is metrizable exactly when its topology is generated by a countable family of seminorms; it is Fréchet when it is complete for the resulting uniform structure.

## Examples

Every [Banach space](https://knowlpedia.com/linear-algebra/banach-space/) is Fréchet. The space \(C^\infty(M)\) on a compact [smooth manifold](https://knowlpedia.com/fiber-bundles/smooth-manifold/) is Fréchet for seminorms controlling derivatives of increasing order. The [Schwartz space](https://knowlpedia.com/functional-analysis/schwartz-space/) \(\mathcal S(\mathbb R^n)\) is another basic example. These spaces are usually not normable because convergence must simultaneously control infinitely many derivatives or weighted derivatives.

## Mapping theorems and terminology

The Baire property of [complete metric spaces](https://knowlpedia.com/topology/complete-metric-space/) gives Fréchet spaces versions of the open mapping, closed graph, and [uniform boundedness theorems](https://knowlpedia.com/functional-analysis/uniform-boundedness-principle/). These results make the category much more rigid than arbitrary locally convex spaces.

**Terminology warning.** Some authors use **F-space** for any complete metrizable [topological vector space](https://knowlpedia.com/functional-analysis/topological-vector-space/), without local convexity. Under that convention, every Fréchet space is an F-space but not conversely.

## References

1. Walter Rudin, *Functional Analysis*, 2nd ed., McGraw–Hill, 1991. [WorldCat record](https://search.worldcat.org/title/Functional-analysis/oclc/21163277). Relevant: Chapter 1.
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 9.
