Canonical: https://knowlpedia.com/lie-groups/completely-reducible-representation-lie/

# Completely reducible Lie representation

A representation that splits as a direct sum of irreducible subrepresentations.

Let \(\rho\) be a finite-dimensional representation, either of a [Lie algebra](https://knowlpedia.com/lie-groups/lie-algebra/) \(\mathfrak{g}\) (a [Lie algebra representation](https://knowlpedia.com/lie-groups/representation-of-a-lie-algebra/)) or of a [Lie group](https://knowlpedia.com/fiber-bundles/lie-group/) \(G\) (a [Lie group representation](https://knowlpedia.com/lie-groups/representation-of-a-lie-group/)).

**Definition.** \(\rho\) is **completely reducible** if for every [invariant subspace](https://knowlpedia.com/linear-algebra/operator-invariant-subspace/) \(W\subset V\) (a [subrepresentation](https://knowlpedia.com/lie-groups/subrepresentation-lie-algebra/)), there exists an invariant complement \(W'\subset V\) such that
\[
V = W \oplus W'
\]
as representations.

Equivalently, \(V\) can be written as a finite direct sum of [irreducible](https://knowlpedia.com/lie-groups/irreducible-representation-lie-algebra/) subrepresentations.

**Context and key theorems.**
- If \(\mathfrak{g}\) is [semisimple](https://knowlpedia.com/lie-groups/semisimple-lie-algebra/) over \(\mathbb{C}\) (or \(\mathbb{R}\) with suitable hypotheses), then every finite-dimensional representation is completely reducible; this is [Weyl’s complete reducibility theorem](https://knowlpedia.com/lie-groups/weyls-theorem-complete-reducibility/).
- If \(G\) is [compact](https://knowlpedia.com/lie-groups/compact-lie-group/), then every finite-dimensional continuous representation is completely reducible, via averaging an [inner product](https://knowlpedia.com/linear-algebra/inner-product/) (a Lie-group analogue of Maschke’s theorem), and more globally by [Peter–Weyl](https://knowlpedia.com/lie-groups/peter-weyl-theorem/).

Complete reducibility is the structural reason [highest-weight](https://knowlpedia.com/lie-groups/highest-weight/) classifications work for compact and semisimple settings.
