Definition

Let π:GU(H)\pi:G\to U(H) be a on a nonzero . It is irreducible if its only are {0}\{0\} and HH. Closedness is essential in infinite-dimensional Hilbert spaces: the definition is topological, not purely algebraic. Because every π(g)\pi(g) is unitary and π(g)1=π(g1)\pi(g)^{-1}=\pi(g^{-1}), a closed invariant subspace is automatically reducing, so its is invariant as well.

Commutant characterization

The representation is irreducible if and only if every bounded operator TT satisfying Tπ(g)=π(g)TT\pi(g)=\pi(g)T for all gGg\in G is a scalar multiple of the identity. Equivalently, the of π(G)\pi(G) is CI\mathbb C I. This is the unitary form of . A nontrivial invariant subspace yields a nonscalar commuting , while the spectral projections of a nonscalar self-adjoint operator in the commutant recover a nontrivial invariant subspace Folland, §3.1.

Examples and reducible cases

Every one-dimensional unitary representation is irreducible. The trivial representation on C\mathbb C is the simplest example. A nontrivial orthogonal direct sum π1π2\pi_1\oplus\pi_2 is reducible because each summand is a proper closed invariant subspace. An cannot have two orthogonal nonzero invariant summands.

Conventions and scope
References
  1. Gerald B. Folland, A Course in Abstract Harmonic Analysis, 2nd ed., CRC Press, 2016. DOI record. Relevant: §3.1 on invariant subspaces, irreducibility, and Schur's lemma.