| spectral theorem | |
|---|---|
| Name | Spectral theorem |
| Field | Functional analysis; Quantum mechanics |
| Introduced | 19th century |
| Notable figures | John von Neumann; David Hilbert; Erhard Schmidt; Marshall H. Stone |
spectral theorem
The spectral theorem provides a characterization of linear operators on a Hilbert space in terms of their spectra and associated projection operators. In Quantum mechanics it underpins the representation of observables as self-adjoint operators and explains decomposition of states into measurement outcomes. The result unifies diagonalization in finite dimensions with a measure-theoretic functional calculus for unbounded operators in infinite dimensions.
The spectral theorem has multiple equivalent formulations depending on context: for normal operators on finite-dimensional complex Hilbert spaces it reduces to diagonalization by a unitary matrix; for bounded normal operators on separable Hilbert spaces it yields a representation as an integral of the identity function against a projection-valued measure; for unbounded self-adjoint operators (typical for quantum observables) it gives a unique spectral resolution and allows domain considerations. Key named variants include the Finite-dimensional spectral theorem (matrix diagonalization), the bounded operator (or continuous functional calculus) version, and the unbounded self-adjoint version typically attributed to John von Neumann and developed using ideas from David Hilbert and Erhard Schmidt.
In the axioms of quantum theory an observable corresponds to a self-adjoint operator on a Hilbert space of states (e.g., the Schrödinger equation state space). Self-adjointness ensures a real spectrum (possible measurement outcomes) and supports a projection-valued spectral decomposition. Normal operators (operators commuting with their adjoint) generalize this to complex-valued spectral measures; examples include unitary evolution operators like the time-evolution operator U(t)=e^{-iHt/ħ} where H is self-adjoint. Prominent practitioners who formalized these links were John von Neumann and Paul Dirac; Dirac's heuristic bra–ket and delta-normalized eigenvectors motivated the rigor provided by the spectral theorem.
A central object is the projection-valued measure (PVM) E(·) on the Borel subsets of the complex plane or real line that assigns orthogonal projections on the Hilbert space. For a self-adjoint operator A one obtains A = ∫_ℝ λ dE(λ), interpreted via strong operator integrals. This permits a functional calculus f(A)=∫ f(λ)dE(λ) for bounded Borel functions f, enabling construction of exponentials e^{itA} and resolvents (A−zI)^{-1}. The PVM formalism connects to the theory of measure theory and to operator algebras: the spectral measure generates a commutative von Neumann algebra (a W*-algebra) and is central to the Gelfand–Naimark framework for C*-algebras.
In finite dimensions the spectral theorem is the classical result that every normal matrix M ∈ M_n(ℂ) is unitarily diagonalizable: M = U D U^*, with U unitary and D diagonal containing eigenvalues. For Hermitian (self-adjoint) matrices eigenvalues are real and eigenvectors form an orthonormal basis; multiplicity and degeneracy are handled by orthogonal projections onto eigenspaces. This underlies matrix mechanics of Werner Heisenberg and practical computations in atomic and molecular spectroscopy (e.g., computing energy levels in Hartree–Fock approximations).
Infinite-dimensional operators often have continuous spectrum components without normalizable eigenvectors, as in the position and momentum operators of a particle on ℝ. The spectral theorem represents such operators via PVMs that assign projections onto spectral subspaces rather than single eigenvectors. To accommodate Dirac delta "eigenvectors" and generalized eigenfunctions, the rigged Hilbert space (Gelfand triple) formalism extends the Hilbert space by a space of test functions and its dual, reconciling physicists’ distributions with the rigorous spectral decomposition. Examples include the momentum operator and the Hamiltonian of free particles with continuous energy spectrum.
Spectral decompositions encode the Born rule: the probability of measuring a value in a Borel set B for observable A in state |ψ⟩ is ⟨ψ|E(B)|ψ⟩. Projection postulate collapses states onto ranges of E(B) for ideal measurements. The functional calculus yields unitary dynamics e^{-iHt/ħ} used in Schrödinger picture and constructs propagators and Green's functions via the resolvent, linking to scattering theory in Lippmann–Schwinger equation contexts. In quantum information, spectral properties determine entanglement measures via reduced density operators, whose eigenvalues (the von Neumann entropy spectrum) control entropy and resource quantification.
Proofs exploit different frameworks: Hilbert space methods use orthogonal decomposition, approximate eigenvectors, and the Riesz representation theorem; measure-theoretic proofs construct projection-valued measures via resolution of the identity and functional calculus for normal operators. C*-algebraic approaches use the Gelfand representation for commutative C*-algebras to identify a bounded normal operator with multiplication by the coordinate function on a measure space, giving an abstract spectral theorem applicable to elements of a C*-algebra and to representations on Hilbert spaces. Foundational contributors and texts include Marshall H. Stone (Stone's theorem on one-parameter unitary groups), von Neumann's mathematical foundations of quantum mechanics, and standard functional analysis expositions by Reed and Simon.