Observable algebra

The algebra of operators used to represent observables of a quantum system; in finite dimensions typically all linear operators on the Hilbert space.
Observable algebra

Let H\mathcal{H} be the system Hilbert space (see ). An observable algebra is a unital *-algebra A\mathcal{A} of operators acting on H\mathcal{H}, meaning:

  1. A\mathcal{A} is closed under addition and scalar multiplication.
  2. If A,BAA,B\in\mathcal{A} then ABAAB\in\mathcal{A} (closure under multiplication).
  3. If AAA\in\mathcal{A} then AAA^\dagger \in \mathcal{A} (closure under adjoint).
  4. The identity operator II belongs to A\mathcal{A}.

In the standard finite-dimensional setting one takes A=B(H)\mathcal{A} = \mathcal{B}(\mathcal{H}), the algebra of all on H\mathcal{H} (equivalently, all complex matrices in a chosen basis; see ).

An observable is a element AAA \in \mathcal{A}; its spectral projections encode measurement outcomes.

A state on A\mathcal{A} can be defined abstractly as a linear map ω:AC\omega:\mathcal{A}\to\mathbb{C} that is positive (ω(AA)0\omega(A^\dagger A)\ge 0) and normalized (ω(I)=1\omega(I)=1). In finite dimensions, every such state is represented by a ρ\rho via

ω(A)=Tr(ρA), \omega(A)=\operatorname{Tr}(\rho A),

using the .

Physical interpretation

  • A\mathcal{A} is the mathematical container for “what you can measure” and “how measurements combine.”
  • Noncommutativity (ABBAAB\neq BA) encodes incompatible observables and the impossibility of assigning sharp values to all observables simultaneously.
  • Commutative subalgebras behave like classical algebras of random variables, describing compatible families of measurements.

Key properties

  • *-structure and reality: Self-adjoint elements correspond to real-valued measurement outcomes.
  • Positivity: Operators of the form AAA^\dagger A are positive; states must assign them nonnegative expectation values.
  • Expectation values: The operational content of the algebra is summarized by the Aω(A)A\mapsto \omega(A).
  • Subsystem structure: For composite systems, observables typically include tensor-product forms; reduced descriptions use the to produce subsystem states.