Hermitian Operators & Real Eigenvalues
Also known as: Self-adjoint operators · Observable reality theorem
Measured quantities are always real numbers — you never read '3+2i volts' off a meter. Quantum mechanics guarantees this by demanding every observable be a Hermitian (self-adjoint) operator. Hermiticity forces the eigenvalues (the possible measurement outcomes) to be real and the eigenstates to be orthogonal, so different outcomes are perfectly distinguishable. It's the algebraic condition that makes operators safe to call 'observables.'
A real eigenvalue keeps an eigenstate's phase spinning on the real axis; a non-Hermitian part would make it grow or decay.
Equivalent forms
One symmetry condition,  = †, simultaneously guarantees real outcomes and a clean orthonormal basis to measure in.
Where it holds
Dimensional analysis
Hilbert's spectral theory of self-adjoint operators, sharpened by von Neumann's rigorous 1932 formulation of quantum mechanics in Hilbert space, established that observables must be self-adjoint to have real spectra. It gave the measurement postulate its mathematical backbone.
- Guaranteeing real energies in every quantum simulation
- Building measurement bases in quantum computing
- Diagonalizing Hamiltonians in condensed-matter and chemistry codes
- Hermitian and 'symmetric' differ for unbounded operators; self-adjointness is the precise requirement
- Real eigenvalues do NOT require a real matrix — complex Hermitian matrices also have real eigenvalues
- Orthogonality is automatic only for distinct eigenvalues; degenerate ones must be orthogonalized by hand
What if…
Eigenvalues can be complex, so it cannot represent a directly measured quantity (though non-Hermitian 'PT-symmetric' operators can still have real spectra in special cases).
Their eigenstates span a subspace; you can always choose an orthonormal basis within it, preserving the clean measurement structure.
Pauli-Z eigenvalues
- is Hermitian (equal to its conjugate transpose)
- , real; |↑⟩|↓⟩