Chapter 13: Hilbert Spaces & Quantum Mechanics
1900–1932
David Hilbert & the 23 Problems
David Hilbert (1862–1943) stood at the International Congress of Mathematicians in Paris in 1900 and announced 23 problems that would shape mathematics for the coming century. His closing words captured his philosophy of mathematics:
“Wir müssen wissen, wir werden wissen.” — We must know, we will know.
Among those 23 problems, several drove the development of functional analysis and the very machinery that would later describe quantum mechanics: integral equations (Problem 19), the rigorous foundations of physics (Problem 6), and the axiomatization of mathematics itself.
Hilbert's Göttingen became the world capital of mathematics. The work he did on integral equations from 1904 to 1910 — studying infinite systems of linear equations — led him, almost by accident, to invent the infinite-dimensional inner product space that now bears his name.
Hilbert Spaces: Complete Inner Product Spaces
A Hilbert space \(\mathcal{H}\) is a vector space equipped with an inner product \(\langle \cdot, \cdot \rangle\) that is complete with respect to the induced norm. Completeness means every Cauchy sequence converges: there are no “missing points.”
\( \langle f, g \rangle = \int_{-\infty}^{\infty} f^*(x)\, g(x)\, dx \)
The space \(L^2(\mathbb{R})\) of square-integrable functions is the canonical example: functions \(f\) with \(\int |f|^2 < \infty\), equipped with the inner product above. This is Schrödinger's arena.
An orthonormal basis \(\{e_n\}\) satisfies\(\langle e_m, e_n \rangle = \delta_{mn}\) and spans the whole space: every vector \(\psi\) can be written
\( \psi = \sum_n \langle e_n, \psi \rangle\, e_n \)
This is the Hilbert-space version of Fourier series: any quantum state can be expanded in any complete basis of eigenstates.
The Spectral Theorem
Hilbert's crowning achievement in this framework was the spectral theorem: every self-adjoint (Hermitian) operator \(A = A^\dagger\) on a Hilbert space has a spectral decomposition. In the discrete case:
\( A = \sum_n \lambda_n \,|\lambda_n\rangle\langle \lambda_n| \)
where \(\lambda_n \in \mathbb{R}\) are the eigenvalues and\(|\lambda_n\rangle\) the orthonormal eigenvectors. For operators with continuous spectrum (like the position operator) the sum becomes an integral over a projection-valued measure.
Hilbert proved this in the context of integral equations. Twenty years later, quantum mechanics would reveal that it described physical measurement itself.
Heisenberg's Matrix Mechanics Was Hilbert Space Theory
In June 1925, Werner Heisenberg, recuperating from hay fever on the island of Helgoland, wrote down the rules of quantum mechanics in terms of infinite arrays of numbers. He multiplied them by a rule he found “rather strange” — not commutative, but yielding the right physics. He had reinvented matrix multiplication.
When Heisenberg showed his work to Max Born, Born immediately recognized what Heisenberg had found: operators on an infinite-dimensional space. Born and Jordan (and independently Dirac) formulated matrix mechanics properly: quantum observables are operators, quantum states are vectors, and the fundamental commutation relation is
\( [\hat{x},\, \hat{p}] = \hat{x}\hat{p} - \hat{p}\hat{x} = i\hbar \)
This is not an anomaly of quantum mechanics — it is a theorem of functional analysis. Position and momentum are unbounded self-adjoint operators on \(L^2(\mathbb{R})\), and they cannot both be simultaneously diagonalizable. Heisenberg's “strange multiplication” was operator composition in a Hilbert space.
Schrödinger's Wave Mechanics & \(L^2(\mathbb{R})\)
In January 1926, Erwin Schrödinger, working independently (and motivated by de Broglie's matter waves), wrote down the equation governing quantum wave functions:
\( i\hbar\, \frac{\partial \psi}{\partial t} = \hat{H}\psi = \left(-\frac{\hbar^2}{2m}\nabla^2 + V\right)\psi \)
The wave function \(\psi(x,t)\) lives in \(L^2(\mathbb{R})\): it must be square-integrable so that \(\int |\psi|^2\, dx = 1\) (normalization). The Hamiltonian \(\hat{H}\) is a self-adjoint operator, and its eigenvalues are the allowed energy levels. Schrödinger's equation is the Hilbert-space spectral problem in disguise.
Schrödinger himself showed that his wave mechanics and Heisenberg's matrix mechanics gave identical results — but he could not explain why. The answer, given by von Neumann in 1932, was that both are representations of the same abstract Hilbert-space structure.
The Inner Product & the Born Rule
The inner product on \(L^2(\mathbb{R})\) is not merely an abstract structure — it is physical measurement. Max Born proposed in 1926 that
\( \langle \psi | \phi \rangle = \int \psi^*(x)\,\phi(x)\, dx \)
and \(|\langle \lambda_n | \psi \rangle|^2\) gives the probability of obtaining eigenvalue \(\lambda_n\) when measuring observable \(A\)in state \(|\psi\rangle\). This is the Born rule, one of the postulates of quantum mechanics.
The Bridge: Geometry Is Physics
Quantum observables are self-adjoint operators on a Hilbert space. Measurement outcomes are their eigenvalues — which are real, by the spectral theorem. The probability of each outcome is the squared magnitude of the inner product. Hilbert's pure mathematics of infinite-dimensional geometry turned out to be the exact language of quantum reality.