The Mathematics of Quantum Mechanics

The mathematics of quantum mechanics is, at its core, linear algebra dressed up in physics clothing. States live as vectors in a complex Hilbert space; observables are Hermitian operators acting on those vectors; measurement projects the state onto an eigenvector, and the squared amplitude gives the probability. That last move (the Born rule) is the only piece that isn’t pure math. Schrödinger’s equation is just unitary rotation of the state vector through time. Everything else (superposition, entanglement, uncertainty) falls out of vectors that don’t commute. The physics is in the postulates; the rest is bookkeeping.

Saying the physics is in the postulates means the empirical content of quantum mechanics fits on a notecard, while the rest is theorems. Linear algebra provides the machinery for free: Hilbert spaces, Hermitian operators, eigenvalues, tensor products, unitary groups. None of that knows anything about electrons. The physics is in the identification rules. This vector is a spin-up electron. That operator measures position. Probabilities are squared inner products (the Born rule, which no math derivation produces). Once those bridges are nailed down, the formalism runs on its own and predicts double-slit fringes without further input.

Superposition, entanglement, and uncertainty each fall out of a different feature of the formalism, but all three are theorems before they are physical surprises. Superposition is the fact that any sum of state vectors is itself a state vector; Schrödinger cats exist because Hilbert space is closed under addition. Entanglement is the observation that most vectors in a tensor product $\mathcal{H}_A \otimes \mathcal{H}_B$ cannot be written as a single product $|\psi\rangle|\phi\rangle$; correlation is the generic case, separability the exception. Uncertainty is the commutator: when $[A, B] \neq 0$, no vector is an eigenstate of both, so sharp values can’t coexist.

The canonical commutator $[x, p] = i\hbar$ is special because position and momentum are conjugate variables: each generates translations in the other. Momentum is the generator of spatial shifts; position is the generator of momentum shifts. Classically this is the Poisson bracket $\{x, p\} = 1$; quantization promotes Poisson brackets to commutators times $i\hbar$, which is the only constant with units of action available. The general uncertainty relation $\sigma_A \sigma_B \geq |\langle [A, B] \rangle|/2$ then forces $\sigma_x \sigma_p \geq \hbar/2$, and every conjugate pair (energy/time, angle/angular momentum) inherits the same $\hbar$-scale bound. Sending $\hbar \to 0$ recovers classical mechanics.

At the edge of the classical limit, three lenses explain why quantum weirdness vanishes in everyday life. The correspondence principle is the constraint: in the high-quantum-number limit, observable predictions must reduce to classical ones; any new theory has to recover Newton where Newton worked. WKB is the calculation: writing $\psi \sim \exp(iS/\hbar)$ and taking $\hbar \to 0$, the phase oscillates so fast that everything cancels except the path where $\delta S = 0$, and least-action falls out. Decoherence is the physical answer: $\hbar$ stays finite, but entanglement with the environment suppresses off-diagonal density-matrix elements exponentially, so the pointer basis looks classical.