Hilbert space
/ HIL-bert /
Every quantum state is a vector, and Hilbert space is the arena those vectors live in. If you know how ordinary arrows add and how to take their dot product, you already have the picture — Hilbert space is that idea generalized to allow complex components, infinitely many dimensions, and a notion of length so that limits behave. It is the stage on which the whole play of quantum mechanics is performed.
Formally, a Hilbert space is a complex vector space equipped with an inner product <phi|psi>, which assigns a complex number to each pair of vectors, together with the property of completeness: every Cauchy sequence of vectors converges to a vector still inside the space. The inner product gives a norm ||psi|| = sqrt(<psi|psi>), lets you say when two states are orthogonal (<phi|psi> = 0), and underlies probability amplitudes. For a single particle in one dimension the relevant space is L^2, the square-integrable functions psi(x) with integral |psi|^2 dx finite; for a spin-1/2 it is just the two-dimensional space C^2.
Why the abstraction earns its keep: superposition is simply vector addition, measurement is projection onto a subspace, observables are operators acting on the space, and completeness guarantees that the limits and infinite sums you write down (Fourier expansions, spectral decompositions) actually land back in the space. Dirac's bra-ket notation is the working language of these vectors, and it lets you compute without ever committing to a particular coordinate representation.
The polarization states of a single photon form a two-dimensional Hilbert space C^2 spanned by |horizontal> and |vertical>; a diagonal photon is the superposition (|H> + |V>)/sqrt(2).
Even the simplest quantum system is a vector space with an inner product; superposition is addition.
Completeness (Cauchy sequences converging inside the space) is the technical heart; it is what lets infinite superpositions and spectral expansions be trusted rather than merely formal.