Pointing at a direction without naming it
Suppose I hand you an arrow in Hilbert space and ask you to describe it precisely. "It points that way" is no good over the phone. The trick everyone uses, in ordinary space and in quantum mechanics alike, is to agree on a few reference directions first — east, north, up — and then describe any arrow by how far it reaches along each. "Three steps east, two north" pins a point down exactly. That agreed set of reference directions is called a basis, and learning to use one is the practical skill at the heart of this whole rung.
In quantum language, a basis is a set of special states — call them |1⟩, |2⟩, |3⟩, and so on — chosen as the reference directions. Any state |ψ⟩ at all can then be written as a weighted sum of them: |ψ⟩ = c₁|1⟩ + c₂|2⟩ + c₃|3⟩ + ... You will recognize the right-hand side at once: it is a superposition. So "writing a state in a basis" and "expressing a state as a superposition of basis states" are the very same act.
What makes a good basis: orthonormal
Not every set of reference directions is equally convenient. The nicest choice is an orthonormal basis, a forbidding word for two simple demands. *Ortho*-: the basis states are mutually at right angles, so their overlaps are zero — ⟨i|j⟩ = 0 whenever i and j differ. *Normal*: each basis state has length one — ⟨i|i⟩ = 1. Together: the building blocks are all the same size and all perfectly distinct, like the clean east-north-up axes rather than a set of lopsided, overlapping directions.
The payoff of orthonormality is a small miracle of convenience: each coefficient cᵢ in the expansion is just an overlap you can read off directly, cᵢ = ⟨i|ψ⟩. To find "how much |1⟩" is in your state, you simply take the bracket ⟨1|ψ⟩ — no solving, no fuss. This is the quantum echo of a trick you may know from ordinary vectors: to get a component, project onto the matching axis.
|psi> = c1|1> + c2|2> + c3|3> + ... each coefficient: c_i = <i|psi> (just an overlap!) orthonormal means: <i|j> = 0 if i != j , <i|i> = 1
Complete: nothing left out
A basis also has to be complete — it must offer enough directions that *every* possible state can be reached by combining them, with nothing left over that the basis cannot describe. East and north alone are not complete in three-dimensional space; you would miss everything up and down. A complete set leaves no gaps. Completeness is what guarantees that the expansion |ψ⟩ = Σ cᵢ|i⟩ truly works for any state at all, not just lucky ones.
Physicists wrap up "orthonormal and complete" in a single elegant statement called the completeness relation. You will meet its precise form later; the plain-language version is just "these building blocks are enough, and they overlap cleanly." Whenever you see a sum over a basis suddenly inserted into the middle of an expression — a very common move — that is the completeness relation quietly at work, saying "I am allowed to slip in a full set of building blocks here, because they add up to nothing-changed."
Why measurement loves bases
Here is the part that turns bookkeeping into physics. When you measure something, the possible outcomes correspond to a particular basis — the natural "axes" for that measurement. Measure an electron's spin up-or-down, and the relevant basis is just {|up⟩, |down⟩}. Now expand the state in that basis, |ψ⟩ = c↑|up⟩ + c↓|down⟩, and the Born rule says the probability of each outcome is the size-squared of its coefficient: |c↑|² for up, |c↓|² for down. The coefficients you found by projecting are, after squaring, the very odds the experiment will show.
One last liberating idea: the choice of basis is *yours*. The same physical state can be written in the up/down basis or, just as legitimately, in a left/right basis — different reference axes, different coefficients, but the very same underlying arrow. A measurement simply asks, "which basis are you about to interrogate me in?" Changing your question changes the numbers you predict, even though the state has not moved. Keeping the arrow fixed while swapping the axes is one of the most powerful habits in all of quantum mechanics.