Inner Product Spaces

A bare complex vector space tells you how to add vectors and scale them, but it gives you no notion of length or angle. To measure overlap between quantum states — and ultimately to compute probabilities — we equip the space with an inner product.

The inner-product axioms

An inner product on a complex vector space $V$ is a map that takes two vectors $|u\rangle, |v\rangle$ to a complex number $\langle u | v \rangle$ , subject to three axioms. We use the physics convention, which is conjugate-linear in the first slot and linear in the second.

1. Conjugate symmetry.

\langle u | v \rangle = \overline{\langle v | u \rangle}.

Swapping the arguments conjugates the result. A consequence is that $\langle v | v \rangle$ equals its own conjugate, so it is always real.

2. Linearity in the second argument.

\langle u \,|\, \alpha v + \beta w \rangle = \alpha \langle u | v \rangle + \beta \langle u | w \rangle, \qquad \alpha, \beta \in \mathbb{C}.

Combined with conjugate symmetry, this forces conjugate-linearity in the first argument: $\langle \alpha u + \beta w \,|\, v \rangle = \overline{\alpha}\langle u | v \rangle + \overline{\beta}\langle w | v \rangle$ .

3. Positive-definiteness.

\langle v | v \rangle \ge 0, \qquad \text{with } \langle v | v \rangle = 0 \iff |v\rangle = |0\rangle.

This is the axiom that makes "length" meaningful, and it is the reason the first slot is conjugated: in $\mathbb{C}^n$ we get $\langle v | v \rangle = \sum_k \overline{v_k} v_k = \sum_k |v_k|^2 \ge 0$ .

The induced norm

From the inner product we define the norm (length) of a vector by

\| v \| = \sqrt{\langle v | v \rangle}.

Positive-definiteness guarantees the quantity under the root is real and non-negative, so the norm is a well-defined real number. A vector with $\|v\| = 1$ is normalized — exactly the condition we impose on physical quantum states so that the Born-rule probabilities sum to 1.

The Cauchy–Schwarz inequality

Every inner product satisfies the fundamental bound

|\langle u | v \rangle| \le \|u\|\,\|v\|,

with equality if and only if $|u\rangle$ and $|v\rangle$ are linearly dependent (parallel). This single inequality underlies the triangle inequality $\|u + v\| \le \|u\| + \|v\|$ , the definition of the angle between vectors, and the bound $|\langle \phi | \psi \rangle|^2 \le 1$ for normalized states that caps transition probabilities at 1.

A worked example

Let $|v\rangle = (2 + i,\; 1 - 2i)$ in $\mathbb{C}^2$ . Its squared norm is

\langle v | v \rangle = \overline{(2+i)}(2+i) + \overline{(1-2i)}(1-2i) = |2+i|^2 + |1-2i|^2.

Computing each modulus-squared: $|2+i|^2 = 2^2 + 1^2 = 5$ and $|1-2i|^2 = 1^2 + (-2)^2 = 5$ . Therefore

\langle v | v \rangle = 5 + 5 = 10, \qquad \|v\| = \sqrt{10}.

Note the result is real and positive, as positive-definiteness demands.

Try it

This is a numerical exercise — your code should return a number, not a circuit. For $|v\rangle = (2+i,\; 1-2i)$ , compute the squared norm $\langle v | v \rangle$ .

Run your code to see the quantum state.