Skip to content
Tutorials
Mathematics
What Do the Pauli Matrices Mean?

What Do the Pauli Matrices Mean?

February 6, 2025

The Pauli matrices show up everywhere in quantum computing, but they are often introduced as “here are three matrices, now use them.” This page explains what they mean for a single qubit.

The three Pauli matrices

σx=(0110),σy=(0−ii0),σz=(100−1). \sigma_x= \begin{pmatrix} 0&1\\ 1&0 \end{pmatrix},\quad \sigma_y= \begin{pmatrix} 0&-i\\ i&0 \end{pmatrix},\quad \sigma_z= \begin{pmatrix} 1&0\\ 0&-1 \end{pmatrix}.

You can think of them as the “X, Y, Z axes operators” for a qubit.

A qubit is a 2D quantum system

A qubit state ∣ψ⟩\lvert \psi\rangle lives in a 2D complex vector space. A spinor is just a coordinate representation of that abstract vector in some chosen basis (commonly the σz\sigma_z eigenbasis).

Nothing in the column (ab)\begin{pmatrix}a\\b\end{pmatrix} means “spin value in the x direction” directly. It just means:

  • the amplitude to get one measurement outcome in a chosen basis
  • and the amplitude to get the other outcome

Pauli matrices as observables (measurements)

In quantum mechanics, measurable quantities are represented by Hermitian operators. The Pauli matrices are Hermitian, so they can be treated as observables.

Each Pauli matrix has eigenvalues ±1\pm 1. For example:

  • measuring σz\sigma_z yields either +1+1 or −1-1
  • the corresponding eigenstates are ∣0⟩\lvert 0\rangle and ∣1⟩\lvert 1\rangle (in the usual convention)

More explicitly:

σz∣0⟩=+∣0⟩,σz∣1⟩=−∣1⟩. \sigma_z\lvert 0\rangle = +\lvert 0\rangle,\quad \sigma_z\lvert 1\rangle = -\lvert 1\rangle.

Similarly, σx\sigma_x has eigenstates ∣+⟩\lvert +\rangle and ∣−⟩\lvert -\rangle, where:

∣±⟩=12(∣0⟩±∣1⟩),σx∣±⟩=±∣±⟩. \lvert \pm \rangle=\frac{1}{\sqrt{2}}(\lvert 0\rangle\pm \lvert 1\rangle), \quad \sigma_x\lvert \pm\rangle=\pm\lvert \pm\rangle.

This gives a concrete rule for probabilities:

If your state is ∣ψ⟩\lvert \psi\rangle, then the probability of getting +1+1 when measuring σx\sigma_x is ∣⟨+∣ψ⟩∣2\lvert\langle + \vert \psi\rangle\rvert^2, and for −1-1 it is ∣⟨−∣ψ⟩∣2\lvert\langle - \vert \psi\rangle\rvert^2.

This is the general measurement recipe: probabilities are squared overlaps with eigenstates.

Pauli matrices as rotations (gates)

In quantum computing, Pauli matrices also act as gates:

  • XX flips ∣0⟩↔∣1⟩\lvert 0\rangle \leftrightarrow \lvert 1\rangle
  • ZZ applies a relative phase ∣1⟩↦−∣1⟩\lvert 1\rangle \mapsto -\lvert 1\rangle

On the Bloch sphere, these correspond to 180° rotations about the X, Y, Z axes.

If you have read the Bloch sphere page, this is why Pauli matrices feel “geometric”: they are the simplest nontrivial rotations of a qubit.

Why do they come in a triple?

There are three because a single qubit has three independent real degrees of freedom once you factor out normalization and global phase. The Pauli matrices form a natural basis for qubit operators:

Any 2×22\times 2 Hermitian matrix can be written as a real linear combination of:

I, σx, σy, σz. I,\ \sigma_x,\ \sigma_y,\ \sigma_z.

That’s why Hamiltonians, noise models, and measurement operators are often expressed in the Pauli basis.

Commutation (why measurement axes conflict)

The Pauli matrices do not commute:

[σx,σy]=2iσz(and cyclic permutations). [\sigma_x,\sigma_y]=2i\sigma_z \quad \text{(and cyclic permutations)}.

Non-commutation is the algebraic way to say:

  • measuring along one axis and then along another can change the state,
  • you cannot assign consistent pre-existing values to all three components at once.

Quick example: “spin up in X” then measure Z

If the qubit is prepared in ∣+⟩\lvert +\rangle (the +1+1 eigenstate of σx\sigma_x), then measuring σz\sigma_z gives:

P(0)=∣⟨0∣+⟩∣2=12,P(1)=∣⟨1∣+⟩∣2=12. P(0)=\left|\langle 0 \vert + \rangle\right|^2=\frac{1}{2},\quad P(1)=\left|\langle 1 \vert + \rangle\right|^2=\frac{1}{2}.

So “definite along X” implies “random along Z.”

Next

Single-Qubit Gates (X, Y, Z, H, S, T, and Rotations)What Is a Hadamard Gate?