A controlled gate is usually described as acting on its target. When the target is in a state the gate only multiplies by a phase, the gate changes nothing about the target, and the phase lands on the control instead. That is *phase kickback*.

The state |−⟩ = (|0⟩ − |1⟩)/√2 is such a state for `x`: X|−⟩ = −|−⟩. So `cx q[0], q[1]` with `q[1]` in |−⟩ multiplies the part of the state where `q[0]` is |1⟩ by −1, and leaves `q[1]` in |−⟩.

```qasm
OPENQASM 3.0;
include "stdgates.inc";
qubit[1] q;
x q[0];
h q[0];
```

This prepares |−⟩. Phase kickback is how the algorithms in the next unit read a function's answer: the answer becomes a phase on the input qubits, and interference turns the phases into an outcome. The debugger shows it step by step: `wai-quantum debug` prints the amplitudes before any line, and the Studio's **Step through** pane draws each qubit's state.