Measurement: The Curse of the Qubit
A qubit is a private creature. As long as nobody interacts with it, it can sit in superposition — a blend of 0 and 1. But the moment anything tries to record which state it is in, the blend ends. That recording event is a measurement, and it is the curse of the qubit: you gain one classical answer, and you lose the quantum-ness you were trying to use.
Here is the part that surprises people: a measurement does not have to be a scientist peering through an instrument. Any intrusion into the qubit's privacy counts. If the qubit interacts with another atom, bumps into a stray electron, or leaks even a single photon that carries information about its state — that is a measurement too. The universe does not care whether a human was watching.
Before measurement
The qubit holds a blend of possibilities. The amplitudes can still interfere.
After measurement
One outcome is now a fact. The blend is gone, and no further interference is possible.
So measurement is not a passive reading, like checking a thermometer. It is an active disturbance: to look is to change. This is why quantum algorithm designers treat measurement as the final, irreversible step — they arrange everything beforehand with gates and interference, and measure only when the answer has been coaxed into the open. For the full story, see Quantum Measurement and the overview in 10 questions.
Common myth: “Measurement just reveals what was already there.” Not for a qubit. Before measurement there is no hidden fact of the matter — the outcome is genuinely created in the act of measuring, with probabilities set by the amplitudes.
Go deeper — the math & the rigor
The measurement postulate says: when you measure a qubit in a chosen basis, the outcome is probabilistic and the state collapses onto the observed basis state. For a qubit \(|\psi\rangle = \alpha|0\rangle + \beta|1\rangle\) measured in the computational basis, the Born rule gives outcome \(|0\rangle\) with probability \(|\alpha|^2\) and \(|1\rangle\) with probability \(|\beta|^2\). Afterward the state is whichever outcome you saw — \(\alpha\) and \(\beta\) are gone.
Collapse is basis-dependent: measuring in a different basis (say the \(\{|+\rangle, |-\rangle\}\) basis) collapses onto a different pair of states, with different probabilities. The qubit does not "have" an outcome waiting to be found; the basis you choose decides the question being asked, and the state answers it once.
There is a fundamental limit to what readout can extract. However large the \(2^n\) amplitudes of \(n\) qubits are, a measurement of all \(n\) qubits returns at most \(n\) classical bits — one binary outcome per qubit. (This is the qualitative content of the Holevo bound.) That is why algorithms cannot simply compute every answer and print them out: the bottleneck is not computing power, it is the readout itself.
Finally, the environment is always "measuring." When a qubit entangles with surrounding atoms, fields, or photons in an uncontrolled way, information about its state leaks outward — an unobserved measurement by the universe. This is decoherence: the superposition degrades into ordinary classical uncertainty without anyone deliberately looking. Keeping qubits isolated enough to compute before this happens is the central engineering battle of quantum hardware.
Key takeaways
- A measurement is any event that records a qubit's state — a scientist observing, or a stray atom, electron, or photon interacting with it.
- Measuring collapses superposition: one classical outcome becomes fact, and interference is no longer possible.
- Collapse is basis-dependent: the basis you measure in determines which question the qubit answers.
- n qubits yield at most n classical bits of readout (Holevo bound) — readout, not computing power, is the fundamental bottleneck.
- Decoherence is the environment measuring the qubit uncontrollably, leaking information and destroying superposition.
Check your understanding
Q1.A qubit in superposition drifts near a stray electron and they interact briefly, exchanging information about the qubit's state. What happened?
Any event that records information about the qubit's state is a measurement — no human observer is required.
Q2.You have 10 qubits in a rich entangled state. You measure all of them. How many classical bits of information can you extract at most?
The Holevo bound says n qubits yield at most n classical bits of readout, however large the state space is.
Q3.Why is decoherence described as the environment 'measuring' the qubit?
Decoherence is information about the qubit leaking into the environment — an unobserved measurement that degrades superposition into classical uncertainty.
References
The books, papers, and articles this lesson drew on — with a note on what each one was used for.
- Nielsen & Chuang, Quantum Computation and Quantum Information, Ch. 2 (measurement postulate, Born rule, Holevo bound).
Used for: Measurement postulate, basis-dependence, and the readout limit.
Tip: press ← / → to turn pages.