Postulates of Quantum Mechanics

Attempt these from memory before reading.

• State the five postulates.
• Why must the state vector be complex-valued? What observable effect would be lost with real amplitudes alone?
• Write the Born rule. What is the probability that the state is left in $|a_n\rangle$ after measuring $\hat{A}$?
• State the trace rule for expectation values.
• What exactly is the tension the measurement problem names?

Check your answers against the text. Prompt 2 is the one that catches people: the complex phase is not decoration, it is the only thing interference needs.

Classical mechanics is a theory: it tells you what happens and why. Quantum mechanics is a set of rules. The postulates say what a state is, how observables act, what a measurement returns, how states evolve, and how composite systems combine — and then stop.

That looks like a failure of taste, and for most of the twentieth century it was treated as one: the postulates were presented as a recipe to be applied without asking why any of it is true. But the rules are not arbitrary, and three of them are forced by experiment rather than chosen:

• Linearity is not postulated; it is what an experiment with a slit and a probability amplitude implies. Add a second slit and the amplitudes add.
• Hermiticity is not postulated either. An observable must return a real number, and the only operators guaranteed to have a real spectrum are the self-adjoint ones.
• The Born rule is the one postulate with no derivation from the others. It is the “measure what you compute” rule, and it is what makes the whole scheme predictive.

What the postulates do not do is explain collapse. That is not a missing detail; it is the one genuine conceptual crisis in the subject, and section 2.2 is devoted to it.

Postulate 1 (State Space). The state of a quantum system is completely described by a normalised Vector $|\psi\rangle$ in a complex Hilbert space $\mathcal{H}$.

Physical motivation. Superposition is observed in interference experiments (e.g., double-slit), where a particle can traverse multiple paths simultaneously. The complex-valued nature of the state is essential: relative phases between superposition components produce observable interference patterns that cannot be replicated with real amplitudes alone.

Postulate 2 (Observables). Every measurable quantity (observable) is represented by a Hermitian (self-adjoint) operator $\hat{A} = \hat{A}^\dagger$ acting on $\mathcal{H}$.

Physical motivation. Hermitian operators have real eigenvalues, matching the fact that measurement outcomes are real numbers. The spectral theorem guarantees a complete set of eigenstates, providing a basis for expansion.

Postulate 3 (Measurement). A measurement of observable $\hat{A}$ yields one of the eigenvalues $a_n$ of $\hat{A}$. The probability of measuring $a_n$ when the system is in state $|\psi\rangle$ is

Where $|a_n\rangle$ is the eigenstate corresponding to $a_n$. After measurement, the state collapses to $|a_n\rangle$.

Physical motivation. The Born rule $P = |\langle a_n|\psi\rangle|^2$ was postulated by Born (1926) to connect wave functions to observable probabilities. It correctly predicts the intensity distribution in diffraction experiments and the statistics of particle detections.

Postulate 4 (Time Evolution). The time evolution of the state is governed by the time-dependent Schrodinger equation:

Where $\hat{H}$ is the Hamiltonian (energy operator).

Physical motivation. This is the quantum analogue of Hamilton’s equations in classical mechanics. The Schrodinger equation is linear, guaranteeing superposition is preserved. Energy conservation is built in: for a time-independent Hamiltonian, $\langle H \rangle$ is constant.

Postulate 5 (Composite Systems). The state space of a composite system is the tensor product of the state spaces of the components.

Physical motivation. This postulate produces entangled states, which have been confirmed experimentally (Bell inequality violations, quantum teleportation). The tensor product structure ensures that measurements on subsystems can exhibit correlations stronger than any classical theory permits.

The measurement postulate (Postulate 3) introduces a fundamental tension: the Schrodinger equation describes deterministic, unitary evolution, but measurement produces probabilistic, non-unitary Collapse. This is the measurement problem.

The conflict. Consider a system in a superposition $|\psi\rangle = \alpha|a_1\rangle + \beta|a_2\rangle$. Under unitary evolution, the state remains a superposition. But a measurement of $\hat{A}$ yields either $a_1$ or $a_2$ with probabilities $|\alpha|^2$ and $|\beta|^2$, and the state collapses to the corresponding eigenstate. No unitary operator can map a superposition to a single eigenstate with the correct probabilities.

Major interpretational approaches:

• Copenhagen interpretation. Collapse is a fundamental, irreducible process. The classical measuring apparatus triggers the collapse. No further mechanism is specified.

• Many-worlds interpretation (Everett, 1957). The universal wave function never collapses. Instead, measurement causes the observer and system to entangle, branching into multiple non-interacting “worlds,” each corresponding to one measurement outcome.

• Decoherence (Zurek). Interaction with the environment rapidly suppresses off-diagonal elements of the reduced density matrix in a preferred basis (“einselection”), explaining the emergence of classical behaviour from unitary quantum mechanics.

• Bohmian mechanics. Particles have definite positions guided by the wave function via the “pilot wave.” The wave function never collapses, but the effective description reproduces the Born rule.

The measurement problem remains an active area of research in the foundations of quantum mechanics.

For systems where the state is not known precisely (statistical mixtures), the density operator Provides a more general description than the state vector.

Definition. For a pure state $|\psi\rangle$, the density operator is $\hat{\rho} = |\psi\rangle\langle\psi|$. For a statistical mixture of states $|\psi_i\rangle$ with probabilities $p_i$:

Properties:

• $\mathrm{Tr}(\hat{\rho}) = 1$ (normalisation)
• $\hat{\rho}^\dagger = \hat{\rho}$ (Hermitian)
• $\hat{\rho}^2 = \hat{\rho}$ if and only if the state is pure; $\hat{\rho}^2 \lt \hat{\rho}$ for mixed states
• Expectation values: $\langle A \rangle = \mathrm{Tr}(\hat{\rho}\hat{A})$

Time evolution: $i\hbar\,d\hat{\rho}/dt = [\hat{H}, \hat{\rho}]$ (Liouville-von Neumann equation).

The density matrix is essential for describing subsystems of entangled states (reduced density matrices via partial trace), open quantum systems, and decoherence.

• Superposition: A system can be in a linear combination of eigenstates: $|\psi\rangle = \sum_n c_n |a_n\rangle$.
• Uncertainty Principle: non-commuting observables cannot be simultaneously measured with arbitrary precision.
• Probabilistic Nature: Quantum mechanics predicts probabilities, not deterministic outcomes.
• No-cloning theorem. There is no unitary operation that copies an arbitrary unknown quantum state $|\psi\rangle$. This follows from the linearity of quantum mechanics and has profound implications for quantum information.

• Electronic Band Structure: Applies the quantum postulates to electrons in periodic potentials, yielding Bloch states and energy bands.

• Vectors and Vector Spaces

• Hilbert Spaces

• Angular Momentum and the Hydrogen Atom

• Thermodynamics and Statistical Mechanics

The postulates of quantum mechanics are like the rules of a game that nature plays at the smallest scales. The state vector is your scoreboard: it contains everything knowable about the system. Hermitian operators are the measuring instruments, and their real eigenvalues are the only possible outcomes you can read. The Born rule tells you that the square of the amplitude gives the probability, like measuring how much of a wave is present at each point. Time evolution is smooth and deterministic, like a river flowing according to a fixed current. The measurement problem is the deep mystery: the smooth flow of the Schrodinger equation suddenly jerks when you look, collapsing the wave function to a single outcome. This tension between smooth evolution and sudden collapse is still debated. Decoherence explains why we do not see quantum weirdness in everyday life: the environment constantly measures the system, washing out quantum superpositions.

Mistake 1: Confusing probability amplitude with probability The probability of measuring eigenvalue $a_n$ is $|\langle a_n|\psi\rangle|^2$, not $\langle a_n|\psi\rangle$. The amplitude $\langle a_n|\psi\rangle$ is a complex number, and probabilities must be real and non-negative. Students often forget to take the modulus squared, leading to complex or negative “probabilities.”

Mistake 2: Assuming measurement always yields the expectation value The expectation value $\langle \hat{A} \rangle = \sum_n a_n P(a_n)$ is the average over many measurements, not the result of a single measurement. A single measurement of $\hat{A}$ yields one of the eigenvalues $a_n$, not the expectation value. The expectation value is what you would get on average if you repeated the experiment many times.

Mistake 3: Forgetting that measurement collapses the state After measuring $\hat{A}$ and obtaining $a_n$, the state collapses to $|a_n\rangle$, not to some other state. This collapse is irreversible in the sense that information about the original superposition is lost. Subsequent measurements of $\hat{A}$ will yield $a_n$ with certainty until the system evolves or is disturbed.

The first two rows are the ones that matter most. The projector row in particular kills a very common confusion: $\lvert\psi\rangle$ is a state, not something you can measure, and no experiment distinguishes one normalised vector from a scalar multiple of it.

The last row is the right way to think about interpretation. Many-worlds, Bohmian mechanics and collapse all agree on every experimental prediction and disagree only about what is real. The postulates are the predictions; the interpretation is extra structure layered on top.

• Measurement-device-independent (MDI) rules. Instead of assuming the measurement operators, derive them from observed statistics. Several currently used DI protocols have since been attacked, and the surviving challenge is closing every device-independent loophole.
• Landauer’s principle at the nanoscale. The bound is uncontroversial in the reversible limit; whether it survives when the memory is unreliable or the erasure is many-body is still argued.
• Quantum error correction and thresholds. The threshold theorem says below a critical error rate, arbitrarily long computation is possible. The frontier is pushing the threshold up to make the requirement physically attainable, and understanding decoherence in the codes that do the work.
• Contextuality as a resource. Kochen-Specker contextuality was a obstruction to classical models and has become the resource that powers certain advantage-based computation and measurement tasks.
• The foundations of collapse models. Objective collapse is testable in principle; the open question is whether any parameter range survives experimental bounds.

Mix these with Linear Algebra, Angular Momentum and the Hydrogen Atom, Symplectic Structure and KAM Theory and Hamiltonian Mechanics.

1. (Linear algebra.) Prove that a self-adjoint operator on a finite-dimensional Hilbert space has an orthonormal eigenbasis, and identify exactly which step fails for an unbounded self-adjoint operator. That failure is why the spectral theorem is a separate theorem.
2. (Hamiltonian mechanics.) The Schrodinger equation is $i\hbar\,d|\psi\rangle/dt = \hat{H}|\psi\rangle$ with $\hat{H}$ the quantisation of the classical Hamiltonian. Show that quantising $p^2/2m$ gives $-\hbar^2\partial_x^2/2m$, and explain why the replacement $p \mapsto -i\hbar\partial_x$ is forced by the commutator structure of translations rather than chosen.
3. (Symplectic structure.) Write the state vector as a point of projective Hilbert space and identify the natural symplectic form there. Then say what the Hamiltonian flow on that space has in common with classical Hamiltonian mechanics, and where the analogy stops.
4. (Angular momentum.) Show that $\lvert L^2\rvert = \hbar^2 l(l+1)$ and not $\hbar^2 l^2$, using only the postulates and the commutation relations. The extra factor is a pure consequence of quantisation.
5. (Density matrices.) Show that $\hat{\rho} = \lvert\psi\rangle\langle\psi\rvert$ satisfies $\mathrm{Tr}(\hat{\rho}^2) = 1$ for any normalised state, and that for a mixture $\mathrm{Tr}(\hat{\rho}^2) = \sum_i p_i^2 \le 1$. Purity is therefore $\mathrm{Tr}(\hat{\rho}^2) = 1$, with equality only when the mixture is trivial.
6. (Research.) Kochen-Specker contextuality rules out noncontextual hidden variable models. State the result precisely enough to see why it requires dimension at least three — the two-dimensional case has an explicit noncontextual model, and the difference is instructive.

• Angular Momentum and the Hydrogen Atom
• Approximation Methods
• Density Functional Theory: Conceptual Overview
• Historical Motivation