The Core Ideas in One Page

Quantum mechanics is a framework for assigning probabilities to the outcomes of interventions on physical systems. Its many applications look different, but they repeatedly use the same ten ideas: states, superposition, measurement structures, the Born rule, Hamiltonian dynamics, tensor products, exchange symmetry, entanglement, classical limits, and equivalent formulations.

This page is a compact map. Each section states what an idea does, gives the minimum mathematical form needed to recognize it, and links to the canonical treatment. It is not a substitute for the derivations on those pages.

The Prediction Grammar

A finite-dimensional quantum model can be summarized by three ingredients.

1. A state is a positive trace-one operator $\rho$.
2. A transformation is a physically allowed map $\mathcal E$ on states.
3. A measurement has positive effects $\{E_a\}$ that sum to the identity.

Their basic relations are

$$\begin{aligned} \rho&\geq0, & \operatorname{Tr}\rho&=1, \\ E_a&\geq0, & \sum_a E_a&=I, \\ p(a)&=\operatorname{Tr}(\rho E_a). \end{aligned}$$

For a closed system, the transformation is unitary:

$$\rho' = U\rho U^\dagger.$$

For an open or noisy system, $\mathcal E$ is more generally a completely positive trace-preserving map. This state–transformation–measurement grammar is representation-independent. Wavefunctions, matrices, spinors, and phase-space functions are ways of expressing its ingredients in particular problems.

1. A System Has a Quantum State

A state records all probabilities that a specified model can assign to future measurements, conditional on a preparation. It is not a list of definite values possessed by every observable.

A pure state can be represented by a normalized ket $|\psi\rangle$, but the physical state is the ray

$$\left\lbrace e^{i\chi}|\psi\rangle : \chi\in\mathbb R \right\rbrace .$$

The common phase $e^{i\chi}$ is global and cannot affect any prediction. A density operator is the more general state representation. It includes pure states,

$$\rho_\psi=|\psi\rangle\langle\psi|,$$

as well as statistical mixtures, thermal states, reduced subsystem states, and states degraded by noise.

The state is meaningful only relative to a declared system, Hilbert space, preparation, and time. A wavefunction $\psi(x)$ is not a different kind of state; it is the position-basis representation $\langle x|\psi\rangle$ of a state vector.

Continue with State Vectors, Rays and Global Phase, and Density Operators.

2. States Can Be Superposed

The state-vector space is linear. If $|0\rangle$ and $|1\rangle$ are vectors in the same Hilbert space, then

$$|\psi\rangle = \alpha|0\rangle +\beta e^{i\phi}|1\rangle, \qquad |\alpha|^2+|\beta|^2=1,$$

is another normalized state vector when $\alpha$ and $\beta$ are chosen appropriately.

The crucial physical datum is the relative phase $\phi$, not the fact that the same vector can be expanded in many bases. A measurement in the $\{|0\rangle,|1\rangle\}$ basis sees only $|\alpha|^2$ and $|\beta|^2$. Measurements in other bases can make the two components interfere and reveal $\phi$.

A coherent superposition is not the same state as a classical mixture. For equal weights,

$$\begin{aligned} \rho_{\mathrm{coh}} &= |+\rangle\langle+|, \\ \rho_{\mathrm{mix}} &= \frac12|0\rangle\langle0| +\frac12|1\rangle\langle1|, \end{aligned}$$

where $|+\rangle=(|0\rangle+|1\rangle)/\sqrt2$. Both give equal probabilities in the computational basis, but an $X$-basis measurement distinguishes them.

The canonical distinction is developed in Superposition and Relative Phase and Classical Mixtures vs Quantum Superpositions.

3. Observables Are Measurement Structures

For a sharp real-valued observable, a self-adjoint operator $A$ and its spectral projectors $\{P_a\}$ encode possible outcomes:

$$A=\sum_a aP_a$$

in the discrete case. The eigenvalues $a$ label outcomes; the projectors, not the eigenvalues alone, determine their probabilities.

Not every measurement is sharp or naturally represented by one self-adjoint operator. A positive-operator-valued measure, or POVM, uses effects $\{E_a\}$ with

$$E_a\geq0, \qquad \sum_aE_a=I.$$

This covers noisy detection, nonorthogonal outcomes, coarse graining, and measurements whose classical labels need not be numerical.

An outcome-probability rule does not by itself specify what happens after the outcome. Effects determine probabilities; instruments determine conditional state updates. Keeping those roles separate prevents a common confusion in measurement theory.

See Observables, POVMs: First Encounter, and State Update Rule.

4. Measurement Outcomes Are Probabilistic

The Born rule pairs a state with a measurement event:

$$p(a)=\operatorname{Tr}(\rho E_a).$$

For a pure state and a rank-one projector $E_a=|a\rangle\langle a|$, this becomes

$$p(a|\psi)=|\langle a|\psi\rangle|^2.$$

Quantum probabilities are conditional on both the preparation $\rho$ and the measurement $\{E_a\}$. Saying “the probability of spin up” is incomplete until the spin axis and measurement procedure are fixed.

Probabilistic does not mean every result is uncertain. If $\rho=|a\rangle\langle a|$ and the measurement contains the same projector, then $p(a)=1$. Nor does the Born rule say that an unmeasured observable always carried an unknown classical value. It supplies operational outcome probabilities without that additional assumption.

Expectation values summarize repeated statistics,

$$\langle A\rangle_\rho = \operatorname{Tr}(\rho A),$$

but one expectation value does not determine a full probability distribution. The Born Rule is the canonical probability postulate.

5. Time Evolution Is Generated by Energy

For a closed system with Hamiltonian $H(t)$, a state vector obeys the Schrödinger equation

$$i\hbar\frac{d}{dt}|\psi(t)\rangle = H(t)|\psi(t)\rangle.$$

The corresponding propagator is unitary. If $H$ is time independent,

$$U(t,t_0) = \exp\left[ -\frac{i}{\hbar}H(t-t_0) \right].$$

Unitarity preserves inner products, normalization, and distinguishability. For an autonomous closed model, the Hamiltonian is both the energy observable and the generator of time translations. With explicit driving or changing frames, the relation between a generator and measured energy requires more care.

A subsystem interacting with an unobserved environment need not evolve unitarily by itself. Its reduced dynamics can be a quantum channel,

$$\rho(t)=\mathcal E_{t,t_0}\!\left(\rho(t_0)\right),$$

and may display decoherence, dissipation, or noise. Closed and open dynamics are therefore two levels of description, not competing postulates.

See Unitary Time Evolution and Closed vs Open Systems.

6. Composite Systems Use Tensor Products

For distinguishable subsystems $A$ and $B$, the composite state space is

$$\mathcal H_{AB} = \mathcal H_A\otimes\mathcal H_B.$$

A product state has the form $|\psi_A\rangle\otimes|\phi_B\rangle$, often abbreviated $|\psi_A\phi_B\rangle$. The tensor product also contains linear combinations of product vectors, including entangled states.

Local observables act as $A\otimes I_B$ or $I_A\otimes B$. A typical composite Hamiltonian separates local and interaction terms:

$$H_{AB} = H_A\otimes I_B +I_A\otimes H_B +H_{\mathrm{int}}.$$

The tensor product is not a Cartesian product and not a direct sum. It permits joint alternatives such as $|00\rangle$, $|01\rangle$, $|10\rangle$, and $|11\rangle$ for two qubits, and its dimension multiplies: $\dim\mathcal H_{AB}=(\dim\mathcal H_A)(\dim\mathcal H_B)$.

Continue with Tensor Products of Hilbert Spaces.

7. Identical Particles Require Exchange Symmetry

Labels such as “particle 1” and “particle 2” are bookkeeping devices for identical particles, not experimentally distinguishable identities. In ordinary nonrelativistic quantum mechanics in three spatial dimensions, physical states of identical bosons are symmetric and states of identical fermions are antisymmetric under exchange:

$$P_{12}|\Psi\rangle = \begin{cases} +|\Psi\rangle, & \text{bosons},\\ -|\Psi\rangle, & \text{fermions}. \end{cases}$$

The minus sign for fermions produces the Pauli exclusion principle and Slater determinants. The plus sign for bosons permits macroscopic occupation of one mode. Exchange symmetry shapes atomic shells, molecular structure, quantum statistics, and many-body phases.

This rule is distinct from entanglement. Symmetrizing particle labels can produce nonfactorized expressions, but entanglement for identical particles requires a carefully specified subsystem or mode decomposition.

See the Symmetrization Postulate and Identical-Particle Entanglement Cautions.

8. Entanglement Is Structural, Not a Force

A pure bipartite state is entangled across the split $A|B$ if it cannot be written as a product:

$$|\Psi\rangle \neq |\psi_A\rangle\otimes|\phi_B\rangle.$$

For example,

$$|\Phi^+\rangle = \frac{|00\rangle+|11\rangle}{\sqrt2}$$

is entangled. Its reduced states are

$$\rho_A=\rho_B=\frac{I}{2},$$

even though the joint state is pure. The whole can therefore carry information that is absent from either subsystem alone.

Entanglement changes which joint probability distributions are possible. It does not act as a force, carry energy by itself, or permit controllable faster-than-light signaling. Its definition also depends on the chosen decomposition into subsystems or observable algebras.

For mixed states, “not a product” is insufficient: classically correlated mixtures can be nonproduct but separable. The canonical definitions begin at Entangled States and Separable Mixed States.

9. Classical Physics Emerges as a Regime

Quantum mechanics does not predict that every macroscopic system visibly interferes. Classical descriptions emerge when particular states, observables, scales, environments, and resolutions make quantum corrections negligible.

Several mechanisms can cooperate:

• actions large compared with $\hbar$ support semiclassical expansions;
• narrow wave packets can follow approximately classical trajectories for a limited time;
• decoherence suppresses interference between selected alternatives in a reduced description;
• coarse graining discards phase information below experimental resolution;
• large occupation numbers or large quantum numbers can make relative fluctuations small;
• stationary-phase structure selects classical paths in suitable path integrals.

No single item is a universal classical-limit theorem. Ehrenfest’s theorem alone does not guarantee a narrow packet, and decoherence alone does not select a unique measurement outcome. “Large object” is likewise not a mathematical condition.

Use the Classical Limit for the mechanism map and Ehrenfest Theorem Overview for one important but limited bridge.

10. Quantum Mechanics Has Equivalent Formulations

The same predictions can be organized through state vectors, wavefunctions, matrices, density operators, Heisenberg-picture observables, path integrals, phase-space functions, or operator algebras. These formulations emphasize different structures:

• wave mechanics makes spatial boundary conditions concrete;
• matrix and operator mechanics foreground spectra and commutators;
• density operators handle mixtures, subsystems, and open dynamics;
• the Heisenberg picture makes operator evolution and symmetry transparent;
• path integrals organize composition, semiclassics, and field-theory bridges;
• algebraic formulations emphasize observables, states, and locality.

Equivalence means that corresponding preparations, transformations, and measurements produce the same outcome statistics in their common domain. It does not mean every formulation is equally convenient, equally rigorous, or available under every set of assumptions.

A basis change is also not a change of physics. If state and observable representatives are transformed consistently, probabilities remain invariant:

$$\operatorname{Tr}(\rho E_a) = \operatorname{Tr} \left( V\rho V^\dagger \,V E_a V^\dagger \right).$$

See Equivalent Formulations and the Representation Translation Table.

Worked Example: One Qubit Interferometer

A two-level system shows the first five ideas in one calculation. Prepare

$$|\psi(0)\rangle = |+\rangle = \frac{|0\rangle+|1\rangle}{\sqrt2}.$$

Let the Hamiltonian be

$$H=\frac{\hbar\omega}{2}\sigma_z.$$

After time $t$, and after dropping an irrelevant global phase,

$$|\psi(t)\rangle = \frac{ |0\rangle+e^{i\omega t}|1\rangle }{\sqrt2}.$$

A computational-basis measurement always gives $0$ and $1$ with equal probability. An $X$-basis measurement instead gives

$$\begin{aligned} p(+|t) &= |\langle+|\psi(t)\rangle|^2 \\ &= \frac{1+\cos(\omega t)}{2} \\ &= \cos^2\left(\frac{\omega t}{2}\right). \end{aligned}$$

The state supplies amplitudes, linearity preserves a relative phase, the Hamiltonian evolves that phase, the measurement basis recombines the amplitudes, and the Born rule converts the final amplitude into a probability. Nothing in the state alone predicts an outcome distribution until the measurement is specified.

Worked Example: Composition Creates Entanglement

Prepare two qubits in $|+\rangle_A|0\rangle_B$. A controlled-NOT operation maps

$$\begin{aligned} |00\rangle&\longmapsto|00\rangle, \\ |10\rangle&\longmapsto|11\rangle. \end{aligned}$$

By linearity,

$$|+\rangle_A|0\rangle_B \longmapsto \frac{|00\rangle+|11\rangle}{\sqrt2}.$$

The input is a product state; the output is entangled. Neither subsystem has a pure state after the operation, although the pair does. This example uses superposition, tensor-product composition, unitary dynamics, and entanglement without invoking any force between the qubits.

A Minimal Model-Building Checklist

When approaching a new quantum problem, identify the following before calculating:

1. System and regime: What degrees of freedom are retained, and what has been neglected?
2. State space: Is it finite-dimensional, a wavefunction space, a Fock space, or an effective subspace?
3. Preparation: What density operator or ensemble describes the initial state?
4. Dynamics: What Hamiltonian, unitary, channel, or master equation applies, and over what time range?
5. Measurement: What outcomes and effects model the actual readout?
6. Composition: Which tensor-product split or mode algebra defines the subsystems?
7. Statistics and symmetry: Are particles identical? Which symmetries and constraints restrict the state?
8. Approximation and limit: What scale hierarchy controls the model, and what check would reveal its failure?

Most apparent paradoxes become more precise once one of these declarations is made explicit.

Common Mistakes

• Treating a wavefunction as a physical field in space rather than a representation of a state.
• Calling every basis expansion a distinct physical superposition.
• Confusing a coherent superposition with ignorance about a classical alternative.
• Equating observables only with finite Hermitian matrices and overlooking spectral measures, POVMs, and domain issues.
• Discussing a probability without specifying both state preparation and measurement.
• Assuming the POVM effects uniquely determine the post-measurement state.
• Saying all dynamics are unitary without identifying whether the system is closed.
• Using a direct sum where composition requires a tensor product.
• Treating exchange symmetry and entanglement as the same structure.
• Interpreting entanglement as a signal or interaction.
• Reducing the classical limit to $\hbar\to0$, large mass, or decoherence alone.
• Mistaking a change of basis or picture for a change in physical predictions.

Exercises

Exercise 1: Global and relative phase

Compare

$$|\psi\rangle = \frac{|0\rangle+|1\rangle}{\sqrt2}, \qquad |\phi\rangle = \frac{|0\rangle+i|1\rangle}{\sqrt2}.$$

Are the states related by a global phase? Give one measurement that distinguishes them.

Solution

They are not related by a global phase: matching the coefficient of $|0\rangle$ would fix the global factor to $1$, but then the coefficients of $|1\rangle$ would still differ. A $Y$-basis measurement distinguishes them. The state $|\phi\rangle$ is the $+1$ eigenstate of $\sigma_y$, so it gives the $+y$ outcome with probability one. The state $|\psi\rangle$ gives the two $Y$-basis outcomes with equal probability.

Exercise 2: Probability is not update

Two measuring devices have the same effects $\{E_a\}$ and therefore the same outcome probabilities for every input state. Must they leave the same conditional state after outcome $a$?

Solution

No. A POVM specifies $p(a)=\operatorname{Tr}(\rho E_a)$ but does not uniquely specify the operation applied when $a$ occurs. Different instruments can realize the same effects while disturbing, resetting, or coupling the system in different ways. A state-update claim therefore requires instrument or Kraus-operator data in addition to the effects.

Exercise 3: Read the interference fringe

In the qubit interferometer above, for which times is the $+x$ result certain, and for which times is the $-x$ result certain?

Solution

Since

$$p(+|t)=\cos^2\left(\frac{\omega t}{2}\right),$$

the $+x$ result is certain when $\omega t=2\pi n$. It is impossible when $\omega t=(2n+1)\pi$, at which times the $-x$ result is certain. Here $n$ is any integer.

Exercise 4: Reduced states of a Bell pair

Show that tracing either qubit out of $|\Phi^+\rangle=(|00\rangle+|11\rangle)/\sqrt2$ gives $I/2$.

Solution

The joint density operator is

$$\begin{aligned} \rho_{AB} = \frac12\Bigl( &|00\rangle\langle00| +|11\rangle\langle11| \\ &+|00\rangle\langle11| +|11\rangle\langle00| \Bigr). \end{aligned}$$

Under the partial trace over $B$, the two cross terms vanish because $\langle1|0\rangle=\langle0|1\rangle=0$. The remaining terms give

$$\rho_A = \frac12 \left( |0\rangle\langle0| +|1\rangle\langle1| \right) = \frac{I}{2}.$$

The same argument applies to $\rho_B$.

Exercise 5: Exchange symmetry

Let $|u\rangle$ and $|v\rangle$ be orthonormal one-particle states. Write the normalized two-particle spatial states for identical bosons and identical fermions with one particle in each orbital. What happens to the fermionic state if $|v\rangle=|u\rangle$?

Solution

For distinct orthonormal orbitals,

$$\begin{aligned} |\Psi_B\rangle &= \frac{|u\rangle|v\rangle+|v\rangle|u\rangle}{\sqrt2}, \\ |\Psi_F\rangle &= \frac{|u\rangle|v\rangle-|v\rangle|u\rangle}{\sqrt2}. \end{aligned}$$

If $|v\rangle=|u\rangle$, the antisymmetric numerator vanishes. Two identical fermions cannot occupy the same complete one-particle state, which is the Pauli exclusion principle.

References

• P. A. M. Dirac, The Principles of Quantum Mechanics, 4th ed., Oxford University Press (1958) — foundational operator and transformation theory.
• J. von Neumann, Mathematical Foundations of Quantum Mechanics, Princeton University Press (1955) — Hilbert-space, statistical, and measurement structure.
• L. E. Ballentine, Quantum Mechanics: A Modern Development, World Scientific (1998) — ensemble interpretation, measurement, symmetry, and classical limits.
• A. Peres, Quantum Theory: Concepts and Methods, Kluwer (1995) — states, measurements, composite systems, and operational distinctions.
• J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, 3rd ed., Cambridge University Press (2020) — standard graduate formalism, symmetry, dynamics, and identical particles.
• M. A. Nielsen and I. L. Chuang, Quantum Computation and Quantum Information, 10th anniversary ed., Cambridge University Press (2010) — density operators, channels, tensor products, and entanglement.
• P. Busch, P. Lahti, J.-P. Pellonpää, and K. Ylinen, Quantum Measurement, Springer (2016) — observables, POVMs, instruments, and measurement structure.