What Is Quantum Mechanics?

Quantum mechanics is the framework in which physical systems are represented by states in Hilbert space, measurement alternatives by operators or more general measurement procedures, outcome probabilities by the Born rule, and closed-system dynamics by unitary transformations generated by the Hamiltonian. Composite systems are represented by tensor products.

That definition is deliberately structural. Quantum mechanics is not defined as “the physics of small things,” a catalogue of surprising effects, or one particular interpretation of the formalism. It is a predictive framework that connects preparations, transformations, measurements, and statistical outcomes.

The Short Definition

For a finite or countably described system, the minimal grammar is:

Physical role	Mathematical representation
Pure state	A ray represented by $\lvert\psi\rangle$ in a Hilbert space $\mathcal H$
General state	A positive trace-one density operator $\rho$
Sharp observable	A self-adjoint operator and its spectral projectors
General measurement	A collection of positive effects $\{E_a\}$ with $\sum_aE_a=I$
Outcome probability	$p(a)=\operatorname{Tr}(\rho E_a)$
Closed evolution	A unitary operator $U(t)$ generated by $H$
Composite system	A tensor-product space $\mathcal H_A\otimes\mathcal H_B$

The table is a map, not a complete postulate set. Domains matter for unbounded operators, continuous outcomes require measures, and state update requires more information than outcome probabilities alone. Core Formalism owns those qualifications.

The Prediction Cycle

A quantum model turns a laboratory or physical question into probabilities:

1. Specify a preparation and assign a state $\rho$.
2. Specify the relevant physical process and its transformation $\mathcal E$.
3. Specify a measurement through effects $E_a$.
4. Compute the outcome law

$$p(a) = \operatorname{Tr} \left[ E_a\mathcal E(\rho) \right].$$

5. Compare that law with observed frequencies, including statistical and modeling uncertainty.

For a closed system,

$$\mathcal E(\rho) = U\rho U^\dagger.$$

For an open system, noisy process, or measurement-conditioned evolution, $\mathcal E$ may be a more general quantum channel or instrument.

The state is therefore neither a list of observed outcomes nor a classical trajectory. It is the mathematical object from which the theory generates probabilities for stated procedures.

What Quantum Mechanics Replaced

Classical mechanics represents a point particle by a phase-space point $(q,p)$, or an uncertain ensemble by a probability distribution over phase space. Given forces and initial data, Hamilton’s equations determine the trajectory.

Classical wave theory instead uses fields whose amplitudes can interfere and whose intensities are often quadratic in those amplitudes. It explains diffraction and interference without quantizing matter or radiation.

Quantum mechanics combines and changes aspects of both:

• complex amplitudes can interfere;
• observables need not possess simultaneously sharp values;
• probabilities are calculated from the quantum state and measurement;
• the state can evolve deterministically between interventions;
• composite systems can occupy entangled states that are not assembled from independent subsystem states.

The standard formalism does not identify a quantum state with an unknown point in ordinary classical phase space. Classical descriptions can emerge as controlled approximations, but the approximation must be derived for the states, observables, resolution, and timescale in question.

Core Ingredient: States

A normalized pure-state vector satisfies

$$\langle\psi\mid\psi\rangle = 1.$$

Vectors that differ only by a global phase represent the same pure state:

$$\lvert\psi\rangle \sim e^{i\alpha} \lvert\psi\rangle.$$

The physical pure state is therefore a ray rather than one uniquely phased vector.

A general state is a density operator satisfying

$$\rho \geq 0, \qquad \operatorname{Tr}\rho = 1.$$

Density operators describe statistical mixtures, reduced states of entangled systems, and pure states through

$$\rho_\psi = \lvert\psi\rangle\langle\psi\rvert.$$

A wavefunction is a representation of a state in a chosen basis:

$$\psi(x) = \langle x\mid\psi\rangle.$$

It is not an additional kind of state alongside vectors and density operators. States and Representations develops this distinction.

Core Ingredient: Observables and Measurements

For a sharp observable $A$, the spectral theorem associates measurable sets $\Delta$ with projectors $E_A(\Delta)$. In a discrete case,

$$A = \sum_a aP_a.$$

For a state $\rho$, the probability of outcome $a$ is

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

Not every apparatus is a sharp projective measurement. A general positive-operator-valued measure uses effects $E_a$ such that

$$E_a \geq 0, \qquad \sum_a E_a = I,$$

and

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

The observable or effects determine outcome probabilities. To determine the post-measurement state conditioned on an outcome, one needs an instrument or an explicit measurement model.

This separation matters: a measurement is a physical procedure represented by mathematical objects, not merely a passive revelation of a value assumed to have been present in every context.

Core Ingredient: Amplitudes and Probability

Quantum theory assigns complex amplitudes before assigning probabilities. For a pure state and outcome vector $\lvert a\rangle$, the transition amplitude is

$$\mathcal A(a) = \langle a\mid\psi\rangle,$$

while the Born rule gives

$$p(a) = \lvert\mathcal A(a)\rvert^2.$$

If two alternatives are physically indistinguishable, their amplitudes can add:

$$\mathcal A = \mathcal A_1 + \mathcal A_2.$$

The probability then contains an interference term:

$$\begin{aligned} \lvert\mathcal A\rvert^2 &= \lvert\mathcal A_1\rvert^2 + \lvert\mathcal A_2\rvert^2\\ &\quad+ 2\operatorname{Re} \left( \mathcal A_1^*\mathcal A_2 \right). \end{aligned}$$

If the alternatives become distinguishable through a record or environment, that interference can be reduced or lost in the relevant subsystem description. The rule for adding amplitudes is thus tied to the physical distinguishability of alternatives.

Core Ingredient: Dynamics

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

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

When $H$ is time independent,

$$\lvert\psi(t)\rangle = e^{-iHt/\hbar} \lvert\psi(0)\rangle.$$

The evolution is unitary and preserves inner products and normalization. Probabilities may change because the state changes relative to the chosen measurement basis, even though total probability remains one.

For a density operator,

$$i\hbar \frac{d\rho}{dt} = [H,\rho].$$

Open systems require more general dynamics because correlations with unobserved degrees of freedom can produce noise, dissipation, and decoherence.

Core Ingredient: Composition

If systems $A$ and $B$ have Hilbert spaces $\mathcal H_A$ and $\mathcal H_B$, the composite space is

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

Product states have the form

$$\lvert\psi\rangle_A \otimes \lvert\phi\rangle_B,$$

but not every composite state is a product. An entangled state cannot be written that way:

$$\lvert\Psi\rangle_{AB} \neq \lvert\psi\rangle_A \otimes \lvert\phi\rangle_B.$$

Entanglement allows correlations unavailable to product preparations. It does not permit controllable faster-than-light signaling, and it should not be described as ordinary ignorance about two independent local states.

Composite Systems and Entanglement owns this structure and its information-theoretic consequences.

What Is Distinctly Quantum?

No single slogan captures the distinction, but several structural features recur:

• Superposition: nonzero linear combinations of state vectors are valid state vectors before normalization.
• Interference: probabilities can depend on relative phases between amplitudes.
• Noncommutativity: measurement statistics and joint sharpness depend on operator order and compatibility.
• Uncertainty: some pairs of observables cannot both have arbitrarily small dispersion in one state.
• Entanglement: a composite state can fail to reduce to independent local state assignments.
• Context-dependent measurement statistics: the predicted distribution depends on which physical measurement is performed.

These features are related but not interchangeable. A state can be unentangled and still show single-particle interference. Two observables can fail to commute without every state saturating an uncertainty bound.

Not Merely the Mechanics of Small Objects

Quantum mechanics is often introduced through atoms, electrons, and photons because quantum effects are difficult to ignore at those scales. Size, however, is not the definition.

Large systems can preserve interference, squeezing, tunneling, or entanglement when sufficiently isolated and controlled. Microscopic systems can admit excellent classical approximations when actions are large relative to $\hbar$, states remain localized, measurements are coarse, or environmental decoherence suppresses accessible interference.

The useful question is not simply “Is it small?” It is:

Which quantum predictions differ from the relevant classical model at the available preparation quality, resolution, and timescale?

Classical Limit maps the mechanisms that make those differences negligible in controlled regimes.

What Quantum Mechanics Explains

Quantum mechanics accounts for:

• atomic stability and discrete spectra;
• chemical bonding and molecular structure;
• electronic bands, semiconductors, magnetism, and superconductivity;
• lasers, spontaneous emission, and quantum optical effects;
• nuclear magnetic resonance and other spin-based spectroscopy;
• scattering, tunneling, and interference of matter;
• quantum statistics of identical particles;
• entanglement, quantum communication, and quantum computation;
• the nonrelativistic structure underlying much of atomic, molecular, condensed-matter, and materials physics.

These achievements use many approximations and effective models. The formalism does not replace the work of choosing a Hamiltonian, estimating errors, or validating a model against experiment.

Why Quantum Mechanics Matters gives the application map without treating every emerging technology claim as settled.

What It Does Not Settle by Itself

The predictive formalism does not, by itself, provide a unique answer to every question about:

• what the quantum state represents ontologically;
• whether probabilities are fundamental or emergent;
• how to describe individual outcomes beyond the operational measurement rules;
• which interpretation of the formalism should be preferred;
• why a particular effective Hamiltonian is accurate for a physical device;
• how quantum mechanics and gravitation fit into a final fundamental theory.

Interpretations can matter for conceptual and foundational work, but one does not need to select an interpretation before calculating spectra, transition probabilities, or unitary dynamics correctly.

Decoherence explains important parts of the emergence of stable classical records, but it is not by itself a complete interpretation of measurement. The foundations and open-systems pages keep those claims separate.

Scope and the Boundary with Field Theory

Nonrelativistic quantum mechanics often assumes a fixed set of degrees of freedom or particles. Relativistic wave equations extend some aspects, but processes involving particle creation and annihilation are naturally described by quantum fields.

Quantum field theory adds local fields, relativistic covariance, variable particle number, renormalization, and gauge structure. It does not discard the Hilbert-space, operator, amplitude, and measurement grammar; it embeds that grammar in a broader framework.

This section states the boundary, and Relativistic Quantum Mechanics develops the continuation. Relationship to the QFT Site explains which questions belong on each side of that boundary.

Where to Go Next

• For the shortest guided continuation, use the First Quantum Mechanics Roadmap.
• For the state-observable-probability framework, enter Core Formalism.
• For coordinate-space calculations and standard models, enter Wave Mechanics and Model Systems.
• For current no-go constraints and interpretive limits, enter Foundations and Reality.

Common Mistakes

• Defining quantum mechanics only as the physics of microscopic objects.
• Treating a wavefunction as a literal classical material wave in ordinary space for every system.
• Treating a state vector’s global phase as observable.
• Adding probabilities when indistinguishable alternatives require adding amplitudes.
• Assuming every observable has a sharp value in every state.
• Treating measurement as passive observation rather than a modeled physical procedure.
• Assuming a density operator always represents ordinary ignorance rather than a reduced entangled state.
• Describing entanglement as a mechanism for faster-than-light signaling.
• Treating an interpretation as a prerequisite for using the standard predictive rules.
• Assuming quantum mechanics supplies the correct Hamiltonian without physical modeling and experimental validation.

References

• L. E. Ballentine, Quantum Mechanics: A Modern Development, 2nd ed., World Scientific (2014).
• P. A. M. Dirac, The Principles of Quantum Mechanics, 4th ed., Oxford University Press (1958).
• M. A. Nielsen and I. L. Chuang, Quantum Computation and Quantum Information, Cambridge University Press (2010).
• J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, 3rd ed., Cambridge University Press (2020).
• R. Shankar, Principles of Quantum Mechanics, 2nd ed., Springer (1994).
• J. von Neumann, Mathematical Foundations of Quantum Mechanics, Princeton University Press (1955).

Exercises

1. Match each physical role to its mathematical object: pure state, general state, sharp observable, general measurement, closed evolution, and composite system.

Solution

A pure state is a ray represented by a normalized vector $\lvert\psi\rangle$. A general state is a positive trace-one density operator $\rho$. A sharp observable is represented by a self-adjoint operator and its spectral measure. A general measurement is represented by positive effects $\{E_a\}$ summing to the identity. Closed evolution is unitary, and a composite system uses a tensor-product Hilbert space.

2. Why do $\lvert\psi\rangle$ and $e^{i\alpha}\lvert\psi\rangle$ represent the same pure state?

Solution

For any effect $E_a$,

$$\begin{aligned} &\left( e^{-i\alpha}\langle\psi\rvert \right) E_a \left( e^{i\alpha}\lvert\psi\rangle \right)\\ &\qquad= \langle\psi\rvert E_a \lvert\psi\rangle. \end{aligned}$$

The global phases cancel from every Born probability. Relative phases between components can affect interference, but one phase multiplying the entire state vector does not.

3. Two indistinguishable alternatives have amplitudes $\mathcal A_1$ and $\mathcal A_2$. Identify the term that would be missed by adding the two probabilities separately.

Solution

Adding amplitudes gives

$$\begin{aligned} \left| \mathcal A_1+\mathcal A_2 \right|^2 &= \lvert\mathcal A_1\rvert^2 + \lvert\mathcal A_2\rvert^2\\ &\quad+ 2\operatorname{Re} \left( \mathcal A_1^*\mathcal A_2 \right). \end{aligned}$$

The last term is the interference term. Adding probabilities separately would omit it.

4. Explain why an exactly classical-looking trajectory for $\langle x\rangle$ does not prove that the full state is classical.

Solution

An expectation value is only one moment of a distribution. A state may be broad, split into separated packets, or contain interference while retaining a centroid that follows a classical equation. Classical behavior also requires appropriate localization, fluctuation control, resolution, and often decoherence over the timescale of interest.

5. A bipartite pure state cannot be written as $\lvert\psi\rangle_A\otimes\lvert\phi\rangle_B$. What structural feature does this demonstrate, and what does it not permit?

Solution

The failure of product factorization demonstrates entanglement. Measurements on the subsystems can then have correlations not obtainable from an independent product preparation. Entanglement does not permit controllable faster-than-light signaling; local outcome probabilities cannot be selected remotely to transmit a message.