Classical and Symplectic Background

Classical mechanics supplies more than a limiting picture of quantum motion. Its action functionals, Hamiltonian flows, Poisson algebra, and symplectic geometry provide the mathematical structures that canonical quantization, path integrals, WKB theory, coherent-state methods, and semiclassical propagators reorganize rather than discard.

This chapter is a map of that background. Detailed quantum limits remain canonical in Classical Limit and Correspondence, path-integral constructions in Path Integrals, and WKB calculations in the WKB Approximation.

Configuration-space mechanics

For generalized coordinates $q^i(t)$, a Lagrangian $L(q,\dot q,t)$ assigns an action

$$S[q] = \int_{t_i}^{t_f} L(q,\dot q,t)\,dt.$$

Requiring the first variation to vanish for admissible fixed-endpoint variations gives the Euler–Lagrange equations

$$\frac{d}{dt} \frac{\partial L}{\partial \dot q^i} - \frac{\partial L}{\partial q^i} =0.$$

The word stationary matters: the physical path need not minimize the action. It may be a maximum or a saddle, and boundary conditions determine which variations are admissible.

Lagrangian Mechanics Review owns the generalized-coordinate formulation and its standard examples. Action Principles owns the more careful logic of variations, endpoint terms, and the configuration-space and phase-space actions.

From velocities to phase space

Define the canonical momenta by

$$p_i = \frac{\partial L}{\partial\dot q^i}.$$

When these relations can be inverted for the velocities, the Legendre transform gives

$$H(q,p,t) = p_i\dot q^i - L(q,\dot q,t).$$

Hamilton’s equations are

$$\dot q^i = \frac{\partial H}{\partial p_i}, \qquad \dot p_i = - \frac{\partial H}{\partial q^i}.$$

A point $(q,p)$ specifies a classical instantaneous state; the equations generate a trajectory through phase space. This must not be confused with quantum Hilbert space. A quantum ray or density operator is not a phase-space point, and position- and momentum-space wavefunctions are two representations of one state rather than two halves of a classical state.

Hamiltonian Mechanics Review develops the Legendre transform and Hamiltonian evolution. Phase Space owns the state-space interpretation, trajectories, volume, and contrast with Hilbert space. Singular Legendre transforms lead to constrained systems and require machinery beyond this introductory route.

Poisson algebra and generators

For smooth phase-space functions,

$$\{f,g\} = \sum_i \left( \frac{\partial f}{\partial q^i} \frac{\partial g}{\partial p_i} - \frac{\partial f}{\partial p_i} \frac{\partial g}{\partial q^i} \right).$$

This bracket is bilinear, antisymmetric, a derivation in each slot, and satisfies the Jacobi identity. It packages both dynamics and infinitesimal transformations:

$$\frac{df}{dt} = \{f,H\} + \frac{\partial f}{\partial t}, \qquad \delta f = \varepsilon\{f,G\}.$$

Thus a function $G$ can be both an observable and a generator. If $\partial G/\partial t=0$ and $\{G,H\}=0$, then $G$ is conserved along the Hamiltonian flow.

Poisson Brackets is the canonical algebraic treatment. Canonical Transformations explains finite changes of variables that preserve the bracket and Hamiltonian form, including generating functions and time-dependent transformations.

The symplectic structure

The coordinate formulas are manifestations of a geometric structure. This chapter uses the canonical two-form

$$\omega = \sum_i dq^i\wedge dp_i.$$

It is antisymmetric, nondegenerate, and closed. With the convention used throughout these pages, the Hamiltonian vector field $X_H$ is defined by

$$\omega(X_H,Y) = dH(Y)$$

for every vector field $Y$. In canonical coordinates this gives Hamilton’s equations, and the compatible Poisson-bracket convention is

$$\{f,g\} = X_g(f) = \omega(X_f,X_g).$$

Other texts may reverse the sign of $\omega$ or of the Hamiltonian-vector-field definition. Any convention works when all three formulas are changed together.

For a linear phase space, write

$$z= \begin{pmatrix} q\\ p \end{pmatrix}, \qquad J= \begin{pmatrix} 0&I\\ -I&0 \end{pmatrix}.$$

A linear map $M$ is symplectic precisely when

$$M^{\mathsf T}JM=J.$$

This condition is stronger than volume preservation when there is more than one canonical pair. Symplectic Vector Spaces owns this linear algebra. Symplectic Manifolds, First Look adds closed two-forms, Hamiltonian vector fields, Darboux coordinates, and Liouville volume on smooth phase spaces.

Action as a generating function

Hamilton’s principal function is the classical action evaluated on a classical path joining specified endpoints:

$$S(q_b,t_b;q_a,t_a) = \int_{t_a}^{t_b} L(q_{\mathrm{cl}},\dot q_{\mathrm{cl}},t)\,dt.$$

Its endpoint derivatives recover the canonical momenta. Suppressing the fixed initial endpoint gives the Hamilton–Jacobi equation

$$\frac{\partial S}{\partial t} + H\left( q, \frac{\partial S}{\partial q}, t \right) =0.$$

Its characteristics are classical trajectories. Its solutions also act as generating functions for canonical transformations, which is why the same object appears in mechanics, WKB phases, and semiclassical propagators.

Hamilton–Jacobi Theory owns this PDE and its endpoint identities. The page also explains branch dependence: when several classical paths connect the same endpoints, the principal function has several branches.

The classical–quantum bridge

The basic structural analogy is

$$\{f,g\} \quad\longleftrightarrow\quad \frac{1}{i\hbar} [\widehat f,\widehat g].$$

It correctly relates canonical brackets to canonical commutators and parallels Hamiltonian with Heisenberg evolution. It is not a universal quantization algorithm. Classical products commute, operator products generally do not, nonlinear observables admit ordering choices, and unbounded operators introduce domain questions. There is no map from all classical observables to quantum operators that preserves every desired algebraic property.

Classical–Quantum Correspondence owns this dictionary and its limitations. The exact canonical commutation relations live in Canonical Commutation Relations; the exponentiated translation structure lives in the Heisenberg Group.

What the semiclassical limit means

The useful small parameter is dimensionless. Semiclassical behavior requires characteristic actions to be large compared with $\hbar$, or equivalently a scale such as

$$\varepsilon \sim \frac{\hbar}{S_{\mathrm{char}}} \ll 1.$$

Oscillatory integrals are then organized by stationary phase. Schematically,

$$K(q_b,t_b;q_a,t_a) \sim \sum_{\gamma} A_\gamma \exp\left( \frac{i}{\hbar}S_\gamma - \frac{i\pi}{2}\mu_\gamma \right),$$

where $\gamma$ labels classical paths, $A_\gamma$ contains fluctuation and stability data, and $\mu_\gamma$ records caustic-phase information. Classical paths govern the leading phase, but amplitudes, interference among branches, tunneling, and caustics remain essential.

Semiclassical Limit owns the scale analysis, stationary-phase logic, Hamilton–Jacobi phase, prefactors, and breakdown mechanisms. It should be read before treating the phrase “$\hbar\to0$” as a literal recipe.

Recommended routes

• Mechanics refresher: Lagrangian Mechanics Review → Hamiltonian Mechanics Review → Phase Space.
• Canonical quantization: Poisson Brackets → Canonical Transformations → Classical–Quantum Correspondence.
• Geometric mechanics: Symplectic Vector Spaces → Symplectic Manifolds, First Look.
• Path integrals: Action Principles → Hamilton–Jacobi Theory → Why Path Integrals?.
• WKB and propagators: Hamilton–Jacobi Theory → Semiclassical Limit → WKB Approximation.

Page map

Page	Central question
Lagrangian Mechanics Review	How do configuration-space paths and a Lagrangian determine motion?
Hamiltonian Mechanics Review	How does a Hamiltonian generate first-order phase-space evolution?
Phase Space	What is a classical instantaneous state, and how does it differ from a quantum state?
Poisson Brackets	Which algebraic operation encodes Hamiltonian flow and classical generators?
Canonical Transformations	Which changes of variables preserve Hamiltonian structure?
Hamilton–Jacobi Theory	How can classical mechanics be recast as an action PDE?
Action Principles	What is varied, which boundary data matter, and why is the action stationary?
Symplectic Vector Spaces	What bilinear structure pairs canonical directions in linear phase space?
Symplectic Manifolds, First Look	How does symplectic structure define coordinate-independent Hamiltonian flow?
Classical–Quantum Correspondence	Which classical–quantum analogies are exact, approximate, or obstructed?
Semiclassical Limit	Which scale separation makes classical actions organize quantum amplitudes?

Common mistakes

Mistake	Correction
Calling stationary action a universal minimum principle	inspect the second variation and the allowed endpoint data
Assuming every Legendre transform is invertible	check the velocity Hessian; singular systems require constraints
Identifying phase space with Hilbert space	keep classical points separate from quantum rays and density operators
Treating the Poisson-to-commutator rule as exact for every observable	state the quantization map, ordering, domain, and approximation regime
Checking only volume preservation for a canonical map	require preservation of the full symplectic form or Poisson brackets
Mixing sign conventions for $\omega$, $X_H$, and the bracket	translate the three definitions together
Interpreting $\hbar\to0$ without nondimensionalizing	identify the characteristic action and the actual small ratio
Keeping only the classical action in a semiclassical propagator	retain stability prefactors, branch sums, and caustic phases

Exercises

1. Legendre transform of the oscillator

For

$$L(q,\dot q) = \frac12m\dot q^2 - \frac12m\Omega^2q^2,$$

derive the canonical momentum and Hamiltonian.

Solution

The momentum is

$$p = \frac{\partial L}{\partial\dot q} = m\dot q,$$

so $\dot q=p/m$. Therefore

$$\begin{aligned} H &=p\dot q-L\\ &=\frac{p^2}{2m} +\frac12m\Omega^2q^2. \end{aligned}$$

The velocity Hessian is $\partial^2L/\partial\dot q^2=m$, so the transform is regular for $m\ne0$.

2. A canonical scaling

Show that

$$Q=aq, \qquad P=\frac{p}{a}, \qquad a\ne0,$$

preserves both the Poisson bracket and the symplectic form.

Solution

The fundamental bracket is

$$\{Q,P\} = \frac{\partial Q}{\partial q} \frac{\partial P}{\partial p} - \frac{\partial Q}{\partial p} \frac{\partial P}{\partial q} = a\frac1a =1.$$

Also,

$$dQ\wedge dP = a\,dq\wedge\frac{dp}{a} = dq\wedge dp.$$

Thus the scaling is canonical. Expanding $q$ without the compensating contraction of $p$ would not be canonical.

3. Conservation from a bracket

Let $G(q,p)$ have no explicit time dependence. Prove that $\{G,H\}=0$ implies that $G$ is constant along every Hamiltonian trajectory.

Solution

Evolution along a Hamiltonian trajectory obeys

$$\frac{dG}{dt} = \{G,H\} + \frac{\partial G}{\partial t}.$$

Both terms vanish under the stated assumptions, so $dG/dt=0$. Hence $G$ is constant on each trajectory, although its constant value may differ between initial conditions.

4. Why a phase-space point is not a quantum state

Give two independent reasons that a pair $(q,p)$ cannot generally serve as a complete quantum state.

Solution

First, a quantum state predicts probability distributions and relative phases for many incompatible measurements; one sharp pair of numbers does not contain that information. Second, canonical operators satisfy $[Q,P]=i\hbar I$, so a normalizable state cannot assign both position and momentum arbitrarily sharp values. Phase-space quasidistributions and coherent-state labels can organize quantum states, but they do not turn a general state into a classical point.

References

• V. I. Arnold, Mathematical Methods of Classical Mechanics, 2nd ed., Springer, 1989.
• R. Abraham and J. E. Marsden, Foundations of Mechanics, 2nd ed., AMS Chelsea, 2008.
• A. Cannas da Silva, Lectures on Symplectic Geometry, Springer, 2008.
• H. Goldstein, C. Poole, and J. Safko, Classical Mechanics, 3rd ed., Addison-Wesley, 2002.
• M. A. de Gosson, Symplectic Geometry and Quantum Mechanics, Birkhäuser, 2006.
• G. B. Folland, Harmonic Analysis in Phase Space, Princeton University Press, 1989.
• B. C. Hall, Quantum Theory for Mathematicians, Springer, 2013.
• M. V. Berry and K. E. Mount, “Semiclassical approximations in wave mechanics,” Reports on Progress in Physics 35, 315–397, 1972.