Phase Space and Its Geometry
~6 min read · Tier toggle changes depth (T1 plain · T2 standard · T3 rigour)
You should be comfortable with Hamilton's canonical equations and the concept of the Legendre transform. Familiarity with basic vector calculus and the idea of a differential form is helpful but not strictly required.
Why this matters
Phase space provides the geometric stage upon which the entire history of a mechanical system is played out.
In Newtonian mechanics, we track position and velocity. In Hamiltonian mechanics, we track position and momentum. The collection of all possible pairs \((q, p)\) forms a mathematical space known as phase space. Understanding the geometry of this space is crucial because it reveals that the flow of a system is not just a set of trajectories, but a structured, volume-preserving map. This geometric perspective is the direct precursor to quantum mechanics, where the commutation relations of operators mirror the Poisson bracket structure of classical phase space.
The Geometry of State
Imagine a simple pendulum. In Newtonian terms, you describe it by its angle \(\theta\) and its angular velocity \(\dot{\theta}\). In Hamiltonian terms, you describe it by \(\theta\) and its conjugate momentum \(p_\theta\). The set of all possible values of \(\theta\) and \(p_\theta\) constitutes a two-dimensional plane. This plane is the phase space for that single degree of freedom.
For a system with \(N\) degrees of freedom, the phase space is \(2N\)-dimensional. Each point in this space represents a complete state of the system. As time evolves, the system traces a curve through this space. Unlike configuration space, which only cares about where things are, phase space cares about where things are and how hard they are moving. This distinction is vital because it allows us to treat the evolution of the system as a flow in a geometric manifold, rather than just a solution to a differential equation.
Hamiltonian Vector Fields
The motion of the system in phase space is governed by Hamilton's canonical equations. These equations define a vector field on the phase space. At every point \((q, p)\), the vector field specifies the instantaneous rate of change of the coordinates.
Given the Hamiltonian \(H(q, p, t)\), the time evolution is given by the system of first-order differential equations:
This pair of equations can be viewed as defining a vector field \(\mathbf{X}_H\) on the phase space. The components of this vector field are the partial derivatives of the Hamiltonian. The trajectory of the system is the integral curve of this vector field. The beauty of this formulation is that the geometry of the phase space dictates the dynamics. The Hamiltonian function \(H\) acts as a generator of time translations, and its gradients determine the direction of the flow.
Liouville's Theorem and Volume Preservation
A profound consequence of the symplectic structure of phase space is Liouville's Theorem. It states that the volume of any region in phase space is conserved as the system evolves. This is not merely a conservation of energy; it is a conservation of phase space volume.
To see why, consider the divergence of the Hamiltonian vector field \(\mathbf{X}_H\). The divergence is given by the sum of the partial derivatives of the components of the vector field with respect to their corresponding coordinates:
Substituting Hamilton's equations into this expression, we find:
Assuming \(H\) is a smooth function, the mixed partial derivatives are equal. Therefore, each term in the sum cancels out, resulting in a divergence of zero. A vector field with zero divergence is incompressible. This means that if you take a cloud of points in phase space, its volume remains constant as it flows, even though its shape may distort and stretch. This incompressibility is a fundamental property of Hamiltonian systems and has deep implications for statistical mechanics, where it ensures that the density of states in phase space is preserved over time.
The Symplectic Form
The geometric structure that underlies Hamiltonian mechanics is the symplectic form. In \(2N\)-dimensional phase space, the symplectic form \(\omega\) is a closed, non-degenerate 2-form. In local coordinates \((q_1, \dots, q_N, p_1, \dots, p_N)\), it is written as:
This form defines a natural area element on the phase space. The flow generated by any Hamiltonian preserves the symplectic form, and this implies Liouville's theorem, because the phase-space volume element is built from \(\omega\) (it is \(\omega^N/N!\) up to sign); for one degree of freedom the two statements coincide. The symplectic structure is what distinguishes phase space from a generic \(2N\)-dimensional manifold. It provides a canonical way to pair positions and momenta, ensuring that the equations of motion are reversible and that the geometry is consistent with the physical principles of mechanics.
The derivation
Hamilton's Canonical Equations
Given a non-singular Lagrangian L(q, q̇, t) with conjugate momentum p = ∂L/∂q̇ and Hamiltonian H(q, p, t) = p q̇ − L obtained by Legendre transform, the dynamics of the system are governed by the pair of symmetric first-order equations q̇ = ∂H/∂p and ṗ = − ∂H/∂q, together with ∂H/∂t = − ∂L/∂t. These are Hamilton's canonical equations, equivalent to the single second-order Euler–Lagrange equation.
Do not confuse phase space volume with configuration space volume. A system can have a constant energy surface in phase space, but the volume of the region enclosed by that surface is not necessarily related to the volume in configuration space. The symplectic structure couples \(q\) and \(p\), so you cannot treat them as independent variables in the same way you might in a Cartesian coordinate system. C
Worked example
Problem. Consider a one-dimensional harmonic oscillator with mass \(m = 2\ \mathrm{kg}\) and spring constant \(k = 8\ \mathrm{N/m}\). The Hamiltonian is \(H = \frac{p^2}{2m} + \frac{1}{2}kq^2\). Calculate the rate of change of the phase space area element \(dq \wedge dp\) along the trajectory of the system.
- Identify the Hamiltonian vector field components: \(\dot{q} = \frac{\partial H}{\partial p} = \frac{p}{m}\) and \(\dot{p} = -\frac{\partial H}{\partial q} = -kq\).
- Calculate the divergence of the vector field in one dimension: \(\nabla \cdot \mathbf{X}_H = \frac{\partial \dot{q}}{\partial q} + \frac{\partial \dot{p}}{\partial p}\).
- Compute the partial derivatives: \(\frac{\partial \dot{q}}{\partial q} = \frac{\partial}{\partial q}\left(\frac{p}{m}\right) = 0\) and \(\frac{\partial \dot{p}}{\partial p} = \frac{\partial}{\partial p}(-kq) = 0\).
- Sum the derivatives to find the divergence: \(0 + 0 = 0\).
- Conclude that the rate of change of the phase space area element is zero, consistent with Liouville's theorem.
Units check. The divergence is a rate, in \(\mathrm{s^{-1}}\); here it is exactly zero, so phase-space area is conserved. (The numbers \(m\) and \(k\) are not needed: the result holds for every Hamiltonian.)
Three quick tests
- What is the dimension of the phase space for a system with \(N\) degrees of freedom?
Answer
2N - What is the divergence of the Hamiltonian vector field?
Answer
Zero - What geometric object defines the symplectic structure of phase space?
Answer
The symplectic form \(\omega\)
Problem set for PU-201 →
Each derivation in this unit ends with five graded problems with full solutions.
Lecture notes drafted with a local model and checked line by line by the editor; every number in the worked example was recomputed.