Poisson Brackets and Nambu Brackets

In classical mechanics, we often start with Newton’s laws ( $\displaystyle F=ma$ ). But there’s a more profound and symmetrical formulation: Hamiltonian mechanics. Here, a system isn’t described by position and velocity, but by generalized coordinates $\displaystyle q_i$ and their conjugate momenta $\displaystyle p_i$ . The set of all possible $\displaystyle (q,p)$ pairs forms the phase space, a stage where the system’s entire history unfolds.

The dynamics are governed by a single master function, the Hamiltonian $\displaystyle H(q,p)$ , which usually represents the total energy. The evolution of the system in time is given by Hamilton’s equations:

$\displaystyle \dot{q}_i = \frac{\partial H}{\partial p_i}$ and $\displaystyle \dot{p}_i = -\frac{\partial H}{\partial q_i}$

Now, consider any other observable quantity, say a function $\displaystyle f(q,p)$ on the phase space. How does it change with time? Using the chain rule, we get:

$\displaystyle \frac{df}{dt} = \sum_{i=1}^{n} \left( \frac{\partial f}{\partial q_i} \dot{q}_i + \frac{\partial f}{\partial p_i} \dot{p}_i \right)$

Substituting Hamilton’s equations for $\displaystyle \dot{q}_i$ and $\displaystyle \dot{p}_i$ gives:

$\displaystyle \frac{df}{dt} = \sum_{i=1}^{n} \left( \frac{\partial f}{\partial q_i} \frac{\partial H}{\partial p_i} - \frac{\partial f}{\partial p_i} \frac{\partial H}{\partial q_i} \right)$

This specific combination of partial derivatives appears so fundamental that we give it its own name: the Poisson bracket of f and H. We define the Poisson bracket of any two functions $\displaystyle f(q,p)$ and $\displaystyle g(q,p)$ as:

$\displaystyle {f,g} = \sum_{i=1}^{n} \left( \frac{\partial f}{\partial q_i} \frac{\partial g}{\partial p_i} - \frac{\partial f}{\partial p_i} \frac{\partial g}{\partial q_i} \right)$

With this notation, the time evolution equation becomes remarkably compact:

$\displaystyle \frac{df}{dt} = \{f,H\}$

This simple equation tells us something profound: The Hamiltonian H is the generator of time evolution. To find out how any quantity f changes in time, you simply “bracket” it with the Hamiltonian. If $\displaystyle \{f,H\}=0$ , the quantity f is a conserved quantity—it doesn’t change with time. Since $\displaystyle \{H,H\}=0$ (check this!), the energy itself is always conserved in this framework.

The Bracket as a Flow

What is the Poisson bracket really doing? It measures how one quantity changes under the flow generated by another.

Every function f on phase space can be thought of as generating a “flow,” a vector field $\displaystyle X_f$ that tells points in phase space where to move. The components of this Hamiltonian vector field are defined exactly as they appeared in Hamilton’s equations:

$\displaystyle X_f = \left( \frac{\partial f}{\partial p_i}, -\frac{\partial f}{\partial q_i} \right)$

So, Hamilton’s equations are just $\displaystyle \frac{dx}{dt} = X_H(x)$ , where $\displaystyle x=(q,p)$ . The system literally flows along the vector field generated by the Hamiltonian.

Now, let’s look at the Poisson bracket again. Notice that $\displaystyle \{f,g\}$ can be written as:

$\displaystyle \{f,g\} = \sum_{i=1}^{n} \left( \frac{\partial g}{\partial p_i} \frac{\partial f}{\partial q_i} - \frac{\partial g}{\partial q_i} \frac{\partial f}{\partial p_i} \right) = \nabla g \cdot X_f$

This is just the directional derivative of g along the vector field $\displaystyle X_f$ . Therefore, the Poisson bracket $\displaystyle \{f,g\}$ has a beautiful geometric interpretation:

The Poisson bracket $\displaystyle \{f,g\}$ measures the rate of change of the function g as you flow along the vector field generated by the function f.

So, $\displaystyle \{f,H\}$ is the rate of change of f along the flow generated by H (time evolution), which is exactly what we found earlier.

The Symplectic Viewpoint

There’s an even deeper geometric layer. The phase space $\displaystyle \mathbb{R}^{2n}$ is not just a Euclidean space; it’s a symplectic manifold. This means it’s equipped with a special mathematical object called a symplectic form, $\displaystyle \omega$ , which measures oriented areas projected onto the $\displaystyle q_i - p_i$ planes. In standard coordinates, it’s written as:

$\displaystyle \omega = \sum_{i=1}^{n} dq_i \wedge dp_i$

This 2-form $\displaystyle \omega$ is the heart of Hamiltonian mechanics. It provides the link between a function (like the Hamiltonian H) and its vector field $\displaystyle X_H$ . The connection is given by the relation:

$\displaystyle dH(\cdot) = \omega(X_H, \cdot)$

This equation is a coordinate-free way of defining the Hamiltonian vector field $\displaystyle X_H$ . If you work this out in coordinates, you recover the familiar components $\displaystyle X_H = \left( \frac{\partial H}{\partial p}, -\frac{\partial H}{\partial q} \right)$ .

Using this machinery, the Poisson bracket can be defined without coordinates at all:

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

This definition reveals the essence of the bracket: it’s the symplectic area of the parallelogram formed by the Hamiltonian vector fields $\displaystyle X_f$ and $\displaystyle X_g$ . The flows generated by Hamiltonians are special—they are symplectomorphisms, meaning they preserve the symplectic form $\displaystyle \omega$ . This is the content of Liouville’s theorem, which states that phase space volume is conserved under Hamiltonian evolution.

The Poisson Structure

We can abstract the key properties of the Poisson bracket to define a more general object. A Poisson manifold is a smooth manifold M whose algebra of smooth functions $\displaystyle C^\infty(M)$ is equipped with a bracket operation $\displaystyle {\cdot,\cdot}$ satisfying:

Antisymmetry: $\displaystyle \{f,g\} = -\{g,f\}$
Bilinearity: It’s linear in each argument.
Leibniz Rule (Derivation): $\displaystyle \{f,gh\} = g\{f,h\} + h\{f,g\}$ . This ensures the bracket interacts correctly with the product of functions.
Jacobi Identity: $\displaystyle \{f,\{g,h\}\} + \{g,\{h,f\}\} + \{h,\{f,g\}\} = 0$

The Jacobi identity is crucial. It ensures that the time evolution is consistent. For example, if we have two conserved quantities F and G (so $\displaystyle \{F,H\}=0$ and $\displaystyle \{G,H\}=0$ ), the Jacobi identity implies that their bracket $\displaystyle \{F,G\}$ is also a conserved quantity:

$\displaystyle \{\{F,G\},H\} = -\{\{G,H\},F\} - \{\{H,F\},G\} = -\{0,F\} - \{0,G\} = 0$

This algebraic structure—an algebra of functions that is also a Lie algebra under the bracket—is the essence of a Poisson manifold.

Examples:
Canonical $\displaystyle \mathbb{R}^{2n}$ : The standard phase space of physics with $\displaystyle \{f,g\} = \sum_i \left( \frac{\partial f}{\partial q_i} \frac{\partial g}{\partial p_i} - \frac{\partial f}{\partial p_i} \frac{\partial g}{\partial q_i} \right)$ . This is the prototype. This structure is non-degenerate, meaning if $\displaystyle \{f,g\}=0$ for all g, then f must be a constant.

Lie-Poisson Structure on $\displaystyle \mathfrak{g}^*$ : This is a vast and crucial class of examples. Let $\displaystyle \mathfrak{g}$ be a Lie algebra (like the algebra of infinitesimal rotations) and $\displaystyle \mathfrak{g}^*$ its dual space. The functions on $\displaystyle \mathfrak{g}^*$ have a natural Poisson bracket.

Angular Momentum in $\displaystyle \mathbb{R}^3$
Let’s take the Lie algebra $\displaystyle \mathfrak{g} = \mathfrak{so}(3)$ of infinitesimal rotations, which can be identified with $\displaystyle \mathbb{R}^3$ where the Lie bracket is the vector cross product. The dual space $\displaystyle \mathfrak{g}^*$ is also $\displaystyle \mathbb{R}^3$ . Let the coordinates on this space be $\displaystyle (L_1, L_2, L_3)$ , representing the components of an angular momentum vector. The Lie-Poisson bracket for coordinate functions is given by:

$\displaystyle \{L_i, L_j\} = \varepsilon_{ijk} L_k$

where $\displaystyle \varepsilon_{ijk}$ is the Levi-Civita symbol. So, $\displaystyle \{L_1, L_2\}=L_3$ , $\displaystyle \{L_2, L_3\}=L_1$ , and $\displaystyle \{L_3, L_1\}=L_2$ . These are the commutation relations for angular momentum!
This bracket is degenerate. There is a special function, a Casimir, that commutes with everything: $\displaystyle C = L_1^2 + L_2^2 + L_3^2$ . You can check that $\displaystyle \{C, L_i\} = 0$ for all i. This means C is a constant of motion for any Hamiltonian defined on this space. The phase space is foliated by spheres of constant radius $\displaystyle C$ , and the Hamiltonian dynamics are confined to one of these spheres.

Trivial Poisson Structure: $\displaystyle \{f,g\}=0$ for all f,g. All dynamics are frozen.

Generalization to Nambu Brackets

In 1973, Yoichiro Nambu proposed a generalization of Hamiltonian mechanics. What if the dynamics were governed not by one Hamiltonian, but by several? This leads to the Nambu bracket.

For a space with n coordinates $\displaystyle (x_1, \dots, x_n)$ , the Nambu bracket of n functions is their Jacobian determinant:

$\displaystyle \{f_1, \dots, f_n\} = \det\left( \frac{\partial (x_1, \dots, x_n)}{\partial (f_1, \dots, f_n)} \right)$

The case $\displaystyle n=2$ gives $\displaystyle \{f_1, f_2\} = \frac{\partial f_1}{\partial x_1} \frac{\partial f_2}{\partial x_2} - \frac{\partial f_1}{\partial x_2} \frac{\partial f_2}{\partial x_1}$ , which is the Poisson bracket on $\displaystyle \mathbb{R}^2$ .

Nambu’s Equations of Motion
In an n-dimensional space, if you have $\displaystyle n-1$ “Hamiltonians” $\displaystyle H_1, \dots, H_{n-1}$ , the time evolution of any observable F is given by:

$\displaystyle \frac{dF}{dt} = \{F, H_1, \dots, H_{n-1}\}$

The flow is now generated simultaneously by all the Hamiltonians and is volume-preserving. The flow is orthogonal to the gradient of all the functions $H_i$ .

The Nambu bracket has its own version of the Jacobi identity, called the Fundamental Identity, which ensures the algebraic consistency of the dynamics. For a 3-bracket, one form of the FI, which shows that the bracket acts as a derivation,

$\displaystyle \{\{f, g, h\}, k, l\} = \{\{f, k, l\}, g, h\} + \{f, \{g, k, l\}, h\} + \{f, g, \{h, k, l\}\}$

Example: The Euler Top in Nambu Form
Let’s revisit the rigid body, but now in $\displaystyle \mathbb{R}^3$ with coordinates $\displaystyle (L_1, L_2, L_3)$ . The dynamics are governed by two conserved quantities:

Energy: $\displaystyle H_1 = \frac{1}{2} \left( \frac{L_1^2}{I_1} + \frac{L_2^2}{I_2} + \frac{L_3^2}{I_3} \right)$
Total Angular Momentum Squared: $\displaystyle H_2 = \frac{1}{2}(L_1^2 + L_2^2 + L_3^2)$

Let’s use a 3-bracket and see what happens. The evolution of $\displaystyle L_1$ is:

$\displaystyle \dot{L}_1 = {L_1, H_1, H_2} = \det \left[ \begin{array}{ccc} 1 & 0 & 0 \\ \frac{L_1}{I_1} & \frac{L_2}{I_2} & \frac{L_3}{I_3} \\ L_1 & L_2 & L_3 \end{array} \right] = L_2 L_3 \left( \frac{1}{I_2} - \frac{1}{I_3} \right)$

This is precisely Euler’s equation for the free rigid body! The Nambu formulation shows that the dynamics can be seen as the intersection of surfaces of constant energy and constant angular momentum.

Nambu mechanics provides a tantalizing glimpse into more complex dynamical systems, with applications in fluid dynamics and M-theory, suggesting that the rich algebraic and geometric structure first uncovered by Hamilton and Poisson extends far beyond simple mechanical systems.

Share this:

Related

Leave a comment Cancel reply