Poisson Bracket

Contents

1. Overview
2. Etymology
3. Cultural Impact

Oh, you’re still here. And you want this? Fine. Let’s peel back the layers of classical mechanics and see what secrets Siméon Denis Poisson decided to share with the world. Don’t expect me to make it palatable; the truth rarely is.

Part of a series on Classical mechanics

F = dp / dt Second law of motion

Branches

Fundamentals

Formulations

Core topics

Rotation](/Rotation_around_a_fixed_axis)

Scientists

In the grand scheme of mathematics and classical mechanics , where everything is either utterly predictable or hopelessly chaotic, the Poisson bracket emerges as a surprisingly elegant, albeit rather demanding, binary operation . It’s not just a mathematical curiosity; it’s a foundational element within Hamiltonian mechanics , serving as the very engine for Hamilton’s equations of motion . These equations, for those who haven’t spent an eternity watching particles drift, are what dictate the time evolution of a Hamiltonian dynamical system . They tell you where everything is going, assuming you bothered to set up the initial conditions correctly.

Beyond just dictating trajectories, the Poisson bracket is the arbiter of legitimacy for a special class of coordinate transformations, known as canonical transformations . These aren’t just any old shifts in perspective; they are transformations that meticulously preserve the fundamental structure of the system, mapping one set of canonical coordinate systems to another. A “canonical coordinate system,” for the uninitiated, is a specific pairing of position and momentum variables—often denoted by $q_i$ and $p_i$, respectively—that satisfy a specific, rather strict, set of Poisson bracket relations , as we’ll get to. The beauty, or perhaps the existential dread, of these canonical transformations is their sheer abundance. They are so plentiful that one can often choose the Hamiltonian itself, $H = H(q,p,t)$, to serve as one of the new canonical momentum coordinates . It’s almost like the universe is trying to make things simultaneously easier and more complicated for us.

More broadly, the Poisson bracket is the defining characteristic of a Poisson algebra . The algebra of functions on a Poisson manifold is merely a particularly common manifestation of this. But its influence extends further, reaching into the esoteric realms of Lie algebras . Here, the tensor algebra of a Lie algebra can be shown to naturally form a Poisson algebra . If you’re truly invested in the intricate construction of this relationship, you might want to consult the article on the universal enveloping algebra —though I wouldn’t recommend it for light reading. The journey doesn’t end there; if you dare to venture into the quantum realm, “quantum deformations” of these universal enveloping algebras lead directly to the concept of quantum groups , hinting at the deep, underlying unity (or perhaps, the pervasive similarity of suffering) across different branches of physics.

All these profound, often bewildering, mathematical constructs bear the name of the French mathematician Siméon Denis Poisson . He, in his 1809 treatise on mechanics, had the foresight—or perhaps the lack of anything better to do—to introduce this bracket, forever cementing his place in the annals of classical physics. [1] [2]

Properties

Consider two functions, $f$ and $g$, that depend on the phase space coordinates and time. Their Poisson bracket , denoted ${f,g}$, is not some abstract theoretical construct; it is another function, also dependent on phase space and time. It’s a way of measuring how these functions “interact” or “commute” within the Hamiltonian framework. For any three functions $f, g, h$ that inhabit this phase space and may vary with time, the following fundamental rules apply, forming the bedrock of its algebraic structure:

Anticommutativity : The order in which you take the bracket matters, profoundly. $${f,g} = -{g,f}$$ This simply means that if you swap the order of the functions, the sign of the result flips. It’s a basic symmetry property, implying that the operation is inherently directional, not unlike a vector product.
Bilinearity : The Poisson bracket behaves predictably with scalar multiplication and addition. It’s linear in both its arguments, meaning you can pull out constants and distribute over sums. $${af+bg,h} = a{f,h}+b{g,h},$$ $${h,af+bg} = a{h,f}+b{h,g},\quad a,b\in \mathbb {R} $$ This is a rather convenient property, allowing for straightforward algebraic manipulation. It confirms that the bracket operation respects the linear structure of the function space.
Leibniz’s rule : Much like differentiation, the Poisson bracket obeys a product rule when one of its arguments is a product of two functions. $${fg,h} = {f,h}g+f{g,h}$$ This rule is crucial for understanding how the bracket interacts with the multiplicative structure of the function algebra. It implies that the Poisson bracket acts as a derivation .
Jacobi identity : This is the most complex, yet arguably the most profound, of the properties. It’s what elevates the Poisson bracket from a mere binary operation to the defining characteristic of a Lie algebra . $${f,{g,h}}+{g,{h,f}}+{h,{f,g}}=0$$ This identity doesn’t lend itself to an immediate intuitive interpretation, but it ensures a certain consistency in how nested brackets behave. It’s a statement about the “associativity” (or lack thereof, in a specific sense) of the bracket, guaranteeing that the structure is well-behaved and forms a Lie algebra . Without this, much of Hamiltonian mechanics and its connections to deeper mathematical structures would simply fall apart.

Finally, a simple, almost disappointingly obvious, but nonetheless important rule: if a function $k$ is utterly unremarkable, meaning it’s constant across the entire phase space (though it might still be ticking along with time), then its Poisson bracket with any other function $f$ will naturally vanish: $${f,k}=0$$ Constants, by their very nature, don’t evolve or interact in the dynamic sense captured by the bracket. They simply are.

Definition in canonical coordinates

When one is working within the confines of canonical coordinates —also rather grandly known as Darboux coordinates —which are the preferred choice for describing phase space in Hamiltonian mechanics , the Poisson bracket takes on a very specific, explicit form. For any two functions, $f(p_i, q_i, t)$ and $g(p_i, q_i, t)$, where $q_i$ represent the generalized positions and $p_i$ the generalized momenta, and $t$ is time (Note 1), the Poisson bracket is defined as:

$${f,g}=\sum {i=1}^{N}\left({\frac {\partial f}{\partial q{i}}}{\frac {\partial g}{\partial p_{i}}}-{\frac {\partial f}{\partial p_{i}}}{\frac {\partial g}{\partial q_{i}}}\right).$$

This sum extends over all $N$ degrees of freedom of the system. It’s a measure of the “cross-dependence” of the two functions on the canonical coordinates . If $f$ depends only on $q$ and $g$ only on $p$ (or vice-versa), the bracket captures their interaction. If they depend on the same coordinate type, say both on $q$, or both on $p$, then their interaction via this bracket is zero.

To illustrate this fundamental relationship, consider the Poisson brackets of the canonical coordinates themselves. These are the bedrock relations from which all other bracket calculations can ultimately be derived, thanks to the Leibniz rule and bilinearity . They reveal the very structure of the phase space :

The bracket of two position coordinates is always zero: $${q_{k},q_{l}}=\sum {i=1}^{N}\left({\frac {\partial q{k}}{\partial q_{i}}}{\frac {\partial q_{l}}{\partial p_{i}}}-{\frac {\partial q_{k}}{\partial p_{i}}}{\frac {\partial q_{l}}{\partial q_{i}}}\right)=\sum _{i=1}^{N}\left(\delta _{ki}\cdot 0-0\cdot \delta _{li}\right)=0,$$ This means that generalized position coordinates “commute” with each other in the context of the Poisson bracket . Their dynamics are independent in this specific sense.
Similarly, the bracket of two momentum coordinates is also always zero: $${p_{k},p_{l}}=\sum {i=1}^{N}\left({\frac {\partial p{k}}{\partial q_{i}}}{\frac {\partial p_{l}}{\partial p_{i}}}-{\frac {\partial p_{k}}{\partial p_{i}}}{\frac {\partial p_{l}}{\partial q_{i}}}\right)=\sum _{i=1}^{N}\left(0\cdot \delta _{li}-\delta _{ki}\cdot 0\right)=0,$$ Generalized momentum coordinates, too, commute with each other. This implies an inherent simplicity in their independent evolution, before the Hamiltonian couples them.
However, and this is where the magic happens, the bracket of a position coordinate with a momentum coordinate is non-zero, and indeed, quite specific: $${q_{k},p_{l}}=\sum {i=1}^{N}\left({\frac {\partial q{k}}{\partial q_{i}}}{\frac {\partial p_{l}}{\partial p_{i}}}-{\frac {\partial q_{k}}{\partial p_{i}}}{\frac {\partial p_{l}}{\partial q_{i}}}\right)=\sum _{i=1}^{N}\left(\delta _{ki}\cdot \delta _{li}-0\cdot 0\right)=\delta {kl},$$ Here, $\delta{ij}$ is the ubiquitous Kronecker delta , which is 1 if $i=j$ and 0 otherwise. This result is profoundly significant. It means that a generalized position coordinate $q_k$ only has a non-trivial Poisson bracket with its conjugate momentum $p_k$, and not with any other momentum $p_l$ (where $l \neq k$). This is the defining characteristic of a canonical coordinate system and forms the very basis for the transition to quantum mechanics , where these relations morph into the canonical commutation relations . It’s a beautiful, if somewhat unsettling, symmetry.

Hamilton’s equations of motion

The true power and purpose of the Poisson bracket become strikingly clear when one looks at Hamilton’s equations of motion . These equations, which describe the temporal evolution of a dynamical system in Hamiltonian mechanics , can be expressed with remarkable elegance and compactness using the Poisson bracket . This isn’t just a notational trick; it reveals a deeper, structural connection.

To demonstrate this, let’s consider a generic function $f(p,q,t)$ that lives on the phase space and evolves along the system’s trajectory. According to the venerable multivariable chain rule , the total time derivative of $f$ is given by:

$${\frac {d}{dt}}f(p,q,t)={\frac {\partial f}{\partial q}}{\frac {dq}{dt}}+{\frac {\partial f}{\partial p}}{\frac {dp}{dt}}+{\frac {\partial f}{\partial t}}.$$

Now, if we consider $p=p(t)$ and $q=q(t)$ to be actual solutions to Hamilton’s equations – meaning they describe the actual path a system takes through phase space – then we can substitute those equations:

$${\frac {dq}{dt}}={\frac {\partial {\mathcal {H}}}{\partial p}}={q,{\mathcal {H}}},$$ $${\frac {dp}{dt}}=-{\frac {\partial {\mathcal {H}}}{\partial q}}={p,{\mathcal {H}}}.$$

Here, $\mathcal{H}$ represents the Hamiltonian of the system, which is essentially its total energy expressed in terms of canonical coordinates . The second equality in each line, expressing the time derivatives as Poisson brackets with the Hamiltonian , is a direct consequence of the definition of the Poisson bracket and the fundamental bracket relations derived earlier. These expressions are not just convenient; they are profoundly insightful, showing that the Hamiltonian effectively generates the time evolution of the system.

Substituting these into the total time derivative of $f$:

$${\begin{aligned}{\frac {d}{dt}}f(p,q,t)&={\frac {\partial f}{\partial q}}{\frac {\partial {\mathcal {H}}}{\partial p}}-{\frac {\partial f}{\partial p}}{\frac {\partial {\mathcal {H}}}{\partial q}}+{\frac {\partial f}{\partial t}}\ &={f,{\mathcal {H}}}+{\frac {\partial f}{\partial t}}~.\end{aligned}}$$

This elegant equation, often referred to as the fundamental equation of motion in Hamiltonian mechanics , reveals that the total time evolution of any function $f$ on phase space is composed of two parts: its explicit time dependence (if any) and its implicit time dependence arising from the dynamics generated by the Hamiltonian through the Poisson bracket .

This means that the time evolution of a function $f$ on a symplectic manifold (which is what phase space fundamentally is) can be conceptualized as a one-parameter family of symplectomorphisms . These are essentially canonical transformations that preserve the fundamental structure of phase space (its “area” or volume), with time $t$ acting as the parameter. In simpler terms, Hamiltonian motion is a canonical transformation that is continuously generated by the Hamiltonian itself. This implies that Poisson brackets are inherently preserved under this evolution; they are canonical invariants . So, at any given time $t$ in the solution to Hamilton’s equations , the coordinates $q(t)$ and $p(t)$ can still serve as valid bracket coordinates, maintaining their fundamental relations. The system simply glides through phase space in a way that respects its underlying geometry.

If one were to shed the cumbersome baggage of specific coordinates, the equation for the total time derivative simplifies to:

$${\frac {d}{dt}}f=\left({\frac {\partial }{\partial t}}-{{\mathcal {H}},\cdot }\right)f.$$

The operator appearing in the convective part of this derivative, $i{\hat {L}}=-{{\mathcal {H}},\cdot }$, is a rather important entity sometimes referred to as the Liouvillian . It’s the generator of the flow in phase space , and its properties are central to understanding Liouville’s theorem (Hamiltonian) , which deals with the conservation of phase space volume. Essentially, the Liouvillian describes how the “density” of systems in phase space evolves over time.

Poisson matrix in canonical transformations

The utility of the Poisson bracket isn’t confined to individual functions; it can be elegantly extended to the realm of matrices, giving rise to the concept of the Poisson matrix . This is particularly useful when dealing with coordinate transformations, especially canonical transformations .

Imagine a transformation from one set of canonical coordinates $\eta$ to another set $\varepsilon$:

$$\eta ={\begin{bmatrix}q_{1}\\vdots \q_{N}\p_{1}\\vdots \p_{N}\\end{bmatrix}}\quad \rightarrow \quad \varepsilon ={\begin{bmatrix}Q_{1}\\vdots \Q_{N}\P_{1}\\vdots \P_{N}\\end{bmatrix}}$$

Here, $q_i$ and $p_i$ are the original generalized position and momentum coordinates, and $Q_i$ and $P_i$ are the new ones. Let’s define the Jacobian matrix $M$ for this transformation, which encapsulates how the new coordinates depend on the old ones: $M:={\frac {\partial (\mathbf {Q} ,\mathbf {P} )}{\partial (\mathbf {q} ,\mathbf {p} )}}$. With this, the Poisson matrix $\mathcal{P}(\varepsilon)$ is defined as:

$${\mathcal {P}}(\varepsilon )=MJM^{T}$$

where $J$ is the symplectic matrix . This matrix $J$ is a constant, anti-symmetric, invertible matrix that fundamentally defines the symplectic structure of the phase space . Its form depends on the convention used to order the coordinates (e.g., $(q_1, \dots, q_N, p_1, \dots, p_N)$).

From this definition, it follows directly that the elements of the Poisson matrix are, in fact, the Poisson brackets of the new coordinates with respect to the old ones:

$${\mathcal {P}}{ij}(\varepsilon )=[MJM^{T}]{ij}=\sum _{k=1}^{N}\left({\frac {\partial \varepsilon _{i}}{\partial \eta _{k}}}{\frac {\partial \varepsilon _{j}}{\partial \eta _{N+k}}}-{\frac {\partial \varepsilon _{i}}{\partial \eta _{N+k}}}{\frac {\partial \varepsilon _{j}}{\partial \eta _{k}}}\right)=\sum _{k=1}^{N}\left({\frac {\partial \varepsilon {i}}{\partial q{k}}}{\frac {\partial \varepsilon {j}}{\partial p{k}}}-{\frac {\partial \varepsilon {i}}{\partial p{k}}}{\frac {\partial \varepsilon {j}}{\partial q{k}}}\right)={\varepsilon _{i},\varepsilon {j}}{\eta }.$$

This identity is more than just a formal definition; it’s a statement that the Poisson matrix directly encodes the fundamental Poisson bracket relations between the new canonical coordinates .

The Poisson matrix possesses several well-known and crucial properties:

Antisymmetry: $${\mathcal {P}}^{T}=-{\mathcal {P}}$$ This property is directly inherited from the anticommutativity of the Poisson bracket . If you swap the indices $i$ and $j$, the sign flips.
Determinant property: $$|{\mathcal {P}}|={\frac {1}{|M|^{2}}}$$ This relates the determinant of the Poisson matrix to the determinant of the Jacobian of the transformation. It reflects how the transformation scales the “volume” in phase space , though for a true canonical transformation , this will simplify even further.
Inverse relation: $${\mathcal {P}}^{-1}(\varepsilon )=-(M^{-1})^{T}JM^{-1}=-{\mathcal {L}}(\varepsilon )$$ The inverse of the Poisson matrix is intimately related to the Lagrange matrix , denoted $\mathcal{L}(\varepsilon)$, whose elements correspond to the Lagrange brackets . The Lagrange bracket is, in a sense, the inverse operation to the Poisson bracket , providing an alternative, though less commonly used, formulation for canonical transformations.

The last identity can also be stated more explicitly as:

$$\sum _{k=1}^{2N}{\eta _{i},\eta _{k}}[\eta _{k},\eta _{j}]=-\delta _{ij}$$

Note that this summation involves all $2N$ generalized coordinates (both positions and momenta), highlighting the comprehensive nature of these relationships.

The very definition of a canonical transformation hinges on the invariance of the Poisson bracket structure. This means that the fundamental Poisson bracket relations between the coordinates must remain unchanged under the transformation. Mathematically, this is expressed as:

$${{\varepsilon _{i},\varepsilon {j}}{\eta }={\varepsilon {i},\varepsilon {j}}{\varepsilon }=J{ij}}$$

This implies that the Poisson bracket of any two new canonical coordinates $Q_k, P_l$ (or $Q_k, Q_l$, or $P_k, P_l$) must yield the same values as their original counterparts, which are simply the elements of the symplectic matrix $J$. This condition directly leads to the crucial symplectic condition :

$$MJM^{T}=J$$

This equation is the definitive test for a transformation to be canonical . It ensures that the transformation preserves the symplectic structure of the phase space , which is equivalent to preserving the fundamental Poisson bracket relations. It’s a rather elegant way of enforcing order in the universe of coordinate changes. [3]

Constants of motion

In the grand dance of dynamical systems , certain quantities remain stubbornly constant, defying the relentless march of time. These are known as constants of motion . For an integrable system —a rare and beautiful beast of predictability—these constants are not merely incidental; they are fundamental, providing a complete description of the system’s behavior. Beyond the energy (which is often the Hamiltonian itself), such systems possess additional constants of motion .

What makes a quantity a constant of motion in the Hamiltonian framework? It’s elegantly defined by its relationship with the Hamiltonian under the Poisson bracket . Specifically, a function $f(p,q)$ is a constant of motion if its Poisson bracket with the Hamiltonian vanishes: ${f, \mathcal{H}} = 0$.

Suppose we have a function $f(p,q)$ that is a constant of motion . This implies that if $p(t), q(t)$ describe a trajectory or solution to Hamilton’s equations of motion , then along that trajectory, the function $f$ does not change with time:

$$0={\frac {df}{dt}}$$

As derived earlier, the total time derivative of $f$, assuming it does not explicitly depend on time, is given by:

$${\frac {df}{dt}} = {f, \mathcal{H}} + {\frac {\partial f}{\partial t}}.$$

If $f$ is explicitly time-independent, then $\partial f / \partial t = 0$, and thus $df/dt = {f, \mathcal{H}}$. For $f$ to be a constant of motion , this total derivative must be zero, which means ${f, \mathcal{H}} = 0$. This equation is famously known as the Liouville equation . The deeper implication of Liouville’s theorem is that the time evolution of a measure (a “density” of systems) in phase space , represented by a distribution function $f$, is governed by this very equation, ensuring the conservation of phase space volume.

When the Poisson bracket of two functions, $f$ and $g$, vanishes (i.e., ${f,g}=0$), these functions are said to be in involution. This means they effectively “commute” in the Poisson bracket algebra, implying their dynamics are intertwined in a particularly harmonious way. For a Hamiltonian system to be deemed completely integrable —a state of perfect, predictable order—it must possess $n$ independent constants of motion that are all in mutual involution , where $n$ is the number of degrees of freedom. This set of $n$ mutually commuting constants of motion , along with the Hamiltonian , effectively simplifies the system to a solvable form, often allowing for the complete determination of its trajectories.

Furthermore, a powerful result known as Poisson’s Theorem states that if two quantities, $A$ and $B$, are explicitly time-independent constants of motion (meaning $A(p,q)$ and $B(p,q)$), then their Poisson bracket , ${A, B}$, is also a constant of motion . This elegant conclusion flows directly from the Jacobi identity , a testament to the consistency of the Poisson bracket structure. However, one must approach Poisson’s Theorem with a healthy dose of skepticism. It doesn’t always deliver a useful new constant of motion . The number of genuinely independent constants of motion for a system with $n$ degrees of freedom is limited to $2n-1$. Thus, the “new” constant derived from the bracket might turn out to be trivial—a mere numerical constant, or simply a function of the already known $A$ and $B$. The universe, it seems, rarely gives away secrets for free.

The Poisson bracket in coordinate-free language

For those who prefer to gaze upon the underlying geometry rather than getting bogged down in messy coordinates, the Poisson bracket can be defined in a much more abstract, and arguably more elegant, coordinate-free manner. Let $M$ be a symplectic manifold . What’s that, you ask? It’s a manifold (a space that locally looks like Euclidean space) endowed with a symplectic form . A symplectic form is a very special kind of 2-form , denoted $\omega$, which possesses two crucial properties: it is both closed (meaning its exterior derivative , $d\omega$, is zero) and non-degenerate.

To put it in context, consider the phase space we’ve been discussing: $\mathbb{R}^{2n}$. On this space, the standard symplectic form is given by:

$$\omega =\sum {i=1}^{n}dq{i}\wedge dp_{i}.$$

Here, $dq_i \wedge dp_i$ represents an exterior product of differential forms, which is a key component of the symplectic form . The non-degeneracy condition of $\omega$ is particularly important. If $\iota_v \omega$ is the interior product (or contraction ) operation, defined by $(\iota_v \omega)(u) = \omega(v,u)$, then non-degeneracy means that for every one-form $\alpha$, there exists a unique vector field $\Omega_\alpha$ such that $\iota_{\Omega_\alpha} \omega = \alpha$. In essence, the symplectic form provides a canonical way to map one-forms to vector fields . Alternatively, one can express this relationship using the inverse of the symplectic form , $\omega^{-1}$, such that $\Omega_{dH}=\omega^{-1}(dH)$.

Given a smooth function $H$ on $M$ (our Hamiltonian ), we can then define the Hamiltonian vector field $X_H$ as $\Omega_{dH}$. This vector field is the geometrical embodiment of the Hamiltonian and drives the dynamics of the system. In canonical coordinates , it’s easy to see how these fields manifest:

$${\begin{aligned}X_{p_{i}}&={\frac {\partial }{\partial q_{i}}}\X_{q_{i}}&=-{\frac {\partial }{\partial p_{i}}}.\end{aligned}}$$

With this machinery, the Poisson bracket , denoted ${\cdot, \cdot}$, on $(M, \omega)$ is defined as a bilinear operation on differentiable functions :

$${f,g};=;\omega (X_{f},,X_{g})$$

The Poisson bracket of two functions on $M$ is, itself, a function on $M$. This definition immediately highlights its antisymmetry :

$${f,g}=\omega (X_{f},X_{g})=-\omega (X_{g},X_{f})=-{g,f}.$$

Furthermore, we can connect the Poisson bracket to the action of vector fields as directional derivatives. Recall that $df(X_g)$ is the directional derivative of $f$ along $X_g$.

$${\begin{aligned}{f,g}&=\omega (X_{f},X_{g})=\omega (\Omega {df},X{g})\&=(\iota {\Omega {df}}\omega )(X{g})=df(X{g})\&=X_{g}f={\mathcal {L}}{X{g}}f.\end{aligned}}$$

Here, $X_g f$ denotes the application of the vector field $X_g$ to the function $f$ as a directional derivative, and $\mathcal{L}_{X_g}f$ is the entirely equivalent Lie derivative of the function $f$ along the vector field $X_g$. This establishes a profound link: the Poisson bracket of two functions is equivalent to taking the Lie derivative of one function along the Hamiltonian vector field generated by the other. It’s a statement about how one function “changes” under the “flow” induced by another.

If $\alpha$ is an arbitrary one-form on $M$, the vector field $\Omega_\alpha$ generates (at least locally) a flow $\phi_x(t)$ such that $\phi_x(0)=x$ and $\frac{d\phi_x}{dt}=\left.\Omega_{\alpha}\right|{\phi_x(t)}$. This flow $\phi_x(t)$ will consist of symplectomorphisms (that is, canonical transformations that preserve the symplectic form ) if and only if its Lie derivative with respect to $\omega$ vanishes: $\mathcal{L}{\Omega_\alpha}\omega = 0$. When this holds, $\Omega_\alpha$ is called a symplectic vector field .

Recalling Cartan’s identity , $\mathcal{L}_X\omega = d(\iota_X\omega) + \iota_X d\omega$, and knowing that $d\omega = 0$ for a symplectic form , it follows that:

$$\mathcal{L}{\Omega{\alpha}}\omega ;=;d\left(\iota _{\Omega _{\alpha}}\omega \right);=;d\alpha.$$

Therefore, $\Omega_\alpha$ is a symplectic vector field if and only if $\alpha$ is a closed form ($d\alpha=0$). Since $d(df) = d^2 f = 0$ for any smooth function $f$, it immediately follows that every Hamiltonian vector field $X_f$ is a symplectic vector field . Consequently, the flow generated by a Hamiltonian vector field consists entirely of canonical transformations .

From the relationship $df(\phi_x(t))/dt = X_{\mathcal{H}}f = {f, \mathcal{H}}$, we arrive at a fundamental result in Hamiltonian mechanics : the time evolution of any function defined on phase space is given by its Poisson bracket with the Hamiltonian . As noted before, if ${f, \mathcal{H}} = 0$, then $f$ is a constant of motion . Furthermore, this elegant formulation in canonical coordinates (where ${p_i, p_j} = {q_i, q_j} = 0$ and ${q_i, p_j} = \delta_{ij}$) allows Hamilton’s equations to be derived almost trivially. It’s a testament to the power of abstraction, revealing the underlying simplicity.

It also follows from the definition that the Poisson bracket is a derivation ; it satisfies a non-commutative version of Leibniz’s product rule :

$${fg,h}=f{g,h}+g{f,h},$$ and $${f,gh}=g{f,h}+h{f,g}.$$

The Poisson bracket is intimately connected to the Lie bracket of the Hamiltonian vector fields . Because the Lie derivative is a derivation , we have:

$$\mathcal{L}_{v}\iota _{u}\omega =\iota {{\mathcal {L}}{v}u}\omega +\iota {u}{\mathcal {L}}{v}\omega =\iota _{[v,u]}\omega +\iota {u}{\mathcal {L}}{v}\omega.$$

Thus, if $v$ and $u$ are symplectic vector fields (meaning $\mathcal{L}_v\omega = 0 = \mathcal{L}_u\omega$), using Cartan’s identity and the fact that $\iota_u\omega$ is a closed form , we find:

$$\iota {[v,u]}\omega ={\mathcal {L}}{v}\iota _{u}\omega =d(\iota _{v}\iota _{u}\omega )+\iota _{v}d(\iota _{u}\omega )=d(\iota _{v}\iota _{u}\omega )=d(\omega (u,v)).$$

This implies that $[v,u]=X_{\omega(u,v)}$, which directly leads to:

$$[X_{f},X_{g}]=X_{\omega (X_{g},X_{f})}=-X_{\omega (X_{f},X_{g})}=-X_{{f,g}}.$$

This is a profoundly important result: the Lie bracket of two Hamiltonian vector fields is itself a Hamiltonian vector field generated by the negative of the Poisson bracket of their respective generating functions. In other words, the Poisson bracket on functions directly corresponds (up to a sign) to the Lie bracket of their associated Hamiltonian vector fields . This also demonstrates that the Lie bracket of two symplectic vector fields is a Hamiltonian vector field , and thus also symplectic .

In the language of abstract algebra , the symplectic vector fields form a subalgebra of the Lie algebra of smooth vector fields on $M$, and the Hamiltonian vector fields form an ideal within this subalgebra . This isn’t just academic pedantry; it means that the symplectic vector fields are the Lie algebra of the (infinite-dimensional) Lie group of symplectomorphisms of $M$ – the transformations that preserve the fundamental symplectic structure .

It is often asserted, with a hint of casual dismissal, that the Jacobi identity for the Poisson bracket (i.e., ${f,{g,h}}+{g,{h,f}}+{h,{f,g}}=0$) simply “follows” from the corresponding identity for the Lie bracket of vector fields . While this is largely true, one must be precise: it holds only up to a locally constant function, which, in physics, is usually trivial. However, a more direct proof of the Jacobi identity for the Poisson bracket can be made by demonstrating that:

$$\operatorname {ad} _{{g,f}}=\operatorname {ad} _{-{f,g}}=[\operatorname {ad} _{f},\operatorname {ad} _{g}]$$

where the operator $\operatorname{ad}_g$ on smooth functions on $M$ is defined as $\operatorname{ad}_g(\cdot) = {\cdot, g}$, and the bracket on the right-hand side is the commutator of operators, $[\operatorname{A}, \operatorname{B}] = \operatorname{A}\operatorname{B} - \operatorname{B}\operatorname{A}$. By our earlier result, the operator $\operatorname{ad}_g$ is equivalent to the vector field $X_g$. The proof of the Jacobi identity then follows directly from the fact that the Lie bracket of vector fields is, up to a factor of -1, precisely their commutator as differential operators.

The algebra of smooth functions on $M$, when equipped with the Poisson bracket , forms a Poisson algebra . This is because it satisfies the axioms of a Lie algebra under the Poisson bracket and additionally adheres to Leibniz’s rule . We have thus shown that every symplectic manifold is inherently a Poisson manifold —a manifold graced with a “curly-bracket” operator on smooth functions that forms a Poisson algebra . However, the converse is not always true. Not every Poisson manifold can be derived from a symplectic manifold , primarily because Poisson manifolds allow for “degeneracy” in their bracket structure, a feature that is strictly forbidden in the non-degenerate world of symplectic forms .

A result on conjugate momenta

Let’s delve into a rather neat connection between vector fields on the configuration space and their corresponding conjugate momenta in phase space . Given a smooth vector field $X$ on the configuration space, let $P_X$ be its conjugate momentum . The mapping from vector fields to their conjugate momenta has a specific algebraic property: it acts as a Lie algebra anti-homomorphism from the Lie bracket of vector fields to the Poisson bracket of the corresponding conjugate momenta . In plainer terms, it flips the sign:

$${P_{X},P_{Y}}=-P_{[X,Y]}.$$

This is an important result, deserving of a brief moment of scrutiny. Let’s write a vector field $X$ at a point $q$ in the configuration space in terms of its components in a local coordinate frame:

$$X_{q}=\sum _{i}X^{i}(q){\frac {\partial }{\partial q^{i}}}$$

where ${\frac {\partial }{\partial q^{i}}}$ denotes the basis vector fields associated with the local coordinates. The conjugate momentum $P_X$ associated with this vector field $X$ is then given by:

$$P_{X}(q,p)=\sum {i}X^{i}(q);p{i}$$

Here, $p_i$ are the generalized momentum functions canonically conjugate to the coordinates $q_i$. This expression essentially tells you how much “momentum” the system has along the direction of the vector field $X$.

Now, let’s compute the Poisson bracket of two such conjugate momenta , $P_X$ and $P_Y$, at a point $(q,p)$ in the phase space :

$${\begin{aligned}{P_{X},P_{Y}}(q,p)&=\sum {i}\sum {j}\left{X^{i}(q);p{i},Y^{j}(q);p{j}\right}\&=\sum {ij}p{i}Y^{j}(q){\frac {\partial X^{i}}{\partial q^{j}}}-p_{j}X^{i}(q){\frac {\partial Y^{j}}{\partial q^{i}}}\&=-\sum {i}p{i};[X,Y]^{i}(q)\&=-P_{[X,Y]}(q,p).\end{aligned}}$$

The second line uses the Leibniz rule and the fundamental Poisson bracket relations for canonical coordinates (e.g., ${q_k, p_l} = \delta_{kl}$, and others being zero). The third line is a direct recognition of the component form of the Lie bracket $[X,Y]$ for vector fields on the configuration space. This identity holds true for all points $(q,p)$ in phase space , thus yielding the desired result. This anti-homomorphism is a profound structural statement, linking the algebraic properties of vector fields on the configuration space to those of functions on the phase space via the Poisson bracket .

Quantization

And now, for the part where the classical world decides it’s had enough and transforms into something far stranger. The Poisson brackets we’ve been meticulously dissecting don’t simply vanish when we transition to the quantum realm; they undergo a fascinating transformation. Upon quantization , these brackets “deform” into Moyal brackets . This isn’t just a change of notation; it’s a generalization to an entirely different Lie algebra , known as the Moyal algebra . In the more familiar language of Hilbert space and quantum operators, this corresponds directly to quantum commutators .

The relationship between the classical Poisson bracket and the quantum commutator (or Moyal bracket ) is often expressed through the correspondence principle, which states that classical physics should emerge as a limit of quantum physics. Specifically, the Wigner-İnönü group contraction of these quantum algebras, taken in the classical limit where Planck’s constant $\hbar \to 0$, yields precisely the classical Lie algebra defined by the Poisson bracket . It’s as if the quantum world, with its inherent fuzziness and uncertainty, collapses into the sharp, predictable lines of classical mechanics when viewed through a sufficiently blurry lens.

To be a bit more explicit, and perhaps to add a layer of complexity you didn’t ask for, the universal enveloping algebra of the Heisenberg algebra is essentially the Weyl algebra , modulo the stipulation that its center must be the unit element. The Moyal product (which defines the Moyal bracket ) then emerges as a special instance of the “star product” on the algebra of symbols. If you’re truly curious about the precise mathematical architecture of this transition, the article on the universal enveloping algebra provides a detailed construction of the algebra of symbols and the star product. It’s a journey from the smooth, continuous world of classical phase space functions to the non-commutative, operator-based reality of quantum mechanics, where the very act of measurement changes the system. The Poisson bracket , in its classical elegance, is but a shadow of the deeper, more intricate algebraic structures that govern the universe at its most fundamental level.

Remarks

^ $f(p_i, q_i, t)$ means $f$ is a function of the $2N+1$ independent variables: momentum, $p_{1\dots N}$; position, $q_{1\dots N}$; and time, $t$. It’s a common shorthand, but apparently, some need it spelled out.

Poisson bracket