Constrained Systems and Dirac Brackets

Constrained Systems and Dirac Brackets in advanced mechanics. Theory, formulas, worked examples and applications.

Conceptual context

What happens when a system's Hamiltonian cannot be built the usual way because some momenta cannot be inverted in terms of the velocities? This happens, for instance, in Maxwell electrodynamics in Hamiltonian form, or in any theory with gauge symmetry: the standard Legendre transform fails because relations appear — primary constraints — between the coordinates and momenta themselves, instead of equations solving for the velocities. Paul Dirac developed, in the late 1950s, the systematic procedure to handle these cases without breaking the Hamiltonian structure.

The central idea is that the Hamiltonian actually generating the system's evolution is not the naively obtained canonical one ($H_c$), but a total Hamiltonian ($H_T$) that adds the contribution of each constraint multiplied by an undetermined factor. This leaf's quantities, $H_T$, $H_c$, $\lambda^\alpha\phi_\alpha$, $\phi_\alpha$, $\lambda_\alpha$, $C_{\alpha\beta}$, $F$, and $\{F,G\}_D$, trace exactly that chain: from the canonical Hamiltonian to the total one, from individual constraints to their bracket structure, and from there to the corrected bracket that does respect the constraints during time evolution.

🟢 Essential level

The essential idea, without algebra, is this: when a system has restrictions that were already present even before solving the equations of motion (primary constraints), it is not enough to compute the Hamiltonian the usual way and evolve with it. A term must be added by hand for each constraint, multiplied by a number the formalism itself determines (or leaves free, depending on the case) by requiring that the constraints remain valid at every instant.

That consistency requirement is the heart of the method. If a constraint does not automatically stay satisfied as the system evolves, the formalism forces new information to appear: either it fixes the value of the corresponding multiplier, or it reveals an additional constraint that was not previously evident. Distinguishing these two possibilities — second-class constraints, with a fixed multiplier, and first-class ones, with a free multiplier — is the underlying question a student must be able to answer before operating with the formulas.

First-class constraints are not a defect of the formalism: they reflect gauge freedom, the same freedom that in electromagnetism allows choosing different potentials for the same physical field. Second-class ones, in contrast, genuinely reduce the number of independent degrees of freedom of the system.

🔵 Formal level

Basic formulas

This first formal layer presents the reference relation used to recognize the model and read its variables before entering more abstract derivations. Identify what each term represents, which assumptions make it valid, and how the chosen coordinates connect to the physical motion that the model is intended to describe.

The formal level rests on four relations:

Dirac's total Hamiltonian

Real student questions

"Why isn't it enough to impose the constraint at the end, substituting its zero value after solving the equations?" Because the multiplier associated with each second-class constraint affects the dynamics of the other variables during the evolution, not just at the end; ignoring it during the calculation produces a trajectory that does not respect the constraints at any intermediate instant.

"How do I know if a constraint is first or second class without computing the whole $C_{\alpha\beta}$ matrix?" There is no general shortcut: the Poisson brackets between all identified primary constraints must be explicitly computed and it must be checked whether the resulting matrix is invertible.

Final synthesis

By the end of this leaf, the student should be able to explain why the canonical Hamiltonian is not enough when primary constraints are present, build the total Hamiltonian by adding the constraint contribution, and distinguish first-class constraints (free multiplier, gauge freedom) from second-class ones (determined multiplier, invertible constraint matrix, Dirac bracket needed).

The relevant final check is one of consistency, not experimental in the usual sense: verifying that the Dirac bracket of any constraint with any observable vanishes identically is the proof that the formalism has been applied correctly.