A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Lagrangian mechanics reformulates Newton's laws using L = T − V (kinetic minus potential energy) and the principle of stationary action. The Euler–Lagrange equation d/dt(∂L/∂ẋ) − ∂L/∂x = 0 yields equations of motion without explicitly calculating forces or accelerations. This powerful approach naturally incorporates constraints via generalized coordinates and reveals symmetries that lead to conservation laws.
Your work on Newton's laws approached mechanics through forces: identify all forces acting on a body, sum them to get the net force, and integrate F = ma to get the trajectory. This works beautifully for a single particle in a simple geometry, but it becomes unwieldy fast. A pendulum requires resolving tension along a curved path. A bead constrained to a wire requires computing a constraint force that does no work. A double pendulum requires coupling multiple force equations. Lagrangian mechanics reformulates the same physics using energy rather than force, and the result is a method that handles constraints and complex geometries almost automatically.
The central object is the Lagrangian L = T − V: kinetic energy minus potential energy. From your work on conservation of energy, you know that total mechanical energy E = T + V is conserved in many systems. The Lagrangian is the difference, not the sum — a quantity whose integral over time, called the action S = ∫L dt, measures something like the "cost" of a trajectory. The principle of stationary action (Hamilton's principle) states that nature takes the path for which the action is stationary — not necessarily minimized, but neither increased nor decreased by small variations. Among all the possible paths a system could take between two configurations, the one that actually occurs makes the action stationary. This principle is deep: it reframes physics as a global optimization over trajectories, not a local rule about forces.
Applying calculus of variations to make the action stationary (which is where your work on partial derivatives becomes essential) yields the Euler–Lagrange equation: d/dt(∂L/∂q̇) − ∂L/∂q = 0 for each generalized coordinate q. The beauty is in what q can be. Instead of being tied to Cartesian coordinates, you can choose any coordinates that naturally describe your system — the angle of a pendulum, the distance along a constrained track, the joint angles of a robot arm. Constraints are handled by simply choosing coordinates that satisfy them from the start, eliminating constraint forces from the problem entirely. For a pendulum, one generalized coordinate (the angle θ) fully describes the system; the Euler–Lagrange equation in θ directly yields the equation of motion without ever mentioning tension.
The deepest result in the Lagrangian framework is Noether's theorem: every continuous symmetry of the Lagrangian corresponds to a conserved quantity. If L does not change when you translate the system in time (time-translation symmetry), energy is conserved. If L does not change when you translate in space, momentum is conserved. If L does not change under rotation, angular momentum is conserved. Conservation laws are not separate empirical facts to be discovered one by one — they are consequences of symmetry, and the Lagrangian is the object that makes the symmetry manifest. This is why Lagrangian mechanics is not just a computational convenience but a conceptual reorganization of physics: it reveals the structural reasons why conserved quantities exist.