Module 2: Classical mechanics · Lesson 2.1 · about 50 min
Lagrangian mechanics
Generalised coordinates, constraints and degrees of freedom, d'Alembert's principle, the Euler-Lagrange equations, cyclic coordinates and conserved momenta, Noether's theorem, velocity-dependent potentials, Lagrange multipliers and Hamilton's principle, with the shortcuts GATE and NET reward.
Module 2 · about 50 minutes. By the end you will be able to count degrees of freedom, choose generalised coordinates, write a Lagrangian and get the equations of motion from it, spot conserved quantities without solving anything, handle a charge in a magnetic field, and find a constraint force with a Lagrange multiplier.
Newton’s laws in vectors are covered in the JAM lesson on Newton’s laws and frames. Here the same physics is written in a form that works in any coordinates and makes conservation laws easy to see.
1. Constraints and generalised coordinates
A system of particles needs coordinates. A holonomic constraint is an equation that removes one of them: a rigid rod fixes a distance, a bead on a wire must sit on a curve. With such constraints there are degrees of freedom, and any independent quantities that fix the configuration are generalised coordinates: angles, distances, or anything else.
Constraints that cannot be written that way are non-holonomic, for example (a particle outside a sphere) or rolling without slipping on a plane, where the condition ties velocities together and cannot be integrated. A constraint that depends on time (a bead on a wire that is being turned) is rheonomic; one that does not is scleronomic.
Worked example. A rigid body is particles with fixed distances between every pair. Fix three points that are not in a line: that is 9 coordinates and 3 distances, so 6 degrees of freedom, and every other particle is then fixed. Three of them say where the centre is and three (for example the Euler angles) say how the body is turned.
2. From d’Alembert to the Euler-Lagrange equations
Ideal constraint forces (a smooth wire, a rigid rod) do no work in a virtual displacement, an instant small change allowed by the constraints. Newton’s second law then gives d’Alembert’s principle, , with only the applied forces. Rewriting it in the :
Change variables. ri = ri(q, t), so δri = Σj (∂ri/∂qj) δqj, with no δt term (virtual means at one instant).
Generalised force. The applied part becomes Σj Qj δqj, with Qj = Σi Fi·∂ri/∂qj. For a conservative force, Qj = −∂V/∂qj.
Two identities. ∂ṙi/∂q̇j = ∂ri/∂qj and d/dt(∂ri/∂qj) = ∂ṙi/∂qj turn the inertia term into d/dt(∂T/∂q̇j) − ∂T/∂qj.
Independence. The δqj are independent, so each coefficient vanishes separately. With V independent of q̇, put L = T − V.
- qj, q̇j
- a generalised coordinate and its generalised velocity, treated as independent variables in L
- pj = ∂L/∂q̇j
- the canonical (conjugate) momentum; it is mẋ for a Cartesian coordinate and an angular momentum for an angle
- ∂L/∂qj
- the generalised force, including the "fictitious" centrifugal terms that come from T
- n equations
- one second-order equation for each degree of freedom, with no constraint forces in sight
- Physical meaning
- Each equation says "rate of change of the momentum belonging to qj = the force belonging to qj", which is Newton's second law in whatever coordinates you chose.
Worked example: the Atwood machine with a heavy pulley. Masses hang over a pulley of moment of inertia and radius . One coordinate, the drop of , fixes everything: the pulley turns through . So and . The Euler-Lagrange equation gives
with no string tensions anywhere. A pulley made of a uniform disc has .
3. A spinning hoop: equilibria that move
A bead slides on a smooth vertical hoop of radius that is forced to spin at a fixed rate about its vertical diameter. Measure from the bottom. The bead moves along the hoop at and around the axis at , so
Equilibria are where : the bottom, the top, and, only if , the angle with . Spin the hoop past and the bottom stops being stable: the bead climbs to . This is a bifurcation, the same mathematics as a magnet ordering below its Curie point.
Here the time-dependent constraint does work on the bead, so is not conserved. But has no explicit , so is. Section 5 explains why.
4. Cyclic coordinates and Noether’s theorem
If a coordinate is missing from (it is cyclic, or ignorable), the Euler-Lagrange equation reads : its conjugate momentum is conserved. This is the simplest case of Noether’s theorem: every continuous symmetry of the Lagrangian gives a conserved quantity.
| Symmetry of L | Conserved quantity |
|---|---|
| shift every position by the same vector (homogeneity of space) | total linear momentum |
| turn the system about an axis (isotropy of space) | angular momentum about that axis |
| shift the time origin (L has no explicit t) | the energy function h |
Worked example. A particle moves on the surface of a cone under gravity. With and as coordinates, . is cyclic, so is conserved: the vertical component of angular momentum, because the cone is symmetric about its axis. The -momentum is not conserved, since gravity and the cone break that symmetry.
A symmetry tells you a conserved quantity before you solve anything. For a charge near an infinite straight wire, the system is unchanged by moving along the wire and turning around it, so and are conserved even though the motion is complicated.
5. The energy function and when it is the energy
Define the energy function
So is conserved whenever has no explicit time dependence. It equals the total energy only when is a quadratic form in the velocities () and does not depend on velocities, which is true when the transformation has no explicit . Euler’s theorem on homogeneous functions then gives , so .
On the spinning hoop, the kinetic energy has a velocity-free part , so : is conserved, is not. Two separate questions are worth keeping apart: is h conserved? (look for explicit in ) and is h the energy? (look for explicit in the coordinates).
6. Velocity-dependent potentials: a charge in a field
The Lorentz force depends on velocity, yet it fits the Lagrangian scheme with a generalised potential :
The Euler-Lagrange equations give with and . The canonical momentum is not ; it is the one that becomes in quantum mechanics. A gauge change , adds to , a total derivative, so the motion is unchanged.
Worked example. A uniform field can be written with . Then . Both and are cyclic, so and are conserved. The first says : the orbit centre sits at , the canonical momentum labels the guiding centre. With the symmetric gauge the cyclic coordinate is instead; different gauges show different conserved quantities.
“The canonical momentum is always mass times velocity.” Only for a Cartesian coordinate with a velocity-free potential. For an angle it is an angular momentum, for a charge in a magnetic field it is (which depends on the gauge), and on the spinning hoop . Always compute .
7. Constraint forces with Lagrange multipliers
Choosing coordinates that satisfy the constraints hides the constraint forces. To find one, keep an extra coordinate and impose the constraint with a multiplier :
The right side is the generalised constraint force. The method also handles non-holonomic constraints written in differential form, , by replacing with .
Worked example: when does a particle leave a sphere? A particle slides from rest at the top of a smooth sphere of radius . Keep with the constraint , and measure from the top, so the height is and . The equation at is . Energy gives , so the normal force is . It reaches zero at , about 48.2° from the top, where the particle flies off.
8. Hamilton’s principle
The Euler-Lagrange equations are exactly the condition that the action is stationary for paths with fixed end points:
This is the cleanest statement of classical mechanics: it needs no forces or coordinates, it shows at once that adding to changes nothing (it only adds a constant to ), and it carries straight into field theory and, via Feynman’s path integral, quantum mechanics. The same calculus of variations solves geometric problems: the shortest path on a surface, the soap film between two rings (a catenoid), and the brachistochrone, the curve of fastest descent, which is a cycloid.
Worked example. A Lagrangian gives SHM. Multiply it by any constant, or add , and the equation of motion is unchanged. A test question that offers two Lagrangians differing by such a term expects you to call them equivalent.
- Degrees of freedom for holonomic constraints; rolling without slipping on a plane is non-holonomic.
- Euler-Lagrange: , ; ideal constraint forces drop out.
- Canonical momentum ; a cyclic coordinate means a conserved (Noether: symmetry gives a conservation law).
- is conserved when ; it equals only when the coordinates have no explicit time dependence.
- A charge: , ; gauge changes add a total derivative.
- Constraint forces: . A particle leaves a smooth sphere at .
- Hamilton’s principle ; and give the same motion.
Practise: exam questions on this lesson
They start easy and climb to top-rank level. More on the Lagrangian mechanics practice page.
Basics: the JAM lessons on Newton’s laws and frames and potential energy. Next: lesson 2.2, Hamiltonian mechanics.
Check yourself
Sources
- H. Goldstein, C. Poole and J. Safko, Classical Mechanics, 3rd ed. (2002), chapters 1 and 2
- L. D. Landau and E. M. Lifshitz, Mechanics, 3rd ed. (1976), chapters 1 and 2
- J. R. Taylor, Classical Mechanics (2005), chapters 6 and 7
- GATE Physics syllabus (Classical Mechanics), IIT organising institute; CSIR-UGC NET Physical Sciences syllabus, Part A Classical Mechanics
Your ticks stay in this browser. Save them as a file on the course home.
Updated 2026-11-22. Not affiliated with the IITs, IISc or the GATE organising institute. The general aptitude part of GATE is not covered here. Check the official GATE website for the current pattern. All questions are original.